Auerbach_Computer_Notebook_International_1969 Auerbach Computer Notebook International 1969

Auerbach_Computer_Notebook_International_1969 Auerbach_Computer_Notebook_International_1969

User Manual: Auerbach_Computer_Notebook_International_1969

Open the PDF directly: View PDF PDF.
Page Count: 1280

DownloadAuerbach_Computer_Notebook_International_1969 Auerbach Computer Notebook International 1969
Open PDF In BrowserView PDF
AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

Prepared and Published by
AUERBACH INFO, Inc.
121 North aroad Street
Philadelphia, Pa. 19107
Phone
215·491·8200

AUERBACH
(!)

AUERBACH INFO, INC.
AUERBACH INFO, INC. publishes periodically updated looseleaf reference works for current
awareness in the field of information processing, data communications, and graphics.
• AUERBACH Standard EDP Reports

An eight-volume analytical service providing detailed, objective reports on the major U.S. computer systems. Hardware and software are analyzed in a standardized report format that facilitates comparisons.
Benchmark problems are used to measure overall system performance in typical commercial and scientific
applications. Updated twelve times per year.
• AUERBACH Scientific and Control Computer Reports

A two-volume extension of AUERBACH Standard EDP Reports containing detailed, objective analyses of
the U.S. computer systems that are specialized, by hardware design or software support, for scientific,
control, and other nonbusiness-oriented applications. Updated six times per year.
• AUERBACH Computer Notebook
A two-volume current awareness service on more than 80 U.S. computer systems, updated twelve times
a year with new reports and revisions. Contains individual descriptions of the features and limitations of
each system, over 100 pages of objective hardware and performance comparison charts, and complete
price lists.
• AUERBACH Computer Notebook International

A two-volume guide which has the entire contents of AUERBACH Computer Notebook plus coverage expanded to include computer systems manufactured outside the United States. It also includes Special
Reports which provide facts and guidelines on topics of current interest. Updated twelve times a year.
• AUERBACH Software Notebook
A detailed guide to the manufacturer-supplied software for third-generation U.S. computer systems. This
service contains individual, analytical reports designed to supply the facts you need to understand, evaluate, and utilize computer software. Updated six times per year.
• AUERBACH Data Communications Reports

A complete reference source on digital data communications equipment and techniques. Contains individual, analytical reports on more than 50 different types of communications terminals and processing
equipment, detailed reports on common-carrier facilities, and a guide to the design of effective data communications systems. Updated four times per year.
• AUERBACH Data Handling Reports
A comprehensive guide to selecting and applying the wide range of support equipment and supplies used
in conjunction with computer systems. Detailed, analytical reports, in a standardized format for easy com·
parisons, describe equipment used for capturing data at its source, preparing input to computers, performing media conversions, processing unit records, and handling printed forms. Updated four times
per year.
• AUERBACH Graphic Processing Reports

A comprehensive analytical service on information storage and retrieval systems, microform reader/
printers, plotters, computer-driven displays, and photo typesetters. Includes system design tools, operational procedures, performance, support requirements and prices. Updated four times per year.
• AUERBACH Time-Sharing Reports

A two-volume service covering all aspects of commercial time-sharing. It includes reports on the state of
the time-sharing art, time-sharing languages, applications, equipment and individual reports on commercial time-sharing services. Updated four times per year .

•••••••••

AUERBACH Computer Notebook International
THE INFORMATION CONTAINED HEREIN HAS BEEN OBTAINED FROM RELIABLE SOURCES AND HAS BEEN EVALUATED BY EXPERIENCED
TECHNICAL PERSONNEL. DUE TO THE RAPIDLY CHANGING NATURE OFTHE TECHNOLOGY AND EQUIPMENT, HOWEVER, THE INFORMATION
CANNOT BE GUARANTEED.

,,"ted

In

U.S.A.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

A

1:001. oot

AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

CONTENTS

AUERBACH

"
CONTENTS
Page

BINDER 1
T ABLE OF MONETARY CONVERSIONS . . . . . . . . . . . . . . .

1:002.001

USERS' GUIDE . . . . . . .

4:001. 100

GLOSSARY . . . . . . . . . .

7:001. 001

COMPARISON CHARTS11:001. 002
Quick Reference Index to All Charts . . . ...
U. S. A. Computers Configuration Rentals . . . . . . . . . . . . . .
11:010. 101
Hardware Characteristics Central Processors and Working Storage.
. . . . . . . . . . . . 11:210.101
Auxiliary Storage and Magnetic Tape. . . .
. . . . . . . . . . . . . . . . .. . 11:220. 101
Punched Card and Punched Tape Input-Output . . . . . . . . . . . . . . . . . . . . . . . 11:230.101
Printers and Specialized Input-Output Equipment . . . . . . . . . . . . . . . . . . . . . 11:240.100
Software. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . . . . . . . . . . . . . . .. 11:300.100
System Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11:400.101
Non- U. S. A. Computers Hardware Characteristics Central Processors and Working Storage . . . . . . . . . . . . . . . . . . . . . . . . .
Auxiliary Storage and Magnetic Tape . . . . . . . . . . . . . . . . . . . . . . . . . . .
Punched Card and Punched Tape Input-Output . . . . . . . . . . . . . . . . . . . . . .
Printers and Specialized Input-Output Equipment . . . . . . . . . . . . . . . . . . .

.
.
.
.

11:510. 101
11:520.101
11:530.101
11:540.101

Computer Contracts - A Survey and Analysis . . . . . . . . . . . . . . . . . . . . . . . .
A Survey of the Character Recognition Field . . . . . . . . . . . . . . . . . . . . . . . . . .
Decision Tables: A State-of-the-Art Report . . . . . . . . . . . . . . . . . . . . . .
Magnetic Tape Recording: A State-of-the-Art Report " . . . . . . . . . . . .
High-Speed Printers: A State-of-the-Art Report . . . . . . . . . . . . . . . . . . . .
Random Access Storage: A State-of-the-Art Report . . . . . . . . . . . . . . . . . .
Digital Plotters: A State-of-the-Art Report . . . . . . . . . . . . . . . . . . . . . .
Data Collection Systems: A State-of-the-Art Report . . . . . . . . . . . . . . . . .
The Selection and Use of a Data Processing Service Center . . . . . . . . . . . . .
Data Communications -What It's All About . . . . . . . . . . . . . . . . . . . . . .
Source Data Automation Techniques and Equipment . . . . . . . . . . . . . . . . . . .. .
Design and Applications of Automated Display Systems . . . . . . . . . . . . . . . . . .
Keyboard to Magnetic Tape Encoders . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . .

23:010.001
23:020.001
23:030.001
23:040.001
23:050.001
23:060.001
23:070.001
23:080.001
23:090.001
23:100.001
23:110.001
23:120.001
23:130.001

SPECIAL REPORTS -

COMPUTER SUMMARIES AND PRICE LISTS U.S.A.
Burroughs
Burroughs
Burroughs
Burroughs

B
B
B
B

100/200/300 Series
............................
5500 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6500 and B 7500 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2500 and B 3500. . . . . . . .
. ........................

201:011.100
203:011. 100
204:011. 100
210:011. 100

Burroughs Series E Computers . . . . . .
. ........................
Burroughs E2000/3000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Burroughs E4000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Burroughs E6000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

220:011.
222:011.
223:011.
224:011.

CDC
CDC
CDC
CDC

241:011.100
242:011. 100
243:011. 100
244:011.100

1604 (Control Data Corporation) . . . . . . . . . . . . . . . . . . . . . . . . . . . ' . . . .
160 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1604-A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
160-A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

100
100
100
100

11/69

1:001. 002

AUERBACH COMPUTER NOTEBOOK IN1'ERNATIONAL

BINDER 1 (Contd. )
COMPUTER SUMMARIES AND PRICE LISTS (Contd. ) U. S. A. (Contd.)
CDC 3400/3600/3800. . . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
CDC 3400 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
CDC 3600 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
CDC 3800 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ' .. .

245:011.100
246:011. 100
247:011.100
248:011. 100

CDC 3100/3300/3500 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
CDC 6000 Series . . . . . . . . . . . . . . . . . . . . : . . . . '. . . . . . . . . . . . . . . . . .
CDC 7600 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

250:011. 100
260:011. 100
270:011. 100

GE-105
GE-115
GE 130
GE-200
GE-400
GE-600

309:011. 100
310:011. 010
311:011. 100
320:011. 100
330:011. 100
340:011.100

(General Electric) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
................................................ .
.....•..•........................................
Series . . . . . . . . • . • . . . . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Series . . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
BINDER 2
~

11/69

(Contd.)

IBM 1401 (International Business Machines Corp.) .. . • . . . . . . . . . . . . . . ..
IBM 1410 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . ..

401 :011. 100
402:011. 100

IBM 7070
IBM 7072
IBM 7074

403:011. 100
404:011. 100
405:011. 100

IBM 7090 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . . . . . . .
IBM 7094 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
IBM 7040/7044 . . . • . • . . . . . . . . . . • . . . • . . . . • . . • • . . . • . • . . . • . . • • .

408:011. 100
409:011. 100
410: 011.100

IBM
IBM
IBM
IBM

412:011.
413:011.
414:011.
415:011.

1620
1620
1440
1460

ModelL . . . . . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . . . . . .
Model 2 . . . . . . • . . . . . . . . . • . . . . . . . . . . . • . . . . . • . • . . . . . .
..................•..••......•.........•........
•....•.........•..•.............•....•..........

100
100
100
100

416:011. 100
417:011.100
418:011. 100

IBM 7010 • . . . . . . • . . . . . . . . . • . . . . . . . • . . • . . . . . . . . • . . . . . . . . . .
IBM 7080 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . .
IBM 1130 . . . . . . . • . . . . . . • . . . • . . . . . . . . . . • . . . . • . . . . . . . . . . . . .
IBM System/360 Models 30, 40, 50, 65, 75 . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . .
Model 20 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . . .
Model 67 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Model 85 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Model 25 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . . . . . .
Model 44 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
IBM System/3 . . . . . . . • . . • . . • . • . . • • • • . • . • • • . • . . . . • . . • . . • • . • .

420:011. 100
422:011. 100
427:011. 100
430:011. 100
432:011. 100
435:011.100
450: 011.100

Honeywell
Honeywell
Honeywell
Honeywell
Honeywell

400 (Honeywell EDP Division) . • • . . . . . . . . . . . . . . . . . . . . . . . .
800 . . . . . . . . . . . . . . . . . . . • . • . . . . . . . . . • • . . . . . . . . . . . .
1800 . . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . .
1400 . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Series 200 . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . .

501:011.100
502:011. 100
503:011. 100
505:011. 100
510:011. 100

Monrobot XI (Litton Industries, Inc.) . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

531. OIL 100

NCR 315 (National Cash Register Company) . . • . . . . . . . . . . . . . . . . . . . . . .
NCR 315-100 • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
NCR 315-RMC
.................•......................•..

601:011. 100
602:011. 100
603:011.100

NCR Small Computers . . . . . . . . . . . . . • . . . . . . . . . • • . . . . . . . . . . . . . .
NCR 395 . . . . . . . . . . . . . . . . . . . . . . . . . • . . . . • . . . . . . . . . . . . . . . .
NCR 400 . . . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
NCR 500 . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . . . . .
NCR Century Series . . . . . . . .

610:011.100
612:011. 100
613:011. 100
614:011.100
620:011. 100

A

AUERBACH
®

(Contd. )

1:001.003

CONTENTS

BINDER 2 (Contd.)
U. S. A. (Contd.)
RCA 301 (Radio Corporation of America) . . . . . . • . . . • . . . . . . . . . . . . . . . .
RCA 3301 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . . .
RCA Spectra 70 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Spectra
Spectra
Spectra
Spectra
Spectra
Spectra
Spectra

70/15
70/25
70/35
70/45
70/55
70/46
70/60

..............•.........•......•..........•.
.......••.....••.•....•.•.••..•......•......
.......•.........•...•..•.•..............•..
......••............................•.......
.......•......••.•.....••......•............
........................•.....••..•.......•.
...............•........••.•.•........•.....

701:011.100
703:011.100
'/10:011.100
712:011.100
713:011.100
714: 011.100
715:011.100
716: 011.100
717: OIl. 100
718: 011. 100

SDS Sigma 7 (Scientific Data Systems)

740:001. 010

UNIVAC
UNIVAC
UNIVAC
UNIVAC

1004 (Sperry Rand Corp.) . . . . . . . . . . . . • . . . • • . • . . . . . . . . . . .
SS 80/90 Model I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
SS 80/90 Model II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
ill . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

770:011.100
771:011. 100
772:011.100
774:011.100

UNIVAC 1050
............................................ .
UNIVAC 1107
...................................•.........
UNIVAC 1108
............................................ .
UNIVAC 418 Series
........................................ .
UNIVAC 418-1/11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
UNIVAC 418-I11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . . .

777:011.100
784:011. 100
785:011. 100
790:011.100
791:011.100
792:011.100

UNIVAC 490 Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
UNIVAC 9000 Series (9200/9300/9400) . . . . . . . . . . . . . . . . . . . . . . . . . • . . ..

800:011. 100
810:011.100

Non-U. S. A. Computers
DENMARK
RC 4000 (A/S Regnecentra1en) . . . . . . . . . . . . . . . . . . . . • . . . . . . • . . • • • . . . . . 1300: 011.100
FRANCE
Bull GE Gamma 10 (Compagnie Bull General Electric)
Bull GE 55

1440:011.100
1445:011.100

ISRAEL
1490:011. 100

Elbit 100 (Elbit Computers Ltd.)
JAPAN
Fujitsu F ACOM 270 Series (Fujitsu, Ltd.) . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
Fujitsu FACOM 230 Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . .
Hitachi Hitac 3010 (Hitachi, Ltd.) . . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . . . .
Hitachi Hitac 8000
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
Nippon Electric NEAC-Series 2200 (Nippon Electric Company) . . . . . . . . . . . . . • .
Nippon Electric NEAC-Series 2200/50 . . . . . . . . . . . . . . • . . . . • . . . . . • • . . . .

1540: OIl. 100
1541:011.100
1555:011. 100
1557:011.100
1575:011. 100
1576:011.100

THE NETHERLANDS
Philips P1000 Series (NV Philips-Electrologica)

1620:011.100

UNITED KINGDOM
ICL System 4 (International Computers Ltd.) . . . . . . . . . . . . . . . . . . . . . . . . . ..
ICL 1900 Series
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . . ..

1850:011. 100
1855:011. 100

WEST GERMANY
Siemens System 4004 (Siemens AG) . . . . . . . • . . . . . . . . • . • . . • • . • . . • . . . . . .: 1950: 011. 100
Siemens System 300 . • . • . . . . . . . • . . . . . . . . . . . . . . • . . . • . . . • . . . • • • . .. 1955: OIl. 100

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

11/69

fA
a

AUERBACH

1 :002.001

COMPUTER

NOTEBOOK
INTERNATIONAL

MONETARY CONVERSION TABLE

AUERBACH

'"
MONETARY CONVERSION TABLE

Country

Currency Unit

Par Value of Unit,
U. S. $

U. S. A. Selling Price of Unit
on October 24, 1969, U. S. $

Denmark

Krone

0.1335

0.133075

France

Franc, F

0.18004

0.178975

Israel

Pound

0.2900

0.2875

Japan

Yen, ¥

0.00277778

0.002798

Netherlands

Guilder

0.276243

0.2785

United Kingdom

Pound, £

2.40

2.3921

Deutschmark, DM

0.25

0.2715

West Germany

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

11/69

USERS' GUIDE

AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

AUERBACH

@)

P.i" .... ...1 .i_II C. A

A
fa!

..

AUERBACH

4:001;100

AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

USERS' GUIDE

USERS' GUIDE
.1

INTRODUCTION
AUERBACH Computer Notebook International is a looseleaf reference service that provides
the facts and insights you need to understand digital computer systems manufactured throughout the world and to make straightforward, objective comparisons of their capabilities and
costs. Monthly supplements keep the Notebook comprehensive and up-to-date - and keep
you informed of significant new developments in the computer field.
AUERBACH Computer Notebook International consists of five major sections:
•

Users' Guide (this section)

•

Glossary

•

Special Reports

•

Comparison Charts

•

Computer Descriptions and Price Lists (behind the tabs labeled BURROUGHS
through UNIVAC and Denmark through West Germany).

The contents and purpose of each of these sections are
follow. Next, beginning on page 4:001. 800, you'll find
the information in this Notebook to solve various types
in evaluating and using computers and in auditing their

explained in the paragraphs that
straightforward guidelines for using
of problems that are often encountered
operations.

AUERBACH Computer Notebook International is a uniquely useful tool for everyone who needs
to understand and use digital computer systems. Like most tools, it will be of some value
to nearly everyone who uses it, but it will be of far greater value to those who are willing to
invest a little time and effort in learning how to use it most effectively. To ensure that all
of the information in this Notebook can be effectively employed in solving your data processing
problems, we strongly recommend a thorough reading of the remainder of this Users' Guide .
•2

GLOSSARY
Everyone who is called upon to participate in the selection, application, or auditing of digital
computer systems needs to develop an understanding of the meaning and significance of a
large number of technical terms. The Glossary in this Notebook has been specifically
designed to fill that need.
Approximately 700 terms related to digital computers and their applications are listed in
straightforward alphabetical order. The terms are not merely defined; their Significance,
implications, and interrelationships are discussed, and examples and cross-references are
liberally employed to clarify the presentation. Thus, this Glossary can serve not only as a
reference tool, but also as a tutorial guide to many aspects of the computer field.
The first page of the Glossary (page 7:001.001) contains brief instructions for using it
effectively•

•3

SPECIAL REPORTS
This section of the Notebook contains tutorial papers and state-of-the-art reports on timely
topics in computer technology and applications. These reports will help you keep informed
on developments and trends in the data proceSSing field. The Special Reports have been
extracted from AUERBACH Standard EDP Reports and will be revised and updated when
necessary to maintain their value as up-to::aate reference sources .

•4

COMPARISON CHARTS
The" Comparison Charts in this volume are divided into four basic categories:
Configuration Rentals'
© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

5/69

4:001.400

.4

USERS' GUIDE

COMP.i\RISON CHARTS (Contd.)
•

Hardware Characteristics

•

Software

•

System Performance.

The Quick Reference Index on page 11:001. 002 will guide you quickly to all of the Comparison
Chart entries that pertain to any specific computer system.
The organization of the comparison charts can be seen by looking at the Table of Contents
for the Comparison charts on page 11:001. 101. The Introductions identified in the Table
of Contents present the structure of the Charts and the meaning of each entry .
•5

COMPUTER DESCRIPTIONS AND PRICE LISTS
The largest section of this Notebook contains Computer Descriptions and Price Lists. These
are extracted from the detailed Computer System Reports for the commercially important
U. S. -manufactured digital computer systems, and are especially prepared for this Notebook
for non-U. S. -manufactured computer systems.
The U. S. Computer Descriptions and Price Lists are arranged alphabetically by manufacturer, with certain exceptions as indicated in the Table of Contents on page 1:001. 001.
Non-U. S. Computer Descriptions and Price Lists are arranged alphabetically by manufacturer within country of manufacture. Divider tabs for the major U. S. computer manufacturers, from BURROUGHS to UNIVAC, and for countries, from DENMARK to WEST
GERMANY, make it easy to locate the information you need.
Each Computer Description summarizes the general scale and orientation of the system, its
degree of compatibility with other equipment, the capabilities of the central processor and
peripheral devices, the optional features, the capabilities for simultaneous operations, the
available software (compilers, assemblers, operating systems, etc.), and the principal
advantages and drawbacks relative to competitive systems. Unlike the other sections of this
volume, the Computer Descriptions do not conform to a rigidly standardized format. The
aim is to summarize and interpret the important features of each system in a concise,
readable report.
Unless otherwise noted, the Price List for each system contains single-shift monthly rental
prices, purchase prices, and monthly maintenance charges for each hardware component
and optional feature. (Remember that the total monthly rentals for several representative
configurations of each U. S. system can be found in the Configuration Rentals Comparison
Chart on Page 11:010.101.)

•6

REGULAR SUPPLEMENTS
Your copy of AUERBACH Computer Notebook International will be kept comprehensive and
up to date by means of monthly supplements. Each supplement contains new. reports on
recently announced equipment and/or revised versions of previously published reports that
reflect changes in equipment characteristics and in the state of the art.. These supplements
serve an important current-awareness function by keeping you informed .of Significant new
developments in the computer field.
A cover sheet containing a summary of the new information and easy-to-follow filing instructions accompanies each supplement. We recommend that you set up.a standard procedure to
ensure that each new supplement will be filed promptly in your binder. (Note that the page
numbering system is far simpler than it may appear to be at first glance; all page numbers
are arranged in strict numerical sequence, although there are many "gaps, " or omitted page
numbers, to facilitate the insertion of new material in the most appropriate places.)

.7

DERIVATION AND RELIABILITY
AUERBACH Computer Notebook International is prepared and edited by experienced computer
system analysts. Most of the material on U. S. computers in this Notebook is extracted from
AUERBACH Standard EDP Reports, an 8-volume analytical reference service that for more
than five years has served as an authoritative source of information on computer equipment,
software, and performance for computer users, manufacturers, and consultants.
In gathering, analyzing, and evaluating material for these reports, our staff starts with the
specifications and manuals issued by the equipment manufacturers and other reliable sources.

5/69

fA

AUERBACH
®

(Contd. )

:RS' GUIDE

.7

4:001.700

DERIVATION AND RELIABILITY (Contd.)
Advance information from the manufacturers frequently enables us to publish a detailed
analysis immediately after the official announcement of a new computer system. Extensive
amplification and clarification of the generally-available specifications are usually obtained
through visits to or correspondence with the manufacturers. Users of the equipment arc
also interviewed whenever practical.
Every report describing a specific manufacturer's equipment or services is sent to the manufacturer for review prior to publication. We invite the manufacturer's comments regarding
the completeness and accuracy of the report. Where differences of opinion exist between a
manufacturer and our staff, however, the published material always reflects the opinion of
our staff.
Comments and suggestions from our subscribers are always welcome because they help us
to make this publication even more effective in meeting the needs of its users. We welcome
notification of any errors or omissions, as well as suggestions for additions to the Notebook
or improvements in its clarity or balance .

•8

HOW TO USE THIS NOTEBOOK EFFECTIVELY
The information in this volume can meet many different needs, and you will probably find
new uses nearly every time you open it. There are many possibilities for casual yet
rewarding "browsing" that will enrich your overall understanding of computers and their
applications. Most of your computer information problems, however, will probably fall
within one of the following three classes:
(1)

Details are needed on certain characteristics or capabilities of one or more specific
computer systems. How can they be found most efficiently?

(2)

The required equipment configuration and price range for a computer system are
known. Which computers fit into this class, and what are their capabilities?

(3)

The performance requirements for a computer system are known. Which computers can meet these requirements, and how much will they cost?

Suggested procedures for using the information in this Notebook to solve each of these three
types of problems are described in the following paragraphs .
. 81

When Details on a Specific Computer System Are Needed:
Use the Table of Contents on page 1:001. 001 to guide you to the appropriate Computer
Description and Price List. Here you will find a concise report covering the system's
design orientation, software, features, and limitations, plus detailed cost data. Also, be
sure to check the appropriate columns in the Hardware Characteristics Comparison Charts
for details on the central processor, storage devices, and input/output equipment; use the
Quick Reference Index on page 11:001. 002 to guide you to all the pertinent pages. For
standardized measurements of overall processing speeds for U. S. computers, the System
Performance Comparison Charts, beginning on page 11:400.101, are the place to turn .

. 82

When the Required Configuration and Rental Range Are Known (for U. S. Computers):
Turn first to the Introduction to the Configuration Rentals Comparison Charts on page
11:010.101. There you will find the specifications for each of our 13 standard configurations.
Find the standard configuration (identified by a Roman numeral) that most closely matches
your needs. Now turn to the Configuration Rentals Comparison Chart on page 11:010.101.
The column corresponding to your standard configuration contains the monthly rentals for a
number of computer systems that are suitable for use in the type of equipment configuration
you need. Those systems that fall within the allowable rental range can now be further
investigated in any or all of the following ways:
•

Use the System Performance Comparison Charts (for U. S. systems), beginning on
page 11:400.101, to check each system's overall performance on typical business
and scientific problems.

•

Use the Hardware Characteristics Charts, beginning on page 11:210.101, to compare the important characteristics of the central processors and peripheral devices
available for each system. Use the Quick Index begiuning on page 11:001. 002 to
locate the specific information you want.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

5/69

4:001.820

. 82

USERS' GUIDE

When the Required Configuration and Rental Range Are Known (for U. S. Computers): (Contd.)
•

. 83

Turn to the individual Computer Descriptions and Price Lists (using the Table of
Contents on page 1:001. 001) for a summary of the important features and drawbacks
of each system plus detailed pricing information .

When the Performance Requirements are Known:
First, relate the specific performance requirements for U. S. computers to one or more of
the five standard benchmark problems described beginning on page 11:400.100 of the Comparison Charts. If your workload consists mainly of commercial data processing applications, this should not be difficult because most of the runs will probably be basically similar
to the File Processing or Sorting standard problems. For scientific applications, check the
descriptions of the Matrix Inversion and Generalized Mathematical Processing problems to
see which one(s) are most like your principal applications.
When you have determined which standard problem, or appropriately-weighted combination
of problems, best approximates your requirements, turn to the System Performance Comparison Charts, beginning on page 11:400.104. From the listed processing times for the
standard problems, find which computer systems appear to be able to do the job within the
allowable time, and their monthly rentals.
The systems that survive this "screening" process can now be further investigated by turning
to either the Hardware Characteristics Comparison Charts (using the Quick Reference Index
on page 11:001. 002) or the individual Computer Descriptions and Price Lists (using the Table
of Contents on page 1:001. 001) .

.9

WHAT THIS NOTEBOOK CAN - AND CANNOT - DO FOR YOU
The facts, evaluations, and insights in this Notebook can:
•

Provide the background information you need to understand and apply digital
computers.

•

Serve as a ready reference to answer specific questions posed by your associates
or your clients.

•

Keep you informed of new developments in the fast-moving computer field.

•

Provide useful indications of the prices and performance of competitive computer
systems in applications similar to your own or your clients'.

•

Help to narrow the range of choices and aid in the decision-making process whenever computer equipment must be selected.

•

Assist you in preparing requests for proposals and in evaluating proposals from
computer manufacturers.

•

Allow you to compare computer systems on an international basis.

It is important to remember, however, that it would be impossible to include all of the
pertinent information about computers in a compact volume such as this, or toensure that
all of the published information is completely up-to-date at all times. One important aspect
that cannot feasibly be included in a generalized publication such as this is the availability
and quality of local support - both maintenance service and programming assistance - for
each computer system.

Therefore, you should keep in mind, and utilize when necessary, other sources of information about computer systems. The possibilities include: computer manufacturers' representatives, documentation published by the manufacturers, users of the computers being
investigated, independent computer consultants, and more detailed computer reference
services such as AUERBACH Standard EDP Reports.
No matter which of these alternative sources of information you decide to utilize, you'll find
that your AUERBACH Computer Notebook International will give you a good "head start, " so
that you can effiCiently gather the remaining information you need to solve the problem or
complete the evaluation.

5/69

fA.

AUERBACH
®

GLOSSARY

AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

AUERBACH

~

7:001. 001

A

STAIIDARD

EDP

GL.OSSARY

REPORTS

AUERBACH
~

GLOSSARY
.1

PURPOSE AND SCOPE

.2

This Glossary has been compiled with three main objectives in mind:
(1) To define, in a precise and consistent manner, the meanings of more than
700 words and phrases as used in the AUERBACH Computer Technology Reference Services.
(2) To aid the novice in understanding what he reads and hears about the data processing field by providing clearcut, up-to-date explanations of the terminology,
with liberal use of illustrative examples and cross-references.
(3) To guide the expert in choosing the correct term to express a given concept by
clarifying the distinctions and similarities among related terms.
Although a number of other glossaries of data processing terms have been prepared (see
the Annotated Bibliography in Paragraph. 3), none of them was found to be satisfactory
to meet the objectives stated above. Therefore, this Glossary was compiled through
careful analysis of the definitions in previous glossaries and of current usage in the data
processing field; it contains many terms, examples, comments, and cross-references
that are not included in any of the glossaries listed in the Bibliography.
With respect to scope, this Glossary defines more than 700 terms whose meanings in the
data processing field are different from their meanings in the general U. S. vocabulary.
It does not include terms whose meanings are obvious or are the same as in the everyday,
nontechnical vocabulary. Also excluded, in general, are specialized terms that have been
arbitrarily coined by individual computer manufacturers or users and have not found
widespread acceptance .
ORGANIZATION AND USE
In compiling a Glossary, it is necessary to choose one of two basic forms of organization:
dictionary form, in which the entries are arranged alphabetically, or thesaurus form, in
which the entries are arranged in logical groups to keep related terms close together.
Although the thesaurus form has certain advantages, it has one major disadvantage: to
locate the definition of a particular term, one must first consult an alphabetical index.
Therefore, to facilitate rapid references, we have chosen the dictionary form, with aU
terms arranged in straightforward alphabetical order.
All multi-word phrases are listed in their natural order (e. g., "absolute address" rather
than "address, absolute").
In the case of terms with two or more distinct meanings, the meanings are numbered sequentially, and the first meaning listed is the most common or most general one.
Numerous examples and comments follow the formal definitions to clarify the meanings,
usage, and significance of the concepts and entities that are defined in this Glossary.
Several different types of cross-references are used. Their meanings are as follows:
•
Same as - indicates that the referenced term has the same meaning as (i. e., is
synonymous with) the term containing the reference, and that the referenced
term is the preferred one.
•
Synonymous with - indicates that the referenced term has the same meaning 3.S
the term containing the reference, and that the term containing the reference is
the preferred term.
•
Contrast with - indicates that the referenced term is a related term that has a
meaning significantly different from that of the term containing the reference.
•
See also - indicates that the referenced term is a related term whose definition
will provide additional background or clarification.
•
See - indicates that the referenced term is an alternative or qualified form of
the term containing the reference.
•
Underline - indicates that the underlined term is Significant in the definition
and is defined elsewhere in the Glossary.
Because of its straightforward alphabetical arrangement, this Glossary can be used in the
same way as a dictionary: simply turn to the term of interest and read its definition. All
of the underlined terms used in the definition are defined elsewhere in the Glossary, and
if you are not sure of their precise meanings, you may want to turn to their definitions.

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

5/67

AUERBACH STANDARD EDP REPORTS

7:001. 002
.2

ORGANIZATION AND USE (Contd.)
In this way, a "chain" of cross-references can guide you in learning all the important
terms and concepts that are associated with some particular aspect of the data processing
field.
Thus, the many cross-references, coupled with the examples and comments, make this a
Glossary that can serve not only as a reference work, but as an educational guide to many
aspects of computers and data processing.

.3

ANNOTATED BIBLIOGRAPHY
Th0 principal sources of input for this Glossary are listed below, along with brief comments
about their format, content, and usefulness. It should be noted that numerous other glossaries and dictionaries were studied, but the others were found to be either highly specialized or significantly less precise and authoritative than the sources listed below.
(1) AUERBACH Corporation, "Glossary," AUERBACH Standard EDP Reports, May
1962.
This earlier version of the AUERBACH Glossary served as the principal basis
for the new edition because of the importance of maintaining consistency with the
terminology that has been used in AUERBACH Standard EDP Reports throughout
its five-year history. However, the earlier version was in thesaurus form, contained no cross-references, had fewer examples and comments, and, of course,
did not include the many important terms that have been introduced since 1962.
(2) AUERBACH Corporation, "Users' Guide," AUERBACH Standard EDP Reports,
May 1962.
This 162-page Users' Guide, though not a glossary, contains expanded definitions
and practical examples of many of the important terms and concepts. As such,
it served as a valuable reference in the compilation of this Glossary.
(3)

Bureau of the Budget, Automatic Data Processing Glossary, U. S. Government
Printing Office, Washington, D. C. , December 1962 (also published as the Datamation ADP Glossary).
This 62-page glossary, arranged in dictionary form, was compiled in 1962 to
serve as a U. S. Government reference on data processing terminology. Although
the coverage is quite broad, many of the definitions are somewhat imprecise and
inconsistent, there are few cross-references, and many terms of current significance are not included.
(4) Honeywell, Inc., Glossary of Data Processing and Communications Terms,
Third Edition, April 1966.
This 88-lJage publication combines selected definitions from references (3) and
(6) with original, none-too-precise definitions of terms related to the communications field. It uses the dictionary form.
(5) IFIP/ICC (International Federation for Information Processing and International
Computation Centre), IFIP/ICC Vocabulary of Information ProceSSing, NorthHolland Publishing Company, Amsterdam, 1966.
This 208-page hardbound volume probably represents the most impressive and
successful effort to date to develop a truly definitive glossary of data proceSSing
terminology. Its definitions are quite precise, and examples and explanatory comments are liberally employed. The thesaurus format permits clear distinctions
among related or contrasting terms. For our purposes, however, the volume has
three major drawbacks: (1) the thesaurus format requires use of an index to locate
specific terms; (2) the distinctly British flavor of the language, terminology, and
usage can distract - and in some cases mislead - American readers; and (3) terms
describing specific languages, codes, and applications (e. g., COBOL, ASCII, message switching) are conspicuously absent.
(6)

U. S. A. Standards Institute, American Standard Vocabulary for Information Processing, U. S. A. Standard X3. 12-1966, New York, 1966.
This 32-page vocabulary was prepared by ASA (now USASI) Subcommittee X3. 5,
Terminology and Glossary, and approved as a U. S. A. Standard on June 14, 1966.
Its purpose is to define current usage and encourage standardization of terms and
their meanings. The dictionary form is used, with enough cross-references to
clarify most of the interrelationships among terms. Several significant drawbacks,
however, detract from the usefulness of the American Standard Vocabulary: (1) most
of the definitions are brief, with relatively few examples and explanatory notes;
(2) many of the definitions are so terse that they are totally unsatisfactory; e. g .•
"COBOL. (Common Business Oriented Language. ) A business data proceSSing
language"; and (3) numerous terms of considerable current importance have been
omitted; e. g., background program, EAM, emUlator, integrated circuit, PL/I,
privileged instruction, systems analYSiS, virtual address, etc.
(Contd. )

5/67

A

AUERBACH

'"

7:001. 101

GLOSSARY

A
absolute address
An address that is permanently assigned by the
machine designer to a particular storage location.
For example, the addresses 0000, 0001, and 0002
might be assigned to the first three locations in a
computer's working storage.
absolute coding
Coding that uses machine instructions and absolute
addresses; therefore, it can be directly executed
by a computer without prior translatwn to a different
form. Contrast with relative coding and symbolic
coding.
access mode
A technique used to obtain a specific record from,
or to place a specific record into, a specific file.
See random access and serial ~
access time
The time interval between the instant when a computer or control unit calls for a transfer of data to
or from a storage device and the instant when this
operation is completed. Thus, access time is the
sum of the waiting time and transfer time. Note: In
some types of storage, such as disc and drum storage, the access time depends upon the location specified and/or upon preceding events; in other types,
such as core storage, the access time is essentially
constant.
accounting machine
(1) A keyboard-actuated machine used to prepare
accounting records. (2) Same as tabulator.
accumulator
A register that holds one operand, with means for
performing various arithmetic and/or logical operations involving that operand and (where appropriate)
another operand; usually, the result of the operation
is formed in the accumulator, replacing the original
operand. Note: Among computers currently in use,
some have a single accumulator, others have multiple accumulators, and still others (especially those
that use two-address or three-address instructions)
have no accumulator as such; in the latter case, the
results of arithmetic and logical results are usually
formed in programmer-specified locations in the
computer's main storage.
accuracy
The degree of freedom from error; a measure of the
smallness of error or the range of error. Thus,
high accuracy implies small error. Note: Accuracy
should be carefully distinguished from precision,
which is the degree of discrimination with which a
quantity is stated. For example, a 6-digit numeral
is more precise than a 4-digit numeral, but a properly computed 4-digit result may be more accurate than
an improperly computed 6-digit result.
acronym
A word formed from the initial letter or letters of
the words in a name or phrase; e. g., ALGOL from
ALGOrithmic Language, COBOL from COmmon
Business Oriented Language.

activity
The degree of frequency with which individual records
in a file are used, modified, or referred to. For
example, an "activity factor" of 0.10 (or 10 per cent)
denotes that an average of 1 of cvery 10 master- file
records is referenced or affected by a transaction
during a run.
adder
--X-device capable of forming a representation of the
sum of two or more numbers whose representations
are supplied as inputs.
address
A name, numeral, or other reference that deSignates
a particular location in a store or some other data
source or destination. Note: Numerous types of
addresses are employed in computer programming;
see, for example: absolute address, base address,
direct address, effective address, immediate address, indirect address, relative address, symbolic
address.
address format
The arrangement of the address parts of an instruction. Among the commonly-used address formats
one-address, one-pIus-one, two-address, and
three-address. Note: In some computers all of the
instructions employ the same address format, while
in other computers two or more different address
formats are used with the various types of instructions.

are

address modification
An operation that causes an address to be altered in
a prescribed way by a stored-program computer.
Note: The address upon which modification is performed is called the presumptive address, and the
address that results is called the effective address.
See also index and indirect address - the two most
common forms of address modification.
address register
A register capable of holding the address of a location
in a store or of some other data source or destination.
ADP (Automatic Data Processing)
Data processing performed largely by automatic
means; i. e., by a system of electronic or electrical
machines which require little human ;:.ssistance or
intervention.
alarm
--rsignal that warns a human operator of an equipment
fault or some other abnormal condition; e. g., a
warning lamp or buzzer.
ALGOL (ALGOrithmic Language)
A process oriented language developed as a result of
international cooperation to develop a standard language for expreSSing computational algorithms.
ALGOL is designed to serve as a means for communicating computational procedures among humans,
as well as to facilitate the preparation of such procedures for execution on any computer for which a
suitable ALGOL compiler exists. Note: The basic
elements of ALGOL are arithmetic expressions containing numbers, variables, and functions. These are

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

5/67

7:001. 102

AUERBACH STANDARD EDP REPORTS

combined to form self-contained units called assignment statements. Declarations are non-computational
instructions which inform the compiler of characteristics such as the dimensions of an array or the class
of a variable. A sequence of declarations followed by
a sequence of statements, all enclosed within "begin"
and "end" instructions, constitute an ALGOL program
block. ALGOL is not widely used in the United
States, but is very popular in Europe.

application package
A computer routine or set of routines designed for a
specific application (e. g., inventory control, on-line
savings accounting, linear programming, etc.) Note:
In most cases, the routines in the application packages are necessarily written in a generalized way and
will need to be modified to meet each user's own
specific needs.

algorithm
A set of well-defined rules for the solution of a
problem in a finite number of steps; e. g., a full
statement of an arithmetic procedure for evaluating
the sine of an angle to a stated precision, or a full
statement of a procedure for computing a rate of return. Contrast with heuristic.

argument
An independent variable. For example, in table
look-up operations, the arguments are the numbers
that are used to identify the locations of the desired
items in the table.
arithmetic unit
A section of a computer in which arithmetic, logical,
and/or shift operations are performed.

allocation
The assignment of specific portions of storage de-·
vices or specific input-output devices to hold specific
programs and/or data files. Note: Allocation of
storage and input-output devices may be performed:
(1) by the programmer, when he writes a program;
(2) by the operator, when he loads a program for execution; or (3) automatically, by an operating system.

array
A group of items arranged in a meaningful pattern.
Example 1: A one-dimensional array (i. e., a list) of
one-word items:
Adams
Baker

Collins
Dorsey
Example 2: A two-dimensional array (i. e. , a matrix)
of one-digit items:
7 254
3
8
0
9
6
9
1
6
4
7
3
8

alphabet
An ordered set of characters used for the representation of sounds in a spoken language; in English, the
26 letters A through Z.
alphameric
Same as alphanumeric.
alphanumeric
Pertaining to a character set that includes both alphabetic characters (letters) and numeric characters
(digits). Note: Most alphanumeric character sets
also contain special characters.
analog
-pertaining to data represented in the form of continuously variable physical quantities (e. g., voltage or
angular position). Contrast with digital.
analog computer
A computer that operates on analog data by performing physical processes on the data. Contrast with
digital computer.
analyst
----xperson skilled in defining problems and developing
algorithms or other systematic procedures for their
solution. See also programmer.
AND
--(1) A logical operator which has the property that if
P is a statement and Q is a statement, then IIp AND Q"

is true if both P and Q are true, and false if either P
or Q, or both P and Q, are false. lip AND Q" is often
represented by PI\Q, p. Q, or PQ. (2) The logical
operation that uses the AND operator; also called
logical product, logical multiplication, and conjunction.
application
The problem or system to which a computer (or other
proceSSing equipment or technique) is applied.

artificial intelligence
The capability of computers or other devices to perform functions that are normally associated with
human intelligence, such as reasoning, learning,
adapting to environmental changes, and self-improvement.
ASCII (American Standard Code for Information Interchange)
~-bit code adopted as a U. S. A. Standard in order to
facilitate the interchange of data among various types
of data processing and data communications equipment.
Note: Because of the very large investment in equipment and programs which use earlier codes, ASCII
has not been Widely used to date, but a steady trend
toward its usage may be expected.
assemble
To prepare a machine language program from a program written in symbolic coding by substituting absolute operation codes for symbolic operation codes
and absolute or relocatable addresses for symbolic
addresses. For example, the symbolic instruction
ADD TAX might be assembled into the machine instruction 24 1365, where 24 is the operation code for
addition and 1365 is the address of the storage location labeled TAX. Contrast with compile and generate.
assembler
A computer program that assembles programs written
in symbolic coding to produce machine language programs. Note: Assemblers are an important part of
the basic software for most computers; their use can
greatly reduce the human effort required to prepare

(eontd. )
5/67

A

AUERBACH
@

GLOSSARY

7:001. 103

and debug computer programs by enabling the coder
to use a symbolic language which is simpler and more
meaningful to him than the computer's machine
language.

instead, data is transferred in blocks between auxiliary and working storage. Disc storage and drum
storage are the most common types of auxiliary
storage.

associative memory
A storage device whose storage locations are identified by their contents (rather than by names or positions, as in most computer storage devices). Synonymous with content-addressable memory. Note:
Associative memories can facilitate programming
and increase computer efficiencies by eliminating
the need for item-by-item search operations, but
the high cost of implementing such memories limits
their current use to specialized functions such as
holding small, frequently-referenced tables.

availability
The period of time, usually quoted by an equipment
manufacturer, that can be expected to elapse between
placement of a non-priority order for a particular
type of equipment and delivery of the eqUipment to
the user.

asynchronous computer
A computer in which each operation starts as a result of a signal generated by the completion of the
previous operation or by the availability of the equipment required for the next operation. Contrast with
synchronous computer.

B box
------same as index register.

attribute
A characteristic of an entity; e. g., the attributes
of a file might include its name, use, creation date,
record length, record format, and the device to
which it is currently assigned.
audio response unit
Same as voice response unit.
audit trail
A means for systematically tracing the progress of
specific items of data through the steps of a process
(particularly from a machine-generated report or
other output back to the original source document) in
order to verify the validity and accuracy of the
process.
automatic check
A check performed by a facility that is built into
equipment specifically for checking purposes. Also
called a "built-in check" or "hardware check. "
Contrast with programmed check.
automatic data processing
See ADP.

available time
Same as uptime.

B

background program
A program, usually of the batch processing type,
that is not subject to any real-time constraints and
can be executed whenever the facilities of a multiprogramming computer system are not required by
real-time programs or other programs of higher
priority. Contrast with foreground program.
backspace
To move an input or output medium backward for a
distance of one unit; e. g., one character position on
a typewriter, one row on punched tape, or one block
on magnetic tape.
backup
---pertaining to eqUipment or procedures that are available for use in the event of failure or overloading of
the normally-used equipment or procedures. Note:
The provision of adequate backup facilities is an
important factor in the design of every data processing system, and is especially vital in the design of
real-time systems, where a system failure may
bring the total operations of a business to a virtual
standstill.
backup storage
Same as auxiliary storage.
band

---0:) A group of tracks

automatic programming
(1) The use of a computer to perform some stages of
the work involved in preparing programs. (2) In
particular, the use of a computer to translate programs expressed in a process oriented language into
machine language or a machine oriented language
(i. e., to compile).
automation
The theory, art, or technique of making processes
more automatic, thereby reducing or eliminating the
need for human intervention.
auxiliary storage
Storage that supplements a computer's working stor~. Synonymous with backup storage, mass stor~, and secondary storage. Note: In general, the
auxiliary storage has a much larger capacity but a
longer access time than the working storage. Usually, the computer cannot access auxiliary storage
directly for instructions or instruction operands;

(usually in a disc storage or
drum storage unit) which are associated for some
specific purpose; e. g. , a group of 8 tracks which are
read and recorded upon in parallel to permit highspeed transfers of 8-bit bytes of data. (2) The range
of frequencies between two defined limits.

base

--n.) A reference value.
(2) Same as radix.

base address
A specified address (often held in a "base address
register") which is combined with a relative address
(usually contained in an instruction) to form the absolute
address of a particular storage location. Synonymous
with origin.
batch processing
A technique in which items to be processed are collected into groups (i. e., "batched") to permit convenient and efficient processing. Note: Most business

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

5/67

7:001. 104

AUERBACH STANDARD EDP REPORTS

blank
----X-character used to produce a character space (io e. ,
no mark) on an output medium.

applications are of the batch processing type; the
records of all transactions affecting a particular
master file are accumulated over a period of time
(e. g., one day), then they are arranged in sequence
and processed against the master file.

block
--X-group of words, characters, or digits that are held
in one section of an input/output medium or store and
handled as a unit; e. g., the data recorded on a
punched card, or the data recorded between two interblock gaps on a magnetic tape.

batch total
A sum of a set of items which is used to check the
accuracy of operations on a particular batch of
records.
baud
--;;;: unit of signalling speed equal to the number of discrete conditions or signal events per second. Note:
In the case of a train of binary signals, and therefore
in most data communications applications, one baud
equals one bit per second.
Baudot code
A 5-bit code used in telegraphy for more than 100
years and still widely used in data communications
and punched tape. Note: The Baudot code has two significant disadvantages: the limitation to 5 bits per
character (i. e., 32 code combinations) requires
frequent shifts between the "letters" and "figures"
cases, and there is no provision for a parity check.
The Baudot code is gradually being replaced by ASCII
and other codes.

block diagram
A diagram of a system, instrument, computer, or
program in which selected portions are represented by
annotated boxes and interconnecting lines. Note: A
flowchart is a special type of block diagram that shows
the structure and general sequence of operations of a
program or process.
blocking
Combining two or more records into one block. Note:
The principal purpose of blocking is to increase the
efficiency of computer input and output operations.
For example, the effective data transfer rates of
most magnetic tape units can be greatly increased by
reducing the need for frequent tape stops and starts
through combining multiple short records into blocks
which are several thousand characters in length.

BCD (Binary Coded Decimal)
Pertaining to a method of representing each of the
decimal digits 0 through 9 by a distinct group of
binary digits. For example, in the "8-4-2-1" BCD
notation, which is used in numerous digital computers,
the decimal number 39 is represented as 0011 1001
(whereas in pure binary notation it would be represented as 100111).
benchmark problem
A precisely defined problem that is coded and timed
for a number of computers in order to measure their
performance in a meaningful and directly comparable
manner. Note: The benchmark problem may be one
of the user's own specific applications, or (as in the
case of the AUERBACH System Performance comparisons) it may be representative of a class of
typical computer applications.
binary
---pertaining to the number system with a radix of two,
or to a characteristic or property involving a choice
or condition in which there are two possibilities,
Note: The binary number system is widely used in
digital computers because most computer components
(e. g., vacuum tubes, transistors, flip-flops, and
magnetic cores) are essentially binary in that they
have two stable states. Example: The binary numeral 1101 means: (1 x 2 3) + (1 x 22) + (0 x 21) + (1 x 20)
which is equivalent to decimal 13.
binary coded decimal
See BCD.
binary search
A search technique in which a set of items is divided
into two parts, one of the parts is rejected, and the
process is repeatcd on the accepted part until the
item or itcms with the desired property are found;
also called "dichotomizing search. "
bit
- A binary digit; a digit (0 or 1) in the representation
of a number in binary notation.

5/67

A

Boolean operation
A logical operation on single bits, Note: The term
"Boolean" refers to the processes used in a special
type of algebra formulated by George Boole.
bootstrap
(1) A form of loader whose first few instructions are
sufficient to bring the rest of itself into the computer's
storage from an input device. (2) More generally, a
technique or device designed to bring itself into a desired state by means of its own action.
branch
-WSame as conditional transfer. (2) A set of instructions that are executed between two successive conditional transfer instructions.
breakpoint
A specified point in a program where the program may
be interrupted by manual intervention or by a monitor
routine. Note: Breakpoints are usually used as an aid
in testing and debugging programs; they facilitate
halting a computer or triggering a printout at a particular point so that specific conditions can be
examined.
brush

--xu electrical conductor used to sense the presence or
absence of holes in a punched card.

buffer
---X-Storage device used to compensate for differences
in the rates of flow of data or in the times of occurrence of events when transmitting data from one device to another. For example, a buffer holding one
line is associated with most line printers to compensate for the large difference between the high speed
at which the computer can transmit data to the printer
and the relatively low speed of the printing operation
itself.

~
A mistake in the design of a program or a computer
system, or an equipment fault.
(Contd. )

AUERBACH

'"

GLOSSARY

7:001. 105

bus
--A major path used to transmit signals from one or
more sources to one or more destinations. Synonymouswith~

byte
A group of adjacent bits operated upon as a unit and
usually shorter than a ~ Note: In a number of
important current computer systems, the term "byte"
has been assigned the more specific meaning of a
group of eight adjacent bits, which can represent one
alphanumeric character or two decimal digits.

c
calculator
A device capable of performing arithmetic operations.
Note: This term is generally applied only to devices
that require frequent intervention by a human operator; contrast with computer.
call
--(1) In computer programming, to transfer control to
a subroutine, usually by supplying the required
parameters and executing a jump to the entry point of
the subroutine. (2) In communications, the actions
performed by the party initiating a connection, or
the effective use that is made of a temporary connection between two stations.
calling seguence
A specified set of instructions and data necessary to
call a given subroutine.
card
-Usually same as punched card; see also edge-notched
card, edge-punched card, magnetic card.
card field
In a punched card, a group of columns (or parts of
columns) whose punchings represent one item. For
example, a three-column field might hold an item
representing order quantity, whose value ranges from
000 to 999.
card image
A direct, one-to-one representation of the contents of
a punched card; e. g., a matrix, in core storage or on
magnetic tape, in which a "1" bit represents a
punched hole and a "0" bit represents the absence of
a hole.
card punch
A machine that punches holes in punched cards. Note:
The data to be punched in the cards may be transmitted to the punch by a computer, by EAM equipment, or by an operator's keystrokes (see keypunch).
card reader
A machine that senses the holes in punched cards to
provide input to a computer or EAM equipment.
carry
(1) A signal that arises when the sum or product of
two or more digits in one digit position equals or exceeds the radix of the number system in use; the
carry is forwarded to the next more Significant digit
position for processing there. (2) To forward a carry
as defined in (1).

cartridge
A unit of a storage medium which can be conveniently
removed from the storage device and replaced by
other similar cartridges, without loss of the data recorned in it. Note: A variety of interchangeablecartridge storage devices are now in use. They permit rapid random access to their on-line contents,
while providing economical off-line storage for
virtually unlimited volumes of data. The cartridges
usually consist of single or multiple magnetic discs
or multiple magnetic cards.
cathode ray tube
An electronic vacuum tube containing a screen on
which information can be stored or displayed. The
abbreviation CRT is frequently used. Note: Cathode
ray tubes served as the principal storage medium in
some of the early digital computers; they now serve
as the basic component of most display units.
cell
--See storage cell.
central processor
The unit of a computer system that includes the circuits which control the interpretation and execution
of instructions. Synonymous with Q1!!l (central
processing unit) and main frame.
chad
--;;. piece of material that is removed in the process of
forming a hole or notch in a medium such as punched
cards or punched tape.
chadless
Pertaining to the punching of tape in such a way that
no chad results because each hole is only partially
perforated. Note: Chadless perforation makes the
full surface of a punched tape available for interpreting (i. e., printing) the characters represented by the
punching; but many high-speed punched tape readers
cannot read chadless tape.
chain printer
A line printer in which the type slugs are mounted on
a chain that moves horizontally past the printing positions. Note: Chain printers generally provide more
accurate vertical registration than the more commonly used drum printers, and interchangeable chains
often permit rapid changes in the size or make-up of
the character set.
chaining
(1) The linking together of a sequence of instructions
or commands, usually for the purpose 0f simplifying
the coding process or reducing execution time and/or
storage requirements. (2) The division of a program
into a number of sequential segments, only one of
which resides in working storage at a time; each
segment uses the output from the previous segment
as its input.
channel
--;;.path or group of parallel paths for carrying signals
between a source and a destination. See also inputoutput channel.
character
A member of a set of mutually distinct marks or signals used to represent data. Each member has one

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

5/67

7:001. 106

AUERBACH STANDARD EDP REPORTS

synchronous computer. (2) A device that records the
progress of real time, or some approximation of it,
and whose contents are available to a computer program (frequently in a special register); the clock may
also be capable of initiating a program interrupt when
a specified period of time has elapsed.

or more conventional representations on paper (e. g. ,
a letter of the ordinary alphabet) and/or in data processing equipment (e. g., a particular configuration of
and ] bits).

°

character recognition
The identification of graphic characters by automatic
means. See MICR and OCR.

closed loop
(1) A loop from which there is no exit other than by
intervention from outside the program; such a loop is
usually the result of a programming error or machine
fault. (2) Pertaining to a process control system that
utilizes feedback (i. e., information about the condi-tions being controlled) in order to exert self-correcting
influences upon its own operation.

character set
A set of mutually distinct marks or signals used to
"represent data; e. g., a typical character set for a
printer might include the digits through 9, the
letter.:l A through Z, and the common punctuation
marks.

°

closed shop
A computer installation that may be operated (and, in
some cases, programmed) only by personnel on the
staff of the associated computer department. Contrast
with open shop.

characteristic
Same as exponent.
check
--:;;:-general term meaning a partial or complete test for
the absence of certain classes of errors or for the
correct performance of a process. Note: A check
may be either an automatic check or a programmed
check. Among the types of checking commonly performed in computers are echo checks, parity checks,
read-after-write checks, residue checks, summation
checks, and validity checks.

closed subroutine
A subroutine that can be stored in one place and connected to a program by means of linkages at one or
more points in the program. Contrast with open subroutine. Note: The use of closed subroutines tends
to save storage space whenever a particular subroutine must be used at two or more different pOints
in a program.

check bit
A binary check digit. Note: A parity check usually
involves appending a check bit of the appropriate
value to an array of bits.
.

clutch point
In a clutch-operated input or output device (e. g. ,
most card readers and punches), one of the instants
at which it is possible to engage the clutch. For example, in a card reader which has a 3-point clutch
and a 600-millisecond clutch cycle, the clutch points
occur at intervals of 200 milliseconds. Therefore,
it is possible to engage the clutch (and thereby initiate
the feeding of a card) every 600, 800, 1000, 1200, or
1400 . . . milliseconds.

check digit
A digit associated with a word or part of a word for
the purpose of checking for the absence of certain
classes of errors.
check problem
A problem whose correct results are known, and
which is used to determine whether a computer and/or
a program are operating correctly.
check protection
The insertion of a character, most commonly an
asterisk, in place of one or more suppressed zeros to
guard against tampering with the amount printed on a
check. For example, the amount $
9. 98 with
check protection added becomes $****9.98.
check sum
Sec summation check.
checkpoint
(1) A place in a program where the results of one or
morc checks are examined. (2) Same as rerun point.
circuit

-wA

system of conductors and related elements
through which electrical current flows. (2) A communications link between two or more points.

clear
To "erase" (i. e., delete) the data in a storage location or device by bringing all of the storage cells involved to a prescribed state - usually to the state
denoting zero or blank.

COBOL (COmmon Business Oriented Language)
A process oriented language developed to facilitate
the preparation and interchange of programs to perform business data processing functions. Note: Designed in 1959 by a committee representing the U. S.
Government and several computer manufacturers,
COBOL has evolved through several versions (e. g. ,
COBOL-60, COBOL-61, COBOL-61 Extended,
COBOL-65). COBOL-65 forms the basis for a proposed standard version of the language which will
probably soon be adopted as an official U. S. A.
Standard. Every COBOL source program has four
divisions, whose names and functions are as follows:
(1) Identification Division - identifies the source program and the output of a compilation. (2) Environment
Division - specifies those aspects of a data processing problem that are dependent upon the physical
characteristics of a particular computer. (3) Data
Division - describes the data that the object program
is to accept as input, manipulate, create, or produce
as output. (4) Procedure Division - specifies the
procedures to be performed by the object program,
by means of English-like statements such as:
SUBTRACT TAX FROM GROSS-PAY GIVING NET-PAY.
PERFORM PROC-A THRU PROC-B UNTIL X IS
GREATER THAN Y.
code

clock
(1) A timing device that generates the basic periodic
signal used to control the timing of all operations in a
5/67

--"A

A

/Itb\

AUERBACH
(OJ

set of unambiguous rules that specifies the exact
manner in which data is to be represented by the characters of a character set; e. g. , ASCII, Hollerith code.
(eontd.)

GLOSSARY

7:001 107

code translation
The act of converting data from one code to another.
coder
A person who prepares coding for a computer. See
also programmer.
coding
(1) An ordered list or lists of the successive instructions which will cause a computer to perform a
particular process. See also absolute coding, relative coding, skeletal coding, symbolic coding. (2) The
act of preparing coding as defined in (1) above.
collate
~e as merge (i. e., to form a single sequenced file
by combining two or more Similarly sequenced files).
collating sequence
The ranking, or precedence with respect to each
other, of all the characters in a character set that
can be used to constitute a key used for sequencing
purposes. Note: Most collating sequences are arranged so that the digits 0 through 9 and the letters
A through Z fall into their natural sequences. However, either the digits or letters may come first, and
the handling of special characters varies widely.
collator
A machine that feeds and compares two or more files
of punched cards or other documents in order to
match or merge them or to check their sequence.
Note: The cards that match can be separated from
those that do not match, making it possible to select
specific cards as well as to file cards automatically.

communication
The transfer of information from one person, place,
or device to another. See also data communications.
communications link
The physical means of connecting one location to another for the purpose of transmitting information
between them; e. g., a telegraph, telephone, radio,
or microwave circuit.
compare
To examine two words or items to discover whether
they are identical, or to discover their relative magnitudes or relative order in a sequence.
compatibili ty
The characteristic that enables one device to accept
and process data prepared by another device wlthout
prior code translation, data transcription, or other
modifications. Thus, one computer system is "data
compatible" with another if it can read and process
the punched cards, magnetic tape, etc., produced by
the other computer system. See also program compatibility.
compile
To prepare a machine language program ,(or a program expressed in symbolic coding) from a program
written in another programming language (usually a
process oriented language such as COBOL or
FORTRAN). The compilation process usually involves examining and making use of the overall structure of the program, or generating more than one
object program instruction for each source program
statement, or both. Contrast with assemble and
generate.

column

~ vertical arrangement of characters or other

symbols. (2) A location capable of holding one digit
or character, especially in a punched card; e. g., one
of the SO groups of 12 punch positions in a standard
SO-column card.
column binary
Pertaining to a method for representing binary data
on punched cards in which adjacent positions in a
card column correspond to adjacent bits of data. For
example, in a standard SO-column, 12-row card,
each column may be used to represent 12 consecutive
bits of a 36-bit word. Sometimes called "Chinese
binary." Contrast with row binary.
command
(1) A control signal, especially one transmitted from
a computer to a peripheral device or input-output
channel. (2) Loosely, an instruction.
comment
An explanation or identification, for human use, of a
step in a routine; the comment has no effect upon the
operations of the computer that executes the routine.
common language
(1) A programming language that can be used to prepare programs for a number of different computer
systems; examples include ALGOL, COBOL, and
FORTRAN. (2) Loosely, an input-output medium and
code that can be read or recorded upon by a variety
of business machines, thereby facilitating intercommunication among them.

compiler
A computer program that compiles. Note: Compilers
are an important part of the basic software for most
computers; they permit the use of process oriented
languages which can greatly reduce the human effort
required to prepare computer programs. However,
the computer time required to perform the compilation process may be exceSSive, and the object programs produced by the compiler usually require more
execution time and more storage space than programs
written in machine language or symbolic coding.
complement
A number whose representation is derived from the
representation of another number by one of the following rules (or by some equivalent process): (1) To derive the "radix complement" or "true complement, "
subtract each digit from one less than the radiX, then
add 1 to the least significant digit, checuting all carries required. Thus, 830 is the "tens complement"
of 170 in decimal notation using three digits. (2) To
derive the "radix-minus-one" complement, subtract
each digit from one less than the radix. Thus, 829
is the "nines complement" of 170 in decimal notation
using three digits, while 01110 is the "ones complement" of 10001 in five-digit binary notation. Note: In
many computers, the absolute value of a negative
number is represented as a complement of the corresponding positive number.
computer
A device capable of solving problems by accepting
data, performing prescribed operations on the data,

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

5/67

7:00 I. 108

AUERBACH STANDARD EDP REPORTS

and supplying the results of these operations, all without intervention by a human operator. See also analog
computer, digital computer, general-purpose computer, special-purpose computer, stored-program
computer.

control program
A routine, usually contained within an operating system, that aids in controlling the operations and managing the resources of a computer system.

concatenate
To unite in a series; to link together.

control sequence
The normal order of selection of instructions for execution. See also sequential control.

conditional transfer
An instruction that mayor may not cause a jump
(i. e., a departure from the normal sequence of executing instructions) depending upon the result of
some operation, the contents of some register, or
the setting of some indicator. Contrast with unconditional transfer. Note: Conditional transfer instructions are the basic means for implementing decisionmaking processes in stored-program computers.

control unit
(1) A section of a computer that effects the retrieval
of instructions in the proper sequence, interprets
each instruction, and stimulates the proper circuits
to execute each instruction. (2) A device that controls
the operation of one or more units of peripheral equipment under the overall direction of the central
processor.

configuration
A specific set of equipment units which are interconnected and (in the case of a computer) programmed to
operate as a system. Thus, a computer configuration
consists of one or more central processors, one or
more storage deVices, and one or more input-output
devices. Synonymous with system configuration.

conversational mode
A mode of operation that implies a "dialogue" between
a computer and its user, in which the computer program examines the input supplied by the user and
formulates questions or comments which are directed
back to the user.

connector
In a flowchart, a means of representing the convergence of two or more paths into one, the divergence
of one path into two or more paths, or a "break" in a
single path which is continued in another area.

convert
TO'transform data according to some criteria while
preserving its information content; e. g., radix conversion from decimal to binary , code translation from
Hollerith to EBCDIC, data transcription from punched
cards to magnetic tape, conversion from analog to
digital representation, etc.
---

console
~rtion of a computer that is used for communication between operators or maintenance engineers and
the computer, usually by means of displays and
manual controls.
constant
A quantity whose value does not vary.
variable.

converter
A device that converts data from one form to another
in order to make it available or acceptable to another device; e. g., a "card-to-tape" converter that transcribes
data from punched cards to magnetic tape so that the data
can be read into a computer system at high speed.

Contrast with

copy
To reproduce data in a new location, leaving the original data unchanged.

constant-ratio code
A code in which all of the valid characters have the
same number of 1 bits, thereby facilitating the performance of a Validity check. For example, in the
"4-of-8" code, frequently used in data communications, each of the valid characters is represented by
a combination of four 1 bits and four 0 bits.

core storage
A type of storage that uses an array of magnetic cores,
each capable of storing one bit of data. Note: Most
current computers use magnetic core storage as their
main working storage. This widespread acceptance is
due to the fact that magnetic cores require no power
while storing data, can be switched rapidly from one
state to the other by relatively small currents, and
can tolerate adverse environmental conditions.

content-addressable memory
Same as associative memory.
contents
A general term for the data contained in any storage
device, location, or medium.
control card
A punched card that contains input data required for a
specific application of a general routine such as a
generator or operating system; e. g., one of a series
of cards that direct an operating system to load and
initiate execution of a particular program.
control counter
Same as sequence counter.
control panel
(1) A part of a computer console that contains manual
controls such as switches, buttons, and dials.
(2) Same as plugboard.
5/67

corrective maintenance
Maintenance that is carried out to correct a fault.
Contrast with preventive maintenance.
-counter

~vice, such as a register or storage location, that

holds a number, permits this number to be increased
by one or by an arbitrary constant, and is often capable of being reset to zero. See also sequence counter.
CPU (Central Processing Unit)
Same as central processor.
CRT
-See cathode ray tube.

A

AUERBACH

'"

(Contd. )

GLOSSARY

7:001. 109

cryogenics
The study and use of devices which utilize the properties assumed by materials at temperatures near
absolute zero. Note: Certain materials become
"superconductive" at very low temperatures; i. e. ,
their resistance falls to zero, so they can maintain
(i. e., store) a current indefinitely. Cryogenic techniques have found little practical application in computer design to date, but they represent a promising
area for research and development.
cybernetics
The science of exploring analogies between organic
and machine processes. Emphasis is upon comparative study of control and communication In machines
and in the nervous systems of animals and man.
cycle
(1) An interval of time or space in which one set of
events or phenomena is completed. (2) A set of
operations that is repeated regularly in thc same sequence; the operations may be subject to variations
during eac h repeti tion.
cycle time
The minimum time interval between the starts of
successive accesses to a storage location. Contrast
with access time. For example, if it takes 2 microseconds to read a word out of a core storage unit
and 3 more microseconds to rewrite the word before
another read operation can be initiatL>'Tam computers. Indexing can greatly simplify programming
by facilitating the handling of loops, arrays, and
other repetitive processes. Some computers have
many index registers, some have only one, and
others have none.
indexed address
An address that will be or has been modified by addition or subtraction of the contents of an index
register.
-indicator
(1) A device that can be set into a prescribed state,
often according to the results of a previous process,
and which can subsequently be used by a control unit
to determine a selection from alternative processes;
e. g., an overflow indicator is set whenever an overflow occurs. (2) A device (e. g., a lamp) that informs
an operator of the existence of a particular condition;
e. g., power on, stacker full, hopper empty.
indirect address
An address that specifies a storage location that contains either a direct address (i. e., an address that
specifies the location of an operand) or another indirect address. Note: Indirect addressing (also
called "multilevel addressing") is a form of address
modification possible in many, but not all, digital
computers; it can Simplify programming and increase
execution speeds in certain applications by permitting
the effective addresses of many instructions to be
modified by changing the contents of a single storage
location.

information retrieval
The methods, procedures, and equipment for recovering specific information from stored data, especially from collections of documents or other
graphic records.

5/67

EDP REPORTS

information theory
A branch of mathematics that is concerned with the
factors affecting the transmission of information,
such as transmission rate, channel Width, noise,
distortion, and the probabilities of errors.
initialize
To set the variable items of a process at initial
values before the process is started; e. g., to set
counters, indicators, and addresses to the appropriate starting values at the beginning or other prescribed points of a computer program.
in-line subroutine
Same as open subroutine.

index

information
The meaning that humans assign to data by means of
the known conventions used in its representation.

STAND~RD

input

--0)

The process of transferring data from external
storage or peripheral equipment to internal storage
(e. g., from punched cards or magnetic tape to core
storage). (2) Data that is transferred by an input
process. (3) Pertaining to an input process (e. g. ,
input channel, input medium). (4) To perform an
input process. (5) A signal received by a device or
component. Note: As the above definitions indicate,
"input" is the general term applied to any technique,
device, or medium used to enter data into data processing equipment, and also to the data so entered.

input area
An internal storage area used for the receipt of input
data which is transmitted as the immediate result of
execution of an input instruction.
input-output
A general term for the techniques, devices, and
media used to communicate with data processing
equipment and for the data involved in these communications. Depending upon the context, the term
may mean either "input and Jutput" or "input or
output." Synonymous with I O.
input-output channel
A channel that transmits input data to, or output data
from, a computer. Note: Usually a given channel
can transmit data to or from only one peripheral device at a time. However, some current computers
have multiplexor channels, each of which can service
a number of Simultaneously operating peripheral
devices.
input-output control system
See IOCS.
inquiry station
An input-output device that permits a human operator
to interrogate a computer system and receive prompt
replies in a convenient form. Note: Frequently, the
inquiries are entered from a keyboard and the
computer-generated replies are typed and/or displayed. Inquiry stations may be located remotely
from the computer. An airline reservation system,
for example, usually includes multiple inquiry stations in widely scattered locations.
inscribe
To read the data recorded on a document (e. g., a
check) and write the same data on the same document,
but in a form that makes the document suitable for
proceSSing by automatic character recognition equipment.

A

AUERBACH

'"

(Contd. )

GL.OSSARY

7;001.117

instruction
A set of characters that specifies an operation to be
performed and, usually, the values or locations of
one or more of its operands. Note: In this context,
the term instruction is preferable to the terms command and order, which are sometimes used
-synonymously.

interface
A shared boundary; e. g., the boundary between two
systems) or between a computer and one of its
peripheral devices.

instruction code
Same as operation code.

interleave
(1) To assign successive addresses to locations separatcd physically or in time by other locations. (2) To
allocate digits to storage cells on a ~ so that cells
allocated to successive digits of a particular word are
separated by a specific number of intermediate cells
which may be allocated Similarly to the digits of
other words.

instruction counter
Same as sequence counter.
instruction format
The allocation of the characters comprising an instruction between the component parts of the instruction (e. g., the operation part and one or more
address parts). See also address format.
instruction register
A register that stores the current instruction of a
computerTs program so that it can be interpreted
by the control unit.
instruction repertoire
The set of all the different types of instructions that
can be executed by a particular computer or used in
a particular programming language. Synonymous
with "instruction repertory" and "instruction set. "
instruction time
See execution time.
integrated circuit
A complete, complex electronic circuit, capable of
performing all the functions of a conventional circuit
containing numerous discrete tranSistors, diodes,
capacitors, and/or reSistors, all of whose component
parts are fabricated and assembled in a single integrated process. The resultant assembly cannot be
disassembled without destroying it. Note: Integrated
circuits, now coming into widespread use in commercially available computers, promise dramatic
improvements in speed, economy, reliability, and
compactness.
integrated configuration
A computer configuration in which input-output functions such as card reading and printing are performed
by peripheral equipment connected di.rectly to the
central processor. Contrast with paired configuration.
integrated data processing.
See IDP.
interblock gap
The distance between the end of one block and the
. beginning of the next block on a magn~ tape. The
tape can be stopped and brought up to normal speed
again in this distance, and no reading or writing is
permitted in the interblock gap because the tape
speed may be changing. Synonymous with interrecord gap and record gap (but use of these two
terms is not recommended because of the important
distinction between blocks and records).
interchangeable-cartridge storage
A storage device that uses cartridges which can be
conveniently removed from the device and replaced
by other similar cartridges.

interference
Same as demand on processor.

interlock
A protective facility that prevents one devicc or operation from interfering with another; e. g., by locking
the keys o[ a console typewriter to prevent manual
entry of data while the computer is transferring data
to the typewriter.
internal storage
A storage device that is permanently linked to a computer and directly controlled by it; e. g., core storage
and drum storage. Contrast with external
storage.
interpret
(1) To translate, explain, or tell the meaning of.
(2) To print on a punched card the data already
punched in the card.
interpreter
(1) A punched card machine that is capable of sensing
the data punched into a card and printing it on the
card. (2) Same as interpretive routine.
interpretive routine
A routine that deals with the execution of a program
by translating each instruction of the source language
into a sequence of machine instructions and executing
them before translating the next instruction. Thus,
each instruction must be translated every time it is
to be executed - an inherently inefficient process.
See also simulator.
inter-record gap
Same as inter block gap.
interrupt
A Signal, condition, or event that causes an interruption; e. g., completion of an input or output operation,
detection of incorrect parity, or an attempt to execute
an illegal instruction or to write in a protected location.
interruption
A temporary suspension of the execution of a sequence
of instructions as a result of the occurrence of some
prescribed event or condition. Note: The interrupt
usually triggers an unconditional transfer to a predetermined location, where a special routine (usually
part of an operating system) determines the cause of
the interruption, takes the appropriate action, and
then transfers control back to the point where the
program was interrupted - or, in some cases, to
another program of higher priority. Effective interruption facilities are a vital ingredient of computers
that are to operate in a multiprogramming or realtime mode.
--

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

5/67

AUERBACH STANDARD EDP REPORTS

7:001. 118

I/o

symbols may be combined into a meaningful communication. Note: An unambiguous language used to
express computer programs is called a programming
language.

--Same as input-output.
lOCS nput/Output Control System)
A standar routine or set 0 routines designed to initiate and control the input and output processes of a
computer system, thereby making it unnecessary for
users to prepare detailed coding for these processes.
item
----Xn arbitrary quantity of data that is treated as a unit.
Note: a record, in turn,i"S"a collection of related
items, while a file is a collection of related records.
Thus, in payroll processing, an employee's pay rate
forms an item, all of the items relating to one employee form a record, and the complete set of employee recorcls forms a file.

latency
Same as waiting time.
lateral parity check
Synonymous with row parity check.
leader
--uf A blank or unused length of tape at the beginning
of a reel of tape. (2) A record that precedes a group
of "detail records" and contains data about the group
which is not contained in the individual detail records.
letter

iterative
Pertaining to a process in which a sequence of operations is executed repeatedly until some condition is
satisfied (e. g., until all items have been processed,
or until a certain variable reaches a specified value).
Note: In computer programs, iterative processes are
normally implemented by means of loops.

J
~
A unit of work for a data processing system, especially from the standpoint of installation scheduling and
accounting.
jump

--X departure from the normal sequence of executing
instructions in a computer. See also conditional
transfer and unconditional transfer.
justify
--ufTo adjust the position of words on a printed page
so that the left-hand or right-hand margin is regular.
(2) By extenSion, to shift an item in a register so
that the most or least significant digit is at some
specified position in the register.

K
key
--One or more characters associated with a particular
item or record and used to identify that item or
record, especially in sorting Qr collating operations.
Note: The key mayor may not be attached to the
record or item it identifies. Contrast label and tag.

library
---xnDrganized collection of information for study and
reference purposes. See also program library.
library routine
A tested routine that is maintained in a program
library (in contrast to a routine written especially
for a particular job).
line printer
A printer that prints all the characters comprising
one line during each cycle of its action. Synonymous
with "line-at-a-time printer." Note: Two widely
used types of line printers are chain printers and
drum printers.
linear programming
An operations research technique that involves locating the maximum or minimum of a linear function of
variables which are subject to linear constraints and
inequalities. Note: Linear programming (often abbreviated "LP") is useful for solving certain problems
involving many variables whose optimum values must
be found (e. g. , many distribution, blending, and
resource allocation problems).

list
-(1) An ordered set of items.

(2) To print the items
and records that comprise a file or the instructions
that comprise a program.

L
label
---;;: name that is attached to or written alongside the
entity it identifies; e. g., a key that is attached to the
item or record it identifies;or a name written alongside a state:ffient on a coding sheet.

5/67

of sounds in a spoken language; in English, one of the
26 characters A through Z.

linkage
Coding that connects two separately-coded routines;
e. g., the coding that links a subroutine to the program
with which it is to be used. See also calling sequence.

keypunch
A keyboard-actuated card punch. The punching in
each column is determined by the key depressed by
the operator.

language
A defined set of symbols and of rules or conventions
governing the manner and sequence in which the

---xn alphabetic character used for the representation

listing
A printed list of the instructions or statements that
comprise a program.
literal
----WIn a programming language, an item whose representation in characters remains essentially unaltered
during the operation of the appropriate compiler.
(2) In a machine language, a numeral that is embedded within an instruction and used directly as an
operand of that instruction.

fA

AUERBACH

"

(Contd. )

GLOSSARY

7:001. 119

load
--(1) The quantity of data transferred in a single input

or output operation. (2) To read a program into
internal storage in preparation for its execution.
(3) To insert a supply of an input or output medium
(e. g. , punched cards or a reel of magnetic tape) into
II peripheral device.
load-and-go
An operating technique in which the loading and execution phases of a program are performed in one continuous run. The "loading" phase frequently includes
performance of the functions of an assembler, compiler, or generator. Note: The load-and-go technique is especially effective when a program must be
compiled or generated for a one-time application,
such as the production of a flpecial report.
loader
------x6ervice routine designed to read programs into
internal storage in preparation for their execution.
location
(1) A part of a store which can be explicitly and
uniquely specified by means of an address, and which
holds a word or part of a word. (2) Loosely, any
place in which data can be stored.
lockout

~ inhibition of all or certain types of references

to a particular part of a computer system (e. g., a
magnetic tape unit, or certain areas of core storage).
Lockout may be effected by means of either instructions or manual switches. Note: A "write lockout"
inhibits writing in specific areas of storage while
permitting reading of data stored in those areas. See
also storage protection.
log
-A record of the operations of data processing equipment, which lists each job or run, the time it required, operator action~and other pertinent data.
logical operation
(1) An operation whose ~rands and result are single
digits. (2) By extension, an operation with operands
and result of any length in which each digit of the result depends on not more than one digit of anyone
operand. Usually the same operation is performed
on all corresponding digits of the operands. The
most common logical operations are AND, exclusive
OR, inclusive OR, NOR, and NOT.
logical record
Same as record; contrast with physical record, which
is synonymous with block.
longitudinal parity check
A parity check performed on the bits in each track of
magnetic tape or punched tape. At the end of each
block, the parity bits that have been generated for
each of the tracks are recorded simultaneously in the
form of a "longitudinal check character," which is
regenerated and checked when the block is read.
Synonymous with track parity check.
look-up
---seetable look-up.
loop

-----x sequence of instructions that can be executed repetitively, usually with modified addresses or modified data values. Each repetition is called a cycle.

Cycling continues until a specified critenon is satisfied (c. g., until a countcr reaches a predetermined
value). Note: The use of loops greatly facilitates the
coding of any iterative process.
low-order
---pertaining to the digit or (ltgits or a number that have
the least weight or SIgnificance; e. g., in the number
53276, the low-order digit is 6. Contrast with highorder.
--

M
machine address
Same as absolute address.
machine instruction
An instruction that a computer can directly recognize
and execute.
machine language
A ~!1guagc that is used directly by a computer. Thus,
a "machinc language program" is a set of instructions
which a computer can directly recognize and execute,
and which will cause it to perform a particular
process.
machine oriented la~
A language in which there is a general (though not
necessarily strict) one-to-one correspondence between the statements of the source program and the
instructions of the ~bje~~.am (which will normally be a machine language program ready for execution on a particular computer). Note: The input
to an assembler is usually expressed in a machine
oriented language. Contrast with process oriented
language.
machine-readable
Pertaining to data represented in a form that can be
sensed by a data processing machine (e. g., by a card
reader, magnetic tape unit, or optical character
reader).
machine word
Same as word (i. e., a group of bits or characters
treated as a unit and capable of being stored in one
storage cell).
macro instruction
An instruction written in a machine oriented language
that has no equivalent operation in the computer, and
is replaced in the object program by a predetermined
set of machine instructions. Note: Macro instruction
facilities can ease the task of coding in a machine
oriented language by precluding the need for detailed
coding of input and output operations, blocking, format
control, checking for errors, etc.
magazine
See cartridge, hopper, stacker; "magazine" is sometimes used as a synonym for any of these three terms.
magnetic card
A thin, flexible card with a magnetic surface upon
which data can be stored. Note: Some large-capacity
auxiliarystorage devices use a large number of magnetic cards, contained in interchangeable cartridges.
One card at a time is extracted from the cartridge,
transported to a read/write station where data is
read and/or recorded, and then returned to the
cartridge.

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

5/67

7:001. 120

AUERBACH STANDARD EDP REPORTS

magnetic core
A small piece of magnetic material, often toroidal in
shape (i. e., doughnut-shaped), whose magnetic properties make it suitable for storing one bit of data.
See also core storage.

map

--X list that indicates the areas of storage occupied by
various elements of a program and its data.
mark
-same as flag.

magnetic ink character recognition
See MICR.

mark sensing
A technique for detecting pencil marks entered by
hand in prescribed places on punched cards or other
documents. The marked data may be converted into
punched holes in the same cards, recorded on another medium, or transmitted directly to a computer.

magnetic storage
A storage device that uses the magnetic properties of
materials to store data. Note: Most of the storage
devices currently used with computers fall into this
broad category. Magnetic storage embraces two distinct types of storage devices: those in which there is
relative movement between the heads and the magnetic
medium (e. g., drum storage an(f"(i"iSc storage), and
those in which no such movement occurs (e. g., ~
storage).

mask

-rmachine word containing a pattern of characters or
bits that is used to extract or select parts of other
machine words by controlling the retention or elimination of selected characters or bits.

mass storage
Same as auxiliary storage.

magnetiC tape
A tape with a magnetic surface on which data can be
stored by selective polarization of portioiiS'Of the surface. Note: The magnetic tape currently in widest
use with computers is made of a polyester plastic, is
one-half inch in Width, is supplied in 2400-foot reels
with a diameter of 10.5 inches, and is recorded with
7 or 9 tracks across the tape at a recording density
of 200, 556, 800, or 1600 rows per inch.

master file
A file containing relatively permanent information
which is used as a source of reference and (usually)
is periodically updated. Contrast with detail file.
~

(1) In mathematics, a two-dimensional rectangular
array of quantities that is manipulated according to
defined rules. (2) By extenSion, an array of any number of dimensions.

magnetic tape unit
A device, used to read data from or record data on
magnetic tape, that contains a tape transport mechanism, reading and writing heads, and associated
controls.
--

medium
Any agency or means for representing data; usually,
a material on which data is recorded. Note: Among
the most widely used media are punched cards,
punched tape, magnetic tape, and printed forms.

main frame
(1) Same as central processor. (2) That portion of a
computer system which is not considered peripheral
eqUipment.
main storage
Same as working storage.

memory
Same as store (i. e., a device into which data can be
inserted and retained, and from which the data can
be obtained at a later time).

maintenance
Tests, measurements, adjustments, repairs, and
replacements intended to keep equipment in satisfactory working order. Note: All maintenance can be
classified as either corrective maintenance or
prjlventive maintenance.

merge
To form a single sequenced file by combining two or
more similarly sequenced files. Note: Merging may
be performed manually, by a collator, or by a computer system for which a "merge routine" is available.
Repeated merging, splitting, and remerging of strings
of records can be used to arrange the records in sequence; this process. called a "merging sort, " is
frequently used as the basis for sorting operations on
computer systems.

malfunction
Same as fault.
management information system
A system designed to supply the managers of a business with the information they need to keep informed
of the current status of the business, to understand
its implications, and to make and implement the appropriate operating decisions.

message
An arbitrary amount of information (e. g., a group of
characters or words) that is transmitted as a unit.

mantissa
Same as fixed-point part.
manual input
(1) The entry of data into a device by manual means
at the time of processing. (2) Data entered into a device by manual means at the time of processing; e. g. ,
data entered by means of a keyboard, or by setting
switches, dials. or levers.

5/67

See also group mark, word mark.

message switching
A technique for controlling the traffic within a data
communications network that involves: the reception
of messages from various sources at a switching
center, the storage of each message until the proper
outgoing communications link is available, and the
ultimate retransmission of each message to its destination or destinations.

A

(Contd. )

AUERBACH
®

GLOSSARY

7: 001. 121

module

~An incremental block of storage or some other

microprogramming
A method of operation of the control unit of a computer in which each instruction, instead of being used
to initiate control signals directly, starts the execution of a sequence of "microinstructions" at a more
elementary level. The microinstructions are usually
stored in a special read-only storage unit. Note:
The instruction repertOire of a microprogrammed
computer can be altered to suit particular requirements by simply changing the stored microinstructions.
microsecond
One millionth of a second, abbreviated IJ.sec or IJ.S.
millisecond
One thousandth of a second, abbreviated msec or ms.
minimum-latency coding
A method of coding used for those computers (no
longer in common use) in which the waiting time for
a word in working storage depends upon its location;
locations for both instructions and operands are so
chosen that access times are reduced or minimized.
Synonymous with optimum coding.
misfeed
The failure of a punched card or other document to
pass through a machine in the prescribed manner.
mistake
The failure of a human to carry out an operation in
the required manner (e. g., in writing a program or
in operating equipment). Contrast with fault. See
also error.
mnemonic
Pertaining to a technique used to assist human memory. Note: Most symbolic assembly languages use
mnemonic operation codes, which are typically abbreviations such as MPY for multiply and SUB for
subtract.
mode

---0:-)

A system of data representation used in a computer; e. g., binary mode, decimal mode. (2) See
access mode.

modem (modulator-demodulator)
A device that provides the appropriate interface between a communications link and a data proceSSing
machine or system by serving as a modulator and/or
as a demodulator.
modify
To alter an instruction or address in a prescribed
way. See also address mo~on.
modulator
A device that receives electrical pulses, or bits,
from a data proceSSing machine and converts them
into signals suitable for transmission over a communications link. Contrast with demodulator.--

"building block" that can be used to expand the capacity of a computer system. (2) An interchangeable,
plug-in unit containing electronic components.
modulo N check
Same as residue check.
monitor
---;YO;)bserve the state of a system or the execution of
a program and indicate Significant departures from
the normal or expected conditions.
,

monitor routine
(1) A routine designed to indicate the progress of
work in a computer system. (2) Formerly, same as
executive routine.
Monte Carlo method
A trial-and-error technique of repeated calculations,
based on the concept of randomness, that can be used
to solve problems containing a large number of
variables with interrelationships so complex that a
straightforward analytical solution is impossible or
impractical.

multi -precision
Pertaining to the use of two or more computer words
to represent a number in order to gain increased
precision.
multiple address
Pertaining to an instruction containing more than one
address; e. g., one-plus-one address, two-address,
three-address.
multiplex
To transmit two or more messages simultaneously
over a single channel or other transmission facility.
This can be accomplished either by splitting the
channel's frequency band into two or more narrower
bands ("frequency-division multiplexing") or by interleaving the bits, characters, or words that make up
the various messages ("time-division multiplexing").
multiplexor
A device that makes it possible to transmit two or
more messages Simultaneously over a single channel
or other transmission facility.
--multiplexor channel
A special type of input-output channel that can transmit data between a computer and a number of simultaneously operating peripheral devic<)s.
multiprocessing
The simultaneous execution of two or more sequences
of instructions in a single computer system. This
may be accomplished through the use of either two or
more central processors (i. e., a multiprocessor
system) or a single central processor with several
instruction registers and several sequence counters.
Synonymous with parallel processing.
multiprocessor
Pertaining to a computer system that contains two or
more central processors.

© 1967 AUERBACH Corporation and AUERBACH Info. Inc.

5/67

AUERBACH STANDARD EDP REPORTS

7:001. 122

or both, are true. "P NOR Q" is often represented
by PVQ. (2) The logical operation that uses the NOR
operator.

multiprogramming
A teclmique for handling two or more independent
programs simultaneously by overlapping or interleaving their execution. The overlapping or interleaving of the execution of the various programs is
usually controlled by an operating system which attempts to optimize the overall performance of the
computer system in accordance with the priority requirements of the various jobs.

normalize
To adjust the exponent and fixed-point part of a number
in floating-point representation so that the new fixedpoint part lies within a prescribed standard range.

multi sequencing
The simultaneous execution of two or more parts of a
program by separate central processors.

N

NOT
--(1) A logical operator which has the property that if
P is a statement, then the NOT of P is true if P is
false, and false if P is true. "NOT P" is often represented by P, - P, or iP. (2) The logical operation that uses the NOT operator; also called
"negation. "

name

--p; word

or phrase that constitutes the distinctive designation of an entity and is generally used in referring
to that entity; e. g., a person's name, or a symbol
used to identify a particular data item.
---

nanosecond
One billionth of a second (i. e., 10- 9 second), abbreviated nsec or ns.
NDRO
Same as nondestructive readout.
nest
--(1) To embed a structure (such as a subroutine or
block of data) within another structure of the same
form. (2) To evaluate a polynomial of the Nth degree
by an algorithm that consists of (N-1) multiply operations and (N-1) add operations in succession.
ninety-column card
A punched card containing 45 vertical columns and 12
rows. Each column is divided into an upper and a
lower half, and each half-column is capable of holding
one character. Thus, the card is logically equivalent
to a 90-column card, which accounts for its name.
Note: In each half-column, the digits 0, 1, 3, 5, 7,
or 9 can be represented by a single punched hole in the
appropriate pos ition; the digits 2, 4, 6, or 8 or other
characters are represented by a combination of two
or more punched holes. The popularity of 90-column
cards has been declining steadily in recent years.
Contrast with cighty-column card.
noise

--W Random variations of one or more characteristics
of any entity such as voltage, current, or data. (2)
Loosely, any disturbance that tends to interfere with
the normal operation of a device or system.

number
~A mathematical entity that may indicate a quantity
or amount of units. (2) Loosely, a numeral.
number system
A system for the representation of numbers according
to an agreed set of rules. Note: All number systems
used in data processing utilize "radix notation, "
which means that there is a fixed ratio between the
significance of each digit position and the significance
of the previous digit position. This ratio is called the
radix or base of the number system, and the significances of successive digit pOSitions are successive
integral powers of the radix. For example, in the
decimal number system, the radix is 10, and the
numeral 5762 means:
(5 x 10 3) + (7 x 10 2) + (6 x 10 1) + (2 x 10 0).
The decimal number system is generally used by
humans, whereas computers frequently employ the
binary (radix 2), octal (radix 8), decimal (radix 10),
and hexadecimal (radix 16) number systems.
numeral
A representation of a number, usually by means of
one or more digits. - - numerical analysis
The study of methods of obtaining useful quantitative
solutions to problems that have been expressed
mathematically, including the study of the errors
and bounds on errors in obtaining such solutions.
numerical control
The automatic control of operations (such as those of
milling or boriIig machines) wherein the control is
applied at discrete points in the operation through
proper interpretation of numerical data. Contrast
with process control.

nondestructive readout
A reading process that does not erase the data which
has been read. Contrast with destructive readout.

o

nonerasable storage
A storage device or medium whose contents are not
erasable; i. e. , the stored data can only be changed
by replacing the storage medium with new medium
bearing the new data. Contrast with erasable storage.
See also read-only storage.
NOR
--::\ logical operator which has the property that if P and
Q are statements, then the NOR of P and Q is true
if both P and Q are false, and false if either P or Q,

5/67

object language
A language that is an output from a translation
process. Contrast with source language.
object program
A program expressed in an object language (e. g. , a
machine language program that can be directly executed by a particular computer.
OCR (Optical Character Recognition)
The automatic reading by machine of graphic characters through use of light-sensitive devices.
(Contd. )

A

AUERBACH

'"

GLOSSARY

7:001.123

octal
"Pertaining to the number system with a radix of eight,
or to a characteristic or property involving a choice
or condition in which there are eight possibilities.
Note: Octal numerals are frequently used as a "shorthand" representation for binary numerals, with each
octal digit representing a group of three bits (binary
digits); e. g. , the binary numeral 110 101010 can be
represented as octal 652.
odd-even check
Same as parity check.
odd parity
See parity bit.
off-line
Pertaining to equipment or devices which are not in
direct communication with the central processor of
a computer system. Contrast with on-line. Note:
Off-line devices cannot be controlled by a computer
except through human intervention.
one-address
Pertaining to an address format in which each instruction contains one address part, which normally
specifies the location of an operand.
one-plus-one
Pertaining to an address format in which each instruction contains two address parts, one of which
normally specifies the location of an operand while
the other (the ''Plus-one'' address) specifies the
location of the next instruction to be executed in the
normal sequence. Contrast with two-address. Note:
One-plus-one addressing was commonly used in computers which used magnetic drums for working storage.
on-line
Pertaining to equipment or devices which are in
direct communication with the central processor of
a computer system. Contrast with off-line. Note:
On-line devices are usually under the direct control
of the computer with which they are in communication.
on-the-fly printer
A printer in which the type remains in motion during
the printing process; at the appropriate instants
during its movement, the paper and type are forced
together to cause the desired characters to be printed.
Note: Most high-speed chain printers and drum
printers are of the on-the-fly type.
open-ended
Pertaining to a process or system that can conveniently
be augmented or improved.
open shop
A computer installation that may be programmed and
operated by any qualified employee of the organization.
Contrast with closed shop.
open subroutine
A subroutine that must be inserted directly into a
program at each point where it is to be used. Synonymous with in-line subroutine. Contrast with
closed subroutine.
operand
A unit of data upon which an operation is performed.
Note: The operand of a computer instruction may also

be an equipment item such as an indicator, switch,
or peripheral device.
operating environment
A collective term for all of the facilities that contribute
to the efficient and convenient execution of programs in
a computer system.
operating system
An organizeJ collection oft routines and procedures
for operating a computer. These routines and procedures will normally perform some or all of the
following functions: (1) Scheduling, loading, initiating,
and supervising the execution of programs. (2) Allocating storage, input-output units, and other facilities
of the computer system. (3) Initiating and controlling
input-output operations. (4) Handling errors and
restarts. (5) Coordinating communications between
the human operator and the computer system. (6)
Maintaining a log of system operations. (7) Controlling operations in a multiprogramming, multiproceSSing, or time-sharing mode. Note: Among the
facilities frequently included within an operating system
are an executive routine, a scheduler, an 10CS, utility
routines, and monitor routines.
operation
(1) A general term for any well-defined action. (2) The
derivation of a unit of data (the "result") from one or
more given units of data (the "operands") according to
rules that completely specify the result for any permissible combination of values of the operands. (3) A
program step undertaken or executed by a computer
(e. g., addition, multiplication, comparison, shift,
transfer).
operation code
A code used to represent the specific operations of a
computer.
operations research
The use of analytical techniques to solve operational
problems in order to provide management with a
sound, logical basis for making decisions and predictions. Among the common techniques of operations
research are linear programming, Monte Carlo
methods, information theory, and queueing theory.
operator
(1) A person who operates a machine. (2) A symbol
that indicates an action to be performed on one or
more operands (e. g. , the logical operators AND
and OR).
optical character recognition
See OCR.
optical scanner
A device that scans printed or written data, using
optical techniques, and converts the data into digital
representation.
optimum coding
Same as minimum-latency coding.
OR
-See exclusive OR and inclusive OR.
Note: When OR is used without qualification,
"inclusive OR" is implied.
order

-W

To arrange items in a specified sequence.
Loosely, an instruction.

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

(2)

5/67

AUERBACH STANDARD EDP REPORTS

7:001. 124

origin
Same as base address.

~

A segment of a program or data, usually of fixed
length, that has a fixed virtual address but can in
fact reside in any region of the computer's working
storage. Note: The division of every program and
its data into pages can facilitate the control of timesharing operations by permitting straightforward
"swapping" of pages belonging to various programs
between working storage and auxiliary storage.

output

~The

process of transferring data from internal
storage to external storage or to peripheral equipment
(e. g. , from core storage to magnetic tape or a printer).
(2) Data that is transferred by an output process. (3)
Pertaining to an output process (e. g. , output channel,
output medium). (4) To perform an output process.
(5) A signat transmitted from a device or component.
Note: As t e above definitions indicate, "output" is
the general term applied to any technique, device, or
medium used to take data out of data processing equipment, and also to the data so transferred.

output area
An internal storage area used for the release of
output data; the area occupied by output data at the
time when execution of an output instruction is
initiated.
overflow
In an arithmetic operation, the generation of a quantity
beyond the capacity of the register or storage location
which is to receive the result.
overhead
A collective term for the factors which cause the performance of a device or program to be lower then it
would be in the ideal case; e. g. , the start and stop
times which can cause a magnetic tape unit's effective
speed to be far lower than its rated speed; and the
time and storage space required by an operating system to perform its functions.
overlay
To transfer segments of programs from auxiliary
storage into working storage for execution, so that
two or more segments occupy the same working storage locations at different times. Note: This technique
makes it possible to execute programs which are too
large to fit into the computer's working storage at one
time; it is also of great importance in multiprogramming and time-sharing operations.
overpunch
To change the data represented in a punched card
column or punched tape row by punching one or more
additional holes into the column or row.

page printer
(1) A printer in which the pattern of characters for an
entire page is determined prior to printing. Synonymous with "page-at-a-time printer." (2) A
widely-used but misleading term for teleprinters
(i. e. , the character-at-a-time printers commonly
used in low-speed communications networks).
paired configuration
A computer configuration that includes two central
processors: a "main" processor all of whose input
and output is from and to magnetic tape (or some
other high-speed medium), and a "satellite" processor
equipped to perform the necessary data transcription
functions (e. g. , punched cards to magnetic tape,
magnetic tape to printer).
paper tape
Same as punched tape.
parallel
Dealing with the elements of a word or message
(e. g. , the bits or characters) simultaneously, each
element in a different device. Contrast with serial.
parallel processing
Same as mUltiprocessing.
parameter
A variable that is assigned a constant value for a
particular purpose or process; e. g. , the re-order
level for a particular item in an inventory control
program, the matrix size in a generalized matrix
inversion program, the record length in a sort program generator.
parity bit
A bit (binary digit) that is appended to an array of
bits to make the sum of all the "1" bits in the array
either always even ("even parity") or always odd
("odd parity"). For' example:
Even Parity
Odd Parity
o 1 1
0
1
1
o 1 0
0
1
0
o 1 0
0
1
0
Data bits
0
1
1
0
1
1

own coding
Coding supplied by the user that causes a generalized
program to perform a function tailored to the user's
specific needs; e. g. , coding which alters the output
format of a manufacturer-supplied sort routine to
conform with a user's file format.

o

p

Parity bit

pack
---oro store several short units of data in a single storage
cell in such a way that the individual units can later be
recovered; e. g. , to store two 4-bit BCD digits in one
8-bit storage location or one magnetic tape row.
packing density
Same as. recording density.
padding
Dummy characters, items, or records used to fill
out a fixed-length block of information.
5/67

A

1

1
1

1
0

0
1

1
1

1
0

1

0

1

0

1

0

parity check
A check that tests whether the number of "1" bits in
an array is either even ("even parity check") or odd
("odd parity check"). Synonymous with odd-even
check. See also row parity check and longitudinal
parity check.
pass
-one complete cycle of input, processing, and output
in the execution of a computer program. For example,
a "one-pass compiler" reads the source program,
(Contd.)

AUERBACH

co

7: 001. 125

GL.OSSARY

compiles it, and writes the object program without
intermediate input-output operations or human
intervention.
patch
To correct or modify a program in a rough or expedient way by adding new sections of coding.
pattern recognition
The identification of shapes, forms, configurations,
or sounds by automatic means; e. g. , optical character
recognition (OCR), machine recognition of human
speech.
peak speed
The maximum instantaneous speed which a device
is capable of achieving when no allowances are made
for factors such as start times, stop times, interblock gaps, etc. This is the speed usually quoted in
manufacturers' specifications, but it may differ
substantially from the device's effective speed in
typical applications.
perforated tape
Same as punched tape.
performance
The execution of the functions required of a device or
system; the degree of speed or effectiveness with
which these required functions are carried out. See
also system performance.
peripheral equipment
All of the input-output units and auxiliary storage
units of a computer system. Note: The central
processor and its associated working storage and
control units are the only parts of a computer system
which are not considered peripheral equipment.
physical characteristics
The dimensions, weight, heat dissipation, and electrical power requirements of each unit of a computer
system.
physical record
Same as block; contrast with logical record. Note:
To avoid the need for distinguishing between physical
records and logical records, use of the alternative
terms "block" and "record", respectively, is
recommended.
picosecond
One thousandth of a nanosecond (1. e. , 10- 12 second),
abbreviated psec.
pinboard
A perforated board used to control the operation of
some automatic data processing equipment through
manual insertion of cordless pins in the appropriate
holes. See also plugboard.
pitch
The distance between corresponding points of
adjacent characters, rows, tracks, etc. ; e. g. , most
high-speed line printers have a character pitch (1. e. ,
horizontal spacing) of 10 characters per inch and a
line pitch (i. e. , vertical spacing) of 6 or 8 lines per
inch.
PL/I (Programming Language I)
A process oriented language designed to facilitate the
preparation of computer programs to perform both

business and scientific functions. Note: Developed
jointly by IBM and the SHARE users' organization
between 1964 and 1966, PL/I represents an attempt
to combine the best features of existing programming
languages (such as ALGOL, COBOL, and FORTRAN)
with a number of facilities not available in previous
languages. However, it has not yet been demonstrated
that an efficient compiler for the PL/I language can
be developed, and PLj I has to date made only limited
inroads upon the popularity of other programming
languages.
plotter
A device that produces a graphical representation of a
dependent variable, as a function of one or more other
variables, by means of an automatically controlled
pen or pencil. See also XY plotter.
plugboard
A perforated board used to control the operation of
some automatic data processing equipment. The holes
in the board (called "hubs" or "sockets") are manually
interconnected, in a manner appropriate to the job to
be performed, by means of wires terminating in
plugs (called "patchcords"). Synonymous with control
panel (2). See also pinboard.
pocket
Same as stacker.
postmortem routine
A diagnostic routine, often a dump, that is used after
a program has failed to operate as intended.
precision
The degree of discrimination with which a quantity is
stated. For example, a three-decimal-digit numeral
permits discrimination among 1000 possible values.
Precision should be carefully distinguished from
accuracy, which is the degree of freedom from error.
For example, a 6 -digit numeral is more precise than
a 4-digit numeral, but a properly computed 4-digit
result may be more accurate than an improperly
computed 6-digit result.
preset
--""Pertaining to a condition or variable whose value is
established prior to the initiation of a !:!!!l.
presumptive address
An address that is altered through address modification to form an effective address which is actually
used to identify an operand.
preventive maintenance
Maintenance that is carried out to keep equipment
in proper operating condition and to prevent faults
from occurring during subsequent operationSContrast with corrective maintenance.
printer
A machine that produces a printed record of the data
with which it is fed, usually in the form of discrete
graphic characters that can be conveniently read by
humans. See also chain printer, drum printer, line
printer, page printer.
print pOSition
In a line printer, a position in which anyone of the
members of the printer's character set can be
printed in each line. Note: Most of the current line

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

5/67

-7:001. 126

AUERBACH STANDARD EDP REPORTS

printers have between 80 and 160 print positions;
i. e. , they can print between 80 and 160 characters
per line.
priority
A preferential rating that specifies the relative urgency
or importance of a particular job or task. Note: In
some operating systems, the entry oIaliigh-priority
job can cause immediate suspension of the processing
of jobs of lower priority.
privileged instruction
A computer instruction that is not available for use in
ordinary programs written by users; its use is restricted to the routines of the operating system. Note:
Input-output, priority control, and storage protection
instructions are in the "privileged" category in many
of the current computers.
problem oriented facilities
A collective term for the standard software other than
assemblers, compilers, and operating systems that is
available for a particular computer system. Included
are utility routines (such as simulators, sort and
merge routines, report program generators, data
transcription routines, and file maintenance routines)
plus application packages and problem oriented
languages.
problem oriented language
A language whose design is oriented toward the specification of a particular class of problems, such as
numerical control of machine tools. Contrast with
process oriented language.
procedure
The course of action taken to solve a problem
procedure oriented language
Same as process oriented language
process
A system of operations designed to solve a problem
or lead to a particular result.
process control
The automatic regulation of a process (such as the
production of chemicals or the generation of power)
wherein the control is applied continuously and adjustments are made to keep the values of one or more
controlled variables (such as temperature or flow
rate) constant. Contrast with numerical control.
process oriented language
A language designed to permit convenient specification, in terms of procedural or algorithmic steps,
of data processing or computational processes.
Examples include ALGOL, COBOL, and FORTRAN.
Contrast with probJ:eillOriented1ailguage and machine
oriented language.
processor
A device or system capable of performing operations
upon data Note: The term may refer to either
hardware (see central processor) or software (an
assembler or compiler is sometimes referred to as
a "language processor").

program
(1) A plan for solving a problem. (2) To devise a
plan for solving a problem. (3) A computer routine;
i. e. , a set of instructions arranged in properse.::-quence to cause a computer to perform a particular
process. (4) To write a computer routine.
program compatibility
The characteristic that enables one computer system
to execute programs written for another computer
system and obtain identical results. See also compatibility. Note: Program compatibility can beachieved through the use of two computer systems with
similar instruction repertoires and facilities; or between dissimilar computers - through emulators,
simulators, translators, or coding in a common
language.
program interrupt
See interrupt.
program library
An organized collectIon of tested programs, together
with sufficient documentation to permit their use by
users other than their authors.
program step
A single instruction or operation in a program.
programmed check
A check that is carried out by a series of instructions
in a program. Contrast with automatic check"
programmer
A person who devises programs. Note: The term
"programmer" is most suitably applied to a person
who is mainly involved in formulating programs.
particularly at the level of flowchart preparation. A
person mainly involved in the definition of problems
is called an analyst, while a person mainly involved
in converting programs into coding suitable for entry
into a computer system is called a coder. In many
organizations. all three of these functions are performed by "programmers."
programming language
An unambiguous language used to express programs
for a computer.
protected location
A location whose contents are protected against
accidental or improper alteration. See also storage
protection.
pseudocode
A programming language whose instructions are not
directly executable by a computer.
pseudo instruction
An instruction that has the same general form as a
machine instruction but is not directly executable by
a computer. Pseudo instructions are commonly used
in machine oriented languages to control the operation
of a translator. Synonymous with directive.
pulse

---P: sudden, significant change of short duration in the
processor demand
Same as demand on processor.

5/67

value of some variable (most commonly the voltage
in an electrical circuit).
(Contd. )

A

AUERBACH


VIm

s

~
.£
~iZ

.!--o

.1
§os
r;~

~~

~~

~ ~

~E

.1

§OS

~'i:

~~

.!--o

~

" '"
~~

Q~

-

-

2,700
4,350

.~~

-

-

-

r1S2

13,975
22,220

-

8,712 9,351 14,402 12,421
13,845 15,569 18,915 17,425
19,000
- 20,400 7,125
19,780
-

491/492
494
1004
1050
1107

1/69

til

]

-

UNIVAC
UNIVAC
UNIVAC
UNIVAC
UNIVAC

-

·a'cution

C'

:=

Times, jispc (G

Digits :'v1 in.
Precision)

a +b
Floating
Point F>,('cution

Lil,~l,.

c =- ah

TinlC's, I-Lsec

c

= a/b

Checking of Dat.:l Transfers

I/o

CEl\TIL\L
PI{OCESSOI{

Program Interrupt Facility

sensing

Number of Index i{egisters

Indir('ct AddrC'ssing

bpecia! Echtin!;

AND, INC OR, EXCOR

AND, INC OR, EXC OR

AND, INC OR, EXC OR

None

None

No

C"pabiliti~s

Boolean Operations
Table Look-up
Console T\1'e\\'1'iter

Yes

Yes

Yes

1 integrated nonsimultaneous channel

I integrated non-simultaneous channel; buffered
I/O units are available

8 input and 8 output

Two magnetic tape
operations can be performed
simultaneously

Program-compatible with
Honeywell 400

H-800 and 1800 are programcompatible; each can run up
to 8 progTams concurrently

402

1402

802 (H-800)

1802 (H-1800)

Input-Output Channels

Features and C0111ments

Model Number

Core

Core

Core

Core

1,024

4,096

4,096

8,192

I\'Iinimum

4,096

32,768

2S,672

G5,536

Maximum

49,152

393,216

345,144

786,432

Decimal Digits

32,768

262,144

229,376

524,288

Characters

9. 25 per 24 bits

6.5 per 24 bits

6

2

210,000

310,000

533,333

1,600,000

Parity

Parity

Parity

Parity

None

None

Tvpe of Storage
Number of \Vords

I

IVla.xinlum

Total Storage
Cycle Time, /-Lsec

Effective Transfer Hate, char/sec

None

A 256-word control memory
is also utilized

WORKIi\G
STORAGE

Checking
Storage Protection

Features and Comments

*

\Vith optlOnal equipment.
(s) Using subroutine.

01969 AUERBACH CorporatIOn and AUERBACH Info. Inc

1/69

COMPQRliO~1

11 :210.118

System Identity

Word Length

Model 25

Binary Bits

8 per byte

8 per byte

8 per byte

8 per byte

Decimal Digits

2 per byte

2 per byte

2 per byte

2 per byte

Characters

DATA
STRUCTURE

1 per byte

Fraction Size

-----

Exponent Size

---

Radix
Floating Point
Representation

2020
~dls 1&2 Mdls 3&4

Model Number
Arithmetic Radix

Decimal

Operand Length, Words

Variable

Instruction Length, Words

2,4, or 6 bytes

Addresses per Instruction

0, lor 2

Likely Fixed
Point Execution
Times, J,lsec (5
Digits Min.
Precision)
Likely Floating
Point Execution
TImes, Msec

Model 40

1 per byte

1 per byte

1 per byte

Binary

Binary

Binary

24 or 56 bits

24 or 56 bits

24 or 56 bits

7 bits

7 bits

7 bits

2025
Binary or
decimal

2030
Binary
(decimal*)

2040
Binary
(decimal*)

Variable
2, 4, or
6 bytes

Variable
2, 4, or
6 bytes

Variable
2, 4, or
6 bytes
0, I, or 2

I

0, 1, or 2

0, I, or 2

c :::: a + b

675

1,207

113 or 182

78 or 96

36 or 64

c = ab

7,000

7,530

616 or 645

296 or 395

113 or 178

c=a/b

10,810

11,340

805 or 1,308

481 or 767

216 or 349

c = a +b

---

---

303 or 369*

107 or 161*

43 or 62*

c = ab

---

---

730\ or I,

295 or 874

105 or 294'

---

154~

664 or 1,839*

350 or 1,717

157 or 511

Checking of Data Transfers

Parity

Parity

Parity

Parity

Program Interrupt Facility

Yes, I/o
only

Yes, 5
classes

Yes, 5

classes

Yes, 5
classes

c = alb
CENTRAL
PROCESSOB

IDM System/360
Model 30

Model 20

Number of Index Registers

8 max.

16 max.

16 max.

16 max.

IndIrect Addressing

None

None

None

None

Special Editing Capabilities

Good

Good

Good

Good

AND, INC OR,
EXCOR

AND, INC OR,
EXC OR

AND, INC OR,
EXC'OR

Boolean Operations

AND, INC OR

Table Look-up ..

None

None

None

None

Console Tvpewriter

None

Yes

Optional

Optional

Integrated channels
permit sharing of

1 selector
channel or 1

o to 2

o to

Input-Output Channels

Features and Comments
Model Number
Type of Storage

core storage

models
Limited program
compatibility with
other System/360
models

mh~~;e,\eXor
ch

selector
channels; 1

:;;,~~;e,llexor

2 selector
channels; 1 multiplexor channel

These models have a high degree of program compatibility

2020

2025

2030

2040

Core

Core

Core

Core

8,192 bytes

16,384 bytes

l\Iinimum

4,096 bytes

16,384 bytes

Maximum

16,384 bytes

49, 152 bytes

65, 536 bytes

262,144 bytes

Decimal Digits

32,768

98,304

131,072

524,288

Characters

Number of \Vords

Maximum
Towl Storage
WORKING
STORAGE

16,384

49,152

65,536

262,144

Cycle TIme, J,Lsec

3. 6 per half-byte

0.9perlor
2 bytes

1. 5 per 1
byte

~Je~er

Effective Transfer Rate, char/sec

62,500 max.

185,000 max.

321,000 max.

390,000 max.

Checking

Parity

Parity

Parity

Paritv

Storage Protection

None

Write only*

Write only*

Write onlv*

Features and Comments

Certain I/O units can
be connected without
their usual control
units or channels

Wlth optional equipment.
(s) l'sing subroutine.

1/69

A

AUERBAC~

GW

2

0

RTS

11 210119

CENTRAL PROCESSORS AND WORKING STORAGE

Model 44

Model 50

IBM System/360
Model 65

System Identity

Model 67

32 + 4 parity

8 pcr bytc

8 per bytc

8 per byte

Binary Bits

9.2

2 per byte

2 per byte

2 pel' bytc

Decimal Digits

4

1 per byte

1 pCI' byle

1 pcr bylc

Characters

Word Length
DATA
STRUCTURE

Radix

Binary

Binary

BInary

Binary

24. 32, 40, 48 or 56 b,ts

24 or 56 bits

24 or 56 bits

24 or 56 bits

Fraction Size

7 b,ls

7 b1ls

7 bits

7 bits

Exponent Size

2044

2050

2065

2067

Bmary (decllnnl *)

BInary or
deC1l11nl

Bmuryor
decimal

Bmary,

1 or 1/2 word

Vanable

1 or 1/2 word

2, 4, or
6 byles

Floating Point
Representation

Model Number
dCCl111Ul

Arithmetic Radix

Vanable

Variable

2, 4, 01'
6 byles

2, 4, or 6 bytcs

Instruction Length, Words

Operand Length, Words

Addresses per Instruction

2

0, 1, or 2

0, 1, or 2

0, 1, or 2

13.0; 7.0*

12 or 35

3.5 or 9.0

4.201'9.7

26.3; 20.5*

40 or 86

7.001'32

7.7 or 3:3

41.0; 33.8*

44 or 97

11 or 47

12 or 48

18.8 or 11. 6*

14 or 21

4.701'4. 8

5.401'5.5

73.6 or 21. 8*

29 or 49

6.1 or 9.7

6.8 or 10.4

137.5 or 31. 0

30 or 81

9.3 or 16

10.0 or 16.9

Parity

Parity

Panty

Panty

Checking of Data Transfers

Yes, 5 classes

Program Interrupt Facility

Yes, 5 classes
16

Yes, 5

Yes, 5

classes

classes

16 max.

16 max.

25 max.

None

Yes; 8 register
USSOCIUtl ve memory

c

~

c

a +b
~

Likely Fixed
Point Execution
Times, J.,Lsec (5
Digits Min.
PreciSion)

ab

c ~ alb
c

=a +b
c

~

Likely Floating
Point Execution
Times, iJ.sec

ab

c ~ alb

Number of Index Registers
Indirect Addressing

None

None

RestrlCted

Good

Good

Good

AND, INC OR, EXC OR

AND, INC OR
EXC on

AND, INC OR,
EXC OR

AND, INC 01{,
EXC on

None

None

None

None

Standard

Optional

Opbonal

Optional

1 multiplexor channel
with 64 subchannels; 1*
or 2* hi~l-speed multiolexor c annels

o to

o to

1 or 2 Channel Conlrolle1's; up 10 7 channels
per controller

Input- Output Channels

SpeCIal hardware facil1tales hme-sharing
opcratlOns

Features and Comments

3 seleclor
channels; 1
mulhplexor
ch"nncl

6 selector
channels; 0 or
1 mu~~tcxor

Special Editing Capabilities
Boolean Operations
"

Table Look-up

Console Typewriter

~h"nn

Limited program compatibility with other
~stem/360 models
2044

2050

2065

2067

Model Number

Core

Core

Core

Core

Type of Storage

8,192 4-byte words

65,536 bytes

131,072 bytes

262,144 bytes

Minimum

65,536 4-byte words

524,288
b'tes

1,048,576
bytes

2,097,152

Maximum

524,288

524,288

2, 0~7, 152

4,194,304

Decimal Digits

262,144

262,144

1,048,576

2,097,152

Characters

1. 0 per 4 -byte
word

2.0 per 4
bytes
851,000
max.

0.75 per 8
bvtes4,760,000
m"x.

121,200 max.

Number of Words

0.75 per 8
4,760,000 max.

Parity

Parity

Parity

Write only

Read and write

Read and write

Interleaving
improves
sequenhal
access rate

1 or 2 central procesSOl'S and 1 to 8 independent 262K modules
per syslem

Maximum
Total Storage
Cycle Time,

hvtPR'

Read* and write*
Standard genernl registel'S are in extended
core storage; HighSpeed Registers are
optional

CENTRAL
PROCESSOH

~sec

Effective Transfer Rate, char/sec

WORKING
STOIlAGE

Checking

Parily
Sto~age

Protection

Features and Comments

*

With optlOnal equipment.
(s) Using subroutine.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc

1/69

11

210130

COMPARISON CHARTS

Svstem Identity

Word Length

Model 75

8 per byte

6 + parity + word mark 6 + parity + word mark

Decimal Digits

2 per byte

2 per byte

1

1

1 per byte

1 per byte

1

1

Binary

Binary

Decimal

Decimal

Fraction Size

24 or 56 bits

24,56, or 112 bits

8 dIgltS(S)

8 diglts(s)

Exponent Size

7 bIts

7 bits

2 digIts(s)

2 diglts(s)

Model Number

2075

2085

1401

1411

Arithmetic Hadi"

Bmaryor
deCImal

Binary, Decimal

Decimal

Decimal

Operand 1.ength, Words

VarIablc

Variable

1 to N char

1 to N char

Instruction Length, Words

2, 4, or
6 bytes

2,4. or 6 bytes

1 to 8 char

1 to 12 char

Addresses per Instruction

LiI.pd
Point E'\:C'cution

=:lb

C

~S('C

TinH..'s,

(5

l)igit~

9,930 (8)

c

(s)

?>.1in.
Precision)

alb

c

-------

124 (short); 116 (long)

Parity

Parity

Check ing of Data Transfers

Yes; 4 typos

Yes; 4 types

Program Interrupt Facility

63

63

~umber

None

c

30 (short); 52 (long)

=

~l

Ii)

c

189 (short); 752 (long)

~

rt/b

Yes; up to 5 levels
Good

None

Optional

None

Scan instructions

Optional

Sk15, 000

Peak Speed, bits/sec

Handles large number
of remote termmals

64 lines

Features and Comments

Model Number

B 9351

Model Number

2,000

Capacity, char

MultI-statlOn unit

DATA
COMMUNICATIONS
CONTROLLER

CRT
DISPLAY

Features and Comments
Model Number
Peak Speed, pOints/sec

PLOTTER

Features and Comments

Model Number
Name

OTHER
INPUTOUTPUT
DEVICES

Features and Comments
*With optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc

1/69

COMPARISON CHARTS

11 :240.104

CDC 160 & 160-A

System Identity
Model Number

,.1

CDC 1604 & 1604-A

CDC 3100, 3300, & 3500

1612 G

166G-2

1612 G

505

501

501

505

3152

1

1

24

24

24

S/ch

S/ch

l/ch

Single Spacing

1000

150

1000

500

1000

1000

500

150

I-inch Spacing

500

130

500

375

571

571

375

150

1

0.2

2. S Max.

2. S Max.

2.S Max.

<0.1

<0.1

<0.1

120

120

120

136

136

136

136

120

64

64

64

64

64

64

64

64

Checking

None

None

Echo

Echo

Echo

Features and Comments

Higher speeds possible
when restricted character sets are used.

Maximum Number On-Line

Speed,
lines/min

Demands on Processor, %
PRINTED
OUTPUT
Number of Print Positions
Character Set Size

None

Echo

Echo

Increased speed is possible
with restricted character
set; Dual-channel controller
provided with 501 and 505.

Dual channel
controller
provided

Model Number
MICR
READER

Peak Speed, documents/min
Features and Comments

OPTICAL
CHARACTER
READER

Model Number

Page Reader

Peak Speed, documents/min

370 char/sec

915

Features and Comments

DATA
COMMUNICATIONS
CONTROLLER

CRT
DISPLAY

PLOTTER

Model Number

3266-A

3274

3275

Peak Speed, bits/sec

2400/
lme

2,500,000

230.400

Features and Comments

Multilme

Multiple lines,
1 at a time

Model Number

210

217

Capacity, char

1,000

1,000

Features and Comments

Multistation

Single
station

Model Number

3293

Peak Speed, points/ sec

300

Features and Comments

Incremental

Model Number
OTHER
INPUTOUTPUT
DEVICES

Name
Features and Comments

*With optional equipment.

1/69

A.

AUERBACH

PRINTERS AND SPECIALIZED INPUT-OUTPUT

11 :240.105

CDC 3'400, 3600, & 3S00

System Identity

CDC 6000 Series

501

505

501

S/ch

S/ch

8/ch

1000

500

1000

Model Number

512

Maximum Number On-Line

8/ch

Sing Ie Spacing

1200

Speed,
lines/min
I-inch Spacing

571

685

<0.1

0

0

Demands on Processor, %

136

136

136

Number of Print Positions

64

64

48

Character Set Size

Echo

Echo

571

375

PRINTED
OUTPUT

Checking

Echo

Dnal channel
controller provided

Features and Comments

Model Number
Peak Speed, documents/min

MICR
READER

Features and Comments
Model Number

915 Page Reader

Peak Speed, documents/min

370 char/sec

OPTICAL
CHARACTER
READER

Features and Comments
3266-A

3274

3275

2400/
line

2,500,000 230,400

Multiline

Multiple lines
1 at a time

210

217

Mode I Number

1,000

1,000

Capacity, char

MultistatIon

Single
station

3266-A

3276

6671

2400/
line

2400/
lme

2400/
line

Multi-line
controllers

6673
40,800

6674

6676

Model Number

40,SOO

110

Peak Speed, bits/sec

Handles multiple lines
1 at a time

DATA
COMMUNICATIONS
CONTROLLER

Features and Comments

CRT
DISPLAY

Features and Comments

3293

Model Number

300 steps/sec

Peak Speed, paints / sec

Incremental

Features and Comments

PLOTTER

Model Number

6411 & 6416
I/O Buffer & Control
Doubles I/O Capability

Name

OTHER
IN PUTOUTPUT
DEVICES

Features and Comments
*Wlth optional equipment.

fodels
I, 2

1403

Maximum Number On-Line

1

1

2

2

2

2

150 (430
numeric)

240 (600
numeric)

600

1100

600

1100

132

196

480

750

480

750

Demands on Processor, %

94 or 0.6*

90 or 1. 0*

0.3 to O. 84

0.73

0.7 max.

1.3 max.

Number of Print Positions

120 or 144*

100 or 132

132

100 or 132

132

Character Set Size

13,39,52 or 63 characters*

48

48

Checking

Echo

Echo, validity

Echo, \'alidity

Features and Comments

Interchangeable
horizontal
typebar

Model Number

1009

7750

1009

Peak Speed, bits/sec

2,400

22,400

2,400

Independent computer; up to 112 lines

Uses 4-of-8 code;
synchronous

Single Spacing
Speed,
lines/min
I-inch Spacing

PRIXTED
OCTPCT

3

1403
1 2

1403 Arodel 3

Model Number
MICR
READER

Peak Speed, documents/min
Features and Comments
Model Number

OPTICAL
CHARACTER
READER

Peak Speed, documents/min
Features and Comments

DATA
COMMUNICATIONS
CONTROLLER

4-of8 code

Features and Comments
Model Number
CRT
DISPLAY

Capacity, char
Features and Comments

PLOTTER

Model Number

1627

Peak Speed, points/ sec

200 or 300

Features and Comments

Incremental

Model Number
OTHER
INPUTOUTPUT
DEVICES

Name
Features and Comments

*With optional equipment.

1/69

fA.

AUERBACH

PRINTERS AND SPECIALIZED INPUT·OUTPUT DEVICES

11 :240.115

m:\l7070.7072. & 7074

m"'17080

IBM 7090 & 7094

7400

717

720

716

3

10

10

1

150

150

500

75 to 150

150

150

400

75 to 150

1.5 max.

100

100

< 1

120

120

120

120

48

48

48

48

"alidit"

Echo

Echo

Programmed echo

System Identity
Model Number
Maximum Number On-Line

Single Spacing
Speed,
lines/min
I-inch Spacing

I

:\ot usable with 7072

Demands on Processor, o/c
PRINTED
OUTPUT
Number of Print Positions
Character Set Size
Checking

"'fuximum of 72 characters
per print cycle
Features and Comments

Model Number
Peak Speed, documents/min

MICR
READER

Features and Comments
Model Number
Peak Speed, documents/min

OPTICAL
CHARACTER
READER

Features and Comments
1009

7750

1009

2,400

22,400

2,400

4-of-8
code

Independent
computer

Uses 4-of-8 code;
synchronous

Model Number
Peak Speed, bits/sec

DATA
COMMUNICATIONS
CONTROLLER

Features and Comments
Model Number
Capacity, char

CRT
DISPLAY

Features and Comments
Model Number
Peak Speed, pOints/sec

PWTTER

Features and Comments
Model Number
Name

OTHER
INPUTOUTPUT
DEVICES

Features and Comments
*With optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

1/69

COMPARISON CHARTS

11:240.116

Model Number

640102

Maximum Number On-Line

a/trunk

640210

640200

640300

4

4

690/
940

650/
805

805

1000

407

400

400

520

1.4 Max.

81 Max.

81 Max.

1.5 Max.

120

120

120

120 or
132

Validity

300/
450

820/
1180

820/
1180

?

Demands on Processor, %

8 Max.

17 Max.

21 Max.

?

Number of Print Positions

132

132

160

132

Character Set Size

64 or
51

64 or
51

64 or
51

Up to
128

Checking

Echo

Validity

One 640-102 is "integrated" into every Century
100 system: one 640-200, 210 or 300 is
"integrated" into every Century 200 system:
other 'hrinters are connected via buffer controis: igher speeds are for all-numeric printin~

340-512 can operate as a 24-position numeric lister at 1850 Ipm.

Validity

Validity

Listed speeds
are based on
use of a restricted 42character
set

Model 340644 can function as a2000
lpm numeric

lister

Model Number

670-101

671-101

402-3, 402-4

407-1

Peak Speed, documents/min

600

1,200

750

1,200

Usable off-line

Usable off-line

Model Number

420-2 Optical Reader

420-2 Optical Reader

Peak Speed, documents/min

1,664 char/sec

1,664 char/sec

Features and Comments

Reads journal tapes

Reads journal tapes

Model Number

621-101

621-201

321-1(315); 327-3 (315-RMC)

Peak Speed, bits/sec

50,000/
line

50,000/
line

40, 800/line

Features and Comments

Controls up
to 15 lines

Controls up
to 255 lines

Controls up to 99 lines

Model Number

795

Capacity, char

256, 512, or 1,024

Features and Comments

Extensive editing controls

Peak Speed, pOints/ sec

Model Number
Name
Features and Comments
*With optional equipment.

1/69

4

I-inch Spacing

Features and Comments

OTHER
INPUTOUTPUT
DEVICES

4
600/
1200

Model Number
PLOTTER

340-601,
-632,-644

1500/
3000

Features and Comments

CRT
DISPLAY

340-503

1500/
3000

Features and Comments

DATA
COMMUNICATIONS
CONTROLLER

340-502
340-512

450/
900

PRINTED
OUTPUT

OPTICAL
CHARACTER
READER

340-3

Single Spacing
Speed,
lines/min

MICR
READER

NCR 315, 315-100, & 315-RMC

NCR Century Series

System Identity

A.

AUERBACH

PRINTERS AND SPECIALIZED INPUT-OUTPUT DEVICES

RCA SPECTRA 70
70/242

11:240.117

System Identity

RCA 301
70/243

70/248

l/Trunk

333

335

2

2

Model Number
Maximum Number On-Line

625

1,250

600

800 to 1,000

835 to 1,075

Single Spacing

508

715

480

500

572

I-inch Spacing

85/22*

84/32*

Demands on Processor, %

160

Number of Print Positions

Varies

132/160*

132

132

120

64

64

4B

47; 64

Timing

Timing

None

None

"Quietized"

Versions are
available with
96-character
print drums

Can print
on punched
cards

1000 Ipm with
47 character
character sets.

versions are
available

Speed
lines/min

PRINTED
OUTPUT

Character Set Size
Checking

Features and Comments

Model Number

Burroughs BI02

Peak Speed, documents/min

1,560

MICR
READER

Features and Comments
70/251

5820 VIDEOSCAN

1,300

1,500

Model Number
Peak Speed, documents/min

OPTICAL
CHARACTER
READER

Features and Comments
70/627

70/653

70/668

378 CMC

320,000
bytes/sec

5100
char/sec

6000
bytes/sec

2,400 per line

Direct computer
line

Single line

48 lines

Up to 80 lines

Model Number
Peak Speed, bits/ sec

DATA
COMMUNICATIONS
CONTROLLER

Features and Comments

r

70/ 52 Video Data Terminal

Model Number

1,080

Capacity, char

Single-station unit; can be multiplexed
via 70/755 Video Data Switch

CRT
DISPLAY

Features and Comments
Model Number
Peak Speed, POints/sec

PLOTTER

Features and Comments
70/510

70/630

328

338

Voice Response Unit

Data Gathering System

Jnterro~ating

TypewrIter

Monitor
Printer

IB9-word
vocabulary

120 char/sec

10 char/sec

10 char/sec

Model Number
Name

OTHER
IN PUTOUTPUT
DEVICES

Features and Comments

*With optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

1/69

COMPARISON CHARTS

11-:240.118

RCA 3301

System "Identity

333

335

4152

7912

Maximum Number On-Line

2

2

8

1

Single Spacing

800

800

700/922

600

I-inch Spacing

540

540

480/670

430

Demands on Processor, %

<0.1

<0.1

0.16 to 0.22

14

Number of Print Positions

120

160

128

100 to 130

Character Set Size

64

54

57

Checking

None

None

Echo

1,000 Ipm with restricted
(47 characters) character
set

No form
control loop

Features and Comments

No form
control loop

PRINTED
OUTPUT

Model Number
Peak Speed, documents/min
Features and Comments
Model Number
OPTICAL
CHARACTER
READER

Peak Speed, documents/min
Features and Comments

DATA
COMMUNICATIONS
CONTROLLER

Model Number

3376

3377

3378

Peak Speed, bits/sec

40,800

276,000

2,400/
line

Features and Comments

Single
line

Computer link

Up td 160
lines

Model Number
CRT
DISPLAY

Capacity, char
Features and Comments
Model Nnmber

PWTTER

Peak Speed, pOints/sec
Features and Comments
Model Number

OTHER
INPUTOUTPUT
DEVICES

Name
Features and Comments

*With optional equipment.

1/69

UNIVAC S. S. 80/90

Model Number

Speed,
lines/min

MICR
READER

UNIVAC ill

A

·

AUERBACH

PRINTERS AND SPECIALIZED INPUT-OUTPUT DEVICES

UNIVAC 418 Series

11:240_119

System Identity

UNIVAC 490 Series

Model Number

0755

0758

0751,0755,8121

0758-00

16

8

l/ch

l/ch

400

700/922 Numeric

700/922

1600

Single Spacing

900

I-inch Spacing

340

472/563 Numeric

472/484

0_2toO.7

0.2toO.7

0.051 to 0.41

Maximum Number On-Line

Speed
lines/min

Demands on Processor, %
PRINTED
OUTPUT

132

132

63

63

None

None

Number of Print Positions
Character Set Size
Checking

Higher speeds can be
obtained with restricted
character set

Features and Comments

Model Number
Peak Speed, documents/min

MICR
READER

Features and Comments
Model Number
Peak Speed, documents/min

OPTICAL
CHARACTER
READER

Features and Comments
CTMC

WTS

CTS

CTMC

WTS

CTS

Up to 4,800
per line

40,800

40,800

Up to 4,800
per line

40,800

40,800

Up to 32
lines

Synchronous

Asynchronous

Up to 32
lines

Synchronous

Asynchronous

Model Number
Peak Speed, bits/sec

DATA
COMMUNICATIONS
CONTROLLER

Features and Comments
Model Number
Capacity, char

CRT
DISPLAY

Features and Comments
Model Number
Peak Speed, pOints/sec

PLOTTER

Features and Comments
Model Number
Name

OTHER
INPUTOUTPUT
DEVICES

Features and Comments
*With optional equipment.

© 1969 AUE RBACH Corporation and AUE RBACH Info. Inc_

1/69

COMPARISON CHARTS

11 :240.120

UNIVAC 1004

System Identity
Model Number

1004 I

Maximum Number On-Line

1

UNIVAC 1050
1004 II, III

0755-01

UNIVAC 1107
0755-02

4 or 8

7400

7418

15

Single Spacing

400

600

600/750

70,0/922

700 to 922

600

I-inch Spacing

340

380

422

468

475

424

0.7 max.

0.12 max.

0.09

Speed,
lines/min

Demands on Processor, %

100

0.6 max.

Number of Print Positions

132

128

100 to 130

128

Chara~ter

63

63

63

51

None

Validity

None

Higher rates
obtainable with
restricted character sets

No form control loop

PRINTED
OUTPUT

Set Size

Checking

Features and Comments

Model Number
MICR
READER

Peak Speed, documents/min
Features and Comments
Model Number

OPTICAL
CHARACTER
READER

Peak Speed, documents/min
Features and Comments

DATA
COMMUNICATIONS
CONTROLLER

Model Number

DLT Series

Standard Communications
Subsystem

Standard Communications
Subsystem

Peak Speed, bits/sec

Varies widely;
many models

Up to 4,800 per line

Up to 4,800 per line

Features and Comments

Single-line

Up to 32 lines

Up to 32 lines

Model Number
CRT
DISPLAY

Capacity, char
Features and Comments
Model Number

PLOTTER

Peak Speed, pOints/sec
Features and Comments
Model Number

OTHER
INPUTOUTPUT
DEVICES

Name
Features and Comments

*With optional equipment.

1/69

fA.

AUERBAC~

PRINTERS AND SPECIALIZED INPUT-OUTPUT DEVICES

UNIVAC 1108
7299-03

0758-00

,

4/channel

11:240.121

UNIVAC 9200 & 9300

UNIVAC 9400

3030-00

3030-02

768-00

1

1

7

System Identl ty
Model Number

768-99

Maximum Number On-Line

1200/1600

700/922

250/500*

600/1200*

900/1100

1200/1600

Single Spacing

800/834

472/484

220

451

652/670

800/810

1-inch Spacing

0.025 max.

13

31

<1.0

Demands on Processor, %

132

96/132*

120/132*

132

Number of Print Positions

63

63

63

63

None

Echo

Echo

Echo

For use on
9200

For use on
9300

Speed
lines/min

C~acter

PRINTED
OUTPUT

Set Size

Checking

Features and Comments

Model Number
Peak Speed, documents/min

MICR
READER

Features and Comments
Model Number
Peak Speed, documents/min

OPTICAL
CHARACTER
READER

Features and Comments
CTMC

WTS

CTS

DCS-1

DCS-1

4,800
per line

40,800

40,800

50,000

50,000

230,400 per line

Peak Speed, bits/sec

Single-line

Singleline

For 4 or 16 lines

Features and Comments

32 lines

Synchro- Asynnous
chronous

DCS-4, -16

Model Number
DATA
COMMUNICATIONS
CONTROLLER

Mode I Number
Capacity, char

CRT
DISPLAY

Features and Comments
Model Number
Peak Speed, points/sec

PLOTTER

Features and Comments
Model Number
Name

OTHER
INPUTOUTPUT
DEVICES

Features and Comments
*Wlth optlonal equipment.

V 1969 AUERBACH Corporation and AUERBACH Info,lnc.

1/69

•• :.,UU.IUU

A

COMPARISON CHARTS
SOFTWARE

STANDARD

EDP

AUERBACH

REPDRIS

SOFTWARE COMPARISON CHARTS
INTRODUCTION
The charts on the following pages show the
principal software facilities available from the
manufacturers for use on nearly 100 U. S.manufactured digital computer systems. These
charts, arranged in alphabetical order by manufacturer, enable you to make direct comparisons of the type and extent of software support
f~cilities furnished by the manufacturers of
competitive computers. Moreover, the charts

provide valuable indications of the age of each
computer system and the type of circuitry it
employs.
In the Software Comparison Charts, a "bullseye" (large black dot) denotes the availability
or use of a particular facility, while a blank
space denotes its absence. Explanations of the
specific chart entries follow.

SYSTEM CHARACTERISTICS
Identity

Manufacturer and model number of the computer system.

Date of First Customer
Delivery ,

Month and year in which the first successful
installation was made or is scheduled to be
made.

Solid state

Uses electronic components whose operation
depends on the control of electric or magnetic
phenomena in solids (e. g., transistors, crystal
diodes, ferrite cores), as distinguished from
the earlier vacuum-tube technology.

Integrated Circuits

Uses complete, miniaturized electronic circuits, all of whose component parts are fabricated and assembled in a single integrated
process, so that the resultant assembly cannot
be disassembled without destroying it.
LANGUAGE PROCESSORS

These are specialized computer routines
which translate programs written in languages
designed for programming convenience into
machine-Iangu~ programs suitable for execution by computers.
Language processors can be grouped into two
major categories, assemblers and compilers,
depending upon the type of source language

1/69

they accept as input. An assembler accepts
programs written in a symbolic code that is
closely related to the computer's own machine
language. A compiler accepts programs
written in a "process oriented language" such
as COBOL or FORTRAN, which permits convenient specification of data processing or
computational processes in terms of procedural or algorithmic steps rather than specific
computer operations.

Assembler

Assembles programs written in a symbolic
language that is similar to machine language
but simpler and more meaningful, thereby
greatly reducing the human effort required to
prepare and debug programs.

ALGOL Compiler

Compiles programs written in ALGOl, an
international language designed for convenient
expression of computational procedures.
ALGOL is very popular in Europe but is not
as widely used in the United States.

COBOL Compiler

Compiles programs written in COBOL, the
COmmon Business Oriented Language designed
in 1959 and accepted as a USA Standard. COBOL
uses English-like procedural statements and is
by far the most widely used process oriented
language for business applications.

A

AUERBACH

~

11:300.101

SOFTWARE

FORTRAN Compiler

Compiles programs written in FORTRAN, a
language designed to facilitate the preparation
of scientific programs through the use of expressions and symbols similar to those of
algebra. FORTRAN has been accepted as a
USA Standard language in two versions
(FORTRAN and Basic FORTRAN), and is by
far the most popular scientific programming
language.

PL/I Compiler

Compiles programs written in PL/I, a multipurpose language developed jointly by IBM and
the SHARE users' organization between 1964
and 1966. PL/I represents an attempt to combine the best features of ALGOL, COBOL, and
FORTRAN with a number of facilities not available in previous languages.
OPERATING SYSTEMS

An operating system is an organized collection
of routines and/or procedures for operating a
computer. It will normally handle some or
'all of the following functions: (1) scheduling,
loading, initiating, and supervising the execution of programs; (2) allocating storage, inputoutput units, and other facilities of the computer system; (3) initiating and controlling

input-output operations; (4) handling error
conditions and restarts; (5) coordinating communications between operator and computer;
(6) maintaining a log of system operations;
and (7) controlling operations in a multiprogramming, multiprocessing, time-sharing, or
data communications mode.

Tape Operating System

Resides on magnetic tape and performs some or
all of functions (1) through (6) above; randomaccess storage devices are not required.

Disc Operating System

Resides on a random-access storage medium
(disc, drum, or magnetic strip) and performs
some or all of functions (1) through (6) above;
usually more efficient than an equivalent tape
operating system.

Multiprogramming

Support for handling two or more independent
programs simultaneously by overlapping or
interleaving their execution.

Multiprocessing

Support for controlling the simultaneous execution of two or more sequences of instructions
in a single computer system, usually through
the use of two or more central processors.

Time-Sharing

Support for furnishing computing services to
multiple simultaneous users at remote
terminals, while providing rapid responses
to each of the users.

Data Communications

Support for controlling the transmission of
digital data between the computer site and one
or more remote locations, usually via a communications medium such as a telephone,
telegraph, or microwave circuit.

In addition a column is included to indicate the

presence or absense of a Report Generator, a
routine that constructs programs, based upon

problem parameters supplied as input, to perform routine report-writing functions.

© 1969 AUERB~CH Corporation and AUERBACH Info, Inc.

1/69

11 :300.102

COMPARISON CHARTS

SOFTWARE COMPARISON CHART

OPERATING
SYSTEMS

LANGUAGE
PROCESSORS

SYSTEM
CHARACTERISTICS

a

J.I

~
~

;l::

...

.!:f
u

III

.!:f

::eo ~

s;::::
as

Gl

00
Burroughs
Burroughs
Burroughs
Burroughs
Burroughs

B 100
B 200
B 300
B 2500
B3500

12/63
10/62
5/65
5/67
5/67

Burroughs B 5500
Burroughs B 6500
Burroughs B 7500
CDC 160/160A
CDC 1604/1604A

12/64
3/69
3/69
5/60
1/60

CDC 3100
CDC 3300
CDC 3500
CDC 3400
CDC 3600

1/65
3/66
7/68
11/64
6/63

CDC 3800
CDC 6400
CDC 6500
CDC 6600
GE-115

2/66
3/66
7/67
8/64
12/65

GE-130
GE-200 Series
GE-400 Series
GE-600 Series
Honeywell 110

4/69
4/61
5/64
4/65
8/68

Honeywell
Honeywell
Honeywell
Honeywell
Honeywell

120
125
200
1200
1250

2/66
3/68
7/64
1/66
7/68

Honeywell
Honeywell
Honeywell
Honeywell
Honeywell

2200
4200
8200
400
1400

12/65
3/68
6/68
12/61

Honeywell 800
Honeywell 1800
mM 360, Model 20
mM 360, Model 25
mM 360, Model 30

1960
1963
12/65
1/68
5/65

mM
mM
mM
mM
mM

4/65
7/66
8/65
11/65
10/66

360,
360,
360,
360,
360,

Model
Model
Model
Model
Model

40
44
50
65
67

~

00

:sl
'0
00

"CI

!

t'o

~

I<

.SI

'S

Gl
III
III

<

•••
•••
•• •• ••
•• •
•• • •
•
•
••
•••
•••
••
••
•
•
••
••
•
•

•••
• •
••• •• •••
••
••
•• • ••
•• • ••
•
•
••
•
••
•
•
••
••
•
••
••
•

·(1)
·(1)
·(1)

.(1)
.(1)

.(1)

.(1)
.(1)

I<

I<

a-0

.!!l
.....

.!!l

6

8
IDENTITY

S

I<

.!!l
....

III

i

0

a-0

u

S

~

Z

u

u
..:l

0

t!l
..:l

~

i5r..

u

<

•••
•
••
•

••
••
•••
•
••
•••

••
•••
••
••
••
•
•••
•
••
••

£.
I<

.!!l

i

.~
~

a

S

£.
.~
':;j

I<

u

~

~

~

....

••

••
••
••

0

....
Po<

••

E-<

~
C)

III

0

•

•
•• ••
•• •
•

i
~

J.I

.a.
~

~

.~
III
III

8
0

~

~

•••
•
••

bD

.S

~
I

.§E-<

•
•
• •

•

••
•

••
••
••

••
•
••
•• ••
•
• ••
•• ••
•
• •
•
•• ••
•••
••• ••• •••
• ••
•• •• ••
••• •
•• •• ••
•• ••
• • •
•• •• ••
•• ••
•
•• ••
••
• •
•
•• •
••
••
• •• •• •• •• •• ••
• • • • • •• •
•• •• ••• •• •• ••• •••
•

.~

':;j

••
•

••

••
••
•••
••
••
•

••
•

••

oS

I
0

u

as

~

0

~

':;j
I<

til

~

t:

~

~

••
••
•••

••
••
•

•••
••
••
•••
••
•••
•••
••
••
•
•

•
•

•••

••
••
•••
••
•••
••
••
•• •••
• •
•• ••
• • •

(1) mM uses "hybrid" electronic circuitry in the System/360; this is a compromise between solid-state and integrated circuitry.

1/69

A

AUERBACH

e

(Cont'd)

11:300.103

SOFTWARE

SOFTWARE COMPARISON CHART (CaNT'O)

\

~

~

fI.l

:',:l

fI.l

8....

5
is

~

~

fI.l
~

....~!>.~
o ~
.2l;:::

~

"d

Sh

Q)

:s.....

QQ

[/J

os

S
.2l

~

.s
IDENTITY

OPERATING
SYSTEMS

LANGUAGE
PROCESSORS

SYSTEM
CHARACTERISTICS

mM 360, Model 75
mM 360, Model 85
mM 1401
mM 1410
mM 1440

11/61
4/63

mM 1460
mM 1620 Model 1
mM 1620 Model 2
mM 7010
mM 7040

10/63
9/60
10/62
10/63
7/62

mM 7044
mM 7070
mM 7072
mM 7074
mM7080

7/62
3/60
6/62
11/61
8/61

mM7090
mM 7094
NCR 315
NCR 315-100
NCR 315 RMC

7/62
10/62
2/62
12/64
9/65

NCR
NCR
RCA
RCA
RCA

Century 100
Century 200
Spectra 70/15
Spectra 70/25
Spectra 70/35

9/68
3/69
9/65
12/65
2/27

RCA
RCA
RCA
RCA
RCA

Spectra 70/45
Spectra 70/46
Spectra 70/55
301
3301

2/66
1/69
9/66
2/61
7/64

2/66

9/60

UNIVAC
UNIVAC
UNIVAC
UNIVAC
UNIVAC

m
SS 80/90 I
SS 80/90 IT
4181
418IT

UNIVAC
UNIVAC
UNIVAC
UNIVAC
UNIVAC

418m
490
491/492
494
1004

6/69
12/61
12/65
6/66
1/63

UNIVAC
UNIVAC
UNIVAC
UNIVAC
UNIVAC

1050
1107
1108
9200
9300

6/63
9/62
12/65
6/67
9/67

UNIVAC 9400

6/69

8/62
8/58
6/62
6/63
?/64

~
0

~

~

.(1)

••
•
•

•

••
••
•••
••
••
•
•••
••

••
•••
•
•

••
••
•
••
•

••
•
•••
••

••
•

~

.!II
.....

0

0

S'

..sI

Co)

!RfI.l

d

-a

...:l
0
...:l

-<

-<

•
••
••
••
•
••
••
•••
•••
•

••

••
•

••
••
••

•

Co)

...:l
0

~
Co)

•
••
••
•
•

•
••
•••
•

••
•
••
••
••
•

S'

••
••
•
•••
••

•
••
•

••
••
•

~

.!II
.....

.!II

]:
0

Co)

z

-<
~
~
r...

••
••

~
~

.!II
.....

S'

••

•

~

[/J

r:::

bO

bO

:cl

os
~

:cl

Q)

r:::

os

~

~

fI.l

.§

S
os

Sh
0

p..

....fI.l

~

Q

~

•
•

••

•
•

••

Co)

....

~

<:.)

........6

bO

8~
Eo<

0

.~

.~
fI.l
fI.l

80
~

.S<

~

.~
1a

••
•
••
•••
•
••
••
••
•
••
••
•
•
••

•
••

•
•
•

.sos

S
0

Q)

.~
Co)

.Eo<...~

~
Q

I

•
•••
••

•
••
••

•
•
•••

•
••

•••
••
•
••
•
.(2)
.(2)
.(2)
.(2)
.(2)
.(2)

•

.(2)
.(2)

•

•

•
•••
•
•
•

•
••
•
•••

•
•

••
••
•
•
••
•
•

•

•
•

••

••
•••
•
•
•••
••
•

It

••

•
••

•
•

••
•
•
•

~

~

~

•
••

•• ••
• •
•
•••
•

S
.2l

~

~

d

1::0

P..

Q)

~

•
•
•
•
••
••
••
•
•
•
••
••
•
••
••
•
II

•
•
••
•

(2) Drum oriented.

© 1969 AUERBAC,H, Corporation and AUERBACH Info, Inc.

1/69

11 :400.100

4

COMPARISON CHARTS
SYSTEM PERFORMANCE

STANDARD

EDP

AUERBACH

REPORTS

SYSTEM PERFORMANCE COMPARISON CHARTS
INTRODUCTION
These unique charts list the total processing
times for five standard "benchmark" problems
which represent typical computer workloads in
both business and scientific applications. Each
line on the System Performance charts shows
the cost and calculated performance of a particular computer system arranged in a particular standard equipment configuration. (The
standard configurations are defined in Table I
of the Configuration Rentals Comparison
Charts section.)

Processing times are listed for the following
standard benchmark problems:
(1)

The essence of most business data
processing applications is the updating of files to reflect the effects
of various types of transactions.
This benchmark problem is a file
processing run in which transaction data in a detail file is used to
update a master file, and a record
of each transaction is written in a
report file or journal (Figure 1).
This type of run forms the bulk of
the workload for manY' computer
systems, in diverse applications
such as billing, payroll, and inventory control.

The System Performance charts are particularly useful when you need to make cvmparisons of the performance and cost of competitive
computer systems (or different configurations
of the same system) in applications similar to
your own or your client's.
Each of the standard benchmark problems has
been coded and timed in detail by experienced
programmer/analysts. Each computer system's central processor speeds, input-output
speeds, and capabilities for simultaneous operations have been carefully considered to
determine the overall time required to process
each problem.

The listed "Activity" factors of 0.0,
0.1, and 1.0 refer to cases in which
an average of 0, 0.1, and 1.0 transaction record, respectively, mustbe
processed for each record in the
master file. Low activities are characteristic of applications such as
inventory control, whereas a payroll
run might well have an activity factor of 1.0. All calculated processing
times are reported in terms of the
number of minutes required to
process 10,000 master-file records.

To minimize subjective errors and ensure
valid performance comparisons, the input,
output, and basic computational procedure for
each benchmark problem are rigidly specified.
Conversely, the details of the computational
procedure are left flexible so that useful features of specific computer systems can be
effectively utilized.
All of the processing times shown in the System Performance charts are idealized times
with no allowance for set-up times, equipme~t
failures, inefficient coding, software inefficiencies, operator errors, or idle time. The
degrading effects of these factors are difficult
to estimate and tend to vary widely from installation to installation, but it is important to note
that they can cause a computer system's overall throughput to be substantially lower than our
published processing times for individual runs
might seem to indicate.

1/69

Generalized File Processing
Problem A

A

Detail
(Transaction)
File
Figure 1. Run Diagram for Generalized
File Processing Problem A

AUERBACH

~

11:400.101

SYSTEM PERFORMANCE

Figure 2 is a general flowchart
that summarizes the computational
process. Both the master file and
detail file are sequentially arranged, and conventional batch
processing techniques are employed. Record lengths are 108
characters for the master file, 80
characters (1 card) for the detail
file, and 120 characters (1 line) for
the report file. Record layouts are
fixed for the detail and report files,
but are left flexible for the master
file in order to take advantage of
the specific capabilities of each
computer system.
Card reading and printing are performed on-line in all standard configurations except paired configurations VIIB and VIIIB, in which cardto-tape and tape-to-printer transcriptions are performed off-line,
usually by a separate small-scale
computer. The master file is on
magnetic tape in all standard configurations except Configuration I,
where it is on punched cards.
(2)

Random Access File Processing
Problem

upon considerations of economy,
system throughput, software support, and reliability. Therefore,
disc files will normally be chosen
in preference to drums (which are
relatively expensive) or magnetic
strip devices (which tend to be
relatively slow and of lower
reliability).

Look for next
L"ccord from
Old !\laster F ,Ie

Input nc:\.1.
block from
Old l\Im;ter File

Unpack Master
Record and form
control totals

This benchmark problem represents a wide range of real-time
computer applications in which an
on-line master file is accessed to
answer inquiries and/or updated
to reflect various types of transactions. Figure 3 shows the basic
run diagram. Examples of this
type of processing include real-time
inventory control, credit checking,
airline and hotel reservations, online savings systems, etc.
In contrast to Generalized File

Processing Problem A, described
above, this problem uses random
access storage to hoW the entire
master file ~line, and processes
all transactions as they occur,
without prior sorting. All calculated
times are reported in terms of the
time in milliseconds required to
process each transaction and the
total time in minutes required to
process 10,000 transactions.
This problem is evaluated for one
or more of the three Random Access
standard configurations (IIIR, IVR,
and VIIIR). Where there are two or
more random access devices that
could satisfy the specified capacity
requirements, our choice is based

Figure 2. General Flowchart for Generalized
File Processing Problem A

© 1969 AUERBACH Corporation and AUERBACH Info, .Inc.

1/69

11:400.102

COMPARISON CHARTS

Master
File

remote terminal that initiated the
transaction (though the processor
time required to effect this transaction is not included in the published timing figures).
(3)

Detail
Transactions

Because conventional data processing techniques usually require all
records to be arranged in a particular sequence, sorting operations
are an important and timeconsuming part of the workload in
most business computer installations. This benchmark problem
requires that a file consisting of
10,000 records, each SO characters in length, be arranged sequentially according to an S-digit key,
such as an account number.

Report
File

Figure 3. Run Diagram for Random
Access File Processing Problem
Figure 4 is a general flowchart
that summarizes the computational
process. The master file is sequentially arranged in random access storage, and a two-stage
indexing procedure is used to determine the location of each masterfile record that needs to be accessed.
Record lengths are lOS characters
for the master file, SO characters
(1 card) for the detail transactions,
and 120 characters (1 line) for the
report file. Record layouts are
fixed for the detail and report files,
but are left flexible for the master
file so that the specific features of
each computer system can be advantageously utilized.

Use two-stage
indexing procedure
to obtain Master
Record address

The detail transactions (e. g., inquiries, orders, or deposits) are
assumed to be arriving in a random
sequence and at a continuous rate
that is high enough to ensure that
one or more transactions are always waiting to be processed.
Therefore, it makes no difference
whether the transactions enter the
system via an on-line card reader,
a simple remote inquiry terminal,
or a multi-terminal data communications network. This assumption
means that the Random Access File
Processing Problem does not attempt the highly complex and variable task of measuring the efficiency
of real-time data communications
networks; .it simply measures the
central computer system's ability
to locate and update randomlyaddressed master-file records.

Update Master
Record to reflect
Detail Transaction

The report file is written on either
magnetic tape or a random access
device, presumably for printing at
some later time. Each report
record is also made available for
optional transmission back to the

1/69

Sorting

Figure 4. General Flowchart for Random
Access File Processing Problem

fA.

AUERBACH

I>

11:400.103

SYSTEM PERFORMANCE

The "standard Estimate" column
lists the estimated sorting times
calculated by our analysts for sorting operations that use straightforward magnetic tape merging
techniques. Two-way tape merging
is used in the four-tape Standard
Configuration IT and ~ree-way
merging in all of the larger systems.
Whenever timing data is available
for a standard, manufacturersupplied sort routine, the time required to perform the same 10,000record sort is listed in the "Available Routines" column. Because
most manufacturer-supplied sort
routines now use internal sorting
and merging techniques which are
more sophisticated than those used
to prepare our estimates, the
"Available Routines" sort time will
often be substantially less than the
"Standard Estimate" time for a
given configuration. Nevertheless,
the Standard Estimates provide
useful, directly comparable indications of each computer system's
basic capabilities to perform magnetic tape input-output operations.
(4)

form with a precision equivalent to
at least eight decimal digits.
The "standard Estimate" columns
list the matrix inversion times calculated by our analysts through a
simple estimating procedure that
uses the system's floating-point
arithmetic speeds. Whenevertiming
data is available for a standard,
manufacturer-supplied matrix inversion routine, it is reported in
the "Available Routines" columns.
(5)

Generalized Mathematical
Problem A
Another frequently-encountered
scientific problem involves the
evaluation of polynomial equations
of the type Y = A + Bx + Cx2 + Dx3
+ Ex4 + Fx5. This benchmark problem includes the follOWing basic
steps:
•

Read an input record consisting
of 10 eight-digit numbers.

•

Perform a floating-point calculation that consists of evaluating
five 5th-order polynomials,
executing five division operations, and evaluating one
square root.

•

For every 10 input records,
form and print one output
record consisting of 10 eightdigit numbers.

Matrix Inversion
In many scientific and operations
research applications, such as
multiple regression, linear programming, and the solution of
simultaneous equations, the bulk of
the central processor's time is
spent in inverting large matrices.
This benchmark problem involves
the inversion of 10-by-10 and 40-by40 matrices. It measures the speed
•'1of the central processor on floatingpoint calculations; no input or output
operations are involved. All matrix
elements are held within the system's
main storage unit in floating-point

The "Computation "Factors" of 1,
10, and 100 mean that the standard
calculation described above is performed 1, 10, or 100 times,
respectively, for each input record
to show the effects of varying ratios
of computation to input-output volume. Processing times are listed
in terms of milliseconds per input
record.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

1/69

11 :400.104

COMPARISON CHARTS

SYSTEM PERFORMANCE COMPARISONS
GENERALIZED FILE
PROCESSINq PROBLEM A
STANDARD
CONFIGURATION
NUMBER

SYSTEM
IDENTITY

MONTHLY
RENTAL,

$

RANDOM ACCESS
BENCHMARK PROBLEM

Activity
0.0

0.1

Timing Summary
1.0

,Minutes per 10,000 Records
Burroughs B 200

I
II
III

4,525
5,895
8,840

-

-

Time per
Time per
Transaction 10,000 Records
Milbseconds

SORTING
10,000 80-Char. Records
standard
Estimate

Minutes

Minutes

2.2
1.4

2.9
2.8

67.
26.
26.

-

-

22.
9.5

1.9
1.8

18.
18.

-

-

-

7.5
5.0

125.

21.

-

-

-

-

125.

21.

-

-

-

14.

Burroughs B 2500

II
III
IVR

4,910
6,415
10,130

0.95
0.75

Burroughs B 3500

IVR
VIlA
VIIIR

11,630
15,480
19,680

0.37
-

2.0

18.

-

301.

50.

-

III
V
VIlA
VIIB

23,340
25,250
30,995
28,705

1.2
1.2
0.55
0.55

2.0
2.0
1.7
0.69

19.
19.
17.
1.8

-

-

-

-

-

-

2.9
2.9

2.8
2.8

IIIR
IVR
VI
VIlA

9,390
14,250
14,610
20,375

-

0.94
0.36

IVR
VI
VIlA

15,980
16,240
22,025

-

-

0.94
0.36

2.0
2.0

20.
20.

-

VI
VIlA
VIIB
VIIIB

16,640
22,110
23,511
39,045

0.56
0.56
0.56
0.29

1. 96
1.62
0.77
0.33

16.
16.
2.6
1.0

-

VIB
VIIB
VIIIB

40,110
40,671
57,045

0.19
0.19
0.19

0.28
0.28
0.19

1.2
1.2
1.0

-

-

-

1.4
2.0
1.4

-

VIlA
VIIIA

42,100
54,540

0.38*
0.19*

0.38*
0.19*

2.0*
1. 0*

-

-

2.5
1.3

..

VIlA
VIIIA

64,100
76,625

0.38*
0.19*

0.38*
0.19*

2.0*
1. 0*

-

-

2.5
1.3

-

-

-

-

-

Burroughs B 5500

CDC 3100

CDC 3300

CDC 3400

CDC 3600

CDC 6400
CDC 6600

GE 215

GE 225

GE 235

GE 415

GE 425

I

-

-

-

-

-

I
II
III
VI

4,905
6,250
7,375
8,325

I
II
III
IV
VI

5,085
6,450
9,155
15,620
11,985

3.7
1.6
0.80
1.6

III
IV
VI

11,870
18,385
15.120

1.5
0.77
1.5

I
II
III
IV
VIlA

5,135
6,955
8,255
13,950
15,245

-

2.4
1.8
0.47
0.47

I
II
III
IV
VIlA

6,120
7,940
9,240
14,935
16,545

2.4
1.8
0.47
0.47

-

-

2.0
2.0

20.
20.

-

3.7
3.7
3.7

-

-

..

-

-

-

-

-

-

-

-

-

6.1
2.7

-

-

-

-

3.7
2.5
1.8
2.5

-

-

2.5
1.7
2.5

25.
17.
25.

-

-

-

75.

-

2.4

15.
15.
15.

-

1.8
1.5
1.5

-

2.4
1.8
1.4
1.4

15.
61.

15.
15.
15.
15.

-

-

-

A .,

AUERBACH

-

-

67.
25.
25.
18.
25.

-

-

-

-

-

-

2.5

-

67.
28.
28.
28.

5.4
3.7
3.7

I

-

-

..
..

-

..

-

-

6.1
2.7

3.7
3.7
3.7
1.8

37.
25.
25.

-

37.
10.
5.3
10.
10.

5.
10.

-

-

..

-

-

24.
14.
8.5
14.

-

-

-

-

-

24.
13.
3.
3.

25.
13.
3.1
3.1

*Indicated time is for the tape-to-tape main processing run only; it is assumed that the required on-line card-to-tape and
tape-to-printer transcriptions will be performed with these or other programs.

1/69

Available
Routines

-

11:400.105

SYSTEM PERFORMANCE

SYSTEM PERFORMANCE COMPARISONS (Contd.)
MATRIX INVERSION
Standard Estimate
SYSTEM
IDENTITY

STANDARD
CONFIGURATION
NUMBER

MONTHLY
RENTAL,

GENERALIZED
MATHEMATICAL
PROBLEMA

Available
Routines

Computation Factor
for 10% Output

Array Size

$
10

40

10

40

1

4,525
5,895
8,840

--

4,910
6,415
10,130

0.026

IVR
IVR
VIIA
VIIIR

11,630
15,480
19,680

0.013

Burroughs B 5500

III
V
VIlA
VIlB

23,340
25,250
30,995
28,705

0.0025
0.0025
0.0025
0.0025

CDC 3100

IIIR
IVR
VI
VIIA

9,390
14,250
14,610
20,375

-

0.0013
0.0013

IVR
VI
VIIA

15,980
16,240
22,025

0.0008
0.0008

0.046
0.046

CDC 3400

VI
VIlA
VIIB
VIIIB

16,640
22,110
23,511
39,045

0.0004
0.0004
0.0004
0.0004

CDC 3600

VIB
VIIB
VIIIB

40,110
40,671
57,045

CDC 6400

VIIA
VIIIA

CDC 6600

VIIA
VillA

GE 215

I
II

I
II

m

Burroughs B 2500

Burroughs B 3500

CDC 3300

II

m

-

0.14
0.14
0.14
0.14

0.006
0.006
0.006
0.006

0.25
0.25
0.25
0.25

-

0.026
0.026
0.026
0.026

-

0.0003
0.0003
0.0003

0.017
0.017
0.017

--

42,100
54,540

0.00022
0.00022

0.011
0.011

--

64,100
76,625

0.00003
0.00003

0.0014
0.0014

-

-

0.75

-

0.08
0.08

-

4,095
6,250
7,375
8,325

0.70
0.70
0.70
0.07

33.
33.
33.
3.2

IV
VI

5,085
6,450
9,155
15,620
11,985

0.31
0.31
0.31
0.31
0.033

15.
15.
15.
15.
1.7

GE 235

III
IV
VI

11,870
18,385
15,120

0.07
0.07
0.005

3.5
3.5
0.22

GE 415

I
II

5,135
6,955
8,255
13,980
15,245

--

--

m
VI

GE 225

I
II

m

m
IV
VIIA

GE 425

I
II
III

IV
VIIA

--

-

-

1.5
-

6,120
7,940
9,240
14,935
16,545

-

0.0029

-

-

-

0.0021

-

0.17

-

--

0.12

100

Milliseconds

Minutes
Burroughs B 200

10

-

-

-

--

0.60
0.60
0.60
0.60
0.030

38.
38.
38.
38.
1.9

-

-

-

--

-

78.
-

-

78.

-

74.
7474.
9.5*

-

50.
50.

-

-

-

700.

350.
74.
74.
74.
39.*

-

50.
50.

-

6,900.
-

-

3,500.

-

330.
330.
330.
330.

--

330.
330.

-

50.
50.

50.
50.

265.
265.

65.
65.
12.
9.9

65.
65.
23.
23.

145.
145.
145.
145.

6.0
6.0
6.0

6.5
6.5
6.5

13.*
6.2*

13.*
6.2*

13.*
6.2*

13.*
6.2*

13.*
6.2*

13.*
6.2*

--

-

-

-

--

120.

-

100.

-

--

-

61.
61.
61.

-

-

-

--

-

280:

1,800.

-

-

240.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

1,400.

1/69

COMPARISON CHARTS

11 :400.106

SYSTEM PERFORMANCE COMPARISONS (Contd.)
GENERALIZED FILE
PROCESSING PROBLEM A
SYSTEM
IDENTITY

STANDARD
CONFIGURATION
NUMBER

MONTHLY
RENTAL,

$

RANDOM ACCESS
BENCHMARK PROBLEM

Activity
0.0

0.1

1.0

Minutes per 10,000 Records
GE 435

m
IV
VlIA

13,400
19,095
19,975

1.8
0.47
0.47

1.8
1.4
1.4

GE 625

VlIA
VIIIA

39,505
54,275

0.47*
0.26*

0.70*
0.26·

2.8*
1.4*

GE 635

VIlA
VIllA

40,630
55,400

0.47·
0.26*

0.70·
0.26·

Honeywell 110

I
II

IIIR
Honeywell 120

I
II

m
mR
Honeywell 200

I
II

m
mR
IV
Honeywell 1200

I

Honeywell 4200

Honeywell 8200

Honeywell 400

Honeywell 800

4,185
4,995
7,415
6,285
14,640
5,060
5,870
7,875
7,665
14,945
11,765
10,985
16,195
15,805

m
IV
IVR
VlIA
VIlB

-

-

4.0
2.1

-

-

-

-

-

148.6

-

3.4
0.9

3.4
2.1

160.
21.
21.

1.7

-

148.6

0.39

17.

-

148.6
111.1
-

-

-

0.39

1.7

17.

0.9
0.39
0.39

2.1
2.1
0.5

21.
21.
2.

8,935
16,095
13,435
18,330
17,685

0.9
0.39

2.1
1.7

21.
17.

IV
IVR
VlIA
VIlB
VIlIB
VlIIR

25,805
21,995
23,345
24,170
36,425
28,225

0.39

1.7

17.

0.39
0.39
0.30

2.1
0.49
0.30

21.
2.
1.1

VlIA
VlIIA
VIIIR

39,120
51,360
40,670

0.35·
0.28*

0.35*
0.28*

II

7,695
9,815
15,590
11,015

2.0
2.0
1.1
2.0

4.0
3.0
2.4
3.0

24.
20.
20.
20.

IV
Vl

10,750
12,290
20,980
14,530

1.6
1.6
0.57
1.6

3.7
2.8
1.9
2.8

24.
20.
20.
20.

Vl
VIlA
VllB
VIIIA
VIIIB

19,329
36,679
27,795
53,600
46,325

0.60
0.34
0.30
0.20
0.20

2.0
2.0
0.42
2.0
0.42

17.
17.
3.1
17.
3.1

-

-

-

0.39
0.39

-

-

-

-

-

-

-

2.1
0.5

21.
2.

-

-

-

-

0.43*
0.33*

-

-

-

-

-

111.1
-

111.1

Minutes
13.
3.1
3.1

A

AUERBACH

'"

-

-

-

-

3.1
1.7

-

-

-

-

-

-

-

24.7

24.7
-

-

41.
14.

-

-

7.9

2.5

2.8

-

-

-

-

33.
6.8

-

-

-

33.
6.8

-

2.5

-

-

6.8
2.5
2.5

-

18.5
-

6.8
2.5

7.1
2.7

2.5
2.5

2.8
2.8

6.8

-

24.7

-

18.5

-

18.5

-

-

-

-

2.5
2.5
2.1

-

-

-

-

226.3

37.7

-

2.3
1.8

226.3

-

37.7

-

-

-

-

12.
8.9
5.2
8.9

-

-

-

--

9.5
8.0
4.4
8.0
6.3
2.4
2.4
1.5
1.5

*Indicated time is for the tape-to-tape main processing run only; it is assumed that the required on-line card-to-tape and
tape-to-printer transcriptions will be performed with these or other programs.

1/69

-

3.1
1.7

-

160.
21.
21.

m

--

-

3.4
2.1

II

-

-

-

-

-

Minutes

2.8*
1.4*

190.
28.
27.

6.4
4.7

Milliseconds

-

-

-

Timing Summary
10,000 SO-Char. Records
Time per
Time per
standard
Available
Transaction 10,000 Records Estimate
Routines

-

3.4
0.9

m
IV
Vl
Honeywell 1400

3,835
3,465
6,180
5,070

m
mR
IV
IVR
Vl
VIlA
VIlB

II

Honeywell 2200

2,405
2,855
4,520

15.
15.
15.

SORTING

-

-

-

11 :400.107

SYSTEM PERFORMANCE

SYSTEM PERFORMANCE COMPARISONS (Contd.)
MATRIX INVERSION

SYSTEM
IDENTITY

STANDARD
CONFIGURATION
NUMBER

MONTHLY
RENTAL,

Standard Estimate

GENERALIZED
MATHEMATICAL
PROBLEM A

Avallable
Routines

Computation Factor
for 10% Output

Array Size

$
10

40

10

40

1

Minutes
GE 435

III
IV
VIlA

-

0.0016

.028
.028

-

.021
.021

--

GE 625

VIlA
VIIIA

39,505
54,275

GE 635

VIlA
vmA

40,630
55,400

0.0004
0.0004

Honeywell 110

I
II
IIIR

2,405
2,855
4,520

Honeywell 120

I
II
III
IIIR

3,835
3,465
6,180
5,070

Honeywell 200

I
II
III
mR
IV

4,185
4,995
7,415
6,285
14,640

I
II
III
IIIR
IV
IVR
VI
VIlA
VIIB

5,060
5,870
7,875
7,665
14,945
11,765
10,985
16,195
15,805

m
IV
IVR
VIlA
VIIB

8,935
16,095
13,435
1B,330
17,685

IV
IVR
VIlA
VIIB
vmB
VIIIR

25,805
21,995
23,345
24,170
36,425
28,225

VIlA
VIllA"
vmR

Honeywell 2200

Honeywell 4200

Honeywell 8200

Honeywell 400

Honeywell 1400

Honeywell 800

--

13,400
19,095
19,975

0.0005
0.0005

Honeywell 1200

-

--

--

0.0043
0.0043
0.0043

-

0.0028
0.002B

-

0.09

-

-

--

-

-

0.23
0.23
0.23

-

0.17
0.17

-

0.002
0.002

0.10
0.10

39,120
51,360
40,670

0.0002
0.0002

0.012
0.012

II
III
IV
VI

7,695
9,815
15,590
11,015

0.15
0.15
0.15
0.15

8.0
8.0
8.0
8.0

II
m
IV
VI

10,750
12,290
20,980
14,530

0.16
0.16
0.16
0.035

8.5
8.5
8.5
2.0

VI
VIlA
VllB
VIllA
VIIIB

19,329
36,679
27,975
53,600
46,325

0.003
0.003
0.003
0.003
0.003

0.17
0.17
0.17
0.17
0.17

-

-

-

--

-

10

100

Mlillseconds

-

-

-

-

--

--

-

74.

149.*
149.*

13.*
8.*

14.*
14.*

113.*
113.*

-

-

--

-

-

-

-

1,300.

1B.*
18.*

-

-

190.

-

13.*
8.*

-

-

-

-

-

88.
88.
lB.

--

88.
15.

-

-

--

B8.
88.
BO.

-

-

88.
58.

-

--

-

-

720.
720.
720.

-

490.
490.

-

88.
21.

-

-

200.
200.

75.
75.

75.
75.

75.
75.

88.
5.1

-

-

-

-

90.
90.

90.
90.

600.
600.

72.

90.

600.

-

-

-

-

-

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

-

-

-

-

-

-

1/69

COMPARISON CHARTS

11:400.108

SYSTEM PERFORMANCE COMPARISONS (Contd.)
GENERALIZED FILE
PROCESSING PROBLEM A
SYSTEM
IDENTITY

STANDARD
CONFIGURATION
NUMBER

MONTHLY
RENTAL,

$

Activity
0.0

0.1

1.0

Minutes per 10.000 Records
Honeywell 1800

IBM 360, Model 20

IBM 360, Model 25

IBM 360 Model 30

VI
VIlA
VIlE
VIllA
VIllE
I
II
IIIR

2,776
3,558
3,630

I
II
III
IIIR
IVR

3,555
4,945
6,445
5,280
9,635

I

IBM 360, Model 44

IBM 360, Model 50

IBM 360, Model 65

IBM 360. Model 75

IBM 360, Model 85
IBM 1401

-

-

1.8
0.33
1.8
0.22

18.
1.5
18.
1.5

6.0

-

-

7.0

-

-

-

2.8
1.4

2.8
2.1

-

-

-

67.
21.

-

67.
21.
21.

-

4,097
4,714
6,956
6,111
11,656

II
III
IIIR
IVR
VI

7,221
8,208
7,343
13,032
11,601

1.5
1.5
1.5

2.0

20.

V
VI
VIlA
XI

11,723
10,802
14,531
9,717

1.5
1.5
0.38
1.5

2.0
2.0
2.0
5.0

20.
20.
20.
50.

III
IV
IVR
VIlA
VIlE
VIIIR

15,400
21,564
18,399
19,720
21,837
26,773

1.5
0.38

2.0
1.5

20.
15.

VIlA
VIlE
VIllE
VIIlR

34,585
35,187
51,944
43,388

0.40
0.22

VIlA
VIIB
VIlIB
VIlIR

47,298
47,900
64,657
56,101

-

-

0.40
0.22

0.59
0.22

VIlIB
VIlIR

92,177
87,736

I

4,320
5,920
10,810
11,485

II

m
IV

IBM 1401-G

I

2,270

IBM 1410

I

6,115
8,415
12,240
19,060
15,790
23,560

II

III
IV
VI
VIIB

1/69

0.33
0.33
0.22
0.22

III
IIIR
IVR

II

IBM 360, Model 40

27,150
36,650
34,725
54,950
53,575

3.7
1.5

-

-

-

0.38
0.38

-

-

0.23

-

3.7
2.0

-

2.0
2.0

-

-

-

-

20.
20.

-

-

SORTING

TImmg Summary

10,000 80-Char. Records

TIme per
TIme per
Standard
Transaction 10,000 Records Estimate
MillIseconds

-

-

Mmutes

-

-

-

-

-

30.
9.5

-

148.
109.

25.
18.

-

-

-

-

-

-

-

40.
9.7

148.
109.

25.
18.

-

25.
9.2
5.0
3.0

-

-

13.
9.7

148.
109.

25.
18.

-

-

109.
-

-

-

-

-

-

0.59
0.22

2.0
1.1

-

0.23

-

-

7.5
4.2
2.6

-

-

3.2
2.0
2.0
2.0
1.2

A.

AUERBACH

2.4
1.8

-

1.8
1.7
1.4

-

1.6

-

-

117.

20.

139.
80.
20.
20.
20.
20.
3.3

-

-

2.7
2.1
2.7
2.7
1.9

-

-

100.
40.
26.
20.

-

-

-

-

-

-

55.
55.
28.
55.

10.4
8.6
4.0
3.0
3.8

2.0
1.8
1.7

-

117.

-

9.7

-

-

20.

-

-

-

2.4
1.8

117.

1.5

-

-

-

-

-

2.3
2.3

-

-

27.
10.

9.7
2.3

-

2.0
1.1

61.

-

18.
20.

-

-

-

32.

-

20.
2.0

-

Mmutes

-

117.

-

--

Available
Routmes

186.

-

2.0
0.58

3.7
2.4
2.0

2.7
1.4
1.0
1.4
0.85

67.
20.
20.

RANDOM ACCESS
BENCHMARK PROBLEM

20.

-

-

-

-

-

-

-

-

41.
15.
12.

35.
13.
10.

-

-

30.
9.0
6.0
9.0

-

9.7
6.0
7.0
7.0

11:400.109

SYSTEM PERFORMANCE

SYSTEM PERFORMANCE COMPARISONS (Contd.)
MATRIX INVERSION

SYST(,M
Im:NTITY

STANDARD
CONFIGURATION
NUMBER

MONTHLY
RENTAL,

stan~ard

Estimate

I

GENERALIZED
MATHEMATICAL
PROBLEM A

Available
Iloutines

Computation Factor
for 10% Output

Array Size

$
10

40

10

40

1

Honey" e 11 1800

IllM 360, Model 20

VI
VIlA
VIIll
VIllA
VIIIB

27,150
36,650
34,725
54,950
53,575

0.0013
0.0013
0.0013
0.0013
0.0013

-

0.066
0.066
0.066
0.066
0.066

-

I
III
IIIR

2,776
3,55R
3,630

I
II
III
IIIIl
IVil

3,555
4,945
6,445
5,280
9,635

0.032
0.032
0.032
0.032
0.032

1.7
1.7
1.7
1.7
1.7

I
II
III
IIIIl
IVil

4,097
4,714
6,956
6,111
11,656

0.025
0.025
0.025

1.2
1.2
1.2

II
III
lIm
IVR
VI

7,221
8,208
7,343
13,032
11,601

0.0071
0.0071
0.0071

0.39

IBM 360, Model 44

V
VI
VIIA
XI

11,723
10,802
14,531
9,717

0.0017
0.0017
0.0017
0.0017

0.10
0.10
0.10
0.10

IBM 360, Model 50

III
IV
IVR
VITA
VIIB
VIIIR

15,400
21,564
18,399
21,720
21,837
26,773

0.0017
0.0017

0.07
0.07

0.0017
0.0017

0.07
0.07

VIlA
VITB
VIIIB
VIIIR

34,585
35,187
51,944
43,388

-

-

0.00022
0.00022

0.012
0.012

VITA
VIIB
VIIIB
VIIm

47,298
47,900
64,657
56,101

-

-

0.00016
0.00016

0.0089
0.0089

IBM 360, Model 85

VIllB
VIIIR

92,177
87,736

0.00007
0.00007

0.0036
0.0036

IBM 1401

I
II
III
IV

4,320
5,920
10,810
11,485

0.33
0.33
0.33
0.33

2,270

-

-

6,115
8,415
12,240
19,060
15,790
23,560

0.17
0.17
0.17
0.17
0.17
0.17

9.0
9.0
9.0
9.0
9.0
9.0

IllM 360, Model 25

IBM 360, Model 30

IllM 360, Model 40

IBM 360, Model 65

IBM 360, Model 75

IBM 1401-G

I

IBM 1410

I
II
III
IV
VI
VIIB

-

-

-

-

-

-

-

-

-

0.39
0.39

-

-

-

-

10

100

M1lliseconds

Minutes

-

-

-

-

-

-

---

-

-

-

-

75.
6.7
75.
5.9

75.
14.
75.
14.

-

-

-

-

-

-

130.
130.
130.

-

100.
100.
100.

-

-

-

100.
100.

-

-

-

-

-

-

-

-

-

-

12,000.
12,000.
12,000.

-

-

-

480.
480.
480.

4,230.
4,230.
4,230.

-

-

150.
150.

2,000.
2,000.

-

-

-

150.

2,000.

100.
100.
100.
100.

100.
100.
100.
100.

280.
280.
280.
280.

100.
100.

100.
100.

400.
400.

100.
9.7

100.
31.

400.
280.

-

-64.

-

-

9.7
6.5

-

9.7
6.5

-

-

-

-

-

8.7

-

520.

520.
-

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

-

-

9.7
6.5

-

-

-

-

100.

-

1,200.
1,200.
1,200.

-

130.
130.
130.
130.

-

9.7
6.5
8.7

-

5,000.

5,000.
-

--

-

64.

-

35.
35.

-

20.

-

50,000.

-

50,000.

-

-

-

1/69

COMPARISON CHARTS

11:400.110

SYSTEM PERFORMANCE COMPAIUSONS (Contd.)
RANDOM ACCESS
BENCHMARK PROBLEM

GENERALIZED FILE
PROCESSING PROBLEM A
SYSTEM
IDENTITY

STANDARD
CONFIGURATION
NUMBER

MONTHLY
RENTAL,

$

Tlnl1ng Summary

ActiVity
0.0

0.1

1.0

Minutes per 10,000 Records
IBM 1440

I

1I**
III**

-

-

3,295
4.050
5,920

3.8
2.9

10.7
5.1

135.
73.
48.

Tmle pe,'
Time per
Transaction 10.000 !tecord"
MillIsecond,

-

~

OOOTmc

10.000 80-Char. Record.
Standard
Estimate

AV'1I1abk
/{outmeh

1-I

~hnutes

l\lmutes

-

-

-

40.
HI.

-

-

-

-

-

-

!l.1

-

8.5

-

-

-

-

-

-

-

:J. ~
2.2

-

-

-

2.7
1.!l

-

-

-

2.4

-

~.3

-

-

-

H.3

-

IBM 1460

III

11,735

1.4

3.6

26.

IBM 7010

III
IV
VI
VIIB

19,175
27,225
22,175
28,355

1.4
0.56
1.4
0.64

2.0
1.3
2.0
0.96

20.
13.
20.
3.2

VI
VIlA
VIIIB

20.715
27,190
47,145

1.4

2.3

20.

0.33

0.75

5.5

VIlA
VIIIB

36,690
56,645

-

-

0.39

0.39

1.9

III
VIIB
VIIIB

19,400
29,775
45,030

1.3
0.45
0.38

6.7
0.80
0.80

67.
4..5

VIIB
VIIIB

32,915
49,890

1.2
1.2

1.7
1.7

5.7
5.7

IBM 7074

VIIB
VIIIB

40,465
72,840

0.45
0.18

0.6
0.18

2.2
1.7

-

-

3.0
1.2

1.5
1.2

IBM 7080

VIIB
VIIIB

51,745
79,325

0.42
0.18

0.58
0.2

2.
1.4

-

-

2.1;
1.3

1.2
0.12

IBM 7090

VIIB
VIIIB

66,770
89,215

0.47
0.21

0.61
0.21

1.9
1. (;

-

-

-

3.2
1.5

-

IBM 7094-1

VIIB
VIIIB

72,395
95,065

0.47
0.21

0.61
0.21

1.9
0.96

-

-

3.2
1.5

-

I
II
III
IIIC
IV
IVR

5,450'
4,775
7,695
7,800
19,040
12,445

I
II
IIIC

4,750
3,975
7,300

m
mc
IV
IVR

IBM 7040

IBM 7044
IBM 7070

IBM 7072

NCR 315

NCR 315-100

NCR 315 RMC

RCA Spectra 70/15
RCA Spectra 70/25

RCA Spectra 70/35

I
II

-

,

3.3
1.5
1.3
0.4

-

-

-

-

5.1
3.0
3.7
1.9

-

-

4.5

80.
29.
26.
24.
18.

-

3.6
1.6

7.8
3.8

80.
41.
25.

9,970
10,175
19,140
13,820

1.7
1.7
0.35

1.7
1.7
1.9

19.
19.
19.

3,470
4,815

-

II
III
IV

5,990
6,610
12,585

I
II
m
mR
IVR
VI
VIlA

5,420
6,896
7,616
8,336
10,791
9,046
13,022

-

1.~

1.4
1.4
0.7

1.4
1.4

-

-

1.4
0.7

-

2.2
2.2
2.2
1.3

-

2.2
2.2

-

2.2
1.3

-

66.
22.
22.
22.
13.
64.
22.
22.

-

22.
13.

-

-

-

-

-

-

235.

-

-

A.

AUERBACH

-

-

15.

8.5
3.0

20.
5.8

I

-

-

-

,

i

7. !J

-

-

-

-

-

4.

-

-

-

149.5
109.5

26.0
18.0

-

-

I

-

-

2.7
-

-

-

-

19.

15.

15.
10.
2.5

15.
10.
2.5

-

!

2li.

-

-

i

-

2. (,

-

I

;

-

-

-

!

5.7
2.0
2.0

-

-

,
,

-

3.8

-

-

R. {)
4. H

-

-

-

3.8

39.

-

**Using 1311 Disk storage Drives in place of magnetic tape.

1/69

-

39.

-

i
!

77.

-

I

-

15.0
10.0
-

-

-

-

10.0
10.0

-

-

II

11:400.111

SYSTEM PERFORMANCE

SYSYTEM PERFORMANCE COMPARISONS (Contd.)
MATRIX INVERSION

SYSTEM
IDENTITY

STANDARD
CONFIGURATION
NUMBER

Standard Estimate
MONTHLY
RENTAL,

$

I

GENERALIZED
MATHEMATICAL
PROBLEM A

Available
Routines

Computation Factor
for 10% Output

Array Size
10

10

40

40

1

-

IBM 1460

m

11,735

0.17

-

IBM 7010

m
IV
VI
VIm

19,175
27,225
22,175
28,355

0.06
0.06
0.06
0.06

3.5
3.4
3.4
3.4

IBM 7040

VI
VIlA
VIIIB

20,715
27,190
47,145

0.002
0.002
0.002

0.10
0.10
0.10

IBM 7044

VIlA
vmB

36,690
56,645

0.001
0.0010

0.068
0.068

IBM 7070

m
VIm
VIIIB

19,400
29,755
45,030

0.037
0.037
0.037

2.1
2.1
2.1

IBM 7072

VIm
VIIm

32,915
49,890

0.0037
0.0037

0.24
0.24

IBM 7074

VIm
VIIIB

40,465
72,840

0.003
0.003

0.17
0.17

IBM 7080

VIIB
VIIIB

51,745
79,325

IBM 7090

VIIB
VIIIB

66,770
89,215

0.001
0.001

0.062
0.062

IBM 7094-1

VIm
VIIIB

72,395
95,065

0.0004
0.0004

0.029
0.029

NCR 315

I
II
m
mc
IV
IVR

5,450
4,7'15
7,695
7,800
19,040
12,445

0.09
0.09
0.09
0.09
0.09

I
II
mc

4,750
3,975
7,300

0.09
0.09
0.09

5.
-

III
mc
IV
IVR

9,970
10,175
19,140
13,820

--

0.4
0.4
0.4

IBM 1440

I
ll*

111*

NCR 315-100

NCR 315 RMC

3,295
4,050
5,920

RCA Spectra 70/15

I
II

3,470
4,815

RCA Spectra 70/25

II

5,990
6,610
12,585

m

IV
RCA Spectra 70/35

I
II

m

IIIR
IVR
VI
VIlA

5,420
6,896
7,616
8,336
10,791
9,046
13,022

-

-

-

-

-

-

0.014
0.014
0.014

0.70
0.70
0.70

0.014
0.014

0.70
0.70

-

-

10

100

Milliseconds

Minutes

--

--

-

-

-

--

--

-

-

100.
17.
16.
13.
7.7

-

-

-

--

--

150.
150.
150.

1,300.
1,300.
1,300.

47.
47.

450.
400.

-

.

-

0.056
0.055
0.055

3.6
3.6
3.6

63.
63.

600.
600.

6,000.
6,000.

-

-

25.
25.

45.
45.

400.
400.

-

11.
11.

37.
37.

350.
350.

-

-

-

8.5
7.7

30.
30.

270.
270.

7.7
7.7

17.
17.

140.
140.

23.
32.
32.
32.
23.

190.
200.
200.
200.
190.

2,000.
2,000.
2,000.
2,000.
2,000.

-

-

-

-

-

45.0
45.0
45.0

230.
230.
230.

2,000.
2,000.
2,000.

-

47.
47.
47.

350.
350.
350.

3,300.
3,300.
3,300.

47.
47.

350.
350.

3,300.
3,300.

-0.077
0.077
0.077
0.077
0.077

-

-

--

-

-

-

4.

-

-

-

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

-

-

-

-

-

-

-

-

-

1/69

COMPARISON CHARTS

11 :400.112

SYSTEM PERFORMANCE COMPARISONS (Contd.)
GENERALIZED FILE
PROCESSING PROBLEM A
SYSTEM
IDENTITY

STANDARD
CONFIGURATION
NUMBER

MONTHLY
RENTAL,

$

0.0

RANDOM ACCESS
BENCHMARK PROBLEM

SORTING

Activity

Timing Summary

10,00080-Char. Records

0.1

Time per
Time per
Transactlon 10,000 Records

1.0

Minutes per 10,000 Records
RCA Spectra 70/45

MIlliseconds

-

standard
Estimate

Minutes

MInutes

-

-

9.4
2.4
2.4

-

III
IIIR
IV
IVR
VI
VilA
VIlB

8,712
9,351
14,402
12,421
10,567
14,156
16,142

III
IIIR
IV
IVR
VI
VIlA
VIIB
VIIIB

13,840
15,564
18,915
17,425
13,845
17,345
19,425
33,975

I
II
III
IV
VI

4,271
5,084
9,687
20,290
12,880

-

-

5.7
1.5
1.5
1.5

10.1
4.3
4.3
4.3

III
IV
VI
VIlA
VilB

11,390
18,940
14,265
21,265
21. 604

0.61
0.37
0.61
0.29
0.29

1.9
1.9
1.9
1.9
0.29

18.
18.
18.
18.
1.3

-

-

4.0
2.7
4.0
1.9
1.9

-

UNIVAC III

III
VI
VIlA
VIIB

19,000
20,400
25,000
38,730

0.19
0.19
0.19
0.19

2.1
2.1
2.1
0.19

20.
20.
20.
1.5

-

-

1.7
1.7
1.2
1.2

1.2
1.2
1.2
1.2

UNIVAC 418

III
VIlA

7,125
17,875

1.6
0.42*

2.4
0.68*

24.
3.7

-

III
VIlA
VIllA

19,780
31,270
48,120

2.3
0.27*
0.27*

2.3
0.42*
0.42*

21.
2.4*
2.4*

III
VIlA
VIllA

14,290
25,085
43,755

0.82
0.32*
0.32*

2.2
0.50*
0.34*

III
VIlA
VIllA

32,270
39,405
49,555

0.82
0.32*
0.32*

2.2
0.50*
0.34*

3.2

5.3

RCA Spectra 70/55

RCA 301

RCA 3301

UNIVAC 490

UNIVAC 491/492

UNIVAC 494

UNIVAC 1004
UNIVAC 1050

UNIVAC 1108

1.4

-

2.2

22.

-

-

0.36

1.3

1.4
0.36
0.36

2.2
2.2
0.52

1.4

2.2
1.3

22.

0.36
1.4
0.36
0.36
0.18

2.2
2.2
0.52
0.18

22.
22.
2.1
0.85

-

-

I
II
III
IV

3,470
5,030
6,660
18,720

1.0
0.82
0.53

2.9
2.4
2.1

VilA
VIllA

50,365
65,075

0.27*
0.27*

0.43*
0.27*

I

1,290

-

UNIVAC 9300

I
II
III
IV

1,740
3,610
4,545
7,810

-

-

200.
49.
32.
32.
32.

-

1,800
2,725

UNIVAC 9200

-

13.

-

I
II

-

12.
22.
22.
2.1

-

-

149.5

26.0

109.5

18.0

-

-

149.5

-

-

-

-

26.0

-

109.5
-

18.0
-

-

-

-

-

-

-

-

-

-

9.4

-

2.4

9.4

-

-

-

9.4
2.4
2.4
1.2

-

-

60.
15.
13.
15.

-

-

-

15.
1.7
1.7

-

-

22.
2.2*
2.2*

-

-

-

-

5.1
2.1
2.1

-

20.
1. 9*
1. 7*

-

-

5.1
2.1
2.1

-

-

-

100.
27.

-

-

-

100.
24.
24.
21.

-

-

-

10.
5.5
3.6

-

-

-

1.9
1.9

-

-

-

-

-

-

-

-

-

206.
21. 2
21.2
21. 2

-

-

6.5
4.7
4.7

@

-

-

206.

A

-

11.
2.8

-

AUERBACH

-

-

1. 5*
1. 3*

2.1
2.1
2.1

-

2.4

-

*Indicated time is for the tape-to-tape mam processing run only; it IS assumed that the required on-line card-to-tape and
tape-to-printer transcriptions will be performed with these or other programs.

1/69

Available
Routmes

-

-

-

-

-

11:400.113

SYSTEM PERFORMANCE

SYSTEM PERFORMANCE COMPARISONS (Contd.)
MATRIX INVERSION

SYSTEM
IDENTITY

STANDARD
CONFIGURATION
NUMBER

MONTHLY
RENTAL,

GENERALIZED
MATHEMATICAL
PROBLEM A

Available
Routines

Standard Estimate

Computation Factor
for 10% Output

Array Size

$
10

40

10

40

1

Minutes
RCA Spectra 70/45

III
IIIR
IV
IVR
VI
VIlA
VIIB

8,712
9,351
14,402
12,421
10,567
14,156
16,142

0.0053

0.30

0.0053

0.30

0.0053
0.0053
0.0053

0.30
0.30
0.30

III
IIIR
IV
IVR
VI
VIlA
VIIB
VIIIB

13,840
15,564
18,915
17,425
13,845
17,345
19,425
33,975

0.0015

0.08

0.0015
0.0015
0.0015
0.0015

I
II
III
IV
VI

4,271
5,084
9,687
20,290
12,880

0.37
0.37
0.37
0.37
0.020

III
IV
VI
VIlA
VIIB

11,390
18,940
14,265
21,265
21,604

0.0010
0.0010
0.0010

0.040
0.040
0.040

III
VI
VIlA
VIIIB

19,000
20,400
25,000
38,730

0.024
0.024
0.024
0.024

1.4
1.4
1.4
1.4

UNIVAC 418

III
VIlA

7,125
17,875

UNIVAC 490

III
VIlA
VillA

19,780
31,270
48,120

0.023
0.023
0.023

1.0
1.0
1.0

UNIVAC 491/492

III
VIA
VillA

14,290
25,085
43,755

0.018
0.018
0.018

0.8
0.8
0.8

UNIVAC 494

III
VIlA
VIIIA

32,270
39,405
49,555

0.001
0.001
0.001

0.05
0.05
0.05

UNIVAC 1004

I
II

1,800
2,725

UNIVAC 1050

I
II
III
IV

UNIVAC 1108

VIlA
VillA

UNIVAC 9200
UNIVAC 9300

RCA Spectra 70/55

RCA 301

RCA 3301

UNIVAC III

-

-

-

0.0015

-

--

-

-

0.08
-

0.08
0.08
0.08
0.08
20.
20.
20.
20.
1.0

-

-

3,470
5,030
6,600
18,720

-

-

50,365
65,075

0.00017
0.00017

0.0089
0.0089

I

1,290

I
II
III
IV

1,740
3,610
4,545
7,810

-

-

10

100

Milliseconds

-

-

--

1,150.

100.

1,150.

47.
47.
9.5

100.
100.
100.

1,150.
1,150.
1,150.

-

47.

47.

280.

42.

280.

47.
47.
9.5
4.8

47.
47.
29.
29.

280.
280.
280.
280.

--

-

--590.

--

3,700.

--

-

-

-

-

42.
-

-

300.

-

-

-

-

-

100.

42.

11.
11.
11.
11.

---

-

47.

-

0.19
0.19
0.19
0.19

-

-

-

-

-

-

,-

65.
65.
8.3
25.
25.
25.

-

-

-

-

-

-

65.
65.
26.

210.
210.
210.

250.
250.
250.

2,500.
2,500.
2,500.

-

-

--

100.
55.
55.

290.
290.
290.

3,400.
3,400.
3,400.

75.

290.
290.
290.

2,700.
2,700.
2,700.

45.
45.
75.
7.3*
7.3*

-

--

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

-

-

-

7.0*
7.0*

-

75.
7.3*
7.3*

75.
58.*
58.*

-

--

7.0*
7.0*

-

-

21. *
21.*

-

-

-

1/69

£
J6.\

AUERBAC~
@

11:510.101
STANDm

ED]?
REPORTS

i,.....---~------I

COMPARISON CHARTS - NON--U. S. A. COMPUTERS
CENTRAL PROCESSORS AND WORKING STORAGE

COMPARISON CHARTS - NON-U.S.A. COMPUTERS
CENTRAL PROCESSORS AND WORKING STORAGE

An introduction to the Central Processor and Working Storage Section of the
Comparison Charts, giving the precise meaning of each entry, can be found on
Page 11 :210.101.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

11/69

COMPARISON CHARTS -

11 :510.102

A/s Regnecentralen Re 4000

System Identity

Word Length
DATA
STRUCTURE

Floating Point
Representation

Binary Bits

(Denmark)
plus 1 panty and :l
protectIon
~4

NON-U,S.A. COMPUTERS

GE-55 (France)

Bull-GE GAMMA 10 (France)
6

8 + 1 parity per byte

Decimal Digits

6

1

2

Characters

3

1

1

Radix

Binary

-

Decimal

Fraction Size

36

-

7 deCImal digits

Exponent Size

12

-

2 hexadecimal dIgits

Model Number

RC 4005

GAMMA 10

GE-55

Arithmetic Radix

Binary

Decimal

Decimal, binary

Operand Length, Words

1/2, 1, 2

Variable

Variable

Instruction Length, Words

1

3

1 to 8 bytes

Addresses per Instruction

1

2

0, 1, or 2

c=a+b

?

?

7

c:::: ab

?

?

?

~

?

?

(s)

c=a+b

?

-

(s)

c

?

-

(8)

?

-

(s)

Checking of Data Transfers

Yes

Yes

Parity

Program Interrupt Facility

Yes

No

Yes

Number of Index Registers

3 or 4

None

10

Likely Fixed
Point Execution
Times, J.Lsec (5
Digits Min.
Precision)
Likely Floating
Point Execution

c

~

alb

ab

Times, j.Lsec
c~a/b

CENTRAL
PROCESSOR

Indirect Addressing

Yes

Yes

No

Special Editing Capabilities

None

?

?

Boolean Operations

AND, OR, EXC OR

?

AND, INC OR, EXC OR

Table Look-up

Yes

None

None

Console Typewriter

Yes

None

None

Input-Output Channels

2, maXlmum 64 controllers

?

Up to 3

Features and Comments

High-speed (max 16 controllers simultaneously) and lowspeed mput/ output channels

Model Number
f---Type of Storage

Selected configurations
marketed in some U. S. A.
areas

RC 4081

GAMMA 10

GE-55

Core

Core

Core

Minimum

16,384

1024

2,500

Maximum

131,072

4096

10,000

Decimal Digits

786,432

4096

20,000

Characters

393,216

4096

10,000

eye Ie Time, Msec

1.5

7

7.9

Effective Transfer Rate, char/sec

1,500,000

?

?

Checking

Parity

Yes

Parity

Storage Protection

Yes

No

None

Number of Words

Maximum
Total Storage
WORKING
STORAGE

Features and Comments

*

With optional equipment.
(s) Using subroutine.

11/69

fA

AUERBACH
®

(Contd.)

11:510.103

CENTRAL PROCESSORS AND WORKING STORAGE

ElbIt 100 (Israel)

270-10

12

16 + parIty

3

4.6

2

2

Bmary

-

Hexadecimal

FUjItsu FACOM 270 Senes (Japan)
270-20
16 + parity + memory

System Identity

270-30

Protect bit

Binary Bits
Word Length

Decimal Digits
Characters

DATA
STRUCTURE

Radix

Binary

Bmary

24 or 56 bIts

24 or 56 bIts

Fraction Size

7 bIts

7 b,ts

Exponent Size

Bmary

Bmary

Binary

1

0.5 or 1

1,2,or4

1. 2, or4

1

1

lor 2

lor 2

Instruction Length, Words

1

1

1

1

Addresses per Instruction

14.0

312

21. 6

8.1

-

(s) 15,000

37.2

16.2

-

(s) 15,000

55.8

22

-

*43.2

*15.7

c=a+b

-

*76.8

*27.9

c = ab

-

*117.6

*43.2

c=a/b

None

ParIty

ParIty

Parity

Cbecking of Data Transfers

5 or 8 dIgIts
1 digIt

Floating Point
Representation

100

Model Number
Arithmetic Radix
Operand Lengtb, Words

Likely Fixed
Point Execution
Times, /-Lsee (5
Digits Min.
Precision)

c = a +b
c = ab
c=a/b

Likely Floating
Point Execution
Times, #lsee

1 level

1 level

12 levels

12 levels

Program Interrupt Facility

None

3

3

3

Number of Index Registers

1 level

No

Yes

Yes

None

(s) none

(s) none

(s) none

None

AND, EXC OR

AND, INC OR,
EXC OR

AND, INC OR,
EXCOR

No

None

None

None

Yes

Standard

Standard

standard

Maximum of 256

1 direct channel

1 direct channel
3 selector channels

1 direct channel
multiplexor
7 selector channels

CENTRAL
PROCESSOR

Indirect Addres sing
Special Editing Capabilities
Boolean Operations
Table Look-up
Console Typewriter

Input-Output Channels

Features and Comments
Model Number
Core

Core

Core

Core

Type of Storage

1,024

1,048

4,096

8,192

Minimum

8,192

4,096

32,768

65,536

Maximum

24,576

4.6 x4,096

147 x 10 3

299 x 103

16,384

8,192

65,536

131,072

2.0

2/word

2.4/word

.9/word

10 x 105

833 x 103

2,222 x 103

Number of Words

None

Panty

Parity

Parity

None

None

Write only

Write only

Includes a drum
(131 x 103 words,
20 mBec access)

Includes a drum
(262 X 10 3 words,
10 msee access)

Decimal Digits
Characters

Maximum
Total Storage
Cycle Time, J,.Lsec

Effective Transfer Rate, char/sec

WORKING
STORAGE

Checking
Storage Protection

Features and Comments

*

With optlOnal equipment.
(s) Using subroutine.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

11/69

COMPARISON CHARTS - NON-U. S. A. COMPUTERS

11:510.104

System Identity

Word Length
DATA
STRUCTURE

FUjitsu F ACOM 230 Series (Japan)
230-20

230-10

230-3~

Binary Bits

8 +W.M. +P

4+W.M. +P

4+W.M. +P

Decimal Digits

2/byte

1

1

Characters

l/byte

0.5

0.5

Radix

(s) decimal

Hexadecimal

*Hexadecimal

Fraction Size

1-20 dIgitS

Variable 2-123 digits

Variable 2-123 digits

Exponent Size

(s) 2 dIgIts

2 dIgIts

2 digIts

Floating Point
Representation

Model Number
Arithmetic Radix

Bmary & decimal

HexadecImal & decimal

Hexadecimal & decimal

Operand Length, Words

Variable

VarIable

Variable

Instruction Length, Words

I, 2, 3. 4. 5. 6 bytes

4. 8. or 12 digits

4. 8. or 12 digits

Addresses per Instruction

O. 1. or 2

O. 1. or 2

O. 1. or 2

c=a+b

96

98.1

80.3

c = ab

7100 (s)

639.0

263.3

c=a/b

8500 (s)

1308.15

553.8

c = a +b

8300 (s)

917.1

155.9

c = ab

7900 (s)

1703.25

425.3

Likely Fixed
Point Execution
Times. I'sec (5
Digits Min.
Precision)
Likely Floating
Point Execution
Times, J.l.sec

11.000 (s)

2421. 0

696.2

Checking of Data Transfers

Parity

Parity

Parity

Program Interrupt Facility

Yes

Yes

Yes

Number of Index Registers

-

2

2

Indirect Addressing

None

Yes

Yes

Special Editing Capabilities

(s)

Yes

Yes

Boolean Operations

None

Yes

Yes

Table Look-up

None

None

None

Console Typewriter

*optional

*optional

*optional

Input-Output Channels

1 dIrect
2 data

1 direct
4 { selector x 4
multiplexor

1 direct
4 { selector x 4
multiplexor

Features and Comments

W.M. = Word mark
P
= ParIty check

c = alb
CENTRAL
PROCESSOR

Model Number
Type of Storage

Core

Core

Core

Minimum

4 x 103 bytes

4 x 10 3 bytes

4 x 103 bytes

Maximum

8 x 103 bytes

32 x 103 bytes

32 x 103 bytes

Decimal Digits

16 x 103

65 x 103

65 x 103

Characters

8 x 103

32 x 103

32 x 103

Number of Words

Maximum
Total Storage

WORKING
STORAGE;

Cycle Time, IJsec

2/byte

1.8/byte

2.2/byte

Effective Transfer Rate, char/sec

125 x 103

277 x 10 3

455 x 103

Checking

ParIty

Parity

Parity

Storage Protection

None

Yes, write only

Yes, wrIte only

Features and Comments

Including a drum
(65 x 10 bytes
15 msec)

* With optional equipment.
(s) USing subroutine.

11/69

fA

AUERBACH
®

(Contd.)

CENTRAL PROCESSORS AND WORKING STORAGE

11:510,105

FUjitsu F Al OM 200 ""nes Japan)
230-50

System Identity

230-60

36 + 4 + 2 (P)

36 + 4 + 2 (P)

10.5

10.5

Binary Bits
Word Length

Decimal Digits

6

6

Bmary

BInary

27 or 62 bits

27 or 62 bits

Fraction Size

9 bits

9 bits

Exponent Size

Characters

DATA
STRUCTURE

Radix
Floating Point
Representation

Model Number
Bmary

Binary

lor 2

lor 2

1

1

1

1

13.2

3.92

c=a+b

23.7

6.72

c = ab

Arithmetic Radix
operand Length, Words
Instruction Length, Words
Addresses per Instruction

Likely Fixed
Point Execution
Times, J,.lsec (5
Digits Min.
Precision)

44

12.84

23.7

4.93

39.6

6.34

48.4

7.78

Parity

Parity

Checking of Data Transfers

Yes, 8 classes

Yes, 5 classes

Program Interrupt Facility

8

8

Number of Index Registers

Yes

Yes

Good

Good

AND, INC OR NOT, EXC OR

AND, INC OR NOT, EXC OR

Good

Good

'OPtional

'OPtional

7 data channels

18 channels
{ selector or
multiplexor

selector or
multiplexor

Multlprocessor capability

c=a/b

c

= a +b
Likely Floating
Point Execution
Times, /-Lsec

c = ab
c=a/b

CENTRAL
PROCESSOR

Indirect Addressing

Special Editing Capabilities
Boolean Operations

Table Look-up
Console Typewriter

Input-Output Channels

Features and Comments

Model Number
Core

Core

16 x 103

32 x 103

Minimum

65 x 10"

262 x 10"

Maximum

682.5 x 103

2.3 x 106

Decimal Digits

390 x 103

1. 57 x 106

2.2/word

.92/bank

2.7 x 106

40 x 106

Parity

ParIty

Yes, write only

Yes, good

Type of Storage
Number of Words

Multi-Bank memory optional

Characters

Maximum
Total Storage
Cycle Time, IJ-sec

Effective Transfer Rate, char/sec

WORKING
STORAGE

Checking
Storage Protection

Features and Comments

*

With optIonal equipment.
(s) Using subroutine,

© 1969 AUERBACH Corporation and AUERBACH Info, Inc

11/69

11:510,106

COMPARISON CHARTS - NON-U.S.A. COMPUTERS

System Identity

H-8300

Binary Bits

8 per byte

8 per byte

8 per byte

Decimal Digits

2

2

2

Characters

1

1

1

Radix

-

-

Binary

Fraction Size

-

-

24 or 56

Exponent Size

-

-

7

Model Number

H-8210

H-8200

H-8300

Arithmetic Radix

Binary, decimal

Binary, decimal

Binary, decimal

Operand Length, Words

Variable

Variable

Variable

Instruction Length, Words

4 or 6

4 or 6

2, 4, or 6

Word Length
DATA
STRUCTURE
Floating Point
Representation

lor 2

lor 2

0, 1, or 2

c=a+b

63

88

51

c = ab

416

(s)

141

c=a/b

648

(s)

232

c = a +b

-

-

79 or 114

c = ab

-

Addresses per Instruction
Likely Fixed
Point Execution
Times, )lsec (5
Digits Min.
Precision)
Likely Floating
Point Execution

-

-

Checking of Data Transfers

Panty

Parity

Parity

Program Interrupt Facility

Yes

Yes

Yes. multilevel

Times, ,",sec

c=a/b
CENTRAL
PROCESSOR

Hitachi HITAC 8000 Series (Japan)
H-8200

H-8210

182 or 465
394 or 1218

Number of Index Registers

0

0

16 max

Indirect Addressing

None

None

None

Special Editing Capabilities

Good

Good

Good

Boolean Operations

AND, INC OR, EXC OR

AND, INC OR, EXC OR

AND, INC OR, EXC OR

Table Look-up

None

None

None

Console Typewriter

Optional

Optional

Optional

Input-Output Channels

1 selector;
1 multiplexor

1 with
6 truuks

o to 2 selector;
Program compatible with
IBM System/360

Features and Comments

H-8210

Model Number

1 multiplexor

H-8200

H-8300
Core

Core

Core

Minimum

8192

8192

32,768

Maximum

32,768

16,384

65,536

Decimal Digits

65,536

32,768

131,072

Characters

32,768

16,384

65,536

Cycle Time, IIsee

1.4 per byte

2.0/byte

1.44/2 bytes

Effective Transfer Rate, char/sec

750,000

250,000

695,000

Checking

Parity

Parity

Parity

Storage Protection

None

None

Write only

Type of Storage
Number of Words

Maximum
Total Storage
WORKING
STORAGE

16 general-purpose
registers in core
storage

Features and Comments

*

WIth optional equipment.
(s) Using subroutine.

11/69

fA.

AUERBACH
@

(Contd.)

CENTRAL PROCESSORS AND WORKING STORAGE

Hitaclu HIT AC 8000 Series (Japan
H-8400
H-8500

11:510.107

Hitaciu HIT AC 3010 (Japan)

System Identity

8 per byte

8 per byte

6 + panty

2

2

1

Decimal Digits

Binary Bits

Characters

Word Length

1

1

1

Bmary

Binary

DeClmal

24 or 56 bits

24 or 56 bits

8 dIgIts

Fraction Size

7 bits

7 b,ts

2 digits

Exponent Size

H-8400

H-8500

H-3045, 3055

Binary, decimal

Binary, deCImal

Decimal

Variable

Variable

1 to 44 char

2, 4, or 6

2, 4, or 6

10 char

Instruction Length, Words

2

Addresses per lnstruction

DATA
STRUCTURE

Radix

Floating Point
Representation

Model Number
Arithmetic Radix
Operand Length, Words

0, 1, or 2

0, 1, or 2

25

5.9

147

82

11.9

4200 (s)

c = ab
c=a/b

Likely Fixed
Point Execution
Times, #lsec (5
Digits Min.
Precision)

c=a+b

111

15.8

9000 (s)

37 or 53

8.6 or 11.3

(s)

c=a+b

68 or 212

13.4 or 26.9

(s)

c = ah

101 or 305

17.0 or42.9

(s)

c=a/b

Parity

Parity

Parity

Checking of Data Transfers

Yes, multilevel

Yes, multilevel

None

Program lnterrupt Facility

16 max

16 max

3*

Number of lndex Registers

None

None

Yes

Good

Good

Fair

AND, INC OR, EXC OR

AND, INC OR, EXC OR

AND, INC OR, EXC OR

None

Single char only

Optional

Optional

None

o to 3 selector;

o to 6 selector;
1 multiplexor

1 simultaneous
channel*

Program compatible with IBM System/360

CENTRAL
PROCESSOR

Indirect Addres sing

None

1 multiplexor

Likely Floating
Point Execution
Times, J1.sec

High-speed arithmetic
cirCUIts optional

Special Editing Capabilities
Boolean Operations
Table Look-up
Console Typewriter
Input-Output Channels

Features and Comments

H-8400

H-8500

H-3045, 3055

Core

Core

Core

32,768

65,536

20,000

Minimum

262,144

524,288

40,000

Maximum

Model Number
Type of Storage

Number of Words

524,288

1,048,576

40,000

Decimal Digits

262,144

524,288

40,000

Characters

1. 44 per 2 bytes

0.84/4 bytes

3.5/2 char

510,000

1,058,000

1~5,

Parity

Parity

Parity

Write only*

Write only'

None

16 general-purpose regIsters in fast scratchpad
memory

200

Other models have 7microsecond cycle time
per 2 characters

Maximum
Total Storage
Cycle Time, ILsec

Effective Transfer Rate, char/sec

WORKING
STORAGE

Checking
Storage Protection

Features and Comments

*

With optIOnal equipment.
(8) Using subroutine.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

11/69

COMPARISON CHARTS - NON-U, S, A, COMPUTERS

11;510,108

System Identity

Word Length
DATA
STRUCTURE
Floating Point
Representation

Binary Bits

6 + 2 punctuation + parity

Decimal Digits

1

Characters

1

Radis

Binary

Fraction Size

36 bits

Exponent Size

12 bits
E050

NIOO

N200

Arithmetic Radix

Decimsl

Decimal

Decimsl

1toN

1 to N

Operand Length, Words

1 to N

Instruction Length, Words

1 to N

1toN

1 to N

Addresses per Instruction

2

0, 1, or 2

0, I, or 2

c=a+b

123

123

84

c = ab

3100 (s)

3100 (s)

500

c=a/b

3700 (s)

3700 (s)

1134

c=a+b

-

-

-

c = ab

-

-

-

c=a/b

-

-

-

ParIty

Parity

Parity

Likely Floating
Point Execution
Times, J,lsec

Checking of Data Transfers
Program Interrupt Facility

Yes

Yes

Yes

Number of Index Registers

6

6

6, 15*

Indirect Addressing

Optionsl

OptIOnal

Yes

Special Editing Capabilities

Poor; excellent*

Poorj excellent*

Excellent

Boolean Operations

AND, EXC OR

AND, EXC OR

AND, EXC OR

Table Look-up

None

None

None

Console Typewriter

None

Yes

Yes

Input-Output Channels

2; 3*

2; 3*

3; 4*

Includes built-in
I/O control

IBM 1401/1410/7010
compatible through
software
N200M

Features and Comments
Model Number

E050M

N100M

Type of Storage

Core

Core

Core

Minimum

4096

2048

4096

Maximum

16,384

32,768

65,536

Decimal Digits

16,384

32,768

65,536

Characters

Number of Words

Maximum
Total Storage

WORKING
STORAGE

2200 200

Model Number

Likely Fised
Point Execution
Times, /Lsec (5
Digits Min.
Precision)

CENTRAL
PROCESSOR

Nippon Electric NEAC Senes 2200 IJaDanl
2200 100

2200 50

16,384

32,768

65,536

Cycle Time, IJsec

2.0/char

2.0/char

2.0/char

Effective Transfer Rate, char/sec

167,000

167,000

250,000

Parity

Parity

Parity

None

None

None

Checking

,

Storage Protection

Features and Comments

*

WIth optional equIpment.
(s) Using subroutine.

11/69

A

(Contd.)

AUERBACH
I!)

11:510.109

CENTRAL PROCESSORS AND WORKING STORAGE

2200 300

NIppon Electric NEAC Series 2200 (Japan)
2200 400

System Identity

22001500

6 + 2 punctuation + parity

Binary Bits

1

Decimal Digits

1

Characters

BInary

Radix

36 bIts

Fraction Size

12 bits

Exponent Size

N300

N400

Word Length
DATA
STRUCTURE
Floating Point
Representation

N500

Model Number

Decimal, bmary

Arithmetic Radix

1 to N

Operand Length, Words

1toN

Instruction Length, Words

0, 1, or 2

Addresses per Instruction

61.5

43

12

c=a+b

363

216

96

c = ab

850

612

196

c=a/b

c

=a

Likely Fixed
Point Execution
Times, "sec (5
Digits Min.
Precision)

+b

34.5*

26*

6

46.5'

32*

12

c = ab

51*

45*

21

c = alb

Parity

Parity

Panty

Checking of Data Transfers

Yes

Yes

Yes

Program Interrupt Facility

15; 30*

15; 30*

15 + 15 per program

Number of Index Registers

Yes

Yes

Yes

Excellent

Excellent

Excellent

Likely Floating
Point Execution
Times, IJ.sec

Indirect Addres sing

Special Editing Capabilities

AND, EXC OR

AND, EXC OR

AND, EXC OR

Optional

Yes

Yes

Table Look-up

Yes

Yes

Yes

Console Typewriter

4

4; 8*

8, 16*

IBM 1401/1410/7010 compatIble through software

CENTRAL
PROCESSOR

Boolean Operations

Input-Output Channels

Features and Comments

N300M

N400M

N500M

Core

Core

Core

Model Number
Type of Storage

16,384

16,384

65,536

Minimum

131,072

262,144

524,288

Maximum

131,072

262,144

524,288

Decimal Digits

131,072

262,144

524,288

Characters

1.5/char

l/char

1.5/8 char

333,000

500,000

1,777,000

Number of Words

Parity

Panty

Parity

Yes*

Yes*

Yes*

Maximum

Total Storage
Cycle Time, j.Lsec

Effective Transfer Hate, char/sec

WORKING
STORAGE

Checking
Storage Protection

Features and Comments

*

With optIOnal equipment.
(s) Using subroutine.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

11/69

11:510.110

COMPARISON CHARTS -

System Identity

S. A. COMPUTERS

Phillps PI000 Senes (Netherlands)
Binary Bits

8

Decimal Digits

2

Word Length
DATA
STRUCTURE
Floating Point
Representation

Characters

1

Radix

Bmary

Fraction Size

24 or 56 blts

Exponent Size

7 bits

Model Number

PllOO

Arithmetic Radix

Binary or decimal

Operand Length, Words

Up to 204

Instruction Length, Words

4 or 8

Addresses per Instruction
Likely Fixed
Point Execution
Times, jJsec (5
Digits Min.
Precision)
Likely Floating
Point Execution
Times, /Jsec
CENTRAL
PROCESSOR

NO~U.

P1200

P1400

lor 2

c=a+b

53.5

17.5

7.5

c = ab

167.5

69.5

17.0

c=a/b

217

?

16.5

c=a+b

67

26

9

c = ab

138

59

17

c=a/b

182

96

19

5

7

Checking of Data Transfers

Parlty

Program Interrupt Facility

Yes

Number of Index Registers

14

Indirect Addressing

Yes

Special Editing Capabilities

Excellent

Boolean Operations

AND, EXCOR

Table Look-up

Yes

Console Typewriter

Yes

Input-Output Channels

3

Features and Comments

Special hardware for automatIc alignment, rounding-off, and truncanon of decimal
numbers

Model Number

1100

Type of Storage

1200

1400

Core

Core

Core

Minimum

16,384

65,536

131,072

Maximum

65,536

262,144

524,288

Decimal Digits

131,072

524,288

1,048,576

Characters

65,536

262,144

524,288

Cycle Time, /Jsec

1

1

1

Effective Transfer Rate, char/sec

?

?

?

Checking

Parity

Parity

Parity

Storage Protection

Yes

Yes

Yes

Number of Words

Maximum
Total Storage

WORKING
STORAGE

1200 and 1400 can be extended with 2. 5-l'sec
core storage up to 14,680,064 bytes in mod-

Features and Comments

ules of 2, 097,152

*

With optional equipment.
(s) Using subroutine.

11/69

fA

AUERBACH
®

(Contd.)

CENTRAL PROCESSORS AND WORKING STORAGE

~CL

4-40

4-30

System 4 (United Kmgdom
4-50
4 70

11:510.111

System Identity

4-75

8 + 1 panty per byte

Binary Bits

2

Decimal Digits

1

Characters

Word Lengtb
DATA
STRUCTURE

Radix

Bmary fraction, hexadeCImal exponent

24 or 56 bits

Floating Point
Representation

Fraction Size

7 bits

Exponent Size

4-30

4-40

4-50

4-70

Model Number

4-75

Binary t deClmal

Arithmetic Radix

Variable

Operand Length, Words

2, 4, or 6

Instruction Length, Words

0, 1, or 2

Addresses per Instruction

50

33.8

25.2

4.82

6.12

c=a+b

673

119.4

82.0

9.17

10.5

c = ab

691

162.1

111.2

14.1

15.4

-

40.30r55.5

37.4 or 52.6

6.82or8.68

8.17 or 10.0

c = a +b

-

70.5 or 214

67.7 or 211

9.90 or 16.4

11.2 or 17.7

c = ab

-

104 or 308

101 or 305

13.7 or 23.6

15.0or25.0

Likely Fixed
Point Execution
Times, J,lsec (5
Digits Min.
Precision)

c = alb

Likely Floating
Point Execution
Times, /-Lsee

c = alb

Parity

Checking of Data Transfers

Yes, multilevel

Program Interrupt Facility

16/processor state

Number of Index Registers

None

Indirect Addressing

Good

Special Editing Capabilities

CENTRAL
PROCESSOR

Boolean Operations

AND, INC OR, EXC OR

Table Look-up

None

Console Typewriter

Yes
1 to 16 channels; combination
of selector, and 1 or 2
multIplexors

2 to 8 selector;
1 multiplexor

2 or 3 selectors;
1 multiplexor

Multiply/divide
decimal only

Program compatible with IBM System/360; 4-75 same as 4-70, but
has special pagmg hardware

4-30

4-40

Core

Core

32,192
65,536

4-50

4-70

InPut-Output Channels

Features and Comments

4-75

Model Number

Core

Core

Core

65,536

65,536

65,536

65,536

Minimum

131,072

262,144

1,048,576

1,048,576

Maximum

131,072

262,144

524,288

2,097,152

2,097,152

Decimal Digits

65,536

131,072

262,144

1,048,576

1,048,576

Characters

1.5/2 bytes

1. 5/2 bytes

1.44/2 bytes

0.9/4 bytes

0.9/4 bytes

513,000

465,000

694,000

2,222,000

1,900,000

Parity

ParIty

Parity

Parity

Parity

None

WrIte only*

Write only*

Write only

WrIte only

Type of Storage
Number of Words

EffectIve cycle time is 0.65 per
4 bytes with full mterleavmg

Maximum
Total Storage
Cycle Time, /.Lsee

Effective Transfer Rate, char/sec

WORKING
STORAGE

Checking
Storage Protection

Features and Comments

*

With optional equipment.
(s) Using subroutine.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

11/69

11:510.112

COMPARISON CHARTS - NON-U. S. A. COMPUTERS

System Identity

Word Length
DATA
STRUCTURE

Binary Bits

24 + parity

Decimal Digits

6.9

Characters

4

Radix

Binary

Fraction Size

37 bits

Floating Point
Representation

Exponent Size

2010

Arithmetic Radix

Binary

Operand Length, Words

lor 2

2020

Instruction Length, Words

1

Addresses per Instruction

0, 1, or 2

Likely Floating
Point Execution
Times, Ilaec

ICL 1900 Senes Umu,d_ KillKdom
1902A
1903A

1904A

8 bits

Model Number

Likely Fixed
Point Execution
Times I /.laee (5
Digits Min.
Precision)

CENTRAL
PROCESSOR

I

1901A

2030

2040

c=a+b

57

23

11

5.9

c = ab

128'

64'

32'

16.5

c=a/b

133'

70'

35'

21. 6

0= a +b

103'

83'

41'

11'

c = ab

203'

180'

90'

17'

o=a/b

233'

208'

104'

32'

4; 12'

10; 31'

Cheoking of Data Transfers

Parity

Program Interrupt Faoility

Yes

Number of Index Registers

3

Indireot Addressing

No

Speoial Editing Capabilities

Good

Yes, multIlevel

Boolean Operations

AND, INC OR, EXC OR

Table Look-up

None

Console Typewriter

Optional

Yes

Input-Output Channels

4; 7'

4; 8'

Features and Comments
Model Number

2010

2020

2030

2040

Type of Storage

Core

Core

Core

Core

Minimum

4,096

8,192

16,384

65,536

Maximum

16,384

32,768

65,536

262,144

Deoimal Digits

113,049

226,099

452,198

1,808,793

Characters

65,536

131,072

262,144

1,048,576

Cycle Time, /Jsec

4.0

3.0

1.5

0.75

Effective Transfer Rate, ohar/seo

37,000

615,000

1,230,000

2,220,000

Cheoking

Parity

Number of Words

Maximum
Total Storage
WORKING
STORAGE

Storage Protection

Features and Comments

*

With optional eqUIpment.

(8) USing subroutine.

11/69

A

(Contd.)

AUERBACH

'"

CENTRAL PROCESSORS AND WORKING STORAGE

ICL 1900 Serles (United Kmgdom)
1904 5F
1906E 1907E

904, 5E

11:510,113

1906F

System Identity

1907F

24 + parlty

Binary Bits

6.9

Word Length

Decimal Digits

Characters

4

DATA
STRUCTURE

Radix

BlUary

38 blts + slgn

Floating Point
Representation

Fraction Size

8 bits + sign

Exponent Size

2040, 2050

2042, 2052

2060, 2070

2062, 2072

Model Number

Binary

Arithmetic Radix

lor 2

Operand Length, Words

1

Instruction Length, Words

0, lor 2

Addresses per Instruction

11.8

7.2

14.4

9.4

29.9

23.6

31. 8

26.3

36

28.2

39.5

32.1

31.6

15

37.6

21

c=a+b

47.6

22

53.6

28

c = ab

69.6

36

75.6

42

Likely Fixed
Point Execution
Times, /Jsec (5
Digits Min.
Precision)

c=a+b
c = ab
c=a/b

Likely Floating
Point Execution
Times, llsec

c=a/b

Panty

Checking of Data Transfers

MultIlevel

Program Interrupt Facility

3

Number of Index Registers

Yes

CENTRAL
PROCESSOR

Indirect Addres sing

Good

Special Editing Capabilities

AND, EXC OR, INC OR

Boolean Operations

None

Table Look-up

Standard

Console Typewr!ter

6 to 30'

12 to 60'

2050 and 2052 have autonomous floatmg
point unit; multiprogramming;
program compatlble.

Dual processor systems. 2070 & 2072
have autonomous floating pomt UnIt on
each processor.

2040, 2050

2042, 2052

2060, 2070

2062, 2072

32,768

32,768

65,536

65,536

Minimum

262,144

262,144

262,144

262,144

Maximum

1,808,793

1,808,793

1,808,793

1,808,793

Decimal Digits

1,048,576

1,048,576

1,048,576

1,048,576

Characters

1.8

0.75

1.8

0.75

1,110,000

1,700,000

830,000

1,140,000

Input-Output Channels

Features and Comments
Mode! Number

Core

Type of Storage
Number of Words

Maximum

Total Storage
Cycle Time, IJsec

Effective Transfer Rate, char/sec

Parity

WORKING
STORAGE

Checking

Variable datum and hmit registers

Storage Protection

Dual processor systems sharing
common core store

Features and Comments

*

With ophonat equipment.
(s) Using subroutine,

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

11/69

11:510.114

COMPARISON CHARTS - NON-U.S,A. COMPUTERS

System Identity

4004 55

Binary Bit&

8/byte

8/byte

8/byte

8/byte

8/byte

Decimal Digits

2/byte

2/byte

2/byte

2/byte

2/byte

Characters

l/byte

l/byte

l/byte

l/byte

l/byte

Radix

-

Binary

Bmary

Binary

Fraction Size

-

24 or 56 bIts

24 or 56 bIts

24 or 56 bIts

Exponent Size

-

-

7 bIts

7 bits

7 bits

Model Number

4004/15

4004/25

4004/35

4004/45

4004/55

Arithmetic Radix

Binary, decimal

Binary, decimal

Binary, decimal

Binary, decimal

Binary, decimal

Operand Length, Words

Variable

Variable

Variable

Variable

Variable

Instruction Length, Words

4 or 6 bytes

2, 4, or 6 bytes

2, 4, or 6 bytes

2, 4, or 6 bytes

2, 4, or 6 bytes

Addresses per Instruction

0, lor 2

0, lor 2

0, lor 2

0, lor 2

0, lor 2

c=a+b

76

49

51 or 80

25 or 42

7, 8 or 20

Word Length
DATA
STRUCTURE
Floating Point
Representation

Likely Fixed
Point Execution
Times, /Jsec (5
Digits Min.
Precision)
Likely Floating
Point Execution
Times, /Lsec

c = ab

(5)

445

163 or 287

82 or 134

18 or 62

c=a/b

(5)

185

243 or 206

106 or 111

23 or 25

c = a +b

-

81 or 116

37 or 53

13 or 18

c = ab

-

-

c=a/b
CENTRAL
PROCESSOR

Siemens System 4004 (West Germanyl
4004 25
4004 35
4004 15

4004 15

Checking of Data Transfers

203 or 536

68 or 212

23 or 50

-

445 or 1282

101 or 305

28 or 84

Parity

Panty

Parity

Parity

Parity
Yes, multilevel

Program Interrupt Facility

Yes, limIted

Yes,4 levels

Yes, multilevel

Yes I multilevel

Number of Index Registers

None

15 max

16 max

16 max

16 max

Indirect Addressing

None

None

None

None

None

Special Editing Capabilities

Fair

Fair

Good

Good

Good

Boolean Operations

AND, INC OR
EXC OR,

AND,INC OR
EXCOR,

AND, INC OR
EXC OR

AND, INC OR,
EXC OR

AND, INC OR,
EXC OR

,

Table Leok-up

None

None

None

None

None

Console Typewriter

Optional

Optional

Optional

Optional

Optional

1 with 6 subchannels. 2
simultaneous

4408 selector
channels, 0 or 1
multiplexor

o to 2 selector
channels; 1
multiplexor

o to 3 selector

o to 6

Input- Output Channels

channels; 1
multiplexor

selector
channels; 1
multiplexor

Model Number

4004/15

4004/25

4004/35

4004/45

4004/55

Type of Storage

Core

Core

Core

Core

Core

4,096 bytes

16,384 bytes

16,384 bytes

16,384 bytes

65, 536 bytes

Maximum

16,384 bytes

65,536 bytes

65,563 bytes

262, 144 bytes

524,288 bytes

Decimal Digits

32,768

131,072

131,072

524,288

1,048,576

Characters

16,384

65,536

65,536

262,144

524,288

Features and Comments

Minimum

Number of Words

Maximum
Total Storage

WORKING
STORAGE

Cycle Time, ILsec

2.0/byte

1. 5/4 bytes

1. 44/2 bytes

1.44/2 bytes

.84/4 bytes

Effective Transfer Rate, char/sec

250,000

1,333,333

507,000

679,000

1,201,000

Checking

Parity

Parity

Panty

Parity

Parity

Storage Protection

None

None

Write only*

Write only'

Wnte only'

No generalpurpose
registers

15 general-

16 generalpurpose
registers In
core storage

16 general-

Features and Comments

purpose

16 generalpurpose
registers In
core storage

purpose

registers in
core storage

registers in
core storage

'" With optional equipment.
(s) llslOg subroutine.

11/69

fA

AUERBACfl
@

(Contd.)

11:510.115

CENTRAL PROCESSORS AND WORKING STORAGE

SIemens System 300 (West Germany)

301

303

302

304

305

306

-

System Identity

24

24

24

24

24

24

4/word

4/word

4/word

4/word

4/word

4/word

Decimal Digits

4/word

4/word

4/word

4/word

4/word

4/word

Characters

-

-

-

-

Bmary

Bmary

Radix

-

-

-

24 bits

24 or 34 bits

Fraction Size

-

-

-

-

10 bits

10 bits

Exponent Size

301

302

303

304

305

306

Bmary

Bmary

Bmary

Binary

Binary

Binary

1

1

1

1

1

1

Operand Length, Words

1

1

1

1

1

1

Instruction Length, Words

1

11

1

1

1

1

Addresses per Instruction

11

I

9

30

8

8

4

-

-

29

23

23

13

-

-

-

23

23

13

Binary Bits

Word Length
DATA
STRUCTURE
Floating Point

Representation

Model Number
Arithmetic Radix

c=a+b
c

Likely Fixed
Point Execution
Times, f.lsec (5
Digits Min.
Precision)

= ab

c~a/b

-

-

-

-

29

14 or 15

-

-

14 or 18

-

-

26

-

-

26

11 or 13

None

None

None

None

None

Parity

Yes, 2 levels

Yes, multilevel

Yes, multllevel Yes, multIlevel Yes, multIlevel Yes, multilevel

Program Interrupt Facility
Number of Index Registers

None

None

None

None

None

16

Yes

Yes

Yes

Yes

Yes

Yes

None

None

None

None

None

None

AND,OR

AND, OR

AND, OR

AND, OR

AND, OR

AND, OR

None

None

None

None

None

None

Optional

Standard

Standard

Standard

1 multiplexorchannel wlth
6 trunks

1 multIplexor-channel with 10 trunks, 1 highspeed channel with 5 trunks

1 integrated
channel, 1
high-speed
!channel

1 multiplexor
channel with
5 trunks

Standard

c::::: a +b

c

=

Likely Floftting
Point Execution
Times, /-Lsee

ab

c ~ alb

Checking of Data Transfers

CENTRAL
PROCESSOR

Indirect Addressing
Special Editing Capabilities

Boolean Operations
Table Look-up

Console Typewriter

Standard

Input-Output Channels

Features and Comments

:,01<

Model Number

301

302

303

304

Core

Core

Core

Core

Core

Core

4,096

8,192

4,096

8,192

8,192

16,384

Minimum

16,384

16,384

16,384

16,384

16,384

65,536

Maximum

65,536

65,536

65,536

65,536

65,536

262,144

Decimal Digits

65,536

65,536

65,536

65,536

65,536

262,144

Characters

1,6/word

L 5/word

8,3/6'bits

1,5/word

L5/word

0,6/word

Cycle Time, J.Lsec

2,600,000

668,000

120,000

2,668,000

2,668,000

2,220,000

Effective Transfer Rate, char/sec

None

None

None

None

None

Parity

None

None

Yes, write
only

Yes, write
only

Yes, write
only

Yes, write
only

305

Type of Storage

Number of \Vards

Maximum
Total Storage
WORKING
STORAGE

Checking

Storage Protection

Features a.nd Comments

*

\Vlth optIOnal eqUlpment.

(5) Using subroutine.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc,

11/69

-A ;;D~p

-

AUERBAC~

•

REPORTS

11:520.101
COMPARISON CHARTS - NON-U. S. A. COMPUTERSI
AUXILIARY STORAGE AND MAGNETIC TAPE

COMPARISON CHARTS NON-U. S. A. COMPUTERS
AUXILIARY STORAGE AND MAGNETIC TAPE

An introduction to the Auxiliary Storage and Magnetic Tape Section of the Comparison
Charts, giving the precise meaning of each entry, can be found on Page 11:220. 101.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

COMPARISON CHARTS -

11;520.102

System Identity

NON-U.S.A. COMPUTERS

A/S Regnecentralen RC 4000 (Denmark)

Model Number

RC 4320

RC 433

Type of Storage

Drum

Disc

Units On-Line

16

6/controller

Read/Write
Operations

16

l/controller (max 15)

Minimum

65,536

2,048,000

Maximum

524,288

12,288, 000

10. 5

94

Maximum
Number

Seek Operations
Number of
Words per
Unit
Maximum
Total
Storage
AUXILIARY
STORAGE

Decimal Digits
Characters

Rotational Time, msec
13

Minimum
Waiting
Time,
msec

Average (Random)

163

Maximum

Effective Transfer Rate, char/sec

150,000

Sector Size, char

256 plus 1 parity word

120,000

Transfer Load Size I char
Parity plus status word

Checking

Features and Comments

Model Number

Maximum
Number
of Units

On-Line

1/controller (max 64 controllers)

Reading/Writing

Maximum 16

Searching
Rewinding

Demands on
Processor,

%
Transfer
Rate, kilochar/sec
MAGNETIC
TAPE

RC 709

RC 707

All

Reading/Writing
Starting/ Stopping
Peak

36

25

9

36

1, OOO-char blocks

22.5

17.5

7.8

22.5

5.1

4.8

3.6

5.1

100-char blocks
Tape Speed, inches/sec

45

Data Tracks

6

Data Rows per Block

4 to 65,538

Data Rows per Inch

800

IBM 729 Compatible

Yes

8
4 to 87,380
556

200

800

IBM 2400 Compatible
Reading

Parity

Writing

Parity, read after write

Checking

Read Reverse

Feature s and Comments

"'With optional equipment.
AUE RBACH Computer Characteristics Digest

11/69

A

(Contd.)

AUERBACH
@

AUXILIARY STORAGE AND MAGNETIC TAPE

Bull Gamma 10
(France)

11:520.103

Bull-GE-GE-55
(France)

Elbit 100 (Israel)

GE-55

CLC-1

Drum

Magnetic drum

System Identity
Model Number

2

Type of storage
Units On-Line

1

1/trunk

2

l/unit

89 600

50, 000

Read/Write
Operations

Number

Seek Operations
Minimum

89 600

100 000

Maximum

358,400

300,000

Decimal Digits

200,000

Characters

179 200

Maximum

Number of
Words per
Umt
MaxImum
Total
Storage

Rotational Time, mseo

30

AUXILIARY
STORAGE

Minimum

0
15

WaIting
Time,
msec

Average (Random)

30

Maximum

70,000

Effective Transfer Rate, char/sec
Variable

1,400

Sector Size, char

?

Transfer Load Size, char

?

Checking

Features and Comments

No auxiliary storage
devices as yet

Kennedy 1400/360

MFU 35

Peripheral Equipment

Model Number
On-Line

2

Reading/Writing

2

l/unlt

1/unit

0

1/unit

1/unit

Searching

2

l/unit

1/unit

Rewinding

?

Reading/Writing

?

Starting/Stopping

Maximum
Number
of Units

Demands on

Processor,
500 bytes/sec

34

25 inches/sec

%

Peak
1, OOO-char blocks

Transfer
Rate, kilochar/sec

lOa-char blocks
4

25
9

Tape Speed, incbes/sec

7 or 9

MAGNETIC
TAPE

Data Tracks

Variable

Data Rows per Block

100

800

800

Data Rows per Incb

No

Yes

No

IBM 729 Compatible

No

No

No

IBM 2400 Compatible

Parity

Parity

Reading

Parity

Parity

Writing

No

No

Checking

No magnetic units
announced as yet

Read Reverse

Features and Comments

*With optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

11/69

COMPARISON CHARTS - NON-U. S. A. COMPUTERS

11:520.104

Fujitsu FACOM 270 Series (Japan)

System Identity
Model Number

F627A

F631A

F631B

Twe of Storage

Drum

Dlsc

Disc

Maximum

Number

Numher of
Words per
Unit

Maximum
Total
Storage
AUXILIARY
STORAGE

Units On-Line

8/channel

8/channel

8/channel

~/write
rations

l/channel

l/channel

l/channel

Seek Operations

l/channel

l/channel

l/channel

Minimum

524 x 183 hytes

33.5 x 106 bytes

67.1 x ~06 bytes

Maximum

524 x 183 bytes

33.5 x 106 bytes

67. 1 x 106 bytes

Decimal Digits

1, 048 x 8 x 103 / channel

67. 1 x 106 x 8/channel

134 x 106 x 8/channel

Characters

524 x 8 x 103 / channel

33.5 x 106 x 8/channel

67. 1 x 106 x 8/ channel

0

Rotational Time, msec
Minimum

0

0

Average (Random)

8.4

150

150

Maximum

17

290

290

150 x 103

56 x 103

In 95 x 103

Transfer Load Size, char

16 - 65 x 103 bytes

256 - 131 x 103 bytes

256 - 262 x 103 bytes

CheCking

Check bytes

Check bytes

Check bytes

Features and Comments

Floating head

Model Number

F603D

F603E

F603F

On-Line

8/ controller

8/ controller

8/controller

Reading/Writing

1/controller

1/controller

l/controller

Searching

1/controller

1/controller

l/controller

Rewinding

8/controller

8/controller

8/controller

Reading/Writing

CPU dependent

Starting/ Stopping

CPU dependent

Waiting
Time,
msec

Effective Transfer Rate, char/sec

Out 19.0 x 103

Sector Size, char

Maximum
Number
of Units

Demands on
Processor,

%
Transfer
Rate~ kilochar/sec
MAGNETIC
TAPE

Peak

41. 7/60

66.7/96

60

I, OOO-char blocks

28.9/36.7

46.4/59

39.5

100-char blocks

7.67/8.12

12.4/13.2

9.7

Tape Speed, inches/sec

75

120

5

Data Tracks

7

7

9

Data Rows per Block

Variable

Data Rows per Inch

55G/800

556/800

800

IBM 729 Compatible

Yes

Yes

No

IBM 2400 Compatible

Yes

Yes

Reading

Lateral and IOllgltudmal parity

Writing

Read-after-write parity

Checking
Read Reverse

Yes

Yes

Features and Comments

Cross call

Cross call

'WiU, optional equipment.
AUERBACH Computer Characteristics Digest

11/69

A

(Contd.)

AUERBACH

'"

AUXILIARY STORAGE AND MAGNETIC TAPE

11:520.105

Fujitsu FACOM 270 Series (Japan)

System Identity
Model Number

F631K

F461K

Disc

Disc pack

4/cbannel

4/channel

Units On-Lme

l/cbannel

l/channel

Read/Write
Operations

l/channel

l/channel

Seek Operations

90 x 106 bytes

7.25 x 106 bytes

90 x 106 bytes

7.25 x 106 bytes

180 x 106 x a/channel

14.5 x 106 x 8/channel

Decimal Digits

90 x 106 x 8/channel

7.25 x 106 x a/channel

Characters

Type of storage

Minimum
Maximum

Maximum
Number

Number of
Words per
Unit
Maximum
Total
Storage

Rotational Time, msec

0

0

130

87.5

270

160

130 x 103 bytes

156 x 103 bytes

Variable

Variable

Check bytes

Check bytes

AUXILIARY
STORAGE

Minimum
Walbng
Time,
msee

Average (Random)
Maximum

Effective Transfer Rate, char/sec
Sector Size, char
Transfer Load Size, char

Checking

Features and Comments

Model Number

F603G

F401A

8/controller

3/controller

On-Line

1/controller

l/oontroller

Reading/Writing

1/cOlltroller

None

Searching

8/eontroller

None

Rewinding

CPU dependent

Reading/Writing

CPU dependent

Starting/Stopping

96

1. 67

Peak

63.7

1.18

I, OOO-char blocks

15.8

0.32

100-char blocks

120

30

9

4

Variable

Variable

Maximum
Number
of Units

Demands on
Processor,

%
Transfer
Rate, kilochar/sec

MAGNETIC
TAPE

Tape Speed, Inches/sec
Data Tracks
Data Rows per Block

800

333

Data Rows per Inch

No

None

IBM 729 Compatible

Yes

None

IBM 2400 Compatible

Lateral, longitudinal and diagonal parity

Track parity

Reading

Read-atter-wrIte parity

Double write

Writing

Checking
Read Reverse

Features and Comments

'Wlth optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

11/69

COMPARISON CHARTS - NON-U. S.A. COMPUTERS

11:520.106

System Identity
Model Number

F622D

F623A

F624B

Type of storage

Drum

Drum

Drum

Units On-Line

8/channel

8/channel

8/channel

Read/Write
Doerations

Maximum

l/channel

l/channel

l/cbannel

Seek Operations

l/channel

l/cbannel

l/channel

Number of
Words per
Unit

Minimum

131 x 103 bytes

262 x 103 bytes

2,096 x 103 bytes

Maximum

131 x 103 bytes

262 x 103 bytes

2,096 x 103 bytes

Maximum
Total
Storage

Decimal Digits

262 x 8 x 103 / channel

524 x 8 x 103 / channel

4,192 x 8 x 103 /channel

Characters

131 x 8 x 103 / channel

262 x 8 x 103 / channel

2,096 x 8 x 103 /channel

0

0

0

Average (Random)

10

20

17

Maximum

20

40

34

25 x 103 bytes

27 x 103 b~ tes

120 x 103 bytes

Transfer Load Size, char

1 - 16 x 103 bytes

256 - 262 x 103 bytes

256 - 131 x 103 bytes

Checking

Parity

Parity

Check bytes

Features and Comments

Fixed head

Fixed head

Floating head

Model Number

F606A

F603B

F603C

6/controller

8/ controller

8/controller

1/controller

1/controller

1/controller

Searching

l/controller

1/controller

1/controller

Rewinding

6/controller

8/controller

8/controller

Reading/Writing

CPU dependent

CPU dependent

CPU dependent

Number

AUXILIARY
STORAGE

Fujitsu FACOM 270 Series (Japan)

Rotational Time, msec
Minimum

Waiting
Time,
msee

Effective Transfer Rate, char/sec
Sector Size, char

Do-Line
Maximum
Number
of Units

Demands on
Processor,

%
Transfer
Rate l kilocbar/sec
MAGNETIC
TAPE

Reading/Writing

Starting/Stopping

CPU dependent

CPU dependent

CPU dependent

Peak

15/25

15/41. 7

24/66.7

I, OOO-char blocks

11.8/15.6

12.9/28.9

20.7/46.4

100-char blocks

3.26/3.57

5.77/7/67

9.33/12.4

Tape Speed, inches/sec

45

75

120

Data Tracks

7

7

7

Data Rows per Block

Variable

Variable

Variable

Data Rows per Inch

333/556

200/556

200/556

IBM 729 Compatible

Yes

Yes

Yes

IBM 2400 Compatible

Yes

Yes

Yes

Yes

Yes

Cross call

Cross call

Reading

Lateral and longitudinal parity

Writing

Read-after-write parity

Checking
Read Reverse

Yes

Features and Comments

'With optional eqUipment,
AUERBACH Computer Characteristics DIgest

A

(Contd.)

AUERBACH

'"

AUXILIARY STORAGE AND MAGNETIC TAPE

11:520.107

System Identity

FUjItsU FACOM 230 Series (Japan)
F624B

F627A

F631A

Model Number
Type of Storage

Drum

Drum

DISC

8/channel

8/channel

8/channel

UnIts On-Lme

l/channel

l/channel

l/channel

Read/Write
Operations
Seek Operations

l/channel

l/channel

l/channel

2,096 x 103 bytes

524 x 183 bytes

33. 5 x 106 bytes

Minimum

2, OS6 x 103 bytes

524 x 103 bytes

33. 5 x 106 bytes

Maximum

4,192 x 8 x 103 /channel

1, 048 x 8 x 103 / channel

67.1 x 106 x 8/channel

Decimal Digits

2,096 x 8 x 103 /channel

524 x 8 x 103 /channel

33.5 x 106 x 8/channel

Characters

Maximum
Number

Number of
Words per
Vmt
Mrunmum

Total
Storage

Rotational Time, msec
0

0

0

17

8.4

150

34

17

290

120 x 103 bytes

150 x 103 bytes

56 x 103 bytes

256 - 131 x 103 bytes

16 - 65 x 103 bytes

256 - 131 x 103 bytes

Check bytes

Check bytes

Check bytes

Floating head

Floating head

F603D

F603E

F603F

8/controller

8/controller

8/controller

On-Line

1/controller

1/controller

l/controller

Reading/Writing

l/controller

l/controller

1/controller

Searching

8/controller

8/ controller

8/controller

Rewinding

AUXILIARY
STORAGE

Minimum

WaIting
Time,
mseo

Average (Random)
Maximum

Effective Transfer Rate, char/sec

Sector Size, char
Transfer Load Size, char
Checking

Features and Comments

Model Number

Reading/Writing

CPU dependent

Starting/Stopping

CPU dependent
41. 7/60

66.7/96

60

28.9/36.7

46.4/59

39.5

7.67/8.12

12.4/13.2

9.7

75

120

5

7

7

9

Maximum
Number
of Units

Demands on
Processor,

%

Peak
1, OOO-char blocks

Transfer
Rate, kilochar/sec

100-char blocks

MAGNETIC
TAPE

Tape Speed, inches/sec

Data Tracks
Data Rows per Block

Variable
556/800

556/800

800

Yes

Yes

No

Data Rows per Inch
IBM 729 Compatible
IBM 2400 Compatible

Yes
Lateral and longitudinal panty

Reading

Read-after-write parIty

Writing

Checking
Read Reverse

Yes

Yes

Cross call

Cross call

Features and Comments

'With optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

11/69

AUXILIARY STORAGE AND MAGNETIC TAPE

11:520.108

Fujitsu FACOM 230 Series (Japan)

System Identity
Model Number

F631B

F631K

F461K

Type of Storage

Disc

Disc

DISC pack

Maximum
Number

AUXILIARY
STORAGE

Units On-Line

8/chanoel

4/channel

8/chanoel

Read/Write
Operations

l/chanoel

l/chanoel

l/chanoel

Seek Operations

l/channel

l/chanoel

l/chanoel

Number of
Words per
Unit

Minimum

67. 1 x 106 bytes

90 x 106 bytes

7.25 x 106 bytes

Maximum

67. 1 x 106 bytes

90 x 106 bytes

7.25 x 106 bytes

Maximum
Total
Storage

Decimal Digits

134 x 106 x 8/chanoel

180 x 106 x 8/chanoel

14.5 x 106 x 8/chanoel

Characters

67. 1 x 106 " 8/chanoel

90 x 106 x 8/ chanoel

7.25 x 106 x 8/channel

Rotational Time, meeo
Minimum

0

0

0

Average (Random)

150

130

87.5

Maximum

290

270

160

In 95 x 103 bytes
Out 120 x 103 bytes

130 x 103 bytes

156 x 103 bytes

Transfer Load Size, char

256 to 262 x 103 bytes

Variable

VarIable

Checking

Check bytes

Check bytes

Check bytes

Waiting
Time,
msec

Effective Transfer Rate, char/sec

Sector Size, char

Features and Comments

Model Number

Maximum
Number
of Units

Demands on
Processor,

%
Transfer
Rate, kilochar/sec
MAGNETIC
TAPE

F603G

F401A
3/controller

On-Line

8/controller

Reading/Writing

1/controller

1/controller

Searching

1/controller

None

Rewinding

8/controller

None

Reading/Writing

CPU dependent

Starting/Stopping

CPU dependent

Peak

96

1. 67

1, ODD-char plocks

63.7

1.18

100-char blocks

15.8

0.32

Tape Speed, inches/sec

120

30

Data Tracks

9

4

Data Rows per Block

Variable

Variable

Data Rows per Inch

800

333

IBM 729 Compatible

No

None

IBM 2400 Compatible

Yes

None

Reading

Lateral, longitudinal and diagonal parity

Track parity

Writing

Read-after-write panty

Double write

Checking
Read Reverse

Features and Comments

*Wl th optional equipment.
AUERBACH Computer Characteristics Digest

11/69

A

(Contd.)

AUERBACH

'"

11:520.109

AUXILIARY STORAGE AND MAGNETIC TAPE

FUjItsu F ACOM 230 Series (Japan)

System Identlty

F622D

F623A

Model Number

Core

Drum

Drum

Type of storage

3

8/channel

8/channel

Units On-Lme

6

l/channel

l/channel

Read/Write
Operations

-

l/channel

l/channel

Seek operations

262 x 103

131 x 103 bytes

262 x 103 bytes

262 x 103

131 x 103 bytes

262 x 103 bytes

262 x 9 x 3 x 103

262 x 8 x 103 / channel

524 x 8 x 103/ channel

Decimal Digits

262 x 6 x 3 x 103

131 x 8 x 103/channel

262 x 8 x 103 /channel

Characters

Minimum

Maximum

Maximum
Number

Number of
Words per
Umt
MaJomum

Total
Storage

Rotational Time, msee
6 Ilsec

0

0

6 Ilsec

10

20

Average (Random)

6 J1,sec

20

40

Maximum

6 x 10 6 bytes

25 x 103 bytes

27 x 103 bytes

Variable

1 - 16 x 103 bytes

256 - 262 x 103 bytes

Parity

Parity

Parity

Fixed head

Fixed head

F606A

F603B

F603C

6/controller

8/controller

8/controller

On-Line

l/controller

1/controller

l/controller

Reading/Writing

l/controller

1/controller

l/controller

Searching

6/controller

8/controller

8/ controller

CPU dependent

CPU dependent

CPU dependent

Reading/Writing

CPU dependent

CPU dependent

CPU dependent

Starting/Stopping

15/25

15/41. 7

24/66.7

11. 8/15. 6

12.9/28.9

20.7/46.4

1, OOO-char blocks

3.26/3.57

5.77/7.67

9.33/12.4

100-char blocks

45

75

120

7

7

7

AUXILIARY
STORAGE

Minimum

Waitmg
Time,
msec

Effective Transfer Rate, char/sec
Sector Size, char

Variable

Variable

Variable

333/556

200/556

200/556

Yes

Yes

Yes

Yes

Yes

Yes

Transfer Load Size, char
Checking

Features and Comments

Model Number

Maximum

Number
of Units

Rewinding
Demands on
Processor,

%

Peak
Transfer
Rate, kilochar/sec
MAGNETIC
TAPE

Tape SPeed, inches/sec
Data Tracks
Data Rows per Block
Data Rows per Inch
ffiM 729 Compatible
IBM 2400 Compatible

Lateral and longitudinal par;ty

Reading

Read-after-write parity

Writing

Checking
Yes

Yes

Yes

Cross call

Cross call

Read Reverse

Features and Comments

'Wi th optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

11/69

COMPARISON CHARTS - NON-U. S. A. COMPUTERS

11:520.110

Hitachi HITAC 3010 (Japan)

System Identity
Model Number

H-366

Type of Storage

Disc

Units On- Line
Maximum
Number

Number of
Words per
Unit
Maximum
Total
Storage
AUXILIARY
STORAGE

2

Seek Operations

2

Minimum

22,118,400

Maximum

88,473,600

Decimal Digits

177 x 10 6

Characters

177 x 10 6

Rotational Time, rnsec
Waiting
Time,
msec

2

Read/Write
Operations

50

Minimum

0

Average (Rnndom)

105

Maximum

Effective Transfer Rate, char! sec

150
25,400

Sector Size, char

160

Transfer Load Size, char

1 to 1,600

Checking

Parity

Features and Comments

Model Number

Maximum
Number
of Units

Demands on
Processor,

%

H-382

H-197

H-581

H-582

H-3485

On-Line

12

12

14

14

Reading/Writing

2

2

2

2

Searching

0

0

0

0

Rewinding

All

All

All

All

Reading/Writing

Varies

Varies

Varies

Varies

Starting/Stopping

Varies

Vanes

Varies

Varies

Peak

10.0

30.0

55.0

33.3

66.7

120

1, OOO-char blocks

9.0

25.0

41. 6

30.0

42.0

75

100-char blocks

5.0

15.0

12.8

15.0

15.0

17

Tape Speed, inches/sec

30

60

100

100

100

Data Tracks

6

6

6 (2 bands)

Data Rows per Block

Variable

Variable

Variable

Data Rows per Inch

333

500

555

IBM 729 Compatible

No

No

No

Yes

No

No

No

No

Transfer
Rate, kilochar/sec
MAGNETIC
TAPE

H-381

IBM 2400 Compatible
Reading

Row panty

Writing

Read-after-write

150
6
VarIable

333

667

800

Checking
Read Reverse

Yes

Features and Comments

I
*With optional equipment.

11/69

fA

AUERBACH

'"

(Contd. )

AUXILIARY STORAGE AND MAGNETIC TAPE

11.520:111

Hitachi HITAC 8000 Series (Japan)
H-8564

H-8564-l2

H-8564-11

H-8564-21

System Identity

H-8566

H-8577

H-8568-11

Drum

Disc

Magnetic
cards

Model Number

Disc

Disc

DISC

Disc

8/trunk

2/trunk

2/trunk

l/trunk

16/trunk

8/channel

32/trunk

l/channel

l/channel

l/channel

l/channel

l/channel

l/channel

l/channel

l/unit

l/unit

l/unit

l/unit

10/unit

8/unit

l/unit

7. 25x 106

2.56xl0 6

5.12x 106

5.12x 106

1.6xl06

29. 2x 106

537x 106

Minimum

7. 25x 106

2.56xl0S

5.12x 106

5.12x 106

1. 6 x 106

233. 4x 106

537x lOS

Maximum

Type of Storage
Umts On-Line

Read/Write
Operations

Number

Seek Operations

116 x 106

10.2xl06

20. 5x 106

10. 2x 106

51. 2x 106

3. 7x 10 9

34.4x 109

Decimal Digits

58x 10 6

5.l2xlO S

10. 2x 10 6

5.12x 106

25.6xl0 6

1. 9x 10 9

17.2xl09

Characters

25

25

25

25

17.2

25

60

Number of

Words per
Umt

Mwnmum
Total
Storage

Rotational Time, msec

0

0

0

0

0

0

0

87.5

72.5

87.5

87.5

8.6

87.5

500

160

115

160

160

17.2

ISO

550

15S, 000

15S, 000

15S, 000

15S, 000

210,000

312,000

70,000

Variable

200

200

200

Vanable

Varlable

Variable

!to 36, 250

200

200

200

1 to 36, 250

1 to 146, 880

lt~

AUXILIARY
STORAGE

Minimum

WaIting
Time,
msec

Average (Random)
Maximum

16,384

Effective Transfer Rate, char/sec
Sector Size, char
Transfer Load Size, char

Cyclic check code

Checking

H-85S4 series uses changeable disc packs (similar to IBM 2311); H-8564-21 has 2 disc drives
per unit; H-8568-11 uses changeable cartrIdges holdmg 256 cards each; all mrunmum storage
capacities are based on 1 trunk.
H-8422

H-8432

8/trunk

16/trunk

16/trunk

l/channel

l/channel

l/channel

H-8442

H-8445

H-8451

Features and Comments

H-8453

Model Number
On-Line
Reading/Writing

0

0

0

Searching

All

All

All

Rewinding

Varies

Varies

VarIes

Reading/Writing

Varies

Varies

VarIes

Starting/Stopping

15.0

30.0

SO.O

120

no.o

120

12.7

20.4

40.0

81. 3

:10.6

61. 2

1, OOO-char blocks
100-char blocks

5.4

5.2

10.3

20.8

ri,6

11. 3

37.5

37.5

75

150

:17.5

75

8

8 Or 6*

8

Variable

Variable

Variable

400

800 (200, 556 or 800)*

1,600

No

Yes*

No

No

Yes

Yes

Track, row
parity

Tracks, row and diagonal

Row parity

Read-after-write

Maximum

Maximum
Number
of Units

Demands on
Processor,

%

Peak

Transfer
Rate, kilochar/sec
MAGNETIC
TAPE

Tape Speed, inches/sec
Data Tracks
Data Rows per Block
Data Rows per Inch
IBM 729 Compatible
IBM 2400 Compatible
Reading
Checking
Writing
Read Reverse

Yes

Features and Comments

'With optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc

11/69

11:520.112

COMPARISON CHARTS - NON-U. S. A. COMPUTERS

System Identity
Model Number

N271B

N274A-l

N274A-2

E271

E26l

Type of Storage

Drum

Disc

Disc

Drum

Disc

Units On-Line

4/control

l/control

l/control

4/ controller

4/ controller

~:m/Write

l/control

l/control

l/control

l/controller

1/controller

Seek Operations

-

l6/control

8/control

-

l/unit

Minimum

1.7x106

294 x 106

147 x 106

82,000

820,000

Maximum

1. 7 x 106

294 x 106

147 x 106

82,000

820,000

Decimal Digits

6.8 x 106/cont.

294 x 10B/cont.

147 x lOB /cont.

.34 x lOB /cont.

3.3 x lOB /cont.

Characters

B.8 x 106/cont.

294' x 106/ cont.

147 x 106 /cont.

. 34 x 106/cont.

3.3 x 106/cont .

34

25

16.7

34

0

0

0

0

Average (Random)

17

62.5

8.3

?

Maximum

34

135

16.7

534

Effective Transfer Rate, char/sec

900,000

208,330

103,000

83,333

Maximum
Number

Number of
Words per
Unit
Maximum
Total
Storage

AUXILIARY
STORAGE

Nippon Electric NEAC Series-2200 (Japan)

erations

Rotational Time, rosee
Minimum

Waiting
Time,
rosee

Sector Size, char

Variable

Variable

128

100

Transfer Load Size, char

Variable

Variable

1 to N

ltoN

Checking

Cyclic

Cyclic

Parity

Parity

Features and Comments

For
2200/700
only

16 drives per N274A-1;
8 drives per N274A-2

E261 uses changeable disc
packs; both units for 2200/50
only

N204A Scncs

Model Number

-1
On-Line

Maximum
Number
of Units

-3

4/controller

N204C Series
-13

-15

4/controller

E204
4/controller

Reading/Writing

l/controller

1/controller

1/controller

Searching

0

0

0

Rewinding

All

All

All

Reading/Writing

Varies widely

Varies widely

2 max

Starting/Stopping

0

0

Peak

32

64

88.8

I, OOO-char blocks

23.5

47

100-char blocks

7

14

Tape Speed, inches/sec

60

120

Data Tracks

8

8

6

Data Rows per Block

Variable

Variable

Variable

Data Rows per Inch

400

800

556

IBM 729 Compatible

No

No

Yes

IBM 2400 Compatible

No

Yes

No

Reading

Track and row parity;
Orthotronic system

Track and row parity;
cyclic redundancy

Track and row parity

Writing

None

Read-after-write

Read-after-write

No

Yes

Demands on
Processor,

%
Transfer
Rate, 1010char/sec
MAGNETIC
TAPE

-2

Checking
Read Reverse

400

0

28.8

64

65.3

19.4

43.2

7.1

19.4

5.0

11

2.54

120

36

80

16

556

Yes

Features and Comments

8.9

For 2200/50 only

*With 0Jltional eqUipment.
AUERBACH Computer Charactenstlcs Digest

11/69

A.

AUERBACH

(Contd. )

AUXILIARY STORAGE AND MAGNETIC TAPE

11 :520. 113

Nippon Electric NEAC Senes-2200 (Japan)

System Identity

N259

N261

N262

N271

N271A

Model Number

Disc

DISC

DISC

Drum

Drum

Type of storage

8/control

8/conLrol

4/control

8/control

8/control

Units On-Line

l/control

l/control

l/control

l/control

l/control

Read/Write
Operations

Maximum
Number

l/unit

l/unit

2/unit

-

-

9.2x106

134 x 10 6

268 x 10 6

2.6x10 6

327,700

9.2 x 10 6

134 x 106

268 x 10 6

2.6 x 10 6

327,700

73.6 x 10 6 /cont.

1.07

10 9 /cont.

20.8 x 10 6 /cont.

2.6 x 10 6 /cont.

Decimal Digits

73.6 x 10 6 /cont.

1. 07 x 10 9/cont.

1. 07 x 10 9 /cont.

20.8 x 10 6 /cont.

2.6 x 10 6 /cont.

Characters

25

51. 4

51. 4

50

16.7

0

0

0

0

0

?

104

104

27.5

8.3

190

171

125

55

16.7

208,000

188,000

188,000

106, 000

103,000

1 to N

1 to N

1 to N

128

?

1toN

1 to N

1 to N

1 to N

1 to N

Transfer Load Size, char

Cyclic

Cyclic

Cyclic

Cyclic

Parity

Checking

X

10 9 /cont.

1. 07

X

Seek Operations
Number of
Words per
Unit

Minimum
Maximum

MaxImum
Total
Storage

Rotational Time, msec

AUXILIARY
STORAGE

Minimum
Watting
Time,

Average (Random)

msec

Uses
changeable
disc pack

-1, -2

Maximum

Effective Transfer Rate, char/sec
Sector Size, char

Features and Comments

-2, -4

I -5

I

-7

N204 B Series
-8

I -9

Model Number

-11, -12

8/controller

On-Line

1/controller

Reading/Writing

0

Searching

All

Rewinding

Varies widely

Maximum
Number
of Units

Demands on
Processor,

Reading/Writing

0

%

Starting/Stopping

20

44

67

29

64

96

13.3

Peak

16

33.4

48

21.2

43.2

61. 5

10.7

1, OOO-char blocks

5.7

10.3

13. 6

6.3

11

14.6

3.8

36

80

120

36

80

120

24

Transfer
Rate, kilochar/sec

lOO-char blocks

MAGNETIC
TAPE

Tape Speed, inches/sec

6

Data Tracks

Variable
200,556

Data Rows per Block
200, 556, 800

Data Rows per Inch

556

Yes*

IBM 729 Compatible

No

IBM 2400 Compatible

Track and row parity

Reading

Read-after-write

Writing

Yes (optional on N204B-11/-12)

I

Checking

Read Reverse

N103 with characteristics simIlar 10 N204B-ll/-12 available for 2200/100 only

Features and Comments

*With optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc

11/69

11:520.114

COMPARISON CHARTS - NON-U. S. A. COMPUTERS

Philips P1000 Series
(Netherlands)

System Identity
Model Number

P1041

Type of storage

DiSC

Disc

Disc

4/trunk

8/control

8/trunk

1

l/trunk

l/trunk

Seek Operations

8

l/unit

l/unit

Minimum

7,250,000

7.25 x 106

5.84 x 106

7,250,000

7.'25 x 106

5.85 x 106

Number of
Words per
Unit

Maximum

Maximum
Totsl
Storage

Decimal Digits

116 x 106/control

116 x 106 /trunk

46. 7 x 106 /trunk

Characters

58 x 106/control

58 x 106/trunk

23.4 x 106/trunk

25

25

25

Minimum

25

6

0

Average (Random)

107.5

112.5

112.5
170

Rotational Time, maee
Waiting
Time,
msce

Maximum

175

170

Effective Transfer Rate, char/sec

156,000

156,000

156,000

Sector Size, char

Variable record length

3,600

3,600

Transfer Load Size, char

?

1 to 36,890

1 to 320

Checking

Check characters

Cyclic

Removable disc pack; Model
P1041-002 has reduced seek
time (775 mace average access
time)

Uses changeable disc pack;
Model 4-30 only

Uses changeable disc pack;
4-40 and higher numbered
models

P1064-1,
-2, -3

4450

4453

4454

8/trunk

60

120

30

40.5

81.1

20.2

10.3

20.7

5.2

Features and Comments

P1061-1,
-2 -3

Model Number

Maximum
Number
of Units

Demands on
Processor,

%
Transfer
Rate, kilochar/sec
MAGNETIC
TAPE

4425

Read/Write
Operations

Units On-LIne
Maximum
Number

AUXILIARY
STORAGE

ICL System 4 (United Kingdom)

On-Line

8/control

8/control

Reading/Writing

2/control

2/control

l/trunk

Searching

2/oontrol

2/control

0

Rewinding

8/control

a/control

All

Reading/Writing

Varies

Varies

Varies

Stsrtlng/Stopping

None

None

Varies

Peak

30

160

1, OOO-char blocks

20

140

I

90

J

61

6011201180

60

30 162 197

~7.

5

8.6

100-char blocks

4452

Tape f'lJeed, inches/sec

37.5175 1112• 5 37.5175 1112. 5 75

75

150

37.5

Dats Tracks

8 or 6*

8 or 6*

8

8

a

Dats Rows per Block

Varies

Varies

18-65K

Dats Rows per Inch

800

1,600

m,556, or

800

800

800

IBM 729 Compatible

Yes*

Yes*

Yes

No

No

No

IBM 2400 Compatible

Yes

Yes

No

Yes

Yes

Yes

6

Reading

Cyclic redundancy

Writing

Read-after-write

Read-after-write

Yes

Yes

Track and row parity; cyclic redundancy for 8-track

Checking
Read Heverse

Features and Comments

... With optional equipment.

AUERBACH Computer Characteristics DIgest

11/69

A

(Contrl. )

AUERBACH
®

AUXILIARY STORAGE AND MAGNETIC TAPE

11:520.115

System Identity

ICL System 4 (Umted Kmgdom)
4440

4441

4442

4443

Model Number

Disc

Disc

DISC

Disc

Type of Storage

4/trunk

8/trunk

4/trunk

8/trunk

Umts On-Lme

l/trunk

l/trunk

1/irunk

l/trunk

Read/Write
Operations

l/unit

l/unit

l/unit

l/unit

600 x 106

300 x.106

600 x 106

300 x 106

Minimum

600 x 106

300 x 106

600 x 106

300 x 106

Maximum

Seek Operations

5.6 x 109/trunk

Decimal Digits

2.8 x 19 9/trunk

Characters

MaxImum
Total
Storage

AUXILIARY
STORAGE

Minimum

0

Walting
Time,
msee

Average (Random)

80
100

Maximum

530,000

530,000

265,000

265,000

21 142

21,142

10,567

10,567

Effective Transfer Rate I char/sec
Sector Size, char
Transfer Load Size, char

1 to 338,000 (one cylmder)

Checking

Cyclic check

For all models except 4-30

For 4-70 and 4-75 only

4460

4461

Features and Comments

Model Number

4462

8/trunk

On-Line

l/trunk

Reading/Writing
Searching

0

Maximum
Number
of Units

Rewinding

All

Varies

Reading/Writing

Varies

Starting/Stopping

60

60

120

200

Peak

37.5

30

61

102

1, ODD-char blocks

8.6

5.6

11.3

19

75

37.5

75

125

6

8

8

8

Demands on
Processor,

%
Transfer
Rate, kilochar/sec

100-char blocks

MAGNETIC
TAPE

Tape Speed, inches/sec
Data Tracks
Data Rows per Block

18-65, 536
800

Number of
Words per
Unit

Rotational Time I msee

41.1

4458

Maximum
Number

Data Rows per Inch

1600

Yes

No

No

Yes

IBM 729 Compatible
IBM 2400 Compatible

Track and row parity; cyclic redundancy for 8-track

Reading

Read-after-write

Writing

Checking
Read Reverse

Yes

4460, 4461, and 4462 utihze phase encoded recordinl\ (NItZ recording optional)

Features and Comments

"'With op1icmal equipment.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc

11/69

11: 520. 116

COMPARISON CHARTS -

ICL 1900 Series (United Kingdom)

System Identity
Model Number

2801

2802

2820

2821

2805 Series

Type of Storage

Disc

DISC

Disc

Disc

Disc

8/control

8/control

4

4

14/control

Read/Write
Operations

l/control

l/control

1

1

l/control

Units On- Line
Maximum
Number

Number of
Words per

Unit

Maximum
Total
Storage
AUXILIARY
STORAGE

NON-U.S.A. COMPUTERS

Seek Operations

8/control

8/control

4

4

14/control

Minimum

1,048,576

2,097,152

409,600

819,200

26,214,400

Maximum

1,048,576

2,097, 152

409,600

819,200

104,857,600

Decimal Digits

57,881,395

115,762,790

11,304,960

22,609,920

10, 129,244, 160

Characters

33,554,432

67, 108,864

6,553,600

13,107,200

5,872,025,600

0

Rotational Time, msee
Minimum

0

0

0

0

Average (Random)

102.5

102.5

195

195

150

Maximum

180

180

325

325

240

208,000

208,000

208, 000

208,000

76,000 to 190,000

Transfer Load Size, char

512 to 2,097,152

512 t02, 097, 152

512 to 40,960

512 to 40,960

512 to 1, 638, 400

Checking

Check
characters

Check
characters

Check

characters

Check
characters

Cyclic check
code

Exchangeable
disc packs

Exchangeable

Twin exchangeable disc

disc packs

stores

Waiting
Time,

msee
Effective Transfer Rate, char/sec
Sector Size, char

Features and Comments

Model Number

1971

1972

1973

On-Line

6/control

6/control

6/control

Reading/Writing

l/control

l/control

1/control

Searching

0

0

0

Rewinding

6/control

6/control

6/control

Reading/Writing

Varies with processor model

Sharting/ Stopping

Varies with processor model

Peak

20.8

41. 7

60.

1, OOO-char blocks

15.1/13.6

28.0/26.4

33.4

100-char blocks

4.3/3.3

7.0/6.1

6.7

Tape Speed, inches/sec

37.5

75

75

Data Tracks

6

6

6

Data Rows per Block

Variable

Maximum
Number
of Units

Demands on
Processor,

%
Transfer
Rate, kilochar/sec
MAGNETIC
TAPE

2806 Series is
same but has
about double the
data capacities

Daha Rows per !uch

200,556

IDM 729 Compatible

Yes

IDM 2400 Compatible

No

a

200,556,800

Reading

Track and row parity

Writing

Read-after-write

Features and Comments

Short gap facility

Checking

Read Reverse

tWith optIonal equipment.
AUERBACH Computer Characteristics Digest

11/69

fA

AUERBACH

'"

(Contd.)

AUXILIARY STORAGE AND MAGNETIC TAPE

11:520.117

System Identity

ICL 1900 Series (Umted Kingdom)
1962

1963

1964

2851

Model Number

Drum

Drum

Drum

Drum

Type of storage

4/control

4/control

4/control

8/control

Units On- Lme

l/control

l/control

l/control

l/control

Read/Write
Operations

-

-

-

-

32,768

131,072

524,288

524,288

32 768

131,288

524,288

524,288

904,396

3,617,587

14,470,348

29,940,697

Decimal Digits

524,288

2,097,152

8,388,608

16,777,216

Characters

Maximum
Number

Seek Operations
Minimum
Maximum

Number of
Words per
Unit
Maximum
Total
Storage

Rotational Time, msee
Minimum

0

0

0

10

10

20

6.5

20

20

40

13

50,000

100,000

100,000

1,400,000

4 to 131,072

4 to 524,288

4 tc 2,097,152

512 tc 2, 097, 152

Character
parity

Character
parity

Character
parity

CyclIc check
code

0

AUXILIARY
STORAGE

Walting
Time,
msee

Average (Random)
Maximum

Effective Transfer Rate, char/sec
Sector Size, char

Transfer Load Size, char
Checking

Features and Comments

2505

2506

Model Number

2507

2501

2504

4/control

4/control

On-Line

l/control

l/control

Reading/Writing

0

0

Searching

No rewind

4/control

Rewinding

16.3

Varies with processor model

Reading/Writing

-

Varies wIth processor model

Starting/Stopping

20.0

80.0

160.0

40.0

80.0

8.8

48.0/30.8

88.3/53.2

30.4/22.4

58.5/40.6

4.5

10.5/4.8

17.5/7.7

9.6/4.5

17.1/7.5

75

37.5

75

Maximum
Number
of Units

Demands on

Processor,

%

Peak
1, OOO-char blocks

Transfer
Rate, kilochar/sec

100-char blocks

MAGNETIC
TAPE

Tape Speed, Inches/sec

150

37.5

8

8

-

variable

-

1600

No

No

IBM 729 Compatible

No

Yes

IBM 2400 Compatible

Cyclic check code

Vertical redundancy check

Read-after-write

Read-atter-write

Data Tracks
Data Rows per Block
800

IVertlCal
redur.dancy
dih~~ flc~c
redundancy
& ongitu On c eo cars

Data Rows per Inch

Reading
Checking
Writing

Read Reverse

Bit serial recording.
Cassette loaded. No
longer marketed

Read reverse and error correctIon faCIlitles are standard.
Run-on-facility

Features and Comments

>I<\Vi1h optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

11/69

COMPARISON CHARTS - NON-U.S.A. COMPUTERS

11:520.118

Siemens System 4004 (West Germany)

System Identity
Model Number

4004/564

4004/568-11

Type of storage

Disc

Maguetic cards
4/trunk

Units On-Line

8/trunk

Read/Write
Operations

l/chaunel

l/cbaunel

Seek Operations

l/unit

l/unit

Number of
Words per
Unit

Minimum

7.25 x 106 bytes

536 x 106 bytes

Maximum

7.25 x 106 bytes

536 x 106 bytes

Maximum
Total
Storage

Decimal Digits

116 x 106/trunk

4, 288 x 106/trunk

Characters

58 x 106/trunk

2,144 x 106 /trunk

Maximum
Number

AUXILIARY
STORAGE

?

?

Minimum

0

0

Average (Random)

87.5

523

Maximum

160

557

Effective Transfer Rate, char/sec

156,000

70,000

Sector Size, char

?

?

Transfer Load Size, char

1 - 36,260

1 - 16,384

Checking

Cyclic check

CycliC cneck

Features and Comments

Changeable "Disc Packs" (ruM
2311 Disk Storage Drive)

Changeable cartridges hold
256 cards each

Model Number

4004/432

4004/4443

4004/4446

4004/441

16/trunk

16/trunk

16/trunk

16/trunk

l/chaunel

l/channel

l/chaunel

Rotational Time, msee
Waiting
Time,
msee

On-Line
Maximum
Number
of Units

Demands on
Processor,

%

Reading/Writing

l/chaunel

Searching

0

0

0

0

Rewinding

All

All

All

All

Reading/Writing

Varies

Varies

Varies

Varies

Starting/Stopping

Varies

Varies

Varies

Varies

Peak

30.0

60.0

120.0

25.0

1, OOO-char blocks

20.4

40.5

81. 1

20.7

100-char blocks

5.2

10.3

20.7

8.1

Tapc Speed, inches/sec

?

?

?

?

Data Tracks

8 (6 Optional)

6

Data Rows per Block

Variable

Variable

Data Rows per Inch

800 (200, 556, or 800 with 7-channel feature)

333 or 500

ruM 729 Compatible

Only when 7-channel tape feature is Installed

ruM 2400 Compatible

Yes

Transfer

Rate, blochar/sec
MAGNETIC
TAPE

Yes

Reading

Track, row, and diagonal parity

Writing

Read-after-write

No
Yes

No
Track parity

Checking
Read-after-write

Read H('verse

Yes

Features and Comments

Dual chaunel controllers are available; backward
reading is standard; compatible with mM 2400
Series Tape Units

Yes

Yes

?

'tWith optional equipment.
AUE RBACH Computer Charactenstlcs DIgest

11/69

A

(Contd. )

.,

AUERBACH

AUXILIARY STORAGE AND MAGNETIC TAPE

11:520.119

Siemens System 300 (West Germany)

System Identity

2013

2014

2015

2027

2051

Model Number

Drum

Drum

Drum

Core

Disc

Type of storage

4/trunk

4/trunk

4/trunk

l/trunk

l/trunk

l/channel

l/channel

l/channel

l/channel

l/channel

l/unit

l/unit

l/unit

l/unit

l/unit

65,536

131,072

262,144

16,384

1,792,000

65,536

131,072

262,144

16,384

1,792,000

Maximum

1,168,000

Decimal Digits

1,168,000

Characters

262,144

524,288

1,048,576

65,536

262,144

524,288

1,048,576

62

62

62

0

0

0

32

32

32

64

64

64

-

147.5

72,000

72,000

72,000

2,668,000

208,000

64

64

64

256

1- 4, 096

1 - 4, 096

1 - 4, 096

1 - 35,840

Cyclic check code

Cyclic check code Cyclic check code

-

-

-

65,536

-

25

0
87.5

Parity

Cyclic check

-

-

Units On-Line
Maximum
Number

Read/Write
Operations
Seek Operations

Number of
Words per
Unit

Minimum

Maximum
Total
Storage

Rotational Time, msec

AUXILIARY
STORAGE

Minimum

Waiting
Time,
msee

Average (Random)
Maximum

Effective Transfer Rate. char/sec

Sector Size, char
Transfer Load Size, char
Checking

Features and Comments

Model Number
On-Line
Maximum
Number
of Units

Reading/Writing
Searching
Rewinding

Demands on
Processor,

Reading/Writing
Starting/Stopping

%

Peak
1, OOO-char blocks

Transfer
Rate, kilochar/sec

100-char blocks

MAGNETIC
TAPE

Tape Speed, inches/sec
Data Tracks
Data Rows per Block

Data Rows per Inch
mM 729 Compatible
mM 2400 Compatible
Reading
Writing

I

Checking

Read Reverse

Features and Comments

.With optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

11/69

A

11:530.101

AUERBACH
COMPUTER

AUERBACH
~

NOTEBOOK
INTERNATIONAL

COMPARISON CHARTS - NON-U.S.A. COMPUTERS
PUNCHED CARD AND PUNCHED TAPE INPUT-OUTPUT

COMPARISON CHARTS - NON-U.S.A. COMPUTERS
PUNCHED CARD AND PUNCHED TAPE INPUT-OUTPUT

An introduction to the Punched Card and Punched Tape Input-Output Section of
the Comparison Charts, giving the precise meaning of each entry, will be found
on Page 11:230.101.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

11/69

11:530.102

COMPARISON CHARTS - NON-U. S. A. COMPUTERS

A/S Regnecentralen RC 4000 (Denmark)

Bull Gamma -10
(France)

Bull-GE GE-55
(France)

Model Number

RC 405

300

L617

Maximum Number On-Line

l/controller (maximum 64 controllers)

1

1

Peak Speed, cards/min

1200 for 80 columns, 1500 for 51 columns

300

150

?

?

System Identity

Demands on Processor, %
PUNCHED
CARD
INPUT

Code Translation

Yes

Automatic

Automatic

Checking

Dual read comparison

?

Character validity

Features and Comments

Reads cards
column by column

Model Number

PUNCHED
CARD
OUTPUT

300

P540

Maximum Number On-Line

1

1

Peak Speed, cards/min

300

60

Demands on Processor, %

?

?

Code Translation

Yes

Automatic

Checking

?

Hole check

Punches cards
column by column,
can interpret

Features and Comments

Model Number
Maximum

PUNCHED
TAPE
INPUT

PUNCHED
TAPE
OUTPUT

Num~cOn-Line

RC 2000

300

l/controller (maximumum 64 controllers)

1

Number of Channels

5, 6, 7, or 8

5, 7, or 8

Peak Speed, char/sec

2000

300

Demands on Prooessor, %

?

Code Translation

Programmed

Checking

?

Features and Comments

Can handle 6channel
Olivetti tape

PTR055

PTP 55

Model Number

RC 150

Maximum Number On-Line

l/controller (maximum 64 controllers)

Number of Channels

5, 6, 7, or 8

5, 7, or 8

Peak Speed, char/sec

150

105

Demands on Processor, %
Code Translation

Automatic

Checking

Features and Comments

'With optional equipment.
AUE RBACH Computer Characterostics Digest

11/69

fA..

AUERBACH

'"

(Contd.)

PUNCHED CARD AND PUNCHED TAPE INPUT/OUTPUT

11:530.103

System Identity

Elblt 100 (Israel)

Model Number
Maximum Number On-Line
Peak Speed, cards/min
Demands on Processor, %
Code Translation
Checking

PUNCHED
CARD
INPUT

Features and Comments

Model Number
Maximum Number On- Line
Peak Speed, cards/min
Demands on Processor, %
Code Translation

PUNCHED
CARD
OUTPUT

Checking

Features and Comments

ASR 33/620

Model Number

Dlgitronics 2500

256

256

5, 6, 7, or 8

5, 6, 7, or 8

10

Maximum Number On-Line
Number of Channels
Peak Speed, char/sec
Demands on Proces Bor, %

Automatic

Automatic

Parity

Parity

Code Translation

PUNCHED
TAPE
INPUT

Checking

Features and Comments

Tally P120

ASR 33/620

256

256

5, 6, 7, or 8

5, 6, 7, or 8

120

10

Model Number
Maximum Number On-Line
Number of Channels
Peak Speed, char/sec

Demands on Processor, %
Automatic

Automatic

Panty

Panty

Code Translation

PUNCHED
TAPE
OUTPUT

Checking

Features and Comments

• WI th optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

11/69

11:530.104

COMPARISON CHARTS -

System Identity

PUNCHED
CARD
INPUT

NON-U.S.A. COMPUTERS

Fujitsu F ACOM 230 Series and 270 Series (Japan)

Model Number

F664A

Maximum Number On-Line

8/channel + 2 direct channels

Peak Speed, cards/min

800

Demands on Processor, %

0.5

Code Translation

Yes

Checking

Dual read

Features and Comments

PUNCHED
CARD
OUTPUT

Model Number

F683A

Maximum Number On-Line

8/channel + 2 Direct Channels

Peak Speed, cards/min

250

Demands on Processor, %

0.2

Code Translation

Yes

Checking

Read after punch

Features and Comments

Model Number

F749A

Maximum Number On-Line

8/ channel + 2 Direct Channels

F749E

F750A

F748A

Number of Channels
Peak Speed, char/sec
PUNCHED
TAPE
INPUT

200/400

600/1,200

240

1,000

Code Translation

None

None

None

None

Checking

Dual read

Dual read

Dual read

Dual read

Features and Comments

Console
type

Console
type

Model Number

F766A

Maximum Number On-Line

8/channel + 2 Direct Channels

Demands on Processor, %

Free-

Free-

standing
type

standing
type

F767A

Number of Channels
Peak Speed, char/sec
PUNCHED
TAPE
OUTPUT

200

100

Code Translation

None

None

Checking

Feed check'

Feed check

Demands on Processor I %

Features and Comments

*With optional equipment.
AUERBACH Computer Characteristics Digest

11/69

A

(Contd.)

AUERBACH

'"

PUNCHED CARD AND PUNCHED TAPE INPUT/OUTPUT

11:530.105

HItachi HITAC 8000 Series (Japan)
H-8239-11, -21

H-8233

System Identity
11-8238

Model Number

l/trunk

l/trunk

l/trunk

Maximum Number On- Line

400

750

1,470

Peak Speed, cards/min
Demands on Processor, o/r

Vanes

Varies

Vanes

Automatic

Automatic

AutomatIc

Hole count

ValIdity;
echo check

ValIdIty;
echo check

Mark reading
feature optIOnal

Mark reading
feature optional

H-8234

H-8235

Model Number
Maximum Number On-Line

H-8239-11, -31
1 per trunk

l/trunk

l/trunk

91 (160 col/sec)

100

250

Code Translation

Checking

PUNCHED
CARn
INPUT

Features and Comments

Peak Speed, cards/mm
Demands on Processor, %

Varies

Varies

Varles

Automatic

AutomatlC

AutomatlC

Code Translation

Echo check

Hole count

Hole count

Checking

PUNCHED
CARD
OUTPUT

Features and Comments

H-8226-1

H-8229-22

H-8221

H-8222

Model Number

Maximum Number On-Line

l/trunk

l/trunk

l/trunk

l/trunk

5, 6, 7, or 8

5, 6, 7, or 8

5, 6, 7, or 8

5, 6, 7, or 8

500

200

200

1, 000

Varies

Varies

Varies

Vanes

Automatic

Automatic

AutomatlC

AutomatIc

ParIty

Parity

Parity

Parity

CombinatIon
reader and
punch

Combination
reader and
punch

For H-8210 Processor only

Number of Channels
Peak Speed, char/sec
Demands on Proces sor, %
Code Translation

PUNCHED
TAPE
INPUT

Checking

Features and Comments

H-8227-1

H-8229-22

H-8221

H-8222

Model Number

l/trunk

l/trunk

l/trunk

l/trunk

Maximum Number On-Line

5, 6, 7, or 8

5, 6, 7, or 8

5, 6, 7, or 8

5, 6, 7, or 8

110

100

100

100

Varies

Varies

Varies

VarIes

Automatic

Automatic

Automatic

Automatic

None

None

Echo check

Echo check

CombinatIOn
reader and
punch

Combination
reader and
punch

Number of Channels
Peak Speed, char/sec
Demands on Processor, %

Code Translation

PUNCHED
TAPE
OUTPUT

C,hecking

Featur(':.; ,lud Comments

- ..
'With optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc,

11/69

11:530.106

COMPARISON CHARTS - NON-U. S. A. COMPUTERS

HitachI IUTAC 3010 (Japan)

System Identity

PUNCHED
CARll

Model Number

H-323

H-329B

Maximum Number On-Line

Nippon Electric 2200/50
(Japan)
E214

2

2

8

Peak Speed, cards/min

600

1,470

400

Demands on Processor, %

6.72

VaTies

0.1 max

Code Translation

Automatic

Automatic

Automatic

Checking

Hole count;

character

Character validity

validity

Echo check;
character
valIdity

Model Number

H-334

H-336

E214

Maximum Number On-Line

INPU'1'

,
Features and Comments

PUNCHED
CARD
OUTPUT

1

1

l/address assignment

Peak Speed, cards/min

100

200

100 to 400

Demands on Processor, %

Varies

Varies

0.1 max

Code Translation

Automatjc

Automatic

Automatic

Checking

Hole count

Hole count

Punch die activation

Model Number

H-322

H-176

E209

Maximum Number On-Line

1

1

l/address assignment

Number of Channels

0,

8

5, 6, 7, or 8
300

Features and Comments

PUNCHED
TAPE
INPUT

6, 7, or 8

Peak Speed, char/sec

1,000

200

Demands on Processor, %

Varies

Varies

0.1 max

Code Translation

Matched codes

Matched codes

Programmed

Checking

Parity

Parity

Parity; dual-read compare

Features and Comments

PUNCHED
TAPE
OUTPUT

Model Number

H-331

E209

Maximum Number On-Line

1

1/ address assignment

Number of Channels

5, 6, or 7

5, 6, 7, or 8

Peak Speed, char/sec

100

110

Demands on Processor, %

Varies

0.1 max

Code Translation

Matched codes

Programmed

Checking

None

None

Features and Comments

*Wiih optional equipment.

AUE RBACH Computer CharacteristIcs Digest

11/69

A

(Contd.)

AUERBACH
®

11:530.107

PUNCHED CARD AND PUNCHED TAPE INPUT-OUTPUT

System Identity

NEAC Series 2200/100, 200, 300, 400, 500 (Japan)
N214-2

N223

N223-2

1/address assIgnment
800

1/system
400

Model Number

N123**

1,050

400

Varies

0.2 max

Automatic

Maximum Number On-Line
Peak Speed, cards/min
Demands on Processor, %
Code Translation

Validity cycle

Checking

Direct
transcription
feature

Reader/punch,
direct
transcription
feature

Direct
transcription
feature

N224A-1

N214-1

N214-2

PUNCHED
CARD
INPUT

Features and Comments

Model Number

N224A-2

Maximum Number On-Line

l/address assignment

Peak Speed, cards/min
Demands on Processor, %

Varies
Automatic

Code Translation

PUNCHED
CARD
OUTPUT

Checking

Direct
tr anscription

Reader/punch;
Direct transcription feature

Direct
transcript ion
feature

N109A-1**

N209A-1

N209A-2

l/system

1/address assIgnment

Features and Comments

Model Number
Maximum Number On-Line

5, 6, 7, and 8

Number of Channels

300

300

0.2 max

Varies

Peak Speed, char/sec

1000

Demands on Proces sor, %
Code Translation

Programmed
Parity; dual-read
compare

Checking

Special code
detecting

Nll0A-1**

Special code detecting, ISO code prOC('SHmg
feature

NllOA-3**

l/system

N210A-1

60

Programmed

Feature s and Comments

N210A-3

l/address assIgnment

Model Number

Maximum Number On-Line
Number of Channels

5, 6, 7, and 8

0.1 max

PUNCHED
TAPE
INPUT

110

60
Varies

110

Peak Speed, char/sec

Demands on Processor, %
Code Translation

None

PUNCHED
TAPE
OUTPUT

Checking

Fea:ttlr('f> and Comments

"With optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc

11/69

COMPARISON CHARTS - NON-U.S.A. COMPUTERS

11:530.108

System Identity
Model Number

PUNCHED
CARD
INPUT

PUNCHED
CARD
OUTPUT

Philips PlOOO (Netherlands)
PlOlO

PlOll

I

PI012

4512
l/trlDlk

l/trunk

I

1,500/2000

800

1435

4514

Maximum Number On-Line

1/controller

Peak Speed, cards/min

400/500

Demands on Proces sor, %

Less than 1

Varies

Varies

Code Translation

Automatic

Automatic

Automatic

Checking

Character validity

Validity

Validity

Features and Comments

Higher speed is for 51-column cards; lower for
80-column cards

By column
on demand

By column
on demand

Model Number

PlO15

4521

Maximum Number On-Line

l/controller

800/1000

PlO16

l/trlDlk

Peak Speed, cards/min

100

Demands on Processor, %

Less than 1

Varies

Code Translation

Automatic

Automatic

Checking

Character validity

Read after punch

Features and Comments

Higher speed for Pl016 is for
51-column cards

By rows

300/400

Model Number

PUNCHED
TAPE
INPUT

TCL System 4 (United Kingdom)

100

PI020

4580, 4581

Maximum Number On-Line

l/controller

4/trunk

Number of Channels

5, 6, 7, and 8

5, 7 and 8

Peak Speed, char/sec

1,000

1500

Demands on Processor, %

Less than 1

Varies

Code Translation

None

Automatic

Checking

None

Parity, validity

Model Number

PI025

4585

Maximum Number On-Line

1/controller

l/trlDlk

Number of Channels

5, 6, 7, and 8

5, 7, and 8

Peak Speed, char/sec

150

150

Demands on Processor, %

Less than 1

Varies

Code Translation

None

Automatic

Checking

Echo

Parity, valIdity

Features and Comments

PUNCHED
TAPE
OUTPUT

Features and Comments

'With optional eqUipment.
AUE RBACH Computer Characteristics Digest

11/69

A

(Contd.)

AUERBACH

'"

11:530.109

PUNCHED CARD AND PUNCHED TAPE INPUT/OUTPUT

System Identity

ICL 1900 Series (United Kingdom)
2101

I

2103

2104

Varies
1600

I

600

600

Varies according to processor model

2105

2106

1

1

300

600

0.6

1.2

Model Number

Maximum Number On-Line
Peak Speed, cards/min

Demands on Processor, %
Code Translation

Automatic
Proper photocell functioning and correct registration

Optical
binary
image
feature

60-column
cards

1920

Checking

1901A only;
optional binary
image
feature

1901A
only

1922

Features and Comments

Model Number

2151

Maximum Number On-Line

Varies
100

33

Peak Speed, cards/min

300

Demands on Processor, %

Varies accordIng to processor

Code Translation

Automatic
Hole count

Echo check

Hole count

Checking

Row punch

Column punch

Row punch

Features and Comments

1915

I

1916

2601

I

2602

Maximum Number On-Line
Number of Channels

5, 6, 7, and 8

I

1000

250

I

Peak Speed, char / sec

1000

Demands on Proees sor, %

Varies according to processor model

Code Translation

Format board

PUNCHED
TAPE
INPUT

Checking

Parity

Combined reader and punch using one Ilo
channel; for 1901A, 1902A, 1903A only

1925

PUNCHED
CARD
OUTPUT

Model Number

Varies

300

PUNCHED
CARD
INPUT

2601

Features and Comments

2602

Model Number

Maximum Nwnber On-Line

Varies

Number of Channels

5, 6, 7, and 8

Peak Speed, char/sec

110

Demands on Processor, %

Varies according to processor model

Code Translation

Format board

PUNCHED
TAPE
OUTPUT

Checking

None

Combined reader and punch using one
1901A, 1902A, 1903A only

Ilo channel: for

Features and Comments

----'With op\!onaJ cquipment.

© 1969 AUERBACH Corporatton and AUERBACH Info, Inc.

11/69

11:530.110

COMPARISON CHARTS - NON-U. S. A. COMPUTERS

System Identity

PUNCHED
CARD
lNPUT

PUNCHED
CARD
OUTPUT

Siemens System 4004 (West Germany)

Model Number

4004/237

Maximum Number On-Line

l/trunk

l/trunk

Peak Speed, cards/min

1435

666

Demands on Processor, %

Varies

Varies

Code Translation

Automatic'

Automatic *

Checking

Circuit checks; validity check in translate mode

Features and Comments

Optional mark
reading feature

90-column verified
cards feature

Model Number

4004/234

4004/236

4004/4235

Maximum Number On-Line

l/trunk

l/trunk

Peak Speed, cards/min

100

300

Demands on Processor, %

Varies

Varies

Code Translation

Automatic

Automatic

Checking

Hole count

Hole count

Features and Comments

Single stacker

Two stackers;
Read/punch option

Model Number

4004/4226

4004/4227

Maximum Number On-Line

l/trunk

l/trunk

Peak Speed, ohar/.ec

400

1,000

Demands on Processor I %

Varies

Varies

Code Translation

Automatic *

Atutomatic'

Checking

Parity

Panty

Number of Channels

PUNCHED
TAPE
lNPUT

Features and Comments

Model Number

4004/4225

Maximum Number On-Line

l/trunk

Number of Channels

PUNCHED
TAPE
OUTPUT

Peak Speed, char/sec

100

Demands on Processor I %

Varies

Code Translation

Automatic *

Checking

Parity

Features and Comments

iWlth optional eqUipment.
AUE RBACH Computer Characteristics Digest

11/69

fA

AUERBACH


(Contd.)

PUNCHED CARD AND PUNCHED TAPE INPUT/OUTPUT

11:530.111

System Identity

Siemens System 300 (West Germany)
2009

2010

l/channel

l/channel

33

600

Varies

Varies

Automatic

Automatic

Model Number
Maximum Number On-Line

Peak Speed, cards/min
Demands on Processor, %
Code Translation

Validity check in translate mode

Combined
reader/punch

Column
binary feature

Checking

PUNCHED
CARD
INPUT

Features and Comments

2021

Model Number

l/channel

Maximum Number On-Line
Peak Speed, cards/min

245

Demands on Processor, %

Varies

Code Translation

Automatic
Echo

PUNCHED
CARD
OUTPUT

Checking

Column binary feature

Features and Comments

0001 Console Tape Input
only for Mod. 303

0016 Console Tape Input
for Mod. : 302, 304, 305, 306

2006

2008

l/channel

l/channel

l/channel

l/channel

5 and 6 channels

5 and 6 channels

~j,.~n:". 8

~j,.~n:". 8

30

200

400

400 input -max 150

Varies

Varies

Varies

Varies

Automatic

Automatic

Automatic

Automatic

Second read station

Second read station

Parity

Parity

Combined
input/output

Model Number
Maximum Number On-Line
Number of Channels
Peak Speed, char/sec
Demands on Processor, %
Code Translation
Checking

Features and Comments

2007
l/channel
5, 6, 7, 8 channels
150 max
Varies
Automatic

PUNCHED
TAPE
INPUT

Model Number
Maximum Number On-Line
Number of Channels
Peak Speed, char/sec
Demands on Processor, %
Code Translation

Parity

PUNCHED
TAPE
OUTPUT

Checking

Features and Comments

'With optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

11/69

A

AUERBACH
~

11:540.101

AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

COMPARISON CHARTS - NON-U.S.A. COMPUTERS
PRINTERS AND SPECIALIZED INPUT/OUTPUT DEVICES

COMPARISON CHARTS - NON-U. S. A. COMPUTERS
PRINTERS AND SPECIALIZED INPUT/OUTPUT EQUIPMENT

An introduction to the Printers and Specialized Input/Output Equipment Section
of the Comparison Charts, giving the precise meaning of each entry, can be
found on Page 11:240.101.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

11/69

11:540.102

COMPARISON CHARTS - NON-U. S. A. COMPUTERS
\

System Identity

A/S Regnecentralen RC 4000 (Denmark)

-

Model Number

RC 610

Maximum Number On-Line

l/controller (maximum 64
controllers)

Single Spacing

1000

I-inch Spacing

1350

Bull Gamma 10 (France)

1
667

300

Speed,
lines/min

Demands on Processor I %
PRINTED
OUTPUT

Number of Print Positions

132

Character Set Size

64

1~0

96

Checking

Feature s and Comments

MICR
READER

or 144

Yes

One line buffer

Dual feed carriage

Model Number

-

Peak Speed, documents/min

600

Features and Comments

CMC7 characters

Model Number
OPTICAL
CHARACTER
READER

Peak Speed, documents/min

Features and Comments
Model Number
DATA
COMMUNICATIONS
CONTROLLER

Peak Speed, bits/sec
Features and Comments

Model Number
CRT
DISPLAY

RC 4195

Capacity, char
Features and Comments

PWTTER

OTHER
INPUTOUTPUT
DEVICES

Model Number

RC 4193 Controller

Peak Speed, pOints/sec

6400 (160 mm/sec)

200, 300, 350, or 450

Features and Comments

Resolution: 0.025 mm

Incremental plotter

Model Number

RC 450

Name

Strip Printer

Features and Comments
*Wlth optional equipment.

AUERBACH Computer Characteristics Digest

11/69

A

(Contd.)

AUERBACH
®

PRINTERS AND SPECIALIZED INPUT/OUTPUT DEVICES

11:540.103

Elblt 100
(Israel)

Bull-GE GE-55 (France)

PRT 051

I

PRT 052

I

PHT 055

PRT 056

PHT 050

I

Shepard 880

PRT 057

256

1

Typically 83

TypICally 140

System Identity

600

50 char/sec

Model Number
Maximum Number On-Line
Single Spacing
Speed
hnes/min

I

I-inch Spacing

Demands on Processor. %
96

I

128

I

96

l

PRINTED
OUTPUT
128

128

80

Number of Print Positions

64

Character Set Size

Parity

Speed is given for 48-character set;
lower speeds for 64-character set

Serial
printers

Checking

Featw'cN and Comments

Model Number

Peak Speed, documents/min

MICR
READER

Features and Comments
Model Number
Peak Speed, documents/min

OPTICAL
CHARACTER
READER

Features and Comments
Model Number

Peak Speed, bits/sec

DATA
COMMUNICATIONS
CONTROLLER

Features and Comments
Model Number

Capacity, char

CRT
DISPLAY

Features and Comments
Model Number
Peak Speed, pOints/sec

PLOTTER

Features and Comments

Model Number
Name

OTHER
INPUTOUTPUT
DEVICES

Features and Comments
*With optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Illfo, Inc

11/69

COMPARISON CHARTS -

11 :540.104

NON-U. S. A. COMPUTERS

Fujltsu FACOM 230 Series and FACOM 270 Series (Japan)

System Identity
Model Number

F642A

F643A

F643C

Maximum Number On-Line

8/chalUlel + 2 direct channels

1

1

Single Spacing

1,500/1, 000/500

480/240

480/240

I-inch Spacing

750/500

320/200

320/200

Speed,
lines/min

Demands on Processor, %

0.8

Number of Print Positions

136

80

136

Character Set Size

64/128

50/100

50/100

Checking

Parity, validity

Parity, timing, paper feed

Features and Comments

With buffer memory

PRINTED
OUTPUT

Unbuffered control

Flag bit control

Model Number
MICR
IlEADER

Peak Speed, documents/min

Features and Comments
Model Number
OPTICAL
CHARACTER
READER

Peak Speed, documents/min

440

360 lines/min

440 lines/min

Features and Comments

Document scanner

Page scanner

Roll paper reader

i\1odel Number
DATA
COMMUNICATIONS
CONTIlOLLER

Peak Speed, bits/sec
FC'atures and Comments
Model Number

CRT
DISPLAY

Capacity, char

Features and Comments

PLOTTER

Model Number

F620lB

Peak Speed, points/sec

400

Features and Comments
Model Number
OTHER
INPUTOUTPUT
DEVICES

Name

Electronic Printer

Features and Comments

la, 000

hnes/min

*Wlth optional equipment.

AUERBACH Computer Characteristics Digest

11/69

fA

AUERBACH
®

(eontd. )

PRINTERS AND SPECIALIZED INPUT/OUTPUT DEVICES

11:540.105

Hitachi HIT AC 8000 Series (Japan)
8244-31

8244-32

8245-11

8245-12

8246-11

System Identity
8246-12

l/trunk

Model Number
Maximum Number On-Line

300

150

600

300

1250

625

Single Spacing

252

148

435

252

769

476

I-inch Spacing

Speed
lines/min

Varies

Demands on Processor, %

132

132

132 or 160.

132 or 160.

132 or 160*

132 or 160.

63

110

63

110

63

110

PRINTED
OUTPUT
Number of Print Positions
Character Set Size

None

Checking

Includes
Kama
characters

Includes
Kama
characters

Includes
Kama
characters

Features and Comments

Model Number
Peak Speed, documentS/min

MICR
READER

Features and Comments
Model Number
Peak Speed, documents/min

OPTICAL
CHARACTER
READER

Features and Comments
Model Number
Peak Speed, bits/sec

DATA
COMMUNICATIONS
CONTROLLER

Features and Comments
Model Number

i

Capacity, char

CRT
DISPLAY

Features and Comments
Model Number

I

Peak Speed, points/sec

PLOTTER

Features and Comments
Model Number

8212-1

Input/Output typewnter
500 chor/min peak speed

Name

OTHER
INPUTOUTPUT
DEVICES

Features and Comments
*WIth optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

11/69

11:540.106

COMPARISON CHARTS - NON-U. S. A. COMPUTERS

Nippon NEAC Series 2200
(Japan)

Hitachi HITAC 3010 (Japan)

System Identity
Model Number

H-333

Maximum Number On-Line

2

H-333C

E206

1/Address assignment

SingIe Spacing

800 to 1000

521 to 600

333

I-inch Spacing

500

405

310

Speed,
Lmes/min

Demands on Processor, %

Varies

Number of Print Positions

120 or 160*

120 or 160*

120 or 132*

Character Set Size

63

96

60

Checking

None

O. 8 maximum

PRINTED
OUTPUT

Cycle check, printer check

Includes Kama

Features and Comments

characters

Model Number
MICR
READER

PC'ak Speed, documents/min
Features and Comments

OPTICAL
CHARACTER
READER

Model Number

H-5820

N240D-1

Peak Speed, documents/min

1500

1100

Features and Comments

Videoscan document reader

407 Font
N284A

N292

64 lines

max.

256 lines
max.

Model Number

N244A-1

N244A-2

Pc,1k Speed, points/sec

300

200

Model Number
DATA
COMMUNICATIONS
CONTROLLER

PNIk Speed, bits/sec

F('atures and Comments
Model Number
CRT
DISPLAY

Capacity, char
"'('atures and Comments

PLOTTER

Features and Comments

1\10<1el Number
OTHER
INPUTOUTPUT
DEVICES

l":unc

F('atu r<.>s and Comments

*Wlth optional equipment.

AUE RBACH Computer Characteristics Digest

11/69

fA

AUERBACH
®

(Contd.)

PRINTERS AND SPECIALIZED INPUT/OUTPUT DEVICES

Philips P1000 Series
(Nether lands)
P1030-001

P1030-002

11:540.107

ICL System 4 (United Kingdom)

P1030-003

4554

4555

4560

1
1/ control unit

System Identity

1

l/trunk

Maximum Number On-Line

360

600

1000

1350

750

270

380

600

779'

667

Less than 1

Model Number

4561

Single Spacing
Speed
lines/min
l-inch Spacing

Vanes

132

160

64

64

Validity

Vahdlty

Demands on Processor, %

I

PRINTED
OUTPUT
132

160

\132

Speeds based on restricted 48-character
set

Number of Print Positions
Character Set Size
Checking

Features and Comments

Model Number
Peak Speed, documents/min

MICR
READER

Features and Comments
Model Number

2

150

Peak Speed, documents/min

OPTICAL
CHARACTER
READER

I

Features and Comments
Model Number

P1080

Peak Speed, bits/sec

640,000

DATA
COMMUNICATIONS
CONTROLLER

Features and Comments

For up to 16 teletypewrIters

Model Number

Capacity, char

CRT
DISPLAY

Features and Comments
Model Number

P1035

Peak Speed, pOints/sec

300

PLOTTER

Features and Comments
Model Number

2

Name

Lector
9000 documents/hour

OTHER
INPUTOUTPUT
DEVICES

Features and Comment..<:;

*Wit.lt optional equiprnento

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

11/69

11:540.108

COMPARISON CHARTS -

System Identity

NON-U.S.A. COMPUTERS

ICL 1900 Series

Model Number

1933

Maximum Number On-Line

Varies

2401

2402

2404

2405

1

1

,
Single Spacing

1100 or 1350

300

600

300

600

I-inch Spacing

630

231

333

163

326
16.2

Speed,
lines/min

Demands on Processor, %

VarIes according to processor model

8.1

Number of Print Positions

96, 120
or 160

96 or 120

96 or 120

Character Set Size

64

64

64

Checking

Print synchronizatlOn and hammer count

Features and Comments

Automatic
wrIte
feature

Model Number

8500

Peak Speed, documents/min

1200

Features and Comments

Endorser; zeroi..ll; 18 stackers; 6-pocket pull-out; usable off line

Model Number

8101/8201/8301

PRINTED
OUTPUT

MICR
READER

OPTICAL
CHARACTER
READER

DATA
COMMUNICATIONS
CONTROLLER

CRT
DISPLAY

PLOTTER

OTHER
INPUTOUTPUT
DEVICES

Peak Speed, documents/min

1901A only

600

Features and Comments

Control is 8101; character reader is 8201; mark reader is 8301

Model Number

7007/2
Multiplexor

7900
System

7070/1 Single
Channel

7070/2 Data
Terminals

7070/3

7010/3

Peak Speed, bits/sec

50/line to
2400/line

50/line to
4800/line

110

110

1200

2400

Feature. and Comments

Up to 63
lines

Up to 252
lines

5-bit
code

8-bIt
code

For
7152 CRT

For telephone

Model Number

7152

Capacity, char

520 or 1040

Features and Comments

Local or remote

Model Number

1934

Peak Speed, points/ sec

200 or 300

Features and Comments

1004

Model Number

1004 link

Name

Link to UNIVAC 1004 plugboard computer

Features and Comments
*With optional equipment.

AUE RBACH Computer Characteristics Digest

11/69

fA

.,

AUERBACH

(Contd.)

PRINTERS AND SPECIALIZED INPUT/OUTPUT DEVICES

11:540.109

SlOmens System 4004 (West Germany)
4247

243

l/trunk

l/trunk

750

1,250

System Identity
Model Number
Maximum Number On-Line
Single "pacing
Speed
hnes/min

535

715

Varies

Varles

132

132 or 160

64

64

Timing

Timing

1-inch Spacing
Demands on Processor, %

PRINTED
OUTPUT

Number of Print Positions
Character Set Size
Checking

Features and Comments

Model Number
Pesk Speed, documents/min

MICR
READER

Features and Comments
4250

4251

4252

4253

1,600

1,600

750

750

Model Number
Pesk Speed, documents/min

OPTICAL
CHARACTER
READER

Features and Comments
653

66S

4666

300 char/sec

6,000 bytes/sec

8,000 char/sec

Model Number
Pesk Speed, bits/sec

DATA
COMMUNICATIONS
CONTROLLER

Features and Comments
Model Number
Capacity, char

CRT
DISPLAY

Features and Comments
Model Number
Pesk Speed, pOints/sec

PWTTER

Features and Comments
Model Number

752
Video Data Terminal

Name

OTHER
INPUTOUTPUT
DEVICES

Features and Comments
*With optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc

11/69

COMPARISON CHARTS - NON-U.S.A. COMPUTERS

11:540.110

System Identity

Siemens System 300 (West Germany)

Model Number

2022

2023

Maximum Number On-Line

l/cbannel

l/cbanneJ

Single Spacing

750

600

I-inch Spacing

535

378

Demands on Processor, %

Varies

Varies

Number of Print Positions

120

104

Cbaracter Set Size

48

48

Cbecking

Echo/timing

Tinung

Speed,
lines/min

PRINTED
OUTPUT

Features and Comments

Extension print drum of 104 to 120 columns
Extension print drum of 120 to 136 columns

Model Number
MICR
READER

Peak Speed, documents/min

Features and Comments
Model Number
OPTICAL
CHARACTER
READER

Peak Speed, documents/min

Features and Comments
Model Number
DATA
COMMUNICATIONS
CONTROLLER

Peak Speed, bits/ sec

Features and Comments
Model Number
CRT
DISPLAY

Capacity, char

Features and Comments
Model Number
PLOTTER

Peak Speed, points/sec

Features and Comments

OTHER
INPUTOUTPUT
DEVICES

Model Number

Console TypewrIter 2017

Typewriter T 100

Name

69 and 104 char/line

10 char/sec 200 bits/sec

Features and Comments
*With optional equipment.

AUERBACH Computer Characteristics Digest

11/69

A

(Contd.)

AUERBACH
@

SPECIAL REPORTS

AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

AUERBACH
(!)
Print~d

in

IL~_A

--.

2J:001. 001

~

STANDARD

~EDP

AUERBAC~

SPECIAL REPORTS
CONTENTS

REPORTS

®

SPECIAL REPORTS
CONTENTS
Computer Rental Contracts and Proposals - A Survey and Analysis .

· 23:010.001

A Survey of the Character Recognition Field ..

· 23 :020. 001

Decision Tables Symposium. . . . . . . . . . . . .

. 23 :030.001

Magnetic Tape Recording: A State-of-the-Art Report . . . . . . • . . . . . . • . . . . . . . . . . . 23:040.001
High-Speed Printers: A State-of-the-Art Report. . . .

. .23:050.001

Random Access Storage Devices: A State-of-the-Art Report. .

. . 23:060.001

Digital Plotters: A State-of-the-Art Report. . . . . . . . . . . . . . .

. 23:070.001

Data Collection Systems: A State-of-the-Art Report . . . . . . . . .

· 23:080.001

The Selection and Use of a Data Processing Service Center . . . . . .

· 23:090.001

Data Commnnications - What It's All About . . . . . . .

· 23:100.001

Source Data Automation Techniques and Equipment • . . . • . • • . . . . . • . • . . . • • . . . • • . 23:110.001
Design and Applications of Automated Display Systems . . . . • . . . . . • . . . • . • . . . . . . . 23:120.001
Keyboard to Magnetic Tape Encoders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23:130.001

,

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

5/69

23010 ';01

........

~

,."I.~EDP
Aul"At"

•

SPECIAL.. REPORT
COMPUTER CONTRACT'-,

II'" •

AUERBACH SPECIAL REPORT:
COMPUTER RENTAL CONTRACTS AND PROPOSALSA SURVEY AND ANALYSIS

PREPARED BY
THE TECHNICAL.. STAFF OF
AUERBACH CORPORATION

C 1967 AUERBACH Corporation and AUERBACH Info. Inc

-1. "'.....

23:010.002

~EDP

SPECIAL REPORT
COMPUTER CONTRACTS

AIII~

CONTENTS
.1

INTRODUCTION

.2

CONTRACT FACTORS AND CONSIDERATIONS

.21
.211
.212
.213
.214
.215
.216

Specifications
Equipment
Software
Method of payment
Amount of chargeable time
Chargeable time
Assurance of serviceable time

.22
.221
.222
.223

Acceptance
Shipping and installation charges
Delivery and acceptance dates
Acceptance tests

.23

Environment

.24
.241
.242

Maintenance
Reliability
Maintenance responsibility

.25

User's Rights

.26
.261
.262
.263
.264
.265
.266
.267
.268

Additional Factors
Special equipment
System design
Training
Program testing
Special programs
Conversion credits
Investment tax credit
Pu rchase -leaseback

.27

Summary of Considerations

.3

ORGANIZATION OF THE SURVEY

.4

CONCLUSIONS

.41
.42
.43
.44
.5

6/67

Omission of User Safeguards
Variations Among the Manufacturers
Possibilities for Negotiations
Inadequacies of the Standard Contracts
SURVEY COMPARISON CHARTS

A•

AUERBACH

""111

23:010. 100

...A-

Ill.....

EDP

SPECIAL REPORT
COMPUTER CONTRACTS

IIHI"

COMPUTER RENTAL CONTRACTS AND PROPOSALSA SURVEY AND ANALYSIS
.1

INTRODUCTION
The acquisition of an electronic data processing system is a major expenditure for any
company. Therefore, the prospective user should carefully weigh all of the factors which
could have either a direct or indirect influence on determinIng- which system meets his requirements at the lowest overall cost. The difficulty faced by the user in accurately assessing
the merits of various systems offered by competitive manufacturers is compounded by many
intangible factors, such as equipment reliability, availability, competence of the manufacturer's support personnel , software performance, and programmIng difficulty. Even
the true cost of the computer hardwarc itself can be hard to pin down hecause of the effects
of varying extra-use charges, downtime credits, discounts, and purchase options.
An accurate analysis of relative equipment costs involves prOjections of the monthly use of
each system throughout the contract period, plus a study of the Implications of all the
clauses in each contract. The evaluation of such a study is difficult because most of the
standard computer rental contracts fail to cover certain major cost factors. Often the
contracts do not define potentially important points such as whether or not set-up time is
to be included in chargeable machine usage time. Extra-usage charges are often established by individual branch managers rather than by specific terms in the standard contracts. Equipment rentals during the first decade of the computer era have often been
handled in a surprisingly informal way, perhaps because the essential question often was
"Will it ever work?" rather than "Will it always work?"
Now that computers are a vital cog in most business organizations, rental contracts are
more Important than ever. A well-prepared contract should show what costs the user will
need to bear during the installation period and exactly how much help he can count on from
the manufacturer. It should show how much computer time IS allowed under the basic
rental charge, how operational time is to be computed, the cost of extra time, and the discount, if any, that is applicable when the equipment is not fully utilized. It should show
what the user can expect when a breakdown occurs: how soon the service engineer should
arrive and what credits are allowed for time lost due to the breakdown. A well-prepared
computer rental contract should cover all these and numerous other points that may involve major expenditures by one of the contracting parties.
Unfortunately, although contract terms are becoming increasingly important, objective
comparisons between the terms offered by different manufacturers are still very difficult
to make. Differences in termill.llogy and frequent omissions of Important factors from
the standard contracts continuc 10 make it hard for the prospectlve user to evaluate all of
the alternatives.
The objectives of this Special Report are:
(1)

To identify the major considerations in contracting for the rental of a
data processing system; and

(2) To present a clearcut analysis and comparison of the terms and provisions
of the standard commercial and government rental contracts currently being
offered by nine major U. S. manufacturers.
A knowledge of the terms that all nine manufacturers are prepared to offer can clearly
strengthen the prospective user's bargaining position when negotiating with anyone manufacturer .
.2

CONTRACT FACTORS AND CONSIDERATIONS
In analyzing the costs of competing equipment, it is desirable to use an indentical specification base and to examine all manufacturers' equipment suitable for the application against
this common base. The U. S. Government, by employing such a technique through its
yearly "Invitations to Bid, " is able to elicit responses which are clear in detail and highly
conducive to comparative analysis. In contrast, the commercial lease agreements offered
by the major computer manufacturers tend to have a somewhat nebulous quality in that
they neglect to specify certain contract details that are included in government contracts
as a matter of course. Since most manufacturers are willing to negotiate on specific
terms, it is desirable for the user to obtain statements covering all of the pertinent points
discussed in this Special Report as a means of defining the cxact equipment and services
to be provided by the manufacturer.

C 1967 AUERBACH Corporation and AUERBACH Info, Inc.

6/67

AUERBACH STANDARD EDP REPORTS

23:010. ZOO
.2

CONTRACT FACTORS AND CONSIDERATIONS (Contd.)

It .howd be noted that this report does not attempt to cover all of the numerous legal conliderattOilI involved in a contract of the complexity required for a computer system.
Lep! cOUDsel should always be obtained for thorough analysis of a specific cootract. The
balance of thil section. however. Indicates the nature of the contract and proposal terms
which the user should attempt to have clarified •
. 21

Specifications

. 211 Equipment
The manufacturer should provide detailed specifications .:.f the equipment units at contract InUlation. These detailed specifications will permit the user to begin effective preparation for the arrival of the system. If specifications are not complete, programmers
may be unable to complete effective, detailed coding. (This Is less critical, however, If
a firm, process-oriented language Is available.) The hardware unit specifications should
be carefully determined to insure obtaining an expected level of performance .
. 212 Software
In addition to the specification of the equipment configuration, it is reasonable to expect
a complete delineation of the specific program "packages" which are to be made available.
Perhaps the one area that will be most difficult to specify is the software to be provided
by the manufacturer. Software includes program translators (compilers, assemblers,
generators, etc.) as well as utility routines. The ideal objective would be to have all
languages and routines fully documented, completely free of errors, easy to learn and easy
to use. At the present state-of-the-art, provision should be made for additional manufacturer's assistance in utilization and Implementation of these techniques.
The user should determine whether process oriented (compiler) languages will be useful.
Experienced programmers often prefer assemblers. The user should determine that the
translators for the languages to be used are readily available and fully tested. It may be
found that the sort routines, report program generators, debugging routines, etc., will
not fit within the conceptual ideas of the user's intended operational practices. The
manufacturer may, therefore, be asked to modify them as necessary.
The user should assure himself that all of the software he Is obtaining will operate on the
equipment configuration he is to receive .
. 213 Method of payment
The method of payment should be specified in the contract. Apart from outright purchase
and normal rental contracts, it is also possible to obtain a rental contract which includes
a purchase option (usually exercisable within a fixed time period). With a purchase option,
a major portion of the rental charges can be applied to subsequent purchase. The use of
a computer leasing company as a second or third party should also be investigated. See
paragraph.268.
Before a decision is made relative to the type of payment, the user should determine
whethl!r his expected amortization schedule is acceptable to th!' Internal Revenue Service
so that some evaluation can he made of the various alternatives in d'c light of corporate
profits. In some cases, the manufacturer will pass on to til!' customer the 7<;( Investment
Tax Credit, as outlined in paragraph. 267.
The term of the contract should be established. Rental contracts are uf'llally renewable
on a year-to-year basis and cancellable (after an initial period) on 30 to ,1\') days' notice.
One year is the minimum acceptable time by manufacturers as an inithl period in conventional contracts The user can in Borne cases obtain reduced rentals or additional
services by agreeing to a minimum term which is longer than one year.
The responsibility for personal property and sales and use taxes should also be specified
in the contract.
.214 Amount of chargeable time
Rental contracts should clearly define the amount of chargeable time included in the basic rental fee. Some of the more common definitions of the amount of chargeable time
are:
(a) Any 176 hours per month.
(b) Any 200 hours per month.
(c) Any 9 hours per day.
(d) The time wring a specific period such as: 9 am to 5 pm, or 8 am to 5 pm (with
the lunch hour available to the user).
(e)

6/67

Unlimited use.

•
j~\

.

AUERBACH

(Contd. )

23:010.214

SPECIAL REPORT

.214 Amount of Chargeable Time (Contd.)
Additional charges beyond the amount included in the basic rental fee 8hould al80 be defined. These charges are u8Ually a stated percentage of the basic hourly rental rate.
10 to 50 percent being common. Charges for time beyond the basic time are u8ually
based directly on the actual time used on each unit or subsystem .
. 215 Chargeable time
The time to be counted as chargeable time is usually defined similarly to operational use
time. This is the time during which the system is productive or could have been productive
if the user operated efficiently. It is not unusual to declare rerun time as nonchargeable,
provided it is caused by equipment malfunction rather than operator error. Such credit is
usually limited to a maximum of 20 minutes per rerun. Most manufacturers exclude set-up
time from the accrual of rental time; however. a clear statement of the manufacturer's concept of "set-up time" should be obtained in writing by the user. Some manufacturers. for example, consider tape rewinding and program loading as operational use time, and charge
accordingly, while others charge only for program running time .
. 216 Assurance of serviceable time
The manufacturer guarantees (at least implicitly) a certain number of serviceable hours
per day (or month). In some cases, when the number of serviceable hours is less than the
guarantee, the user can reduce his rental pro rata; e. g .• if 176 hours per month are
agreed upon and 6 hours of that time are unavailable, the rental fee can be reduced by
6/176. In case of major failures, a backup facility should be provided .
. 22

Acceptance

.221 Shipping and installation charges
Payment for these services should be mutually agreed upon during contract negotiations.
It is customary for shipping charges to be borne by the prospective user; however, the

costs of in-transit insurance, physical installation and final test of the hardware are
absorbed by the manufacturer.
The site preparation for the equipment is the user's responsibility, but should be designed
in accordance with the manufacturer's recommendations in order to insure proper installation and operating conditions. The manufacturer will usually be most cooperative in
supplying physical installation data and advice. Complete environmental details should be
specified by the manufacturer's site-installation engineering staff and should include: air
conditioning, power, equipment layout, cable lengths, floor loads. special power outlets,
and service area layout. Manufacturers sometimes overstate floor space requirements
(systems can be operated in "crowded" conditions if necessary), but otherwise, provide
good assistance in site design (see Paragraph. 23).
In cases where the manufacturer delivers equipment which differs from that specified and
requires site changes, the manufacturer may then be held responsible for such changes .
. 222 Delivery and acceptance dates
Delivery and acceptance dates should be established. The user can normally postpone the
delivery date with as little as 30 days notice without penalty. Should the cquipment be
delivered before program preparations are completed, a considerable amount of money
can be wasted unnecessarily. Therefore, the delivery date should be carefully reviewed
as the Implementation of the system progresses, and postponed if necessary.
Any program packages specified in the contract should be available at their promised date.
Software should be delivered several months earlier than hardware to permit time for
familiarizatIOn and use.
In, some cases, penalties may be agreed upon for failure to meet hardware or software
schedules, if the user sustams a loss attributable to the delay. The time lapse between the
plaCing of an order and the delivery of equipment often runs between 6 and 24 months. Most
manufacturers have done a good job of meeting delivery schedules for production-model
equipment, but have often had problems when the system includes novel or advanced components or new software.
Quite commonly, users experience difficulties in meeting their own system design and programming schedules, leading to a lack of readiness on the proposed installation date. As
a protE'ctive measure, manufacturers are beginning to include contract clauses specifying
damages to be paid by the user in the event his unpreparedness delays the shipping schedule.
The actual delivery date is not as important as the acceptance date. which is the date before
which the acceptance tests should have been passed. In some cases. manufacturers have
agreed to penalty clauses should the acceptance date be delayed. This is not common, but
penalty clauses as high as $1,000.00 per day have been negotiated.

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

6/67

AUERBACH STANDARD EDP REPORTS

23:010.223

. 223 Acceptance tests
Acceptance tests should be specified and should include additional tests of the system
after it has passed the diagnostic and engineering program tests used by the manufacturer's
Installation team.
It Is Important In any new system to test all components and their Interactions as part

of the overall system. A system should operate without serious equipment failure for a
mutually agreed-upon period (usually 40 to 80 hours) before being considered for final
acceptance tests.
The final acceptance test procedures should be explicit. Good acceptance procedures
involve these factors:
(a) The schedule for the acceptance test period should be clearly defined. This
schedule should show how the time throughout the day should be allocated to
periods of operation, idleness, preventive maintenance, etc. The acceptance test period should last at least 30 days In order to obtain a good estimate of both the mean time between failures (MTBF) and mean time to repair (MTTR).
(b)

During each of the operating periods, the nature of the work which the computer is to be doing should be clearly defined. The work which the computer
should do during the operating period might be divided into cycles. In each
cycle the following should be performed:
(1)

Process actual, but tested, data for key applications.

(2)

Process special data designed to test all of the special features of
the equipment and any program packages supplied. (Experience
has shown that a selection of actual data will not begin to test all
of the possible conditions; therefore, a special input is desirable.
Conversely, a set of special data can never be developed to predict all the unusual conditions which occur in practice; therefore,
a large section of actual data is also desirable).

(3) Use diagnostic routines which exercise all parts of the equipment,
including peripheral units.
By repeating this cycle of tests throughout an operating period, a good test of the system
can be obtained. Of course, each program should be designed to check its own operation
so that any errors which the system makes are promptly reported. Any output should be
checked against specified standard results. The minimum performance level required
for acceptability during the test period must be agreed upon in advance. This agreement
might include minimum mean time between failures, maximum mean repair time, maximum repair time, and mimmum percentage operating time out of total on-time. Estimated performance speeds (as listed in AUERBACH Standard EDP Reports) can be used
as a basis to establish anticipated performance times.
Rental charges for the equipment should not be effective until the system components have
passed the stipulated acceptance tests.
For well-established equipment with many prior satisfactory installations, the acceptance
testing may be conSiderably simplified. A method often used is to operate the system for
a continuous period of one month on the normal work, loaded to the expected schedule.
Rent is then paid retroactively to the beginning of the period, provided a ratio of 0.90
(or better) chargeable time to scheduled operating time has been achieved .
. 23

Environment
The minimum environmental conditions under which the manufacturer's equipment will
perform satisfactorily should be stated. Allowable variations in the following requirements should be specified:
(a) Temperature and humidity
(1)

Equipment - in use and on standby.

(2) Magnetic tape - in use and in storage.
In these two areas. the specification will help determine the amount of air
conditioning that the user wUl have to install.
(b) Power
(1) Voltage requirements and permissible variation.
(2)

Frequency requirements and permissible variation.

(3) Waveform variations allowable.
Advance specification of these factors will help determine requirements for
power transformers and/or a motor-generator set.
6/67

A.

AUERBACH

\

(Contd.)

SPECIAL REPORT

.23

23:010.230

Environment (Contd.)
(c) Space
(1)

. 24

Free floor space around each equipment unit to permit access for
maintenance.
(2) Space to be devoted to the maintenance engineers, equipment, and
spare parts .
Maintenance

.241 Reliability
Reliability is measured as a ratio of serviceable time to the sum of serviceable time
and downtime (time when faults are awaiting repair or are being repaired, or faultcaused rerun time). It is frequently quoted as a percentage and often called percentage
"uptime" (values of 95 to 98 percent are generally expected). In general, only time
that had been scheduled for work by the user is considered in this calculation. A guaranteed uptime should be negotiated at least in the form of minimum serviceable hours
per day (usually equal to the time required by the user for his basic jobs, ranging from
8 to 20 hours).
A more technical method of specifying acceptable reliability is to indicate the mean time
between failures and the mean time to repair equipment failures. Proportions of uptime
and downtime can be estimated from these figures.
Under certain conditions, the importance of the data or of the workload situation will not
permit delays due to equipment (or any other) failure. In such cases, it is desirable to
specify that an emergency or "backup" facility be available. Charges incurred under
such circumstances are usually absorbed by the equipment manufacturer if the emergency
is caused by system failure .
. 242 Maintenance responsib1lity
The contract should define maintenance requirements and procedures, describing the
types of maintenance: fully attended, reSident, non-reSident, unattended, or emergency. In connection with a purchase agreement, there may be a need for a separate
maintenance and spare parts contract. In most rental contracts, the equipment manufacturer guarantees a minimum percentage of uptime or other assurance of usable time.
The responsibility for reliability then rests with the manufacturer. For both rental and
separate maintenance contracts, the level of skill, number of people, and their location
(e. g., user's installation or manufacturer's office) can be considered as discussion
points. In the case of on-site maintenance personnel, facilities such as space, power,
and furniture are usually supplied by the user.
Duration of scheduled maintenance should be specified in the contract after the level of
acceptable reliability has been agreed upon. The user should have the right to establish
his operating hours and the manufacturer should adjust scheduled maintenance times
accordingly. Attention should be given to the availabiiity of maintenance services during
scheduled extra shift operation and also during occasional unscheduled overtime requirements. The maximum time between the call for maintenance and the arrival of maintenance personnel might also be specified.
The method of scheduling and charging the time required to make any changes to equipment and/or engineering improvements should be stipulated. These items are usually
a matter of mutual agreement at the time of occurrence. For rental contracts, however, these usually include modifications or substitutions to maintain the equipment
equivalent to the "current product-line." In any case, an agreement should be reachen
on those types of improvements which will be installed at no cost and those which will
be paid for by the user. When improvements for increased reliability are necessary
(e.g., marginal components or units to be replaced) to maintain the percentage of uptime, they should be made at no cost to the user .
. 25

User's Rights
In the case of rental contracts, the conditions under which the user can modify and/or
maintain the equipment (if any) should be specified. Usually the user may rent time on
his own system to outside users in order to utilize slack periods. Sometimes the manufacturer will agree to buy time. In this case, rates and procedures should be established.

. 26 Additional Factors
.261 Special equipment
If any unit of the system is being constructed especially for the user, the contract should

include complete teclmical performance specifications. If the unit involves the interconnection of equipment from two manufacturers, the individual responsibilities for performance and maintenance should be carefully defined.
C 1967 AUERBACH Corporation and AUERBACH Info. Inc.

6/67

AUERBACH STANDARD EDP REPORTS

23:010.261

.261 Special Equipment (Contd.)

Price, delivery, and acceptance conditions for special units should be stated within the
terms of the contract. The policies adopted for regular equipment can usually be modtfied for special equipment .
. 2(}2 System design
Often the user's system is based on a design outlined in the manufacturer'S proposal. In
this case, the deta1l1ng of the system design and the extension of the system concept
should be accomplished with assistance from the manufacturer. The degree and level of
system design assistance is a point of negotiation. The number, level, and type of skill
of personnel assigned, the aSSignment of specific individuals, the responsibility of the
manufacturer's personnel, as well as their qualifications, are points which should be
considered. The tenure of their aSSignment should also be agreed upon, in addition to
the availability of additional manufacturer's supportpersonnelfor specific needs such as
writing special programs, debugging, or design of difficult parts of the procedures .
. 2(}3 '1 raining
Training courses may be specified to be held on the user's premises and/or at the manufacturer's training centers. The programming language to be used should be decided
upon early in the implementation program, and this language should be used in the training courses. The choice of a programming language is dependent on the avallabllity of
an operational translator p:-ior to the delivery date.
A "reasonable" number of programmers and systems analysts should be trained (usually
as many as the user actually intends to employ in these positions). Training is also
necessary for console operators. Advanced programming courses and orientation programs to be presented to top management personnel should be considered. If good systems courses (as opposed to programming and coding) can be made available, they are
espeCIally desirable for training new analysts.
As part of the training program, it is usual for the manufacturer to provide complete
training materials and reference manuals. Manuals and training materials should apply
to the equipment and the languages to be used, not to earlier systems .
. 264 Program testing
Ideally, the user's first applications should be pre-tested. This might be accomplished
on equipment provided by the manufacturer at another site. Usually no charge is made
for a limited number of machine hours for this purpose. The exact number of hours is
subject to negotiation .
. 2(};) SpeCIal programs
In some cases the user may wish to contract with the manufacturer to supply specific
operational programs (in addition to software packages). In this case, there should be
a firm mutual understanding of: the form of documentation of the programs provided;
delivery date; acceptance date; how changes and improvements will be made after the
program is accepted; how the user can train his own people on the program; and the
maximum permissible processing time or other measure of efficiency. The user will
have to prOVIde firm speCifications for the program early in the schedule and will not
have the same flexibility in changing requirements as he might have if his own group
were doing the programming. Attention should be given to the acceptance tests for such
programs. In general, manufacturers avoid negotiating penalty clauses for late software delivery, or for software that does not "perform as expected."
An industry trend toward the development of highly specialized, "canned" application
packages lends a great deal of sales appeal to some manufacturers' offerings. The user,
however, should not be over-impressed by the quantity and range of such offerings,
unless the packages can directly benefit his particular needs .
. 261) Conversion credits
Where possible, the user should obtain, either in the contract or in a separate agreement,
terms specifying credits for operation of the system in parallel with a replacement system if. at some future date, conversion to a larger system becomes necessary. Typically,
covers ion credit periods last for thirty days .
. 267 Investment tax credit
The Investment Tax Credit applies to the lease or sale of data processing equipment,
\Ilth thE' manufacturers granted the 7% credit by the U. S. Government. The manufacturers retain the option to use the credit themselves or to pass it on to the users. Since
this 7rlc, can represent a Significant amount of money, the user should determine what
the manufacturer's present investment tax credit policies are, and what they are likely
to be in the fu'ure.

6/67

A

(Contd. )

AUERBACH
to

SPECIAL REPORT

23:010.268

.268 Purchase-leaseback
The emergence of computer leasing companies within the past three years offers a third
alternative to the purc;;ase or rental of data procel:lsing systems. By depreciating equipment over longer time periods, leasing companies can often give their customers lower
rental rates than can the original manufacturers.
The most common leasing arrangement is purchase-leaseback, which Involves a threeway exchange. The user buys the equipment from the manufacturer and subsequently
sells it to the leasing company. The lessor then leases it to the user at lower rates than
he originally had been paying or would have paid to the manufacturer. In other cases. the manufacturer sells the equipment directly to the leasing company, which in turn leases It to the user.
Purchase-leaseback should be investigated prior to the acquisition of a data processing
system because in many cases savings ranging from 10 to 30 per cent of the normal
rental rates can result. Some flexibility, however, is sacrificed, since leasing companies generally require longer-term leases .
. 27

Summary of Considerations
As the preceding paragraphs have pointed out, there are many significant factors to be
considered in 'contracting for an electronic data processing system. These factors are
recapitulated below, in a form that may be used as a checklist in negotiating a contract.
•

Basic Specifications
Equipment - the manufacturer should provide detailed specifications of
the equipment units.
Software - specifications should indicate the software to be provided.
The user should assure himself that the software provided
will operate on the equipment configuration selected.
Type of payment - the user should be aware of the various types of payments possible, aside from outright purchases and rental
contracts. The user should also Investigate the tax
implications involved with a particular agreement.
Amount of chargeable time - rental agreements should clearly define the
amount of chargeable time included in the basic rental fee.
In addition, a definition of the "amount of chargeable time"
should be stated.
Chargeable time - a definition should be provided for the time that Is to
be counted as chargeable time.
Assurance of serviceable time - this time should be speCified by the manufacturer; and in the event of a major failure, what backup
facilities are available.

•

Acceptance
Shipping and installation - payment for these services should be mutually
agreed upon during contract negotiations. Some charges
are undertaken by the user while others are absorbed by
the manufacturer.
Delivery and acceptance dates - these dates, and associated penalties,
should be established during contract negotiations. Software packages should be delivered before equipment to
allow for familiarization and use.
Acceptance tests - these tests should be specified and the test procedures made explicit. The amount of time that tests should
run satisfactorily before the equipment is considered
acceptable should be stipulated in the contract.

•

Environment - the minimum environmental conditions under which the manufacturer's equipment will perform satisfactorily should be
stated.

•

Maintenance
Reliability - the minimum level of reliability and methods of maintaining reliable operation should be agreed upon at contract
negotiation.
Maintenance responsibility - maintenance of equipment responsibility
and the types of maintenance provided should be specified.

•

User's Rights - conditions under which the user can modify and/or maintain the equipment and rent time to others should be agreed
upon.

•

Additional Factors
Special equipment - price, delivery, acceptance conditions, and vendor
responslbilities should be specified.

~

1967 AUERBACH Corporation and AUERBACH Info, Inc.

6/67

AUERBACH STANDARD EDP REPORTS

23:010.270

. '!.7

Summary of Considerations (Contd.)
System design - support from the manufacturer may be desirable In
detailing system design and system concepts.
Training - training courses should be provided by the manufacturer, and
the location of the training center be specified.
Program testing - iniUal programs should be pre-tested, perhaps on
equipment provided by the manufacturer at another site.
Special programs - the user may contract with the manufacturer to supply specific operational programs other than the software
packages provided.
Conversion credits - the user should arrange for credit for parallel
operation in the event conversion to a larger system takes
place at some subsequent date.
Investment tax credit - the manufacturer's present and future Investment
Tax Credit poliCies should be considered to determine
whether a saving to the customer is applicable.
Purchase-leaseback - computer leasing arr2ngE'Ments should be considered for possible savings to the user .

.3

ORGANIZATION OF THE SURVEY
The foregoing considerations should be clearly specified by the manufacturer in the form
of contracts, supplementar:r agreements, proposals, or letters of intent. To aid the prospective user in negotiation&, the AUERBACH staff has made a survey in which the standard terms offered by the manufacturers to commercial and government users were
analyzed and summarized.
The arrangement of the tables which summarize the results of this survey is based upon
the U. S. Government's Invitation for Bids to manufacturers of data processing equipment
(General Services Administration Solicitation No. FPNN-E-27332-N-1l-22-65. The
General Services Administration issues such an Invitation for Bids each year; then it
negotiates a one-year contract, running from July 1 to June 30, with each computer
manufacturer. This contract, which in some cases is not finally negotiated until after
July I, then forms the standard contract between all Federal agencies and the manufacturer concerned.
Because the U. S. Government is such an important computer user, the aims of its
negotiators and the contracts which they negotiate are extremely influential in setting
computer marketing trends. The aims of the negotiators are clearly indicated in the
Invitation for Bids, which forms the basic framework for each round of contract negotiations, and the contracts themselves are part of the public records.
The tables summarize the contract terms that were sought by the U. S. Government
negotiators for the currently existing contracts, with references to the particular section
of the Invitation for Bids that provides a detailed explanation of each point. Alongside
the terms sought by the U. S. Government for each contract factor, the tables summarize
the tenns currently offered in the standard government and commercial computer rental
contracts of each of the following manufacturers: Burroughs, Control Data, General
Electric, Honeywell, IBM, NCR, RCA, SOS, and UNIVAC. The tables were prepared by
obtaining, analyzing, and summarizing a copy of each manufacturer's Authorized Federal
Supply Schedule Price List and (where available) a standard commercial contract form.
The material to be published was submitted to each manufacturer for prepublication review and was discussed with the manufacturers' designated representatives for verification and clarification where necessary.
The U. S. Government's Invitation to Bid for fiscal year 1968, covering the period from
July I, 1967 to June 30, 1968, closely parallels last year's solicitation, with the following notable changes:
(1)

The government may terminate the contract after giving 30 days'
notice. Previously, 90 days' notice was required for terminating a
contract involving the removal of an entire system.

(2)

Liquidated damages for failure to deliver software on schedule is the
lesser of the basic daily rental rate or $100 per day per item of software delayed, including all software inoperable as a direct result of
the delay. Previously, the liability equaled the greater of $100 per
day per item of software delayed or the basic daily rental rate.

(3)

\

Acceptance tests require satisfactory performance at a 95% effectiveness le\'el instead of the former 90% level.

!

(4) A liability credit equal to the pro-rated basic rental for service call
response time in excess of one hour has been incorporated into the
solicitation.

6/67

A.

AUERBACH

\
(Contd.)

23:010.300

SPECIAL REPORT

.3

ORGANIZATION OF THE SURVEY (Contd.)
(5) The government may exercise its option to have equipment replaced
when downtime exceeds 5% (formerly 10%) of the total operational
use time per month over a period of three months .

.4

CONCLUSIONS
In compiling and analyzing the tables of computer rental terms, the AUERBACH standard
EDP Reports staff arrived at four signlficant conclusions:
(1) Commercial contracts tend to omit many of the user safeguards that
U. S. Government contracts include.
(2) Terms in the standard contracts, both commercial and government,
vary widely enough so that they may well constitute a decisive factor
in the decision to rent a specific computer system.
(3) Most manufacturers are willing, in varying degrees, to alter the
terms of their standard contracts through clauses which are added
during contract negotiations.
(4) From the user's viewpoint, standard contracts as presently written
are inadequate in a number of important respects •

. 41

Omission of User Safeguards
Among the subjects that simply are not specified in most of the standard commercial
contracts are: firm delivery dates for hardware and software, standards for acceptance
tests (or even the existence of such tests), and guidelines for asseSSing penalties for
nonperformance. It would be nice to believe that all the equipment will be deliverd on
time, that all the required software will be available when needed, and that both the
hardware and software will always perform according to expectations; but these are
assumptions that no businessman can afford to make without some clearly-specified
assurance - such as the terms requested by the U. S. Government negotiators .

. 42

Variations Among the Manufacturers
Areas where the standard contract terms vary among the different manufacturers seem
to be more prevalent than areas where the terms are in agreement. Extra-time
charges (for operation beyond the time allowed by the basic monthly rental) can effectively double the rental cost of some computer systems, while involving no extra cost
on others. Purchase options, by crediting some portion of the previously-paid rental
charges, can reduce the purchase price of a system 75% or more in some cases, or
by a maximum of only 20% in others; the options are free in some cases, but involve
an extra cost in others. Discounts for users who cannot keep their equipment busy
throughout a full shift now appear in some contracts, but not in others .

. 43

Possibilities for Negotiations
r-Iost of the standard commercial contracts are far from sacred, so the user is likely
to find it worthwhile to engage in some bargaining before Signing the contract. During
the preparation of this survey, we received comments from manufacturers' representatives which indicated that they are in a pOSition to offer varying degrees of flexibility
in their contract terms. depending upon the particular user's needs, the competltive
situation. the potential for addItIOnal business, and other variable factors. This
flexibility of terms applies to various manufacturers' poliCies regarding delivery,
extra-time charges, acceptance tests. performance standards, program testing time,
purchase option credits, and nearly every other item in the standard contracts except
the basic monthly rental. Checks among computer users confirmed that contracts
currently in force do vary Significantly from one another as a result of clauses added
during negotiations.

. 44

Most manufacturers are willing to negotiate contract terms with the user until an
agreeable settlement has been reached. While IBM tends to hold firmly to its standard
contract. it is often possible to negotiate certain terms with the branch manager in a
letter of intent. Although the strict legality and enforceability of such a document are
questionable, IBM has tended to honor these as gentlemen's agreements. There are
reliable indications that, when dealing with IBM. negotiations at the branch manager's
level usually produce the best results. When deaUng with other firms, however,
negotiations at higher levels seem to maximize the user's benefits •
Inadequacies of the Standard Contracts
Most of the current standard contracts do not offer the computer user as much protection as he might reasonably expect. None of the standard contracts reviewed in this
survey offers assurance that the program run times or software performance promised
in the manufacturer's proposal will actually be Ilchieved, nor is any penalty specified for
failure to achieve the anticipated throughput in the user's installation. Even where
damages are speCified in the standard contract., the liability rates are generally inC 1967 AUERBACH Corporation and AUERBACH Info. Inc.

6/67

23:010.440

.44

.5

AUERBACH STANDARD EDP REPORTS

Inadequacies of the Standard Contracts (COIltd.)
adequate to compensate for the actual losses; hence, the user generally remains
"locked in" and must try to make the best of a less-than-satisfactory situation. Despite
the current emphasis on "integrated product lines," none of the current standard contracts assures the user that a faster, program-compatlble system will actually be
available to him when he needs it. Such assurance would help the user to formulate his
future expansion plans with far greater confidence •
SURVEY COMPARISON CHARTS
The survey tables that follow summarize the standard contract terms that are currently
applicable when computer systems are rented. The information contained in this Special
Report should be well worth studying at an early stage in every computer procurement
program, and enlightened use of this information (together with appropriate legal counsel) should help to ensure that the resulting contract w1ll be a reasonably comprehensive
and satisfactory one.

/
(

\

6/67

A•

AUERBACH

23:010.501

SPECIAL REPORT
COMPUTER RENTAL TERl\iS

SllBJECT MATTER

What IS tbe minimum rental
period?

TERNS SOUGHT BY U. S. GOV'T
(From GSA SOllcltaliOD of
10/28/65).
PERIOD: JUly I, 1966 to
June 30, 1967
One year or e.s (Sect. A-I. I(a).)

CONTROL DATA
STANDARD TERMS

BURROUGHS STANDARD TERMS

Commercial
(6/67)
year

Goverameld:

(7/66 to 6/67)

u<>" roqueo,.
(Soe 2nd 001. )

I~

Commercial
(6/67)

year.

C".ovemftlent

90 days for a complete computer
system. or 30 days for any
component tbereof.
(Section A-I, l(a).)

90 daya, after
m11limum reDtal
period.

What software is to be suppUed,
and when?

As written 1Dto the COIltract. plus
future work developed by the
manufacturer for geaeral use.
(Section A-I, 2(b).)

Aa GSA roq_". As GSAroq ....... As written into
(See 2nd col. )
(See 2Dd col. )
the contract.

A 8 GSA request •.

Lesser of pro-rated basic monthly
rental or $100 per day per item
of software delayed. IncludtDg
software facilities rendered unusable as a result of the delay of
supportlDg faclUttes. (Section
A-I, 3(b).)

NODe.

As GSA roque.... NODO.
(See 2nd col. )

What is the mmimum acceptable
performance during acceptance
tests ?

90% good time tbrougbout 30 days'
nmning# with at leaat 100 hours
.sed dur!JIg the period.
(Socllon A-I, 4)

Unspecified.

As GSA requests. Unspecified.
(See 2nd col. )

How many hours of operational use
are allowed in the basic monthly
rental (under GSA Opllon B)?

200 hours per mODth.
(Secllon A-I, 5(b).)

176 houra.

What is the standard rate for UDlimited usage (GSA Option A),
expressed in terms of the basic
monthly rental?
5-day week:
6-day week:
7-day week:

This is DOt meDttODed in the
invitation to Bid.

How is the amount of central
processing time used computed for
establishIng the rental due?

OOly that lime between program
START ad program STOP,
measured either by meters or by
userls estimates. (Sect. A-I,
5(a).)

85500-115%

OtIaen IIIIII)MIctflod.
(.7 -day week )

···

1H.

(6/67)

GOVE'rnment
(7/66 to 6/67)

Commerelal
(6/67)

Govel"nment
(7/66 to 6/67)

6 months (200
Serlea). 1 year
(400 and 800)

As GSA rtoquesl.
(Se~ 2nd col )

1 year

90 day., after
minlmum rental
period.

Ae GSA requests
(See 2nd col )

As GSA requests.
90 daye, after
mlnlmum rental (See 2nd col. )
period.

Uupecifled

As GSA request.

Unspecified.

As GSA request •.

commercial

Government

Commercial

Government

Commercial

(6/67)

(7/66 to 6167)

(6/67)

(7/66 to 6167)

(6/67)

I year.

A. GSA requests 1 year.
(See 2nd col. )

All GSA requests UnspecifIed
(See 2nd col. )

As GSA requests UnspeCified.
(See 2nd col )

As GSA requests
(See 2nd col I

As GSA request3. Unspecified.
(Soe 2nd col. )

As GSA requests None.

As GSA requests
(See 2nd col. )

As GSA requpst8
(See 2nd col. )

As GSA requests.

Unspecified.

(See 2nd col. )

As GSA roq ......
(See 2nd col. )

Unspecified.

All GSA requests.
(See 2nd col l

Uupeclfied.

As GSA requests. Unlpecified.
(See 2nd col. I

As GSA requests.
(See 2nd col. )

Unspecifled.

$100 per day
total maximum
liability.

As GSA roque.... Unopectfled.

As GSA requests.

Uupeeified.

As GSA requests. UnspeCified.
(Soe 2nd col.)

As GSA requests.

Unspecified.

(Soe 2nd col: )

As GSA requf'sts
(See 2nd col)

Unspecified.

(Soe 2nd col. )

176 hours.

As GSA reque8ts
for all systems
except NCR 304,
which i. 176 hra.

20%,
Approximately
20%; DO extrau.e charge for
mOlt peripherals.

::::f:
B500

Approximately
20%: DO extraue charp for
mo.t peripheral.

B55OO-115%
Others _pectflecl.
(7 -day weok' I

120% of Item.
120% of Item.
.ubject to extra subject to extra
uoe charp.
use charge.
(7 -day week )
(7 -day week )

110%.
111%.

AeGSA.--.... Aa GSA roqueat. As GSA requests. Unspecified.
(See 2Dd col. )
(See 2nd col. )
(See 2nd col. )

How is the uaage time of peripheral This is not meDliOlled by the
UDits computed for establishing the invitation to Bid.
rental due?

Power on time
les8 user mainterumce and idle
time.

Dlrectu ....
where Ibi. . .
easily meuurable; otbem.e,
aa for OBDtral
proceoaor, If the
perlplleral ..
octually ...ed In
the run.

How loug wUl 1t take a serviceman One hour maximum.
to respond to au emergency service (Sectloo A-I, 6)
call?

UlI8peclfted.

110 apeclfie
1JII8peclfled.
amoUIIII: of time.

What credit is allowed to a user
when a system is down?

Credit at basic reDtal rates for
each machine inoperative u a
result of the malflmcltoo. whenever tbe _ . period
12 boure. (SecIiOll A-I, 6)

None.

Wbnever machlae-flilure dowDtime """_ 10'1. of tota1 _ a tlODal _
time for tb_ caaaacuII... mDlllluo. (Sect. A-I, 8(a).)

Unopeclflecl.

GOVf'rnment
(7 / 66 to 6'67)

As GSA r("quellts
(See 2nd col )

$100 per day total NOlle.
maximum
Itabtltty.

176 hours.

(6/67)

A c; GSA requests 90 days, after
(See 2nd col )
minimum rental
pertod.

I

UnspeCified.

Commercial

Aa GSA rt"quests

All GSA requests. 90 days, after ,Aa GSA requests 90 days after
90 daya. after
minimum period
minimum rental (See 2nd col. )
minimum rental (See 2nd col. )
of rental
period.
period

As GSA requests. As stated In
(See 2nd col.)
contract. •

As GSA requests.
(See 2nd col.)

1 year.

Government
(7/66 to 6167)

UNIVAC STANDARD TERMS

As GSA request. J year
(See 2nd col )

A. GSA requeatll.

(See 2nd col )

(Soe 2nd col. )

(See 2nd clll )

81lS STANDARD TERMS

RCA STANDARD TERMS

NCR STANDARD TERMS

Unspecified.

178bro. ~. CPU - 178110"". CPU -116 bouro: 200 houri.
82500 BlOO
moat perlpherala mOllt peripherals
83500 8200
- unlimited ue. - unlimited ule.
8300
8220
B5500
2S", ~

Aa GSA requests.
(See 2nd co!. )

Commercial

IBM STANDARD TERMS

Ae GSA requests.
(See 2nd col )

Power on time
le.s user ma1ntenance and idle
time.

R.... mucb equipment dowDIIme
before the _ r may, at Id. aptIaD.
elect to have the flllily equ!pmeDt
repIIieed?

Aa GSA requeats.

HONEYWELL STANDARD TERMS

As GSA requeltl. Unspeclfted.
(See 2nd col. )

108%.

ex_

Government
(7/66 to 6167)

(See 2nd 001. )

None.

Wbst damages will be paid If the
software is not dehvered on time?

B5500

year.

Uupeclfted.

As GSA requelt•.
(See 2nd col. )

rMoo
83500

1

AIJ GSA requests.
(See 2nd col. )

As GSA requests. Nooo.
(See 2nd col. )

25%
B200
8300

(6/67)

All GSA reque.ta 90 days, otter
Aa GSA roq_". 90 daye, alter
A. GSA roque....
(Soo 2nd col. )
minimum reatal (See 2nd col. )
minimum rental (See 2nd 001. )
period.
period.

What damages wilt be paid if the
Baste pro-rated rental of the
NODe.
hardware is not delivered on time? system, with a mtntmurn of $100
per day delayed. (Soct. A-I, 3(a).)

This is not mentioned in the
pressed as a percentage of the basi Invitation to Bid.
rental rate?

Commercial

(7/66 to 6/87)
(Soe 2nd col. )

How much notice is needed to
cancel the contract?

What is the extra usage rate ex-

GENERAL ELECTRIC
STANDARD TERMS

(~e

2nd col.)

Non~

(!=\f>f'o

2nd col I

As GSA requests. None.
(See 2nd col. )

As GSA requellts
(See 2nd col )

As GSA requests UnspeCified
(See 2nd col )

As GSA rE'quests Unspecified
(See 2nd col )

As GSA rf"quests
(See 2nd col J

Unlimited use
option only.

Unlimited use.

Unlimited use
optJon only.

Unlimited use
option only.

Unlimited use
option only.

Unhmlted UBe
optIon only

None.

Not applicable.

Not apphcable.

Not applicable

Not applicable

(See 2nd col.)

None.

200 hours.

As GSA requests.
400, 1400. 800,
(See 2nd col. )
1800-176 hours
If on 1 yr CODtract; otherwise#
200 hour./month.
200-200 boursl
month.

176 hours.

20%,

As mut.ally
.,reed upon in
rental contract.

As mutually
agrf"ed upon in
rental contract

IBM-e.tabU.hed Either 10% or 30% 40%,. baaed on
depending on
each indivtdual
bUlable rate••
which generally equipment.
component.
vary from 10%
to 30%.

No extra charif'.

Nfl""

tm.pecifled

Unspecified.

Unspecified.

Not applicable.

No extra charge. No extra charge. No extra charge.

tpeClfied.

As GSA requests
(See 2nd col. )

As GSA requests As GSA requests. As GSA requests As GSA requests. Not appllcable.
(See 2nd col. )
(See 2nd col. )
(See 2nd col. )
(See 2nd col. )

Not appUcahle.

UnspeCified.

Not apphcable.

Not apphcable.

Not apph<-able

Not appltcable.

Unspecified.

Not applicable.

Not applicable.

Not apphcable

Option unavailable Approximately
10%.
, 7 -day week)

No extra charge. No extra charge

No f'xtra charge.

108%.
1I0%,
111%.

AB GSA requests.
(See 2nd col. )

:.;.

Baaed on CPU

No extra charge.

Unopeclfted.

UDipecified.

Direct usage
where this Is
easily measurable: otherwise.
as for central
proces.or. if tbe
peripheral is
actua11y used in
the run.

Direct usage
where practtcable; otherwlle
18 for CPU. if
the peripheral
is actually used
in the run.

Direct usage
where practicable; otherwise,
.. for CPU. If
the peripheral
Is actually used
in the run.

2 bou....

Unspecified.

As GSA requests.

Unspecified.

No definite com- Best effort .•
mitment, but IBM
"ahall alwaya be
responsive to the
Deeds of the Covt. I

Response will be iDD8pectfted.
IDltiated witbln one

As GSA requests. 1;..... time
(Soe 2nd col. )
"redlt eq.iv-

As GSA requests

IBM retains option Onapectfled.
to: (1) provide
backup equtpmenti
(2) provide Oft-site
cuatomer enlt:neer
or (3) replace the
equlp""",t.

NCR reta!no optloa IIlIapaclflod .
to: (1) provide 00site ma1DtelllDce;
(2) provide _ up equipment; or
(3) mike every
effort to rep.....
Il1o equlpmeat.

meat oubject to
extra 1IIIe charge

No defiDlte corn- Uapeclfled.
mitment, but
CDC Illhall
alwaya be re8poDslve to the needs
of the Govt. II

Aa GSA roque.... AaGSA_.. As GSA roquest•.
(See 2Ild col. )
(See 2nd col. ).
(See 2nd col. )
except when
perlod_
24 boure

AaGSArequo.... Unapeclflecl.
(See _001.)

Unapeclfled.

limo for equtp-

Unopectfled.

AaGSArequo.... Uupeclflod.
(See 2nd 001. )

(See 2nd col. )

Option B-Aa GSA Unllpeclfied.
requeo... (See 2nd

As GSA requests.
(See 2nd col.)

Unspecified.

M GSA request•.
(See 2nd col. )

Unspecifted.

~.nt to
~Ime.

001.)

OptI_AadC varte•.

Aa08A-,". ",,-lfiod.

(See_ool.)

From relation to Direct usage
Not applicable.
tme measured where practicable,
or CPU.
otherwise. as for
CPU, If the
peripheral Is
actually used in
the run.

As GSA requests . tunspecified.
(See 2nd col. )

As GSA requests . !Best effort. ..
(See 2nd col. )

As GSA requesta
(See 2nd col.)

As GSA requests None
(Soe 2nd co\. )

hour.

Unspectfled.

(S.. 2nd col.)

When fault
continues for
48 hours after
notification, at
a mutually
.....edupon
rote.

AI GSA reque.ts . Unspecified.
(See 2nd col. )

No definite time
speclfit'd. shall
be responsive to
needs of Govt

As GSA requests.
(See 2nd col. ),
except credit
of 0 511 of
haslc monthly
rental per hour.

As GSA requests . ~ever at customeI' As GSA reqllesta.
(See 2nd col. )
(Soe 2nd col. )
pption .•

© 1967 AUERBACH Corporation and AuERBACH Info, Inc.

6:'67

AUERBACH STANDARD EDP REPORTS

23:010. 502

COMPUTER RENTAL TERM!;l (Contd.)

TERMS SOUGHT BY U S. GOV'T
(From GSA. SolicItation of
10/28/65)

Sl'BJFC r MA rTFR

PERIOD· July 1. 1966 to
June 30. 1967

ttow much computer tlme

1S

pro-

vided frt"t' of chargt' prIor to
inst.lll.J.tlon)

Enough time to allow successful
operation of all specified appUca.hons on Installation day.
(Section A-I. 10)

BURROUGHS STANDARD
TERMS

Commercial
(6/67)

CONTROL DATA STANDARD
TERMS

Government
(7/66 to 6/67)

B5500. B100. 200. B2500 I1I1d B3500
- 6 bours/
82500. 3500 - 6 $1.000 of bu1c
boIIrs/$1.000 of
monthly rental
basic monthly
with a maximum
rental, to a max- oC 60 hourll.
imum IIf 60 bours 05500 - 3 bours/
$1.000.
B100. 200. 3006 boura/$l. 000.
300 - 40 hours.

GENERAL ELECTRIC STANDARD
TERMS

Commercial

Govermnent

Commercial

(6/67)

(1/66 to 6/67)

(6/67)

50 hour. maxl-

100 hra. or

mum, or 2 breI 2 brall1. 000 of
$1.000 of bulc
monthly rental,
whichever Is
leas.

bastc mODthly
l'eDIaI, whlehever Is leaa.

Government
(1/11 to 1/17)

Grealer of 40

As GSA rehou.ra pr 3 hours questa.
per $1.000 of
(See 2nd col )

Commercial
(6/67)

Unapeclfied

Gov.rnment
(1/66 to 1/67)

IBM STANDARD TERMS

Commerctal

(6/67)
Unapeclfled.

G hours/S1. 000

ofbblC

monthly rt'ntal

basiC monthly
rental.

Government

(7/66 to 6/67)

NCR STANDARD TERMS

Commerctal
(6/17)

Varle •• depend10 bouroI$1.000
Ina on equipment. of bas1c monthly
but doea not meet rental
GSA requeat
verbaUm.

Government
(1/66 to 6/67)
10 hours/" .000

RCA STANDARD TERMS

Commerc1al
(6/67)

(7/66 to 6/67)

Unspecifled

Spectra

Government

of bulc monthly

8D8 STANDARD TERMS

Commercial
(6/67)
Unspecified

70/15-10 hr.
70/25-30 hr.
70/3.-40 bra
70/55-50 hrs
20 hw.ra extra
allowed for eom
munieatlons

rontal Minimum
time Is 15 to 40
hour•• dependtna
on equipment

UNIVAC STANDARD TERMS

Government
(7/66 to 6/67)

CommercLal

Governmen&

(6/671

(7/66 to 6/67,

20 hours

unspecified

from 2
I toVaries
7 hour. per

maximum.

i each $1.000 of

bu1c monthly
rental.

i
!

IIYlltema

i

How much computer tlJlle 16
ProVided free of charge after
Installation?

All avatlable tune oulilide baBic
rental period for the firat 90
days. plus COBOL. FORTRAN.
and ALGOL compUation time as
required (SeCtIon A-1. 10)

As GSA requests.
(See 2nd col. )

AaGSAr_ta.
(See 2nd 001. )

All available
time for 90 days
without paylDt!
ema-use
cbarpa.

All available
ttme for 180
daya wltltoot
paying extrause charge•.

Unspecified.

As mutually
"",eedupon,
or twlce the
.....edpr.InatallaUon time

What charge lS made for machme
tllne needed for program testmg
after Cree time allowance has

ThlS 18 not mentioned. by the
Invitation to Bid

7 !% of haslc
rental ratfo 82500 and 83500.
25% ofbasle

Bas Ie rental
rate

Basle l'eDIaI

Baalc rental

Unspecified.

Unspecified

been exhd.usted?

HONF.\·WELL STANDARD
TERMS

rate.

rate.

UnapeeUied.

Storle!. 200 ,
60 days.
All others 90 days.
200 h;"'r. max
IfoneyweU Data
Center rates

Butc rental
rate.

Varies from 30
to ~o days

Any unused

Unspecdled

minimum time
rematntng:

Any ....oed
minimum time
remaming

NCR Data
Center rate..

NCR Data
Center rates

Unspeculed.

As GSA

None·

None.

None .•

None.

requests.
(See 2nd col.)
I

Uupeclfled

Bas 10 rental
rate

BaSIC rental
rate

Unapeclfled

BasiC rental
rote

Vary mg rates

speclfled 1n
prlLe book

Prevaumg
1 priCE'S In
current GSA
bcbedu.le

rental rate B200. B300.
B5500
\\bat reduction In monthly rental
IS allowed lf full utulzallon )5 not
achieved?

&orne defuute reduction IS requU"ed (Section A-I. 5(d) )

\\l1at dIScounts m the rental rate
are apphcable In special

DISCOunts are requested for:

SituatlOllS ')

None.

None

None.

None.

None.

None.

None.

UnspecUled

None.

• Mulllple systems

None.

UnspecifIed.

None.

None.

None.

None.

• Educational use

None

Nont'

None.

20%. plus unlimited ua.,els
pennltted.

Unapeclfled.

50% of bulc
Unapeclfled.
rental rate,
applloable only
to equipment
lnatalled prior
to June 30. 1965.

120-10%
200-25%.
400-50%.
800-25%

• Hospitals

None.

None

None.

20%, plus unlimited us_ Is
permitted.

None.

NODe.

Unspecllfled.

UnspecifIed.

Unspecified.

None.

None

None

None

None

None

None

I

:-'one

'on~

UnspecUled

None

UnspeCified

None.

None

7%

None •

None.

None·

UnlpecUted.

Generally 20%

20%

20%

Unspecified

Spectra 70/35
I1I1d 45-15%
Spectra 70/5520%

NODe.·

None

None·

Unspecdied

10% or 30% per
Item; 20% If
afflhated with
a medical
institution.

None.

None.

Unspecified.

None.

None.·

None.

r-.one •

Unapeclfled

None

None.

None.

Unspecifted

None

None .•

None.

:-O:one •

cIrcumstances:

What 18 the purchase price if
a uller purchases the equipment
"" baa been renttng?

_?-

Does the manufacturer pua on
Invetltment Tax Credit to

-

NODe.

None

NODe.

None.

NODe.

None.

Unspecified.

None.

• Whenever the purchase price
of the equipment 18 reduced.

None.

None.

NODe.

None.

None.

None.

Unspecified.

NODe.

Un8peclfled.

None

None.

None

Unspecified

None.

None •

None.

None·

• As Boon as the eqlnpment bas
become obsolete (Tbls I.
conSidered to occur as aoon
as a Buccessor bas been
announced)
(SectIon A-I. 5(d).

None

None

None.

None.

None.

None.

Unspecified.

None.

Unapeclfled

None

None.

None.

Unspecified.

None.

None·

None

None·

Tbe credit altould talre Into acc_
the pbySleal _ of the ayatem
reoted. I1I1d the total rental paid
hy the user.
(Section A-I, 19).

Free credit of
65% of total rental
darlDt! firat
12m_:80%of
r _ dDrIDC 3Dd
12 mODlba; maxImum credit of
60% of purcltase
J'1'lee.
82500 II 83500l'eDIaI crediIB
acorue for firat 12
montha only: 70%
for 1-8 mOB. , 40%
for6-12m_.

equipment

What credit 18 allowed If a user
purchases the eqJ.ipment he has
been rentmg?

None

I
I

~one

I

None

:
( None.

I
I
I

~Id

Free credit of
65% of total rental
paid durlDt! firBt
12 montbo: 60% of
rental darlDt! 2nd
12 mantha i max..
imum credit of
60% of purchue
Pl'lce.
82500 II 83500 rettta1 credIIB
accrue for fir8t 12
montbo 0DIy: 70%
for 1-8 moe.
fo.6-12_.

60% darinl flrot
12 m _ : 40%
dDrIDC 3Dd 12
m _ . 0pIt0a
good for life of
o_act. Max.
oredll of '0%
ofparcltue
price.

,40%

Free oredlt of
60% of total
reatal paid darlng ftrot 24
montltB:40%
darlng remalalngm0ntb8.
Option laalB for
llfeofOODtract.
MaxImum credit
allowed 18 70%
of purcbaae
prlee.

Free oredlt of
50% of reata1
paid. up to 50%
ofpurcbue
prlee.

The lesser of tile tIlen-current or current price at CUrrent prlee at CUrreIlt price at As GSA requeata. As GSA"-BIB.
the orlglnal pure""e price.
exe»etae of option. e ...olee of optIoD

Free credit of
50% of rental
paid, up to 50%
ofpurcbue
prlee.

fu GSA reque.ta.

'"......lee of
optIa.

(Section A-1, 19).

Not applicable.

'one

"'one

I

,

\\bat rental adjustments may come A rental adJu,sbnent is requested
mto force?
In eaclt of the following

• When the rental paid exceeds
the purchase price of the

'on~
I

!Yea.

Not applicable.

No.

Not applicable.

Uupeolfled.

Not oppIloable.

Serle. 200Free credit of
80% within 12
montltB:60%
within 24
montltB.whlch
ts maximum
credit perlOCi
Serle. 40080% within 12
montha: 50%
wtthln 24
mOlltha

Series 200Free credit of
80% wlthl. 12
months; 60%

within 24
montbo ..... leb
.. maxunum
credit period
Serle. 400 80%Wlthio 12
months. 50%
within 24
montha.

Aa GSA requeata. As GSA requeata.
(Se. 2nd 001. ).
(See 2nd col. ).

Yea. Oft 5 year
leue oontraatll.

Not applicable.

Free credit of
45% to 70% of
total rental paid
Option 18 val"
for 1 year (2
ye8l'll for Btate
and local cav'tS)

Free credit of
from 10% to 75%
of total rental
paid. Varying
Credits for ace.
OptIOD lasts for
contract life.

Varies conalderably. depending OIl all
relevant
factore.

NCR 310. 315.
390. 500. 420:
0.833% of list
price for each
month of rental.
Maximum credit
of 75% Option
lute 24 months.

Free credit of
65% of total
rental paid, to
a maximum of
75% of orlglnal
purcluute prtce.

Free credit of
65% of tota1
rental paid.
Option laaIB for
life of contract.

Sigma Serle. 50%
Nine Serlee 70% of rental
during fir.t 6
montbo: 50%
darlDt! 2nd I
moatba. Max:lmum oredlt of
70% for NIDe
Serle. and 50%
for SIgma Sarle••

As GSA

Orlglnal purebue Aa GSA
price .•
requeata.
(See 2nd col. )

Option lasts

for Ufe of

contract.

Unspecified.

As GSA

Orlgmal price.

requests.

Unapeclfled

!Not oppIleable.

As GSA

Orl1llDal prlee.

requests.

"-ata.
(See 2nd col. )

See 2nd 001.)

Yes.

Not IIJplleable.

(See 2nd col. )

·Uaepeclfled.

Not 8I1I1Ileable.

-.

Daly OIl fouryear or 1 _

fixed-term

6/67

A

Sigma Serle. 50%
Nine Series free oredtt of
70% of rental
paid dDrlDt!
firet 6 months;
50%darlDt! 2nd
6m_a.55%
darlDt! 2nd y.ar,
60% In 3rd year ,
IUld 65% ID BUCce••tve yeara

1-6 months 90%.6-12
montbo - 75%.
12-24 month. 60% of rental
paid on a Doncumulative
baats, maxlmum credit of
70% of _chue
prtce.

1-12 months _
90%. 6-12
mOlltbo - 75%.
12-24 months 60% of rental
paid on a cumulatlve buta.
maximum
credit of 70%
of purchase
prtce.

Aa listed In
cODtract for 1

GSA prtces

year after

for applicable
fiscal year.

rental com ..
menoee,
thereafter
prevalllDt!
ratel.
Not applleable.

Only for 5-)'ear
non-camoeUable
qreeD1et1ta.

Not IIJpllcabie.

(Contd. )

AUERBACH
(' viewed as a fundamental weak point in computer-based information systems - too slow, expensive, and unreliable to be tolerated in applications involving large volumes of input information. The
one solution to this problem is the automatic character reader - a device that has been
developed to the point where it has replaced manual keypunching in selected application
areas, although it still lacks certain functional refinements that will he necessary to make
It o;llltabJr for the full spectrum of computer input operations.
Character readers are machines for directly converting alphanumeric characters or symbols into a machine-readable form. The output of the readers may be in the lorm of
punched cards, punched paper tape, or magnetic tape - or the readers may he operated
on-line (directly connected) to a computer.
1\lost current readers arc severely limited in the type fonts they can read, and, in some
cases, in the size of the character set (alphanumeric vocabulary) they can handle. On the
other hand, character readers are in effective and economically efficient use in several
major industries. Banking is probably the largest current application area for character
readers. The credit-card industry, led by the oil companies, and utility bill processing
are other major application areas. In addition, some retail merchandising firms are now
using character readers, and the United States Post Office Department (which is already
using optical ZIP-code reader/sorters) has expressed interest in seclllg a character reader developed to read hand-written addresses.

.2

Character readers offer the advantages of being faster and morp accurate than manual keypunching, since they permit printed dntn to be entered directly into d:lta-processing systems without any additional human action. The present purchase prices of commercial
magnetic character readers average around $80,000. The prices for optical character
readers range from $80,000 upward, depending upon the speed and sophistication of the
machine (rentals run between $3,000 and $15,000 per month) .
CHARACTER READER TYPES AND FUNCTIONS
There are two basic types of character readers: magnetic and optical. Magnetic character
readers are used almost exclUSively within the banking industry. They can handle only
special type fonts printed in magnetic ink. The font most widely used in the United States,
and adopted as a standard by the American Bankers Association, is Font E-13B - a highly
stylized font that can be used to represent only 10 numeric digits and 4 spe~ial symbols
(Figure 2). Another font, which was developed by Compagnie des Machines Bull-General
Electric, is capable of representing all the characters in the alphabet as well as all the
numeric symbols (Figure 3). However, the Bull font, which has been adopted as a standard
by the European banking community, can at present be read only by the Bull C1\1C-7,
GE MRS200, and Olivetti 7750 magnetic character readers.
Since magnetic readers detect only magnetic marks, non-magnetic dirt or other marks will
not cause reading errors. However, cons iderable care must be taken with the quality of
the printing on the documents. Ink densities and character imagc are both critical.
Relatively high quality-control standards must be maintained in the printing process to
prevent character deterioration and extraneous ink spots.
Optical character readers are used in nearly all major application areas, including banking.
They work on the principle of recognizing the difference in contrast between the characters
and the background on which they are printed. Some optical readers do not require special
fonts and are theoretically capable of reading most type fonts (with suitable adjustments).
So far, however, this theoretical capability is too expensive to realize for most commerical
applications, although there are several optical character readers that can read more than
one type font. The least expensive units are restricted to one font, which is usually
specially designed for low error rates and is often restricted to numerics plus a few
special symbols. Also, optical readers tend to be somewhat less reliable than magnetic
readers because of their greater sensitivity to dirt, document creases, and poor paper
quality. Despite these drawbacks, optical readers Seem to offer the most promise for the
future, and new techniques are being explored and dfilvelOPed to overcome the major functional problems.
© 1967 AUERBACH Corporation and AUER8ACH Info, Inc.

10 67

AUERBACH STANDARD EDP REPORTS

23:020 200

.2

CHARACTER READER TYPES AND FUNCTIONS (Contd. )

All existing commercial character readers, whether magnetic or optical, consist of three
basic functional units.
•

Document transport,

•

Scanner, and

•

Recognition unit.

A functional block diag-ram of a typical character reader is shown in Figure 1.
Document"

Scanner
Umt

TI,ln:-'pOfl

I!ntt

Punched Cards

Document

Output
HO!'Per

ReeogrlltlOn
Glut

M,lgnt'tl(' Tal'£'

Punched I'ape r Tapt.'

Cunt/ol Sq!.n._llb _ _ _ •

To
' - - - - - !Jal.!
Procesbor

IJ.ILI Flo\\' _----+

Figure l.

Functional Diagram of a Character Reader

The function of a character reader's document transport is to move each document to thc
reading station, position it properly, and move it into an "out" hopper. Transport mechanisms can be divided into two basic types: one for handling individual documents (paper
sheets or cards) and the other for handling continuous rolls (cash register or adding
machine tapes).
The function of a character reader's scanner is to convert the alphanumeric characters and
symbols on a document into some analog or digital representation that can be analyzed by
the recognitIOn unit There are two basic methods for accomplishing this: magnetic and
optical.
The recognition unit is the Iwart of the character reader. This unit matches patterns from
the sc:umer against reference patterns stored in the machine and either identifies the
patterns as specific characters or rejects them as being unidentifiable.
n()el' 1\\ r. en THA;-..rSPOHTS

Document transports in character readers designed to handle adding machine or cash
register tapl's cons ist of a tape well in which the paper roll is loaded, paper guides, and
a paper drive control. Once the tape has been manually threaded, the paper is automaticall) moved past the read head in a manner similar to the movement of a fum reel ill a
movie projector. A vacuum system is frequently used to keep the paper flat. Tht: maxImum length of the paper roll that can be handled ranges 1rom 1200 feet for the :\atlOlICll C'lo,h
Register Optical Journal Reader to "any reasonable length" for the Hecognition Equipment Journal Tape Reader. The paper-roll mechanisms are usually designed so that the
roll can be h'lCked up any time rereading is required. A special feature 01 the 1eeder
mechanism used in the Hecognition Equipment .Journal Tape Heac\tor j.. an automatic tape
:Idvance, which speeds up tape movement when there are large spaces between print lines.
1n most other rf'aders, tap" speed is constant at all times.
III character readers designed to hanclle individual sheets or cards, the document-tr'illbport
function Ib divided into two phases: (1) feeding the documents from the input hopper, and (2)
transporting the documents past the reading station. A com1110n device for document feeding
is called a friction feeder. This consists of a belt wound around capstans and partially restmg on the document stack. Constant pressure is exerted against the belt by the document
stack. As the belt moves across the top of the stack, it pushes the top documents into a
separator station, where a combination of rollers and anothpr belt separates the top document from all documents below it. This technique is used in the IBl\\ l-i19 J\Iagnetic Character Reader.
\'acuum or suction feeders are also used to lift documents off the input stack. One example
01 a vacuum fepder is used in the Philco-Ford 6000 Print Reader, which employs a pair of
vacuum belts to lift the document from the stack and carry it for\\'ard to the transport UI1lt.
Both the friction and vacuum devices, however, have problems in handling documents of
thIn paper aJld may occaSionally feed more than one document at a time. A feeder, designed by Rabino\\' Electronics (a subsidiary of Control Data Corporation) uses a set of

10 67

IA

AUERBACH

(l'llntd

I

SPECIAL REPORT

.3

23:020.300

DOCUMENT TRANSPORTS (Contd.)
cone-shaped rollers to feed the documents. The rolling cones engage a corner of the topmost document and roll the corner away from the pile up into paper rollers. which carry
the document to the transport unit. This unit is said to eliminate the possibility of feeding
two sheets at a time.
A popular method for transporting the document to the reading station is a vacuum-drive conveyor belt. Some character readers, such as the IBM 1428, use the conveyor belt to place
the document on a rotating drum, which moves the document past the read head. The paper
is held to the drum by means of a vacuum.
One of the basic disadvantages of the above mechanical techniques is that they cannot move
the document as fast as it can be read. One approach to this problem has been the use of a
high-resolution CRT scanner, developed by Philco Corporation, which can scan the entire
document without requiring any mechanical movement. Another method, used by RCA,
uses a vidicon scanner which takes a picture of the entire document at once. Both of these
systems will be discussed later in this report .

.4

MAGNETIC SCANNER UNITS
Scanner units, as previously mentioned, are divided into two basic categories: magnetic and
optical - and these deSignations are used to characterize the readers themselves.
Since the banking field represents the major application area for magnetic character readers,
all of the magnetic readers produced in the United States have scanning units designed to handle the E-13B font shown in Figure 2.

Figure 2.

Sample of E-13B Font Characters

Most scanning units convert the magnetic characters into an analog voltage waveform for
subsequent identification. The principle used is based on the electrical signals that are
generated by moving the characters past the read head. Each character generates a signal
that has a unique waveform, which the recognition unit matches against reference waveforms. The companies presently using this technique are Burroughs, General Electric,
and National Cash Register.
IBM uses a digital scanning technique, which is exemplified by the IBM 1419 Magnetic Character Reader. In this machine, each character is scanned by 30 magnetic heads stacked
vertically and interconnected to give 10 outputs. The outputs are transmitted to a 70-bit
shift register in the recognition unit, where they are matched against stored reference
patterns .
.5

OPTICAL SCANNER UNITS
Optical scanning methods are based on the differences in contrast between the characters
and the background on which they appear. The function of the scanner is to sample either
portions of a character or a complete character to determine the relationships betweer.
light and dark areas. The common types of scanners used are mechanical discs, flyingspot scanners, parallel photocells, and vidicon scanners .

. 51

Mechanical-Disc Scanner
The mechanical-disc scanner consists of a lens system, a rotating disc, a fixed aperture
plate, and a photomultiplier, as Shown in Figure 4. The characters to be read are flooded
with light, which is reflected from the surface of the document into a rotating disc via the
lens system. The disc has apertures extending from its center toward its periphery. As
the disc rotates, the apertures pick up light samples. A fixed aperture plate regulates the
amount of light and directs the light to a photomultiplier. The photomultiplier tube converts
the light samples into signal pulses. By varying the voltage threshold, the photocell outputs
can be adjusted for different background colors.
The mechanical-disc scanner senses a character of data at a time. Movement between
characters and lines is accomplished either by moving the document, as in the NCR Optical
Journal Reader, or by repOSitioning the lens system, as in the IBM 1428 Alphameric
Optical Reader. Consequently, this type of scanner is relatively slow by comparison
with the other scanners mentioned.

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

10/67

AUERBACH STANDARD EDP REPORTS

23:020.520

~HOTIMIULTI~LII.

_VI." .EPUCTID

LIGHT IN'O ILICTRICAI,.

'MPULlrS

LIGHT
SOUIIe[

Figure 4.
· 52

Mechanical-Disc Scanner

Flying Spot Scanner
The flying-spot scanner consists of a cathode-ray tube, a projection lens, a phototube, and
a control unit. A beam of I ight is generated in the cathode-ray tube and deflected across
the tube in a scan patt(:rn. The lens system projects this scanning light spot onto the document, from which it is reflpcted into a phototube. The phototube generates a voltage signal
whose level is proportional in each instant to the amount of reflected light, thus indicating
light and dark areas. The resulting signals are then either fed dir('ctly to the recognition
unit in analog form or first transformed into digital form.
The flying-spot scanner offt.rs more flexibility than the mechanic:al disc, since its scanning
pattern can be automatically acljusted by the control unit. This p"rmits the use of different
scanning modes (i. e. , scanning certain character fields, scanninb specified portions of the
document). Also, being completdy electronic, it is faster than the mechanical disc and is
generally classified as a medium-speed device.
The introduction of high-resolution cathode-ray tubes (2000 optical lines) has made manufacturers look to the development of a reader in which a complete document can he scanned
without any document motion other than that required to position it under the read station.
A scanner of this type is found in the Model 6000 Print Reader developed by Philco-Ford.
Sylvania Corporation has worked on the development of a similar device, which was expectc·d
to achieve very high reading speeds of up to 6,000 characters per second.

· 53

Parallel Photocells
The use of a vertical grouping of photocells speeds up scanning operations by simultaneously sampling a number of points which, when combined, add up to a complete vertical
slice of the character. The electrical signals generated by the photocells are then quantitized into either black, white, or gray levels. This data is fed into a shift regisll'r and
stored until data on the entire character has been accumulated. Due to the parallel sampling, this type of scanner can achieve higher speeds than the flying-spot scanner.
A variation of this method that eliminates the need for shift rcgisters uses a full "rl'tina"
of photocells to sample an entire character rather than just one vertical slice. Besides
eliminating the shift register, this method also increases reading speed to approximately
2,400 characters per second. Rabinow Electronics (a subsidiary of Control D;~ta Corporation) and Recognition Equipment are two of the companies currently using :1 retina bf
photocells for sanlpling. This sampling technique has the present capability for auhievmg
a higher speed than any of the previously-mentioned techniques.

· 54

Vidicon Scanner
So far, we have discussed scanning methods that read ch..ll'acterf> hy refl,'cting li,;ht frol '
the document to one or more photocells. A totally different method being used is to proj' d
the characters onto a vidicon television camera tube and scan the active surfac(' \11th an
electron beam. Thc resulting video signals are quantitized to digitally il1(II.:alo ilIac:" 01'
white.
This type of scanner is currently heing used by RCA in their 70/251 Docum,'nl Header.
By storing a group of characters on the tube, the need for document movement during
the scanning operation is eliminated in cases whl're the document contains a rl'asonably
small number of characters. The advent of high-resolution vidicon tubt's pould p 'I'mit th,·
character capacity to be increast'd to the point where clocument movement during scanning
will be eliminated on most documents.
Another advantage of the vidicon scanner is spl.'cd. Since it takes only a ft'w milliseconds
for the bean1 to scan the entire lube, a full grouping of stored characters can be rl'ad in that
tin1e. At present the RCA device can scan UP to 1300 characters I1t'r second.

.6

RECOGNITION

U~ITS

/

Recognition units probably represent the areu of greatest technical development in Ihe character reader field. Because of the rapidity tlf the progress being made, we will limit our
discussion to the five most common types of l'l'cognition units now available commt'rciall.\ .
(Contd. )
10/67

SPECIAL REPORT

.61

23:020.610

Optical Matching
Optical matching was one of the earliest recognition methods to be used. It is based on the
use of two photographic masks for each character. One mask is a positive transparency of
the character and the other is a negative transparency. The positive transparency shows
all the Significant areas that should be covered by the character, and the negative transparency shows those areas that should be left blank.
The negative and positive images of the unknown character are projected onto their opposite
masks; 1. e. , the positive image is projected onto the negative mask, and the negative image
onto the positive mask.
Phototubes behind each mask detect any light passing through. A character is identified by
first measuring the total light passing through each of the reference masks and selecting the
one that passes the smallest amount. Character identification or rejection is then made by
comparing the amount of light passed through the selected mask with a threshold value.
Ideally, no light should pass through the reference mask if it matches the character being
identified. In practice, however, the match is seldom precise enough to completely blank
out all light, which is the reason for establishing the threshold value as a tolerance.
The advantages of the optical-matching technique are its ability to idcntify a full alphanumeric character set and its relative SimpliCity, which makes it less expensive than some
of the other techniques. Also, the masks can be manually changed to enable the reader to
handle different character fonts. The major disadvantage is that errors are easily caused
by characters that do not meet strict standards of shape and registration. Also, there may
be problems in distinguishing between such similar letters as "Q" and "0" or between different punctuation marks .

. 62

Analog Waveform Matching
Analog waveform matching is another recognition method that has been in use for some
time, particularly in the magnetic character readers used by the banking industry. It is
based on the principle that each of certain characters passing under a read head will produce a unique voltage waveform as a function of time; that is, the waveform of each character will differ either in shape or length with respect to time. Characters are identified
by matching their waveforms against reference waveforms.

. 63

Machines using this technique have reading speeds of approximately 500 characters per
second. The principal disadvantage of this system is that only a limited number of characters have unique waveforms. Consequently, this technique is found mainly in systems
dealing with a limited character set .
Frequency Analysis
Frequency analysis is a digital recognition method developed for fonts consisting of closelyspaced vertical lines. The outstanding example of this kind of font is the Bull magneticink font shown in Figure 3. The Bull CMC-7 and Olivetti 77 50 magnetic character readers
use this recognition technique. The widths of the gaps betwcen the vertic:ll lines of each
character are measured by variations in magnetic flux. An unknown character is identified
by comparing the sequence and number of its narrow and wide gaps with ~tored codes for
each of the alphanumeric characters. An analog version of this technique IS currently being
used in the General Electric MRS200 Document Reader.
The advantages of the frequency-analysis technique include the ability to accommodate a full
character set, and increased reading speeds .

. 64

Matrix Matching
This technique, one of the more widely-used, stores the scanner signals in a digital register
that is connected to a series of resistor matrices. Each matrix represents a single reference character. The other end of each matrix is connected to a second digital register,
whose voltage outputs are representative of what should be obtained if the reference character were present. Recognition is based upon the resultant output voltage obtained from
each matrix.
The advantage of the matrix-matching technique is that the resistor matrices can be modifield easily, making it easy to change character fonts. In addition, a full alphanumeric
character set can be read. The technique also has the advantage of being quite fast, since
the matching is done by resistor matrices. Reading speeds of up to 2,400 characters per
second have been obtained. The technique is similar in theory to the optical-matching
technique described earlier. but it can handle misregistered characters much more effectively. The numerous machines using this technique are listed in the comparison chart .

. 65

Stroke Analysis
This technique, used by Farrington Electronics, is based on the strokc or line formation
of each character. The characters are differentiated from each othcr by the number and
position of vertical and horizontal strokes. The formation of the unknown character is
matched by a special-purpose computer against a charMter truth table, which indicates

© 1967 AUERBACH Corporation and AUERBACH Info. Inc.

10/67

23:020. 650

. 65

AUERBACH STANDARD EDP REPORTS

Stroke Analysis (Contd. )
the stroke formation for each reference character. At present, this technique is limited
to identifying only a special character font called the Selfchek font, which emphasizes
straight lines. Work is being done to generalize the technique so that it can be applied to
any character font.
Stroke analysis has the advantage of being able to handle a full alphanumeric character set,
but the maximum speeds obtainable by the Farrington character readers are about 300
characters per second, which is low compared to the 2,400 characters per second obtained
by machines using the matrix-matching technique. Also, the stroke-analysis method does
not have the font flexibility of the matrix-matching technique because of the need to change
the wire recognition program in the special-purpose computer every time it is necessary
to switch to a different character font .

.7

ECONOMICS AND SELECTION CRITERIA
The questio., of whether it pays to replace a manual keypunching operation with an automatic
eharacter reader cannot be answered in any general way. The answer depends upon the
characteristics of the specific application - particularly upon the volume of input data that
must be regularly handled, the accuracy requirements of the input operation, and the speed
of the computer. A rule of thumb that can be helpful in reaching a preliminary decision on
whether to seriously investigate the use of a character reader is that an installation preparing 10, 000 input documents per day or requiring 8 to 12 keypunch operators is about the
smallest that might gain from using character-recognition equipment. As the daily input
volume approaches 30,000 documents, character readers tend to cost significantly less
to operate than keypunch devices. The final criterion for making the decision is, of course ,
the number of characters produced per dollar. A simple formula for determining this cost
is to determine:
a
F ~ b+ c
Where:
F = number of characters processed per dollar.
a = total characters processed per month.
b

=

monthly equipment rental and overhead costs.

c = monthly employee salary costs, including supervision and fringe rates.
An example follows for a 10, OOO-document-per-day installation, using the following
parameters:
Keypunching

OCR Reading

No. of keypunch operators = 12

OCR rental = $4000/mo.

No of direct supervisors = 1

OCR machine O/H = $500/mo.

Operators' salaries (including O/H) =
$400/mo.

Operators's salary

Supervisor's salary (including O/H)
$750/mo.

OCR throughput

=

l\Iachine Overhead = 10c;.

= $400/mo.

= 300 char/sec.

Characters/document = 64

l\Iachine (026) rental

= $60/mo.

Keypunch throughput

= 7500 char /hr.

Reject rate = 10'70
(Rejected documents must be keypunched).

Effective hourS/day = 7
Days per month

~

20

The application involves the processing of strictly-controlled, field-typed documents.
CASE I - KEYPUNCHING OPERATION
a = (12)(7500)(7)(20)

12,800,000 char/mo.

b = (12)(66)
c = (12)(400) + (1) (750)

$ 792
4800 + 750

5550
~

F

=

Total cost per month.

12,800,000
$6342
2018 characters per dollar.

10/67

fA

AUERBACH
~

(Contd. )

23:020.700

SPECIAL REPORT

. .,

ECONOMICS AND SELECTION CRITERIA (Contd. )
CASE D - OCR OPERATION
a = 12,800,000 (using same volume as in Case I;
OCR reader's potential throughput is actually
approx. 43 million char/mo.)
b = 4000 + 500

$4500

c '" (1)(400) + (. 2)(750) + 635
operator supervisor keypunching of
rejects (10%)

H8S

$56115 Total cost per month.

F - 12,800,000

-

$5685

= 2251 characters per dollar.
Figure 5 sbows the same relationsbips calculated for volumes of
documents per day.

~,OOO

through 30,000

$,500

$,000

4,500

4,000
NO. OF CHARACTERS
PROCESSED
PER DOLLAR

5,500

loOOO

2,500

2.000

25,000

30,000

35,000

NO OF DOCUMENTS PROCESSED PER DAY

Figure 5. Comparison of keypunching versus OCR costs.
(See text for basis of curves.,
There are four major criteria for evaluating optical character readers. Cost, as discussed
above, is the most obvious one, but it must be carefully rt-Iated to the functional capabilities
of data throughput speed, flexibility, and reliability. tlaturally, all three of thes!' capabilities directly influence the cost of character-reading equipment; but, as is the case with
all equipment, the initial cost is only part of the story.
Throughput speeds are a function of reading speed, document transport speed, data density
on the document, and multi-line or page reading capabilities. The rated rearling speeds of
optical character readers currently on the market range from about 200 to 2,400 characters
per second. You will find, when comparing machines of different speeds and prices, that
the number of characters read per dollar tends to increase at a much faster rate than
machine costs.
Better performance in terms of flexibility and reliability might also save ynu money over
the long run despite the higher initial equipment cost incurred. Flexibility pertains to a
reader's ability to read a variety of character fonts. as well as its rescan ability (i. e .•
ability to re-read a line of characters). paper-bandling capability, and special format
features. The ability to read only selected fields and to skip over crossed-out characters
are two format features that are very useful In some applications.
Reader reliability Is. of course. a fundamental criterion. The reliability of a character
reader is measured by its reject and error rates. The reject rate is generally defined as
the percentage of the total documents read which the reader rejects because it is unable to
recognize one or more characters. The error rate refers to the percentage of documents

C 1967 AUERBACH Corporation and AUERBACH Info, Inc.

10/67

23:020.800

AUERBACH STANDARD EDP REPORTS

.7

ECONOMICS AND SELECTION CRITERIA (Contd. )
containing one or more characters which were incorrectly identified by the reader. The
reject rates of present readers range from about 2% to 20%. while the error rates generally
do not exceed 2%. The best way of Judging the reliability of a character reader is to compare it with the error rate of the current keypunch operation which the machine is being
considered to replace .

.8

TRENDS AND FUTURE DEVELOPMENTS
The scope of applications for character readers is currently limited primarily by their inability to read a variety of different fonts. by their poor performance on handwritten documents. and by the lack of standardization within the industry. Consequently, considerable
development effort is being put into these areas, as well as into improvements in reliability
and speed .

. 81

Multi - Font Capabil ities
'~e work being done on the development of multi-font character readers is taking the
form of three basic approaches: manual, semi-automatic, and fully automatic.

The manual method consists of altering the recognition logic by manually replacing such
machine parts as plugboards and optical masks. This method is low in cost but is clearly
inadequate for reading a stack of documents in which the character fonts are mixed.
The semi-automatic approach consists of effecting changes in the recognition logic by
means of operator controls. This means that either the machine must store all the different reference patterns that can occur, or the recognition parameters must be modified
by means of a special-purpose control unit. The latter technique is used in the presentlyavailable Philco-Ford 6000 Print Reader. Although it has the advantage of being flexible,
it is expensive. The monthly rental for the Philco-Ford character reader is approximately $15,000, as compared with the typical rental charges of around $3,000 for firstgeneration character readers.
The automatic technique demands a recognition unit that can automatically sense a change
in the character style and adjust itself to the change. This is really a self-adaptive or
learning machine, a type of device that is still in the early experimental stages.
82

Recognition of Handwriting
Since each individual has his own style of handwriting, it is difficult to set any recognition
standards for handwritten characters that will not lead to a high reject rate. Consequently,
this problem is even more preplexing than the multi-font recognition problem, because the
recognition logic of the machine can never be set for a particular style.
The work being done on the recognition of handwritten characters can be dIvided into two
classes: hand-printed characters and script. Some of the techniques currently being investigated in connection with hand\\Titten documents are curve tracing, detection of
selected features, and context recognition (which is discussed below). Although a number
of companies are working on the problem, most of the work has been kept confidential.
The primary customer for a reader capable of handling handwritten documents appears to
be the C. S. Post Office Department.
Three companIeS presently offer machines capable of reading a limited hand-prInted character set: Optical Scanning Corporation (OpScan 288), IBM Corporation (1287), and
Recognlt'on Equipment Corporation (ERCR) .

. ti3

Improvements in Reliability
:\aturally, reliability in the form of low error and reject rates is a prime consideration in
all the de\ elopment "ork being done on character readers. One approach that is being
followed to reduce these rates is to improve the resolution of the scanning units and thereby
increase the number of sample points from which the equipment can make an identification.
As previously mentioned, Philco-Ford Corporation is using a cathode-ray tube that has a
resolutIOn of 2,000 optical lines. Even better r('solutJOn can be expected in the near future.
A longer-range approach to the reliability problem is the work being done on "context recognitlOn." This is an attempt to simulate a human being's ability to read by context. When a
person reads, the legibIlity of individual letters or even individual words is usu:tlly not
critlCal This is because human beings read letters within the context of the entire word
and words within the context of the entire sentence, Consequently. the word "Quie" in the
phrase "Ouic and dirty" would easily be identified in context by most human readers as the
word "Quick .. , even though the first letter of the word is an "0" and the last letter is
missmg.
The first thing needed to automate this process of context recognition is a group of fundamental rules that will aid the machine in identifying the characters on the basis of the context in which they are used These context rules must be chosen to agree with the type of
material being read. If a new application is added, then new rules should be instituted.

10/67

A

AUERBACH

(Contd.)

SPECIAL REPORT

. 83

23:020.830

Improvements in Reliability (Contd. )
Chang{:s of these rules can be accomplished by utilizing either hardware (e. g •. plugboards)
or programming techniques.
Although context recognitIOn is not yet sophisticated enough to become the major clement
of a recognition scheme, it can be used as a backup method for identifying illegible characters.
The most obvious advantage is the ability to identify a complete word even if one or two
characters present recognition difficulties. Context recognition will certainly involve an
enormous increase in the storage capacity and logical capabilities of character readers,
but this may be justified by the increase in efficiency that can be attained, Ilowever, tne
economics 01 context-recognition readers will remain highly spec.:ulative until c.:onsiderably
more development work has been undertaken.
Context recognition also prom ises to be useful in the problem 01 reading handwriting. It
could be the basis of a technique for reading complete words rather than a character at a
time. Again, it would radically increase the storage requirements and the cost for a reader,
but the results might well be worth it. Again, the economics will remain unclear, pending
additional development work .

. 8..

Improvement in Speed
Another, though less critical, area of development emphasis in character-reader engineering is speed. The major limitation on reading speed is the amount of time it takes to
mechanically move the document past the reading station. Work now under way indicates
that this limitation will be removed by overlapping the two functions of transporting and
scanning documents. This is already being done in the Document Reader RCA 70/251 through
the use of a vidicon scanner, which photographs an entire card-type document and performs
the scanning function within the cathode-ray tube. This allows a new document to be moved
into place while the previous one is being scanned. Speed can be further increased by the
use of control logic that permits selective scanning; i. e. , scanning only those areas of the
document that contain pertinent information .

. 83

Improvements in Standardization
The "jack of all trades, master of none" theory can certainly be applied to recognition
logic facilities. Great sacrifices in reading reliability and increased costs result from
having to recognize a multitude of font styles. The first giant step toward standardization,
however, has taken place in the acceptance of the USASI standard character set for optical
character recognition. Assuming that all subsequently designed OCR equipment contains
facilities for reading this character set, the industry can expect greater cooperation from
users of business forms and manufacturers of data processing equipment, and much progress is sure to result.

86

Summary
Although the optical character recognition field is still relatively new, and much work
remains to be done in improving equipment performance and developing more flexible
readers at lower cost, the past year has seen some Significant developments. Several
multi-font readers are now available, and three machines capable of recognizing handprinted characters has been introduced on the market. Reliability has improved significantly,
with one manufacturer claiming reject rates of less than 1% and error rates of less than
O. 1% in turnaround applications. As a result of the recently adopted USASI standard OCR
font, a trend toward low-cost, s ingle-font readers may take pl·ace concurrently with the
development of larger multipurpose machines, introducing a wider cost spread than exists
at present. If this happens, and a truly low-cost, reliable optical character reader results,
we can expect to see OCR replacing punched cards as the primary computer input medium .

.9

THE COMPARISON CHART
The accompanying comparison charts summarize the Significant characteristics of representative optical and magnetic character readers in terms of the type of document feed
and transport unit, document size, document speed (documents/minute), types of
scanners and recognition units, type font, character set, and reading speed. It should
be noted that the indicated reading speed usually represents a maximum or potential speed;
the actual speed is dependent on the size and number of documents being read.

~

1967 AUERBACH Corporltlon and

AU~RBACH

Info, Inc.

10.67

AUERBACH STANDARD EDP REPORTS

23:020 901

COMPARISON CHART: OPTICAL CHARACTER READERS
IDENTITY

DOCUMENT
HANDUNG

Document size, inches
(width x length)

Control Data Corp.
915 Page Reader

FarrllllltOn Electronics
Page Reader,
Model 2030

FarrllllltOn Electrorucs
Document Reader,
Model 3010

4.0 x 2.5 10 12.0 x 14.0

4.5 x 5.6 10 8.5 x 13.5

2.0 x 2.25 to 6.0 x 8.5

No. documents/min.

180 lines/minute (approx.)

150-400 lines/minute

440

Traneport type

Conveyor belt

Drive rollers

Drive belt

Feed mechanism

Vacuum

Vacuum

Friction

Sorllng facilities

Dual output stsckers

Dual output stackers

Three output stackers

Max. characters/line

110

75

64

INPUT
FORMAT

Max. hnes/inch

6

6

6

Max. lines/pass

78

70

5

Max. readtng speed,
characters/second

370

400

330

Font styles read

USASI font

Selfcheck 12F and

Character set

Alpnameric

Alphameric

CHARACTER
READING

RECOGNITION SeaMing techruque
Recogrutlon method

I~L

Alphameric

Parallel photocells

Mechanical disc

Mechanical disc

Matrix matching

Stroke analysis

Stroke analYSIS

FLEXlBILITY

Reads selective fields under Format control by plugboard. Format control by plugcontrol of computer program reads selecltve fields
board. reads selective
fields

ERROR CONTROL

Character display; marks
documents; manual correctlOn by keyboard entry. has
res can feature

OUTPUT

Rescan rea ture. character
display. manual correctton
by keyboard entry. rna rks
documents

Data to computer; or punched Data to punched cards.
punched paper tape. or
magnettc tape

en rds, punched paper tape I

or magnetic tape
ioPERATING CONTROL

lRescan feature. data field
check digit

lData to computer. or punched
~3rds. punched paper Lope.
/Dr magnettc tape

On-line with CDC 8090,
3000 Series. or 6000 Series
computers

Off-line

Reads mark-sense

Underscore feature permits
encoding of upper and lower
case characters In output
record

Batch header feature. marksenSing head (optIOnal);
hst-prlnter (opltonal)

$84, 000 (plua control unit
and output unit)

$99,500

$99,500

AV AI LABILITY

4 months

6 to 9 months

6 to 9 months

FIHST DEUVERY

November 1965

April 1967

September 1965

SPECIAL FEATURES

1----~

10/67

USASI, Selfcheck
12F, 12L, or 7B;
IBM 1428

--

PI'RIlXJMA l'E PURCHASE PRICE

A

Off-hne. or on-hne With
any computer

(Contd. )

AUERBACH

'"

23:020 902

SPECIAL REPORT

COMPARISON CHART: OPTICAL CHARACTER READERS (CONTO. )

DOCUMENT
HANDUNG

INPUT
FORMAT

CHARACTER
READING

FarrIngton Electrorucs
Journal Tape Reader,
Model 3040

Farrington Electronics
OpUcalReader/Card
Puncb, Model 3020/3022

.':lrrlllJ,1.on t:h'ctronlcs
Page RI·.lCh·r, MIMlcl 3030

Document Size, \Dcbes
(widtb x Icngth)

Standard 51 or 80 column
tab cards

4.5 x 5.6 to

No. documenta/mID.

550

laO-'101I ~I-,,~/mlnutc

Transporl Iype

Drive bell

Drlvl' rolkrb

Ora ve rollers

Feed meC'h.lnlsm

FrIction

Vacuum

.Journal spools

IDENTITY

H.:. x 13.5

Journal tape, I to 350 ft x
I -5/16 to 4-9/16 inches
400 lines/minute

SorI1Dl! faclhl1es

Dual output stackers

Dual outpul t,l.lC'kcra

-

Max. char.lcters/hDe

65

75

:IL

Max. hnes/lOch

6

6

Ii

Max. lincs/oass

I

70

-

Max. readIng speed,
characlers/second

600

400

\000

Font styles read

USASI; Selfcheck 12F, 7 B.
IBM 1428, 1428E

USASI. Selfch .. ck 12F and
12L

S.. lfcheck 12F /12 L. USASI.
IBM 1428. NCR NOF

Character set

Alphameric

Alphamcrlc

Numeric plus alpha conlrol
symbols

ScanDlng techntque

Mechamcal disc

Mechanical dIsc

FIYlRg spot

RecognItion method

Stroke analysiS

Stroke an,tiys'8

Stroke analYSIS

FLEXIBIUTY

Format control by plugboard. limIted selectIVIty

Reads selccllve fields,
operator programm.lblt.·
by software, formallln!:
and editing faellillen
provIded

Formal control by plugboard or external computer
pro.:ram

ERROR CONTROL

No rescan feature; data
fIeld check dIgIt; automat.c
IRsertlon of correct character; punch check

Character dlspLI)'. marks

Rescan feature. character
d •• play, keyboard insertIon
rna rks Journal lapes

RECOGNlTIO~

documents. manual corrcc-

tlOn by keyboa rd ('nlry.
has rcSC3n fcatur£"

OUTPUT

Punches Hollerith code
mto mput cards

Data to computer or
punched cardb. punchl·d
paper lape, or ma~netlc
tape

0" ta to magnetIc tape or
('omputer

OPERATING CONTROL

Off-hoe

On-hne w.th
computer

Off-hne, or on-hne Wllh
a nv .:omputer

SPECIAL FEATURES

Batch header feature;
serial and sequential
numbering. reads reverse
.mage

Reads mark-sense.
accumulates tot.II~,
extenslv(' rorm,IUIn~ and

APPROXIMATE PURCHASE PRICE

$100,500

$143,000

$107,000

AVAILABIUTY

6 to 9 montbs

6 to !I monthb

6 109 months

FIRST DEUVERY

December 1966

Januarv I(Hi7

March 1967

edltlR~

0~1J 6~O

.Journal tape header enlry.
magnetIC tape label entry

((-.Itun :-.

© 1967 AUERBACH Corporation and AUERBACH Info, Inc

10 67

AUERBACH STANDARD EDP REPORTS

23:020.903

COMPARISON CHART: OPTICAL CHARACTER READERS (CONTD. )
IDENTITY

DOCUMENT
HANDUNG

INPUf

FORMAT

Document Size, Inch ••
(width x length)

General Electric Co.
Bar Font Reader,
Model DRD 200

IBM 1282
Optical Reader Card
Punch

IBM 1285
Optical Reader

2.75 x 4.0 to 3.875 x8.0

Standard 51 or 80 column
tab cards

Journal tape, 36 to 200 ft x
1-5/16 to 3-1/2 Inches

No. documenta/mln.

1200

200

2200 lines/minute

Transport type

Drive belt

Clutch

Conveyor belt

Feed mechanism

Vacuum

Friction

Vacuum

Sorting facilities

Multlstackcr

Single stacker

-

Max. characters/line

65

32

32

Max. lines/Inch

6

10

6

Max. lines/pass

1

I

-

Max. reading speed,
characters/second

2400

267

540

Font styles read

GE COC-5 Bar Font

IBM 1428; SeUcheck 7B

IBM 1428, NatIOnal Cash
Register NOF

Character set

NUJ'1erlC

Numeric

NumerIc

Photocell

Mechamcal disc

Flymg spot

Bar spacmg analysis

Stroke analysIs

Curve tracIng

FLEXlBIUTY

No format control;
limited field selectIvIty

Formatting under control
Reads selectIve fields,
format control by plugboard of external computer
program and program card program; hmlted field
selectiVIty
on drum

ERROR CONTROL

No rescan feature; has
error indicator

Selfcheck dIgIts WIth
automatIc insertlOn of
correct character, res can
feature

Rescan feature, character
display WIth manual keyboard entry. marks

CHARACTER
READING

RECOGNITION Scanning technIque
RecogrutlOn method

documents

OUTPUT

Data to computer

Punches HollerIth code
mto input cards

Data to computer

OPERA TING CONTROL

On-lme WIth GE-4UO
computer

Off-hne

On-line

SPECIAL FEATURES

Mark-sensing feature;
reads bar code

Reads mark-sense, reads
reverse image

APPROXlMATE PURCHASE PRICE

$56,000

$72,000

$84,000

AVAlLABIUTY

12 months

12 months

6 months

FIRST DEUVERY

2nd quarter

March 1965

September 1966

I

1968

A

.,

AUERBACH

(Contd. )

23:020.904

SPECIAL REPORT

COMPARISON CHART: OPTICAL CHARACTER READERS (CONTD. )
IDENTITY

DOCUMENT
HANDlJNG

Documeat alae, IDchea
(wldt1l x leqth)

IBM 1287
OptIcal Reader

IBM 1418
Optical Character Reader

IBM 1428
Alphameric OptIcal
Reader

2.25 x 3.0 10 5.91 x 9.00
or Journal tape.

2.75 x 5.875 10 3.67 x 8.75

3.5 x 2.2510 8.75 x 4.25
400

No. documeDta/mln.

3200 lines/minute

420

Transport type

Conveyor belt

Vacuum drum/conveyor belt Vacuum drum/conveyor belt

Feed mechaDlsm

Vacuum

FricUon

Frlcbon

SorUag faclUtles

MulUstacker

Multistacker

Mul tlstacker

Max, characters/ltoe

85

80

80

Mu, 1I_/iDCh

6

10

10

Max, It_/pa.s

-

2

2

Max, readlac speed,
curacters/second

2000

500

500

Font styles read

IBM 1428, 1428E; Selfcheck IBM 407-1, 407E-l
7B, 12F, 12L; USABI;
NCR NOF; handprinled
3/16 Gothic

IBM 1428

Character set

Numeric, plus 5 letters

Numeric

Alphameric

ScaDDlag technique

Flying spot

Mechanical dISC

Mechanical disc

RecopiUon method

Curve tracing

Matrix matching

Matrix matchlag

FLEXIBIUTY

Formatting under control
of computer program;
selective fIelds

Reads selective f,elds;
format control by external
computer program

Reads selective fields
under control of computer
only

ERROR CONTROL

Rescan feature, character
dIsplay WIth manual
keyboard entry, marks
documents

Resean feature, character
display; kcyboard InS('rtlOn

Itescan feature. error
cht'cklng by external
comput(·r program

OUTPUT

Data to computer

Data to computer

Da ta to computer

OPERATING CONTROL

On-hne WIth IBM
System/360 computer

On-hne WIth IBM 1400
SerIes or System/:160
computer, may be us,'d
off-hne a& 13-pocket
sorter only

On-hne WIth IBM 1400
SerIes .... r System/3GO
computers

SPECIAL FEATURES

Reads mark-sense; resds
hand-prInted digIts 0-9
and 5 alphabetic control
symbols

Reads mark-s,'nse. reads
reverse lma~c

Reads mark-sense, reads
reverse Image

APPROXIMA TE PURCHASE PRICE'

$144,000

$120,000 to 142,000

$1!iO, 000

AVAlLABIUTY

l4 months

6 months

• months

FIRST DEUVERY

2nd quarter 1968

Oclober 1961

October 1962

INPUT
FORMAT

CHARACTER
READING

RECOGNITION

© 1967 AUERBACH Corporation and AljERBACH Info, Inc.

10 67

AUERBACH STANDARD EDP REPORTS

23:020.905

COMPARISON CHART: OPTICAL CHARACTER READERS (CONTD. )
NaUooal Cub Register
420-2 OpUcal Reader

OpUcalSoaDDlng Corp.
OpSoan 288 Character
Reader

Phllco-Ford Corp.
Model 6000
Print Reader

Document alze, Inches
(WIdth x length)

Joornal tape, 1.31 x 10
to 3.25 x 1200

3.5 x 2.5 to 8.5 x 4.5

3x5to8.5x14.0

No. documenta/mln.

3120 linea/minute

200 to 600

180 to 360

Transport type

J OIlrnal spools

Drive belt

Conveyor belt

Feed mechanism

Rollers

Friction and vacuum

Vacuum
01181 output stacker

IDENTITY

DOCUMENT
HANDUNG

INPUT
FORMAT

Sorting facUlties
Max. characters/line
Max. hnes/lnch

-

Dual output stackers
80 (machine-printed).
25 (hand-printed)
3 (machine-printed);
2 (hand-printed)
3

32
4

Max lines/Dass
CHARACTER
READING

90
6
78
1250

Max. reading speed.
characters/second

1664

800

Font styles read

NCR (NOF)

USABI, E-13B, 1428, 407E, Multifont
hand-printed (choice of one)

Character set

Numeric

Numeric plus C, N, S, T,

X, Z,

+, -

Alphameric, upper and
lower case

Scanrung technique

MechanICal disc

Photocells

Flying spot

Recognition method

Matrix matching

Matrix matching

Matrix matching

FLEXIBILITY

Formatting, editing, and
field-selection by plugboard program

Reads selective fields
programmed by plugboard;
reads intermixed fields

Selective fields; reads
intermixed fonts within
a document or batch;
extensive formatting
and editing features

ERROR CONTROL

Character display with
manual keyboard entry;
res can feature, marks
documents

Error character substituted Character display With
for unreadable characters,
manual keyboard entry;
no rea can feature
marks documents

OUTPUT

Data to computer; or
punched paper tape,
tab cards, or magnetic
tape devices

Magnetic tape, 7 or 9 track. Magnetlc tape. punched
cards or paper tape, or
556/800 bpI
data to computer

OPERATING CONTROL

On-line with NCR. IBM 1400 Off-hne
Series, or UNIVAC 9000
Series computers; or
off-line

Off-line

SPECIAL FEATURES

Header-line entry
(Note: Model 420-1, which
IS half a8 fast a8 the 420-2,
is now offered on an Has
returned" baSis at a price
of $60,000)

Handles stock from 20 to
100 pounds; reads handprinted characters

Reads mark-sense; reads
punched holes. Header
documents can be used
for format specificatiOns
to program

APPJWXIMATE PURCHASE PRICE

$80,000

$98,088

$450,000

AVAILABIUTY

30 days

9 months

12 months

FIHST DEUVERY

February 1966 (420-2)
November 1961 (420-1)

1st quarter 1968

May 1965

R ECOGNlTIO~

10 67

IA.
AUERBACH

(Contd. )

23:020.906

SPECIAL REPORT

COMPARISON CHART: OPTICAL CHARACTER READERS (CONTD. )
IDENTITY

DOCUMENT
HANDLING

INPUT
FORMAT

CHARACTER
!READING

Radio Corp. of America
Vid_canW
Document Reader

Recognition Equipment
Electronic RetilUl
Document Carrier

RecognitiOD Equipment
ElpctronlC ReUna
Rapid Index P1I&e Reader

Document size, Inches
(width x length)

2.5 x 4.0 to 2.5 x 8.5

3.25 x 3.25 to 4.25 x 8.50

3.~5

No. documents/min.

1800

I~OO

24

Transport type

Conveyor belt/drum

Conveyor belt

Vacuum drum

Feed mechanism

Vacuum

Vacuum

Vacuum

Sortlag facilities

Dual output stacker

Multistacker

Multlstacker

x 3.26 to 14.0 x 14.0

Max. charscters/line

80

90

ISO

Max. lineS/inch

6

8

8

Max. lines/pass

1

2

100

Max. reading speed,
characters/second

1500

2400

2400

Font styles read

RCA N-2

Multifont. handprinted

Multifont. handprinted

Character set

Numeric

AlphamerIC. upper and
lower case

Alphameric, upper and
lower case

JRECOGNlTlON Scanning technique

VIdicon scanner

Parallel photocells (Retina)

Parallel photocells (Retina)

Stroke analysis

Matrix matching

Matrix matching

FLEXIBILITY

Limited field selectivity
under control of external
computer program

Selective fields; extensive
editing and formatting
features under control of
internal program

Selective fields; extensive
editing and formatting
features under control of
Internal program

ERROR CONTROL

Error character substituted
for unreadable characters,
reject stacker

Programmable actions.
rescan features, sorts
documents mto error
stacker

Data to computer

Any peripheral deVice

Recogrution method

OUTPUT

OPERATING CONTROL

On-line with RCA
Spectra 70 computers

Progra mmable actIOns,
res can features;
marks documents

using magnetic tape,

Any peripheral deVICe
us lni-\ magnetic tape,

punched ca rds, or
paper tape

punched cards, or

Off-line; self-contaIned

Off-hne; self-contaIned

softwa~e

software

paper tape

SPECIAL FEATURES

Reads mark-seIlS";
reads holes

Reads mark-sense,
reads ba r -code;
accumulates totals

Reads mark-sense;
reads bsr-code;
accumulates totals

APPROXIMATE PURCHASE PRICE

$126,900

$550,000

$,;'0.000

AVAILABILITY

24 months

6 to 12 months

6 to I.! months

FIRST DEUVERY

4th quarter 1966

December 1964

November 1964

© 1967 AUERBACH CorporatIon and AUERBACH Info. Inc,

10 67

AUERBACH STANDARD EDP REPORTS

23:020.907

COMPARISON CHART: MAGNETIC CHARACTER READERS
IDENTITY

DOCUMENT
HANDLING

Document lize, iIIohee
(wlddl x leqth)

a_rill £111O&11c Co.
MRB200
Doownent Reader

2.69 x 5.94 t04.06 x 9.06

2.5 x 5.25 to 4. I x 9. 0

2.5 x 4.14 to 5.5 xB.75

No. dooumelltl/mln.

1.565

1.200

600

Trllll8port type

Conveyor belt

Conveyor belt

Conveyor belt and roller

Feed mechanllm

Friction

Vacuum

Friction

SortlJll facUltIel

13 output stackers

12 output stackers

11 output stackers

Max. characters/line

59

64

66

Max. lines/pass

I

1

1

1.300

1.800

1200

Font styles read

E-13B

E-13B
COC-5

E-13B. CMC-7

Character set

Numerals. four
cOl1trol symbols

Numerals. four
control symbols

Numeric + 4 symbols

Magnetic

Magnetic

Analog waveform

Analog waveform matching

Analog waveform matching

Matrix matching

IsPIIU fields using

Early character reader
turnoff; reads COC-5
intermixed wltb E-13B

Split fields using
control symbol.

Validity check

ValidIty check; tIming
check.

Sorted Input documents
land/or data directly to
omputer. if used on-line

Sorted Input documents
and/ or data directly to
computer, if used on-line

Data to computer

Off-lme, or on-hne with
BIOO/B200/B300 and
B2500/B3500 computers

Off-line. or on-lme witb
G E-400 Series computers

On-hne WIth System
360/20, 360/30, or
360/40; or off-hne for
sorting only.

SPECIAL FEATURES

Automatic code translation;
handles Intermixed sizes.
16-pocket version available

Automatic code translation;
handles Intermixed sizes;
endorSing feature; reads
COC-5 optically; TCD
verification; missing digit
detection; multiple digit
selection

Automatic code translation.
handles intermixed sizes.

APPROXIMATE PURCHASE PRICE

$91.200

$80,000

AVAILABILITY

12 months

6 montbs

$49.500 (Mod. 1)
$63,000 (Mod. 2)
9 months

March. 1962

July, 1968

INPUT
FORMAT

CHARACTER Max. reading speed,
READING
characters/second

RECOGNITION Scanning technique
Recognition metbod
FLEXIBILITY

~ontrol symbol

[RROR CONTROL

Ol'TPl'T

OP[RATJ~G

COl'TROL

I

ValIdity check, sIgnal
level amplIfication check

FIRST DELIVERY

10/67

IBM 1259 Magnetic
Character Reader

Burroucba Bloa/IOS
Sorter-Reader.

fA

AUERBACH

'"

(Contd, )

SPECIAL. REPORT

21:020.90a
COMPARISON CHART: MAGNETIC INK CHARACTER READERS (CONTD. )

IDENTITY

DOCUMENT
HANDLING

INPUT
FORMAT

CHARACTER
RFADING

IBM lne . . . .tkl

NCR .oa MICR

CharlCter Reader

CJwoICter Reader

Sorter-Reader

2.7& x 6.0 to 3.67 x 8.75

2.76 x

IBM Ina Mapetkl

Docwnnt size, Inob••
(w1cl&ll x l.ngth)

e. 0 to 3.78 x 8.75

2. Ii x Ii. 21i to '.Ii x 10.0

No. dooum.nta/mln,
Transport type

1600

1,515

750

Conveyor belt

Conveyor belt

Conveyor belt

Feed mechanllm

Friction

Friction

Friction

Sorting f.cllities

13 output stickers

13 output stackers

12 output stackers

Max. character./llne

66

40

56

Max, linea/pass

1

1

1

Max. reading speed,
obaracters/second

2.112

2,112

1,200

Font styles read

E-13B. CMC-7

E-13B

E-13B

Numerals, four

Numerala, four control
symbols

Numerals, four control
symbols

Magnetic

Magnetic

Character set

control symbols

RECOGNlTIm Scanning technique
Recognition method

Magnetic
Matrix matcbing

Matrix matcblng

Analog waveform matcblng

FLEXIBILITY

Spllta fields us Ing oontrol
symbol

Spllta fields using
control symbol

Spllta fields using
control symbol

ERROR CONTROL

Validity cbeck;
timing check.

Validity check.
timing check

Validity cbeck;
tim Ing cbec k

OUTPUT

Data directly to computer

Sorted InPut documents
and/ or data directly to
computer, If used on-line

Sorted Input documents
and/or data directly to
computer. If used on-line

OPERATING CONTROL

On-line. With IBM 360/30.
40. 50. and 65 computers;
or off-line [or sorting only.

Off-line or on-line with
IBM Sy8tem 360/30 or
360/40 computers

Off-line, or on-line with
NCR 315. 315-100, or
315 RMC oomputers

SPECIAL FEATURES

Automatic code translation;
bandle. Interm Ixed al zes

Automatic code tranalation;
bandl. Intermixed alzes

Automatic code translstion,
handles intermixed Sizes

APPROXIMATE PURCHASE PRICE

$110.500

$110,500

$45,000

AV Al LAB ILlTY

9 months

a

Available on an ai-returned
bull.

FIRST DELIVERY

October. 1962

October, 1982

months

CI 1967 AUERBACH Corporation and AUERBACH Info, Inc.

April, 1962

10/67

AUERBACH STANDARD EDP REPORTS

23:020.909

COMPARISON CHART: MAGNETIC INK CHARACTER READERS (CONTD. )
IDENTITY

DOCUMENT
HANDLING

NPUT
IFORMAT

~HARACTER

~EADING

NCR 404 MICR

NCR 407 MICR

Sorter-Reader

Sorter-Reader

Document size, Inches
(width x length)

2. 5 x 5.8 to 3.85 x 8.75

2.75 x 4.0 10 4.5 x 8.75

No. documents/min.

600

1,200

Transport type

Rollers

Conveyor Belt

Feed mechanism

Friction

Vacuum

Sorting facilities

11 output stackers

18 output stackers

Max. characters/line

65

56

Max. lines/pass

1

1

Max. reading speed,
characters / second
Font styles read

1,200

2,400

E-13B

E-13B

Numerals, four control
symbols

Numerals, four control

Magnetic

Magnetic

Character set

RECOGNITION Scanning technlgue
Recognition method

symbols

Matrix matching

Analog waveform matching

FLEXIBILITY

Splits fields us Ing control
symbol

Splits fields using control
symbol

ERROR CONTROL

Validity check, timing check Validity check, timing check

PUTPUT

Sorted input documents
and/or data directly to
computer, If used on-line

Sorted input documents
and/ or data directly 10
computer, If used on-line

OPERATING CONTROL

Off-line, or on-line with
NCR 315 or 315-100
computers

Off-line, or on-line with
NCR 315,315-100, or
315 RMC computers

SPECIAL FEATURES

Automatic code translation;
handles intermixed sizes

Pocket pullout; plugboard
programmable off-line;
batch number advance In
endorser; endorser; field
validatIOn; valid amount.
transact Ion Lode stop;

automat Ie t:ode translation
APPROXIMATE PURCHASE PRICr
AVAILABI LITY
FIRST DELIVERY

10 67

-

$29,900

$95,000

6 months

6 monthe

October 1966

January 1966

A

AUERBACH
~

23:030.001
SPECIAL REPORT
DECISION TABLES

SPECIAL REPORT
DECISION TABLES:
A STATE·OF· THE·ART REPORT
by
the Technical Staff of
AUERBACH Info, Inc.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

1/69

(

\

23:030.100

A•

.... 1IIACIt

snlUID

SPECIAL REPORT
DECISION TABLES

EDP
u'.ln

DECISION TABLES-THEIR GENERAL CONSTRUCTION
AND ACCEPTANCE IN PROGRAMMING
In 1957, dl'l'ISHJn tabLL's W('!"t, applied to probkm formulation for computer programming for
thl' first tl111('. Since thl'n, much progress has been mack in defming the concept for computer
applications. This includes refinement of methods and procedures for constructing and
checking lkcision tabll's for completeness and accuracy. Although limited, some techniques
have been llnplemented for converting decision tables to ('omputer programs.
Though the eXisting literature almost unanimously descnbes the advantages and significant
value of dl'('ision tables, they have not receivcd Wide acceptance. Yet, during the samc time
period, FOHTHAN, COBOL, and many other problem oriented programming languages have
been del doped. \\ldclvaceepted, and used.
Numerous reasons seem to account for this. One is that pr:ogrammers and system analysts are
taught nOl\ eharting as the conventional method uscd in formulating a logical solution to a
probkm, and deCision tables competc with this. A second reason is that decision tablcs are
not well suited for selCntific or mathematic applications which eomprisc a large portion of
computer usage. Finally, it is seldom that a computer application can be formulated entirely
in terms of decision tables.
Thus. the deelslOn table technique must be supplemented with a compatible formulation technique to enable computer applicatIOn, at the present time, no such technique has been developed.
l\Ian\' of the advantages of decision tables coincide with current problem areas in the use of computen; for business applications, such as formulatIOn and documentation. Decision tables
prol"ide an effective means of communication between people in and out of the data processing
field. III defimng both problems and their logical solutions. They provide a concise and compact
form of defmition and description suitable for usc in analysis, programming, and documentatIOn. The extent and naturc of the changes required to update or revise an applications program
is prol'ided by thClr umque form of problem statement.
Howe\,er, the state-of-the-art has not advanced sufficiently to enable economic realization of
these advantages. Though some parts of a problem may be suitable for decision tables, no
technique has been devc!oped which includes decision tahles as a part of the formulatIOn of the
entire problem. Decision table languages developed for use in programming are too restrictive,
thus losmg many of the initial decision table advantages, or they produce object programs which
are not economical in core storage requirements and running time.
Problem oriented languages, such as FDRTRAN and COBOL, are more natural to the programmer
than machine oriented languages, and thus make applications programming machine independent.
Now, a language is needed that will assist in stating the application logically and help the programmer obtain optimum logical flow in the solution. A possible method is the usc of decision
tables for problem definition and formulatlOn.
What Is A Decision Table?
Basically, a decision table is a tabular representation of data arranged in a particular form to
assist in making decisions. The two categories used are "action" and "condition". The form
must takc mto consideration the rules specifying each action, and the specific value of each
condition.
DecisJOn tables are conventionally represented by a rectangle divided into four quadrants,
separated by double lines, as shown in Figure 1. These areas are known as "conditions stub",
"conditions entry", "action stub", and "action entry". Each contains a listing of all appropriate
information relating to the problem. The horizontal items are called rows; the vertical items
are called rules. This is illustrated in Figure 2.
A row is a condition statement which may be Y (yes), N (no), don't care, or an extension of the
conditlOn, as in Figure 3. If a row contains only Y, N, and don't care, it is called a "limitedentry" row; otherwise, it is an "extended-entry" row.
A rule IS a set of condition values and the associated set of actions. If all condition values are
specifically stated. it is a "simple" rule; otherwise (when don't care values are included), it is a
"complex" rule. Conventionally. rules are numbered across the top from left to right and are
tested in that order.
When all sets of condition values satisfy two rules, and actions are identical in both, the rules
are redundant; if the actions are different, the rules are contradictory.
The combinations of rules determine whether a decision is "perfect" or "imperfect". If all
possible combinations of condition values are covered by the rules of the table, it is "perfect";
© 1968 AUERBACH Corporation and AUERBACH Info, Inc,

12/68

23:030.101

AUERBACH STANDARD EDP REPORTS

Statement
CONDITION
STUB

CONDITION
ENTRY

Rulel

Rule 2

Rule 3

Condition Row 1
Condition Row 2
Condition Row 3
Condition Row 4

ACTION
STUB

ACTION
ENTRY

Action Row 1
Action Row 2
Action Row 3
Figure 2: Items of a Decision Table

Figure 1: Quadrants of a Decision Table
Age

<25

25

>25

Salary range

Below $10K

$10 - $20K

Over $20K

Bell is

Alarm

Phone

Door

Code equals

A

B

C

Go to

Work

Conference

Hideout

Figure 3: Examples of extended entry rows
otherwise, it is "imperfect". In an "imperfect" decision table, the last rule must be null (all
"don't care" values). This specifies the action taken when the condition values do not satisfy
any other rule.
Therefore, deCision tables are identified by these three characteristics:
1.

rows
(ho rizontal)

a. ) limited entry
b.) extended entry
c.) mixed entry

2. rules
(vertical)

a. ) Simple
b.) complex

3. complement
of rules

a. ) perfect
b.) imperfect

Though many combinations of the above are possible, the characteristics usually are mixed
entry, complex, imperfect; when a decision table is referred to, this is the norm. OtherWise,
the deviations are speCified (e.g., limited-entry decision table).
A new concept introduced in this paper is that of a generic rule, which covers the speCific
values in a "complex rule". As noted above, a "complex" rule is one where any "don't care"
values appear. In this case, the remaining speCific values form the generic rule. In Figure
9, the generic rule consists of the last two condition values of rule 1, Y, N. Figure 10 illustrates this more completely by showing rules la through ld as the complete family of rules
for the generic rule 1 of Figure 9.
Decision Table Construction
The most basic decision table is limited - entry, simple, perfect, such as Figure 4. Ordering
of rows, rules, and actions are completely irrelevant to the proper solution of the problem.
Construction of such a table is simply an orderly listing of all relevant conditions, actions and
rules. Any decision table can be expanded into this type, and conversely, the smallest decision
table formulating a problem can be derived from a limited-entry. simple. perfect table. However.
a limited -entry, simple, perfect decision table containing N conditions must have 2N rules.
Usually, this is too many for practical workability.
Therefore, the table must be broken down; two methods of doing this are proposed. The first
is to divide the table into a number of smaller, interrelated tables. In addition to being easy
to do, this method retains most of the advantages of working with a limited-entry, simple,
perfect decision table. However, the work load is increased slightly, and the entire problem
cannot be seen at one time. A more sophisticated approach to this method is parsing the
Original table; the success of this depends on the skill of the analyst.

12/68

A

(Contd. )

AUERBACH

'"

SPECIAL REPORT

23:030.102

1

2

3

4

5

6

7

Is phone ringing

y

y

y

y

N

N

N

N

Is alarm clock ringing

y

y

N

N

Y

Y

N

Is door bell ringing

y

N

y

N

y

N

N
y

Turn off alarm clock

X

X

X

X

Answer Phone

X

X

X

X

Answer Door

X

X

X

8

N

X

Ignore

X

Figure 4: Example of a limited-entry simple perfect decision table

The second solution is to construct a higher-level type table, which makes the learning of
decision table techniques almost equivalent to those of flowcharting. Factors requiring constant
attention must be kept in mind when constructing such a sophisticated type table. For example,
Figure 5 shows a mixed-entry, simple, imperfect decision table. The values of the second
condition are not mutually exclus ive, and a rule covering combinations does not exist. Therefore, it docs not contain a perfcct rule set.
The most effective way to reducc the number of rules is to reduce the number of stated conditions and usc extended-entry rows. However, in re-wording statements, care must be taken
not to introduce any meaningful change in the scope of the problem. In using extended-entry
rows, conditlOn values must be mutually exclusive, or an action for reentering the table must
be included. Figures 6, 7 and 8 illustrate these points.
This can result in significant savings in the number of rules. For example, a decision table
with N extended-entry conditions and three values for each condition contains 3N rules. If each
pair of limited-entry conditions can be combined into a single extended-entry condition, a table
with eight conditions requiring 256 rules could be reduced to a table with four conditions and 81
rules.
In general. the number of rules contained in an extended-entry, simple, perfect decision table is
the product of all !\iN values, where !\i is the number of possible condition values, and N is the
number of condition statements having exactly !\i condition values.
Another method of redUCing rules, and consequently the size of the decision table, is to state
complex rules. However, thIS is difficult to implement, since the order in which the rules are
stated is of significant importance.
To combine rules into a complex rule, it is necessary that their action sets be identical. If
two or more rules have identical action sets, they can be combined into a single complex rule
by substituting a "don't care" value for all Y and N condition values.
Though the above applies to limited-entry tables, a complex rule can be created for an
extended-entry table in the same manncr if the rules bemg combined have a generic rule and
do not include all pcssible values for thc remaining conditions. Figures 9 through 12 illustrate
thIS method.
The final way to reduce the rules contained in a decision table is to construct an imperfect
decision table, by establishing the extreme right rule as a null rule, (sometimes referred to as
the else rule). All condition values of the null rule are "don't care", and the actions specified
are those to be taken when no previous rule is satisfied. This poses no problem in construction, except for ascertaining that the preceding rules cover all sets of conditions for which the
action set differs from that specified in the null rule.
In summary, the construction of proper decision table requires a complete set of conditions
and actions be known or developed during the construction process. The construction proccss
is simple with the significant exception of the use and placement of complex rules. The
sequence in which conditions are tested and actlOns taken within a rule is not a property of
any type of decision table. However, the sequence in which rules are stated is important for
complex decision tables. Finally, successful construction and interpretation of most decision
tables hinges on the understanding of the concept of a generic rule, as related to the interpretations of "don't care" values in complex rules.

© 1968 AUERBACH CorporatIOn and AUERBACH Info, Inc.

12/68

23:030.103

AUERBACH STANDARD EDP REPORTS

2

1

Is Bell ringing

Y

Y

Bell is at least

Alarm

Phone

Turn off alarm

X

Answer phone

4

3

Y

-

Door

X

Answer door

X

Ignore

X

Reenter table

X

X

X

Go back to sleep

X

Figurc 5: Example of a mixed-entry simple imperfect decision table

5

1

2

3

4

Salary dOK

Y

Y

Y

Y

Y

Y

Salary 2-10K

Y

Y

Y

Y

N

N

Yrs. of Service <5

Y

Y

N

N

Y

Y

Yrs. of Service 2-5

Y

N

Y

N

Y

N

X

X

Record in File A

X

Record in File B
Record in Error File X

X

9

10

11

12

Y

Y

N

N

N

N

N

Y

Y

Y

N

N

Y

Y

Y

N

Y

N

X

X

X

X
X

X

13

14

15

N

N

N

N

N

Y

N

N

N

N

N

N

Y

Y

N

N

Y

N

Y

N

Y

N

X

X

X

X

X

X

8

7

6

X

16

X

X

X

X

X

X

X

X

Figure 6: Limited-entry decision table of hypothetical problem

1

2

3

4

5

6

7

8

9

Salary (dO, 2-10, 1)

<10

<10

<10

;dO

::0:10

::0:10

I

I

I

Yrs. of Service «5, 2- 5, I)

<5

;;:,;5

I

<5

::0:5

I

<5

X

X

X

X
X

X

Record in File A
Record in File B

X

Record in Error File

(I

;;:,;5 I

'=

Indeterminent)

X

X
X

X

X
X

X

X

Figure 7: Extended-entry decision table of hypothetical problem

Salary <10K or Service <5 yrs.

1

2

3

4

5

6

7

8

Y

Y

Y

Y

N

N

N

N

N

N

Y

N

Salary ;;:,;10K or Service ;;:,;5 yrs.

Y

Y

N

N

Y

Y

Salary or Service Indeterminate

Y

N

Y

N

Y

N

Record in File A

X

X

X

X

Record in File B

X

X

X

X

Record in Error File

X

X

X

X

Impossible

X

Figure 8: Decision table of hypothetical problem with conditions restated

12/68

A

AUERBACH

(Contd. )

(JI

1

2

3

4

5

6

7

Y

Y N

N

-

Y

Salary ;" 10K

-

\,()

Yrs of Service < 5

Y

Y

O'l
00

Yrs of Service;" 5

N N

Y

N Y Y

N Y - - - - -

Record in File A

X

X

X

X

X

X

Salary < 10K
@

N N

Y Y

N Y

N N

9

8

N

1

2

3

4

5

Salary < 10K

Y

Y

N

N

Salary :;, 10K

N N

Y

Y

- - -

6

Yrs of Service < 5

Y

N Y

Yrs of Service :;,5

N

Y

N Y N Y

Record in File A

X

X

X

X

X

7

en

X

::0

co

J>

o

I

oo

.;:;

Record in File B
Record in Error File

X

X

X
X

X
X

Record in File B

X

X

X
X

-

FIGURE 9:

;u
[Tl

~

;U

-I

9

X
X

X

»
r

N Y - - - - -

N Y N

Record in Error File

8

[Tl

()

Y N

J>
C

-0

CORRECT:
The rules are correctly ordered in FIG. 11;
consequently. the expansion of rule 5. mcluding
the remaining values of condition rows 1 and 2
WIth respect to the generic rule. results III
FIG. 12 whIch Illustrates a proper creation of
a complex rule for an extended-entry decision
table.

INCORRECT:
Rule 1 is mconectly placed m FIG. 9.
Its expansIOn In FIG. 10 lead;; to a contl'adlctJon
in the decIsIOn table.

X

X

X

-

X __ X
L

--

FIGURE 11:

o

~

6

::J

'"

la

Ib

lc

Id

2

3

4

5

6

7

Salary < 10K

Y

Salary ;" 10K

Y

Y

N

N

Y

Y N

N

Y

N

N

N Y Y

-

Y N

N

N Y

Yrs of Service < 5

Y

Y

Y

Y

Y

N

Y N

N

Yrs of Service ;,,5

N

N

N

N

N

Y N

Record in File A

X

X

X

X

X

X

X

X

X

X

X

X

X

::J
Q.

J>
C

en
::0
co
J>

o

I

::J

0'
::J

Record in File B

8

9

1

2

Salary <10K

Y

Y

Salary :;,10K

N N

Y N

5b

6

7

N N

Y

N

Y

N

-

Y

Y Y

Y

Y

N

N

N

Y

X

X

X

X

- - - - -

Yrs of Service < 5

Y

Yrs of Service;" 5

N Y N

X

Record in File A

X

Y Y

X

XX

Record in File B

N

4

5a

3

X

X

X

X

Y

9

- - - - N

N Y

X
X

X

8

X

(")

Record in Error File

X
.

FIGURE 10:

X

X

-

-

X

Record in Error File

X

X

X

X

FIGURE 12:

N
W

0

W

-N

a-.

0>

0

r~

23:030.105

AUERBACH STANDARD EDP REPORTS

The Use of Decision Tables in Programming
There are three primary problems relating to the current use of decision tables in programming.
The first is the problem of training analysts and programmers to construct and interpret all
typcs of decision tables. The second is the problem of developing a formulation technique which
includes decision tables, and permits formulation of that part of the problem which is inappropriate for formulation by decision tables. The third is that of converting a decision table into
a computer program.
Although much work has been done on converting decision tables into computer programs, this
problem is not yet solved and is still a major obstacle. Manual conversion poses the least
constraints on the decision table construction, but loses all of the advantages attendant to highorder languages such as COBOL and FORTRAN. Compiler pre-processors which accept decision tables, such as the DETAB/65 pre-processor for COBOL, have been developed, as have
decision table compilers, such as TABSOL. However, both of these methods of conversion
tend to place such restrictions on decision tables that most of their advantages and flexibility
are lost.
A further complication is the fact that while sequencing of condition testing and actions taken
is immaterial to the decision table, sequencing can have a significant influence on the efficiency
of the program generated. This is particularly true in terms of running time. Thus, special
attention must be paid to this point, either during the decision table construction phase, during
the conversion phase, or both. OtherWise, the resulting program may be significantly less
efficient than a program generated by another technique, and therefore more costly in the end.
Pending a solution of the conversion problem, very little work has been done on the first two
problems. This appears to be primarily due to the fact that a solution to the conversion problem
is necessary for the effective use of decision tables in programming and thus it has overshadowed all other problems.
Conclusions
The construction of a decision table per se is not as difficult as some analysts and programmers
claim. The significant difficulty is the concern for efficient object programs, and the requirements of existing decision table pre-processors and compilers. Also, a supplementary formulation technique must be employed in order to develop the problem at hand. Therefore, many
analysts and programmers are not motivated to learn the decision table technique because of the
limited number of instances in which it presents them with a significant advantage.
The communication of a problem and its solution is of significant concern within the computer
industry currently. Here the advantage of decision tables over existing techniques is very
significant. English narratives and program flow-charts of complex problems are difficult to
prepare, and easy to miSinterpret; their length often precludes easy and complete crossreferenCing. Both reflect the capabilities of their creator, and neither can incorporate
changes easily. Decision tables virtually eliminate these problems. But this is more of an
advantage to management than to the analyst or programmer. Therefore, the use of decision
tables to gain this advantage must be imposed by management; then, management must pay the
currently high conversion and operating costs.
It is expected that work toward solving the problems relative to using decision tables in programming will continue. However, decision tables will not enjoy wide acceptance until a
significant breakthrough occurs in the area of automatically converting decision tables into
efficient object programs. That is, until analysts, programmers, or managers can obtain a
consistent and general net advantage from the use of decision tables as a problem formulation
technique.

12/68

(1)

Armerding, G. W., "FORTAB: A Decision Table Language for Scientific Computing
Applications", Proc. Decision Tables SympoSium, (Sept. 20, 1962, New York) pp. 81-87.
Issued also as: Rand Corp. No. RM-3306-PR (Sept. 1962) 39pp.

(2)

Boerdam, W., ''Decision Tables in System Design", unpublished paper, Atlantic
Richfield Co., Los Angeles, Calif., (no date) 9 pp.

(3)

Bromberg, H. "COBOL and Compatibility", Datamation Vol. 7, No.2, (Feb. 1961),
pp. 30-34.

(4)

Brown, Lynn M. "Decision table experience on a file maintenance system", Proc.
Decision Tables SympOSium. (Sept. 20, 1962, New York), pp. 75-80.
--

(5)

Calkins, L. W., "Place of Decision Tables and DETAB-X", Proc. Decision Tables
Symposium, (Sept. 20, 1962, New York) pp. 9-12.

(6)

Canning, Richard G., "Decision Structure Tables", EDP Analyzer, Vol. I, No.4,
(May 1963).

(7)

Canning, Richard G., "How to Use Decision Tables", EDP Analyzer, Vol. 4, No.5,
(May 1966).

fA
AUERBACH

'"

(Contd. )

SPECIAL REPORT

23:030.106

(8)

Cantrell, H. N., "Commercial and Enginccring Applications of Decision Tables",
Proc. Decision Tables Symposium, (Sept. 20, 1962, New York) pp. 55-6l.

(9)

Cantrell, H.N., King, J., and King, F.E.II. (1961). "Lof,ricStructureTables",
Conuu. ACM, Vol. 4, No.6, (June 19(1), pp. 272-5.

(10)

Chapin, Ned. "A guide to deciSion table utilization".
(Oct. 1966, Los Angelcs).

Proc. 1966 Fall DPMA Conf.

(11)

Chapin, Ncd. "Parsing of Decision Tablcs", Comm. ACM, Vol. 10, No.8, (Aug. 19(7),
pp. 507-510, 512.

(12)

Chapin, Ned. "A guide to decision table utilization", In Data Processing Vol. IX, DPMA,
Parik Ridge m., 1967, pp. 327-329.

(13)

Chapin, Ned. "An Introduction To Dccision Tables", DPMA Quartcrly, (April 19(7),
pp. 3-23.

(14)

Chapman, A. E., and Callahan, M. D., "A description of thc basic algorithm used in the
DETAB/65 preproccssor", Systcms Developmcnt Corporation. SP-2534/000/00
(July 7, 19(6), 18pp. (Also Comm. ACM, Vol. 10, No.7, (July 19(7), pp. 441-446.)

(15)

Cunningham, J., "Decision Tables Symposium", Proc. Decision Tables Symposium,
(Sept. 20, 1962, New York) pp. 7-S.

(16)

Devine, Donald, "Decision Tables as the Basis of a Programming Language", Data
Processing, Vol. 7, Data Proccssing Management Association, Park Ridge, ill.: 1965,
pp. 461-466.

(17)

Dixon, P., ''Decision Tables and thcir Application", Computers & Automation, Vol. 13,
No.4, (April 19(4), p. 14.

(IS)

Egler, J. F., (1963). "A Procedure for Converting Logic Table Conditions into an
Efficient Sequence of Tcst Instructions", Comm. ACM, Vol. 6, No.9 (Sept. 1963),
pp. 510-514.

(19)

Evans, O. Y., "An Advanced Analysis Method for Integrated Electronic Data Processing",
paper written in 1959 and published first by the National Machine Accountants Assoc. of
Long Beach, Calif., in March 1960. A condensed version was issued in 1960 as: "IBM
General Information Manual, F20-S047"; and a sequel issued in Sept. 1961 as: "IBM
Ref. No.1 J 1".

(20)

Evans, O. Y., "Decision Tables. A Preliminary Reference Manual". Systems Engineering Services Clearinghouse Report, Ref. No. 1 J 1 (Sept. 19(1), a sequel to "IBM
General Information Manual, F20-S047". (See preceding reference.)

(21)

Evans, O. Y., "GE 225 TABSOL Manual (Preliminary)", General Electric Computer
Dept., Arizona No. CPB-147 (5M 3-(1).

(22)

Fergus, Raymond M., "Decision Tables -- An Application Analyst/Programmer View",
Data Processing, Vol. 12, Data Processing Management Association, Park Ridge, ill.,
1967.

(23)

Fergus, Raymond M., "An Introduction to Decision Tables", Systems and Procedure
Journal, (July-August 1965), p. 24.

(24)

Fergus, Raymond M., "Good Decision Tables and Their Use", Systems and Procedure
Journal, (September-October 19(8), pp. lS-2l.

(25)

Fife, Robert C., "Decision Tables" Univac Application Report (April 19(5), 33 pp.

(26)

Fisher, D. L., (1966). "Data Documentation and Decision Tables", Comm. ACM,
Vol. 9, No.1, (Jan. 1966), pp. 26-3l.

(27)

General Electric Co., "TABSOL Manual", General Electric Co. CPB-147 (1961)
16 pp.

(2S)

Glans, T. B., and Grad, B., "Tabular description language" IBM Tech. Rep. No. 2A5,
(Jan. 19(2).

(29)

Grad, Burton (1961). Tabular Form in Decision Logic", Datamation VoL 7, No.7,
(July 1961), pp. 22-26.

(30)

Grad, Burton, "Structure and Concept of Decision Tables", Proc. Decision Tables
Symposium, (Sept. 20, 1962, New York), pp. 19-2S.

(31)

Grad, Burton, "Using Decision Tables for Product Design Engineering", a paper
prepared for 1962 AlEE Winter General Meeting, NYC Feb. 2, 1962 (CP 62-37S).

(32)

Grad, Burton, ''Decision Tables in Systems Design", Dig. Tech. Papers, ACM Nat'l
Conf., (Sept. 4-7, 1962, Syracuse, N. Y.) pp. 76-77.

(33)

Grad, Burton, "Engineering Data Processing Using Decision Tables", Data Processing,
Vol. 7, Data Processing Management ASSOCiation, Park Ridge, ill., 1965, pp. 467-476.
© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

12/68

23:030.107

AUERBACH STANDARD EDP REPORTS

(34)

Grindley, C. B. B., (1966). "Systematics -- A non-programming language for designing
and specifying commercial systems for computers", The Computer Journal, Vol. 9,
p: 124.

(35)' Hawes, Mary K., 'Decision Table Tutorial Using DETAB-X", developed by Instruction
Task Force of the CODASYL Systems Development Group for the Decision Tables
Symposium of Sept. 20-1, 1962.

12/68

(36)

Hawes, Mary K., "The Need for Precise Problem Definition". Proc. Decision Tables
Symposium, (Scpt. 20, 1962, New York), pp. 13-18.

(37)

Hawes, Mary K. , "Thc Use of Decision Tables for Problem Specification", Univac
Application Report (April 1965).

(38)

Holstein, D., "Decision Tables. A Technique for Minimizing Routine Repetitive
Design", Machine Design, Vol. 34, No. 18, (Aug. 2, 1962), pp. 76-79.

(39)

IBM General Information Manual, "Decision Tables: A Systems Analysis and Documentation Technique", Form Number F20-8102.

(40)

Kavanagh, T. F., and Allcn, M., "The Use of Decision Tables", Proc. of 1963, Conf.
of International Data Processing Management Assn. (Data Processing VI), p. 318.

(41)

Kavanagh, T. F., "TABSOL - A fundamental concept for systems-oriented languages",
Proc. Eastern Joint Computer Conference (Dec. 13-15, 1960, New York), pp. 117-136.

(42)

Kavanagh, T. F., "TABSOL - The Language of decision making" Computers and
Automation, Vol. 10, No.9, (Sept. 1961), pp. 15, 18-22. (A condensation of the previous
paper.) (This is a shortened version of Cantrell's "Logic Structure Tables".)

(43)

Kavanagh, T. F., "Manufacturing Applications of Decision Structure Tables", Proc.
Decision Tables SympOSium, (Sept. 20, 1962, New York), pp. 89-97.

(44)

King, J.E., "LOGTAB: a logic table technique", General Electric Co. Report (no date),
23 pp.

(45)

King, P. J. H., "Conversion of Decision Tables to Computer Programs by Rule Mask
Techniques", Comm. ACM, Vol. 9, No. 11, (Nov. 1966), pp. 796-801.

(46)

King, P.J.H., "Some comments on Systematics", The Computer Journal, Vol. 10,
No.1, (May 1967), pp. 116.

(47)

King, P.J.H., 'Decision Tables", The Computer Journal, Vol. 10, No.9, (August 1967),
pp. 135-142.

(48)

Kirk, H. W. (1965). "Use of Decision Tables in Computer Programming Comm. ACM,
Vol. 8, No.1, (January 1965), pp. 41-43.

(49)

Kramer, F.R. and Kirk, G.J., 'Decision Table Techniques in Computer Control",
IEEE Trans. Power Apparatus & Systems, (May 1966), pp. 495-498.

(50)

Larsen, R. P. (1966). "Data Filtering Applied to Information Storage and Retrieval
Applications", Comm. ACM, Vol. 9, p. 785.

(51)

Ludwig, H.R., "Simulation With Decision Tables", Journal of Data Management, Vol. 6,
(January 1968), pp. 20-27.

(52)

Meyer, H. J., "Decision Tables as an Extension to Programming Languages", Data
Processing, Vol. 7, Data Processing Management Association, Park Ridge, TIL, (1965),
pp. 477-484.

(53)

Montalbano, Michael, "Tables, Flowcharts and Program Logic", IBM Systems Journal,
(Sept. 1962), pp. 51-63.

(54)

Montalbano, Michael, "Letter to Editor (Egler's procedure refuted)", Comm. ACM,
Vol. 7, No.1, (January 1964), p. 1.

(55)

Morgan, J.J., 'Decision tables", Management Services (January-February 1965),
pp. 13-18.

(56)

Naramore, Frederick. "Applications of deCision tables to management information
systems", Proc. Decision Tables Symposium, (Sept. 20, 1962, New York), pp. 63-74.

(57)

Nickerson, R. C •• "An Engineering Application of Logic structure Tables", Comm.
ACM, Vol. 4, No. 11, (Nov. 1961), pp. 516-520.
---

(58)

Phillips, C. A., "Current Status of COBOL", Proc. USA and World Wide Data Systems
and statistics Conf. (Oct. 26, 1961).

(59)

Pollack, S. L., "Conversion of Limited Entry Decision Tables to Computer Programs",
Comm. ACM, Vol. 8, No. 11, (Nov. 1965), pp. 677-682.

A

(Contd. )

AUERBACH
~

SPECIAL REPORT

23:030.108

(60)

Pollack, S. L., "DETAB-X: An Improved Business-OIiented Computer Language",
Mem. RM-3273-PR, Rand Corp., Santa Monica, Aug. 1962.

(61)

Pollack, S. L., and Wright, K.R., ''Data Description for DETAB-X", Mem.
RM-3010-PR, Rand Corp., Santa Monica, March 1962.

(62)

Pollack, S. L., "Analysis of the Decision Ru1es in Decision Tables", Mem. RM-3669-PR,
Rand Corp., Santa Monica, May 1963, 69 pp.

(63)

Pollack, S. L., "What is DETAB-X?", Proc. Decision Tables Symposium, (Sept. 20,
1962, New York), pp. 29-39.

(64)

Pollack, S. L. and Grad, B., ''DETAB-X, Preliminary Spccifications for a Decision
Table st ructured Language", Data Desc Iiption and Transformation Logic Task Forces
of the CODASYL Systems Group, Sept. 1962.

(65)

Pollack, S. L., "Question and Answer Period. . . 9/20/62". Proc. Decision Table
Symposium, (Sept. 20, 1962, New York), pp. 9-12.

(66)

Pollack, S. L., "TABSOL Application Manual, Introduction to TABSOL", General
Electric Computer Dept., AIizona. No. CPB-147A (8M 6-61).

(67)

Pollack, Sol, "Decision Tables for System Design", Data Processing, Vol. 7, Data
Processing Management Association, Park Ridge, TIL, 1965, pp. 485-492.

(68)

Pollack, Sol, "How to Build and Analyze Decision Tables", Federal Clearinghouse
report AD-425027, November 1963, 17 pp.

(69)

Pollack, S. L., "Analysis of the Decision Rules in Decision Tables" Rand Corp. ,
RM-3669-PR, May 1963, 69 pp.

(70)

Pomeroy, CPR L. K., Jr. USN., "Road Maps to DeciSions", Navy Management Review,
(January 1965) pp. 4-5.

(71)

Po rten , Charles, "Automated Planning of ManufactuIing Operations" Automation,

(72)

Press, Laurence 1., "Conversion of Decision Tables to Computer Programs", Comm.
ACM, Vol. 8, No.6, (June 1965), pp. 385-390.
--

(73)

Reinwald, L. T •• "An Introduction to TAB40: a processor for table-written FORTRAN IV
programs" Research Analysis Corp. AD-647418 (Nov. 1966) 46 pp.

(74)

Reinwald, Lewis T. and Soland, Richard M, "Conversion of Limited Entry Decision
Tables to Optimal Computer Programs I: Minimum Average Processing Time",
Jour. ACM, Vol. 13, No.3, (Ju1y 1966), pp. 339-358.

(75)

Reinwald, Lewis T., and Soland, Richard M., "Conversion of limited entry decision
tables to optimal computer programs (TP-197)". Dept. of Commerce Clearing House
Document AD-632 972 (Feb. 1966), 23 pp.

(76)

Schmidt, D.T. and Kavanagh, Thomas F. (1964). "Using Decision structure tables",
Datamation, (Feb., 1964), 42-46, 48-49, 52: (March 1964), 48-49, 52-54.

(77)

Shaw, C.J., ''Decision Tables - an annotated Bibliography", Report TM-2288/000/00,
System Development Corp., Santa Monica, Calif., (Dec. 4, 1965).

(78)

Shober, J.A.H., ''Decision Tables for Better Management Systems", Systems and
Procedures Journal, March-April 1966, pp. 28-32.

(79)

Sprague, V.G., "Letter to Editor (On storage Space of Decision Tables)", Comm. ACM,
Vol. 9, No.5, (May 1966), p. 319.

(80)

Sprague, V. G. and Pollack, Solomon L., "On storage space of decision tables", Comm.
ACM, Vol. 9, No.5, (May 1966), pp. 319-320.

(81)

"The Decision Logic Table Technique", AFP 5-1-1, Department of the Air Force,
September 1965, 42 pp.

(82)

Veinott, Cyril G., "Programming Decision Tables in FORTRAN, COBOL, or ALGOL",
Comm. ACM, Vol. 9, No.1, (January: 1966), pp. 31-35.

(83)

Veinott, Cyril G., "Letter to Editor (More on Programming Decision Tables)", Comm.
ACM, Vol. 9, No.7, (Ju1y 1966), p. 485.
--

(84)

Veitch, E. W., "A chart method for Simplifying truth functions", Proc. ACM (1952
Pittsburg), pp. 127-134.

(85)

William, W. K., "Decision Structure Tables, NAA Bulletin, No.9, (May 1965),
pp. 58-62.

(86)

Wright, K.R., "Approaches to Decision Table Processors", Proc. Decision Table
Symposium, (Sept. 20, 1962, New York), pp. 41-44.

(May 1966), pp. 85-89.

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

12/68

'l,

,<

23:040.001

~~ EDP
1.

AU.

........

•

SPECIAL REPORT
MAGNETIC TAPE RECORDING

10""

MAGNETIC TAPE RECORDING:
A STATE-OF-THE-ART REPORT

by
The Technical Staff of

AUERBACH mfo, Inc.

C 1968 AUERBACH Corporation and AUERBACH Info, Inc.

7/68

23;040,002

A.

SPECIAL REPORT
MAGNETIC TAPE RECORDING

AUERBACH

CONTENTS
•1

INTIWDUCTION

.2

MAGNETIC TAPE VS. DISK PACKS

.3

NEW USES FOR MAGNETIC TAPE

.:11
.32
.33
.-1

.41
. 42
· -1:l
·H

.45
.5

.51
.52
.6

.61
· G2
· G3
.7

768

Keyboard-to-Tape Encoding
Source Da ta Automation
Informution Interchange
THE MAGNETIC TAPE MEDIUM

Tape Chnracteristics .
New Tnpe Formulations
Causes of Tape Errors
Tape Handling and Storage
Shopping for Magnetic Tape
RECORDING ON MAGNETIC TAPE

Data Recording Techniques
Validity Checking Techniques
'MAGNETIC,tAPAE' HANDLERS

' "

'Meehrlllical DeSl'gb' ',:,'
, .•
Characteristics of Current Tape Handlers
Significant Hecent Developments
THE FUTURE OF MAGNETIC TAPE

A

AUfRBACH

,>

SfU""
EDP
"'011$

23:040. 100

mllm

EDP

SPECIAL REPORT
MAGNETIC TAPE RECORDING

1101111

MAGNETIC TAPE RECORDING: A STATE-OF-THE-ART REPORT
.1

INTRODUCTION
Recently-compiled evidence indicates that reports of the death of magnetic tape were premature and greatly exaggerated.
When the swing toward removable disk packs began, there were numerous predictions that
this new glamour medium would soon displace magnetic tape from its long-held position as
the primary high-speed computer input-output medium. Now that the  l 200 bits per inch in a format compatible with the IBM 729 Magnetic Tape
Units and the tape handlerll used in most other second-generation computers. Moreover.
the same unit can be used to verify the accuracy of previously recorded data.

7/68

A

(Contd.)

AUERBACH

"

•
23:040.310

SPECIAL REPORT

.31

Keyboard-to-Tape Encoding (Contd.)
Initial deliveries of the Mohawk Data-Recorders were made in April 1965, and customer
response was enthusiastic (I, 2). Since then, the Mohawk product Une has been expanded
to include both 7-track and 9-track Data-Recorders, as well as models equipped with auxUiary input or output units, special controls, and communications interfaces. The Mohawk
Data-Recorders are also marketed by the National Cash Register Company as the NCR 735
Magnetic Tape Encoders.
Mohawk has convincingly demonstrated the practicality and economic advantages of keyboard-to-tape encoding. Users cite significant cost savings through increased operator
productivity, elimination of card costs, reduced down-time, and increased computer
throughput (1). As a result, it seems likely that punched cards will gradually be phased
out of most computer installations that do not actually require a machine-processable unit
record of each transaction.
The Honeywell Keytare units, announced in January 1968, are functionally similar to the
Mohawk Data-Recorders. though Honeywell claims improved performance through features
such as higher tape speed, vacuum-drive tape handler, take-up reel (vs. Mohawk's tape
bin), improved displays, and simpUfied operation. Honeywell, like Mohawk. offers a broad
range of models for both 7- and 9-track mM-compatible tape. As an indication of the
potential size of the market for keyboard-to-tape encoders, Honeywell points out that between 350,000 and 500,000 keypunches are now in use, representing a yearly rental income
of up to $450 million for mM.
Sangamo ElectriC Company announced a line of Data Stations at the Spring Joint Computer
Conference last April. Sangamo's machines are functionally similar to the Mohawk and
Honeywell units and are offered in a similar range of models. Their chief advantage is a
continuous alphanumeric display of the present location and the content of the data in the
machine'S buffer memory.
The IBM 50 Magnetic Tape Inscriber, also introduced in April 1968, differs from the
Mohawk, Honeywell, and Sangamo units in that it uses special magnetic tape cartridges
that are not compatible with standard 7- or 9-track tape handlers. These cartridges can
be read into a System/360 computer at the comparatively slow speed of 900 characters per
second by the 2495 Tape Cartridge Reader, which was announced concurrently. The tape
cartridges are identical with those used by the mM Magnetic Tape Selectric Typewriter.
Nine tracks are recorded across the tape's 16-millimeter width at a density of 20 characters
per inch. The capacity of each cartridge is 23,000 characters. One Significant advantage
of the mM 50 Inscriber is its capab1llty to store eight dlfferent format programs, anyone
of which may be selected by the operator.
Communitype and Tally Corporation also market keyboard-to-tape encoders, though both
these units were designoJ primarily for data communications applications and record on
"DOn-compatible" tape. Other entries into this burgeoning new market can be expected
soon•

• 32

Source Data Automation
Magnetic tape is one of the very few input-output media that can come reasonably close to
keeping up with the high internal processing speeds of digital computers. It permits large
quantities of information to be stored in a highly compact form. Moreover. the tape itself
is relatively inexpensive and can be reused many times.
For all of these reasons, magnetic tape would be a highly desirable output medium for
equipment used to record data describing events or transactions at the time and place
where they occur. Yet few of the source data automation devices currently on the market
use magnetic tape. Computer-compatible digital recording equipment has generally been
considered too complex, too bulky, and too expensive for practical use in connection with
individual cash registers, typewriters, or other point-of-transaction devices.
Some recent developments point to a change in this situation and indicate that magnetic tape
w1ll soon occupy a key position in the mushrooming field of source data automation.
Magnetic' tape is being used effectively as the output medium in many transmitting data
collection systems, in which the data entered at multiple input stations is transmitted via
cables or communications lines to a central recording unit. The data transmitted from all
C 1968 AUERBACH Corporation and AUERBACH Info, Inc.

7/68

'1

23:040.320

AUERBACH STANDARD EDP REPORTS

.32

Source Data Automation (Contd.)
the input stations can be captured on a single reel of tape and then read into a computer
system at high speeds.
Some source data automation equipment records its output on narrow, low-density tape that
is not suitable for direct input to computer systems. An example is the Digitronics DataVerter system, which records adding-machine data on 1/4-inch tape for later transmission
to a remote computer system. This type of magnetic tape has many of the same advantages
and disadvantages as punched paper tape, plus two significant additional features: the tape
is usually supplied in conveniently interchangeable cartridges, and it can be reused
indefinitely.
Peripheral equipment manufacturers are striving to perfect low-priced magnetic tape
recorders that can handle computer-compatible tape. Potter Instrument Company recently
announced the ME-4210, a desk-top incremental magnetic tape unit that records up to 60
characters per second asynchronously at 200 bits per inch on IBM-compatible tape. The
price of this unit, complete with electronics and power supply, is said to be less than
$1,400 in production quantitie3. Ampex Corporation offers a line of Buffered Tape
Memories designed primarily for data acquisition applications. The buffers in these units
make it possible to record data from real-time processes in formats that are suitable for
direct input to computers .

. 33

Information Interchange
Magnetic tape has long served as an effective medium for interchanging information between
different computers or different locations within individual corporations. Now its role is
expanding to include broader types of information interchange. In these new applications,
the compactness and machine-readability of magnetic tape can lead to large savings in
clerical and shipping costa.
For example, the U.S. Internal Revenue Service now accepts tax data on magnetic tape.
The submission of tax data in this form relieves both the computer-using corporation and
the IRS of many hours of clerical labor that would otherwise be required to prepare, mail,
and transcribe hard-copy records of the data.
Magnetic tape information interchanges of this sort, as well as conversions to new computers, have often been hindered by discoveries that tape compatibility does not necessarily
guarantee compatibility with respect to recording format, character code, and collation
Requenee. Thesll prohlems, however, are gradually being overcome through steady
progress on two fronts:

.4

•

The work of the USA Standards Institute in developing the USA Standard Code for
Information Interchange (USASCII) and standards for implementing this code on
magnetic tape, paper tape, and punched cards (3, 4).

•

The gradual industry-wide swing toward the IBM way of doing things, which has
led to a fairly high degree of de facto standardization of tape, recording formats,
and codes •

THE MAGNETIC TAPE MEDHlM
Magnetic tape is a vital factor in the performance of every computer system that uses it,
yet many computer users pay surprisingly little attention to the tape until it starts giving
them serious trouble. To help you get the most out of your tape - and, in turn, your
computer - the following paragraphs review the characteristics of current and recentlyannounced magnetic tapes, the causes and remedies for tape errors, and some suggestions
for purchasing, storing, and handling tape •

. 41

Tape Characteristics
Nearly all magnetic tape currently used in data proceSSing applications consists of a base,
or substrate, of polyester film (oriented polyethylene terephthalate or equivalent) coated
on one side with a layer of ferromagnetic oxide. The oxide coating is a disperSion of
ferromagnetic material with thermosetting binders, lubricants, solvents, dispersing
agents, and other additives. The formulation of the coating is a key factor in determining
the durability, flexibility, and performance of the tape.
Standard half-inch-wide computer tape has a nominal thickness of 1.9 mils (0.0019 inch),
with a tolerance of ± 0.3 mil. The thickness of the polyester base is a nominal!. 42 mils
(about as thick as a piece of cigarette paper), while the thickness of the coating averages
about 0.4 mils and may not exceed 0.6 mils (4, 5).

7/68

A

(Contd.)

AUfRBACH

'"

SPECIAL REPORT

.41

23:040.410

Tape Characteristics (Contrl.)
The coating is applied to one side of a wide roll of polyester film which has previously been
coated with a very thin prImer coat to assure a strong bond between the base and coating.
Then the coated film passes through a strong ma.gnetic field that orienta the magnetic particles in the proper direction. After the coated film has dried, it is slit into multiple halfinch widths. All of these manufacturing operations must be carefully and constantly
controlled; otherwise. the tape will contain "built-in" defects that will prevent it from
meeting the stringent demands of data processing applications.
Most current computer tape has a nominal length of 2400 feet and is wound on reels with an
outside diameter of 10.5 inches. EIght-inch reels holding 1200 feet and other smaller reel
sizes are used in some installations. Most tape Is tested and certified by the manufacturer
for operation under one or more of the following conditions:

• 42

•

7-track tape handlers at 200, 556. or 800 bits per inch.

•

9-track tape handlers ('total surface" or "full width" testing) at 800 bits per
inch or 3200 nux changes per Inch (1. e •• 1600 bpi lUling the phase modulation
reoording technique) •

New Tape Formulations
Only three U.S. companies - Celanese Plastics Company, DuPont, and 3M Company ourrenUy produce the polyester base magnetic tape, but more than a dozen companies manufacture. and apply the oxide coating and market the finished tape (6). As a result of this
competitive pressure, plWJ the need for improved tape to keep up with advances in tape
handler technology, several significant new tape developments have been announced during
the past year.
Only one of these new formulations - DuPont's Crolyn - represents a deviation from the
usual combination of polyester base and ferromagnetic oxide coating. Crolyn, announced
in mid-1967, uses chromium dioxide as the magnetic medium. DuPont states that the
greater magnetic strength of chromium dioxide, coupled with precise control of particle
size and shape, offers two principal advantages over. conventional coatings: a higher Signal output at the same degree of resolution, and improved resolution at any given signal
level. These improvements should permit reliable operation at higher recording
deDliittlesj in fact, Crolyn tape Is saId to provide the same performance at 1600 bpi as
conventional tape at 800 bpl.
Although Crolyn tape can be used interchangeably with conventional tape in many applications, DuPont states that greater performance benefits can be obtained on equipment
designed or adapted for use with the new tape.· Honeywell became the first computer
manufacturer to offer such speclally-adapUKi equipment in November 1967, when it
announced a special feature that permtts Crolyn tape to be used at a density of 1200 bpi in
the 204B-9 Tape Unit. The feature costs $25 per month and increases the unit's data
transfer rate from 96,000 to 144,000 characters per second. Honeywell's list price for
the Crolyn tape itself was quoted as $56 per 2400-foot reel, compared to $38 for conventional tape. This cost premium, though Significant, could be more than offset by the
increased storage capacity and performance of the Crolyn tape.
IBM, after marketing tape made largely by 3M Company for many years, finally entered
the tape manufacturing business in October 1967 when it introduced Series/500 tape. This
new tape, the product of a two-year joint development effort with Sony Corporation, is .
said to have a formulation that provides an optimum balance among all the important properties such as signal strength, signal quality. pulse width, noise, binder strength, surface
toughness, durabtUty, tear reSistance, elasticity, and coating adhesion.
The 3M Company caused a major stir in the tape industry last May by adding a "guaranteed
performance" tape to its product line. The new tape, called "Scotch" Brand 777GP, uses
the same binder formulation and oxide coating as 3M's older Brand 777 certified tape. But
the new tape. according to 3M, Is so carefully controlled and tested throughout the manufacturing process that individual bit-by-bit certification of each reel is no longer necessary.
As a result, a cost saving of about $3 to $4 per reel is being passed on to buyers of Brand
777GP tape. Brand 777 tape will continue to receive bit-by-bit certification and is still
offered "for those who feel the added cost of certification is warranted because the information bc3ing recorded or stored is irretrievable and even the remotest chance of a writeskip cannot be risked."

C 1968 AUERBACH Corporation and AUERBACH Info, Inc.

7/68

23:040.420

.42

AUERBACH STANDARD EDP REPORTS

New Tape Formulations (Contct.)
Shortly after 3M Company announced its lower-coat uncertified tape, Ampex Corporation
introduced a new 870 Series tape that will be sold at a "slight premium." Ampex claims
an improved binder formula that provides longer tape life and greatly reduced head wear.
Moreover, Ampex guarantees the 870 Series tape to be 100 percent free from original
permanent errors .

• 43

Causes of Tape Errors
Magnetic tape is probably the most efficient medium yet developed for transferring data
into and out of high-speed computer systems - as long as all goes well. But when excessive tape errors are encountered, the throughput of those expensive computers can be
greatly reduced. and thousands of dollars can be lost through wasted machine time. added
labor, and missed deadlines. An essential first step toward overcoming these errors is
a clear understanding of what causes them.
By far the most common cause of tape data errors is the "dropout," a reduction of 50 percent or more in the strength of the signal transferred between the tape and the read/write
head. This loss of signal strength can cause data to be either obliterated or miSinterpreted.
Dropouts can occur either while reading or writing. Sometimes dropouts are caused by
misalignment between the tape and the read/write heads or by bare spots on the tape. By
far the most common cause of dropouts, however, is separation between the tape and the
heads.
A magnetic tape handler cannot function properly unless positive contact between the tape
and the heads is maintained at all times. A gap of as little as 140 millionths of an inch
between tape and head can result in a dropou.t (5). Separation between the tape and heads
is usually caused either by the presence of foreign matter (such as a speck of dust) or by
distortion of the tape so that it will not lie flat against the heads.
Dropouts resulting from separation between tape and heads can be divided into three classes:
..

Permanent dropouts are caused by imperfections in the tape manufacturing
process. These "built-in" dropouts are normally detected and registered when
the tape is certified.

..

Temporary dropouts are caused by bits of dust and other foreign matter that
adhere lightly to the tape surface. These can usually be dislodged when the
tape handler cycles through the re-try operations that follow detected errors.

..

Embedded dropouts are formed when temporary dropouts become permanent.
This occurs when the foreign particles become bonded to the tape surface due
to pressure and heat effects that occur when the tape passes at high speeds
over the heads, tape guides, and other hardware elements.

2mbedded dropouts represent by far the biggest single tape problem in most computer
installations. Though foreign matter such as dust and cigarette ashes are usually blamed
for embedded dropouts, an even more common cause appears to be "self dirt" from the
magnetic tape itself. The two principal types of self dirt are chips of magnetic oxide that
break away from the coating surface rmd chips and burrs left on the edges of the tape due
to faulty slitting. Thus, a key quality criterion for magnetic tape is its freedom from self
dirt, both initially and after long use.
Physical distortion of the tape is another common cause of errors and malfunctions. The
distortion is usually caused by improper winding, which may lead to skewed tape, rippled
edges. or horizontal creases. Longitudinal creases are sometimes caused by badly misaligned tape guides or rollers. Regardless of the cause, the usual effect of such tape
distortion is to prevent the tape from lying flat against the read/write heads, resulting in
dropouts (5, 7) •
• 44

Tape Handling and Storage
Many of the causes of tape errors described above can be minimized through proper care
and handling of the tape. The manufacturers of magnetic tape and tape handlers all supply
detailed instructions for handling their products, but a few basic guidelines seem important
enough to deserve mention here (5, 8).
..

..
7/68

Store tape reels on edge, preferably in a dustproof container that supports the
reel by its hub.
Keep the temperature and humidity of the tape storage area within the range
recommended by the tape manufacturer at all times.

A

(Contd. )

AUERBACH

"'

23:040.440

SPECIAL REPORT

.44

Tape Handling and storage (Contd.)
•

Bring tape stored under different conditions into the computer environment at
least six hours before use to enable it to come to thermal equ1l1brium.

•

Use soUd-flange reels for even rewinding and to keep fingers from touching
the tape edges.

•

Always handle tape reels by their hubs, and never squeeze the flanges together.

•

Never touch the tape between the load points. (If it is absolutely necessary to
do so, wear rubber gloves.)

•

Rewind all tapes in long-term storage at least once a year to relieve the 1L1ternat stresses that build up as a result of temperature variations.

•

Clean all read/write heads, capstans, and tape guides regularly in accordance
with the mallllfacturer's instructions.

•

Make sure rewind mechanisms and tape guides are checked and adjusted regularly to the manufacturer's specifications.

•

Check all tapes and tape reels for contamination regularly in accordance with
the tape manufactu.rer's instructions.

•

Do not store ordinary paper notes inside a canister with a reel of tape; the paper
is likely to shed and cause contamination.

•

Never allow any portion of the tape to touch the operator's clothing or the floor.

•

Keep the computer room as clean, dust-free, and smoke-free as possible.

•

. 45

Consider the advisabllity of a formal program of tape testing, cleaning, and
recertification, usiDg either specialized in-house equipment of the types manufactured by Cybetronlcs and General Kinetics Incorporated, or the tape rehab1l1tation services provided by a firm such as GKI Tape Service Corporation (9) .
Shopping for Magnetic T5!!!
For a number of years, the list price of computer-grade magnetic tape held firm at or
near the level of $40 per reel, and the selliDg price seldom dropped below $30 per reel
regardless of the quantity purchased. DuriDg the past few months, however, there have
been some very significant reductions in magnetic tape prices. It now appears that there
are real bargains to be had, and that the economy-minded tape buyer will be well-advised to
shop around for the supplier who offers the quality he needs at the minimum price.
The U.S. Government, which is by far the largest single user of data processing equipment,
began making significant progress on obtai'ling cost r.eductions on large-volume procurements
of computer equipment and supplies. 'the ueaeral Services Administration established a
centralized magnetic tape purchasiDg system early in 1968, and promptly began obtaining
drastically reduced prices. The lowest price reported to date was $11.50 per reel for an
order of 30,000 reels of certified 7-track tape from Audio Devices, Inc. A price of $12. 00
per reel was obtained on another order for 47,000 reels of 7 - or !I-track tape from 3M Company. Prices of $13. 00 per reel or less have been obtained from several other tape suppliers,
and to date the Government has noticed no deterioration in the quality of the tape obtained at
these low prices (11). It would, of course, be unrealistic for small-volume users of magnetic
tape to expect price quotations as low as those obtained by the Government.
Buyers of magnetic tape should always remember that price is by no means the only factor
in determ1n1Dg the best b~. Quality Is even more important, because the cost of the tape
itself is insignificant compared to the losses that can result from excessive tape errors.
But quality, in the case of magnetic tape, has many aspects which are difficult to measure
and evaluate objectively.

C 1968 AJERBACH Corporation and AUERBACH Info. Inc.

7/68

..P:040.450

.45

AUERBACH STANDARD EDP REPORTS

Shopping for Magnetie

T~

(Contrl.)

Though the buyer should give careful consideration to the manufacturer's warranty and
reputation, the only sure test of quality is the actual performance of the tape in the buyer's
own installation. Thereforp. whenever possible. the buyer should test several reels of a
new brand of tape in his own installation before placing a large order. And, of course.
continuous records should be kept of the number of errors encountered in processing each
reel of tape. These records make it easy to decide:

.5

«I

Which specific reels of tape are flawed and should be rehabilitated or destroyed;
and

...

Which brand of tape provides the best overall performance throughout its lifetime .

RECORDING ON MAGNETIC TAPE
To aid computer users in understanding the functions and effective utilization of magnetic
tape equipment. the following paragraphs briefly explain the techniques that are currently
being used to record data on tape and to ensure its validity. Throughout this section the
emphasis is on techniques used in IBM and IBM-compatible tape handlers (12. 13) .

. 51

Data Recording Techniques
Magnetic tape recording is based upon the interaction between a moving magnetic storage
medium (the tape) and a stationary magnetic transducer (the read/write head). During
recording, the head magnetizes the oxide coating of the tape in a small region immediately
adjacent to the head. During readback of the recorded signals. the head provides an
Induced voltage that reflects the rate of change of magnetization. The path of recorded
signals generated along the surface of the tape by each head is called a "track." Current
IBM-compatible tapes are recorded by either 7 or 9 heads in parallel. and are therefore
referred to as 7 --track or 9-track tapes.
In recording digital data. the two binary digits 0 and 1 must be converted into the appropriate states of magnetic surface saturation (and/or reversals in saturation). The
three most important digital recording techniques are NRZ (non-return to zero), NRZI
(non-return to zero. IBM), and phase modulation (called "phase encoding" by IBM). A
continuous writing current is used in all three methods, although manufacturers' documentation often refers to discrete magnetized "spots" on the tape.
NRZ (non-return to zero) was at one time the most common method of digital recording,
though it is no longer widely used for the reasons discussed below. In NRZ recording.
the direction of the writing current is reversed whenever a 0 is followed by a 1 or a 1 is
followed by a 11 in the input data. Therefore. one direction of surface magnetization
corresponds to a 1 while the opposite direction corresponds to a O. Figure] shows the
relationships between the input data and the writing current for a single track in NRZ
recording. A major disadvantage of the NRZ technique is that if anyone bit is in error,
all the bits that follow '\\111 be read erroneously until the next signal pulse is encountered.
NRZI (non-return to zero, IBM) is the teohnique used in all ourrent IBM magnetic tape
handlers operating at recording densities of 200, 556, and 800 bits per inch. In NRZI
recording. the direction of the writing current is reversed every time a 1 is to be recorded.
Therefore, a 1 is represented by either a positive or negative pUlse, while a 0 is represented by the absence of a pulse. These relationships are illustrated in Figure 1. NRZI
offers one major advantage over NRZ recordIng: if anyone bit is misread in NRZl, the
error will have no effect on the bits that follow.
Phase modulation (or phase encoding) is the technique used in all current IBM magnetic
tape handlers operating at 1600 bits per inch. In this method, a 1 is represented by a
positive change in the writing current, while a 0 is represented by a negative change. As
Figure 1 shows, an additional current reversal must be inserted at mid-bit time whenever
a 1 is followed by another 1 or a 0 is followed by another O. This means that a tape to be
used for lS00-bpi recording in the phase modulation mode must actually be capable of
storing 3200 magnetic flux reversals per inch. The phase modulation technique provides
a clear distinction between 0 bits and no data; the NRZI mode lacks this distinction.
Because the phase modulation technique records a flux reversal in every track position of
every properly-recorded data frame. it has two inherent advantages over the NRZI
technique:

7/68

•

The absence of a flux reversal indicates an error condition; this indication, in
combination with a vertical parity check, permits in-flight correction of singletrack read errors.

•

Each track i8 ;]elf-clocking, so the chances of errors due to skewed recording
are greatly reduced.

A

iContd. )

AUERBACH

'"

SPECIAb REPORT

23:040.520

INPUT DATA

I

0

1

1

0

1

0

0

I
I
I
NRZ

NRZI-;---+-+-r;-+---~~---+--~

PHASE

MODULATION

-t-++++-+-H-+-+-++-HH-

INPUT DATA

o

1

1

o

1

o

o

Figure 1. Comparison of write currents in
three different recording modes .

. 52

Validity Checking Techniques
Adequate provisions for ensuring the accuracy of the recorded information must be provided in every digital tape recording system. The detection of tape errors is usually a
hardware function. while the correction of these errors may be performed by the hardware. by programmed routines. or by a combination of the two methods. The capability
to detect and correct write errors at the time they occur is a particularly desirable
feature; if the error is not detected until the tape is read back. its correction may involve
a long, expensive reconstruction process.
The hardware error checking techniques used in current IBM magnetic tape handlers are
explained in the paragraphs that follow. Table II (13) summarizes the checking schemes
used in the various models and makes It clear that the schemes used in the 7 -track NRZI.
9-track NRZI. and 9-track phase modulation modes all differ very significantly from one
another.
Vertical redundancy checkih!' otten called "row parity checking". is probably the oldest
and most commonty used c eck upon recording accuracy. A parity bit is appended to the
code for each character or byte to be recorded on the tape; the value of this bit is either
o or 1. whichever is needed to make the sum of all the 1 bits in the row either even (in
even-parity checking) or odd (in odd-parity checking). Nine-track tape is always recorded
in the odd-parity mode.
When the tape 1s read. each row is checked to make sure that it contains an even or odd
number of 1 bits. depending upon the parity mode. In the NRZI system. each row is also
read back and checked for proper parity immediately after it is written; this is often
called "read-after-write parity checking." Vertical redundancy checking detects all
single-bit errors. but it may fail to detect errors involving two or more bits in a tape row
unless it is combined with other checking techniques.
tC 1968 AUERBACH Corporation and AUERBACH Info, Inc.

7/68

23:040.521

AUERBACH STANDARD EDP REPORTS

TABLl'; II: ERROR CHECKING IN IBM TAPE HANDLERS
Error Checking Technique
Recording
Te('hnique

NRZI

Phase
Modulation

IBM
Magnetic
Tape Unit

Vertical
RE'dundancy
Check

Longitudinal
Redundancy
Check

Cyclic
Redundancy
Check

Single-Track
Error
Correction

R

R
(Programmed)

Envelope
Check

MultlpleTrack Error
Check

Skew
Check

2401
Models 1-3
(9-track)

R. W

R. W

2401
Models 1-3
(7-track)

R. W

R. W

2401
Models 4-6
(9-track)

R. W

R. W

R

2415
Models 1-6
(9-track)

R. W

R. W

Note 1

2415
Models 1-6
(7-track)

R. W

R. W

2401
Models 4-6
(9-track)

R

R
(Automatic)

W

R. W

R

2415
Models 4-6
(9-track)

R

R
(Automatic)

W

R

R

2420
Model 7
(9-track)

R

R
(Automatic)

W

R. W

R

W

W

W

R
(Programmed)

.

W

W

R - Check is performed during tape Read operations.
W - Check Is performed during tape Write operations.
Note 1 - CRC byte Is written for compatibility with 2401. but is not checked •

. 52

7/68

Validity Checking Techniques (Contd.)
Longitudinal redundancy checking, used in all IBM-compatible. NRZI-mode tape handlers.
monitors all seven or nine tracks to ensure the presence of an even number of 1 bits in
each track of every block recorded on the tape. During writing. a longitudinal redundanoy
check (LRC) character is appended to each block; the LRC character contains either a 0 or
1 bit in each track position, as required to make the total number of 1 bits even In that
track of the block being recorded. As the block is read. the LRC character is regenerated
and checked for agreement with the LRC character read from the tape. Longitudinal
redundancy checking detects all single-bit errors. but it may fail to detect errors involving
two or more bits in one track within a block.
The combination of vertical and longitundinal redundancy checking, combined with the
read-after-write checking made possible by dual-gap read/write heads. proved to be a
very effective system for detecting (but not correcting) tape errors in the 7-track NRZI
tape handlers used in most second-generation computer systems.
Cyclic redundancy checking is used, in addition to vertical and longitudinal redundancy
checking, in the IBM 2401 Magnetic Tape Units when operating in the 9-track NRZI mode
at 800 bpi. This third type of redundancy checking provides additional information that
permits programmed correction of single-track read errors. As each block is written.
a cyclic redundancy check (CRC) character is automatically calculated from the data bytes
and appended to the block. just ahead of the LRC character. During a read operation. the
CRC character is recalculated in the same manner and compared with the CRC character
read from the tape. When a Single-track read error occurs. the track in error can be
identified and the error corrected by means of a programmed reread.
Read errors involving two or more tracks cannot be corrected in this manner, although
IBM emphasizes that they can often be reduced to correctable single-track errors through
repeated reading of the faulty block. In fact, the standard IBM tape error routines read
each error block ]00 times. backspacing the error block across the tape cleaner to dislodge any loose particles after every tenth try, before conceding that the block-contains
an uncorrectable permanent error.

A

(Contd. )

AUERBACH

'"

23:040.600

SPECIAL REPORT

.52

Validity Checking Techniques (Contd.)
Single-track error correction is performed automatically and "in flight" by the 1600-bpi
IBM tape handlers that use the phase modulation recording mode. When reading. these
handlers continuously monitor the signals from all nine tracks. As soon as anyone track
falls to provide II. flux reversal in any data frame, that track is "disabled" for the remainder
of the block (1. e .• its contents are disregarded). Using the vertical redundancy check bit
for each row. the information bits in the disabled track are then regenerated automatically.
This scheme provides automatic correction of lIll errors which are confined to Ii single
track. Errors involving weak signals from two or more tracks within a block are detected
and recognized as uncorrectable. and standard error recovery procedures (backspace and
reread) must be used in these cases.
Envelope checking and multiple-track error checking are two specialized checks employed
in the mM tape handlers that record in the phue modulation mode at 1600 bpi. These
handlers do not perform redundancy checks during write operations. but a weak signal
from any track is indicated by the envelope check on signal amplitudes. The multipletrack error check indicates an abnormal change in data rate during writing or the detection
of weak signals from ~o or more tracks within a block during reading.
Skew checks are designed to detect (and in some cases compensate for) excessive vertical
misalignment of the 7 or 9 recorded bits comprising each row. Excessive skew is usually
a result of improper tape winding or handling techniques .

.6

MAGNETIC TAPE HANDLERS
Havml examined the magnetic tape medium and the techniques used for recording and
cheoking data on it. we will now survey the current status of magnetic tape handlers, the
computer hardware components that transport the tape and read and record information
on it.

. 61

Mechanical Design
Magnetic tape handlers used to record digital data can be divided into two basic classes:
incremental and constant-speed.
In an incremental tape handler, the tape is started and stopped each time a character is
recorded. This permits the unit to operate asynchronously in applications where the
input data is received at widely varying rates, as in telemetry or industrial data collection. But the speed and reliabiUty of incremental data handlers are generally regarded
u inadequate for effective on-line use with high-speed digital computers,

Therefore. virtually all magnetic tape handlers currently used in computer systems tall
into the oonstant-speed class. In these units. the recording of a block of data is not permitted to begin until the tape has been accelerated to a velocity very close to its rated
speed. Once the proper tape velocity has been reached. all of the characters or bytes
comprising the blook are transferred to the tape handler at a constant rate and recorded
on the tape at a constant density. Blank spaces. called interblock gaps. must be left
between the recorded blocke of data so that the tape can be stopped and then reaccelerated
to Its rated velocity betwee!1_ consecutive blocks.
To achieve high performance in computer tape handlers, it is necessary to drive the tape
past the read/write heads at speeds ranging from about 20 to 200 inches per second.
Moreover. the handler must be capable of accelerating the tape to these velocities - and
decelerating it back to a standstill - within a few milliseconds. Providing for these ultrafast starts and stops has severely taxed the current state of the art in electromechanical
design.
The physical form of nearly all computer tape handlers can be characterized by the
methods used to perform two key mechanical functions:
•

Driving the tape past the read/write heads, and

•

Buffering (isolating) the movement of the tape past the heads from the comparatively massive inertia of the reels on which the tape is wound.

The portion of tape to be accelerated can be isolated from the storage reels by vacuum
columns. by swinging tension arms, by a combination of vacuum and arms. or by storage
bins. Nearly all of the tape handlers currently in use with computer systems have buffers
of the vacuum-column type, ranging from a few inches to several feet in length. Tension
arms and storage bins are sUll used in lower-performance tape handlers for off-line use;
the Mohawk Keyed Data-Recorders, for example. use a tape bin in lieu of a take-up reel.
Three basic methods. with numerous variations, are used to drive the tape past the read/
write heads (14). In each of these methods, the tape is held in contact with a rotating
capstan.
C 1968 A'_:'::RBACH Corporation and AUERBACH Info. Inc.

7/68

23:040.610
.61

AUERBACH STANDARD Eoe REPORTS

Mechanical Design (Contd.)
Pinch rollers were among the earliest methods developed for obtaining rapid tape acceleration, and they are stm in widespread use despite the advantages of the newer drive
methods described below. The pinch-roller method uses two capstans rotating in opposite
directions at a constant speed. The tape is driven in either direction by a roller that
clamps it against the appropriate capstan. Tape movement is stopped by a brake or by
another pinch roller that clamps it into contact with a stationary capstan. This method
subjects the tape to very high accelerating forces. Moreover, the rubbing and hammerlIke blows from the pinch rollers have been accused of causing excessive tape wear.
Dual vacuum capstans are used in a newer drive method that imposes lower accelerating
forces and less wear on the tape. Two counter--rotating, slotted capstans are used.
When either capstan is evacuated, the tape adheres to it and is driven in the corresponding
direction. Simultaneously, the other capstan is pressurized so that the tape "floats" over
it on an air film. Vacuum brakes are used to stop the tape. A variation of this method
uses positive air pressure, applied externally, to "clamp" the tape to the capstan. Still
other tape handlers use a combination of vacuum and pressure clamping.
Single-capstan drives are used in many of the high-performance tape handlers announced
during the past four years. This method uses a single capstan, driven by a low-inertia.
high-response servomotor, that accelerates and decelerates in either direction at the
same speed as the tape. The tape is held in contact with the capstan by vacuum teclmiques.
This drive method combines fast starts and stops with low tape wear; in many current
handlers the oxide (recording) surface of the tape touches only the read/write heads and
the tape cleaner.
The principal physical characteristics of most of the current computer tape handlers, as
discussed in the preceding paragraphs, can be summarized by a simple tree diagram
(Figure 2). Each of these tape drive and buffer methods is highly developed today; obtaining further improvements in mechanical performance (as distinguished from increases
in recording density) may demand new approaches to the drive and buffer problems.

PHYSICAL FORM

I
Buffer Method

Tape Drive Method

I

I
Dual Vacuum
Capstans

Pinch
Rollers

Figure 2.

. 62

I

-

Single
Capstan

Vacuum
Columns

I
Tension
Arms

Storage
Bins

Principal physical characteristics of current magnetic tape handlers .

Characteristics of Current Tape Handlers
More than two dozen U. S. companies now manufacture digital tape handlers. and they offer
models with a wide range of performance characteristic!! and prices. The great majority of
the tape handlers now on the market are "IBM compatible." They use standard 1/2-inch tape
and record in on(' 01' more of the following modes:
•

7/68

7-track NRZI at

2{1.I,

556, and/or 800 bits per inch.

A

Alii flHACH

(Contd. )

SPECIAL. REPORT

23:040.620

TABLE

m:

CHARACTERISTICS OF REPRESENTATIVE TAPE HANDLERS

IDENTITY

PERFORMANCE

RECORDING CHARACTERISTICS
No. of
Tracks

Recording
Density.
bits/inch

Interblock
Gap Length.
inches

Tape

Data Transfer

Rewind

Speed.
inches/sec

Rate.
KB or KC/sec

Time,
minutes

0.50
0.50

7
9

200/556/800
200/aOO/1600

0.75
0.60

90
90

18/50/72
18/72/144

.-1.6
1.6

604
809
626

0.50
0.50
1.00

7
9
14

200/556/800
800
800

0.75
0.60
0.75

75
37.5
150

15/42/60
30
240

3.5
1.3

GE

MTH301
MTH3H
MTH412

0.50
0.50
0.50

7
7
9

200/556/800
200/556/800
200/556/800

0.75
0.75
0.60

75
75
150

15/42/60
30/83/120
40/111/160

1.3
1.3
1.3

Honeywell
Honeywell
Honeywell

204A-2
204B-9
204C

0.75
0.50
0.50

9

400
556/800/1200
800

0.67
0.70
0.60

120
120
36

64
24/96/144
29

1.3
1.3

IBM

IBM
IBM

2401-1
2401-6
2420

0.50
0.50
0.60

9
9

9

800
1600
1600

0.60
0.60
0.60

37.5
l12.5
200

30
180
320

3.0
1.0
1.0

NCR
NCR

833-117
833-211

0.50
0.50

7
9

200/556/800
1800

0.75
0.60

50
90

10/28/40
144

3.2
2.0

RCA
RCA

70/441
70/4:45

0.50
0.50

7

9

500
800

0.58
0.65

50
150

25
120

3.2
1.2

UNIVAC
UNIVAC

11 VIC
11vmC

0.50
0.50

800
200/556/800

0.60
0.75

42.7
120

34

7

3.0
1.3

Manufacturer

Model
No.

Tape
Width.
inches

Burroughs
Burroughs

B 9391
B 9393

Control Data
Control Data
Control Data
GE
GE

.62

7

9

9

24/67/96

1.3

Characteristics of Current Tape Handlers (Contd.)
•

9-track NRZI at 800 bits per inch.

•

9-track phase modulation at 1600 bits per inch (3200 fci).

All of the major computer manufacturers are now offering mM -compatible tape handlers
for use with their computers I and it seems apparent that non-compatible tape handlers will
soon be obsolete except in specialized applications.
Table m summarizes the major characteristics of 20 representative magnetic tape handlers
currently being marketed by the major computer manufacturers. All except three of these
units (the Control Data 626, Honeywell 204A. and RCA 70/441) fall into the IBM-compatible
class. Table m is by no means a complete listing; its purpose is simply to illustrate and
compare the main features of some typical third-generation tape units.
Tape speeds for the present computer tape handlers range from less than 20 to 200 inches
per second. Data transfer rates (the product of tape movement speed times recording
density) range from about 4,000 to 320,000 characters or bytes per second. Start and
stop times of 2 to 10 milliseconds are common, Interblock gap length is generally 0.75
inch for 7-track tape and 0,60 inch for 9-track tape. Full-reel rewind times are usually
In t~e range of 1 to 4 minutes. Most of the current tape handlers use vacuum-column
buffers and vacuum tape drives, although pinch-roller drives are still in common use,
especially within the mM line •
. 63

Significant Recent Developments
Four recent developments within the magnetic tape handler market seem to merit discussion:
•
•

The entry of most of the computer manufacturers into the peripheral equipment
business.
The availability of plug-and-program-compatlble tape handlers designed to
replace mM equipment.

•

The announcement of mM's high-performance 2420 Magnetic Tape Unit.

•

The introduction of Burroughs' low-cost, ultra-compact Magnetic Tape Cluster.
C 1968 AUERBACH Corporation and AUERBACH Info, Inc.

7/68

23:040. 630

.63

AUERBACH STANDARD EDP REPORTS

Significant Recent Devel0l!ments (Contd.)
As a result of rapid progress in increasing the performance and reducing the cost of computer mainframes, it is estimated that peripheral equipment now accounts for about 65
percent of the total hardware cost of typical computer systems - and that this figure w1U
rise to 75 percent within the next few years. As a result of this trend, nearly aU of the
major computer manufacturers are now building most of their own peripheral equipment,
and several are now offering their magnetic tape handlers to all comers on an OEM basis.
During the past year, Ampex, MAl, Potter, and Telex have all announced new magnetic
tape handlers that are plug-interchangeable and program-compatible with the IBM 729 and/or
2400 Series Magnetic Tape Units. (The MAl handlers are manufactured by Potter.) Potter
claims to have orders for over 600 of its plug-compatible units to date.
The IBM 2420 Magnetic Tape Unit, announced in January 1968. reads and writes standard
9-track tape at a speed of 320,000 bytes per second while maintainIng compatibility with
IBM's slower 1600-bpi tape hal'dlers. Tape speed is 200 inches per second, and access
time to the next sequential block of data is less than 2 milliseconds. A single-capstan
vacuum tape drive is UBed, and tape threading is automatic. Rewinding is performed at
500 inches per second without removing the tape from the vacuum-column buffers (15).
An optional tape cartridge can be attached to each reel of tape used with an IBM 2420 Tape
Unit. Both reel and cartridge are mounted on the handler as an integral unit. The cartridge
provides a sealed container for the tape and, in conjunction with the automatic threading
feature, eliminates all physical handling of the tape. In view of the improvements it provides in both performance and tape handling, the IBM 2420 must be rated as the most
significant advance in tape handler technology in several years. In the wake of the 2420's
introduction, there were two noteworthy related developments:
•

IBM stopped accepting orders for the 7340 Model 3 Hypertape Drive, which
provided data transfer rates of 340, 000 bytes per second but used noncompatible one-Inch-wide tape in sealed twin-reel cartridges.

•

Telex (formerly Midwestern Instruments) announced a new tape handler that
will serve a·s a plug-compatible replacement for the IBM 2420 Tape Unit and
provide identical performance.

Whereas the IBM 2420 is designed to provide maximum performance, the Burroughs
Magnetic Tape Cluster represents a novel approach to the problem of providing moderate
performance at minimum cost. Announced in 1966, the Magnetic Tape Cluster provides
two, three, or four tape drives in a single compact cabinet (33 inches wide, 30 inches deep.
and 42 inches high). Each tape drive has its own pinch-roller drive mechanism and
vacuum-column buffers, but a single drive motor, power supply, vacuum supply, ventilation system, and electronics unit serve all the tape drives in a clUBter (16). The feed reel
and take-up reel for each drive are mounted on concentric shafts, with the feed reel above
the take-up reel. Tape speed is 45 inches per second, and data transfer rate Is 9,000
bytes per second at 200 bpi, 36,000 bytes per second at 800 bpi, or 72,000 bytes per second
at 1600 bpi. This approach to tape-handler packaging raises the hope of further cost-cutting
innovations yet to come •
•7

THE FUTURE OF MAGNETIC TAPE
This report has presented considerable evidence indicating the continued importance of magnetic tape, as a computer input-output medium. Sales of both tape and tape handlers for data
proceSSing use are expected to increase during the next few years at the rate of 10 to 15 percent per year. As explained in Section .2 of this report, magnetic tape and disk packs will
each find their way into the types of applications for which they are best suited, and computer
systems using both of these media will become increasIngly common.
Further improvements in the performance of magnetic tape handlers will be achieved mainly
through increases in recording density. During the past decade, practical recording densities have increased from 100 or 200 bits per inch to the present 800 or 1600 bits per inch,
and the full potential of achievable tape resolution has not yet been exploited. Many experts
believe that further improvements in both tape formulatioUB and tape handler designs will
make recordIng densities of 3000 to 4000 bits per inch practical within the next few years.

(Contd. )
7/68

AUERBACH

'"

23:040.700

SPECIAL REPORT

.7

THE FUTURE OF MAGNETIC TAPE (Contd.)
Converlely, the advances in the mechanical elements of magnetic tape handlers have been
far less dramattc. The Unlservo I tape drives used with UNIVAC I in 1951 had a tape speed
of 100 inches per second, and the fastest drives available today provide only a 2-to-1
advantage over that speed. No major breakthroughs in mechanical design that are likely
to change this picture can currenUy be foreseen.
Cartridge loading and automatic threading techniques such as those employed in the IBM
2420 Magnetic Tape Unit will probably be more widely used. In fact, such techniques will
be almost mandatory to avoid contamination of the anticipated ultra-high-denslty tapes.
In tape haDdler design, there is still plenty of room for further decreases in cost, improvements In rellabillty, and reductions in tape wear. Progress In any of these areal w1l1 be
warmly welcomed by tape users, and the intensified competition among tape handler
luppllerl lbaWd spur them to Increased efforts.
Rapid growth In the use of magnetic tape in direct keyboard-to-tape encoding, source data
automation, and Information Interchaage, as described In Section. 3 of this report, will
help to elUlure a bright future for magnetic tape in EDP.
REFERENCES

(1)

F.H. Reagan, Jr., "Will Mohawk Make Punched Cards Obsolete?", Data Processing
Mapzitae, December 1886, pp. 46-51.

(2)

D. G. Price, "Whither Keypunch?", Datamation, June 1967, pp. 32-34.

(3)

P. B. Good8tat, "USMCn, What's It All About," Data Procelsl. MagaZine, June 1967,
pp. 20-2".

(4)

"Proposed USA StaIIdard: Recorded Mqnetic Tape for Information Interchange (200 cpl,
NRZI) , " Communications of the ACM, November 1967, pp. 730-737.

(5)

J. M. RiCCi,

(6)

J. Snyders, "Magnetic Tape: A Me.sage About the Medium," Business Automation,
February 1888, pp. 34-39.

(1)

"Manapment Looks at Computer Tape: The Technical View," General Kinetics, Inc.,
1886.

(8)

B. Shapley, "The Care and Storage of Magnetic Tape," Data Proce8siy Magazine,
AprU 1968, pp. 80-81.

(9)

J. J. DeJlanne, "In-House Tape RehabUltattoa," Datamation, August 1965, pp. 51-52.

(10)

"mM Reduces Tape Prices by 13%," Computerworld, November I, 1967, p. 1.

(11)

"Computer Tape Down to $11.50," Computerworld, July 3, 1968, p. 1.

(12)

A. 8. HoaglaDd, Digital Mapetio Reoordlng. Wiley, New York, 1963, pp. 1-26, 125-130.

(13)

"~rec1sion

Magnetic Tape," Datamation, October 1966, pp. 51-60.

IBM 2400 Series Mapetlc Tap! Units aad 2816 Switching Unit, Form A22-6866, mM
t pp. 1.. 11.
.

COJ'l).

(14)

F. Moritz, "Six Ways to Drive Tape," Control Engineering, March 1968, pp. ,82-85.

(15)

mM 2420 Model 7 Magnetic Tape Unit, Form A22-6918, mM Corp., pp. 3-7.

(16) J. T. Gardiner, "The 'Cluster' - Four Tape Stations in a Single Package," AFIPS
Conference Proceedings, Volume 30 (1967 SJCC), pp. 245-252.
--

o 1968 AUERBACH

Corporation and AUERBACH Info, Inc.

7/68

23:050.001
Sill .....

EDP

SPECIAL REPORT
HIG~SPEED PRINTERS

.o,IlS

HIGH-SPEED PRINTERS:
A STATE-OF-THE-ART REPORT

Prepared By
The Technical staff of
AUERBACH Info, Inc.

CJ 1968 AUERBACH Corporation and AUERBACH Info, Inc.

8/68

23:050.002
SPECIAL REPORT
HIGH-SPEED PRINTERS

CONTENTS
•1

BACKGROUND

.2

THE DEVELOPMENT OF HIGH-SPEED PRINTERS

.3

SOME mSTORICAL METHODS OF. PRINTING

.31
.32
.33
.4
.41
.42
• 43
.44
.4fi

Stick Printers
Multiple Typebar or Wheel Printers
Matrix Printers
mGH-SPEED PRINTERS TODAY
On-the-Fly Printing Techniques
Drum Printers
C~ain Printers
Oscillating-Bar Printers
Highf'l' Speeds and Improved Registration

.5

COMPARISON CHART

.6

FUTURE OUT WOK

.61
.62
.7

Conventional Printers
Non-Impact Printers
APPENDIX: UNE PRINTER TERMINOLOGY

I

\
8/68

6

•

-£.

23:050.100
11111"1

~EDP

-

.U£MAC~

SPECIAL REPORT
HIGH-SPEED PRINTERS

millS

HIGH-SPEED PRINTERS: A STATE-OF-THE-ART REPORT
.1

BACKGROUND
Since its inception two decades ago, the computer industry, as a result of numerous advances in electronic technology, has succeeded in developing and producing progressively
faster central processors. Concurrently with, and as a consequence of, these advances,
the industry has been continually faced with the problem of getting information into and out
of these central processors at rates compatible with their ever-increasing internal speeds.
In the early days of the industry, when computers were utilized mostly in scientific and
mathematical applications, single-action character printers were fast enough to cope with
the limited amount of output data that they were required to print. These applications Invol\,ed large amounts of computational time with relatively small volumes of input and output. With the advent of commercial and business applications for electronic computers,
circumstances changcd drastically. Large volumes of data were fed into the computer and
a relatively small amount of computation was perfomed on each data record. In many cases
the data output was voluminous, causing serious problems in producing usable output last
enough.
When input-output devices are connected directly to a computcr system, the throughput of
that system becomes limited by its slowest component, whether that component is a magnetic tape unit, the card reader, the central processor, or the output printer. In most
cases, the card reading and magnetic tape data transfer rates are sufficiently fast to make
the printer the slowest factor. Thus, the use of an on-line printer tends to slow down the
system considerably.
A temporary solution reached in the mid-1950's was to record data at high speeds on magnetic tape, and then, at separate "off-line" stations, transcribe this data from the tape
to various typps of pl'inters. The speeds of these pl'lnters ranged from ;, lines per minute
up to 1,800 lines per minute. However, these statIOns were qUilt' t'xpenSII'(' and performed
only a single function. Accordingly, they have all but (hsappearpd, to be replaced, at least
in larger installations, by small computing systems, thought of as "satellites" to the central
computer. Thus, while the printer is "on-line" to the> small computer, it is "off-line" to the
central system, and does not slow down the throughput rate of a large, expensive computer .

.2

THE DEVELOPMENT OF HIGH-SPEED PRINTERS
Whatever the speci fic method of incorporating a printer in a computer system, speed has
been a major consideration in the development of printing devices. Various mechanisms
have been tried. Generally, the most successful from the standpoint of speed have been
impact-type printers, which print by means of some kind of mechanically-driven typebar
or type-generating device. More specifically, the trend has been to parallel ("line-at-atime") printers, which print an entire line essentially with one stroke, or, at least, in one
complete printer cycle. Since there is no moving carriage in these printers, much greater
speeds are achieved than are possible with serial ("charactet'-at-a-time") printers, which
print each character essentially in a separate cycle, in con'el'l left-to-right sequence across
the print line.
.
These non-sequential printers generally utilize continuous pin-feed forms, and incorporate
some form of high-speed skipping, in which multiple lines can be skipped at several times
the normal printing speed .

.3

SOME HISTORICAL METHODS OF PRINTING
The following paragraphs describe three printing methods which have all but disappeared
from the high-speed printer scene, but which are of historical interest because of their
influence upon the designs of today's high-speed printers .

. 31

Stick Printers (Example: IBM 370)
One technique used for serial printing at intermediate speeds employed a single print
stick, which was normally an eight-sided metal element embossed with eight characters
on each face to provide sixty-four print characters. The character to be printed was selected by the decoding logic, which actuated a rotation and/or an in-out movement of the
stick. At the time of the "dwell" (no movement) of the stick, a single hammer struck the
paper from the rear, moving the paper into contact with an inked ribbon against the printing stick to produce the printed character.
Horizontal positioning and carriage returns were accomplished by moving the entire printing assembly across the platen in a manner somewhat similar to the action of typewriters.

© 1968 AUERBACH Corporation and AUERBACH Info. Inc.

8/68

SPECIAL REPORT

23:050.310

.31

Stick Printers (Example: IBM 370) (Contd.)
The general characteristics of stick printers were:
•

Relati vely low speed (30 to 60 lines per minute).

•

Ribbon motion across the paper, as in a conventional typewriter.

It is to be noted that this general printing method is still widely used in console typewriters

and communication devices, such as the IBM Selectric element and the newer Teletype
model!>, where economy is more important than high speeds •
. 32

Multiple Type-Bar or Wheel Printers (Examples: IBM 403, 407)
Many early line printers, including several that were adapted from punched-card tabulating
machines such as the widely-used IBM 407 and 403, employed a series of type bars or
wheels. Each printing position had a separate bar or wheel containing all characters of the
print set. All positions were printed simultaneously, after the entire line had been decoded
and each bar or wheel had been independently positioned. The actual printing occurred when
hammers, driven by electronic triggers, struck the paper into contact with an inked ribbon
against the type face.
The general characteristics of wheel or bar type printers were:

• 33

•

Relatively low speed (50 to 150 lines per minute).

•

Ribbon motion across the paper, as in a conventional typewriter .

Matrix Printers (Examples: IBM 720, 730)
Since the physical positioning and recoil movement of individual hammers against the embossed characters has been one of the limiting factors in the design of faster printers, a
number of high-speed printers employed matrix-type print heads. Each head consisted of
a small rectangle of fine wires. Characters were formed by-electromechanically actuating
selected individual wires in each print head and, with these wires, striking the ribbon against
the paper. Matrix printers employed either a stationaryheadassembly<>ra moving head assembly.
The stationary assembly had one head for each printing position, while the moving assembly
had one-half or one-fourth as many individual heads spaced farther apart. Each head of the
modng assembly printed in two or four positions in turn after the entire head assembly had
b{'en shifted laterally a short distance.
In general, experience with matrix printers was characterized by frequent and troublesome
mechanical maintenance and service problems.
The general characteristics of matrix printers were:
•

High speed (500 to 1,000 lines per minute).

•

A hidden flat metal platen.

•

Ribbon motion across the paper, as in a conventional typewriter.

•

A relatively poor-quality printed image •

.4

HIGH-SPEED PRINTERS TODAY

.41

On-the- Fly Printing Techniques
The printing techniques described above have given way almost completely to several
variations of the "on-the-fly" approach, in which high print speeds are achieved by extremely rapid hammer action against continuously moving type elements. The principal
variations involve the use of a rotating drum, a horizontally-moving chain, or an oscillating
bar, as detailed below; the actual methods of printing are quite similar in all three techniques.
During each print cycle (normally the time allocated to load the print buffer; decode its
contents; print one line, including hammer action and recoil; and space the paper), all
characters move past the print hammers at each printing position. The character to be
printed is selected by decoding, and a fast-action hammer, controlled by an actuator,
presses the paper against the type slug at the exact moment the required character is in
position. If the machine is printing at 600 lines per minute, each total printing cycle takes
one six-hundredth of a minute. This interval is in turn divided into discrete timing units
for each character which is available, plus several units for paper advance.
In the asynchronous mode of printing, such as is used in the Anelex 5000 drum printers,
thc firing of the hammers does not commence at any fixed point during the rotation of the
character set. Hather, firing commences whenever a signal is received to indicate that
Iine spacing Ins been completed and the print buffer loading is finished. Firing terminates
when a countl'l' indicates that all characters have moved past the hammers or when the buffer
holdin~ the line of characters to be printed has been sensed and found empty.
Hammer action in "on-the-fly" printers is either by: (1) free flight, or "ballistic, ,. hamlllers
(movement stopped by contact with the paper and the type element), or (2) "controlled flight"

8/68

A

AUERBACH,

(Contd.)

HIGI+-SPEED PRINTERS

.41

23:050.410

On-too-Fly Printing Techniques (Contd.)
hammers (fixed spatial movement). The most important advantage claimed for the latter
design principle is positive control over the depth of penetration of hammer action. When
such a printer is operated without paper in the tractor feed, the hammers are prevented
from striking the type element by "end of paper" safety switches.
Vertical format control is generally effected by an 8- or 12-channel paper tape loop. The
vertical spacing of the punches controls the actual spacing on the printed sheet. In some
printers it is necessary to use a loop the exact vertical size of the printed page; in others it
is possible to use loops representing only the vertical area to be imprinted. It is usually
possible to space the printer under program control.
The general characteristics of current "on-the-fly" printers are:
•

. 42

High speeds (300 to 1,200 lines per minute).

•

The absence of a platen.

•

Ribbon movement parallel with paper motion; ribbon width at least
equal to maximum line length.

•

Hammers which strike the paper from behind .

Drum Printers (Examples: Anelex 4000, Honeywell 222, GE PRT201)
A widely-used on-the-fly printing technique is to provide a complete character set (sometimes two or more complete sets) at each print position, and to distribute these character
setR around the circumference of a solid, continuously rotating drum. The timing mechanism
senses the passage of a particular character in front of the hammers, and then fires the
hammers which correspond to the pOSitions in which the given character is to be printed.
Thus, if all the hammers were fired at the same instant, the printed line would consist of
the same character printed at all positions.
Several drum printers have utilized the "shuttle" technique, which cuts in half the number
of hammers needed and hence reduces the cost. In a "shuttle printer" (e. g., Anelex 4000),
the odd-numbered columns are printed in one cycle, then the paper is "shuttled" one column
to the left and the even-numbered columns are printed. Note that since two cycles are needed to print each line, the effective speed is halved .

. 43

Chain Printers (Exflmples: IBM 1403, Potter HSP-3502, CDC 512)
Ina chain printer the hammers must be individually timed, because each character travels
horizontally across many printing positions during the print cycle. Several identical sets of
characters are assembled serially on a horizontally moving chain which resembles a bicycle
chain. At each print position, the paper is forced into contact with the ribbon against the
chain by a solenoid-activated hammer fired as the appropriate character on the chain passes
the printing position. In the IBM 1403 Model 3, the chain has been replaced by a "train"
mechanism in which type slugs move in the same horizontal plane as in the chain at more than
twice the speed of the original 1403 Printer. If all hammers were fired simultaneously in a
chain printer, several sets of sequential characters rather than a line of identical characters
would be printed .

. 44

Oscillating-Bar Printers (Examples: IBM 1443, UNIVAC 3030, Datamark OBP)
An oscillating-bar printer operates much like a chain printer, except that the print slugs are
inserted in a horizontal bar that moves rapidly back and forth instead of being attached to a
continuously-tra veling chain.
The highest printing speeds that can be achieved using this start-stop-reverse type of motion
are considerably lower than those that are possible with a continuously-rotating chain or drum;
the fastest available oscillating-bar printer operates at about 600 alphanumeric lines per
minute. However, a bar printer is likely to cost less than a drum or chain unit of comparable speed, and it offers the added advantage of permitting rapid removal and replacement of type-bars, a valuable asset where an installation'S application mix requires the
use of several different character sets •

. 45

Higher Speeds and Improved Registration
A number of techniques have been implemented to increase effective printing speeds, usually
by making the speed a function of the character set or of the actual number of characters
being printed at a given instant.
One of these techniques is useful where only numeric printing is necessary. Printers designed for all-numeric printing are equipped with drums or chains on which numeric type
faces are repeated several times, often with blank print segments between the groups for
spacing. Such an arrangement (generally with two sets of print characters) permits two
lines to be printed for each drum revolution. Thus, at 1,000 revolutions per minute, 2,000
lines of numeric print per minute can be produced.
Another popular technique is to overlap cycles if only a limited character set is being used
at a given time. With this technique the full character set is present on the drum, but if
C 1968 AUERBACH Corporation and AUERBACH Info. Inc.

8/68

23:050.450

SPECIAL REPORT

TABLE I: CHARACTERISTICS Of CURRENT LINE PRINTERS
PHYSICAL FORM

IDENTITY

Manufacturer

Pl'Intmg

Model

Anele, Corp.

Technique

4000
5000

Anelex Corp.
Burroughs Corp
Burroughs Corp.

B 9240/41/43
B 320/3~1!325
B 328/329

Burroughs Corp.

PRINTING CHARACTERISTICS

Character

VerUcal
Format
Control

Set (No. of
Printable

Tape

Characters)

Drum
Drum

8 ch. (12 opt.)
8 ch (12 opt.)

Drum
Drum
Drum

12 channels

12 channels
12 channels

Horizontal

Vertical

Posit.ons

Spacmg
(char/Inch)

Spacing
(lines /Inch)

120 to 160
80 to 160

10
10

64
64
64

120 to 132
120/120/132
120/132

10
10
10

64
48

136
136

64 (96-128 opt. )
64 (96-128 opt. )

Control nata Corp.
Control Data Corp.

501/505
512

Drum
Horizontal chain

Datamark, Inc.
Datamark, Inc
Datamark, Inc.

300
500
OBP

Drum
Drum
Oscillating bar

Data Products Corp
Data Products Corp
Data Products Corp.

4300
4400
4500

Drum
Drum
Drum

a channels

64 to 128
64 to 128
64 to 128

General Electric Co.

8 channels

12 channels

64
64
64

8 channels

20
20
20

5
5
5

10
10

6 or 8
6 or 8

20
21

5
3

)0
10
10

6
6 or 8

14.88
17.75
17.75

4
4·25

3,5

19
19

3
3

120 or 132
96/108/
120 or 132
120 or 132
120 or 132

10
10

6 or 8
R

20
20

4 5
4.5

10
10

6 or 8
6 or 8

20
20

4.5
4.5

10
10

b
6
6
6

13 to 63

120
100 or 132
120
132
120 or 144
120 or 144

~

8 channels
channels

52 to 64
52 to 64

132
132/160

10
10

Drum
Drum

8 channels
8 channels

63
63 (49 opt.)

Honeywell EDP

222-4

Honeywell tDP

~22-6

Drum
Drum

!j

8 channels
channels

63 (49 opt.)
63

12
12
12
12
12
12

channels
channels
channels
('haMels
channels
channels

48
16 to 240
48
48 to 240
13 to 63

104 to 136

10

10

10
10
10

IBM COIP,

2203-A L -A2

NCR
NCR

640-102/-300

Drum

640-200 / -210

Drum

Potter Instrument Co,

HSP-3502

4 channels

Up to 192

132 or 160

10

RCA
RCA
RCA

70/242
10/243-30/-40
701243-51-61

Drum
Drum
Drum

12 channels
12 channels
12 channels

64
64
96

132 or 160
132/160
132/160

10
10

I SCIentific Data Systems

7440/7460

Drum

8 channels

56

Shepard LaboratOries

400

Drum

8 channels

64

Drum

None
None
None

63
63
63

I

l NIY~C
l·:-;IVAC

075f1-0n

0768-00. -99
3030-00.'-02

I D,IYAC
.45

Drum
Oscillatmg bar

4 25

6

122/122-1
222-1/-2/-3

Horizontal cham

6 or 8

6 or 8
6 or fl

Honeywell EDP

roM

6 or 8

6 or 8
6 or 8

10
10

64
64

Drum

4
4

136
136

12 channels
12 channels

Horizontal cham
Horizontal cham
HOrlzontal tram
OscUlating bar
Oscillatmg bar

20
20

19

Drum
Drum

1132
1403-1/-2
140J-6· -7
14u;J-NI
1443-1'1

6 or 8
6 or 8

22

64/64/48 or 64

Corp.
Corp.
Corp
Corp
Corp

Minimum

10

8 channels

IBM
IBM
IBM
IBM

132
132
132

MaxJmum

10
10

8 channels

Drum

Honeywell EDP

80 to 160
80 to 160
132 to 160

Form Width (Inches)

6
6
6

PRTI00/110/
120
PRTI50
PRT201

General Electric Co
General Electric Co,

Number of
Print

102
Up to 200

In
132
96 to 132

6

or

or
or
or
or
(j or
6 or

19
19

8
8
8
6
8
!:!

6

16 5
18.75
18.75
18.75
16.75
16.75

4.75
~.

fi

3.5
3.5

4
4
3,5

22
22

3.5

6

18.5

2.5

10

6 or ~
6 or 8
6 or 8

18.75
18.75
18.75

10

6

20

4

10

6

21. 5

4.5

10
10
10

G or 8
6 or 8

22
22
22

4

6 or R

6 or 8/6

4
4
4

4
4

Higher Speeds and Improved Registration (Contd.)
only a restricted sct of contiguous characters are being printed, the paper advance and bufferload cycles can take place during the remainder of the drum-rotation cycle, with a consequent
increase in effective speed.
While "on-the-fly" printers have been, from their inception, characterized by frequent misalignment or misregistration of the printed characters, it may be safely stated that printing
quality has been greatly improved, especially on the more expensive printers. For example,
any misregistration in chain-printed copy is horizontal; that is, the spacing between adjacent
characters is uneven. This type of misregistration is less noticeable than the waviness of
the printed line which is characteristic of drum printers. Even the latter are, in general,
now capable of producing high-quality copy with little detectable waviness when properly
adjusted ami maintained .

.5

COMP ARISON CHART
The accompanying comparison chart (Table 1) summarizes the characteristics of more than
50 on-the-fly printer models that are currently being marketed by nine computer manufacturers and five independent suppliers.
The information on the chart is divided into five major categories - Identity, Physical Form,
Printing Characteristics, Performance, and Application - and the individual column headings
are largely self-explanatory. Printing speeds. in lines per minute, are shown for three
separate cases:

•
8/68

Peak speed for single-spaced printing using the full alphanumeric
character set, which usually contains from 48 to 64 characters.

A

AUERBACH

(Contd. )

23:050. !IOO

HIGH-SPEED PRINTERS

TABLE I: CHARACTERISTICS OF CURRENT LINE PRINTERS (CONTD.)
mENTlTY

APPLICATION

PERFORMANCE

Speed (llaea/mtaulel
Peak
(with reatrlcted
character aetl

1-lIlch &peein,
(wItiI full
eharacter aell

Mlxlmum
No. or
Coplea

IIdppln,

Model

Peak
(with full
character letl

4000
1000

300
1250

376
1500

211
740

6
6

21.1
25 (75 opt.)

Bt240/41/43
8320/321/325
B328/329

700/1'lol0/IO'0
475/7,,1/100
1040

700/1040/1040
-115/700/7(,0
1040

550/572/572
380/550/550
572

6
6

25 (75 opt.)
25
25

Control'Data Co.".
Control Data Co.".

SOl/50S
612

800/500
1200

1000/500
1500

Datamark, Inc.
Datamark, Inc.
Datamarl<, Inc.

300
500
OBP

300 min.
1000
300 mm.

Data Produeta Co.".
Data Producta Co.".
Data Produela Co.".

4300
4400
4500

1000
360
600

MIUlUIaclilrer

Allel... Co.".
Anel... Co.".
Burroup Co.".
Burroup Co.".
Burrou(lbs Co.".

General Electric Co. PRTIOO/110/120 300/600/780
General Electric Co.
General Electric Co.

PRTlaO
PRTZOI

571/375

6

Speed

(Inchea/secl

?

6

25
70

300
1200
300

212
545
212

8
6
6

10
25
17

1000
360
600

500
216
300

6
6
6

35
20
20

6

Repreaentallve
Computar Byatems
Va"" Tble Pr_r

Burroup 500 Byltems
Burrouibe 200, 500 Byatema
Burroup 200, 500 Systeml
CDC 3000, 6000 Serlel
CDC 3000, 8000 Serlel

300/600/780

249/414/491

6

14.5 (63 opt.)

GE liS, 130

600
1200

600
1200

484
811

5
5

27.5
27.5

GI 400 Series
GE 400, 600 Serlel

450/300

450/300
1300

8
8

50/20
35

Honeywell 120/110
Honeywell Series 200

Honeywell EDP
Honeywell EDP

122/122-1
222-1/-2/-3

Honeywell EDP
Honeywell IDP

222-4
222-6

9:;0
1100

1266
1100

610
750

8
6

35 to 50
35 to 50

Honeywell Serlel 200
Honeywell Serlea 200

1132
1403-1/-2
1403-6/-7
1403-Nl
U43-Nl
2203-AI/-A2

80
ROO
340/MO
1100
240
350/260

110
340/600
1400
600
750/600

80
500
306/500
805
214

6
6
6
6
6
6

10
33 or 75
33
33 or 75
15
15

mM
mM
IBM
mM
mM
mM

328/400
888

6
10

17
90

IBM
IBM
mM
IBM
IBM
IBM

Corp.
Co.".
Co.".
Corp.
Corp.
Corp.

NCR
NCR

640-102/-300
640-200/-210

Potter Inatnament Co. HSP-3502

fPiO

450/ROO

128~

900/1200

1:"00

3000

380/286
490

450

850

240

6

16 5

RCA
RCA
RCA

70/242
70/243-30/-40
70/243-51/-61

625
1250
R33

625
1250
633

469
750
525

6
6
6

27 5
27.5 (35 opt. I
27.5

Scientific Data
Systems

7440/7460

628/760

Shepard Laboratories
t:NIVAC
UNIVAC
UNIVAC

400
0758-00
0768-00/-99
3030-00/-02

.5

800/1000
2400

1200
12~~

900/1200

250/600

1600
1100/1600
500/1200

6

1130
1400 Series, System/360
140IG, 1440, System/360
Syatem/360
Syltem/360
Syltem/360 Mod. 20

NCR Century Serlel
NCR Century Series

RCA Spectra 70
RCA Spectra 70
RCA Spectra 70
SOS S.gma Series

545

6

13.8

800

6

220/451

6
6

33
33
25

UNIVAC 418, 490, I10S
UNIVAC 9400
UNIVAC 9200, 9300

COMPARISON CHART (Contd.)
•

Peak speed for single-spaced printing using a restricted subset
of the full character set.

•

Effective speed when the average spacing between printed lines is
one inch and the full character set is used.
Prices are not shown on the chart because of the difficulty of obtaining prices that are truly
comparable with respect to the amount of associated control circuitry (controllers, buffers,
etc.) included with the print mechanism •
•6

FUTURE OUTLOOK

.61

Conventional Printers
While future announcements of new impact-printer models are anticipated, it is likely that
future advances will be made prtmarily in the areas of reliability and cost, rather than in
speed.
Since the limiting factors on printing speed tend to be mechanical, associated with paper
handling and hammer motion, it appears that present high-speed print mechanisms are
approaching an upper bound of, perhaps, 2000 alphanumeric lines per minute. However,
there is ample room for further improvements in overall performance through advances in
mechanical reliability and serviceability, and in improved timipg to give better accuracy of
registration. Improvements in these areas have 1leen noticeable over the past few years,
and the upper limits have not yet been reached.

o

1968 AUERBACH Corporation and AUERBACH Info, Inc.

8/68

SPECIAL REPORT

23:050.610
.61

Conventional Printers (Conld.)
High-speed printer mechanisms have tended to be quite expensive ($30,000 to more than
$80,000); and, since no significant price reductions have occurred in the last four to five
years, it may be assumed that there is, in effect, a lower limit to the price of a high-quality
mechanical printer. The industry may experience some price decreases as a result of
improved production methods, but these are not likely to be very dramatic •

• 62

Non-Impact Printers
A breakthrough in printer design may come in the form of non-impact printing techniques
such as photographic, xerographic, or cathode-ray-tube methods. Several non-impact printers are now on the market. One uses an interesting technique in which a character is written on a CRT and then piped through a fiber-optics cord to print on light-sensitive paper.
This unit is reported to run at 6000 alphameric lines per minute, and is an example of the
sort of inspired design of which the industry is capable.

The major hurdles facing non-impact printer manufacturers are: first, that most non-impact
printers require speCially-treated paper, which is expensive and often of an unpleasant consistency to the human touch; and second, that these printers are, at present, incapable of
producing more than one copy of the printout, which is a crippling disadvantage in some
commerical applications. It is interesting to speculate that perhaps non-impact printing
devices will eventually be so inexpensive as to allow the purchase of multiple units, all
driven in parallel by a single set of electronic logic; yet even there, the question of whether
such parallel-produced documents are legal copies of one another will have to be resolved.
Thus, non-impact printing techniques offer the potential for high printing speeds at comparatively low costs, but some serious problems will have to be overcome before they will be
effective across-the-board competitors for mechanical printers. The impact printer is here
to stay, and, while it will be supplemented in certain applications by the newcomers, it is
not likely that it will soon be supplanted by them .
•7

APPENDIX: LINE PRINTER TERMINOLOGY
Alphanumeric
Pertaining to a character set that includes both alphabetic characters (letters) and numeric
characters (digits). Note: Most alphanumeric character sets also contain special characters, such as punctuation or control characters.
Carriage
That portion of a printing device which serves to hold and transport the paper being printed
upon.
Chain printer
A line printer in which the type slugs are mounted on a chain that moves horizontally past
the printing positions. Note: Chain printers generally provide more accurate vertical registration than the more commonly used drum printers, and interchangeable chains often
permit rapid changes in the size or make-up of the character set.

Character set
A set of marks or signals used to represent data; e. g., a typical character set for a printer
might include the digits 0 through 9, the letters A through Z, and the common punctuation
marks.

Control character
A coded character which is part of a computer program or some common-language medium.
Instead of being printed, a control character initiates some kind of mechanical activity on
the part of the device being used for printing (e. g., carriage return, tab, or skip).
Drum

With reference to printing, the imprinting device in an on-the-fly printer, conSisting of
a constantly revolving shaft, drum, or series of interlocked wheels embossed with the
characters which are to be imprinted.
Edit
To rearrange information. Editing may involve the deletion of unwanted data; the selection
of pertinent data; the insertion of various symbols, such as page number and typewriter
characters; and the application of standard processes such as zero suppression.
Font
A family of graphic character representations (1. e., a character set) of a particular size
and style.
~
The total area of a single print position.

8/68

fA

AUlRBACH

(Contd. )

SPECIAL REPORT

.7

23:050.700

APPENDIX: liNE PRINTER TERMINOLOGY (Contd.)
Hard copy
A visible record on a permanent medium.
Line printer
A printer that prints all the characters comprising one line during each cycle of its action.
On-the-fiy printer
A printer in which the type remains in motion during the printing process; at the appropriate
instants during its 'movement, the paper and type are forced together to cause the desired
characters to be printed.
~
The horizontal distance between corresponding points of adjacent type characters; e. g. ,
12-pitch (12 characters per inch) is "eUte" pitch, 10-pitch is "pica" pitch, and 8-pitch is
"billing" pitch.

Platen
An element of the carriage in a typing or printing device which is usaUy (but not necessarily)
a hard rubber cylinder. The function of the platen is to support the paper as it is struck by
the type face, and to guide the paper as it is spaced.
Print position
A position in which anyone of the members of the printer's character set can be printed in
each line. Note: Most of the current line printers have between 80 and 160 print positions,
i. e., they can print between 80 and 160 characters per line.
Registration
The physical pOSitioning of a print line or character (vertical or horizontal registration)
with relation to a form set or the machine itself.
Skip or Slew
To move paper in a f.dnter, without printing, through a distance greater than the normal
line spar.ing, usually at a higher speed than in a single-line advance.
Solenoid
An electro-mechanical actuator used to convert electrical energy into physical movement.
In printers, solenoids are used to fire the print hammers.
Special character
A character that is neither a letter nor a digit; it may be a punctuation mark (e. g., comma)
or a control character that causes a particular operation to be performed (e. g., carriage
return).
Tractor
A device used on printers to control the vertical movement of papd' through the carriage,
normally by means of pinion wheels which engage pinfeed or punched-hole margins.
Vernier
A printer control, normally rotational in nature, used for fine vertical or horizontal
carriage adjustments to align the form being printed while the printer is operating.
Vertical format control tape
A punched paper or plastic tape, usually 8- or 12-channel, formed into a loop and used to
control the spacing and skipping of a line-printer carriage.

C 1968 AUERBACH Corporation and AUERBACH Info. Inc.

8/68

.....

1.

23;060.001
II ......

~EDP

"lItM8Al'~

SPECIAL REPORT
RANDOM ACCESS STORAGE

1t'1I11S

RANDOM ACCESS STORAGE
A STATE-OF-THE-ART REPORT

Prepared hy
tht' Technical Staff of
A VEB BACH Corporation

© 1968 AUERBACH Corporation and AUERBACH Info. Inc.

4/68

--£.

23:060.002

III.....

~EDP

SPECIAL REPO~T
RANDOM ACCESS STORAGE

*"~•

~.".n

CONTENTS

•1

RANDOM ACCESS DEFINED

·2

HARDWARE TYPES

· 21
• 22
· 23
• 231
• 232
· 233
·3

TilE ECONOMICS

· 31
· 32
· 33

Access Times
Storage Costs
Throughput Costs

·4

SYSTEMS CONSIDERATIONS

· 41
• 42

· 43
·5

4/68

Drums
Disc Files
Cartridge-Loaded Units
Magnetic cards
Disc packs
Tape loops

Faster Response
Timely Management Information
Integrated Operations
THE COMPARISON CHART

A ..

AUERBACH

(Contd. )

23:060. 100

£ "......

/A~ EDP

AU(RllAC~

SPECIAL REPORT
RANDOM ACCESS STORM,"

I..---_-----.J
IH'llS

~

RANDOM ACCESS STORAGE:
A STATE-OF-THE-ART REPORT
.1

HANDOM ACCESS DEFINED
Handom access storage is a vital component of most automated systems desih'11ecl to provide
faster response and improved control in an ever-widening scope of applications: management information systems. production control, order processing. inventory manag-pm('nt,
n'scrvations. message switching-. process control. and many more. The computer Sy,;!,·· .. "
that perform these advanced data processing functions generally must employ equipment 01
thl' on-line type. in which the storage files are directly accessible to th(' computer so that
data storag€' and retrieval can be both immediate and automatic. The on-line file concept
calls for a storage medium that permits data to be retrieved rapidly and selectively. on a
random basis.
Handom access storage devices are also desirable for effective utilization of multiprog-rammed computer systems (in which utilization of the equipment is maximized by processing
several independent programs concurrently) and high-performance software (compilers.
operating systems. sorting routines. etc.). The importance of this type of equipment in
the current computer market is illustrated by the fact that IBM's System/360 line indudes
eight different types of random access storage devices with a wide range of data capacities.
access times. and data transfer rates.
The functional meaning of the term "random access" is best understood by comparing random access storage with magnetic tape storage. Data is stored on magnetic tape in serial
form. and the time required to retrieve a certain piece of data is depend€'nt upon its location on the tape. Retrieval time. therefore. can vary widely according to thc location of
the data within the storage medium. In contrast, the time to retrieve data from random access storage is not related to its location in the medium. The retrieval time for anyone
particular item of data is - in thc ideal case - the same as for any other Item of data.
This idealized definition of random access storage does not strictly apply to most of the
existing mass random access storage devices. In these devices the access times to retrieve two different items of data may differ slightly according to the locations of the data.
Time is required to move the section of the storage medium containing the desired data into
position under the read/write head. This is called "latency" or "rotational delay". Latency
is directly dependent upon the relationship between the locations of the desired data and the
data currently under the read/write head; to bring the new data into position under the readwrite head may require a quarter. half. or full turn of the storage medium.
An additional period of time. called "head positioning time". may be required to position
the read/write head over the proper track of the storage medium. In any case, the variance
in access times is measured in milliseconds - whereas several minutes would be required
to search through all the data on a reel of magnetic tape.
One storage medium that does meet the strict definition of random access is the computer's
internal core or thin-film memory. All data contained in it can literally bc a(', '·ss,·,1 in
equal time. regardless of its physical location. Although functionally ideal, core or thinfilm memory is economically impractical for most mass storage purposes because of the
high cost per character stored.
A highly significant recent development in this area is Control Data's Extended Core Storage
Units, which make up to 20 million characters of core storage available to users of the ultralarge-scale Control Data 6600 system. Data transfers between the Central Memory and the
Extended Core Storage Unit start within three microseconds after the instruction is issued
and proceed at the unprecedented rate of 100 million characters per second.
Large-capacity. nonmechanical storage of this type will greatly facilitate efficient utilization
of large-scale computer systems in multiprogramming, time-sharing environments. but as
yet 1ts cost is still too high to justify its use for master-file storage in most applications.
However, current development work in this area indicates that within a few years it may b,
possible to store hundreds of millions of characters in this kind of medium and access thpnt
within a few microseconds - and at a reasonable cost .

.2

HARDWARE TYPES
The most commonly-used mass random access storage devices at the present time arp magnetic drums. magnetic disc files. and cartridge-loaded units. These threc basic types of
devices differ functionally in a number of ways that can be important from an applications
viewpoint.

© 1968 AUERBACH Corporation and AUERBACH Info. Inc.

4 fiB

AUERBACH STANDARD EDP REPORTS

23:060.210

.21

Drums
Magnetic drum devices consist of a revolving drum with a magnetizable surface on which
information is arranged in tracks. Read/write heads pick up and record data as the desired items pass beneath them. This means that there may be a rotational delay of up to
one drum revolution when accessing a given record. In practice. the delay averages out
to one-half revolution. Most magnetic drum devices employ an individual. fixed read/
write head for each track. so that this rotational delay is the only time factor that must be
considered in the accessing operation.
High data transfer rates are, frequently achieved by recording data Simultaneously (in parallel) in two or more adjacent tracks. IBM's 2301 Drum Storag(' unit reads and records
four bits in parallel and transfers 1.200.000 characters per I:wcond. Control Data's 863
Drum reads and records 13 bits (2 characters plus a parity bit) in parallel aDd can transfer
up to 2.000.000 characters per second.
When compared to the other types of random access storage devices. drums have relatively
fast access times and transfer rates. relatively low storage capacities. and a relatively
high cost per character stored. The type of drum memory with a fixed read/write head
serving each data track is partiCUlarly well suited to the storage of systems prof,Tams. address directories for larger-capacity random access units. and for on-line applications
where short response time is more important than large storage capacity.
Though most magnetic drum units use multiple fixed read/write heads. there are some
exceptions. UNIVAC's Fastrand units use movablc access mechanisms to decrease the
number of read/write heads necessary to serve large data stores. In these drum units. as
in most disc files. the access time is significantly increased whenever it is necessary to
move the heads from one data track to another. (Optionally. a small extra storage area is
available which is served by special fixed read/write heads and can always be accessed without head-positioning delays.)
The Fastrand n Mass Storage Unit. used with UNIVAC 418. 494. and 1108 computer systems.
contains two large drums with a total storage capacity of 132 million characters - more than
most disc files. The average time required to position the read/write heads over the selected tracks is 58 milliseconds. followed by a rotational delay that averages 35 milliseconds.
All heads move in unison. and 688.128 characters of data are always under the heads at any
given position of the accchanism
that positions all arms in unison. The General Electric DSU204 Disc Storage Unit. for example. can have from 4 to 16 discs. each served by an independently positionable access
arm. This arrangement provides considerable flexibility: 368.640 characters are available
at any ti me without head repositioning. and there is no need to restrict the data layout to
"cylindprs" in which all the traeks to be accessed at one time are in corresponding positions
on the various discs.
Burroughs. in its extensive line of Disc Files. eliminates all movement of the read/write
heads by providing an individual head for each data track. Consequently. the total access
time is limited. as in fixed-head drum devices. to the rotational delay time. which averages
from 17 to 60 millist->conds in the various models. This is substantially less than the time
required by most of the disc files in which a comb of access arms has to be moved horizontally across the dise surfaces. The head-per-track dcsign used in the Burroughs units
makes their performance characteristics more nearly comparable with those of large drums
than with movable-head disc units. yet their costs are low enough to make them suitable for
many large-volume applications.
Disc file development has been hampered by two major mechanical problems: positioning
movable read/write heads with the desired speed and precision. and maintaining proper
spacing between the heads and the disc surfaces. A number of complex electro-mechanical
techniques have been developed to position the heads quickly and accurately. but their uncertain reliability still causes occasional headaches for both manufacturers and users.
The read/write heads must be kept within a few ten-thousandths of an inch of the magnetic
recording surface in order to achieve the high recording densities required for high data
transfer rates and large storage capacities. To avoid damaging physical contact between
the heads and the rapidly revolving disc surface. many units use the principle of "floating"
the read/write heads on a layer of air generated by the rotational friction of the discs.
Some units also employ solenoids as a fail-safe device that retracts the heads in case of
power failure. Although these solutions are obviously workable. they are mechanically
complex and expensive. Several manufacturers are now developing disc files and/or drums
in which continuous physical contact is maintained between the r(·ad/write heads and the
recording surfaces; in these units the major design problem is the minimization of wear to
ensure reliable long-term performance .

. 23

Cartridge- Loaded Units

The third basic type of random access storage device is the cartridge-loaded units. which
utilize a variety of different types of magnetic media. NCR's CRAM. RCA's Model 3488
and 70/568-11. and IBM's 2321 Data Cell Drive all use magnetic cards or strips. which are
extracted from a replaceable cartridge and wrapped around a revolving drum that carries
them past the read/write heads. The IBM 1311 and 2311. the Control Data 850 Series. and
numerous other units use removable stacks of discs. Potter's RAM unit uses continuous
loops of magnetic tape.
Each of these units represents an attempt to combine the rapid-access capabilities of random access devices with the practically unlimited total storage capacity (on-line plus offline) of magnetic tape. From an applications point of view. the total storage capacity and
flexibility of operation gained by having interchangeable cartridge units must be measured
against the relatively long delays that occur whenever cartridges must be manually interchanged to make new information available on-line .
. 231 Magnetic cards
The tr&il-blazing NCR CRAM (Card Random Access Memory) unit uses flexible magnetic
cards. A cartridge contains 128. 256. or 384 cards. For a read/write operation. the selected card is dropped from the cartridge and held by vacuum against the revolVing drum.
which carries it under the read/write heads. After the card has been read and/or recorded
upon. it is stripped from the drum, and its momentum carries it up through a return chute
and back into the cartridge. There is no need for the cards to be replaced in any particular
sequence; the selector rods can cause the selected card to drop, regardless of its position
in the cartridge, through the use of binary-coded notches in the top of each card.
In the original NCR Model 353-1 CRAM. each cartridge can store over 5.5 million alphanumeric characters. Each card in the cartridge has seven 3.100-character data tracks,
all of which can be read or recorded upon when the card is wrapped around the revolving
drum. The recording mode is similar to that of many magnetic tape systems; there are
eight bit channels per track. and a "read-after-write" check is performed upon recording.
The newer Model 353-2 and 353-3 CRAM units use bit-serial recording, one bit channel per
1. 120-character data track. This change in the recording mode reduces the equipment cost
and increases cartridge capacity to 8 million characters in Model 353-2 and 16 million characters in Model 353-3, but it also results in a lower data transfer rate than that of the Model
353-1.
C 1968 AUERBACH Corporation and AUERBACH Info, Inc.

4/68

23:060.231

AUERBACH STANDARD EDP REPORTS

.231 Magnetic cards (Contd.)
As part of its third-generation Century Series computer line. NCR announced a new. largecapacity CRAM unit. Model 653-101. Each 653-101 unit stores up to 124.416.000 bytes (or
248.832.000 packed decimal digits) in a single 384-card cartridge. The Mylar cards are
3.65 inches wide and 14 inches long. Each card contains 144 tracks. and each track can
store 2.250 bytes. Data is recorded serially by bit at a density of 1. 500 bits per inch.
The time required to drop a selected card from the cartridge and wrap it around the revolving drum. ready for reading or writing. has been reduced to 125 milliseconds (versus
235 milliseconds in earlier CRAM models). A movable head assembly contains 36 read/
write heads which move in unison to one of four positions in order to service all of the 144
tracks on the card.
RCA's Model 3488 Random Access Computer Equipment uses the same basic principles as
CRAM. but each Model 3488 unit can hold 8 or 16 interchangeable card magazines at a time.
Each magazine holds 256 cards and up to 42 million characters of data. Each card contains
64 bands of two tracks each. and each band holds four 650-character blocks of data. Four
pairs of read/write heads are moved. in unison. to one of 16 possible positions so that they
can serve all of the 64 bands. Access time to data on a particular card is normally between
290 and 465 milliseconds. depending upon the position of the addressed magazine. Model
3488 storage is intended for applications where a large volume of relatively inexpensive random access storage is needed. rather than where fast access is important.
The newer Model 70/568-11 Mass Storage Unit, used with RCA's third-generation Spectra
70 computers. is functionally similar to Model 3488. but each of the eight on-line magazines
in the 70/568-11 can store up to 67.1 million bytes. Recording is bit-serial, each card contains 128 data tracks. and each track holds 2.048 bytes. Eight read/write heads move in
unison to one of 16 positions to service all of the tracks.
IBM's 2321 Data Cell Drive. like RCA's Model 3488. provides economical storage for extremely large volumes of data in applications where relatively slow access times can be
tolerated. Each 2321 drive stores up to 400 million characters (or 800 million packed decimal digits) in 10 removable. interchangeable "data cells" with a capacity of 40 million characters each.
Data in the 2321 is recorded on magnetic strips. 13 inches long and 2.25 inches wide, which
are arranged in data cells mounted vertically around the circumference of a cylinder or
"tub file" that can be rotated. Each of the 10 data cells is divided into 20 subcells. and
each subcell contains 10 magnetic strips. There are 100 recording tracks on each strip.
and each track can hold a maximum of 2,000 characters. A bidirectional rotary positioning
system positions the selected subcell beneath an access station. The selected strip is withdrawn from the cell. placed on a separate rotating drum. and moved past the read/write
heads. where reading and/or recording take place. Then the strip is returned to its original
location in the cell. When a previously addressed strip is on the drum. time to access data
on a different strip varies from 375 to 600 milliseconds.
On the basis of direct equipment costs per character stored. the magnetic-card devices
clearly provide the most economical random access storage now available. But the prospective buyer should not overlook: (1) the relatively slow access times of most magneticcard devices. which may lead to intolerably low system throughputs; and (2) the spotty
reliability record of these devices to date - excessive downtime and rapid card wear have
been serious problems in a number of installations .
. 232 Disc packs
During the past few years. much of the action in the random access storage field has involved
"disc pack" drives. Pioneered by IBM with its 1311 Disc Storage Drive in 1962. the disc
pack concept represents a combination of the virtues of discs and magnetic tape that is finding
ever-increasing acceptance among computer installations of all sizes and types. Moreover.
the production of the interchangeable disc pack cartridges is a rapidly growing industry;
companies now marketing disc packs include IBM. Business Supplies Corporation of America.
Caelus Memories. Consolidated Electrodynamics. Control Data. Honeywell. Kee Lox Manufacturing. Mac Panel. Management Assistance Inc .• Memorex. Tab Products. and Wright
Line.
The IBM 1311 and 2311 Disc Storage Drives are patterned after the larger IBM 1301 and 2302
Disc Storage Units. They use the comb-type access mechanism with interchangeable disc
pack cartridges consisting of a stack of six discs. The 10 inside disc surfaces are used for
data recording. A cartridge has a total storage capacity of 2.980.000 characters in the 1311
and 7; 250.000 characters in the 2311. and it can be replaced in about one I'ninute. Compared
with the IBM 1301 and 2302. the 1311 and 2311 have much lower on-line storage capacities
but offer the advantages of cartridge loading and lower price tags.
Control Data Corporation also offers interchangeable-cartridge disc storage drives. with
three models announced to date. Model 852 introduced the concept of "compatibility" into
the random access field for the first time. being functionally identical with the IBM 1311.

4/68

A

(eontd. )

AUERBACH

'"

SPECIAL REPORT

23:060.232

.232 Disc packs (Contd.)
Compatibility in this case refers to the disc pack cartridges. which can be intcrchanged
between IBM 1311 and Control Data 852 drive units. The Control Data version differs from
the 1311 in having a faster head positioning time. The other two Control Data models. the
853 and 854. use the same six-disc cartridges and have the same head-positioning times as
the 852. but have higher data capacities.
A number of other computer manufacturers (including General Electric. Honeywell. HCA.
and UNIVAC) and independent peripheral equipment manufacturers are now marketing disc
storage units that use the same disc packs - though not necessarily the same data recording
formats - as the IBM 2311 Disc Storage Drive. Thus. the disc pack can potentially servc
as a useful medium for inter-computer communication. in the same manner as magnetic
tape has long been used.
In the IBM 2314 Direct Access Storage Facility. IBM provides nine disc storage drives (eight
for on-line use and one spare). each capable of handling a removable 2316 Disk Pack and
storing up to 29.18 million bytes with an average positioning time of 75 milliseconds. Although the 2316 Disc Packs used with the 2314 and the 1316 Disc Packs used with the IBM
1311 and 2311 are conceptually similar. they are not interchangeable. Each 2316 Disc Pack
is divided into 200 "cylinders" holding 129.384 bytes each. The access mechanisms on the
individual drives can move independently and simultaneously. although all of the access arms
on any specific drive always move in unison.
A dual-spindle disc drive is the key peripheral device in the third-generation NCR Century
Series computer line; every Century system will contain at least one disc unit. and all software is disc-oriented. Each NCR disc unit has two vertical spindles. and each spindle
drives an interchangeable disc pack ihat holds 4.2 million bytes. The NCR disc pack. however. is not IBM-compatible. It consists of three discs with six plated metallic recording
surfaces. all of which are used for data storage. Each spindle has a comb-type access
mechanism. and each disc surface is served by 12 read/write heads. arranged in such a
way that the maximum arm movement is only 3/16 inch. As a result. arm movement time
never exceeds 60 milliseconds and averages only 42 milliseconds .
. 233 Tape loops
Potter's RAM unit offers a number of interesting features. Data is recorded on 30-inchlong loops of standard computer-grade magnetic tape held in interchangeable cartridges.
Each tape loop is two inches wide and contains 112 recording tracks. Bit-serial recording
is used. at a density of 1.000 bits per inch. A single cartridge contains 16 tape loops and
can store up to 7.2 million characters. (A newer dual-cartridge RAM unit stores up to 3.6
million characters in each of two 8-100p cartridges.)
Vacuum capstans and "air bearings" are used to reduce wear and contamination of the tape.
Any tape loop not engaged in a data transfer process remains stationary and is drawn away
from both the drive capstan and the read/write heads. Seven reading heads and seven
writing heads serve each of the RAM tape loops. All of the heads move in unison to any
one of 16 discrete positions. Average head positioning time is 62.5 milliseconds. and
average rotational delay is 25 milliseconds. Data transfer rate is 86.000 characters per
second .
.3

THE ECONOMICS
The economics of using random access devices involves conSiderably more than simply
comparing their cost with that of magnetic tape transports. To achieve any sort of valid
economic measurement. it is necessary to make a comparison between the two fundamentally different methods of processing: on-line and batch. On-line processing implies that
all transactions are processed in essentially the order in which they are presented to the
data processing system. so random access to the stored files is a prerequisite. In the
more conventional batch processing approach. the transaction data must be arranged in
the same sequence as the master file before processing. The major economic differences
between the two methods can be determined by comparing their access times. storage
costs. and overall throughput costs .

. 31

Access Times
Comparing the access times of on-line and batch proceSSing really necessitates a comparison between the access time of the random access device and the times for the transactionfile sorting required for batch processing. Once the transaction record is matched or
merged with the master-file record in batch processing. the remaining processing time
required will be about the same as that required for the on-line processing operation. In
making such a comparison. keep in mind that in a well designed on-line system. most of
the access time can probably be overlapped with computer processing; only the non-overlapped access time needs to be measured against the sorting time for the batch processing
case.
These timing factors will vary with the file size. record size. computer system configuration. and type of random access device used. Each case will therefore need to be conSidered separately. and no generalized conclusion can be drawn.
© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

4/68

23:060.320

• 32

AUERBACH STANDARD EDP REPORTS

Storage Costs
Here we must consider the costs of: (1) the storage units themselves. (2) all control units
necessary to connect the storage units to the central processor. and (3) the storage media
(cartridges. tape reels. etc.) required to hold all of the necessary information. both online and off-line. USing currently available equipment. disc files (and large-capacity drum
files) tend to compare favorably in cost with cartridge and tape units for storage requirements of up to around 100 million characters. For storing files of over one billion characters. they tend to become unwieldy because of the large number of physical units required
and their space and maintenance requirements.
When properly used. the best cartridge units can offer significant advantages in storage
cost over both magnetic tape and disc units for storage requirements up to several billion
characters. When total storage requirements exceed this level, tape systems are unmistakably the cheapest. due to the very low cost of the tape itself.
On the basis of relative cost. it would seem that a combination of both serial and random
access storage is likely to become standard practice in many of the EDP installations of
the future. Discs. drums. or future nonmechanical random access stores would be used
for smaller files of up to about 100 million characters. and magnetic tape would be used
for the very large-volume files. Normally. the more active records would be held in random access storage for faster accessibility. while the rest would be stored on magnetic
tape for economy .

. 33

Throughput Costs
In determining the effect that random access storage will have on the number of transactions
your EDP system can process per dollar. you are getting to the crux of whether or not random access storage is practical for your own particular installation. In attempting to make
this decision. you must begin considering some of the broader systems implications of using
random access storage.
It is obvious that a well-designed on-line system is greatly superior to a batch type system
with respect to the total response time required to process a given transaction and update
the necessary files. The advantage might be as much as seconds versus hours or even days.
However. in order to handi., high peak loads without excessive delays. an on-line system
may require significantly more throughput capacity (computer power) than a batch-type system designed to handle the same total workload.

With currently available computer hardware. a system configuration designed for efficient
batch processing generally will be able to process more records per day at a lower cost than
a corresponding random access configuration of the same computer system. This is due not
only to the cost of the random access units themselves. but also to the added core storage
and communications equipment that is usually required for on-line processing.
On the basis of the number of transactions processed per dollar. therefore. batch processing
usually shows a significant advantage over on-line processing with currently available equipment. This advantage may be more than offset. however. by a number of system performance considerations centered around a significant expansion of the data processing system's
utility to the organization .

.4

SYSTEMS CONSID ERA TIONS
The use of random access storage can rarely be justified solely on the basis of the economic
comparisons described above. The user must ultimately decide whether an on-line system
will provide enough added advantages over a batch-type system to justify the added expense.
These advantages take the form of faster response. more timely management information.
and the economies of integrated operations .

. 41

Faster Response
On-line random access files can. of course. provide immediate responses to requests for
information. Because data can be entered into the system on a random basis and filed immediately. as contrasted with the batch processing techniques used in magnetic tape systems.
answers to queries are not only rapid but based on completely up-to-the-minute information.
In cases where different types of data must be supplied to a system user. data retrieval can
usually be accomplished in one pass. whereas a batch processing system might require a
number of separate passes through the different files. The more diverse the data requirements of an organization and the greater the need for up-to-date information, the more
practical an on-line system becomes .

. 42

Timely Management Information
The on-line system's ability to respond quickly to diverse queries with up-to-date information is extremely attractive to management. Not only can the system provide the type of
information needed to tighten the administration and control of operations. but it can provide more pertinent inputs to the management decision-making process.

4/68

fA.

AUERBACH

'"

(Contd. )

SPECIAL REPORT

.42

23:060.420

Timely Management Information (Contd.)
The ability of an on-line system to process transactions as they occur also simplifies the
scheduling problems within the computer facility. Tradeoffs no longer need to be made between regular daily tasks and the occasional tasks such as end-of-month closings and weekly
reports. This tends to reduce peak-load buildups and even out the data processing workload
so that more consistent and efficient use is made of the computing equipment .

. 43

Integrated Operations
Mass random access storage devices are a vital element in the development of modern information systems. By permitting rapid access to all of the pertinent information in the organization's files. random access devices open the door to a total systems concept in which each
individual transaction can immediately trigger the appropriate entries in all of the affected
files. For example. a single sales order might cause changes in inventory, production scheduling, material control, dispatching, billing, accounts receivable, credit, commission, and
other records. Integrated systems will make it possible for large modern corporations to
enjoy the same degree of centralized control and flexibility of operation as small single-proprietor businesses .

.5

THE COMPARISON CHART
The accompanying comparison chart summarizes the Significant characteristics of 26 random
access storage devices. These devices are representative of the equipment currently offered
by the major manufacturers of general-purpose computer systems. Though there are numerous independent suppliers of random access storage units. the devices marketed by the
main-frame manufacturers are believed to be of greater interest to most users of this reference service. More information about these and other random access storage units can,
of course. be found in Volumes 2 through 8 of AUERBACH Standard EDP Reports.
The chart entries have been selected to pinpoint specific advantages or disadvantages of each
device from a user's point of view. An explanation of the meaning and significance of each
comparison chart entry follows.
•

Category - The storage devices included in this chart can be grouped into three
major categories: Magnetic Drums, Magnetic Disc Files. and Cartridge-Loaded
Units (in which the storage medium is conveniently replaceable).

•

Device - Identifies each device by manufacturer. model number. and the name by
which it is commonly known.

•

Representative Computer System - It is difficult (if not meaningless) to evaluate
a random access storage device independently of the computer system to which it
is connected. A single. representative computer system has been selected to
serve as a basis for all the comparison chart entries for each storage device.
The capacity and performance characteristics of some storage devices can be significantly different when they are associated with other computer systems.
Report Reference - Indicates the section where you can find a detailed description of each device in the full. 8- or lO-volume edition of AUERBACH Standard
EDP Reports.

•

•

Storage Medium - The physical medium upon which data is 'recorded.

•

Storage Capacity - The five entries in this general category define data storage
capacity in terms of:
(1)

The number of data discs or drums per physical unit of random access storage (often a variable quantity. in which case the range is indicated).

(2)

The number of logical tracks on each disc surface or drum upon which data
can be recorded. Where a "band", or logical track, is composed of two or
more parallel tracks which are always read and recorded at the same time,
the fact is clearly indicated.

(3)

The maximum number of alphanumeric characters that can be recorded on
a single logical track.

(4)

The maximum number of alphanumeric characters that can be read or recorded without any repOSitioning of the read/write heads (i. e., the "cylinder"
capacity).

(5)

The maximum number of alphanumeric characters (usually six or eight bits
per character) that can be stored in each physical unit of random access storage. The characters are assumed to be in the code and format most commonly used to represent alphanumeric information in the particular system.
It should be noted that in many random access devices, the number of decimal
digits of all-numeric information that can be stored is substantially higher
than the number of alphanumeric characters. For example, in most systems
that use the 8-bit byte representation, two decimal digits can be "packed" into
each byte.
© 1968 AUERBACH Corporation and AUERBACH Info. Inc.

4/68

AUERBACH STANDARD EDP REPORTS

23:060.500

.5

THE COMPARISON CHART (Contd.)
•

Head Positioning Time - For storage devices with movable read/write heads. the
time required to reposition these heads is rt!ported in terms of:
(1)

The minimum time required to move the heads to the next adjacent track
position.

(2)

The average time required to position the heads to read a randomly-placed
record.

(3)

The maximum (worst-case) positioning time.

For the cartridge units that use magnetic cards. the indicated "head positioning
times" actually represent the times required to withdraw a card from the cartridge
and position it on the read/write drum.

4/68

•

Average Rotational Delay - The average time (in milliseconds) required for the
first character of the selected data record to reach the read/write heads after the
heads have been properly positioned (usually one-half revolution in the case of
magnetic disc and drum storage devices). The total average access time for a
randomly-placed record is. of course, the sum of "Average Head Positioning
Time" and "Average Rotational Delay".

•

Peak Data Transfer Rate - The maximum rate at which data is read from or recorded upon the random access storage medium after the desired record has been
located. expressed in characters per second. When large blocks of data must be
read from or recorded in consecutive storage locations. the overall effective data
transfer rate. in some cases. will be significantly lower than the peak rate. due
to rotational delays between records and/or the need for repositioning.

•

Update Cycle Rate - The maximum number of records per second that can be accessed from random storage locations. read into the computer's main storage.
updated. rewritten into the same storage locations. and checked for correct recording. The records must be at least 100 characters in length. All records are
in random locations scattered evenly throughout the storage unit. and no batching
of transactions or overlapping of seek times on multiple storage units is permitted.
This is a useful, standardized measure of a random access storage device's performance in a straightforward on-line file maintenance application.

•

Read-Only Reference Cycle Rate - The maximum number of records per second
that can be accessed from random storage locations and read into the computer's
main storage. In this case, no updating or rewriting is required. All other conditions are the same as for the "Update Cycle Rate" above. This figure measures
a random access storage device's performance in simple inquiry/response applications where no file updating is required.

•

Transfer Load Size - The number of alphanumeric characters that can be transferred to or from the random access storage device in a single read or write
operation. The load size is fixed in some cases and variable in others.

•

Read/Write Checking - The type of checking performed upon the accuracy of data
recording and/or reading. The most commonly employed method is to generate
and record a parity bit for each character. word. or record. and to check the
recorded data for correct parity when it is reread. "Check characters" usually
implies a similar but somewhat more powerful system for detecting errors (and.
in some cases, correcting them). "Read after write" parity checking or separate
(and time-consuming) "write check" operations permit detection of most recording
errors at the time of occurrence - a highly desirable feature.

•

Representative Cost - To complete the picture. a purchase cost figure. expressed
in dollars per character. is listed for each type of random access storage. This
cost is based upon the price of a single physical storage unit of the largest available
capacity. together with any control units that are required to connect it to the specific computer system shown in the chart. (The costs of general-purpose computer
data channels and multiplexors are not included.) It is important to note that the
cost per character may vary significantly when the device is associated with a different computer system, or when more or less storage capacity is required.

fA..

AUERBACH

(Contd. )

SPECIAL REPORT

23:060.900

COMPARISON CHART
CHARACTERISTICS OF RANDOM ACCESS STORAGE DEVICES

@

1968 AUERBACH Corporation and AUERBACH Info, Inc.

4/68

AUERBACH STANDARD EDP REPORTS

23:0&0.901

CHARACTERISTICS OF RANDOM ACCESS STORAGE DEVICES
MAGNETIC DRUMS

C....ory

Device

Control Data 863
Drum Ston,. Unit

IBM 2301 Drum Stora. .

IBM 2303 Drum storage

Repreaentative Computer System

CDC 3300

IBM Syatem/360

IBM Syatem/360

Report RefereDOe

250:042

420:043

420:045

Storace Medium

Drum

Drum

Drum

D.ta Discs or Drums
per Physicsl Ulllt

1

1

1

Data Tracks per Disc
Surface or Drum

64 banda;
13 tracks/band

200 bands;
4 tracks/1xu1d

800

Maximum Characters
per Track

65. 536/band

20. 483/band

4,892

Mall:imum Characters
Accessible Without
Head Repositioning

Total capacity

T~l

Total capacity

Mall:imum Characters
per Physical Unit

4.194.304

4,096.600

3,910.000

Minimum

0

0

0

Average (Random)

0

0

0

Maximum

0

0

0

Average Rotatiooal Delay, Msec

16.7

8.6

8.6

Peak Data Traoafer Rate,
Characters per SecoDd

2.000,000

1,200,000

312,500

Traoafer Load Size, Characters

2 to 4,194,304

1 to 20,483

1 to 4,892

Update Cycle Rate,
References per Second

U.8

22.7

16.5

Read-Only Reference Cycle Rate,
RefereDOes per Second

57.2

116

100

Read/Write CbeckiDg

Parity

Cyclic cbeck cbaracters

CycliC check cbaracters

Representative Cost,
Dollars per Character Stored

0.045

0.050

0.039

Features aDd Commeata

Fixed heads; Interlacing
permits slower data
rates where desirable

Fixed heads; variable
record lengths

Fixed heads; variable
reconllengths

Storage
Capacity

Head
Positioning
Time.
Milliseconds

4/68

fA•

AUERBACH

caP!lcity

(Contd.)

23:060.902

SPECIAL REPORT

CHARACTERISTICS OF RANDOM ACCESS STORAGE DEVICES
MAGNETIC DRUMS

Catet:0ry

Device

UNIVAC FH-432
MaRlletic Drum

UNIVAC FH-1782
Magnetic Drum

UNIVAC Fastrand II
Mass Storage

Representative Computer System

UNIVAC 1108

UNIVAC 1108

UNIVAC 1108

Report Reference

785:042

785:043

785:044

Storage Medium

Drum

Drum

Drums

Data Discs or Drums
per Physical Unit

1

1

2

Data Tracks per Disc
Surface or Drum

128 bands;
3 tracks/band

256 bands;
6 tracks/band

6.144

Maximum Characters
per Track

12. 288/band

49. 152/band

10.752

Maximum Characters
Accessible Without
Head Repositioning

Total capacity

Total capacity

688.128 or 946. 176'

Maximum Characters
per Physical Unit

1.572.864

12.582.912

132.120.576 or
132.358.624'

Minimum

0

0

30

Average (Random)

0

0

58

Maximum

0

0

86

Average Rotational Delay. Msec

4.25

17.0

35

Peak Data Transfer Rate.
Characters per Second

1.440.000

1.440.000

153.750

Transfer Load Size. Characters

\I tQ

6 to 393.216

6t0393.216

Update Cycle Rate.
Referenccs per Second

46.3

11.8

4.1

Read-Only Reference Cycle Rate.
References per Second

217

58.1

10.6

Read/Write Checking

Parity, char. count

Parity. char. count

Check characters

Representative Cost.
Dollars per Character Stored

0.079

0.016

0.0016

Features and Comments

Fixed headlil; drum
search capabiUty

Fixed heads; drum
search capability

Movable access mechanism has 64 read/write
heads.
• With Fastband optional
feature.

Storage
Capacity

Head
Posltloning
Time.
MIlliseconds

393,2~6

© 1968 AUERBACH Corporation and AUERBACH Info. Inc.

4/68

23:060.903

AUERBACH STANDARD EDP REPORTS

CHARACTERISTICS OF RANDOM ACCESS STORAGE DEVICES
DISC FILES (NONREMOVABLE DISCS)

Category

Device

Burroughs
B 9370-2
System Memory

Burroughs
B 9372 Modular
Random Storage

Control Data
Control Data
6638 Disc System 814 Disc File

Representative Computer System

B 2500/3500

B 2500/3500

CDC 6600

CDC 3300

Report Reference

210:042

210:043

260:045

250:043

Storage Medium

Disc

Discs

Discs

Discs

Data Discs or Drums
per Phyaical Unit

1

4/module.
20/benk

72

72

Data Tracu per Disc
Surface or Drum

100

150

192

192

Maximum Characters
per Track

10.000

9.600

81.920 per band
of 12 tracks

8.192

Maximum Characters
Accessible Without
Head Repositioning

Total capacity

Total capacity

5.250.000

4.194.000

Maximum Characters
per Physical Unit

2.000.000

50.000. 000 per
bank

168.000.000

201.326.592

Minimum

0

0

34

34

Average (Random)

0

0

70

75

Maximum

0

0

100

110

Average Rotational Delay. Msec

17

20

26

25

Peak Data Transfer Rate.
Characters per Second

291.000

240.000

1.680.000

196.700

Transfer Load Size. Characters

100 to 10.000

100 to-10.000

5.120

up to 2. 1 million

Update Cycle Rate.
References per Second

11.8

10.0

5.0

5.2

Read-Only Reference Cycle Rate.
References per Second

57.6

49.0

10.4

10.9

Read/Write Checking

Check chars ••
write check

Check chars ••
write check

Parity

Parity

Representative Cost.
Dollars per Character Stored

0.014

0.0048

0.0019

0.0013

Features and Comments

Fixed heads. 1
per track

Fixed heads. 1
per track

horizontallyopposed access
"combs"; records
12 tracks. on 12
disc surfaces. In
parallel

horlzontallyopposed access
"combs"; records
In bit-serial mode

Storage
Capacity

Head
Positioning
Time.
Milliseconds

4/68

A

AUERBACH

•

,

/

(Contd. )

23:060.904

SPECIAL REPORT

CHARACTERISTICS OF RANDOM ACCESS STORAGE DEVICES
DISC FILES (NONREMOVABLE DISCS)

Category

Device

GE DSU204 Disc
Storage Unit

Honeywell 262
Disc File

IBM 2302 Disc
Storage

IfIM 1405 Disc
Storage

Representative Computer System

GE 400 Series

Honeywell 200

IBM System/360

IBM 1401

Report Reference

330:042

510:044

420:044

401:042

Storage Medium

Discs

Discs

Discs

Discs

Data Discs or Drums
per Physical Unit

4 to 16

64

25 or 50

25 or 50

Data Tracks per Disc
Surface or Drum

256

256

500

200

Maximum Characters
per Track

3,840

9,216

4,984

1,000

Maximum Characters
Accessible Without
Head Repositioning

368,640

4,700,000

897,120

2,000 (4,000
with optional 2nd
access arm

Maximum Characters
per Physical Unit

23,592,000

300,000,000

224,280,000

20,000,000

Mimmum

70

15

50

90

Average (Random)

199

78

165

600

MaXimum

305

120

180

800

Average Rotational Delay, Msee

26

26

17

25

Peak Data Transfer Rate,
Characters per Second

75,200

189,000

156,000

22,500

Transfer Load Size, Characters

240 to 7,680

1 to 1,179,648

1 to 224,280

200 or 1,000

Update Cycle Rate,
References per Second

3.0

4.8

4.0

1.4

Read-Only Reference Cycle Rate,
Referenccs per Second

4.4

9.6

5.4

1.6

Read/Write Checking

Check chars.

Cyclic check
characters

Cyclic check
characters

Parity, write
check

Representative Cost,
Dollars per Character Stored

0.0061

0.0010

0.0017

0.0030

Features and Comments

Inell vidually

Two independent
access "combs";
4 read/write
heads serve
each disc face

Two access
"combs" serve
250 track positions each; no
longer in production

Single access
arm servcs all
discs; no longer
in production

Storage
Capacity

Head
PositIOning
Time,
Milliseconds

posiUonable
access arm
serVeS each

disc

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

4/68

AUERBACH STANDARD EDP REPORTS

23:060.905

CHARACTERISTICS OF RANDOM ACCESS STORAGE DEVICES
CARTRIDGE-LOADED UNITS

Category

Honeywell 259
Disc Pack Drive

Storage Drive

GE DSUl60
Removable Disc
Storage Unit

CDC 3300

CDC 3300

GE 400 Series

Honeywell 200

Report Reference

250:044

250:044

330:043

510:042

Storage Medium

Magnetic discs

Magnettc discs

Magnetic discs

Magnetic discs

Data Discs or Drums
per Physical Unit

6 (10 sides)

6 (10 sides)

6 (10 sides)

6 (10 sides)

Data Tracks per Disc
Surface or Drum

100

200

200

200

Maximum Characters
per Track

2.000 or 2.980

4.096

3.840

4.602

Maximum Characters
Accessible Without
Head Repositioning

20.000 or 29.800

40.960

38.400

46.020

Maximum Characters
per Physical Unit

2.000.000 or
2.980.000

8.192.000

7.680.000

9.204.000

Minimum

30

30

30

30

Average (Random)

85

85

85

80

Maximum

145

145

165

150

Average Rotational Delay. Msec

20

12.5

12.5

12.5

Peak Data Transfer Rate.
Characters per Second

77.730

208.000

208;000

208.000

Transfer Load Size. Characters

4 to 29.800

4 to 40.960

384 to 38.400

1 to 46.020

Update Cycle Rate.
Referenccs per Second

6.3

8.3

7.0

7.0

Read-Only Reference Cycle Rate.
References per Second

12.7

14.1

11.0

10.8

Read/Write Checking

Parity. write
check

Cyclic check
characters

Check chars.

Cyclic check
characters

Representative Cost.
Dollars per Character Stored

0.013

0.0053

0.0065

0.0053

Features and Comments

Changeable Disc
Pack cartridges;
format is compatible with IBM

Changeable Disc
Pack cartridges;
format is not
IBM-compatible

Changeable Disc
Pack cartridges;
format is not
IBM -compatible

Changeable Disc
Pack cartridges;
format is not IBMcompatible; variable record lengths

Control Data
852 Disc

Control Data
854 Disc

Storage Drive
Representative Computer System

Device

Storage
Capacity

Head
POSitioning
Time.
Milliseconds

1311

('
\,

4/68

A•

AUERBACH

(Contd. )

23:060.906

SPECIAL REPORT

CHARACTERISTICS OF RANDOM ACCESS STORAGE DEVICES
CARTRIDGE-LOADED UNITS

Category

Device

IBM 1311 Disc
Storage Drive

IBM 2311 Disc
Storage Drive.
Modell

IBM 2314 Direct
Access Storage
Facility

IBM 2321 Data
Cell Drive

Representative Computer System

IBM 1401

IBM System/360

IBM System/360

IBM System/360

Report Reference

401:043

420:046

420:048

420:049

Storage Medium

Magnetic discs

Magnetic discs

Magnetic discs

Magnetic strips

Data Discs 01' Drums
pel' Physical Unit

6 (10 sides)

6 (10 sides)

Eight ll-disc
packs on-line

Ten 200-strip
cartridges

Data Tracks per Disc
Surface or Drum

100

200

200

100 per strip

Maximum Characters
per Track

2.000 or 2.980·

3.625

7.294

2.000

Maximum Characters
Accessible Without
Head RepositIOning

20,000 or
29,800·

36.250

1.167.040
(145, 880 per
pack)

40,000

MaXimum CharactcI's
per Physical Unit

2,000.000 or
2.980.000'

7,250.000

233.408.000

400.0:10,000

l\llmmum

75 or 54'

30

25

375

Average (Random)

250 or 154'

75

75

550

MaXimum

392 or 248'

135

135

600

Average Rotational Delay. MseC'

20

12.5

12.5

25

Peak Data Transfer Rate.
Characters per Second

77.000

156.000

312.000

54.800

Transfer Load Size. Characters

100 to 20.000

1 to 36.250

1 to 145.880

1 to 40.000

Update Cycle Rate.
References per Second

2.801'3.9"

7.0

7.0

1.5

Read-Only Reference Cycle Rate,
References per Second

3.701'5.6"

11.1

11.1

1.8

Read/Wl'lte Checking

Pa rity. write
check

Cyclic check
characters

Cyclic check
characters

Cyclic check
characters

Representative Cost.
Dollars per Character Stored

0.023 or
0.016'

0.0072

0.0010

0.00041

Features and Comments

Changeable Disc
Pack cartridges;
no longer in productlon.
• With optional
feature

Changeable Disc
Pack cartridges;
variable record
lengths

Changeable Disc
Packs; each 2314
has 9 disc drives
(8 on-line and 1
spare)

Changeable Data
Cells hold 200
strips each; 10
cells on-line; 20
movable read!
write heads

Storage
Capacity

Head
POSitIOning
Time,
~1l11lseconds

C 1968 AUERBACH Corporation and AUERBACH Info. Inc.

4/68

AUERBA¢H STANDARD EDP REPORTS

23;060.907

CHARACTERISTICS OF RANDOM ACCESS STORAGE DEVICES
CARTRIDGE-LOADED UNITS

Category

Device

NCR 655-201
Disc Unit

NCR 653-101
CRAM Unit

RCA 70/568-11
Mass Storage
Unit

UNIVAC 8410 Disc
storage System

Representative Computer System

NCR Century 200

NCR Century 200

RCA Spectra 70

UNIVAC 9300

Report Reference

620:011

620:011

710:043

810:042

Storage Medium

Magnetic discs

Magnetic cards

Magnetic cards

Magnetic disc

Eight 256-card
cartridges

Two 1-disc
cartridges

Dats Discs or Drums
per Physical Unit

Two 3-disc packs One 384-card
cartridge
on-line

Data Tracks per Disc
Surface or Drum

192

144 per card

128 per card

100 bands:
2 tracks/band

Maximum Characters
per Track

4.096

2.250

2.048

16.000/band

Maximum Characters
AcceSSible Without
Head Repositioning

524.288
(262. 144 per
pack)

81.000

16.384

32.000
(16.000 per
cartridge)

Maximum Characters
per Physical Unit

8.388.608

124.416.000

536.870.912

3.200.000

Minimum

20

100

439

?

Average (Random)

42

125

488

110

Maximum

60

150

538

245

Average Rotational Delay. Msec

20.8

24

30

25

Peak Data Transfer Rate.
Characters per Second

108.000

71. 250

70.000

136.000

Transfer Load Size. Characters

1 to 4.096

1 to 2.250

1 to 2.048

160

Update Cycle Rate.
References per Second

6.6

5.0

1.6

3.8

Read-Only Reference Cycle Rate.
References per Second

14.7

5.0

1.8

7.2

Read/Write Checking

Parity. write
check

Check chars ••
read after write

Checks chars .•
read after write

Parity. write
check

Representative Cost.
Dollars per Character Stored

0.0048

0.00060

0.00029

0.0067

Features and Comments

Changeable 3disc packs are
not IBM-compattble; each unit
has 2 disc drives

Changeable CRAM
decks hold 384
cards each: 36
movable read!
write heads

Changeable
cartridges hold
256 cards each;
8 cartridges online: 8 movable
heads

Each disc face
holds 1. 600. 000
bytes: only 1 fac e
is accessible at a
time: each unit has
2 disc drives

Storage
Capacity

Head
Positioning
Time.
Milliseconds

4/68

fA.

AUERBACH

-~

23:070.001
II ......

;Q
'u(IIIIAC~

•

EDP

SPECIAL REP'ORT
DIGITAL PLOTTERS

tUml

SPECIAL REPORT
DIGITAL PLOTTERS:
A STATE-OF-THE-ART REPORT
by
the Technical Staff of
AUERBACH Info, Inc.

C 1969 AUERBACH Corporation and AUERBACH Info. Inc.

6/69

23:070.002
SPECIAL REPORT
DIGITAL PLOTTERS

CONTENTS

1.

INTRODUCTION

2.

TYPES OF DIGITAL PLOTTERS

3.

USE CONFIGURATIONS

4.

REPRESENTATIVE SYSTEMS

4.1

4.2
4.3

6/69

Calcomp Model 750 Magnetic Tape Plotting System
Computer Industries, Inc. Magnetic Tape Delta Incremental Plotter
Concord Control, Inc. Coordinatograph

5.

ILLUSTRA TIONS

6.

COMPARISON CHARTS

A

AUERBACH

•

23:070.100

A

AUERBACH

llnOUD

EDP

SPECIAL REPORT
DIGITAL PLOTTERS

U,o.1S

DIGITAL PLOTTERS:
A STATE-OF-THE-ART REPORT
.1

INTRODUCTION
Graphic rpcol'ders have been a principal output device in analog computing systems for a numbC'r of .\'ear8. The rapid increase in the use of these devices in digital systems, however, is
l't'lativpl.\' l'ecent, and present trends indicate a widening range of applications for plotting eqUipment in both the scientific and nonscientific fields. The chief value of a plotter is that large
nmOlUlts of data can be reduced and converted to graphical form for easier study and interpretation. This type of output has proved valuable in stich applications as the plotting of missile trajectories and orbits, the checking and comparing of engineering design calculations, the speeding
lip of the final analysis of scientific evaluation studies. and the automatic plotting of weather
maps.
In the nonscientific areas, plotters are being used to generate sales, inventory, and production
charts that give management a graphic tcol to help forecast future trends. Other uses include
the checking and charting of automatic machine-tool performance, the production of traffic
density pattern data for computer -controlled highway stUdies, and the plotting of earth-moving
and rill problems which are more easily dealt with in graphical form.
The term electromechanical plotters covers virtually all graphic recording devices from continuous strip recorders to digital X/Y plotters. Continuous strip recorders are used basically
for monitoring purposes, such as room temperature and humidity, and for patient monitoring
in biomedical applications. Basically, these devices consist of a paper carriage that transports paper at a constant rate of speed under the recording stylus. The recording stylus is in
contact with the paper at all times and records information by back and forth movement. A
continuous strip recorder may contain a number of styli, thus, Simultaneously recording multiple inputs.
A variation of strip recorders combined with electrostatic printing techniques exists in the
Varian STATOS V Printer/Plotter. This device contains 1024 electrostatic styli across a 12.8inch recording line. Recording is performed by energizing the styli as a dielectricly coated
paper passes underneath; thus, after developing, electrostatic printing of fine dots at a density
of 80 per inch is provided. The paper can be moved synchronously or asynchronously and the
styli can be used as a single set or as a number of continuous channels. This provides multichannel strip recording capability and with proper data structuring, graphic and printer type
output can be achieved.
A similar system, called the LGP-2000, using a laser beam and light sensitive film for recording has been developed by Dresser Systems, Inc., SIE Division. In this system, a laser
beam is swept across the light sensitive film; the beam is turned on at each spot position to be
intensified. Advance to the next line is accomplished by moving the film forward over a roller.
Dresser provides an extensive software package for use on IBM System/360 computers that
allows the user to describe a graphic image in normal graphic coordinate terms. This software package then builds the ordered list of plotter commands that will produce the desired
graphic output.
The first types of plotters developed were analog plotters, which accepted analog input signals
to control pen movement and were generally associated with analog devices. Most continuous
strip recorders are of this type. If an analog plotter is to be connected to a digital computer,
a special digital-to -analog interface is required to convert the digital outputs of the computer
to the analog signals required by the plotter.
Digital plotters were developed for direct connection with digital computers. This not only
eliminated the need for special digital-to-analog converters, but allowed use of the more
precise controls inherent in digital operations. Digital plotters can be used in more general
applications than their analog counterparts, just as digital computers are more flexible than
analog systems. Digital plotters eliminate the problems of drift, dynamic response, and
changing gain settings which are inherent in analog operations.

© 1969 AUERBACH CorporatIon and AUERBACH Info, Inc.

6/69

23:070,101

.1

SPECIAL REPORT

INTRODUCTION (Contd.)
TIl(' devices covered in this special report have been restricted to digital plotters, since this is
the t~l)e primarily used in association with digital computers. Presented in the discussion are
tlw (-(cller:!1 types of digital plotters currently available and a brief description of representative
~ystems. Illustrations of plotter devices and plotter outputs follow.· Finally, a set of compari~on charts including most currently available digital plotters is provided •

.~

1"{PES OF DIGITAL PLOTTERS
I\Iany different digital plotters are available and can be classified by the type ,of plotting surface
pl'ovidcd, the recording technique used, the method of line drawing, and the method of specifyln~ the desired plot. The plotting surfaces currently available are table and drum. Recording
techniques include pen and ink or electrical pulses; the latter recording on sensitized paper.
Line drawing methods arlf: increment, that is, small' generally one unit line segments; or
stroke, long line segments. Line specification methods are: absolute, stating the specific
coordinate to which a line is to be drawn; or relative, indicatmg simply the length and direction of a line.
Table-type plotters utilize a flat plotting surface ranging in size from 30 inches by 30 inches to
5 feet by 24 feet. The paper remains stationary throughout the plotting of a single graphic
image; the writing mechanism performs all necessary movements. The writing mechanism
consists of a carriage and pen assembly that moves along one axiS of the plotting surface; the
pen unit is also free to move along the other axis. Motion in the X or Y direction, or in both
directions simUltaneously, is thus obtained; and the pen can reach any coordinate value contained within the plotting area. The pen also moves up and down to allow both drawing and
positioning.
Generally, table-type plotters are pen and ink type utilizing a fixed coordinate system, which
allows absolute coordinate addressing and drawing straight line vectors of any length. The table
plotters are generally more versatile in that they can be built to meet requirements for high
precision or large size, and can incorporate many supplemental features such as interchangeable recording heads for inking, punching, or scribing, and the ability to plot more than one
curve at a time. All of these advantages are accompanied by proportionately higher costs. Comparative prices of digital plotters alone (no peripheral units included) range from $4,500 to
$22, 000 for the drum type, and from about $15,000 to $80,000 for the table type.
A table-type plotter utilizing an electrostatic recording method has been developed by Ford
Instrument Company of Long Island City, New York. This all-electronic plotting board is
capable of high speeds because all the limiting physical aspects of mechanical plotting systems
are eliminated. A sheet of sensitized paper is sandwiched between two X!Y conductive grids
made of fine wires. When the appropriate X and Y coordinates are chosen and the wires energized, a voltage potential exists at the crossing point of the two wires. The sensitized paper
reacts to this voltage potential to produce a mark at that one point. To produce the next point,
a new pair of grid lines is chosen and energized. At present the Ford Instrument system is
capable of handling about 50 points per second.
The second basic type of plotter, the drum plotter, uses a movable plotting surface in conjunction with a writing carriage to provide the required 2 -dimensional motion. In these units
the writing element moves along one axis while rotation of the drum supplies movement along
the other coordinate. At the present time, California Computer Products (Calcomp). Computer
Industries (formerly Benson Lehner Corporation), and Houston Instrument, Division of Baush
and Lomb. are the major manufacturers that build drum plotters. All of their units employ an
incremental plotting technique that produces a graph by a series of fixed incremental steps of
the drum and/or carriage. Bi-directional motors are used to control motion along both the
X and Y axes so that each input digital signal causes a small incremental step (1/100 inch or
less) of the carriage, the drum, or both. A third (Z-axis) input signal is ~sed to control the
raiSing and lowering of the pen from the surface of the paper.
Drum-type plotters are less expensive than table-type but have some disadvantages. In drumtype plotters each increment must be programmed so that many program steps are required to
produce a long line that can be drawn by one command in a table-type plotter. This not only
causes longer programs for drum -type plotters, but plotting is slower than for table-type.
Furthermore, drum-type plotters do not include an absolute coordinate reference system.
Digital plotters that record their output on film represent a new concept in digital plotting.
This new concept employs the same basic techniques as electromechanical plotters in that
commands from the computer are used to produce discrete incremental steps on the X and Y
axes. The new electronic recording technique uses the incremental plot commands to deflect
a cathode-ray tube (CRT) electron beam in discrete steps. The beam is blanked and unblanked
in place of raising and lowering the pen in an electromechanical plotter. The controlled electron beam from the CRT is used to expose a 35-millimeter film strip, which is advanced at

6/69

A.

AUERBACH

(Contd.)

DIGITAl. PLOTTERS

.2

23:070.200

TYPES OF DIGITAL PLOTTERS (Contd.)
the end of each plot. The exposed film can be automatically processed to produce either positive or negative transparencies for direct viewing or photographic printing .

•3

USE CONFIGURATIONS
The digital plotters on the market today are generally available for off-line use. Magnetic
tape, punched cards, punched paper ta.pe, or manual keyboard provide the input medium. In
most cases, the input can be a computer output specifically prepared for the plotter. This
allows the computer to operate at a higher speed than would be the case if the plotter were
on-line and also eliminates the problem of directly interfacing the plotter with the computer.
Most digital plotters marketed today are adaptable to on-line operation, and i.nterface controllers
are available for them for such widely used computer systems as the IBM System/360, Control
Data 3000 and 6000 Series, GE 400 and 600 Series, RCA Spectra 70, Univac 490 and 1108, and
Burroughs 5500, to mention a few. Most plotter manufacturers offer on-line operation as an
optional input mode and supply the required interface units for widely-used computers such as
the IBM System/360 or Univac 1108. In contrast to the on-Hne mode, all digital plotters on the
market today can operate off-Hne, using either punched cards, punched tape, or magnetic tape
as the chief source of input data. Off-Hne plotters using input from magnetic tape are particularly suitable for use with large, high-speed computers, because they make it unnecessary to
slow the computer down to the relatively low speed of the plotter. In general, most magnetictape handlers available with off-Une plotters are IBM 729 compatible with densities of 200, 556,
or 800 bits per inch.
Some plotter manufacturers offer software support for use with their plotters. For example,
Computer Industri.es, Inc. provides Fortran IV, IBM System/360, and Univac 1108 software
for its Magnetic Tape Delta Incremental Plotter.
Remote graphic output is a new application for digital plotters. A typical example of this is
the California Computer Products Model 210 Remote Plotter Controller. This device permits
any Calcomp 500 or 600 Series Plotter and a separate teleprinter to be connected to a communications Hne. With this configuration, a computer system can service a large number of
'plotter terminals at transmission speeds of up to 300 bits per second, and remote terminals
can interrogate the computer system for graphic information .

.4

REPRESENTATIVE SYSTEMS
Some examples of specific computer-controlled digital plotting systems are presented here to
illustrate the overall relationship between the computer output and the plotting operation. In
the first case, a drum-type incremental plotter manufactured by California Computer Products
is connected on-Hne to an IBM 1130 system. An IBM 1627 Plotter Attachment is used to interface the plotter with the 1130 Processing Unit. The cost of the Plotter Attachment is $675.
Total cost of the Plotter and Plotter Attachment can range from about $5.000 for 11-inch
plotters to $9,000 for 30-inch plotters.
As described earlier in this report, the principles of operation are the same for each of the
models of incremental plotters. The IBM 1130 BCD characters othrough 9 are the only ones
required to control plot operation. Each of the ten characters will cause a distinct plotter
movement, as depicted in Figure 1.
A single output instruction can shift the IBM 1130 to an output plotting operation. The instruction wtll initiate the plotting of one or more points, as controlled by the data stored in the first
locations of the output area. The plotting action is terminated upon receipt of a special character from the core storage of the 1130 Processor .

. 41

Calcomp Model 750 Magnetic Tape Plotting System
A good example of off-line operation is that of the Calcomp Magnetic Tape Plotting System
(Model 750) using tapes prepared by any computer system that employs standard IBM 7- or 9track tape written at densities of 200, 556, or 800 bits per inch. Here automatic plotting is
achieved by including all necessary plotter commands on binary-coded tapes that have been
prepared by appropri.ate computer subroutines.
Each plotter command consists of three bytes. The first byte specifies an incremental step in
the X direction, the second specifies a step in the Y direction, and the third specifies a pen-up
or pen-down command. Up to 93 plot commands can be recorded per inch of tape on the Model
750.
The block address of the data to be plotted can be manually preset and the tape automatically
searched for the required block. Tape is searched at 60 inches per second.
The cost of this Caicomp off-line system is $21,200, excluding the plotter. Any Calcomp 500
series plotter can be used with the 750 System.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

6/69

r~· ·
«

;

23:07.,.420

SPECIAL REPORT

Plotter Movement

'x. -y

,X

IX, IY

0
0

u
0

0
<)

IY

-x.

/

0
0
0
0

0

+x.

IY

0

-Y

0

Instruction Characters
1130 Character:
Plotter Operation:

0

1

2

3

4

5

6

7

8

9

Pen
Down

+Y

+Y.
+X

+X

-Y,
+X

-Y

-Y,
-X

-X

+Y.
-X

Pen
Up

Figure 1. Control Characters for Plot Operation

.42

Computer Industries Inc. Magnetic Tape Delta Incremental Plotter
Another example of off-line plotting is the Computer Industries Inc. Magnetic Tape Delta Incremental Plotter. This unit also uses tape prepared by any computer system that can write in
standard IBM 7-track or 9-track written at densities of 556 or 800 bits per second. With this
system, up to 127 steps in both the X and Y directions can be effected from a single plotter
command recorded at 800 bits per inch. The cost of this system is $27,000 •

. 43

Concord Control, Inc. Coordinatograph
A highly specialized type of off-Une plotter is the Concord Coordinatograph produced by Concord
Control Inc. The Coordinatograph System is designed to prepare final graphics quality copy for
cartographic purposes from data stored on magnetic or paper tape. The system consists of a
Director - a small, special-purpose digital computer - a paper-tape reader, a magnetic-tape
handler (optional), and a typewriter with paper-tape punch (optional). The Director accepts data
from the magnetic tape unit, paper tape reader, or keyboard; processes the accepted data; and
uses the processed data to control the Coordinatograph, which is a table plotter with a five
square-foot plotting surface. An accuracy of 0.005 percent can be maintained over the plotting
surface. A wide variety of instrument heads for scrtbing, printing, projecting, scanning, and
other uses can be installed in the Coordinatograph carriage •

.5

ILLUSTRATIONS
To conclude this Special Report on Digttal Plotters, a sample of the commerically avaUable
digital plotters and associated equipment provided by some of the larger manufacturers is presented. In addition. representative copies of digital-plotter graphic output are included to
illustrate the range of capabilities and job applications that this equipment can meet.

6/69

A

(Contd.)

AUERBACH
~

23:070,502

DIGITAL PLOTTERS
I~,

MAR

66

.:... ,'...... :.: ."
..
:
'",: . .
.:..
.: .. . .. :

Q

'

'. II' '., II·
' • • Cl

'

/j\

'

..

.. ot

......

" "

• • II ""

'

'

.

.....

: :u:
" ' ',
', :
'
It • • It

.. ..
.'
..
.-.....:: : . .

."
:

"

. .:

e:

'

'

.... ,

II.

\I.

"

oJ • • :

'
"

...

.: ..It::.. :-... : ....
. . : :: ..
'",.'
'

'
'

'

'

"

. .. .... :...
'

~,~~.
••••••••••••••••• Q ••

• • • • • • • • • • • • • • eO$ •••
22

Figure 6.

Example of a Backboard Circuit Layout Produced By a Gerber Scientific Plotter

Figure 7. Geometric Pattern Produced By an EIA Datapiotter (Time of Plot: 10 Minutes)
© 1969 AUERBACH Corporation and AUERBACH Info, Inc,

6/69

SPECIAL REPORT

23:070.600

...

'f-:'

\/.",

.~

..

(

.1

Figure 8. Example of Weather Contours Produced on an EIA Dataplotter (Time of Plot:
Three Minutes)
.6

THE COMPARISON CHARTS
The accompanying comparison charts summarize the significant characteristics of representative digital plotting devices. The entries have been selected to describe specific operational
criteria for each device from the user's point of view.

6/69

•

Type: almost all of the plotters included in this chart are of either the table or
drum type as described in the preceding paragraphs. Horizontal positioning of
the plotting table or drum is implied unless otherwise noted.

•

On-line Operation: this entry specifies whether or not a plotter can be connected
to a digital computer data channel for direct, on-line output. At the present
time, only a few computer manufacturers offer digital plotters as part of their
standard line of peripheral equipment, but most plotter manufacturers are prepared to supply interfaces that will adapt their equipment for on-line use with
most digital computers.

•

Input Devices Supplied: several plotters are marketed as integrated systems that
include a magnetic-tape transport, a card reader or a paper-tape reader as a
standard part of the equipment. Some also have rather elaborate operating panels
and provisions for manually entering data through a keyboard.

•

Input Medium: the physical medium on which the data to be plotted can be stored is
listed here. This is generally magnetic tape, paper tape, or punched cards. Facilities for manual input are also indicated here when they are provided.

A

AUERBACH

(Contd.)

DIGITAL PLOTTERS

23:070.500

Model 563 Incremental
Plotter With Model 760 Tape Drive

Model 565 Incremental Plotter

Model 502 Incremental Plotter

Figure 2.

Model 718 Incremental Plotter

California Computer Products, Inc. Digital Plotters

Figure 3. Gerber Scientific Instrument Companies Model 75 Graphic Display Table Shown
with the Series 600 Controller

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

6/69

SPECIAL REPORT

23:070.501

Figure 4. Milgo Electronic Corporation's DPS-6 Digital Plotting System Shown with the
Milgo Model 4066 GL Tape Drive (Ampex TM-707)

Figure 5. A Calcomp Plotter Reproduction of an 18th Century Japanese Woodblock Print

6/69

fA

,.

AUERBACH

(Contd.)

23:070.60!

DIGITAL PLOTTERS

.6

THE COMPARISON CHARTS (Contd.)

•

Input Code: the majority of plotters receive input data in pure binary or in some
binary-coded decimal (BCD) form, depending upon the type of input medium employed. For example, most table-type plotting boards require four decimal digits
to specify each coordinate value, since the matrix range usually extends from
-9999 to +9999. Some drum-type incremental units utilize three successive 2-bit
characters to specify three of the six possible operating movements (+X or -X,
+ Y or - Y. pen up or pen down) for each point. In the case of magnetic tape units,
all manufacturers state provisions for accepting data from IBM-compatible tapes
recorded at a density of 200 bits per inch. A few models also have the ability to
handle tape recorded at 556 or 800 bits per inch.

•

Chart Size: the actual plotting area available is stated in inches. Only the width
dimension is Hsted for drum-type units, since rolls of 120 feet are standard with
these plotters.

•

Plotting Mode: all plotters are capable of operating in a "point" mode in which
a single point is plotted for each pair of input coordinates. This is a relatively
simple operation for the table-type plotters, hut a series of commands (including the pen-up, pen-down control) usually must be given for each point to be
plotted by the drum type. An extension of the point mode is the "continuous"
mode, which yields Significantly higher curve-plotting speeds. However, the
input data must be supplied to the plotter as a continuous train of closely spaced
points. The incremental stepping of the drum units makes them particularly
well suited to this type of operati.on.

•

"Line" or "Line-Drawing" Mode: as defined for table plotters, use of this mode
results in the construction of a straight line between two consecutive pairs of
input data coordinate values. The drum-type plotters cannot operate in this
way, but they can produce lines of any desired length by plotting the required
number of incremental steps with the pen held in the down position (continuous
mode).

•

Accuracy: percentage figures are quoted for full-scale accuracy. For example,
if a plotter wi.th a 30-inch by SO-inch plotting surface has an accuracy figure of
0.05 percent, the plotter is capable of moving the pen to within 0.015 inch (0.05
percent of SO inches) of the true value of any specified coordinate. Where the
accuracy figures vary accordi.ng to the plotting mode, both figures are listed.

•

Speed: for the drum plotters, the speed is fixed for each model according to
the incremental step size. For table units, however, the speed can vary
greatly accordi.ng to the plotting mode and the maximum distance traveled
along either axis to move from one coordinate to the next. To keep the chart
as orderly as possible, all figures given in this column refer to maximum
speeds only, as rated by the manufacturers.

•

Symbol Printing: most plotters offer symbol pri.nting devices as optional equipment, which enable special symbols to be plotted instead of points. Alphanumeric character sets are also available with many plotters so that fully annotated
graphs can be produced to further identify and define the output data.

•

Purchase Price: this column supplies the approximate cost of the plotter but
does not include the cost of the off-line controller or optional i.nput devices.

•

Comments: this column i.s used to mention any additional facts about a particular unit that are unusual or of general interest.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

6/69

23:070.602

SPECIAL REPORT

COMPARISON CHART: DIGITAL PLOTTERS

['"iDL F,\l'rt'HER
I

Auto-trol Corporation
6221 W. 56th Avenue
Arvada, Colorado

I

t"n

Model 6000
Digital Plotter

Model 6030
Digital Plotter

Model 6300
Digital Plotter

!'\ p,'

Tabte

Table

Table

l):\-UNE OPERATION

Yes"

Yes'

Yes'

[, Pl'T IH:VICE SUPPLIED

Magnetic tape transport, (7track), card reader, paper tape
reader. or manual keyboard
optional

Magnetic tape transport, (7track), card reader, paper
tape reader, or manual keyboard optional

Magnetic tape transport, (7track). card reader, paper
tape reader, or manual keyboard optional

Magnetic tape; punched tape;
punched cards; keyboard

Magnetic tape; punched tape;
punched cards; keyboard

Magnetic tape; punched tape;
punched cards; keyboard

[~P[l

r MEDIUM

INPUT CODE 1'0 PLOTTER

BCD or binary

BCD or binary

BCD or bina ry

CI/.UlT SIZE

50 x 72 to 96 x 144 inches

40 x 60 inches

48 x 60 inches

PLOTTING MODE

Point; line

POint: line

POint; line

ACCliHACY (%)

0.025

0.03

0.01

SPf:LD (l\.lAX)

30 inches per second

10 inches per second

10 inches per second

SDIBOL PRINTING

384 characters: alphanumeric
and symbols

Numeric only

64 characters; alphanumeric
and symbols

35. 000 to 50, 000

12, 000

75. 000

I PI"HC/IASE

PfUCE, $

CtnJ:lIE:\TS

I

• Optional at additional cost.

(Contd.)
6/69

AUERBACH

'"

23:070.603

DIGIT AI.. PLOTTERS

COMPARISON CHART: DIGITAL PLOTTERS
MANli FACTtI ItER
California Computer ProducU, JJ\o.
3011 N, Muller Street
Anillheim. California
IDENTITY

TYPE
ON-UNE OPERATION
INPUT DEVICE SUPPUED

incremental Plotter

inoremental Plotter

Incremental Plotter

Model 502

Model 563

Model 565

Table

Drum

Drum

Yes·
Yes·
Yes·
500 series on-line options include a wide range of Interfaoe equipment; 500 series off-line
option. Include 3 different magnetic tape unit ••

INPUT MEDIUM

Magnetic tape; any digital
source

Magnetic tape; any digital
source

Magnetic tape; any digital
source

INPUT CODE TO PLOTTER

6 pulsed lines

6 pulled lines

6 pulsed lines

CHART SIZE

31 x 34 Inches

29. 5 x 120 Inches

11 x 120 Inches

PLOTTING MODE

Point; continuous

Point; continuous

POint; continuous

ACCURACY (%)

0.02

0.1

0.1

SPEED (MAX)

". 2 Inches per second

2.8 inches per second

4. 2 Inehes per second

SYMBOL PRINTING

No

No

No

PURCHASE PRICE, $

17,000

8,000

4,550

COMMENTS

JJ\cremental step al:e for Model 502 II 0,010 in•• 0.005 In., 0.002 In., 0.1 mm, and 0,5 mm;
Incremental step lIizell for Modele 563 and 565 are 0.010 in., 0.005 In, and 0.1 mm. Model 565,
when equipped with II oommunications Interface, 18 marketed as the Model 575 Remote Plotter;
the Model 575 can be connected to the publlc telephone network or a leased voice-band line via
a Bell System Data Set 20lA or 201B and sella for $5,863.

• Optional at additional co.t.

C 1969 AUERBACH Corporation and AUERBACH Info, Inc.

6/69

23:070.604

SPECIAL REPORT

COMPARISON CHART: DIGITAL PLOTTERS
MANllFACTIiRER
California Computer Products, Inc.
305 N. Muller Street
Anaheim, California
IDENTITY

Incremental Plotter

Incremental Plotter

Incremental Plotter

Incremental Plotter

Model 602

Model 618

Model 1>63

Model 665

rYPE

Table

Table

Drum

Drum

llN-UNE OPERATION

Yes'

Yes'

Yes'

Yes'

INPUT DEVICE SUPPLIED

Optional controllers for on-line operation: optional magnetic tape units for off-Une operations.

INPUT MEDIUM

Magnetic tape; any
digital source

Magnetic tape; any
digital source

Magnetic tape; any
digital source

Magnetic tape; any
digital source

INPUT CODE TO PLO'ITER

Option of 6 pulsed
lines or 5-bit binary

Option of 6 pulsed
lines or 5-bit binary

Option of 6 pulsed
lines or 5-blt binary

Option of 6 pulsed
lines or 5-blt binary

CHART SIZE

31 x 34 inches

54 x 72 Inches

29.5 x 120 Inches

11 " 120 Inches

PLOTTING MODE

Point; continuous;
line

Point; continuous;
line

Point; continuous

Point; continuous

ACCURACY (%)

0.1

0,1

0.1

0.1

SPEED (MAX)

3.1 inches per second

1. 4 inches per second

4.9 Inches per second

6.3 inches per second

SYMBOL PRINTING

No

No

No

No

Pl'RCHASE PRICE, $

24,000

40.000

15.000

11,275

COMMENTS

Full-step/half-step operation; Incremental step sizes are 0.005/0.0025 \D., 0.002/0.001 m.,
0.1/0.05 mm or 0.05/0.025 mm (Models 602 and 618) and 0.010 in/a. 005 in, 0.005 in/0.0025
in, 0.0025 infO. 00125 in, (Models 663 and 665); can use 500 or 700 Series input format.

• Optional at additional cost.

6/69

A

AUERBACH

'"

(Contd. )

DIGITAL PLOTTERS

23/070. 60~

COMPARISON CHART: DIGiTAL Pl.OTTERS
MANl1FACTlll\ER

California Computer Products, Inc.
305 N. Muller Street
Anaheim, California
Incremental Plotter

Incremental Plott~r

Inc remental Plotter

Incremental Plotter

Model 702

Model 718

Model 763

Model 765

TYPE

Tabll'

T2ble

Drum

Drum

ON-UNE OPERATION

Yes'

Yes"

Yes'

Yes'

IDENTITY

INPUT DEVICE SUPPUED

Optionsl controllers for on-line operation; optional magnetic tape units for off-line operation

INPUT MEDIUM

Magnetic tape; any
digital source

Magnetic tape; any
digital source

Magnetic tape; any
dlgit.al source

Magnetic tape; any
digital source

INPUT CODE TO PLOTI'ER

5-bit binary

5-blt binary

5-blt binary

5-blt binary

CHART SIZE

31 x 34 Inches

54 x 72 Inches

29.5 x 120 inches

11 x 120 Inches

PLOTTING MODE

Point; continuous;
line

Point; continuous;
line

Point; continuous

Point; continuous

ACCURACY (%)

0.02

0.01

0.1

0.1

SPEED (MAX)

11. 9 inches per second

4. 6 Inches per second

18. 2 inches per second

23.8 Inches per second

SYMBOL PRINTING

No

No

No

No

PURCHASE PRICE, $

31,000

50,000

22.000

18,0(l0

COMMENTS

Full-step!half-step operation; inc,emental step sizes are 0.005 In/O. 0025 In, 0.002 infO. 001
In, 0.1 mm/O.04 mm, and 0.05 mm/C.025 mm (Models 702 and 718) and 0.010 In/0.005 in,
0.005 In/O.0025 In, and 0.0025 In/O, 00125 In (ModelS 763 and 765).

• Optional at additional cost.

© 1969 AUERBACH Corporation and AUERBACH Inlo, Inc.

6/69

SPECIAL REPORT

23:070.606

COMPARISON CHART: DIGITAL PLOTTERS
MA N\' FAc'rl 1m: 1\

Computer Indultrl •• Inc.
(Formerly Benlon-Lehner Corporation)
Graphic Systems Dl.vlaton
14761 Callfa Street
Van Nuys, California 91401

llJENTITY

LTE System

STE System

DDS Drafting System

TYP"

Table

Table

Table

Yes"
Magnetic tape transpert

Yes"
Magnetic tape transpert

No
Magnetic tape transpert; paper
tape reader

rNPUT MEDIUM

Magnetic tape or punched
cards

Magnetic tape or punched
cards

Magnetic tape; paper tape

c--'

ON-UNE OPERATION
INPUT DEVICE SUPPLIED

INPUT CODE TO PLOTTER

6 pulsed lines

6 pulsed lines

6 pulsed lines

CIIART SIZE

42 x 58 Inches

30 x 30 Inches

5 x 5 feet; 5 x 8 feet;
5 x 12 feet; 5 x 16 feet;
5 x 24 feet

PLOTTING MODE

Point; line

Point; line

Point; line

ACCURACY (%)

0.05

0.05

SPEED (MAX)

300 lines per min @
1/4 inch per line

300 lines per min @

See Comments
12 inches per second

SYMBOL PIUNTING

48 characters

48 characters

None

Pt'RCIiASE PIUCE, $

29.000

22.800

no, 000 to

COMMENTS

Has three modes: pelnt. line. and free-run; pelnt mode
operates at 300 pelnts per Inch. Free-run mode can be used
for construction of contour maps •

1/4 Inch per line

144,000

Plotting error not greater than
± 0.002 Inch

• Optional nt additional cost.

6/69

A.

AUERBACH

'"

(Contd. )

I
DIGITAL PLOTTERS

23:070.607

COMPARISON CHART: DIGITAL PLOTTERS
MANll FAC'rUIIER

Computer Industries Inc.
(Formerly Benson-Lehner Corpora.tion)
Gra.phlc Systems Division
14761 CaUfa Street
Van Nuys. California 91401
Digital On- LIne Incremental Plotter

IDENTITY
Model 131

Model 135

Model 145

TYPE

Drum

Drum

Drum

ON-LINE OPERATION
INPUT DEVICE SUPPLlI!:D

YesNone

Yes·
None

Yes·
None

iNPUT MEDIUM

Any digital Input source

Any digital Input source

Any digital Input source

INPUT CODE TO PLOTTER

6 pulsed lines

6 pulsed lines

6 pulsed lines

CHART SIZE

12 Inches x 120 feet

12 Inches x 120 feet

12 Inches x 120 feet

PLOTTING MODE

Point; continuous

Point; continuous

Point; continuous

ACCURACY (%)

0.1

0.1

0.1

SPEED (MAX)

3 Inches per second

1.5 Inches per second

2 inches per second

SYMBOL PRINTiNG

No

No

No

PURCHASE PRICE. $

4,550

4.550

5,000

COMMENTS

IncrementallStep size Is 0.010 Inch per step (Model 131) and 0.005 Inch per step (Models
135 and 145); pen movement time Is 20 msec. (up) and 50 msec. (down).

• Optional at additional coat.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

6/69

SPECIAL REPORT

23:070.608

COMPARISON CHART: DIGITAL PLOTTERS
MANllFAc'rllRER

Computer Indulltries Inc.
(Formerly Benson-Lehner Corporation}
Graphic Systems Division
14761 CaUfa Street
Van Nuya, California 91401

IDENTITY

Digital On- Line lncrementaf Plotter
Model 321

Model 331

Model 335

Model 345

TYPE

Drum

Drum

Drum

Drum

ON-LINE OPERATION

Yes·

Yes·

Yell·

Yes·

INPUT DEVICE SUPPLIED

None

None

None

None

INPUT MEDIUM

Any digital Input

source

Any digital Input
source

Any digital Input
source

Any digital Input
eource

INPUT CODE TO PLOTTER

6 pulsed lines

6 pulsed lines

6 pulsed lines

6 pulsed lines

CHART SIZE

30. 5 inches x
100 feet

30.5 inches x
100 feet

30.5 Inches x
100 feet

30.5 Inches x
100 feet

PLOTTING MODE

Point; continuous

Point; continuous

Point; continuous

Point; continuous

ACCURACY (%)

0.1

0.1

0.1

0.1

SPEED (MAX)

2 Inches per second

3 Inches per second

1. 5 Inches per second

2 Inches per second

SYMBOL PRINTING

No

No

No

No

PURCHASE PIUCE. $

7,500

7,750

7,500

7,750

COMMENTS

Incremental step size i8 0.010 Inch per steP (Models 321 and 331) and 0.005 Inch per
step (Models 335 and 345); pen movement time Is 20 maee. (up) and 50 meee. (down).

• Optional at additional coet.

6/69

fA

AUERBACH

'"

(Contd.)

DIGIT AL PLOTTERS

23:070,609

COM ..ARISON CHART: DIGITAL PLOTTERS
MANI'.'AC'rl1IU:ll

Computer industries Inc.
(Formerly Benson-Lehner Corporation)
Graphic Systems Division
14761 Callfa Street
Van Nuys, California 91401
Card Input Delta Incremental Plotter

IIJEN'rrrV
Model Cm-131

Model CID-135

Model Cm-l4S

TYPE

Drum

Drum

Drum

ON-UNE OPERA'MON

Yes'

Yes'

Yes'

INPUT DEVICE SUPPUED

Card reader and Delta
Control Unit

Card reader and Delta
Control Unit

Cord reader and Delta
Control Unit

INPUT MEDIUM

Punched cards

Punched cards

Punched cards

INPUT CODE TO PLOTTER

6 pulsed lines

6 pulsed lines

6 pulsed lines

CIIART SIZE

11 inches wiele by 120 feet
long

11 inches wide by 120 feet
long

11 inches wide by 120 feet
long

PLOTTING MODE

Point: continuous

Point: continuous

Point; continuous

ACCURACY ('il)

0.1

0.1

0.1

SPEED (MAX)

3 Inches per second

1.5 Inches per second

2 Inches per second

SYMBOL PRINTING

No

No

No

PURCHASE PRICE, $

14,550

14,550

15,000

COMMENTS

Incremental step size Is 0.010 Inch per step (CID-131) and 0.005 inch per step (CID-135
and CID-145); pen movement time Is 20 ms (up) and 50 ms (down): up to 99 steps in x
and/or y from a single plotter command; up to 20 plotter commands per SO-column card.

, Optional at additional COlt,

© 1969 AUERBACH Corporation and AUERBACH Info. Inc,

6/69

23:070.610

SPECIAL REPORT

COMPARISON CHART: DIGITAL PLOTTERS
MANUFACTURER

COIIIP\IkIr IIIdllltrtll blc.
(Form.rly 1e1ll000-Lehllllr CorporattOll)
Graphic 8yJItem1 DlvtllOll
14781 Callfa Street
VIII NUYI, California 91401

IDENTITY

Card Jnput Delta Jnoremental Plotter
Model Cm-Sll

Model Cm-S31

Model Cm-SS5

Model Cm-S45

Drum

Drum

Drum

Drum

ON-UNE OPERATION

Yel·

y •••

y •••

Yel·

INPUT DEVICE SUPPUED

Card re&der and
Delta Control Unit

Card r.ader and
Delta Control Unit

Card reader and
Delta Control Unit

Card reader and
Delta Control Unit

INPUT MEDIUM

Punched card.

Punched card.

Punched card.

Punched card.

TYPE

INPUT CODE TO PLOTTER

6 pul.ed line.

6 pulled lin••

8 pulled line.

6 pul.ed line.

CHART SIZE

SO Inch•• wide by
100 feet 10111

30 Inche. wide by
100 feet lon,

SO Illch•• wide by
100 feet lone

30 Inche. wide by
100 feet lon,

PLOTTING MODE

Point: contlnUOUI

Point: continuoUi

Point: continuous

Point: colltinuoul

ACCURACY (%)

0.1

0.1

0.1

0.1

SPEED (MAX)

2 Inches per seeOlld

S lnebe. per leeOlld

1. 5 Inche. per .econd

2 Incbel per .eeOlld

SYMBOL PRINTING

No

No

No

No

PURCHASE PIUCE, $

17,500

17.750

17.500

17,750

COMMENTS

lDCremelltal.tep Ilze II 0.010 tllCh per ...p (Cm-3U and CIO-3S1) and 0.005 inch per ltap
(CIO-335 and CID-345): pell mOVllllellt time I. 201ll.ee. (Up) and 50 mllC. (dOWll); up to
99 .t.petll x and/or y from a 'lncle plotter oollllllllld: UP to 20 plotter cOllllllandS per
80-column card: ll-Inch paper .apter i ••tlllderd for SO-Inch plotter.

• Optional at additional colt.

6/69

A

(Contd.)

AUERBACH

•

23:070.611

DIGIT AL PLOTTERS

COMPARISON CHART: DIGITAL PLOTTERS
MANllFACTtlHEI\

Computer Indu.trlea Inc.
(Formarly Benlon-Lehner Corporation)
Gr~hlc By_tam. Dlviliton
14761 Callfa Street
Van Nuya, Callforni.. 91401

IDENTITY

Mapetlc

T~

Delta Incremental Plotter

Modal MTD-131-7 and
MTD-131-S

Model MTD-135-7 and
MTD-135-9

Model MTD-145-7 and
MTD-145-9

TYPE

Drum

Drum

Drum

ON-UNE OPERATION

Ye,·

Ye.·

Ye.·

INPUT DEVICE SUPPUED

7 - or 9-tr..ck mllpetlc tape
tun.port and Delt.. Control
Unit

7- or s-trACk mapetic t~
tranlport And Delt.. Control
Unit

7- or 9-tnck magnetlo ~
transport And Delt. Control
Unit

INPUT MEDIUM

Magnetic tape

Magnetic tape

Magnetic

~

INPUT CODE TO PLOTTER

6 pulled linea

S pul.ed linea

6 pulsed lice,

CH},RT SIZE

12 Inches x 120 feet

12 tnahel x 120 feet

12 Inches x 120 feet

PLOTTING MODE

Point; continuous

Point; continuous

Point; contlnuoue

ACCURACY (%)

O. 1

0.1

0.1

SPEED (MAX)

3 Inches per aeconcl

1. 5 inches per second

2 Inchel per second

SYMBOL PRINTING

No

No

No

PURCHASE PIUCE, $

27,000 (-7): 29,000 HI)

27,000 (-7); 29,000 (-9)

27,250 (-7); 29,250 (-9)

COMMENTS

Incremental _tel' 81ze la 0.010 Inch per Itep (MTD-131) and 0.005 Inch per step (MTD-135
and MTD-145): pen movement time I. 20 mlec. (up) and 50 maec. (down): 7-track tape
recorded &t 556 or 800 bpi; 9-track tape recorded at 800 bpi: unique t ..pe form ..t allows up
to 127 steps In x and/or y from .. alnlle command •

• Optional ..t additional COlt.

C 1969 AUERBACH Corporation and AUERBACH Info, Inc.

6/69

23:070.612

SPECIAL REPORT

COMPARISON CHART: DIGITAL PLOTTERS
MANl' FAC1'lJ ItER

Computer industries Inc.
(Formerly Benson-Lehner Corporation)
Graphic Systems DM.slan
14761 Callfa Street
Van Nuys, California 91401

IDENTITY

Magnetic Tape Delta Incremental Plotter
Model MTD-321-7
and MTD-321-9

Model MTD-331-7
and MTD-331-9

Model MTD-335-7
and MTD-335-9

Model MTD-345-7
and MTD-345-9

TYPE

Drum

Drum

Drum

Drum

ON-UNE OPERATION

Yes·

Yes·

Yes"

Yes·

INPUT DEVICE SUPPWED

7- or 9-track
magnetic tape
transport and
Delta Control
Unit

7- or 9-track
magnetic tape
transport and
Delta Control
Unit

7- or 9-track
magnetic tape
transport and
Delta Control
Unit

7- or 9-track
magnetic tape
transport and
Delta Control
Unit

lNPUT MEDWM

Magnetic tape

Magnetic tape

Magnetic tape

Magnetic tape

INPUT CODE TO PLOTTER

6 pulsed lines

6 pulsed lines

6 pulsed lines

6 pulsed lines

CHART SIZE

30,5 Inches x 100
feet

30.5 inches x 100
feet

30. 5 Inches x 100
feet

feet

Point; ccntlnuous

Point; continuous

Point; continuous

PoInt; continuous

0.1

0,1

0.1

0.1

2 Inches psr second

3 Inches psr second

1. 5 inches psr second

2 Inches per second

SYMBOL PRINTlNG

No

No

No

No

PURCHASE PRICE, $

29.000(-7);31,000(-9)

29,000(-7);31,000(-9)

29,000(-7) ;31,000(-9)

29 250(-7);31,250(-9)

PLOTTING MODE

ACCURACY

(%1

SPEED (MAX)

COMMENTS

30.5 Inches x 100

Incremental step size Is 0.010 Inch psr step (MTD-321 and MTD-331) and 0.005 inch psr
step (MTD-335 and MTD-345); psn movement time is 20 maee. (up) and 50 msec. (down);
7-track taps recorded at 556 or 800 bpi; 9-track taps recorded at 800 bpi; unique tape
format allows up to 127 steps In x and/or y from a single command; ll-inch paper
adapter Is standard for 30-tnch plotters.

• Optional at additional ccat.

6/69

A

(Contrl. )

AUERBACH

'"

DIGITAL PLOTTERS

23:070.613

COMPARISON CHART: DIGITAL PLOTTERS
MANII ~'AC'rtIIU:1t
Electronic Associates, Inc.
Long Branch, New Jersey

Concord Control Inc.
1282 Soldiers Field Road
Boston. MaBsachusetta
II> ENTITY

Concord Coordlnatograph

Model 3500 Dataplotter

Series 430

TYPE

Table

Table

Table

ON-UNE OPERATION

Yea

Yee"

?

INPUT DEVICE SUPPUED

Small computer: teletype
ASR 33: mapetlo tape: card
reader

Spec lal Interface unit requ Ired
for on-line operation

?

INPUT MEDIUM

Map_tic tape: punched
carda: paper tape

Mapetlo tape; punched carda:
paper tape manual keyboard:
computer"

?

INPUT CODE TO PLO'M'ER

?

BCD or Binary>

?

CHART SIZE

60 x 60 Inches

30 x 30 Inches: 45 x 60 Inches'

30 x 30 inches

PLOTTING MODE

Point; line

Point; line; continuous

Point; line

ACCURACY (%)

0,001

0.05 (point); O. 1 (line)

0.02

SPEED (MAX)

6 Inches per second (poInt);
2 Inches per second (11l1li);

5. 8 point per second (point):
2.3 Inches per second (lIne);
15 Inches per second (continuous)

20 Inches per second

SYMBOL PRINTING

Yea

4 S characters alphanumeric
BJld symbols

PURCHASE P/uCE, $

140,000 Includes computer BJld
table

48 character alphanumeric
and symbols
18,800 (small taDI~
26.950 ilarge table

COMMENTS

InterchlUlleable hew for line
scribing, photo8crlblng and
photcprojection; head rotation
\a optional; High precision
machine.

17,500

Dataplottera accept 4-dlglt
«-9999 to +9999) Inpute; all
units have manual Input keyboard for selecting one or
more sete of scale factor and
origin values which can then
be c hanged automatic ally by
program; optional program
controlled, 8-pen turret.

• Optional at additional cost.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

6/69

SPECIAL REI'ORT

23:070.614

COMPARISON CHART: DIGITAL PLOTTERS
MANUFACTURER

Gerber SctatSfto
Inlll:rumellt Company
P. O. Box 3015
Hartford, Corm.
IDENTITY

Model 22

Mode13:!

l1ode175

Graphic Dl8play Table

Grephic Dlaplay Table

Graphic DI.play Table

TYPE

Table

Table

Table

ON-UNE OPERATION

Ye,-

Yea·

Yes·

INPUT DEVICE 8UPPIJED

Paper tape; manual slew control.; optional keyboard. magnetic tape reader, and card reader;
Teletype ASR with Serie. 1500 and 2000 control.

INPUT MEDruM

Paper tape; magnetic tape
pllllched carda

Paper tape; magnetic tape;
pllllched card.

Paper tape; magnetic tape;
pllllched cards

I
INPUT CODE TO PLOTTER

EtA (8-level) and USASCJI

EIA (8-1IVel) and USASCII

EIA (8-level) and U8ASCJI

CHART SIZE

50 x 80 Inchee

48 x 80 Inche.

Ii x 8, 5 x 12, 5 x 16, 5 x 20,
or 5 x 24 feet

PLOTTING MODE

Point; line; colltllluoUI

Point; line; oontmuoua

Point; line; contlnuou.

ACCURACY (%)

;to. 007 to +0.009

:to. 0009 to +0.0025

:to. 005 to +0. 009

SPEED (MAX)

3.3, 6.8, or l3.3Inche. per
aecend dependln&: on control
ulled

1.25, 2. SO, or 3.75 Inche.
per second depending on
control uaed

6.6,8.3, or 12.5 Inches per
aecond depending on control
ueed

SYMBOL PRINTING

72 character.

72 character.

72 characterl

PURCHASE PRICE, $

Price on request

Price on request

Price on request

COMMENTS

Incremental !Rep ai.. for Model 22 and 75 Tables 'e 0.0005 Inche •• Step size for Model 32 Table
I. 0, OOOllncbe.; flve individual control IIIIItl including the Series 500, 600, 1500, and 2000
Control. are available at ema coat to be u.ed with anyone of the three Table.; control IIIIIt.
provide Input medium .peclfted above and can a1IO, at extra ocst, Interface a customer specified
compntar.

• Optional at additional coat.

6/69

A

(Contd.)

AUERBACH
to

·"
DIGITAL !"LOTTERS

23:070.11 !I

COMPARISON CHART: DIGITAL PLOTTERS
MANlII"AC'flll\l::l\
Ford Instrument ComplIIY
31-10 ThomBon Avenue
Lone Island City, N. 'i.
mf:NTITY

Electronic Plotter

TYPE
ON-UNE OPERATION
INPUT DEVICE SUPPLIED

Mllgo Electronic Corporation
7620 N. W. 36th Avenue
Miami, Florida

Mode14020D

DPS-6 DICltal plotttnc:
system

Special chart

Table (vertical or horizontal)

Table (vertIcal or horizontal)

'Ie."

Ye.-

Yes"

None

Special Interface

Mapetic tape unit; card
reader; keyboard

Magnetic tape; punched
~ard8

Magnetic tape; punch§d cards;
paper tape; keybosrd;
computer

Mapetlc tape; punched oards;
paper tape; keyboard; computer

INPUT CODE TO PLOTTER

?

Binary

BCD

CHART SIZE

15 x 15 Inohes

30 x 30; 30 x 60;

45 x 60 Inche.

30 x 30; 30 x 60;
45 x 60 Ineb..

Point; line; continuous

Point; line; continuous

INPUT MEDIUM

PLOTTING MODE

Point

ACCURACY (%)

?

0.05

0.05

SPEED (MAX)

50 points per second

30 Inches per second

25 inches per second (lInea)

SYMBOL PRINTING

1N0

No

50 character alphanumeric and
symbols·

PURCHASE PIUCE, $

~Ice upon request

26,000

25,000

COMMENTS

This Is a developmental model
hat features all electronic
operation •

Has two carriages and can plot
two curve8 concurrently.

• Optional at additional cost.

e

1969 AUERBACH Corporation and AUERBACH Info. Inc.

6/69

23:070.616

SPECIAL REPORT

COMPARISON CHART: DIGITAL PLOTTERS
MANUFACTURER
Houllton Instrument
Division of Bausch &. Lomb
4950 Terminal Ave.
Bellaire, Texas 77401
IDENTITY

COMPLOT DP-l

COM PLOT DP-3

COM PLOT DP-5

TYPE

Drum

Drum

Drum

ON-LINE OPERATION

Yes

Yes

Yes

INPUT DEVICE SUPPLIED

Magnetic tape transport"

Magnetic tape transport'

Magnetic tape transport·

INPUT MEDIUM

Magnetic tape; Digital
Computer

Magnetic tape; Digital
Computer

Magnetic t.~pe; Digital
Computer

INPUT CODE TO PLOTTER

?

?

?

CHART SIZE

11 inches x 144 feet

22 Inches x 144 feet

11 inches x 144 feet

PLOTTING MODE

Incremental

Incremental

Incremental

ACCURACY (%)

?

?

?

SPEED (MAX)

300 increments per
second

300 increments per
second

1200 steps per
second

SYMBOL PRINTING

Yes

Yes

Yes

PURCHASE PRICE ($)

3550

6400

11000

COMMENTS

• Optional at additional cost.

6/69

A

AUERBACH

'"

....

-

23:080.001

...A ...."
III....

RDJP

SPECIAL REPORT
DATA COLLECTION

AUERBACH SPECIAL REPORT:
DATA COLLECTION SYSTEMS

PREPARED BY
THE TECHNICAL STAFF OF
AUERBACH CORPORATION

C 1967 AUERBACH Corporation and AUERBACH Info, Inc.

5/67

,/

23:080.100
SPECIAL ,REPORT
DATA COLLECTION

AUERBACH SPECIAL REPORT:
DATA COLLECTION SYSTEMS
.1

INTRODUCTION
Automatic data collection (ADC) implies the recording, in machine-readable form, of the
pertinent data about a transaction at the time the transaction occurs. Some data collection
systems collect and record the transaction data in machine-readable form for later batch
processing; others feed data directly into real-time computer systems to provide up-tothe-minute information for operational decisions.
This Special Report summarizes the results of a comprehensive AUERBACH survey of the
characteristics and applications of the transmitting automatic data collection equipment
that is commereially :l.Vaiiable in the U. S. today. The paragraphs that follow provide background information to aid you in justifying and planning an ADC installation. A comparison chart, arranged in a format designed to facilitate objective comparisons, presents the
key hardware, performance, and cost characteristics of each of seven different transmitting data collection systems. *

.2

WHY AUTOMATIC DATA COLLECTION?
The need for improving the accuracy and reducing the cost of providing the necessary input to automatic data processing systems has long been recognized. Furthermore, modern manufacturing control systems require up-to-the-minute information about what is
happening in the plant, so that operating decisions can be based upon current conditions
rather than upon statistics covering last week's operations.
Transmitting data collection equipment that can meet both these needs is now available
from several manufacturers. Through the use of such equipment, it is now feasible to
design systems that can:
•

Provide tne complete, timely data needed for accurate cost control;

•

Reduce the number of times and places at which data must be transcribed,
thereby cutting clerical costs and error rates;

•

Make and implement operating decisions of a routine nature; and

•

Provide information about current plant conditions upon request.

Actual real-time control of manufacturing operations is still not common, but the other
potential advantages of automatic data collection - reduced clerical costs, increased
accuracy, more effective cost control, and sounder operating decisions - have immediate
significance for nearly every business •
.3

TYPES OF TRANSACTION RECORDING UNITS
Transaction recording units are devices that can record pertinent data about a transaction
in machine-readable form at the time the transaction occurs. The objective of such devices is to collect data accurately and quickly in a form suitable for processing on a computer or tabulating equipment, thus eliminating the need for manual key-punching.
A wide variety of techniques and equipment is currently being employed for transaction
recording. While this report is concerned primarily with transmitting data collection systems designed for industrial use, a review of some of the other techniques and representative equipment used in transaction recording will help to establish the proper perspective •

• 31

Prepunched Tags
One of the Simplest transaction recording techniques has been widely used by retail outlets:
prepunched tags, such as the Dennison and Kimball tags. When an item is sold, the sales
clerk is instructed to tear off one section of the tag (which contains the item identification
and price) and deposit it in a box near the cash register. These tags are collected periodically, carried to the data processing center, and converted to standard punched cards for
use in sales analysis and inventory control applications. Although the method is simple
and inexpensive, it generally involves a high error rate because clerks frequently neglect
to tear off and deposit the required sections. Furthermore, the prepunched tags are difficult to modify for exceptions. The prepunched tag method is very useful for sales analysis
to indicate the fast-moving and slow-moving items. but it has generally been found inadequate ior accurate inventory control.

*A detailed technical

report on each of these systems can be found in AUERBACH Data
Handling Reports, another analytical reference service published by AUERBACH Info,
Inc.
@

1967 AUERBACH Corporation and AUERBACH Info, Inc.

5/67

AUERBACH STANDARD EDP REPORTS

23:080. 320 '.

5/67

• 32

M:mual Recorders
Many organizations employ simple manual devices which record, in machine-readable
form, information coded on embossed cards (e.g., credit cards). Imprinters for this
purpose are produced by Addressograph-Multigraph, Dashew Business Machines, Farrington Electronics, and others. Usually the coded information is read by an optical character
reader to produce input to a computer system. Like the prepunched tags, this system is
simple, relatively inexpensive, provides for capturing a record at the source of certain
relevant information about each transaction, and requires manual transportation of the
recorded data to the processing center. The system is generally suitable only for billing
and sales analysis by territory since only the customer's name, identification number, and
amount of transaction are currently imprinted. The reject rate can be relatively high
because of difficulty in maintaining the required quality of imprinting.
Other variations of this general type of transaction recorder are represented by IBM PortaPunches, in which variable data is encoded by pushing partially punched holes out of a card;
the Wagner Micro-Punch, in which variable data is set up by lever movements and punched
into a card by pulling a handle forward; and the Wright Punch, a simple single-card punch.

• 3:1

1\I ark

Sensing
Mark sensing is a widely-used technique that permits data to be recorded at its source on
standard punched cards, using no special equipment except a pencil that produces electrically conductive marks. After the cards have been carried to the central processing site,
the marked data can be sensed and converted to standard punched-hole form by such
machines as the IBM 514 Reproducing Punch or 519 Document Originating Machine. Newer equipment, such as the IBM 1232 Optical Mark Page Reader, can read and transcribe
marks made by ordinary pencils.

· 34

Byproduct Punched Tape or Cards
Another important transaction recording technique is the connection of paper tape punches
(or, less frequently, card punches) to cash registers, typewriters, savings bank window
machines, and other manually-operated business machines to capture a machine-readable
record of each transaction. As an example of this widely-used technique, let us examine
the use of a cash register with an integrated tape punch. As each sale is rung up, the
clerk records the department number as well as the amount via the register keyboard.
Both are punched into the paper tape, which is collected and carried to the data processing
center at the end of each day to provide input data for sales analysis. Incorporation of the
customer's account number into the paper tape record of each transaction enables billing
to be accomplished from the same input. The obvious advantage of this system is that
source data is captured in machine-readable form as a byproduct of the normal cash register operation. Serious drawbacks to the use of such systems, however, are the cost of
the paper tape punch in each register, the frequency of clerical errors in entering department numbers, and the number of tape rolls that must be collected and spliced for efficient
computer processing.
A variation of this basic technique is the use of optical journal tape readers, such as the
NCR 420-2 and the IBM 1285, to read the printed transaction records produced by many
standard cash registers, adding machines, and accounting machines •

. 35

Non-Transmitting Data Collection Systems
Industrial data collection systems of the non-transmitting type, such as the Standard Register Source Record Punch, are similar to the cash registers described above in that they
produce a record on punched tape or cards of the pertinent data about each transaction,
which must be manually transported to a central location for subsequent processing. The
system response time of such equipment is necessarily long, and it is obviously unsuitable
{or real-time control applications, yet its relatively low cost may make it more suitable
than transmitting equipment for many small-scale installations .

.4

TRANSMITTING DATA COLLECTION SYSTEMS
The highest level of transaction recorders in the field today, and the one that will be of
maximum value to most large manufacturing companies, is represented by the transmitting
data collection systems that are now being used extensively for employee attendance recording, production control, labor distribution, inventory control, and a variety of other
applications. The object of this report is to survey and evaluate the commercially available
data collection systems of this type.
A data collection system of the transmitting type consists of:
• Input units which accept and transmit fixed data from prepunched cards and/or
badges and variable data from dial, lever, or slide settings;
• Output units which record the transmitted data on punched tape, cards, or
magnetic tape, or control its direct entry into a computer system; and
• Cables or communications facilities to transmit the data from the input units
to the output units, which may be located in the same plant or many miles
apart.

fA

AUERBACH

"

(Contd. )

SPECIAL REPORT

.4

23:080.400

TRANSMITTING DATA COLLECTION SYSTEMS (Contd.)
Transmitting data collection systems can be classified as "on-line" systems, which feed
data directly into a computer, or "off-Une" systems, which produce machine-readable
transaction records that will generally be processed later by a computer. Sever.al of the
systems surveyed in this report can be used in either on-line or off-line configurations.
A typical transaction message in a production control and labor distribution application
might consist of: employee number (read from the employee's badge); job number (read
from a prepunched card traveling with the job); machine operation number, transaction
code, and quantity completed (entered by the employee via manually-operated dials or
levers); input station number (transmitted automatically); and time, date, and an error
indicator (added automatically at the central recording unit) .

.5

FACTORS TO CONSIDER IN PLANNING FOR ADC
Enough successful and unsuccessful installations of transmitting data collection systems
have now occurred so that we can list a number of desirable things to do - and to avoid when planning such an installation •

. 51

Detailed Systems Study
The first question is: Do you really need automatic data collection? Instead of installing
an expensive mechanized system to record actual job hours, for example, it might be
better to install a good hourly job standard system and not bother to record actual hours.
The reduced time lags between occurence and reporting of events that automatic data
collection makes possible are of no value unless mangement knows what actions are
dicated by the reports it receives and initiates those actions promptly.
The decision to use automatic data collection equipment in connection with conventional
batch-type processing should be made only after a detailed systems study. (It is assumed
that all real-time information systems will require some form of transaction recording
equipment.) The systems study must determine what information management needs and
the minimum amount of data that must be collected to satisfy those needs. Then a suitable
system must be designed. It is unlikely that straightforward mechanization of existing
manual reporting systems will lead to the most efficient use of automatic equipment.
Existing systems should be streamlined wherever possible, and the full support of top
management is essential.
All potential applications should be carefully considered. For example, an integrated data
collection system in a production plant can be used for attendance reporting, inventory control. parts and material requisitioning, shipping, purchasing, billing, inspection, and
numerous other functions - all in addition to the primary functions of productIOn control
and labor distribution.
Complications will arise from material substitutions, returns, damaged items, obsolete
parts, inaccurate counts, unplanned requiSitions, reworks, etc. Provisions should be
made to handle all such complications without deviating from the cardinal design principle:
send all messages relating to a particular application through the mechanized system.
Don't plan to mechanize only the high-volume transactions and handle the exceptions manually. Il.ial systems will create continual problems and additional expense .

. 52

Configuration
One of the biggest problems in specifying a data collection system is determining system
capacity - how many input stations and central recording units will be needed. The peak
loads that will be imposed on the system must be determined; these will most commonly
occur at cJocking-out time in systems used for attendance reporting. Message lengths
should be minimized to reduce data entry and data transmission times. Message length and
transmission speed will determine the service time per transaction. The service time,
in turn, determines the maximum number of input stations that can be adequately serviced
by each central recorder. In determining the capacity of individual input stations, the time
required to enter the necessary cards, badges, and/or variable data must be added to the
data transmission time.
Closely related to system capacity is the question of where to locate the input stations. You
will need to consider the maximum distance an employee should have to walk to get to an
input station, the maximum waiting times that can be tolerated, and the costs of walking to
the station and waiting to use it as compared to the costs of additional input stations and
transmission lines .

. 53

Cables Versus Two-Wire Transmission
One of the major disadvantages of transmitting data collection systems can be their relatively high cost of installation. The cable cost for systems interconnected by multi -wire cables
can represent a significant portion of the total system cost. A reasonable estimate is about
$1. 00 per foot of cable, with the cost of the cable itself amounting to about one-third of the
total and the labor involved in junction box connections accounting for much of the remainder.
Input stations in most installations will be moved frequently, and each move will usually
require relocation and extension of the existing cables.
C 1967 AUERBACH Corporation and AUERBACH Info, Inc.

5/67

23:080. 530

.53

AUERBACH STANDARD EDP REPORTS

Cables Versus Two-Wire Transmission (Conte:!.)
Since many commercially available data collection systems can utilize two-wire transmission facilities as an alternative to multi-wire cables, the relative merits of the two
transmission modes should be examined. BuildiDgs separated by city streets or plants at
locations remote from the central recording point can be handled more easily with two-wire
hookups. A two-wire system can utilize existing telephone Jines and thereby greatly reduce
installation and maintenance costs. But two-wire systems generally require special adapters
(usually Bell System Data Sets or equivalent modems) to provide for serial transmission of
the bits that make up each character.

· 54

Custom Modifications
Where the published specifications for a particular data collection system do not exactly
coincide with your requirements, remember that most manufacturers will be glad to discuss
potential modifications of their equipment when a sizable installation seems to require such
modifications. It is probable, for example, that most "off-line" systems can be adapted for
on-line use with most digital computer systems, thrugh the user will probably have to bear
the engineering costs of the necessary modifications.

· 55

Training
Another important point to consider is the training and indoctrination that must be given to
each employee who will be using a transaction recorder. With at least 30 minutes of weIlplanned instrUction, it should be possible to reduce the rate of human errors to about 1 per
cent of the total transactions. To ensure acceptance of the mechanized system by the
employees, they must be thoroughly briefed in advance. The briefing should explain why the
system is needed, how it will operate, and how it will affect each employee. Several data
collection installations have failed because the need for pre-installation training and indoctrination was ignored, leading to a strongly rebellious attitude among the workers.

· 56

Reliability
The need for high reliability in a data collection system can hardly be over-emphasized.
Therefore, in evaluating specific equipment, it is wise to ask the manufacturer's representative such questions as:
•

What happens if a single cable breaks? (Is the entire system incapacitated?)

• What happens if a central recorder fails? (Are all connected input stations
incapacitated, or can another recorder pick up the load?)
• Where are the nearest service technicians, and how soon can one be summoned?
.6

THE COMPARISON CHART
The accompanying comparison chart summarizes the key characteristics of seven commercially available transmitting data collection systems, in a concise format designed to
facilitate objective comparisons and pinpoint the specific advantages and disadvantages of
each system. The comparison chart entries are explained below .

. 61

Input
Probably the most importantfactor in determining the success of a data collection installation
is the speed, convenience, and flexibility of data entry. Input data can" be broadly classified
as either "fixed" or "variable." Fixed data is defined as data read from previously prepared punched cards, plastic badges, or other semi-permanent, machine-readable data
storage media. Variable data is data entered manually at transaction time by means of
dial, slide, or lever settings.

· 611 Punched Card Input
All the systems described in this report can accept fixed input data from standard, Hollerithcoded, 80-column punched cards. The method of entry is usually by manual insertion and
removal of one card at a time. The number of columns that can be read from each card and
the number of cards that can be read in a single transaction are indicated.
· 612 Badge Input
Most systems can accept fixed input data from badges or tokens which are manually inserted
into the input device. This capability is particularly valuable for employee attendance recording. The number of columns that can be read from each badge and the number of badges per transaction are indicated.
· 613 Variable Input
The type of facilities that permit the user to enter variable data at transaction time, and
the number of digits that can be entered in a single transaction, are indicated. The variable
data will usually be entered by means of a set of dials, switches, slides, or levers.
(Contd. )
5/67

A.

AUERBACH

23;010.611

SPECIAL REPORT
• 614 Reltrtcted Input

In many appUoationl there wUl be certain lemt -permanent data that tl part of all or mOlt
mellapl from a given Input station. If the Input device Includes meanl for entering variable data and then preventing tt from betng altered by unauthorized personnel or reset to
zero after each transaction, this 1B called "restricted Input." The method of restriction
II ncted; most commonly this consists of a hinged, lockable cover over some of the dials,
levers, or slides used for variable input.
· 615 Transaction Codes
Multi-purpose data collection systems usually utillze a transaction code to specify the
nature and, in many cases, the message format of each transaction. The number of available codes is specified here. In some systems the transaction code Is entered by the same
method as the other variable data; in other systems there are special provisions. Certain
types of transactions may be restricted, requiring Insertion of a supervisor's key or
special badge to Initlate their transmission •
. 616 Automatic Reset
Automatic resetting of the variable dials, levers, or slides to zero after a message has
been transmitted is a feature that will increase input speeds and reduce errors in most data
collection applications.
· 617 Visible Settings
After the variable data for a transaction has been entered, it is important to note whether
the settings are visible to the user so that he can verify that the data has been entered
correctly before the message is transmitted. In systems that employ dials, levers, or
slides for variable input, the settings will generally be visible, though it may not be easy
to read them quickly and reliably. Some input units incorporate a direct, digital display of
the data about to be transmitted .
. 618 Entry Instruction Display
Entry instructions can be displayed in some systems to help the operator enter the correct
data. In several input units, a knob or thumbwheel is used to rotate a cylinder so that
instructions for a particular transaction can be seen through a slot that is normally located
beneath the variable entry dials, switches, or slides .
. 62

Output

.621 Medium
Data collection systems of the transmitting type can be broadly classified as "on-line"
systems, which feed data directly into a computer, and the more common "off-line" systems, which produce a machine-readable record of each transaction for later proceSSing.
Output from an off-line system will generally be on punched tape, punched cards, or magnetic tape. The basic output media for each system are listed here •
. 622 Code
The standard output code (e. g., the number of levels for punched tape output) 1B briefly
described here .
. 623 Maximum Input: Output Unit Ratio
Data collection systems of the transmitting type can assume a wide variety of equipment
configurations, ranging from a single input unit with cable-connected recorder to a far-flung
network with multiple input units transmitting data to multiple recorders or computers by
means of both common carrier facilities and direct cable connections. Probably the most
important parameter in planning the equipment configuration of a system is the maximum
number of input stations that can be connected to a single central recording unit, as indicated in this entry.
· 624 Error Checks
Once a data collection system has been installed and accepted, the operations of an industrial firm will tend to become heavily dependent upon it. Therefore, it is extremely important that the data collection hardware be designed to:
(1) Minimize the occurrence of errors; and
(2) Ensure that virtually all errors that do occur will be detected and corrected.
Minimization of the occurrence of errors involves a great many relatively intangible factors
such as component reliability, mean time between failures, conservatism in circuit design,
transmission line quality, preventive maintenance, proper training of all system users, and
availability and quality of service. The prospective user of any data collection system must
satisfy himself that the incidence of errors and system down-time can be kept low enough to
meet his needs.

C 1967 AUERBACH Corporation and AUERBACH Info. Inc.

5/67

AUERBACH STANDARD EDP REPORTS

23:080.615

COMPARISON CHART: TRANSMITTING DATA COLLECTION SYSTEMS

MANUFACTURER
SYSTEM

CONTROL DATA
TRANSACTER

CONTROL DATA
8010

FRIDEN
COLLECTADATA 30

H('!lOl"t Number·

1130·

1131·

1272*

lliE.lIT

Punched Card Input:
Columns/card
Cards/transaction

15, 22, or 80
l,2,or3

28 to 80
up to 4

up to 76
lor 2

Badge Input:
Columns/badge
Badges/transaction

15- or 22-column cards
used as badges

short cards used
as badges

10

Variable Input:
Type
Digits/transaction

10-position dials
6 or 9

10-position dials
9

l2-position dials
10

Restricted Input:
Type
Digits/transaction

plugboard
10

programmed
no limit

covered dials
8

Transaction Codes (number)
Automatic Reset
Visible Settings
Entry Instruction Display

10
yes
yes
no

9
yes
yes
yes

7
yes
yes
yes

magnetic tape, punched tape,
or CDC computer
7-track, or 5- to 8-level
punched tape
36:1 (2)

magnetic tape or
CDC computer
7-track

punched tape
or computer
8-level

128:1

20:1

Error Checks

input interlocka, message
length, parity, special
circuit checks

parity, message
length, and
special checks

input interlocks,
message length,
parity

Time Recording
Date Recording

yes
yes

yes
yes

yes
yes

60 (3)
34 msec
16- to 60-wire or
2-wire
14,000 ft (7)

54 (4)
none
24-wire or
2-wire
2,500 ft (7)

30
none
l5-wire or
2-wire (6)
2 mUes (7)

$34-89
$390
$70

$40-80
$55
$105

$1,040

$50
$780
$200 (per 16 inputs)
$1,060 (per 128 inputs)
N/A

October 1959
5 to 6 montha

November 1965
4 to 5 months

1961
3 to 6 months

OUTPUT
Medium
Code
MaXimum Ratio of Input
to Output Units

TRANSr.flSSION
Speed (char/sec)
Minimum Polling Delay
Line Requirements
Range
COSTS (PER MONTH)
Input Station
Central Recorder
Control Unit
Typical 10-Station System
AV AILABILITY
First Delivery
Delivery Period

1

N/A

*These references are to AUERBACH Data Handling Reports, another analytical
reference service published by AUERBACH Info, Inc., where a detailed report
on each of these data collection systems can be found.
See facing page for notes (1) through (8).
(Contd. )
5/67

A

AUERBACH

•

SPECIAL. REPORT

23:080.616

COMPARISON CHART; TRANSMITTING DATA COL.L.ECTION SYSTEMS (CoNTD)

IBM

IBM

357

1030

RCA
EDGE

1440-

1441-

1690·

TEXAS INSTRUMENTS
TACTICOM
1810·

up to tlO
unlimited

up to 80
lor2

up to 80
lor 2

up to 79
unlimited

10
unl tnllted

10
1

1 to 12
1

10
1

ll-pos It IOn slides
G. 9. or 12

ll-position slides
12

slides
10

10-position slides
12 (1)

slides c.m he locked

slides can be locked

coded plug
3

emitted from control unit
13

10
yes
no

10
yes
yes
yes

10
yes
yes
no

none
yes
yes
no

punched cards or
IBM 1440 or 1460 computer
Hollerith or 6-bit BCD

punched cards or IBM 1440,
1460, or 360 computer
Hollerith or 6-bit BCD

punched tape or RCA 301,
3301 or Spectra 70 computer
8-level

magnetic tape
or computer
7-track

20: 1

24: 1

25:1

40;1

Input interlocks
message length

input interlocks, parity,
message length, punch
comparison check

input interlocks,
parity, start-end
sequence, message length

input interlocks, parity,
message length, echo
recording check

optional
yes

optional
yes

yes
yes

yes
yes

20
250 msec
41- to 66-wire or
2-wire (6)
5,500 it (7)

60
100
2-wire

27.7 (5)
none
2-wire

125
270 msec
2-wire

8 mUes (7)

(7)

10,000 it (7)

529-62
567 or 87
579
5816

$25-140
$370
none
$1,620

$69-135
$400
$215
$1,305

$60-107
$485

June 1959
6 months

July 1964
6 months

1961
when available (8)

February 1967
3 to 4 months

VI'S

-

(1) Variable input data may be alphanumeric.
:-';0 theoretical1imit on the number of input stations.
From one or two input stations at once.
(4) From up to 15 input stations at once.
(5) From one to four input stations at once.
(6) A data set is required at each transmitting and r.eceiving station for 2-wire operation.
(7) Range is essentially unlimited when telephone lwe. lU'e used.
(8) No longer in production.
(2)
(3)

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

5/67

AUERBACH STANDARD EDP REPORTS

23:010.624

.624 Error Checks (Contd.)
Errors will occur, even in the best-engineered and costliest systems. Therefore it Is
important to detect and correct as many of these errors as possible. The main types of
error checking performed by each system are listed here. The most common checks are:
• Input interlocks - checks which verify that the correct types and amounts of
data have been inserted, in the correct sequence, for each transaction. Such
checks can detect many procedural errors committed by persons entering input
data into the system.
•

Parity - addition of either a "zero" or "one" bit to each character code so that
the total number of "one" bits in every transmitted character code will be either
odd or even. Character parity checking can detect most single-bit transmission
errors, but it will not detect the loss of two bits or of an entire character.

• Message length - checks which involve a comparison of the number of characters
received at the output unit with the correct number of characters as specified
for that particular type of transaction. Message length checks can detect many
errors arising from both improper data entry and equipment or line malfunctions.
· 625 Time and Date Recording
The time of day and/or the day of the week or month form an important part of the record
of each transaction in most data collection applications, so special provisions are frequently
made to supply this information automatically.
. 63

Transmission
These entries describe the available means for connecting and transmitting data between the
input stations and the central recording units, along with the resulting speeds and maximum ranges .

. 631 Speed
This is the normal peak rate of data transmission, in characters per second .
. 632 Minimum Polling Delay
Some control units poll constantly for input station activity, some initiate polling when an
input station requests transmission, and some eliminate polling delays by determining the
station to transmit during the previous transmission or through the use of direct electrical
impluses.
· 633 Line Requirements
Where input and output units can be linked by direct cable connections, the number of conductors required is listed here. In cable-conected systems, data will usually be transmitted in a "parallel by character" mode; i. e., all the bits comprising a single character are
transmitted simultaneously via multiple conductors, and successive characters are transmitted sequentially. Where 2-wire communication lines are employed, data transmission
will necessarily be "serial by bit;" i.e., each bit of each character is transmitted sequentially over the same pair of conductors. A data set is commonly used at each sending and
receiving terminal to perform the necessary conversions between the parallel and serial
transmission modes. Several systems can utilize either multi-conductor cables or 2-wire
communication lines .
. 634 Range
The maximum allowable distances between input stations and central recorders in cableconnected systems are listed here. Where common-carrier telephone lines are used, the
range is essentially unlimited .
. 64

Costs
The approximate single-shift monthly rental prices for each input station, central recorder,
and control unit (when required) are listed here. Where there is a choice of two or more
models with different capabilities, the price range is shown.
The "Typical 10-Station System" is defined as a small, off-line system providing ten input
stations capable of accepting card, badge, and variable input data (where available); one
central recorder; and any required central control units. Costs of cables, transmission
lines, data sets, and installation are .!!.2t included in the indicated monthly rentals.

· 65

5/67

Availability
The first delivery date and the current time from order to delivery are shown, to give you
a good idea of the length of time each system has been in operation and its current production status.

A.
AUERBACH

-A.

23:090.00 I
""1lI1

~'EDP

'UIItI ..

C~

•

1I".tt

SPECIAL REPORT
SERVICE CENTERS

SPECIAL REPORT
THE SELECTION AND USE OF
A DATA PROCESSING SERVICE CENTER

by
Gordon B. Davis
Professor, University of Minnesota
Computer Consultant to the Ame:o-ican Institute
of Certified Public Accountants

C 1967 AUERBACH Corporation and AUERBACH Info, Inc.

8/67

23:090,002
SPECIAL REPORT
SERVICE CENTERS

CONTENTS
.1

.11
.12
.13
.2

.21
.22

Definition
Importance and Need
Purpose of This Report
DESCRIPTION OF DATA PROCESSING SERVICE ORGANIZATIONS
Classification
How They Generally Operate

.3

CHARACTERISTICS OF GOOD APPLICATIONS FOR SERVICE CENTERS

.4

LOCATING A DATA PROCESSING SERVICE CENTER

.41
.42
.43
.44
.5

.51
.52
.53
'.54

.6
.61
.62
.63
.64
.65
.66
.67
.68
.69

.7

.71
.72
.73
.8

.81
.82
.83
.84-

.85
.9

.91
.92

8/67

INTRODUCTION

Standard Business Sources
ADAPSO Directory
Reference Listings
Computer Manufacturers
HOW MUCH THEY CHARGE
Cost Structure of the Service Center
Economics of the Program
Approaches to Charging
Typical Charges
DECISIONS TO BE MADE
Preparation of Machine-Readable Input
Method of Transporting Data
Handling of Input Errors
Storage of the Files
Security Over Files and Data
Special Programs Versus General-Purpose Program
Ownership of Special Programs
Payment for Extras
Ensuring Against Interruption of Service
HOW TO SELECT A DATA PROCESSING SERVICE ORGANIZATION
Preparing a Request for Proposals
Evaluating a Proposal
Completing the Negotiations
CONTROL OVER IMPLEMENTATION
Scheduling
Control Over File Conversion
Sample Run for Acceptance Test
Running in Parallel
Operating Procedures
SELECTING A TIME-SHARING SERVICE CENTER
General Description
Selection Considerations

A•

AUERBACH

23:090 100

A

AUlIlACM

II ......

BDP

SPECIAL REPORT
SERVICE CENTERS

...U1I

THE SELECTION AND USE OF
A DATA PROCESSING SERVICE CENTER
.1

INTRODUCTION

. 11

Definition

. 12

A data processing service center or service bureau is an organization that provides data
processing services to outside clients on a fee basis. These services may be provided
continuously, under contract, or as needed. This definition encompasses centers which
use only unit record equipment as well as centers which have computers. There is considerable diversity in the data processing service center industry, but a central tendency
of established firms is the providing of a complete data processing service rather than
merely renting equipment time. This means that a typical, established data processing
service organization has qualified personnel to analyze customer requirements and write
programs, as well as having control over appropriate equipment ..
Importance and Need
Depending on the definition used, there are some 1200 to 1800 data processing service
centers in the United States which did a business in 1966 of approximately $700 million.
These figures exclude universities, some of which do data processing on a fee basis. In
tallies of service centers, there are differences in the treatment afforded operators of
part-time service centers (often termed "moonlighters") who buy off-shift time on a computer and operate a service center with this equipment.
A data processing service center can be used either to supplant internal manual processing
or to supplement an existing internal machine data processing installation. Those circumstances which justify supplanting internal manual processing are a volume of records,
computations, or tabulations which can be performed at less cost or on a more timely basis
by an organization equipped with data processing machines. Jobs which fit this category
are quantitative or statistical analysis (such as linear programming, critical path scheduling, etc.) or record processing (including such applications as billiug. accounts receivable, sales analysis. payroll. budgets, inventory analysis, etc.).
The reasons why an organization which has its own data processing eqUipment may need to
use a service center include:
(1) Special or periodic overloads.
(2) Projects requiring specialized handling, specialized knowledge. or special
equipment.

. 13

(3) Obtaining experience and assistance in COn!lection with a conversion to new
equipment .
Purpose of This Report
This report is designed to assist a potential user of a data processing servicp center to:
(1)
(2)
(3)
(4)

.2

Locate a suitable service center.
Prepare a request form to use in ohtaining proposals for service.
Evaluate the proposals for service.
Negotiate a contract.

(5) Implement a decision to use a service center.
The report is directed primarily toward the use of a service center for commercial data
processing, although there is some discussion of its use for scientific computation. The
use of a time-sharing service center is considered sufficiently unique that a separate
section is devoted to a summary of considerations in selecting such a service. A subsequent Special Report will cover time-sharing in more detail .
DESCRIPTION OF DATA PROCESSING SERVICE ORGANIZATIONS
The diversity of service center organizations makes it somewhat difficult to categorize them.
Therefore, a classification framework is presented first. followed by a short discussion of
salient points connected with each classification. lje~, there is a description of the way in
which a typical commercial data processing center Will operate for a business-type problem
in which the processing is to be repeated at regul8l' intervals such as weekly, monthly, etc.

© 1967 AUERBACH Corporation and ~l,JERBACH Info. Inc.

8/67

23:090.210

AUERBACH STANDARD EDP REPORTS

. 2I

Classification
Table 1 summarizes the different ways in which data processing service centers may be
classified. These are by ownership, control of equipment, type of equipment and type of
service.
TABLE I: CLASSIFICATION OF DATA PROCESSING SERVICE CENTERS
Ownership

Control of Equipment

Manufacturer

Owner or prime lessor

Independent

Block time lessee

Organization affiliated
University
CPA

Type of Equipment

Type of Service

Unit record equipment

Commercial

Computer

Scientific
Industry specialist

Time-sharing computer

Full-line

The ownership of a data processing center provides a useful background for understanding
the data processing service industry. Most of the major computer manufacturers have
their own data processing centers. In fact, the largest data processing service organization in terms of number of offices and volume of work is Service Bureau Corporation, a
wholly-owned subsidiary of IBM. The independent service organizations vary widely in
size, with the larger ones having offices in many of the major cities.
A Significant portion of the costs of a computer installation are fixed costs. Therefore.
many organizations which have justified a computer for their own use. but which have not
fully utilized the available time. have found it advantageous to enter into a part-time service
bureau arrangement. In some cases. this arrangement involves only the sale of blocks of
time to outside users. with the outside organizations providing their own programming.
staffs. operators. etc. In other cases. however, organizations with large computer installations, such as banks, have organized rather complete data processing services and sell this
in competition with the manufacturer-owned and independent data processing organizations.
There are other service center arrangements which are frequently found. Universities
typically will sell time on their computers although they do not engage in full-service
activities. Some certified public accountants and groups of CPAs have installed equipment
for providing data processing service. The current Code of Ethics of the American
Institute of Certified Public Accountants does not permit its members to advertise. so that
a member CPA having computer facilities will not advertise his data processing service.
except by notifying his own clients and other CPAs.
The control of equipment classification is based upon the fact that the availability of offshift time on computers has made it possible for a person to set up a service center without owning or leasing his own equipment. He leases a block of time from one or more
computer installations, operating these computers with hie own personnel. These businesses
are termed "moonlighters" by the regular data processing centers which own or are prime
lessors of their eqUipment.
The type of equipment found in a service center may be limited to unit record equipment or
it may include a computer. This report is directed primarily toward the centers using
computers. The computer equipment may range from small to large. Several service
centers have been organized which speCialize in time-sharing, and these centers have computers especially suited for time-sharing applications.
Data proceSSing centers tend to specialize in the type of service they offer. Some computer
centers. especially smaller ones, have tended to specIalize in commercial processing of
accounting-type applications. Others tend to specialize in scientific processing. These
centers usually provide personnel with mathematical and analytical ability in the computer
solution of scientific problems. Within these commercial and scientific procesRing
(Contd. )
8/67

A•

AUERBACH

I

\

23:090.211

SPECIAL. REPORT

.21

. 22

Classification (Contd.)
specialties. industry specialistl!l have developed. For example. one service bureau concentrates primarily in the retail business. while another one handles only data processing
for automobile dealerships. Even where there is no announced specialty, the experience
and expertise a service bureau obtains allows it to compete most effectively in the industries
where it has already developed programs and solved problems .
How They Generally Operate
The functions in a data processing service organization are illustrated in Figure 1. Although the organizational structure for centers may vary. there tend to be three major
functions: sales. consulting/programming. and production.

Management

I

I

I

I

Sales

Consulting
and
Programming

Production

I

I

1

Keypunch

Service
and
Quality Control

Data
Processing

I

Figure 1. Functions in a data processing service organization.
The sales function is carried on by sales representatives who call upon customer/! to explain
the services offered by the organization. analyze customer requirements. and present
proposals for performing services.
The purpose of the consulting and programming function is to perform system analysis and
prepare system designs for customers having unique requirements which require speCialized
systems. If the system is accepted by the client. this department also prepares the necessary programs.
The production df'partment performs the data processing activity for the firm. This is
typically divided into three separate areas: keypunch. quality control, and data processing.
The quality control activity is concerned with controlling customer records and ensuring
that the work is done correctly and on time. An account representative or account supervisor in the quality control group Is assigned the responsibility for customer contact
regarding the data processing.
Another way of describing how a service center operates is to trace the handling of a
continuing commercial data processing contract such as preparation of a payroll. preparation of accounts receivable. or a similar application. The salesman who first calls on the
customer may work out the solution and make an estimate. especially if the system is
relatively uncomplicated and fits standard procedures already developed by the center. If
the system is complicated or unique. the salesman will call upon the systems analysts, who
will prepare layouts. system flowcharts. programs. etc.
Once the system has been agreed upon and programs have been prepared. the client's files
are converted to machine-readable form in order to get the system started. Thereafter.
documents received from the client are logged in by the data processing center. checked
for appropriateness by the quality control unit. then keypunched and key-verified. The
data processing group obtains the master records from the quality control group. runs the
program. and turns the results back to quality control. The account representative examines
the master records and the processed reports for completeness and accuracy and returns
the master records to the storap area.
C 1967 AUERBACH Corporation and AUERBACH Info. Inc.

8/67

,

23:090.220

AUERBACH STANDARD EDP REPORTS

.22

How They Generally Operate (Contd.)
The completed reports are either picked up by the client or are transmitted by messenger
service or mail. An error listing accompanies the reports. The client is advised through
the error listing or through a personal call from the account representative as to the action
which should be taken in the next data processing cycle in order to take care of the errors .

.3

CHARACTERISTICS OF GOOD APPLICATIONS FOR SERVICE CENTERS
Obviously, a good application for service center proceSSing is one for which the data
processing service center can perform the processing at a lower cost or on a more timely
basis than can be performed in-house. In general, this suggests that a good apphcation will
have one or more of the following characteristics:

.4

(1)

The volume of records is significant.

(2)

Considerable computation is required.

(3)

The data must be rearranged in several ways to obtain different tabulations or to
perform different computations.

(4)

The time available for processing is too short for the regular in-house processing
staff.

(5)

The user cannot obtain sufficient personnel.

(6)

The data proceSSing center has specialized knowledge not available in-house .

LOCATING A DATA PROCESSING SERVICE CENTER
The emphasis in service bureaus is on service. The equipment is secondary to the quality
of personnel and quality of programs available for use. Therefore, the task of locating a
suitable service center is analogous to locating many other services used by an organization-legal, accounting, medical, etc. In searching out possible service centers, there
are various sources of helpful information .

. 41

Standard Business Sources
The telephone directory classified section lists data processing service centers under data
processing service. Local business directories frequently list data proceSSing service
centers. Other business contacts, especially in the same industry, may be able to provide
names of service bureaus, as may accountants and consultants. CPAs providing data
processing service who are members of the AICPA do not advertise, relying instead on
the recommendations of their clients or other CPAs,

.42

ADAPSO Directory
The Association of Data Processing Service Organizations (ADAPSO) issues a bi-annual
directory of members, ADAPSO membership IS limited to for-profit organizations which
utilize their own equipment on their own premises, assume full responsibility for the
finished product, and have completed one full year of successful operation. This directory
therefore excludes organizations, such 4ls banks, which are only part-time service organizations and moonlighters who do not have their own equipment. Members must subscribe
to a code prescribing standards of conduct. ADAPSO membe~ship is therefore one indication of a stable, bona fide organization. The directory costs $1 and is available from
ADAPSO, 947 Old York Road, Abington, Pennsylvania .

. 43

Reference Listings
The most comprehensive directory of data processing service centers is published annually
in the July issue of Systems magazine. There are much less complete directories in the
June issue of Computers and Automation magazine, and in the Computer Yearbook and
Directory published by American Data Processing, Detroit, Michigan. The September
reference issue of Business Automation for the years 1964 and 1965 contained listings, but
this feature was omitted in 1966 .

. 44

Computer Manufacturers
Computer manufacturers can be a helpful source of information because in many cases they
have their own service centers, and also because they have sold equipment to service
centers, so that they are aware of the centE'rs which are doing business .

.5

HOW MUCH THEY CHARGE
The economics of the service center can he considered under two categories: the cost
structure of the service center and the economics of the programs .

. 51

Cost Structure of the Service Center
Table II divides typical costs of the service center into fixed and variable costs. The
purpose of this tabulation is to provide some understanding of the cost structure of a
service center. Note that a substantial part of the costs of a service center are fixed
(standby, readiness-to-serve, etc.), while the incremental costs of service are fairly

8/67

A

AUERBACH
~

(Contd. )

23:090.510

SPECIAL REPORT

TABLE II: COST FACTORS FOR A DATA PROCESSING SERVICE ORGANIZATION
Fixed Costs

Variable and Semi-Variable Costa

Installation costs

Machine operators
Programmers other than basic staff

Preparation of package programs

Keypunching labor

Start-up costs:

Building rental
Salaried sales, service, and quality control
personnel

Sales commissions
Supplies

Supervision

Customer magnetic tapes, card trays. holders.
etc.
Postage and messenger service

Advertising and promotion

Utilities

Basic programming staff

Equipment rental or depreciation

-

51

Cost Structure of the Service Center (Contd.)
low. One service center manager has estimated that the percentage of cost for two
important elements-equipment rental and labor (other than supervision)-should not
exceed 60 percent of gross income in order for the center to achieve a profit .

. 52

Economics of the Program
The customer's application may be run either on a special program written specifically
for that customer or on a generalized program to which the customer's system has been
adapted. The generalized or package program is used for a number of customers and is
written with that objective in mind.
There are good economic reasons for the use of package programs. The generalized
program spreads the cost of programming over many users; therefore. greater programming effort can go into making the package good. Having a program alreadyavailable makes the system design work with the client easier because the client's system is
adapted to the prolfram rather than the reverse (although most generalized programs do
allow for options with respect to such items as format to suit the individual preferences
of clients). Cost estimates are more certain because of the experience gained from
running similar problems using the generalized program.

. 53

Against the use of package programs is the fact that the program being written with no
single client in mind fits no one exactly and therefore does not completely please anyone.
Even though the recommended approach is to adapt the customer's system to the general
program, it may turn out that this is not feasible, and the service center then adapts the
program to the customer's needs by making changes in the program itself .
Approaches to Charging
There are three basic approaches to cbarging for data processing center services. These
are:
(1) Fixed price.
(2) Time and materials at standard rate.
(3) Cost plus fixed fee or percentage.
The fixed fee is preferred in most cases. with the understanding that changes not agreed on
in advance cost extra. This approach is well suited for standard program packages or for
those cases where specifications for the customer's system are firm and few changes are to
be anticipated. The fixed price may take the form of a fixed charge plus a charge for each
item processed. There may be, in this case. a minimum charge. A typical minimum
charge for work of a recurring nature where the service center must maintain files, controls.
etc .. is $25 per week. This minimum usually applies to the entire set of proceSSing jobs
performed rather than to each report or other item. The minimum reflects the fact that
there is an administrative cost associated with each job no matter how small.
The "time and materials at a standard rate" approach ~!I .suitable where the problem and
procedures are well defined but the running time, nlllJlbe, of runs, or number of transactions are not known. It is also a useful method of cP~1ging in cases where the client's
own program is being used. The "cost plus a fixed fee 9J:' percentage" approach is
applicable where the problem or procedures are not ~J! defined.

© 1967 AUERBACH Corporation and AUE~BACH Info, Inc.

8/67

23:090. 540

AUERBACH STANDARD EDP REPORTS

. 54

Typical Charges
It Is difficult to make statements about typical charges since charges will vary depending on
the portion of the country, type of equipment used, volume of records processed. the extent
of output. and other factors. In order. however, to give some feel for the cost of using
data processing services. the estimates in Table In were prepared. These are based upon
current charges in New York City. Lowest costs are usually obtained by having large
volumes and by using standard packages.
The estimates in Table In include card punching. Since this is a significant factor. some
knowledge of this cost is important. Keypunch operators typically punch some 4, 000 to
10.000 characters or strokes per hour, depending on the type of punching. A rough standard
average rate for pricing purposes for a keypunch operator is 6, 000 numeric strokes or
4.000 mixed alphabetic and numeric strokes per hour from good source documents. The
actual cost for keypunching or verifying will depend on the legibility and format of the
documents. the number of punches per card (since it takes less time to punch SO characters on one card than it does to punch one character on each of SO separate cards). and
the amount of intermixing of alphabetic and numeric characters. A rough rule of thumb
for a quick estimate is one dollar per card column per one thousand cards. Depending on
the characteristics of the job, the rule will tend to give a high figure (by up to 20%). but
is useful for rough estimates. For example, assume a business had the following information to be punched from a document into punched cards:
Digits

Item
Product code
Quantity
Dollar amount

4
4
6

14

Total

Using the rule of thumb, the original keypunching would cost approximately $14 for each
1, 000 documents because there are 14 columns to be punched. Key verification of the
punching would cost an additional $14.
KeypunchIng is usually not considered to be a profitable operation by service centers, but
they offer it as a necessary part of the total service. Large-scale, one-time punching jobs
such as file conversions are sometimes sent to England where keypunching can be performed at a lower price.
One should keep in mind that the manufacturer-rated speed of equipment usually cannot be
maintained as an actual rate in data processing operations, especially in the case of card
handling equipment. For example. a card sorter may be rated at 1,000 cards per minute.
but for all practical purposes (due to handling time. card jams. etc.). the effective throughput rate is only about two-thirds of this figure.
TABLE III: TYPICAL PROCESSING CHARGES
Application
(1)

Payroll service (paycheck, payroll register, quarterly
payroll tax information and W-2 forms). - up to about
500 employees ••

(2)

Sales analysis (assume 1.000 invoices of 2tlines each, 300
products. and 20 salesmen):

(3)

Low

High

30¢

45¢

Basis
per employee per pay
period. $25 minimum
per week is common.

•

Report by product. units and dollars

$20

$27

per report

•

Additional report by salesmen. units and dollars

$ 7

$ 9

per report

Accounts receivable (aged trial balance, aged customer
statements, invoice register. cash receipts journal).

7¢

13¢

per transaction

• Including keypunching and verifying of data .
•• Prices as low as lS¢ have been quoted in New York City.
Prices drop substantially where large volumes of work
are processed.

(Contd. )
8/67

A

AUERBACH
~

23:090.600

SPECIAL REPORT

·6

DECISIONS TO BE MADE

This section describes various key decisions which need to be made when deciding upon
data processing by an outside service center. These are:

. 61

(1)

Preparation of machine-readable input.

(2)

Method of transporting data.

(3)

Handling of input errors.

(4)

Storage of the files.

(5)

Security of files and data.

(6)

Use of a special program versus a general-purpose program.

(7)

Ownership of special programs.

(8)

Payment for extras.

(9)

Ensuring against interruption of service .

Preparation of Machine-Readable Input
One of the options open to the user of a service center is to prepare the data in machinereadable form rather than having the service center do this operation. Methods to perform
this conversion to machine-readable form may be direct or they may be a by-product of
some other data processing operation:
Approach
Direct

Form of Data
Punched cards
{ Paper tape
Magnetic tape
Punched card

Indirect
(By-Product)

Paper tape

1

Optical characters

Equipment
Card punch and verifier
Paper tape punch
Magnetic tape encoder
Punched card attachment on accounting
machine
Paper tape attachment on machine or an add
punch
Optical character printing font on unit such as
cash register or adding machine

In addition there is the possibility of transmitting the data directly to the computer over
communications lines. The question of relative economics of the user preparing the
machine-readable input media versus turning over documents to the data processing center
for conversion is beyond the scope of this report. A succeeding Special Report will
specifically cover an evaluation of equipment for preparing source data input.
· 62

Method of Transporting Data
Although the mails or communications lines are used in some cases. the most common
method of transporting data is by messenger. The question is whose messenger. Where
there is a security problem or considerations of timeliness. a client may choose to use
his own messengers. Otherwise. the service center's messenger or a public messenger
service can be used.

63

Handling of Input Errors
The account representative at the service bureau is responsible for all commumcations
with the client regarding errors or failures in data processing either due to problems at
the data processing center or problems regarding the input furnished to the center. When
an error is detected by the computer. the computer will typically print an error message
and eliminate the item from the processing run. The user then must process the item
manually if it must be done before the next computer processing cycle-as. for example,
is the case with a payroll check. He then sends a change record with the next run in
order to update the files to include the manually processed item.
If the account representative detects an error due to improper processing by the data
center, he will arrange for it to be re-run before it is sent to the client. If an error is
not detected until it reaches the user, he may reject the run if the errors affect so many
parts of it that the results are not usable. Otherwise, the user may accept the run, make
manual adjustments, and send in corrections with the next input batch to be processed. A
question to be discussed with the processing center is responsibility for the cost of re-runs
due to erroneous input data.

· 64

Storage of the Files
Two basic approaches are available. The client may keep his files and take them to the
service center at processing time. If practical, the client's representative may remain
while the data is processed and take the files back with him. The second approach is for
© 1967 AUERBACH Corporation and AUERBACH Info. Inc.

8/67

AUERBACH STANDARD EDP REPORTS

23:090.640

.64

Storage of the Files (Contd.)
the service center to keep all files on its premises. The first approach is used where
there is a necessity to maintain confidentiality in the data processing applications or where
the data processing center has inadequate storage facilities to maintain security and
protection against destruction .

. 65

Security Over Files and Data
Assuming that the data processing center stores the user's files and receives documents
for processing. then a question to be considered is the security over these files and over
the data transmitted to the center. This consideration may be important because of
confidentiality requirements or because of the consequences of loss or destruction.
In the case of confidentiality. one method is to use a code rather than names for such
items as payroll processing. Typically this is not considered necessary. but it is available as a method should it be deemed desirable.
In order to guard against loss or destruction. the client may. as pointed out above. use
his own messenger service and may even store his own files (although this presents many
practical problems). The security arrangements at the data processing center should
include fireproof storage. procedures governing access to records and files. and insurance
to pay claims which may arise. In the case of data being transmitted to the center. the
client should always keep a copy of this data or have some means of reconstructing it in
the event of loss .

. 66

Special Program Versus General-Purpose Program
This was discussed in section. 52. The general-purpose program is to be preferred
because the user knows exactly what he can expect and the costs are firm. A special
program is usually used only when a general-purpose program is unavailable or
unacceptable.

· 67

Ownership of Special Programs
If a client pays for the writing of a special program. the ownership would seem to be his.

However. this should be decided explicitly beforehand. In cases where ownership does
reside with the client. a copy of progress documentation should be obtained as a basis for
progress billings. and the final documentation. including a copy of the program in
machine-sensible form. should be obtained by the client. Provision should also be made
in the contract for restricting. licensing. or otherwise controlling the use of the program
by other users.
· 68

Payment for Extras
Payment for the following are matters for negotiation:
(1) Systems surveys. analysis. etc .. to define the customer's prohlem and to
formulate an approach to processing.
(2) Changes to adapt general-purpose programs for the customer's use.
(3) He-runs necessitated by erroneous input data.
(4)

He-runs necessitated by conditions not anticipated when the system was designed.

There are differences in practice as to payment for systems surveys. If a survey is used
by the service center as a means for obtaining business. there is typically no specific
charge for this service. If the systems survey is requested by the customer in order to
decide how to extend the use of the computer or to alter his processing methods. then this
may be charged as an extra-cost service.
Depending on the type of general-purpose program the service center has. there may be no
changes or there may be minor modifications required to adapt it to the client's problem.
If such minor changes are antiCipated. they may be included in the standard charge for
using the program. However. if the customer wants something not envisioned within the
general-purpose program. this is presumably an extra charge and negotiated accordingly.
As a general principle. re-runs due to erroneous input data or errors caused by the
customer will be charged to him. He-runs caused by the program being unable to handle
conditions which were not excluded when the contract was taken are presumably the
responsibility of the service center. In all cases. however. these should be discussed
beforehand rather than after the fact.
· G9

Ensuring Against Interruption of Service
The service center should itself have made specific arrangements for backup service in
the event of equipment failure or other interruptions of service. It is up to the user to
satisfy himself that these provisions are adequate.
(Contd. )

8/67

fA

AUERBACH
~

23:090.690

SPECIAL REPORT

.69

Ensuring Against Interruption of Service (Ccntd.)

If the user has a copy of a program and the related files, he can, of course, switch at any
time to another service center having similar or compatible equipment. If the user has
paid for a special program. he should also have arranged for a copy of the program. However. the generalized programs provided by the data processing center are usually not
available to its clients. These are considered proprietory.
The ownership of the customer's files should be clearly spelled out so that if the service
center user terminates his relationship for any reason, all master files and data files
maintained by the service center will be returned to him .
.7

HOW TO SELECT A DATA PROCESSING SERVICE ORGANIZATION
In a survey of CPAs regarding the use of data processing service organizations. the following
difficulties were mentioned:
•

Slow service.

•

Lack of accuracy in reports and excessive re-runs.

•

Reports in a format confusing to clients.

•

Insufficient knowledge of accounting by EDP centers.

•

Insufficient planning and preparation.

•

Data transmission difficulties.

•

Additional service costs.

•

Auditing difficulties.

•

Overselling by the centers.

These difficulties can be overcome through a proper approach to the selection of a data
processing bureau and by control over implementation of this decision.
There are three major stpps in selecting a data processing center: (1) preparing a request
for proposals; (2) evaluatmg the proposals; and (3) completing the negotiations. Considerations in the implementation of the decision are covered in section. 8, below. The discussion
that follows is oriented toward the use of a data processing center for a continuing data
processing service rather than for a one-time tabulation or a one-time scientific computation. For the latter case, the general approach is similar, but the amount of investigation may be substantially less because of the one-time nature of the processing .
. 71 Preparing a Reque st for Proposals
The basic idea underlying the request for proposals is that the prospective user of the data
service should define his own data processing reqUirements using his own staff
or a professional advisor. The completeness and detail of the request will depend, in part,
on the capabilities of these individuals. The request should be specific but should allow
the proposals submitted by service centers to suggest either alternative means for
processing or alternative layouts in order to achieve economies or efficiency in processing. The request document should include the following:
proces~ing

(1)

Purpose of the processing.

(2)

A layout of the final reports if the format is important; or a complete description
of content if the format itself is not vital. A sample layout is shown in Figure 2,
and a description of a similar output is given in Figure 3.

(3)

A copy of the input documents (blanks as well as filled-in samples) with a description of the information fields; or a layout of the input data if machine-readable
media will be furnished by the user (Figure 4). If the size of a data item is
variable, a range should he given.

(4)

Number of records to be included in the master file, and the expected growth
factor. The estimate should give a range if there is a considerable difference
in activity for different period!!.

(5)

Handling of exceptions.

(6) Specifications for frequency of processing.
(7) Specifications for timeliness.
(8) Special reqUIrements.

For example:

(a)

extra copies,

(b)

special reports required,

(c)

conversion specifications, including time limits,

prqbl~ms,

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

etc.,
8/67

AUERBACH STANDARD EDP REPORTS
.

23:090.710

10

~ .. • •

•

••

I

j ••• ~ •

••

: : ~: :; i : :

·t·

•

,

,,j

'tt

.... • .. 1·
t

: : 1: : : : : :

I

j

.. ,

••

,tltt+

•• t .1.•. 1 j •
. +., ..... ,.
I I ...

~

,

+~ •
,.

~

+••

t

•••

~

~ >-t~'

-+ .. .

+.,1. t- ...

~ .... .

~*-.

to.

t ...

I • , •••
~-t-.-...j. __ -~

,

..+...-.

.+ ....

,.·t."- .

. l .•••. , .

Note:

•

~-+ ~

The layout shows the format of the report to be
produced. It may be prepared on plain paper
or on a printer layout worksheet such as this one .

••• t ..

.1

PAGE _ _
'_,OF __

.t._____

I

>0.0." •

Figure 2.

Layout of a proposed report.

The job is to prepare a combined trial balance from the punched card trial balance cards
of subsidiary companies. The balancp cards are sorted into account number order.
The output should be labeled with the date of the trial balance and the nate of preparation.
Each page. including the first. should be numbered. The report should list the account
cards for each account number. showing the account number. company code. date. and
amount. Credit balances should be labeled "CR". The total for each account numher
should be labeled.
Acct No
XXXX
XXXX

Company
XX
XX

Date
XX/XX/XX
},."X/XX/XX

Total

XX. XXX. XXX. XX
XX. XXX. XXX. XX
XXX. XXX. XXX. XX

Separate totals should be kept of the debits and credits.
the enn of the report together with a net total.

These totals .,hould be sh,.\vn at

Figure 3. Specifications for a report. in lieu of report layout.
.71

Preparing a Request for Proposals (Contd.)

(9)

8/67

(d)

special security and ('ontl'Ol

(e)

accuracy specificatiuns.

(f)

alternative methods allowed.

~pcclfications.

Acceptance testing requirements such as a test run.

A

AUfRBAcH
~

(Contd. )

23:090.720

SPECIAL REPORT

PUNCHED CARD LAYOUT _C:..o~m:..b:.:ln::;ed;:....T~r..;;lal;;.;...;B:.:a1::;an=.e~_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __

1. General Ledller Year-End Balance Forward Card.
ACCOUNT

#

g

DATE

M

TYPI:

AMOUNT

(123166) (BALFIoIIO)

p

NOTE

X In column 30 meana credtt

A
N
Y
C

0
0
E

.

999999 99 999999 999999 ~ 9 999 999 9 9 99999999999999999999999999999999999999999999999999
I I

I

4 ••

,

•

10 II Iii II •• • .1 17' .. '1 10 1121 ill . . .

... H'.

",

~H"'4M.UH.~4IU~~.~~

• • • ~UUMN.uH.~.a"M.M~ • • M~~nM~~"Nn~

2. __________________________________________________________

99999999999999999999999999999999999999999999
, ••••• , ••

G"~~

••

~~

••••

U"M." ... n_.MNM • •

~M

••~ ••

Figure 4. Specifications for input from cards .
. 71

Preparing a Request for Proposals (Contd.)
This request for proposals should be sent to those organizations which have passed an
initial screening test based on their experience with similar problems. reputation,
financial ability, etc. A week to ten days is usually sufficient for a data processing
center to respond.
If the data processing problem is standard and well defined, the request for proposals can
be used for obtaining firm proposals from service centers. If the problem is such that
considerable systems work is required for the center to make a proposal, the request for
proposals may be used to select a suitable organization with which to work. Where the job
is a small or medium-sized one, a center can be expected to make a suitable proposal if
the job is well-defined in the request, but it may not be willing to invest large amounts of
design time to make a competitive proposal.

.72

Evaluating a Proposal
Figure 5 is an evaluation form listing points which are generally important in evaluating a
data processing service organization. Based on the criteria which the user considers
important for the particular job to be run, a computer service organization can be chosen
from among those which have passed the initial screening. Note that the evaluation form
requires some information which can be obtained only by visiting the data processing
service center. Such a visit is strongly recommended.
A service center should certainly be competitive in its price quotations, but the integrity
and profeSSional character of the operation are equally important. In the course of a

continuing relationship there will be special reports to be prepared, re-runs, changes,
etc., and one should be able to have confidence in the organization.
No weights or method of rating is given in Figure 5. This is left to the user's own
judgement on the basis of his particular needs. In a common approach, the user assigns
an importance weight to each of the criteria and gives each organization a rating based
on its position with respect to each of the items .
. 73

Completing the Negotiations
After a tentative selection of a service organization, a written agreement or contract
should be prepared. Many service centers use a memorandum of understanding in lieu
of a formal contract. The checklist in Figure 6 indicates the types of items to be
considered.

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

8/67

AUERBACH STANDARD EDP REPORTS

23:090.800

Organizations Being Rated

Item

--

--

--

--

Experience with similar problems
Availability of general program packages
Reputation and recommendations
Financial stability
Ratings using the
method preferred by
the user (see text)

Quality of staff
Quality of sales and account representative
Availability of control safeguards
Backup provisions
Proximity and convenience
Quality of proposal
Amount of work the center subcontracts
Time -of -completion quotation
Cost quotation
Ability to meet time and price quotation
Potential for handling requirements in future

Figure 5. Rating form for evaluating a data processing service organization .

.8

CONTROL OVER IMPLEMENTATION
When a decision has been made to accept a data processing organization's proposal.
decisions must be made regarding:
(1) Scheduling of the conversion to the new system.
(2) Conversion of the master files to machine-readable form.
(3) Procedures for acceptance testing.
(4)

Period of parallel operation.

(5) Operating procedures .
. 81

Scheduling
A schedule of events and dates for completion in order to achieve the conversion should be
prepared and should be then used as a basis for reporting progress .

. 82

Control Over File Conversion
One of the major jobs to be undertaken is the conversion of paper document files to
machine-readable files. This usually involves substantial keypunching. In some cases.
this may require the extracting of information from the documents in order to make it
usable for the keypunch operators.
As in the normal operating procedures. irreplaceable files should not be transmitted to
the data center unless there are adequate provisions for reconstructing the documents in
case of loss. If these provisions do not exist. then copies should be made and sent to the
data center for keypunching. Make sure the information transmitted is current. Spot
checks should be made for accuracy of conversion.
As previously mentioned. very large keypunch jobs are frequently sent to England for
punching because of lower labor costs. In such cases. the service center will usually
supervise these contracts.
(Contd. )

8/67

A

AUERBACH
~

SPECIAL REPORT

23:090.830

(1)

Processing to be performed.

(2)

Content and format of input and output.

(3)

Bursting and binding of reports.

(4)

Break-in or parallel runnmp, period.

(5)

Procedures for handlinf\ errors.

(6)

Your responsibilities.

(7)

Person responsible at service center for client contact and person in your organization who is
authorized to deal with the service center.

(8)

Charges including cost of reruns and changes if requested:

Pagination of reports.

(a)

Set-up or one-time programming charge.

(b)

Single charge for control panel and wires.

(c)

Supplies, if additional cost.

(d)

Extra charge for weekend or overnight premium.

(e)

Charges for pickup and delivery.

(f)

Charges for storage, shipping, etc.

(9)

If the contract is not a fixed price contract, a schedule of allowable and non-allowable costs.

(10)

Form and frequency of billings.

(11)

Program ownership (if a specIal program is being written).

(1~)

Delivery of documentation in connection with progress billings. in cases where a special program is
written and billed for.

(13)

Acceptance procedures for speclal programs.

(14)

Conversion of master files to machine-readable form

(15)

Time schedule for converting to tilt; service center processing.

(16)

Work space at sernc£' center if re'luired by client.

(17)

Liability of E>erVlce center for lost data or files and errors in process mg.
center to cover their lIability.

(18)

Security provisions for filf's and records.

(19)

Backup provIsions hy sen'ice cent"r'

(~()

Renewal and cancellation

Insurance carried by

Figure 6. Checklist for completing the negotiations .
. 83

Sample Hun for Acceptance Test
One of the procedures necessary before acceptance is the processing of a sample run in
order to resolve any difficulties before processing begins on a regular basis. The processing of a sample run is especially important where a special-purpose program has been
written .

. 84

Running in Parallel
It is typically not feasible to transfer to the n!'w system Without a brl'ak-in period during
which both the old and new system are operating in parallcl. Except for simplest cases,
a period of dual processing is necessary in order to instruct pf'rsonnel in the new procedun's and to become acquainted WIth error procedures, the new report formats, and
other questIOns whrch may arise. The period of parallel operation is usually two or three
processing periods,
It, 1967 AUERBACH Corporation and AUERBACH Info, Inc

8/67

23:090 • • 50

AUERBACH STANDARD EDP REPORTS

. 85

OperatiDi Procedures
The operating procedures for user personnel will include instructions covering the
following respons ibiUties :
(1) Form of data to be prepared; editing to ensure completeness and legibility.
(2) Schedule for preparation.
(3) Type of control figures and transmission controls.
(4) Person responsible for preparing controls.
(5)

Person responsible for all contacts with the service bureau.

(6) Procedures for handling errors in input data.
(7) Person responsible and procedures for checking output received agamst control
totals.
(8) Procedures for handling delays. errors, or failures in data proceSSing .
.9

SELECTING A TIME-SHARING SERVICE CENTER

.91

General Description
Time-sharing makes possible the simultaneous access to a central computer by many userseach of them operating a different computer program, communicating data and requests
from remote locations, and receiving on-line responses. Each user has his own input-output
device, such as a teletypewriter, which is connected to the computer by a communications line. Each user can make demands at random intervals. This is in sharp contrast
to the scheduled, sequential service offered by the standard service arrangement.
Time-sharing systems may be classified as either computation (SCientific) or record
processing (business) oriented. Most of the time-sharing systems currently operating
are scientific and are designed for use with problems that can be solved by a set of
computational procedures. This class of problems is characterized by very low volumes
of input and output relative to the computations.
Business time-sharing is usually of the opposite character, with many records, each of
which receives relatively little processing. Business data processing applications are
characterized by problems of file management, forms design, report layout, procedures
preparation, and somewhat specialized programming. There are currently very few
time-sharing services for business data processing, and it is too soon to determine how
time-sharing will work out for this type of processing. It is somewhat clearer that timesharing will be significant for small-job scientific processing and for educational use.
Time-sharing users may have a fixed set of programs available to them, or they may be
able to write their own programs from the terminals. In the case of a fixed set of
programs, the processing procedures to be used are fixed, whereas with an open set of
programs, the user can change the processing procedures merely by changing his program.
The fixed set of programs is likely to be more common in systems oriented toward
commercial data processing, whereas the scientific time-sharing systems need to allow
new programs to be written each day, some of which are saved for future use while others
are run and discarded.
Time-sharing is more expensive than regular service center processing, so that other'
factors such as the benefits of immediate response and the ability to alter programs and
interact with the computer form the economic basis for using time-sharing. As examples
of charges, a teletypewriter terminal costs approximately $100 per month, and monthly
minimum charges for the time-shared processing may range from $350 to $800 in those
cases where minimum charges apply.
Time-sharing systems have so many unique characteristics that this article will only
summarize the principal considerations in selecting time-sharing service, and a detailed
study of time-sharing will be the subject of a subsequent Special Report.

. 92

Selection Considerations
Time-sharing computer service is available to anyone who has a telephone. The economic
considerations of long-distance line charges remove from serious consideratIOn timesharing services that are a great distance from the user. If there is a need for on-line
processing, the following are some of the selection considerations:
(1) Response time. A time-sharing system can become overloaded if too many
users are connected to the system. Certain times of the day may present a
serious problem. An investigation should be made into the response time and
the times of day when overload is likely to occur.
(2) Terminal devices. Virtually all time-sharing services provide typewriter inputoutput. If the user requires higher input or output speeds, it is necessary to
make sure that the particular computer system will accept this.

8/S7

A '.

AUERBACH

(Contd. )

23:090.920

SPECIAL REPORT

.92

Selection Considerations (Contd. )
(3)

prO~ramming lan~ages. If the primary purpose of the installation is to do
prolems that can expressed as a set of algebraic procedures. then it is
most important to have access to an algebraic compiler language such as
FORTRAN, ALGOL. or BASIC. If the service center does the programming,
cost estimates should be obtained. The number of standard packages available
to facilitate programming is important. If the processing is primarily of a
research nature. then it is important to have statistical packages so that
programming will not be required for these. If the processing is primarily
commercial. then the existence of prewritten commercial data processing
programs is desirable. However, it frequently happens that a general-purpose
program is unsatisfactory for a particular user; therefore. it is important to
examine whether or not it is possible to easily modify a standard routine furnished by the service center.

(4) Assistance in systems deSign. The typical scientific user of a time-sharing
terminal requires little assistance. A small amount of training is all that is
necessary for him to be able to work out his formula-type problems at a timesharing terminal. In the case of commercial data processing. this will not
usually be true. There are forms design. handling procedures. special programs,
input-output problems, error recovery procedures, and other problems which
require assistance by the service center. The amount and quality of assistance
furnished by the time-sharing service is therefore an important consideration.
(5) Control features. Before committing oneself to a service center and entrusting
data files to it, the user should be satisfied that proper precautions have been
taken against file destruction and against unauthorized use. He also should be
satisfied that, in the event of system failure, there is adequate provision for file
reconstruction. When beginning to use the service center, the same principle
applies as in other computer installations. There should be a period of dual operation during which manual reconstruction will be possible in the event of a failure of
the computer system. This dual operation should not continue for any extended
period of time but should be available during a break-in period of two to four weeks.
(6) Financial integrity. As noted for conventional data processing vices, the failure
of a time-sharing operation will endanger files and future processing, so there
should be an assurance that the operation is properly funded and is likely to
continue in operation.

«:11967 AUERBACH Corporation and AUERBACH Info. Inc.

8/67

•

23: 100 001

A

.utlllAeM

EDP
lI.un

SPECIAL REPORT
DATA COMMUNICATIONS - WHAT Irs ALL ABOUT

PREPARED BY
F H REAGAN. JR.
TECHNICAL STAFI:'
AUERBACH CORPORATiON

© 1966 AUERBACH Corporation and AUERBACH Info, Inc.

4 66

23: 100. 100
SPECIAL REPORT
DATA COMMUNICATIONS

SPECIAL REPORT
DATA COMMUNICATIONS - WHAT ITS ALL ABOUT
.1

INTRODUCTION
Data communications IS a new and rapidly expandlnl!, twIrl that has emer~ed
from a wedding of the communicatIOns and data proct'ssln~ technologies. The
need for rapid, accurate transmission of data hetween the wirlely scattered plants
and offices of modern corporations has imposed stron~ pressu res upon both the
communications common carriers and the computer manufacturers to develop
the necessary techniques and equipment. ImpreSSive progress has been made
durin~ the last few years, so that now nearly every company can find transmission facilities and equipment that will effectively fulfill its data communications needs.
U. S. industry is recognizing, at an ever-increasing rate, the advantages of
company-wide data communications networks and of the closely related concepts
of real-time data proceSSing and integrated management information systems.
Although only about 1 percent of the computers sold in 1965 were linked to a
data communications system, Western Union has predicted that 60 percent of
the computers likely to be sold in 1975 will be so linked. A. T. &T. expects
that the volume of information transmitted In the form of digital data will c\'entuall~' equal the volume tl'ansmitted hy vOice.
A data communications system can be conSidered to consist of a group of functional units whose primary purpose is to transfer digital data between two or
more terminals in a reliable manner. Each unit has a specific set of functions
to perform; the exact functions and the sequence and manner in which they are
enacted are determined by the overall system requirements. Because system
requirements vary from business to business and from application to application, the data communications systems in use today vary Widely in their functions, their structures, and their degree of complexity. Some systems transfer
messages between remote terminals via one or more switching centers where
communications processors are located; other systems transmit inquiries
from numerous remote terminals to a central data processinl: facility, which
generates responses and routes them back to the inquiring terminals. The
design of systems such as these demands a thorough knowledge of both communications and data processing technology.
This report provides an introduction to the concepts and techniques that should
be understood by every prospective user of a data communications system. The
sections that follow describe the types of applications in which data communications systems are being effectively employed, the factors to be considered in
designing a system, the various components of a system and their functions, and
the communications facilities and services provided by the common carriers.

This Special Report, which constitutes an IntroductIOn to the concepts and
equipment involved in th" deSign of modern data commUnIcations systems.
IS also appearing as a feature article In the April 1966 Issue of Data Processing Magazine. The repot-t 18 based upon materIal extracted from the
System DeSign section of AUERBACH Data Communications Reports. another analytICal reference service from AU ERBACH Info, Inc.
AUERBACH Data CommunicatIons Re orts IS deSigned to prOVide the specIalize I ormatIOn that computer users need In order to understand and apply
the current technology and new developments In the rapidly expanding heM
of data communications. Definitive reports and comparison charts describe
the characteristics of commercially available communications terminals
and processors, the data communications faCilities provided by thE' common
carriers, and systematIc techniques for deslgrung data communicatIOns systems and selecting equipment. Regular supplements keep the service comprehensive and up to date. For more Information about AUERBACH Data
Communications Reports, please write or phone the publisher: AUERBACH
Into, Inc .• 121 North Broad Street, Philadelphia. Pa. 19107 (Area Code 215.
LO 7-2930).

© 1966 AUERBACH Corporation and AUERBACH Info, Inc.

4/66

23: 100.200

SPECIAL REPORT

.2

APPLICATIONS
Current applications of data communications systems vary widely in their functions, their scope, and their equipment and programming requirements. New
applications are being developed every day, and It would clearly be impossible
to describe, or even list, all of the specific applications In which data communications equipment is being used. A more rational approach is to divide the
total spectrum of data communications applications into a few fundamental
"application classes. " each performing a certain general function and involving
a certain type of data flow pattern. Most specific applications will then fall
neatly into one application dass or combine the functions of two or more dasses.
Although coarser or finer breakdowns could be justified, it seems reasonable
to consider six fundamental application classes. The function and data flow
pattern of each of these classes are described in the paragraphs that follow .

. 21

Data Collection
The function of this class of applications is the collection and transmission to
a central processing pOint of information concerning the operations of geographically separated manufacturing plants, warehouses, branch and regional
sales offices, and other outlying facilities. The basic data flow pattern is
unidirectional, from multiple remote (and/or local) terminals to the central
processing faCility. This type of system can: (1) prOVide the complete,
timely information about a firm's overall operations that is required for
accurate cost control and Informed management decisions, and (2) reduce
the number of times and places at which data must be manually handled and
transcribed, thereby cutting derical costs and error rates .

4/66

. 22

Data Distribution
In this class of applications, the principal function is the distribution of data
generated and/or processed at a central facility to one or more outlying
locations. Aj1;ain the basic data flow pattern is unidirectional, from the
central facility to the remote (and/or local) terminals. This function, of
course, is the complement of the data collection function described in the
preceding paragraph, and many data communications systems combine the
collection and distribution functions. To appreciate the potential value of a
data distribution system, it is necessary to realize that data has no real value
until it has reached the actual point of application in a useful form. Significant financial benefits can frequently be realized through cutting down the
elapsed time and improving the accuracy of the data dissemination process .

. 23

Inquiry Processing
To meet the competitive demands of modern bUSiness, many firms are
finding It desirable (and in some cases essential) to "go on-line" by establishing central data files that can be randomly acce~sed to provide prompt
responses to inqulriE'sfrom outlying locations. In this class of applications,
the basic flow pattern is bidirectional; inquiry messages are transmitted
from a network of remote terminals to the central proceSSing facility, and
appropriate response messaj1;es are generated and transmitted back to the
inquiring terminals.
The inquiry processing function is frequently combined with real-time file updating;
the appropriate entries in the central data files are modified each time a transaction
occurs so that the central files always reflect the true current status of the business.
Although inquiry processinj1; and real-time file updating systems promise great benefits for nearly every type of business organization, their advantages in terms of
faster response and centralized control should be carefully weiv,hed against their
costs to ensure that the higher direct cost of a real-time system, as compared with
that of a more conventional batch-type proceSSing system, is worthwhIle. Realtime inquiry systems are especially beneficial for organizations such as banks,
brokerage firms, airlines, and hotels, where prompt servicing of customer inquiries is of critical importance .

. 24

Computer Load-Balancing
Organizations that have two or more computers in geographically separated locations may find it advantageous to connect them by means of communications links.
This permits more effective utilization of each of the interconnected computers
because the slack time in one computer's schedule can be used to help smooth out
the peaks in another's. Reliability is greatly enhanced because the communications
links make it easy for one or more computers to take over another computer's
workload when a breakdown occurs. The data flow pattern in this class of application is bidirectional; input data and results are transmitted between each pair
of interconnected computers, and the volume of data flow depends upon their relative workloads at any given time.

A

(Contd. )

AuERBACH

•

23:100.250

DATA COMMUNICATIONS

.25

Computer Time-Sharing
In an effort to make the facIlities of a computer system conveniently available to
multiple users, extensive development work is in progress on "time-sharing"
systems. The design objective of a time-sharing system is to furnish continuous
computing service to many users simultaneously, while providing each user with
virtually instantaneous responses. Multiple consoles, each equipped with appropriate Input and output facilities, are employed, and each console is connected
to the central computer facility by a communications link. (Some or all of the
consoles are likely to be close enough to the central facility so that direct cable
connections can be used.)
The basic data flow pattern in a time-sharing system is bIdIrectIOnal and similar to the pattern for th£' inquiry processing class of applleatlOns; input data
and operating Instructions are transmitted from the ('onsol£'s to the central
computer facility. and the results of computations are transmItted back to
the appropriate ('onsol£'s. The widely-discussed "public utility" computer concept, in whIch multiple suhscribers would share the facilities of a gIant centralized computer comple" on a toll basis, IS a logICal extensIOn of the computer
time-sharing class of applicatIOns .

. ~6

Message Switching
The activities of a modern cOI1loration tend to be spread out over a large number
of widely separalt'd locatIOns, and an efficient system for handling communicatIOns
among all these locatIOns IS VItally necessary. Where commUnications traffic IS
high, a computer- controlled message SWItching system is likely to be the best
overall chOICe. In thIS type of application. the data flow pattern involves twoway message traffIC between a number of termmals and a central switching
center. The sending terminal transmits each message to the center, which
stores it temporarily, performs any processing or code conversion functions
that mayb£' requI\'(·d. and then transmits the messa!1;e to one or more desi!1;nated
receiving termmals. Large networks may utilize two or mor£' switchin!1; centers which ar£' mterconnected by high-speed commUnicatIOns links .

.3

SYSTEM DESIGN
The installation of a data communications system should always be preceded by
a thorough study and re-evaluation of the patterns of information flow throu!1;hout
the organization. r.loney spent in SImply mechaniZIng the existing procedures
for collecting, transmitting, and dlssemmatlng information is likely to be largely
wasted. The real purpose and need for every type of information that is currently being transmitted should be questioned. It is likely that most executives
are regularly receiving some information that IS of little or no value t9 them,
while failin!1; to receive other information which could aId Significantly in
deCiSion-making and cost control, and which could easily be provided if the need
wert: recognized. In some cases, the improve'
420: 102

]:j J

4·66

SPECIAL REPORT

23: 100. 420

.42

Input/CAitput Control {Inits (Contd.)
terminal, the output control unit accepts the received data. stores It temporarily. alld lupplles It to the output devices at the appropriate rate. In control
units which terminate more than one line. the type and capacity of the buffer
storage II a primary con('ern because it determines the frequency at which
each line must be servICed. Various types of buffers are available. such as
magnetic core memoriPR. magnetic drums. transistorized shift rpglsters.
and delay lil'les.
It should be noted that not all data communicatIOns termmals employ buffered

input/output control umts. \Vhen no buffers are used. the input. data transmission. and output functIOns must procl'ed simultaneously and at thp same
speed.
Complex data commumcatlOns systpms that termlOate many lines 10 a central
facility usually usp either a multl-hne communir.ations controller 10 conjunctIOn
with a genl'ral-purpos!' computer or a speclalizE'd. stored-program communicatIOns processor Tht's!' umt!' are ('apable of buffering and controlling simultaneous input/Ql·tput transnlJSSlons on many different lmes. AgalO. a wide variety of
equipment IS now available to pprform these functions. The available devices
differ 10 thE' number and speed of lines they can terminate and in their potential
for performing auxiliary or indE'pendent data processing. Examples include the
three multi-line communications controllers available for use with the generalpurpOSe IBM System/.160 computers and the Collins Data Central system. a computer system designed espeCIally for message switching applications.
AUERBACH Standard EDP Reports contains descriptions of most of the communications control equipmE'nt that is available for use in conjunction with generalpurpose computer systpms. Table II lists some of these controllers and the
report sections where they are described. More detailed analyses of an even
broader range of communications equipment can be found in AUERBACH Data
Communications Reports. the specialized AUERBACH Info, Inc. reference
service for designers and users of data communications systems .
. 43

Error Control Units
The primary purpose of a data communications system is to transmit useful information from one locatIOn to another. To be useful. the received copy of the
transmitted data must constltute an accurate representation of the original input data, Within the accuracy lunits dictated by the application requirements and
the necessary economic tradeoffs. Errors Will occur in every data communications system. ThiS basic truth must be kept m mind throughout the
deSign of every system. Important criteria for pvaluatmg the performance of
any communications system are its degree of freedom from data errors, its
probability of detecting the errors that do occu r. and Its efficiency in overcoming the effects of these errors.
Errors In the receivpd mt-ssages which form the output from a data commulUcations system can result from:
•

Operator errors In prepanng the mput or
or recelvmg terminal.

•

Malfunctioning of the transmitting or receiVing tE'rmlnal pqulpment.

•

Malfunctioning of the commUnications lines. due eIther to random
pulses interfering with data transmiSSion or to a more permanent
condition. such as complE'te failure of the line.

In

opt-I'atlOg the transmitting

Techniques whICh merely det!'ct and indicate errorR are generally less complex
and expensive than techniques which detect errors and then correct them. In
most error control schemes. the digital data at the transmitting termlOal
is encoded to conform to some set pattern. At the receiver, the data is dE'coded
and checked to see whethE'r the received data pattern conforms to the prescribed
rules.
There are two basic. commonly-used methods for automatic checking of data:
validity and redundancy. A validity ('heck ascertalOs whether each data code
is one of a number of permitted bit configurations; this checking is usually
performed on a character basll'l. and any code configuration which does not
represent a legitimate member of the character set is considered an error. In
redundancy checking, one or more additional bits are added to ea('h data configuration in accordance with a specific formulation rule. Checking is accomplished by testing the additional bits to see whether they still conform to the
formulation rule. The most common form of redundancy checking IS parity

4/66

A.

AUERBACf1

(Contd. )

23: 100. 430

• DATA COMMUNICATIONS

TABLE 11: REPRESENTATIVE DATA COMMUNICATIONS CONTROLLERS
Manufacturer

Burroughs

Equipment

Reference

B 100/200/300 Series Data
Communications System
B 5500 Data CommunicatIOns System

203:101

Control Data

:1276 Communication Terminal Controller
6600 Series Data Set Controller

260: 102
260: 103

GE

Datanet-aO Data Communications Processor
Datanet-70 Communications Controller

330'104, 340'101
330'105

Honeywell

281 Single-Channel CommunicatIOn Control
286 Multi-Channel Communication Control

510: 103
510: 104

IBM

1009
1448
2701
2702
2703
7710
7740
7750
7770
7772

401:101
414: 103
420:106
420: 107
420: lOB
401: 106
414: 106, 420: 106
402: 105, 420: 106
420: 103
420: 104

NCR

Teletype Inquiry System
On- Line Savings System

601: 105
601: 106

RCA

3378 Communications Mode Control
3376 Communications Control
70/652 & 70/653 Communication Controls
70/668 Communication Controller
(Multic hannel)

703: 101
703:103
710:101
710:102

UNIVAC

Commumcation Terminal Module Controller
Word Terminal Synchronous
Communication Terminal Synchronous

785:102, 790:101
785: 102, 790'101
785:102, 790: 101

.43

Data Transmission Unit
Transmission Control Unit
Data Adapter Unit
Transmission Control
Transmission Control
Data Communication Unit
Communication Control System
Programmed Transmission Control
Audio Response Unit
Audio Response Unit

~01: 103

Error Control Units (Contd.)
checkmg. in which the total numher of "1" hIts In a data ('onfi~lratlOn of some
arbitrary length is required to he eIther even or odd. Panty checkIng can be
performed on a character basis. on a message basis. or hoth
Error correctIOn procedures may be fully automatIC, or thev may require extensive manual interventIOn by the operators The most common method of error
correctIOn IS retransmIssIon of eIther the complete message or IndIVIdual segments of it untIl the entIre message has heen reCCI\ ed with no detected errors .

. 44

Synchronization {TOlts
Because the data signals are time-dependent (i. e .. the bits are transmitted at
precise time intervals), some means must he provided to ensure synchronization between the transmitting and receiving stations. Two commonly-employed
techniques are referred to as "start/stop synchrOnIzation" and "synchronous
transmission. "
In the start/stop technique, extra signals are transmitted with each character
of data to identify the beginning and the end of the character. The data bits
within each character are transmitted in a strict time sequence, but characters
are transmitted asynchronously, i. e., there is no definite time relationship
between the transmission of successive characters. The advantages of this
method are that it allows data transmission from sources with highly irregular
data input rates (such as manual keyboards), and that the probability of cumulative
errors in synchronization is minimized. The disadvantage of start/stop synchronization is that it increases the required line capacity due to the extra start
and stop bits that need to be transmitted along with the data bits.

© 1966 AUERBACH Corporation and AUERBACH Info. Inc.

4 66

23: 100.440

SPECIAL REPORT

.44

Synchronization Units (eontd.)
In the synchronous transmission tt'chnique, a specific character IS transmitted
to the receiving terminal, which interprets the character and adjusts its
synchronizing circuitry to conform with the transmitted bit rate. The synchronous method is sometimes r!'ferred to as "bit stream synchronizatIOn"
The advantage of thlll tvp" ot synchronization is that It permits higher data
tl'.lIlsmisslOn rat!'s than the start/stop method; the disadvantage is that It
1"t'qulres hIghly preCise and relatively expensive circuitry to mamtam synchronizatIOn throughout the transmissIOn of long messages.

4;'

l\1odulation-Oemodul..tlOn l'lllts
TIU' puis., slgn,d~ gt'llt'rated by busIness machines usually ne!'d to he mcxllfled
to ohtaln gr.'alt'r transmISSion !'fficlency and compatlbil1ty with common-carn!'r
,'OIl1Il1Ulllcatlons tacIiltl!'s. Thl' unit utied at the tr;J.nsmlttlng terminal to ac"'>llll'ltsh thl~ mo80
1,566,000
630
730
1,957,50.0
2,349,000
830
3,132,000 1,030
3,915,000 1,230
4,008,000 1,430
6,264,000 1,830

Thin-FUm Memory (0. 6-microsecond cycle,
for use
B 6506 or B 7506 Basic
System)

'f},th

B6001-3
B6002-3
B6003-3
B6004-3
B6005-3
B6006-3
B6007-3
B6008-3
B6010-3
00012-3
B6016-3
B6020-3
B6024-3
00032-3

98,304
196,608
294,912
393,216
491,520
589.824
688,128
786,432
983,040
I, 179,648
1,572,864
1,966,080
2,359,296
3, 145,728

Bytes
Bytes
Bytes
Bytes
Bytes
Bytes
Bytes
Bytes
Bytes
Byte.
Byte.
Byte.
Bytes
Bytes

Thin-FUm Memory
Thin-FUm Memory
Thin-FUm Memory
Thin- Film Memory
Thin-FUm Memory
Thin-FUm Memory
Thin-FUm Memory
Thin- FUm Memory
Thin- FUm Memory
Thin- FUm Memory
Thin-FUm Memory
Thin-Film Memory
Thin-Film Memory
'l'hm- Film Memory

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

BURROUGHS B 6500/7500

204:221.102

PRICES

IDENTITY OF UNIT
CLASS

Model
Number

B6713
B6713-1
B6714
B6714-1
B6716
B6716-1
B7714
B7714-1
B7716
B7716-1

I Monthly

Monthly
Rental IPurchase Maint.

$

$

2,800

117,600

140

500
3,200

21,000
134,400

20
160

550
3,200

23,100
134,400

25
160

550
4,800

23,100
201,600

25
240

650
4,800

27,300
201,600

30
240

650

27,300

30

850
1,200
4,950
900

44,000
57,600
237,600
45,600

115
125
595
110

3,950
700

224,200
44,840

495
85

2,700
450

129,600
28,800

395
85

650
650
650

31,200
31,200
31,200

80
80
80

650

31,200

80

650

31,200

80

400
100
200
350
10
15

16,800
4,200
8,400
14,700
420
630

20
10
10
10
2
2

900
1,100
1,300
1,100
1,400
1,700

43,200
52,800
62,400
52,800
67,200
81,600
43,200
52,800
62,400
16,800
18,900
16,800
8,400
8,400
1.050

200
230
260
225
260
295
200
230
260
15
15
15
10
10
5

$

Multiplexor with 4 Data Switching
Channels, for B 6503
Additional Data Switching Channel
Multiplexor with 4 Data Switching
Channels, for B 6504
Additional Data Switching Channel
Multiplexor with 4 Data Switching
Channels, for B 6506
Additional Data Switching Channel
Multiplexor with 6 Data Switching
Ch8llJlels, for B 7504
Additional Data Switching Channel
Multiplexor with 6 Data Switching
Channels, for B 7506
Additional Data Switching Channel
Disk storage

MASS
STORAGE
B9372-1
B9372-7
B9375-0
B9376-0
B9375-2
B937G-2
B9375-3
B9376-3
B9371-1
B9371-2
B9371-3
B9371-4
B9371-5

,

B6373
B6471-1
B6471-2
B6471-4
B6674
B6675

INPUTOUTPUT

2/69

Name
I/O Multiplexors and Ch8llJlels

ATTACHMENTS,
ADAPTERS
AND
CHANNELS

0

Feature
Number

10 Million Bytes, 20 msec
20 Million Bytes. 23 msec
100 Million Bytes, 23 msec
Additional 20 Million Byte Increment,
23 msec
100 Million Bytes, 40 msec
Additional 20 Million Byte Increment,
40 msec
100 Million Bytes, 60 msec
Additional 25 Million Byte Increment,
60 msec
Disk File Electronic Unit for B 9372-1
Disk File Electronic Unit for B 9372-7
Optional Additional DFEU for B 9375O/B 9376-0
Optional Additional DFEU for B 93752/B 9376-2
Optional Additional DFEU for B 93753/B 9376-3
Disk File Control (1)
Disk File Exchange; 1 x 2(1)
Disk File Exchange; 2 K 5 (1)
Disk File Exchange; 4 x 10 (1)
Exchange Adapter for B6471-2m
Exchange Adapter for B6471-4
Magnetic Tape

,

B9381-2
B9381-3
B9381-4
B9382-2
B9382-3
B9382-4
B9383-2
B9383-3
B9383-4
B6381-1
B6381-2
B6381-3
B6480
B6481
00681

Tape Clusters:
36KB Cluster; 2 drives
36KB Cluster; 3 drives
36KB Cluster; 4 drives
72KB Cluster; 2 drives
72KB Cluster; 3 drives
72KB Cluster; 4 drives
9/25/36KC Cluster; 2 drives
9/25/36KC Cluster; 3 drives
9/25/36KC Cluster; 4 drives
36KB Cluster Control g~
~KB Cluster Control
9/25/36KC Cluster Control(l)
Cluster Exchange; 7-track, 2 x 8m
Cluster Exchange: 9-track, 2 x 8
200-bpi Adapter for B6381-1 Control

A

AUERBACH
8

900

1,100
1,300
400
450
400
200
200
25

PRICE DATA

PRICES

IDENTITY OF UNIT
CLASS

Model
Number

I

Feature
Number

Name

Monthly
Monthly
Rental ~rchase Maint.
$
$
$
r

IN PUTOUTPUT
(Contd. )

Magnetic Tape
B9391
B9392
B9393
B9394-1
B9394-2
B6391-3
B6391-4
B6393-1
B6393-2
B6393-3
B6490
B6492
B6691
00692

~on1d.)

Free-standing Magnetic Tape Units:
lS/50/72KC Magnetic Tape Unit
72 KB Magnetic Tape Unit
144KB Magnetic Tape Unit
24/66/96KC Magnetic Tape Unit;
96KB MagnetIc Tape Unit
18/50/72KC Magnetic Tape Control(l)
24/66/96KC Magnetic Tape Control(l)
72KB Magnetic Tape Control
144 KB Magnetic Tape Contro~
96 KB Magnetic Tape Control )
Tape Exchange; 2 x 10(1)
Tape Exchange; 4 x 16(1)
200 bpi Adapter for B 6393 -1 Control (1)
200 bpi Adapter Cor B 6393 -3 Control (1)

(lh)

575
575
650
650
650
500
550
500
600
550
250
450
25
25

27,600
27,600
31,200
31,200
31,200
21,000
23,100
21,000
25,200
23,100
10,500
18,900
1,050
1,050

165
165
175
170
170
15
15
15
15
15
10
20
5
5

325
450
100
5

16,250
21,600
4,200
240

83
126
16
2

530
100
15
5

25,440
4,200
630
240

135
15
5
0

300
100
145
260
100
130

16,000
4,200
6,960
15,300
4,200
6,850

70
16
10
65
16
10

850
950
150
60
40
25
55
150

48,000
53,500
6,300
3,000
2,000
1,200
2,640
6,300

180
200
16
20
10
5
15
15

215
325

9,460
14,300

27
27

100

4,400

20

60

2,640
880
1,980

10
3
10,
3
7

Punched Card
B9111
B9112
00110
B9916
B9213
B6210
B6610
B9910

Card Reader (BOO cards/min.)
Card Reader (1400 cards/min.)
Card Reader Control(1)
Validity Check Switch and Indicator
(for B 9111 or B 9112)
Card Punch (300 cards/min. )
Card Punch Control(1)
BCL-BeL Code Translator for 00210(1)
Card Counter
Paper Tape

B9120
B6120
B9926
B9220
B6220
B9928

Paper Tape Reader (500-1000 char/sec)
Paper Tape Reader Control(1)
Input Code Translator
Paper Tape Punch (100 char{seC)
Paper Tape Punch Control (1
Output Code Translator
Printers

B9242
B9243
B6240
B9940
B9941
B9949
B9342
00340
COMMUNICATIONS

Printer (815 lines/min., 120 positions)
Printer (1040 limS/min., 120 positions)
Printer Control
High Speed Slew feature
Additional 12 Print Positions
Powered Forms Stacker
Console Printer and Keyboard
Console Printer Control (1)
Input and Display Subsystems

B9351-1
B9351-2
B9351-3
B9351-4
B9351-5
B9351-6
B9351-7
B9351-8
B9351-9
B9351-10
B9951-1
B9951-2

Control I - Single Input and Display/Location
Control II - Multiple Input and Displays and/or Controls/Location
Control IIA - Ma:x1mum of 3 with
Multiplexor and Control II
Monitor
AlphanumeriC Keyboard
Input and Display Printer
Remote Communications Adapter
Slmplexor
Multiplexor
Multiplexor Extension
Input and Display Printer Adapter
Insert/Delete - Character/Line

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

20

45
35
15
50
50
30
15

1,54Q
660
2,200
2,200
1,320'
660

11

' 7
,.

3
3

2/69

204:221.104

BURROUGHS B 6500/7500

IDENTITY OF UNIT
CLASS

COMl\Il.iNICATIONS
(Conk!. )

Model
Number

Feature
Number

Name

PRICES
Monthly
Monthly
Rental Purchase Maint.

$

$

$

Input and Display Subsystems (Contd.)
B9951-3
B995I-4
B9951-5

Controlled Format
Variable Tab Position
Programmatic Cursor Positioning

20
5
5

880
220
220

3
1
1

900
100

43,200
4,800

110
10

30
30
50
85
60

1,440
1,440
2,400
4,080
2,880

G
G
10
17
12

30
50

1,440
2,400

G
10

50

2,400

10

50
50
85
40
20

2,400
2,400
4,080
1,920
960

10
10
17
10
5

55

2,640

12

Data Communications Controls
136350
BG350-1

Data Communications Processor (1)
Multiline Control; I6-line increment(1)
Communications Line AdaE!ers

BGG51
BGG52
BGG53
B6G54
BGG55
BGG57
136659-1
B6659-2
B6659-3
B6660
BG661
B6662
B6669-1

Typewriter Inquiry station (1)
TWX/Remote Typewriter(l)
B 2500/B 3500 (2400 bits/second)(l)
UNIV AC DCT 2000 (1)
IBM 1050(1)
Model 35 or 8A1 Selective Calling
Service (1)
Input and DisPlad: - Direct Connect
(Multi-wire) )
Input and Display - Data Set
(1200 bits/second) (1)
Input and Display - Data Set
(2400 bits/second) (1)
B 300/B 5500 ~400 bits/second) (1)
Honeywell 120
Model 28/83B3 ( )
Automatic Dial OutC1 )

1

Remote Devices
B9350

Typewriter Inquiry System

NOTES:
(1)

2/69

The units listed are for use in 136500 systems; the correspondil!:g unit for use in B7500 systems
has "7" as its initial digit in place of "6". For example. the B6373 Disk File Control is used
in 136500 systems; the corresponding unit for B7500 systems is the B7373 Disk File Control.

fA.

AUERBACH

-1. "'.....

210:000.001
BURROUGHS B500

. . . EDP

....

~

REPORT UPDATE

IIPlIIS

REPORT UPDATE
~BURROUGHS ANNOUNCES NEW DATA PROCESSING SYSTEM

The B500 System, an extension of the B100/200/300 Series, has been announced by Burroughs
Corporation. Bridging the gap between the B300 and the B2500, the main market of the B500
centers on the high end of the small computer field, its main competitors being the IBM 360/20
and the Honeywell H-120. First deliveries are promised for the fourth quarter of 1968.
Programs written for the B300 are compatible on an assembly-language and object level with the
B500. Since the capacity to handle COBOL is offered on the B500, however, some source language compatibility is offered. The threshold user who confines his programmers to higher-level
business languages may later grow into the B2500/3500 systems, or other machines of higher
power, without painful reprogramming.
Built around the B300 Central Processor, the B500 offers a 6-microsecond memory cycle time,
with main core memories from 9,600 to 19,200 characters. New peripherals announced for the
B500 include the B9370-5 Systems Memory - a version of the head-p.er-track disk file storage
unit used on the B2500/3500 systems. The B9370-5 has an average access time of 23 milliseconds,
and a capacity of 2.4 million 6-bit characters arranged in 240-character segments. The B9245-2
and B9245-3 are low-cost low-speed line printers: 315 lines per minute, 64 character set, with
120 and 132 print positions respectively. Each printer is fully buffered. Print characters are
repeated 3 times around a continuous chain, and individual type slugs may be changed.
The following 7-channel Magnetic Tape Cluster Units are also available:
•

B9384-3, 3-Station, 9-25 KC (200,556 BPI)

•

B9384-4, 4-Station, 9-25 KC (200,556 BPI)

•

B9380-2, 2-Station, 9-25-36 KC (200, 556, 800 BPI)

•

B9380-3, 3-Station, 9-25-36 KC (200, 556, 800 BPI)

•

B9380-4, 4-Station, 9-25-36 KC (200, 556, 800 BPI)

Still another low-cost B500 peripheral is the B9131-1 MICR document Sorter-Reader, a slower
version of the document handler offered with the B2500/3500 systems (1200 DPM at a rental of
$1, 200/month versus the original 1,565 documents at $1, 900/month).
The following rental/price structure applies to a typical small magnetic tape configuration:

Central Processor (9. 6K)
200 CPM Reader
100 CPM Punch
315 LPM Printer (120 print positions)
3 Station Cluster (25KC)
Magnetic Tape Control

Monthly
Rental*

Purchase
Price

$1,015
175
350
500
950
50

$51,450
8,400
18,425
24,000
45,600
2,880

$3,040

$150,755

A basic B500 using Systems Memory would rent for $4,635 per month with a purchase price of
$233,910.
*176-Hour, One-year lease.

o

1968 AUERBACH Corporation and AUERBACH Info, Inc.

3/68

210:001:001

A

AUERBACH

5Un&ll

ED)?

BURROUGHS B 2500 & B 3500
CONTENTS

IUOITS

CONTENTS
Report 210: Burroughs B 2500 & B3500 Systems -

General

Introduction . . . . . . . . . . . . . • . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Data Structure. . • . . . . • • . • • . • • • . '.' • . . . . . . . . • . . . . . . . . . . . . . • . . . .
System COIuiguration (general) . . • • . . . • . . . . . . . . . • . . . . . . . . . . . . . . . • . .
Internal Storage Main (processor) Core Storage . . • . . • • . . • . . . . . . . . . . . . . . . . . . . . . .
B 9370 Systems Memory •••••.•••••.•......••..•.•..•.•.•...
Modular Random Storage Disk Files ..•••..•..•••..••••..••.•.••
Data Memory Bank Disk Files . • • . • . . • . . • . . . . . • . . . . • • • • • . . • . . .
Central Processors (general) . • • . • • • • . • • . . . . . . . . . . . . . • . . . . • . . • . . • .
Console System Control Panel . • . . . . • . . • . . . • . • . . . . . . . . . . . . . . . . . . . . . .
B 9340 Console Printer and Keyboard • • • . • . . . • . . . . . . . . . . . . . . • • . .
Input-Output; Punched Card and Tape B 9110 Card Reader . . . . . . . . • . . . • . • . . . . . . . . . . • . . . . . . • . . ! • • •
B 9111 and B 9112 Card Readers . . • • . . . . . . . • . . • . . . . . . . . . . . • . . . .
B 9212 and B 9213 Card Punches ••.•••••..•.•••.• , • . . . . . . . . . . . .
B 9120 Paper Tape Reader, ••••.••.•••...••••.••••..•...•...•
B 9220 Paper Tape Punch .•.•.••••.•.••......••• r •••••••••••
Input-Output; Printers B 9240, B 9241, B 9242, and B 9243 Line Printers . . . . . . . . . . . • . . • • . •
B 9244-1 and B 9244-2 Multiple Tape Listers • . • . . . . . . . . • . . . . • . . . . •
Input-Output; Magnetic Tape B 9390, B 9391, B 9392, and B 9393 Free-Standing Magnetic Tape Units
B 9381 and B 9382 Clustered Magnetic Tape Units . • . . . • • . . . . . . . . . . . .
Input-Output; others MICR Sorter-Readers . • • • • • • . • . • . . • . . . . . • • . . . . . • . . • . . . . . . . .
Data Communications Subsystems . • • . • • • . . . • . • . . . . . . . . • . . . . . . . .
Display System . • . . . • . • • . • . • . . . . • • . • • • . • . • . . . • . . . . . . . . . . .
Simultaneous Operations (general) • . • • • . • . • . • . • . • . . • • • . . . . . . . • • . . . .
Instruction List . • • • . • . • • • • . • • • . • • • . • • • . • • • • . • . . • . . . . . . . . • • . • .
Compatibility .•••••••.••••••••.•.••••••••••..•. . . . . . • . . . . . •
Data Codes ..•••••••••...•••....•..•....•••••.••••..••.•.•••
Problem Oriented Facilities ...••.•..•••.••••.••.•••..••.•.••.•••
Process Oriented Languages FORTRAN IV••••.• , ••••••.••••. , •...•••. " .••..••••.•••.
COBOL ..••••••.•...••.•.•••.•. '" •••.••...•.••...•...
Machine Oriented Languages Basic Assembly Language .•.•.•..•...•.••••••••.•••.•.••••••
Advanced Assembly Language ••••••.••••.••.•••.••••••.•..•..
Operating Environment Basic Control Program (BCP) .••••••••••••••.••••••.••••••.••
Master Control Program (MCP) .••••••••••••.••••••••••••••••.
System Performance (general) . • . • . • • • • . • • • • • • • • • • • • • • • • . • • • • . . • • .
PhySical Characteristics ..•.•••••••••••.•••.••••••.•••••••• , •...
Price Data .•.••••.••.••••••••••••.••••.•••••••••••.••••••••
Report 213: B 2500 System
Introduction. • • • • • • • . . . • . . . • • • • • • • • • • • . . • • • • • • • . . . • . . • • • • • • • •
System Configuration ••.••••.••••.•••••.•••••••.••••.•..•••••••
Central Processor . • • . . . . • • • . • • • • • • • • • • . . • . • . . • • • • . . • . • . • . . • • •
Simultaneous Operations •.•.•.•.•..•••••••••••.••••••.•••.••.•..
System Performance ••.•...••••••••.•••••••••.•••.•••••...•.•.•

C 1969 AUERBACH Corporation and AUERBACH Info. Inc.

BURROUGHS B 2500 & B 3500

210:001:002

Report 214: B 3500 System
Introduction. . . . • . . . • . • • • . . . . . . • . • • . • . • . • • • . . . • • . • • . . • . . .
System Configuration . . • • . • • • . . . . . . . . • . . • . . . . • • • • . . . . • • . . . . . . • .
Central Processor . . . • • • • . . . • . • • . . • • • • • . • • • • . . • . . . . • . . . . . . . . . .
Simultaneous Operations . • . • . • • . • . . . . . • • . . . . • . . . . • . . • . . • • . . . . . • .
System Performance •••.••..••.......••••..••••••......•.....••

214:011
214:031
214:051
214:111
214:201

A medium-sized B 3500 random-access system with three modules of Disk File
storage and two on-line printers.

The B 2500/3500 Central Processor
features digital displays of register
contents (Section 210: 061).
5/69

This Clustered Magnetic Tape Unit
contains four independent tape
transports (Section 210:092).

A

AUERBACH

"

210:011. 100

A

AUERBACH

STAMQUD

EDP

BURROUGHS B 2500 & B 3500
SUMMARY

REPORTS

~

SUMMARY
.1

SUMMARY
The B 2500 and B 3500 systems constitute Burroughs Corporation's current entry in the
small-scale computer sweepstakes. The two systems are definitely "third generation" in
their use of monolithic integrated circuits, and read-only memory. Their processing
power in business applications is impressive in terms of both internal speed and simultaneity.
Moreover, Burroughs is placing a uniuqely strong emphasis upon multiprogrammed operation - and is supplying the hardware and software required to make multiprogramming
(which Burroughs calls "multiprocessing") a practical reality.
Announced in March 1966, the B 2500/3500 systems were delivered to initial customers in August 1967. Announced software was delivered concurrently with the first
installations. Typical lease prices will range from about $4,200 per month for a
small B 2500 tape system to $20, 700 per month for a large B 3500 system with 100
million bytes of disc storage and 8 tape units.
The B 2500 and B 3500 are fully program-compatible with one another. The few signifi- ____
cant differences between the two systems are in internal speed (the B 3500 is twice as
fast), maximum number of I/O channels (8 for the B 2500 and 20 for the B 3500), and
maximum core memory capacity 120, 000 bytes for the B 2500 and 500,000 bytes for the
B 3500.
On the important matter of compatibility with the IBM System/360, Burroughs has taken
a middle-of-the-road approach. There is no machine-language program compatibility
between the B 2500/3500 and the System/360, but the data formats, codes, and virtually
all input/output media used in the B 2500/3500 are System/360-compatible. Thus,
B 2500/3500 and System/360 installations will be able to interchange data files but not
programs.
Software support for B 2500/3500 systems is divid('d into two )('vels, Rasic and Advanced.
Each level is designed for use with a separate operatin~ syst('m. An ast'l!'mbler, sort
generator, and report generator are providpd at each level. COBOL and FORTRAN compilers, however, are offered only at the Advanced lev()l. Compilers for th() ALGOL and
PL/I languages have not been announced to date.
At the Basic software level - intended for small configurations - the Basic Control Program controls simple stacked-job processing and straightforward I/O operations. Systems
equipped with disc storage and at least 30, 000 bytes of core memory will usually use the
Advanced software. At this level the Master Control Program schedules and controls
multiprogrammed operations and performs many other functions usually found only in the
operating systems for much larger computers .

. 11

The Burroughs 500 Systems
The B 2500 and B 3500 form the lower end of the "Burroughs 500 Systems" computer
family, which currently includes three larger members: the B 5500, B 6500, and B 8500.
Although all five systems utilize many of the same peripheral devices and design concepts,
there is no machine-language program compatibility between the B 2500/3500 and the
larger members of the family.
The Burroughs B 5500 (Report 203:) is a medium-scale computer of highly unorthodox
design. In use since 1963 it has demonstrated the practicality of multi-programmeu opera.tion and of programming and debugging exclusively in higher-level languages (ALGOL,
COBOL, and FORTRAN). As a result, more than twise as many B 55 systems were ordered
during 1966 as during 1965.
The B 6500, announced in June 1966 and scheduled for delivery in late 1969, is a thirdgeneration version of the B 5500. Using integrated circuits and a thin-film main
memory, the B 6500 will provide about eight times the processing power of the B 5500,
while maintaining program compatibility with it and using the same software.
The B 8500 is an ultra-large-scale computer with extensive capabilities for modular
expansion, multiprogramming, and multiprocessing. The B 8500 was announced nearly
two years ago, but only two orders have been placed to date.
Although the B 2500 and B 3500 are distinctly more conventional in design than the
B 5500 and B 8500, it is apparent that their designers have borrowed certain important
con(lepts from the larger Burroughs systems. These key concepts include operation
under an integrated Master Control Program, emphasis upon multiprogrammed operation,
and use of "stacks" to store subroutine parameters.
© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

5/69

210:011. 200

.2

BURROUGHS B 2500 & B 3500

DATA STRUCTURE
The B 2500 and B 3500 provide facilities for convenient handling of variable-length fields
composed of either 8-bit bytes or 4-bit digits. Core memory is addressable by digit
position. The basic unit of data, however, is the "word," which consists of 16 data bits
plus one parity bit and is the aml)unt of information that can be read from or written into
core memory during each cycle. A word can hold either two 8-bit bytes or four 4-bit
digits. Operand lengths can range from 1 to.100 digits or bytes, or up to 10,000 words,
depending upon the operation being performed.
Data represented in the 4-bit format can be either signed (with a 4-bit sign digit preceding
the most significant numeric digit of the field) or unsigned. Data in the 8-bit format is
always unsigned, but all-numeric fields represented in the 8-bit mode can be 'used as
operands in fixed-point arithmetic operations without the need (as in the IBM System/360)
for prior conversion to the 4-bit digit mode.
Floating-point numbers have 2-digit exponents and variable-length mantissas of 1 to 100
digits. Instructions can consist of one, two, three, or four 6-digit "syllables" or a single
8-digit syllable.
As in the System/360, data can be represented internally in either of two codes, EBCDIC
or ASCII, depending upon the setting of a mode flip-flop in the processor. A Translate
instruction uses a table in core memory to accomplish efficient translations from any
4-bit or 8-bit code to any 8-bit code.
A hardware translator provides automatic translations between the EBCDIC internal code
and the 6-bit BCL (Burroughs Common Language) code that can be used by the punched
card, paper tape, and 7-track magnetic tape I/O devlces. The 6-bit BCD code used by
the IBM 1400 Series computers is, in turn, a subset of the BCL code. Thus, the hardware
translator facilitates achieving code compatibility with earlier Burroughs and IBM equipment .

.3

HARDWARE

.31

System Configuration
The B 2500 "central system" consists of a 2501 or 2502 Central Processor and one combined input/output and memory cabinet. Core memory capacity can range from 10,000 to
120.UOO bytes in 10, OOO-byte increments. Four I/O channels (two Type A and two Type B)
are included .n the basic system; a total of up to eight channels can be installed, five of
which can be Type B.
The B 3500 "central system" consists of a 3501 Central Processor, one or two input/output
cabinets, and from one to six memory cabinets. Core memory capacity can range from
10,000 to 500,000 bytes in 18 different sizes. Six I/O channels (three Type A and three
Type ~ are included in the basic system; a maximum of 14 additional channels (seven
Type A and seven Type B) can be installed.
Floating-point arithmetic is available as an,optional feature for both central processors. A standing and a desk-style console unit are offered, one of which is a
system prerequisite.
The operations of all peripheral devices are directed by I/O control units. One I/O
control can be connected to each B 2500 or B 3500 I/O channel, and each type of peripheral
device requires a different I/O control. Most of the I/o controls (card reader, punch,
printer, paper tape, etc.) can accommodate only one peripheral unit each, but the controls
for magnetic tape and disc storage devices can accommodate multiple units, as explained
in the descriptions of these devices. Further flexibility in the control of magnetic tape and
disc operations is provided by a series of Exchange units, which permit a group of tape
or disc units to communicate with the central processor via either of two or more I/O
control units and the associated I/O channel~.
The two types of I/o channels differ in their modes of communication with the processor
and, therefore, in the data rates they can accommodate. Type A channels transfer only
one character at a time to or from core memory and are limited to handling low-speed
peripheral devices such as card readers and punches. Type B channels transfer two
characters in parallel to or from core memory and can handle the faster peripheral
devices such as magnetic tape and disc units .

. 32

Central Processors
The central processors used in B 2500 and B 3500 systems are functionally identical and
completely program-compatible; they differ only in internal speeds, as shown in Table I.
The processor contains the arithmetic unit, logic controls, and hardware interrupt facilities. No facilities for multiple-processor configurations have been announced to date.

5/69

fA

AUERBACH

'"

(Contd.)

210:011. 321

SUMMARY

TABLE I: ARITHMETIC EXECUTION TIMES
TASK (Times expressed in
microseconds)
Fixed Point Binary

Central Processor Model
B 2500

B 3500

not avail.

not avail.

75
66
416
1,810

37.5
33
208
905

102
463
1,861

51
231. 5
930.5

Fixed Point Decimal
c=a+b
b=a+b
c=axb
c = alb
Floating Point Decimal*
c=a+b
c=axb
c = alb
Move a to b
Compare a to b

Note:
.321

54
60

27
30

*With optional feature.
All operand lenths are considered to be five digits in length. (Therefore,
floating-point mantissas are considered to have a precision of five digits.)

Instruction Formats
There are two basic types of instructions: I/O instructions (called "descriptors") and
processor instructions. I/O descriptors vary from one to four 6-digit "syllables" in
length, and are discussed in Paragraph. 35. Processor instructions may be one, two,
three, or four 6-digit syllables in length and may contain - respectively - zero, one,
two, or three core memory addresses. (Branch instructions are eight digits in length,
and therefore constitute an exception to the normal Instruction format.)
The first two 4-bit digits of a processor instruction dpslp,natl' the olwration code' and
initiate execution of the appropriate string of microprop,rams storl'd in t1w system's
Read-Only Memory, a resistive-type memory with a 100-nanosecond acc('ss time.
These microprograms fetch the remainder of the instruction and perform the specified
operation.
In most instructions, the third, fourth, fifth, and sixth digits of the first syllable specify
the lengths of the A and B operands, which can range from 1 to 100 digits or bytes, or
up to 10,000 words (depending upon the operation).
In multi-syllable instructions, the second, third, and fourth syllables (when present)
specify the A-field, B-field, and C-field addresses, respectively. The first digit of
each address syllable specifies: (1) which, if any, of three index registers shall be
used in forming the machine address, and (2) the format of the data field. Each program has its own complement of three 8-digit index registers, which are held in
reserved locations in core memory. There are four format possibilities: signed 4-bit,
unsigned 4-bit, unsigned 8-bit, or indirect address (which means that the data field's
address, rather than the data field itself, will be found in the memory location specified by the address syllable). The remaining five digits of each address syllable
specify the data field address itself.
In many instructions, the second syllable may contain a literal operand rather than the
A-field address. Literal operands are limited to a maximum length of three bytes
(or six digits) .
All addresses in processor instructions are "base-relative." This means that each
program is written as if it started in the first location of core memory. At execution
time, each address is automatically incremented by the contents of a three-digit,
modulo-1000 base register to form an absolute address (which can be further modified
by indexing if desired). The base-relative addressing technique facilitates program
relocation and segmentation; it also permits up to 1 million digits (500,000 bytes) of
core memory to be addressed although the instruction addresses are only five digits
long •

• 322

Processing Facilities
The instruction repertoire'includes efficient facilities for arithmetic, data movement,
comparison, and editing of variable-length decimal and alphanumeric data fields. No
binary arithmetic facilities are available, although logical AND, inclusive OR, and
© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

5/69

210:011. 322

.322

BURROUGHS B 2500 & B 3500

Processing Facilities (Contd.)
exclusive OR instructions are provided. Fixed-point decimal arithmetic instructions
include three-address addition, subtraction, multiplication, and division, as well as
two-address addition and subtraction. The optional Floating-Point feature provides
three-address instructions to add, subtract, multiply, and divide floating-point operands
with 2-digit exponents and mantissas varying from 1 to 100 digits in length. Representative execution times are shown in Table I.
A novel feature of the new Burroughs processors is their ability to combine numeric
operands in the 4-bit and 8-bit data formats in a single operation, without prior transformation. The programmer can specify the format in which the results of such mixedformat operations shall be expressed: signed 4-bit, unsigned 4-bit, or 8-bit.
A group of Scan instructions facilitates the coding of search operations by enabling the
programmer, by means of a single instruction, to search a string of up to 100 characters for the presence (or absence) of a specified character or group of characters.
In addition, table look up operations are facilitated by the Search instruction. A
Translate instruction effects translations from any 4-bit or 8-bit code to any 8-bit
code through the use of a table in core memory. The Edit instruction moves up to 100
characters or digits from a source field to a destination field under the control of a
string of "micro-operators" in core memory, which can specify that any character shall
be inserted, suppressed, or floated under a variety of conditions. This flexible instruction permits normal dollar-and-cent punctuation and either floating dollar sign or check
protection to be accomplished in a single operation.
Two special instructions, Enter and Exit, facilitate entry to and exit from subroutines,
especially when the subroutines are used in nested or recursive fashion. The Enter
instruction causes return control information and subroutine parameters to be moved
into the "stack," which is a core memory area that has been reserved by the programmer. If another subroutine is entered prior to exit from the first subroutine, additional
return control information and parameters will be moved into the stack, causing the
previously-stored parameters to be "pushed down" deeper into the stack. Thus, subroutines can be nested or used recursively to any level up to the capacity of the stack .

. 323

Operational States
The central processor always operates in one of two states: the Normal State, in which
user programs are executed, or the Control State, in which the functions of the MCP or
BCP operating systems are performed. Several "privileged" instructions can be executed only in the Control State. These instructions permit the MCP or BCP to initiate
I/O operations and to control the program mix by setting and clearing registers and
flip- flops.
A powerful interrupt system causes the processor to enter the Control State and branch
to the MCP or BCP whenever any of the folloWing conditions occurs: completion of an
I/O operation, memory parity error, memory address error, invalid instruction
(including attempted execution of a privileged instruction in the Normal State), instruction time out (failure to complete the execution of an instruction within a preset time
limit), and clock interrupt. Memory address errors can result from the formation of
an address beyond the bounds established by the base and limit register settings, a
non-decimal digit in an address, or a "non-synchronized" address (i. e., an address that
is not modulo-2 or modulo-4 when a particular instruction or data format requires such
an address). A clock interrupt occurs when the processor's six-digit timer, which is
incremented once each millisecond, reaches a programmer-specified control value.
Memory protection - an essential feature for successful multiprogramming - is provided by the hardware, using the base and limit registers and the interrupt system.
When the MCP initiates execution of a program, it sets the base register to the program's
initial core memory location. The limit register's setting is made equal to the base
register setting plus the total core memory requirement for the program. Before data
is fetched, all machine addresses are checked against the base and limit register
settings. An out-of-bounds address causes an interrupt and a transfer of control to the
MCP, which suspends execution of the offending program .

. 33

Internal Storage

.331

Core Memory
Core memory cycle times are two microseconds in the B 2500 and one microsecond in
the B 3500. Read access times are 700 nanoseconds and 350 nanoseconds, respectively.
Both processors access one word, consisting of 16 data bits plus one parity bit, per
cycle. Because core memory is addressable by digit position and each word can hold
four digits, all word addresses are modulo-4. Whenever a specific digit or byte position is addressed, the entire word in which it is located is accessed. Data can be
transferred to or from core memory by either the central processor or any of the
I/O control units

5/69

fA

AUERBACH

'"

(Contd. )

SUMMARY

.331

210:011. 331

Core Memory (Contd.)
B 2500 core memory capacities can range from 10,000 to 120,000 bytes in increments of
10,000 bytes (i. e., 5,000 words). B 3500 memory capacities can range from 10,000 to
500,000 bytes in 18 different sizes; see the Price Data section (210:221) for the available capacities and the associated model numbers.
The first 1,220 digit positions of core memory are reserved for use by certain processor instructions, the interrupt system, the I/O control units, and the Master Control
Program (when used). In addition, the first 64 digit positions of the core area assigned
to each program are usually reserved to hold that program's index registers and other
specific information.
To prevent accidental over-writing of one program by another during multiprogrammed
operation, each core memory address is checked to ensure that it lies between the
boundaries established by the base register and the limit register. If not, the program
is interrupted and control is transferred to the Master Control Program .

. 332

Address Memory
The B 2500 and B 3500 Central Processors contain an Address Memory unit with a 100nanosecond cycle time. Address Memory is an array of from 24 to 120 word locations,
expandable in increments of 12 words. Each word is six 4-bit digits in length - long
enough to hold the absolute address of any core memory location. Address Memory's
purposes are to reduce the number of core memory accesses required and to perform
a number of functions that usually require separate processor registers. The first
eight words are used by the processor, and two words are assigned to each installed
I/O channel. The I/O channel words contain the initial and final core memory addresses
for the I/O operation in progress on the associated channel; these words are used by the
I/O control units to determine the core memory locations where output data is to be
accessed or input data stored .

• 34

Disk File Storage
Burroughs offers several versions of its head-per-track Disk Files for use in B 2500/3500
systems. These devices have proved to be among the fastest and most reliable mass
storage units currently available. (No storagc equipment of the interchangeable-cartridge
type is offered at present.)

.341

Systems Memory
Burroughs' 9370 Systems Memory is a new, single-disc storage unit designed to provide, at a fairly low cost, the random-access storage required to hold the systems
software and the user's program library. As in llurroughs' larger Disk Files, one
read/write head serves each track, so no access-arm movement is required. Average
access time is 17 milliseconds, and the peak data transfer rate is 291,000 bytes per
second.
Two models of Systems Memory are offered. Model 9370-1 uses only one face of the
disc and has a capacity of 1,000,000 bytes; Model 9370-2 uses both disc faces and holds
2,000,000 bytes. There are 100 data tracks per disc face. Each track is divided into
100 segments, and each segment holds 100 bytes (or 200 digits) of information. A
longitudinal parity character is recorded at the end of each segment. Systems Memory
is connected to a B 2500 or B 3500 Processor by means of a Systems Memory I/o
Control. Two Systems Memory units can be connected to a single I/O Control by adding
a Systems Memory Exchange •

. 342

Modular Random Storage
The Modular Random Storage subsystem is similar to the Disk Files that have been in
use for several years in second-generation Burroughs systems. The use of a fixed
read/write head for every data track provides fast access times (20 milliseconds
average) and high reliability. Average data transfer rate is 218,000 bytes per second.
The Disk File Subsystem is compatible with the 9370 Systems Memory in segment size
(100 bytes) and addressing structure, so either unit can be used to hold the systems
software and program library.
The Modular Random Storage subsystem is composed of 9371 Electronics Units and
9372 Disk File Modules. Each 9372 Disk File Module contains four non-removable
discs and holds 10 million bytes. From one to five 9372 Modules can be connected to
a 9371 Disk File Electronics Unit, providing up to 50 million bytes of storage. Disk
File Exchanges permit up to twenty 9371- Electronics Units and the associated Disk
File Modules to be connected to from one to four Disk File I/o Controls. Among the
available Exchanges are those designated 1 x 5, 2 x 5, and 4 x 10; the first number is
the maximum number of Controls and the second is the maximum number of Electronics
Units that can be interconnected. Since each I/O Control is connected to a separate
channel, the 2 x 5 and 4 x 10 Exchanges, in effect, provide "floating" I/O channels and
two-way or four-way simultaneity for Disk File operations.
© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

5/69

210:011. 343

. 343

BURROUGHS B 2500 & B 3500

Data Memory Banks
The recently-announced Disk File Data Memory Banks will use the same head-per-track
technique to provide even larger storage capacities at lower costs per byte. Models
9375-0, 9375-2, and 9375-3 will provide average access times of 23, 40, and 60 seconds,
respectively. All three models have a basic storage capacity of 100 million bytes and
can be expanded in modest increments to a maximum of 500 million bytes per I/O
channel in B 2500 systems and 2.5 billion bytes per channel in B 3500 systems .

• 35

Input-Output Equipment
Burroughs offers three card readers, two card punches, a paper tape reader and punch,
four line printers, a tape lister, and three MICR Sorter/Readers for use with B 2500
and B 3500 systems. Except for certain difference in codes, all of these units are
similar to previous Burroughs peripheral units used with B 100/200/300 Series and
B 5500 systems. Newly-developed I/O equipment includes the Magnetic Tape Clusters
(Paragraph.356) and several data communications devices (Paragraph. 359) .

. 351

Card Readers
The 9110 Card Reader is a compact unit that reads standard 80-column cards photoelectrically, in column-by-column fashion, at a peak speed of 200 cards per minute.
An immediate-access clutch permits the reading of an 80-column card to be completed
within a maximum of 350 milliseconds after a "start feed" signal is received. The
input hopper and output stacker have a capacity of 450 cards each.
The 9111 Card Reader uses an immediate-access clutch, a belt drive mechanism, and
13 photoelectric read cells (one for timing) to read cards of 51, 60, 66, or 80 columns
at a peak rate of 800 cards per minute. The input hopper and output stacker can each
hold up to 2400 cards and can be loaded and unloaded while the reader is operating.
Optional features permit 40-column Treasury Checks or the round holes in Postal
Money Orders to be read.
The 9112 Card Reader has a peak reading rate of 1400 cards per minute. Its appearance and physical characteristics are the same as those of the 9111 Card Reader
described above.
All three card readers can read either the EBCDIC or the generally similar BCL
(Burroughs Common Language) card code and translate it automatically to EBCDIC
internal code. Binary card images can be read and stored in memory without conversion; the contents of each card column are stored in the six low-order bit pOSitions
of two consecutive bytes .

• 352

Card Punches
The 9210 and 9211 Card Punches have been replaced by the 9212 and 9213 respectively.
The 9212 Card Punch can punch standard 80-column cards at a peak rate of 100 cards
per minute. Cards are punched one row at a time by a single row of 80 die punches.
Punching accuracy is checked by the hole-count method.
The 9213 Card Punch punches 80-column cards, in row-by-row fashion, at a peak
rate of 300 cards per minute. Punching accuracy is checked by the hole-count method.
Both Card Punches have 1000 card input hoppers and three program selectable output
stackers that hold up to 1200 cards. The three stackers are: primary, auxiliary and
error.
Both card punches can perform automatic translations from the EBCDIC internal code
to EBCDIC card code. Alternatively, binary cards can be punched; the contents of the
six low-order bit positions of two consecutive bytes of core memory are accessed and
punched into each card column. A standard feature in the Card Punch Control permits
automatic translation from EBCDIC internal code to the 64-character BCL (Burroughs
Common Language) code, or optionally to the BULL or ICT card code .

• 353

Paper Tape Units
The 9120 can read punched tape with 5, 6, 7, or 8 code levels at a peak speed of either
500 or 1,000 characters per second. Rated start and stop times are 5 and 20 milliseconds, respectively, and the reader can stop between consecutive characters. A
standard plugboard permits the bit configurations read from tape to be interchanged
and/or inverted. Automatic translation from either 5-level Baudot or 6-level BCL
tape code to EBCDIC internal code is standard. For other punched tape codes, users
can either add the optional 9926 Input Code Translator, which provides flexible code
conversions under plugboard control, or read the tape codes directly and use programmed translation.
The 9220 can punch 5-, 6-, 7-, or 8-level tape at a peak speed of 100 characters per
second. A standard plugboard permits the bit configurations from core memory to be
interchanged and/or inverted prior to punching. Automatic translation from EBCDIC

5/69

fA

AUERBACH

'"

(Contd.)

SUMMARY

2fO:Olf. 353

.353

Paper Tape Units (Contd.)

. 354

internal code to either 5-level Baudot or 6-level BCL tape code is standard. For other
punched tape codes, either programmed translation or the optional 9928 Output Code
Translator can be employed .
Line Printers
Four line printers are available for use with B 2500 and B 3500 systems. Users can
choose a buffered printer with a peak speed (at single spacing) of either 315 lines per
minute, or an unbuffered model with a printer memory option with a peak speed of
either 800 or 1100 lines per minute. Table II lists the peak printing speeds and skipping
speeds of the four models.
Both B 9242-1 and B 9243-1 have 120 print positions, with 12 additional positions available through use of the 9941 option. All models use the 64-character Burroughs Common Language character set. (Other character sets are available on a special-order
basis.) Forms can be from 5 to 20 inches in width and can have a maximum length of
22 inches (when printing at 6 lines per inch) or 16.5 inches (at 8 lines per inch). Vertical format control is provided by a 12-level carriage control tape loop.
B 9245-2 and B 9245-3 printers have 120 and 132 print positions respectively. Forms
can be 4 to 18.5 inches in width and can have a maximum length of 22 inches. Vertical
format control is provided by a four-channel punched format tape.
TABLE II: LINE PRINTER SPEEDS
Model No.

r-----

. 355

Peak Printing
Speed,
lines/minute

B 9245-2
B 9245-3

315
315

B 9242-1
B 9243-1

860
1100

Skipping
Speed,
inches/second
16.5 (minimum)
25 (75 with option)
25 (75 with option)

Buffer
Yes
Yes
optional
optional

9244 Tape Lister
The Burroughs Tape Lister is designed primarily to provide high-speed printed listings
of MICR documents as they are read by a MICR Sorter-Reader. From 6 to 18 listing
tapes can be individually advanced and printed upon, enabling the contents of each
Sorter-Reader pocket to be listed on a separate tape. The six-tape 9244-1 master unit
can be used alone, or one or two six-tape 9244-2 slave units can be connected to it.
The peak printing rate of 1565 lines per minute can be maintained when printing is
restricted to a 16-character set (the digits 0-9 and six special symbols). When the full,
alphanumeric set of 40 characters is used, the rated speed is 600 lines per minute.
Printing is performed on 2. 5-inch-wide, single-ply or two-ply adding machine tapes.
Print lines can extend to 22 print positions on each tape. The 9244-1 unit contains a
44-character buffer. The contents of buffer positions 1 through 22 are simultaneously
printed on the master tape and on a selected detail tape. In addition, when operating
in the "multiprocessing" mode, the contents of buffer positions 23 through 44 can be
simultaneously printed on a third tape •

. 356

Magnetic Tape Clusters
The most novel peripheral units announced for use with B 2500 and B 3500 systems are
the 9381 and 9382 Magnetic Tape Clusters, which provide two, three, or four tape
drives in a single compact unit (33 inches wide, 30 inches deep, and 42 inches high).
Reading and/or writing can be performed simultaneously on any two of the tape drives
in a cluster if two Tape Controls and a 2 x 8 Exchange are employed. Each tape drive
has its own drive mechanism, but the read/write electronics and power supply are
shared. The feed reel and take-up reel for each drive are mounted on concentric
vertical shafts, with the feed reel above the take-up reel. Tape is driven by a capstan
roller and pinch rollers. Normal tape speed is 45 inches per second, and rewind speed
;is 90 inches per second. Tape can be read in either ,the forward or reverse direction.
The standard recording medium for the Magnetic Tape Clusters is 9-track, 1/2-inchwide tape which is compatible with the IBM System/360 9-track tape units. Each unit
can read and record at any two of the following three densities: 200, 800, or 1600 rows
per inch. The associated peak data transfer rates are 9,000, 36,000 or 72,000 bytes
per sE'cond, respectively. A single Tape Control can handle up to 8 tape drives of
either the 36KB or 72KB transfer rate, but not both.
Optionally, individual tape drives can be field-modified to use 7-track tape which is
compatible with earlier Burroughs and IBM systems. Seven-track and nine-track
drives can be intermixed in the same Tape Cluster, although the two types must be
served by separate controls. In the 7-track mode, the available densities are 200, 556,
and 800 rows per inch, and the associated data transfer rates are 9,000, 25,000 and
36,000 characters per second.
© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

5/69

210:011. 357

BURROUGHS B 2500 & B 3500

.357

Free-Standing Magnetic Tape Units
Burroughs also offers four different free-standing tape units for B 2500 and B 3500
systems. Their recording densities and speeds are summarized in Table ITI. Two
of these units use 7-track tape and are similar to second-generation Burroughs units.
The other two units use 9-track tape at recording densities of either 800 or 1600 rows
per inch to achieve compatibility with the IBM System/360. All four models can read
tape in either the forward or reverse direction at 90 inches per second; rewind speed
averages 300 inches per second.
Up to ten free-standing tape units (all of the same type) can be connected to each Tape
Control. Read/write simultaneity within a single group of tape units can be achieved
(in B .3500 systems only) through use of a second Tape Control and a 2 x 10 Magnetic
Tape Exchange unit.
TABLE III: FREE-STANDING MAGNETIC TAPE UNITS
Model No.

Recording
Densities,
rows/inch

Peak Data
Transfer
Rates

B 9390 (7 -track)
B 9391 (7-track)

200/556
200/556/800

18/50 KC/sec
18/50/72 KC/sec

B 9392 (9-track)
B 9393 (9-track)

200/800
200/1600

18/72 KB/sec
18/144 KB/sec

.358 MICR Sorter-Reader
Burroughs, a leader in banking applications and in the design of MICR equipment, offers
four MICR Sorter-Readers for use with the B 2500 and B 3500 systems. All are closely
related to the MICR units used with B 100/200/300 Series systems. The maximum speed
of all three units is 1565 items per minute. Model 9130 is an off-line unit that reads
MICR-encoded documents and sorts them into 13 pockets. Models 9131 and 9132 are both
designed for on-line use and are completely buffered; Model 9131 has 13 pockets and
Model 9132 has 16 pockets. The 9134-1 is an expandable on-line unit that allows additions from a basic four pocket unit to 32 pockets in four pocket modules. It reads
OCR or MICR documents at up to 1565 documents per minute. Read systems are also
modular and may contain any combination of OCR or MICR with a maximum of two
reading systems on any given unit .
• 359 Data Communications
Data communications equipment can be connected to a B 2500 or B 3500 system through
individual Line Adapters and either Single-Line Controls, Multi-Line Controls, or
Terminal Unit Controls. Eleven different Line Adapters permit communications with a
variety of terminal equipment and common-carrier services; see the Price Data section,
210:221, for the function of each adapter. All adapters operate in the half-duplex mode.
The Single-Line Control coordinates the transmission of data between a single communications line equipped with an appropriate Line Adapter and a single B 2500 or B 3500
I/O channel (Type B). The Single-Line Control is designed for limited data communications needs; if more than three lines must be controlled, it will be more economical
to use one Multi-Line Control.
The Multi-Line Control uses two Type B I/O channels and can control simultaneous
transmissions over up to 36 communications lines, each equipped with an appropriate
Line Adapter. One 40-bit word of scratchpad memory holds control information and
data associated with each of the 36 subchannels. Only one Multi-Line Control can be
connected to a B 2500 or B 3500 system.
The Terminal Unit Control, designed for on-line banking applications, can theoretically
control up to 60 communications channels and 960 remote teller consoles.
Burroughs has expanded its line of data communications equipment. The 9351 Input
and Display System, is a CRT terminal that provides keyboard input and alphanumeric
data displays. A 9-by-12-inch screen can display 25 lines of data with up to 80 characters per line. From one to four keyboard/display units can share a control unit
with a 1020-character buffer memory. In addition, Burroughs provides support for
the following communications devices:
•
•
•
•
•
5/69

B 9350 Typewriter Inquiry Station
mM 1030
Burroughs Audio Response System
UNIVAC DCT 2000
TC-500

fA

AUERBACH

'"

(Contd. )

210:01 f. 360

SUMMARY

. 36

•

TC-700

•

Burroughs On-Line Teller Consoles

Simultaneous Operations
One data transfer operation on each installed B 2500 or B 3500 I/O channel can occur
simultaneously with internal processing. All I/o operations, once initiated by the
central processor, are executed independently of the processor under the direction of
an I/O control unit. Requests for access to core memory by the processor and the
various I/O channels are granted by priurity logic; the processor always has the lowest
priority.
Please refer to Paragraph. 31 for the minimum and maximum number of Type A and
Type B channels a system can include, the peripheral devices each type of channel can
service, and other system configuration parameters.
Address Memory (Paragraph. 332) is used to hold the core memory addresses associated
with the I/O operation taking place on each channel. Therefore, only one core memory
cycle (two microseconds in the B 2500, one microsecond in the B 3500) is required for
each unit of I/O data transferred to or from core memory. Type A channels transfer one
character at a time to or from memory, while Type B channels transfer two characters
in parallel. Thus, the Burroughs systems offer a substantially higher degree of simultaneity than most comparably-priced computers.
I/o operations can be initiated only when the central processor is operating in the control
state, and their execution will usually be directed by the Master Control Program. The
processor initiates an-I/O operation by sending an "I/O descriptor" to the appropriate
control unit. I/O descriptors vary from one to four 6-digit syllables in length, and the
core addresses they contain are absolute rather than base-relative. The processor then
proceeds independently while the control unit directs the I/o operation. Upon completion,
the control unit initiates a processor interrupt and sends a 16-bit "result descriptor" to
a reserved location in core memory. The result descriptor informs the processor of
any abnormal conditions or errors that have occurred .

.4

. 41

SOFTWARE
Burroughs offers B 2500 and B 3500 users a choice of two lcvols of softwarc' support: a
Basic package for small-scale configurations and an Advanced packagc for larger systems.
Only the Advanced package permits use of the Master Control Program, the CORO Land
FORTRAN compilers, and the multiprogrammed opcrational modc that Burroughs bills
as principal features of the B 2500 and B 3500 systems. The facilities offered at each of
the two software levels are summarized in the following paragraphs .
Basic Software Package
The Basic package is designed for use with minimum-size B 2500 or B 3500 systems; it
does not require a Disk File or Systems Memory unit or a Console Printer. The main
component of the package is the Basic Control Program. Also included are an assembler,
report generator, sort generator, and utility program generator. All of the Basic software was delivered in August 1967 .

. 411 Basic Control Program (BCP)
The BCP is a group of interrelated loading and I/O routines that reside permanently in
core memory. Burroughs states that the BCP routines and associated tables will require
about 2, 000 byte positions of memory. The BCP's principal functions are:
•

To load and initiate execution of user programs in sequential fashion.

•

To initiate all I/O operations requested by user programs.

•

To service the interrupts that result upon completion of

•

To transfer control to error-handling routines in the user program when errors
or abnormal conditions are detected. (These error-handling routines can be
extracted from the I/o library when the user program is assembled.)

•

To receive control when a user program completes its run, and to load and initiate
the next program.

•

To assist in program debugging by means of trace and dump routines.

I/o operations.

The BCP will accept any user program generated by the facilities of the Basic software
package. In fact, the BCP's presence is required for execution of any program generated
by this package. The BCP can be used with any legitimate B 2500 or B 3500 system
configuration.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

5/69,'

210:011.412

BURROUGHS B 2500 & B 3500

.412 Basic Assembler
The Basic Assembler is a straightforward, tape-oriented symbolic assembly system that
allows the programmer to utilize all the hardware facilities of B 2500 and B 3500 systems.
Its use requires a processor with at least 10, 000 bytes of core memory, two magnetic
tape units, a card reader or paper tape reader for input, a line printer for the listing,
and a card punch, paper tape punch, or a third magnetic tape unit for object-program
output.
The assembly process is essentially a one-for-one translation in which symbolic instructions, coded in a fixed format, are converted into machine language instructions. Inputoutput operations - such as reading, writing, opening and closing of files, and generation
and checking of labels - are coded by means of eight macro-instructions that generate
linkages to the appropriate BCP I/O routines. I/O buffering, blocking, and unblocking
operations, however, must be coded by the user. A group of pseudo-instructions is
provided to control the assembly process. Segmentation of programs can be specified in
the Basic Assembler language, but the actual overlaying of segments at object time must
be coded by the user .
• 413 Problem Oriented Facilities
The Basic Sort Generator accepts parametric input and produces tape sort programs in
Basic Assembler language. Internal sorting is performed by the vector replacement
selection technique, and a backward-read polyphase merge technique is used. From three
to eight magnetic tape units can be used. Records or blocks can be up to 1000 characters
long. Sort keys can be up to 100 characters long and located in up to 10 different areas
of the record. Input and output records may be blocked. Multiple-reel input is allowed,
and restart capabilities are provided.
The Basic Report Generator accepts problem-oriented specifications and generates programs, in Basic Assembler language, to produce the specified reports. Input to the
object programs may be from punched cards or magnetic tape; output may be printed or
on punched cards. Up to four levels of totals are permitted, and up to fifteen 12-position
accumulators are available at each level.
The Basic Utility Program Generator accepts parametric input and generates programs
to perform a variety of media conversion functions .
. 42

Advanced Software Package
The Advanced package is designed for use with B 2500 or B 3500 systems that have at
least 30, 000 bytes of core memory, a Systems Memory unit or Disk File, and a Console Typewriter/Keyboard. Its principal component is the Master Control Program.
Also included are an assembler, COBOL compiler, FORTRAN compiler, report generator, sort generator, and utility program generator. All the Advanced Software
was delivered in 1967. Minimum core memory requirements for compilation are
30, 000 bytes for COBOL and 40,000 bytes for FORTRAN IV .

. 421 Master Control Program (MCP)
The MCP is an integrated operating system that monitors and controls all operations of a
B 2500 or B 3500 system. The MCP consists of a group of interrelated routines that will
permanently occupy approximately 13, 000 bytes of core memory plus 235, 000 bytes of
disc storage. The principal functions of the MCP are:

5/69

•

To schedule the execution of user programs, in a multiprogramming environment,
on the basis of their priorities and memory requirements.

•

To load programs into core memory from disc storage.

•

To allocate core memory and relocate user programs as necessary to achieve
efficient utilization of the available memory.

•

To schedule and initiate all I/O operations, using tables that indicate the status of
each I/O device and the priority of each I/O request awaiting processing.

•

To handle all error conditions that arise, usually by first retrying the operation
and then either initiating a user-supplied error routine or aborting the program.

•

To handle all communications between the system and the operator.

•

To maintain a detailed log of all system operations.

•

To maintain libraries, on disc storage, of user programs and systems software.

•

To control compilations or assemblies, and either insert the resulting object
programs into a library or execute them immediately.

A

AUERBACH
@

(Contd. )

210:01 f. 421

SUMMARY

.421 Master Control Program (MCP) (Contd.)
•

To control program segmentation by loading individual segments of a program
upon request.

•

To perform file control operations such as blocking, unblocking, label generation,
and label checking.

The MCP will accept any user program generated by the facilities of the Advanced software package. In fact, the MCP's presence is required for execution of any program
generated by this package.
For B 2500 and B 3500 systems that use the MCP, Burroughs is placing a greater emphasis upon multiprogrammed operation than has the manufacturer of any previous smallto-medium-scale computer system. Multiprogramming (which Burroughs calls "multiprocessing") can increase a computer system's throughput by increasing the effective
utilization of the processor, core memory, and all peripheral devices; and Burroughs
has already demonstrated the practical value of multiprogramming in its B 5500 system,
which uses a functionally similar Master Control Program and has been in service
since 1963 .
. 422 Advanced Assembler
The Advanced Assembler is a disc-oriented symbolic assembly system that provides all
the facilities of the Basic Assembler plus a few refinements. These refinements include:
•

Facilities to operate under control of the MCP and to establish the necessary
linkages with it.

•

File declarations that specify, for each logical file, its label format, recording
mode, buffer areas, retention period, blocking factor, etc.

•

Macro-instructions to open and close files, initiate I/O operations, block and
unblock logical records, seek disc records, position printer forms, etc.

•

An unlimited number of symbolic labels.

The Advanced and Basic Assemblers use the same mnemonic operation codes for machine
instructions and the same fixed-format coding shcct. Thc differences in their techniques
for handling I/O operations, however, precludc direct compatibility, in l'ither sourcelanguage or object form, between programs coded In th(' Advanc('d and Basic Assembler
languages.
Burroughs states that the Advanced Assembler will usc only 11,000 bytes of core memory (in addition to the 13,000 bytes required by the MCP) .
. 423 Compilers
COBOL and FORTRAN compilers are provided for use with any MCP-equipped B 2500
or B 3500 system that has a minimum of 30K bytes (COBOL) or 40K byte.s (FORTRAN) of
core storage. The COBOL language includes the facilities of the proposed minimum USA
Standard COBOL language, plus useful random access, segmentation, library, and table
handling facilities. Burroughs states that its FORTRAN is compatible with the USA standard
version. Burroughs states that it is practical to run COBOL or FORTRAN compilations
simultaneously with each other and with a mix of production jobs .
• 424 Problem Oriented Facilities
An Advanced Sort Generator, an Advanced Report Generator, and an Advanced Utility
Program Generator will be provided. The input speCifications and functional capabilities
of these generators will be similar to those of their counterparts in the Basic software
package. The essential difference is that the Advanced versions will generate programs
in the Advanced Assembler language for execution under control of the MCP •
•5

COMPATIBILITY
Burroughs has changed its approach to program compatibility from emulation (as originally
announced) to translation and simulation. Emulation is no longer offered; in its place is a
translator that converts IBM 1400 series systems source code to B 3500 Assembly Language
Programs. Languages acceptable as input to the translator are Autocoder, Basic Autocoder,
and SPS, as implemented on the 1401, 1440, and 1460. In addition, Burroughs provides a
translator to convert B 300 Assembly Language to B 2500/B 3500 Assembly Language.
Simulation for the B 2500/B 3500 permits execution of B 100/B 200/B 300/B 500 object
programs on a B 2500/B 3500 directly.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

5/69

-1.

210:221.101
STANDA"

~EDP

-

AUERBAC~

BURROUGHS B 2500 & B 3500

REPORTS

PRICE DATA

~

PRICE DATA
The following pricing schedule accompanied the announcement of the Burroughs
B 2500 and B ,3500 computer systems in March 1966 and still applies to the rental and purchase
of equipment in this series. Notable features of this pricing policy include: (1) a special conversion policy for customers who install a B 2500 or B 3500 in place of an existing computer
system, (2) a purchase option supplement to the basic lease agreement, and (3) extensive use of
one-time field installation charges.
According to the special conversion policy, customers who replace a Burroughs or
other computer system with a B 2500 or B 3500 system are allowed a credit of 90 per cent of
the replaced system's prior month's rent toward rental of the new system. The amount credited,
however, cannot exceed the new system's regular monthly charge.
Customers who rent a B 2500 or B 3500 system can apply, toward the purchase of the
equipment, 70 per cent of all rental charges (excluding taxes) paid during the first six months of
rental and 40 per cent of all rental charges paid during the second six months of rental, if they
opt to purchase the equipment within the first 18 months of rental. A discount of 10 percent of the
purchase price of the equipment is available to customers who decide to purchase after the
eighteenth month of rental.
One-time field installation charges, ranging from $15 to $190, are applicable to additional modules of core storage and almost all supplementary peripheral devices that are added to
already-installed B 2500 and B 3500 systems.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

5/69

210:221.102

A

STANDUD

EDP

AUERBACH

BURROUGHS B2500/3500

REPORTS

~

PRICE DATA

BURROUGHS B 2500/3500
IDENTITY OF UNIT
CLASS
Model
Number

Feature
Number

Name

PRICES
Monthly
Rental
$(1)

Purchase
$(1)

Monthly
Maint.
$(2)

1,195

57,360

125

1,445

69,360

130

25
50
50
15
15

1,200
2,400
2,400
720
720

5
10
7
0
0

1,746

81,360

140

37
67
103
16
16

1,680
3,120
4,800
720
720

5
10
7
0
0

Processing Unit

PROCESSOR
B2501

B2502
B27l0
B2711
B2730
B2740-1
B2740-2
B3501
B37l0
B3711
B3730
B3740-1
B3740-2

B2500:
Central Processor and four I/O Channels
(maximum of two channels may be
Type B)
Central Processor and four I/O Channels
(maximum of two channels may be
Type B)
Type A 1/ a Channel
Type B 1/ a Channel
Floating Point
Console (Standing Level)
Console (Desk Level)
B3500:
Central Processor and six Iio Channels
(maximum of three channels may be
Type B)
Type A I/O Channel
Type B I/O Channel
Floating Point
Console (Standing Level)
Console (Desk Level)
Main Storage

B2001
B2002
B2003
B2004
B2005
B2006
B2007
B2008
B2009
B2012

B2500:
10,000
20, 000
30, 000
40, 000
50, 000
60,000
70,000
80, 000
90,000
120.000

Bytes Core Memory
Bytes Core Memory
Bytes Core Memory
Bytes Core Memory
Bytes Core Memory
Bytes Core Memory
Bytes Core Memory
Bytes Core Memory
Bytes Core Memory
Bytes Core Memory

450
900
1,325
1,725
2,100
2,450
2,775
3,075
3,350
4,175

21,600
43,200
63,600
82,800
100,800
117,600
133,200
147,600
160,800
200,400

20
25
30
40
45
50
60
65
70

B3001
B3002
B3003
B3004
B3005
B3006
B3007
B3008
B3009
B3012
B3015
B3018
B3021
B3024
B3030
B3036
B3045
B3050

B3500:
10,000 Bytes Core Memory
20,000 Bytes Core Memory
30,000 Bytes Core Memory
40, 000 Bytes Core Memory
50, 000 Bytes Core Memory
60, 000 Bytes Core Memory
70,000 Bytes Core Memory
80, 000 Bytes Core Memory
90, 000 Bytes Core Memory
120,000 Bytes Core Memory
150, 000 Bytes Core Memory
180,000 Bytes Core Memory
210,000 Bytes Core Memory
240,000 Bytes Core Memory
300,000 Bytes Core Memory
360, 000 Bytes Core Memory
450, 000 Bytes Core Memory
500,000 Bytes Core Memory

515
1,030
1,520
1,983
2,395
2,781
3,142
3,477
3,786
4,725
5,513
6,300
7,088
7,875
9,450
11,340
14,175
15,750

24,000
48,000
70,800
92,400
111,600
129,600
146,400
162,000
176,400
216,000
252,000
288,000
324,000
360,000
432,000
518,400
648,000
720,000

20
25
30
40
45
50
60
65
70
90
110
140
170
200
260
320
410
440

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

5/69

210:221.103

8URROUGHS 82500/3500

IDENTITY OF UNIT

PRICES

CLASS
Model
Number
MASS
STORAGE

Feature
Number

Purchase
$(1)

Monthly
Maint.
$(2)

150
200
250
55
200

7,200
9,600
12,000
2,640
9,600

12
12
12
10
10

10

480

5

155
258
309
88
200

7,200
12,000
14,400
4,080
9,600

12
12
14
10
12

50

2,400

5

30

1,440

.2

150
650
650
650

7,200
31,200
31,200
31,200

10
80
80
80

650

31,200

80

650

31,200

80

375

18,000

80

450

21,600

90

850
1,650
2,400
3,100
3,775
675
1,200

44,000
85,800
124,800
161,200
196,300
35,100
57,600

115
215
315
415
515
100
125

4,950
900
3,950
700
2,700
450

237,600
45,600
224,200
45,600
129,600
28,800

595
110
495
85
395
85

900

43,200

200

1,100

52,800

230

1,300

62,400

260

1,100

52,800

225

Monthly

Name

R$fffl
Disk Storage

B2371
B2373
B2375
B2473
B2474
B2674
B3371
B3373
B3375
B3473
B3471
B3471-5
B3471-6
B3471-7
B9371-1
B9371-2
B9371-3
B9371-4
B9371-5
B9370-1
B9370-2
B9372-1
B9372-2
B9372-3
B9372-4
B9372-5
B9372-6
B9372-7
B9375-0
B3976-0
B9375-2
B9376-2
B9375-3
B9376-3
INPUTOUTPUT

B2500 Control Units:
Systems Memory Control
Disk File Control
Combination Control
1 x 2 Disk File Exchange
2 x N Disk File Exchange for B 2502
System only
Disk File Exchange Adapter for B 2474
Exchange
B3500 Control Units:
Systems Memory Control
Disk File Control
Combination Control
1 x 2 Disk File Exchange
NIx N 2 Disk File Exchange
(Up to 4 x 20)
Control Adapter (N1 side-up to
4/exchange)
EU Adapter (N2 side-up to 20/
exchange)
Exchange Extension (for over 10 EU's)
DFEU for B 9372-1 to -6
DFEU for B 9372-7
Optional Additional DFEU for B 9375-0
& B 9376-0
Optional Additional DFE U for B 9375-2
& B 9376-2
Optional Additional DFEU for B 9375-3
& B 9376-3
Systems Memory Disk Files:
Systems Memory 17 ms
(1 million bytes)
Systems Memory 17 ms
(2 million bytes)
Modular Random Storage Disk Files:
10 million Byte Storage
20 million Byte Storage
30 million Byte Storage
40 million Byte Storage
50 million Byte Storage
Additional 10 million Byte Increment
20 million Byte Storage
Data Memory Bank Disk Files
100 million Byte Storage
Additional 20 million Byte Increment
100 million Byte Storage
Additional 20 million Byte Increment
100 million Byte Storage
Additional 20 million Byte Increment
Magnetic Tape

B9381-2
B9381-3
B9381-4
B9382-

Tape Units:
36 KB Cluster, 2-Station
(9-Channel-800 bitS/inch)
36 KB Cluster, 3-Station
(9-Channel; 800 bitS/inch)
36 KB Cluster; 4-Station
(9-Channel; 800 bitS/inch)
72 KB Cluster; :='-station
(9-Channel; 1600 bitS/inch)
--

5/69

A

AUERBACH

'"

(eontd. )

PRICE DATA

210:221.1 04

IDENTITY OF UNIT
CLASS

INPUTOUTPUT
(Contd. )

Model
Number

Feature
Number

Name

PRICES
Monthly
Rental
$(1)

Purchase
$(1)

Monthly
Maint.
$(2)

1,400

67,200

260

4S0

23,000

145

575

27,600

165

575

27,600

165

650

31,200

175

650

31,200

170

650

31,200

170

200

9,600

12

350

16,SOO

12

275

13,200

12

300

14,400

12

375

IS, 000

15

325

15,600

12

150

21. 600

12

:l7!,

IH,OOO

15

175
200
250

8,400
9,600
12,000

10
10
10

250

12,000

10

50

2,400

10

25
25
25
10

1,200
1,200
1,200
4S0

5
5
5
1

50

2,400

10

243

11,2S0

12

407

lS,960

12

2S4

13,200

12

309

14,400

12

412

19,200

15

3S7

IS, 000

12

Magnetic Tape (Contd. )
B93S2-3
B93S2-4
B9390
B9391
B9392
B9393
B9394-1
B9394-2
B23S1-1
B23S1-2
B2391-1
B2391-3
B2391-4
B2393-1
B2393-2
B2393-3
B24S0
B24S1
B2490
B2491
B26S0
B26S1
B2691
B2692
B99S0
B99S9
B33S1-1
B33S1-2
B3391-1
B3391-3
B3391-4
B3393-1

72KB Cluster - 3 station
(9-channel - 1600 bits/inch)
72KB Cluster - 4 station
(9-channel - 1600 bits/inch)
IS-50 KC M. T. Unit
(7 -channel; 200/556 bits/inch)
lS-50-72 KC M. T. Unit (7-channel;
200/556/S00 bits/inch)
72 KB M. T. Unit (9-channel;
SOO bits/inch)
144KB M. T. Unit (9-channel;
1600 bits/inch)
24-66-96 KC M. T. Unit (7-channel;
200/556/S00 bits/inch)
96 KB M. T. Unit (9-channel;
SOO bits/inch)
B2500 Controls:
36 KB Cluster Control (9-channel;
SOO bits/inch)
72 KB Cluster Control (9-channel;
1600 bits/inch)
50 KC M. T. Unit Control (7-channel;
200/556 bits/inch)
72 KC M. T. Unit Control (7 -channel;
200/556/S00 bits/inch)
96 KC M. T. Unit Control (7-channel;
200/556/S00 bitS/inch)
72 KB M. T. Unit Control (9-channel;
SOO bits/inch)
144 KB Unit Control (9-channcl;
1600 bits/inch)
96 KB M. T. Unit Control (9-channeJ;
SOO bits/inch)
7 -Channel Cluster Exchangc (2 x Ii)
9-Channel Cluster Exchangc (2 x 8)
7 - or 9 -Channel Magnetic Tape
Exchange (2 x 10) for B 9390
B 9391 and B9392 (B 2502 only)
9 -channel Magnetic Tape Unit Exchange
(2 x 10) for B9393 (B2502)
Adapter to convert B 23S1-1 to
7 -Channel Control
200 bitS/inch Adapter for B 23S1-1 Control
200 bits/inch Adapter for B 2393-1 Control
200 bits/inch Adapter for B 2393-3 Control
Unit Designate Switch for B 93S1
Series Clusters
9-25-36 KC 7-Channel Station Adapter
for B 93S1 Series Clusters
B3500 Controls:
36 KB Cluster Control (9-channel;
SOO bitS/inch)
72 KB Cluster Control (9-channel;
1600 bits/inch)
50 KC M. T. Unit Control (7-channel;
200/556 bits/inch)
72 KC M. T. Unit Control (7-channel;
200/556/S00 bitS/inch)
96 KC M. T. Unit Control (7-channel;
200/556/S00 bits/inch)
72 KB M. T. Unit Control (9-channel;
SOO bitS/inch)

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

5/69

8URROUGHS 82500/3500

210:221.105

PRICES

IDENTITY OF UNIT
CLASS
Model
Number
INPUTOUTPUT
(Contd. )

Feature
Number

Name

Monthly
Monthly
Purchase
Rental
Maint.
$(1)
$(1)
$(2)

B3500 Controls (Contd. ):
B3393-2
B3393-3
B3480
B3481
B3490
B3491
B3680
B3681
B3691
B3692
B9980
B9989

144 KB Unit Control (9-channel;
1600 bits/inch)
96 KB M. T. Unit Control (9-channel;
800 bits/inch)
7 -Channel Cluster Exchange (2 x 8)
9-Channel Cluster Exchange (2 x 8)
7 - or 9 -Channel Magnetic Tape Unit
Exchange (2 x 10) for B 9390, B 9391
and B 9392)
9 -Channel Magnetic Tape Unit Exchange
(2 x 10) for B 9393 (B3500)
Adapter to convert B 3381-1 to 7-Channel
Control
200 bits/inch Adapter for B 3381-1 Control
200 bits/inch Adapter for B 3393-1 Control
200 bits/inch Adapter for B 3393-3 Control
Unit designate Switch for B 9381 Series
Clusters
9-25-36 KC 7-Channel Station Adapter for
B 9381 Series Clusters

490

22,800

12

412

19,200

15

181
206
258

8,400
9,600
12,000

10
10
10

258

12,000

10

52

2,400

10

26
26
26
10

1,200
1,200
1,200
480

5
5
5
1

50

2,400

10

175
325
450
50
52
5

8,400
16,250
21,600
2,400
2,400
240

40
83
126
8
8
2

5
30
0

240
1,440
0

0
5
0

430
530
50

20,640
25,440
2,400

126
135
8

50

2,400

8

52

2,400

8

15

720

5

16

720

5

300
50
52
145
260
50
52
130

16,000
2,400
2,400
6,960
15,300
2,400
2,400
6,850

70
8
8
10
65
8
8
10

860
860

48,.000
4,800

180
180

Punched Cards
B9110
B9111
B9112
B2110
B3110
B9916
B9917
B9918
B9919
B9212
B9213
B2210
B2212
B3212
B2610
B3610

Card Reader 200 cards/min
Card Reader 800 cards/min
Card Reader 1400 cards/min
Card Reader Control for B 2500
Card Reader Control for B 3500
Validity Check Switch and Indicator for
B 9111 and B 9112 only
Card Counter for B 9111 and B 9112 only
P. M. O. Feature for B 9111 and B 9112 only
40 Column Read Switch for B 9111 and
B 9112 only
Card Punch (150 card/min)
Card Punch (300 cards/min)
Card Punch Control for B 2500
(B9212, B9211)
Card Punch Control for B 2500
(B 9213)
Card Punch Control for B 3500
(B 9213)
BCL-BCL Code Translator for
B 2210 Card Punch Control
BCL-BCL Code Translator for
B 3210 Card Punch Control
Paper Tape

B9120
B2120
B3120
B9926
B9220
B2220
B3220
B9928

Paper Tape Reader (500-1000 char/sec)
Paper Tape Reader Control for B 2500
Paper Tape Reader Control for B 3500
Input Code Translator
Paper Tape Punch-(100 char/sec)
Paper Tape Punch Control for B 2500
Paper Tape Punch Control for B 3500
Output Code Translator
~

B9242-1
B9242-2

5/69

860 lines/min - 120 print positions
725 lines/min - OCR "A" Numeric & std. Alpha

fA

AUERBACH

(Contd.)

PRICE DATA

210:221.106

IDENTITY OF UNIT
CLASS
Model
Number
INPUTOUTPUT
(Contd. )

Feature
Number

Name

PRICES
Monthly
Rental
$(1)

Purchase Monthly
Maint.
$(1)
$(2)

Printer (Contd. )
B9242-3
B9243-1
B9243-2
B9243-3
B9245-2
B9245-3
B2240
B2242
B3240
B3242
B9940
B9941
B9943
B9947
B9949
B9340
B2340
B3340

725 lines/min - OCR "B" - alphanumeric
1100 lines/min, 120 print pos. 44 char.
900 lines/min - OCR "A" - Numeric & Std.Alpha
900 lines/min - OCR "B" Alphanumeric
Printer (315 lines/min; 120 print positions)
Printer (315 lines/min; 132 print positions)
Printer Control for B 2500 (B 9240, B 9241,
B 9245-2, -3)
Printer Control for B 2500
(B 9242, B 9243)
Printer Control for B 3500 (B9240
B 9241, B 9245-2, -3)
Printer Control for B 3500
(B9242, B 9243)
High Speed Slew
Additional 12 Print Positions
Printer Memory for Series 9242/3
Dual Printer Control (for B 9240 and
B 9241 only)
Powered Forms Stacker
Console Printer and Keyboard
Console Printer Control for B 2500
Console Printer Control for B 3500
Ma~etic

B9130
B9131
B9131-1
B9132
B9134-1
B2130
B3130
B9930-1
B9930-2
B9931-1
B9931-2
B9932
B9933
B9934
B9935
B9936
B9937
B9938
B9939-1
B9939-2

860
965
965
965
500
540
75

4,800
53,500
53,500
53,500
24,000
25,920
3,600

180
200
200
200
135
135
8

75

3,600

12

78

3,600

8

78

3,600

12

60
40
100
200

2,000
2, 000
4,800
9,600

10
10
10
10

25
55
75
103

1,200
2,640
3,600
4,800

5
15
10
10

1,890
1,900
1,200
2,200
),025
100
103
5

500
500
615
650
390
12
12
0

20
25
200
50

90,720
91,200
91,200
105,600
49,200
4,800
4,800
240
15
960
1,200
9,000
2,400

7
10
10
10
10
5
5

275
450
450
450
450
240
240

0
0
0
0
0
0
0

100
150
100
156
0
200
10
10
5
25

4,800
7,200
4,800
7,200
150
9,000
480
480
240
1,200

12
15
12
15
0
50
1
2
0
5

Ink Character Sorter LReader

13-Pocket Non-System (1565 documents/min)
13-Pocket w/o Endorser (1565 documents/min)
13-Pocket w/o Endorser (1000 documents/min)
16-Pocket w/o Endorser (1565 documents/min)
Reader sorter (1565 documents/min)
MICR Sorter Reader Control for B 2500
MICR Sorter Reader Control for B :3500
Mobile Carrier and Document Tray
Document Tray, additional (for Purchase only)
Document Separators - 13 pocket
Document Separators - 16 pocket
Endorser for B 9131 and B 9132 only (3)
Extended Sort Control (for B 9130
and B 9131 only (3»
Start/Stop Bar (for B 9130 and B9131 only)
Special Field Ending
Override Code (as specified)
Validity Checking
Reverse Override
Item Counter, Resettable
Item Counter, Non -resettable

0
0
50
15

Reader Sorter Controls and Features
(for B9134-1)
B2130'-1
B2130-2
B3130-1
B3130-2
B9930-3
B9932-1
B9932-4
B9932-5
B9932-6
B9933-1

Reader Sorter Control (MICR only)
Reader Sorter Control (OCR/MICR/DUAL)
Reader Sorter Control (MICR only)
Reader Sorter Control (OCR/MICR/DUAL)
Mobile Carrier
Endorser - 1625 doe/min
Batch Ticket Detector
Short Document Read Feature
Short Document Module Expander
Basic Off-line Sort (2 fields only)

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

5/69

210;221.107

8URROUGHS 82500/3500

IDENTITY OF UNIT
CLASS
Model
Number

Feature
Number

Name

PRICES
Monthly
Rental
$(1)

Purchase Monthly
Maint.
$(1)
$(2)

Reader Sorter Controls and Features
(for B9134-1) (Contd.)

INPUTOUTPUT
(Contd. )
B9933-2
B9933-3
.B9933-4
B9933-5
B9933-6
B9933-7
B9933-8
B9933-9
B9933-10
B9935-1
B9935-2
B9935-3
B9936-1
B9937-1
B9938-1
B9938-4
B9938-5
B9938-9
B9939-3
B9939-4
B9939-5

8-Pocket Basic Off-Line Sort (2 fields only)
Expanded Off-Line Field Sort (8 fields max. )
Extended Sort Control
Zero Kill (max. of 3/reader sorter)
No Field - No Digit (max of 3/reader sorter)
Digit Override (max of 3/reader sorter)
Digit Edit (max of 3/reader sorter)
Field Override (max of 3/reader sorter)
Field Edit (max of 3/reader sorter)
Expansion Feature (Pockets 17-32)
Four Pocket Module (Pockets 5-16)
Four Pocket Module (Pockets 17 -32)
Stacker Overflow
Valid Character Check
Multi-Track E13B (1625 OPM)
Numeric OCR "A" (1625 OPM)
Numeric OCR "B" (1625 OPM)
Dual Read Option (1625 OPM)
Resettable Item Counter
Non - Resettable Item Counter
Running Time Meter

30
5
50
10
10
10
10
10
10
100
300
300
10
5
375
650
650
150
5
5
5

1,440
240
2,400
480
480
480
480
480
480
4,800
14,400
14,400
480
240
18,000
31,200
31,200
7,200
240
240
240

5
0
15
1
1
1
1
1
1
10
35
35
1
1
55
98
98
25
1
1
1

1,350
650

67,500
32,500

325
200

75
78

3,600
3,600

8
8

215
325

9,400
14,300

27
27

100

4,400

20

60
20
45
35
195

2,640
880
1,980
1,540
8,775

10
3
10
3
27

195

8,775

27

50
20

2,250
900

11
11

30
15
20
5
5

1,320
660
880
220
220
80
100
25

3
3
3
1
1

Listers
B9244-1
B9244-2
B2244
B3244

Master A/N Lister (1565 lines/min)
Slave A/N Lister (1565 lines/min
two maximum)
Lister Control for B 2500
Lister Control for B 3500
Input and Display Units

B9351-1
B9351-2
B9351-3
B9351-4
B9351-5
B9351-6
B9351-7
B9352-1

B9352-2
B9352-9
B9352-10

Control I - Single Input and Display/Location
Control II - Multiple Input and Displays
and/or Controls/Location
Control IIA - Max. of 3 with Multiplexor
and Control II
Monitor
Alphanumeric Keyboard
Input and Display Printer
Remote Communications Adapter
Input and Display Terminals - Includes:
40 x 20 Monitor, Alphanu Keyboard,
960 char. Memory, Control and
Variable Tab function.
Input and Display Terminal - Same as B9352
except 80 x 12 monitor
Modem Expander
Additional 4 lines to Modem Expander
Options (B9351 Series)

B9951-1
B9951-2
B9951-3
B9951-4
B9951-5
B9951-7
B9951-8
B9951-9

5/69

Input and Display Printer Adapter
Insert/Delete - Character/Line
Controlled Format
Variable Tab Position
Programmatic Cursor Positioning
Display Stand, Low, without work table
Display Stand, High without work table
Display Stand Work Table, Right or Left

A

-

-

(Contd.)

AUERBACH


PRICE DATA

210:221.108

IDENTITY OF UNIT
CLASS
Model
Number
INPUTOUTPUT
(Contd. )

Feature
Number

Name

PRICES
Monthly
Rental
$(1)

Purchase
$(1)

Monthly
Maint.
$(2)

Options (B9352 Series)
B9952-1
B9952-1
B9952-3
B9952-4
B9952-5
B9952-6
B9952-7
B9952-9

Input & Display Printer Adapter
Polling & Select
Controlled Format
Remote Comm. Interface - Direct
Connect 1200 bits/sec
Remote Comm. Interface - Direct
Connect 2400 bits/sec
Remote Comm. Interface - Modem-102
Type-up 300 bits/sec
Remote Comm. Interface - Modem -202
Type-up to 1200 bits/sec
Remote Comm. Interface - Modem-201
Type-up to 2400 bits/sec

10
10
8

675
450
450
360

15
2
2
1

12

540

3

5

225

1

8

360

1

12

540

3

55

2,640

12

175
1,095
125
370
ll5
795
175
1,095
129
370
115
7%

15
100
14
30
10
31
15
100
14
30
10
31

10

8,400
52,560
6, 000
17,760
5,520
37,200
8,400
52,560
6,000
17,760
5,520
37,200
2,575
575
480

35
30
30
45

1,680
1,440
1,440
2,160

5
5
5
10

50
65
50

2,400
3,120
2,400

5
10
5

65

3,120

10

85
100
60
75
50
30
50

4,100
4,800
2,880
3,600
2,400
1,440
2,400

5
10
5
10
5
5
5

50

2,400

5

Remote Unit
B9350
COMMUNICATIONS

Typewriter Inquiry Station
Controllers

B2350-1
B2350-2
B2351
B2353
B2354
B2355-1
B3350-1
B3350-2
B3351
B3353
B3354
B3355-1
B9955-1
B9955-2
B9950

Terminal Control for B 2350-2 CTU
Central Terminal Unit for B 2500
Single Line Control for B 2501 and B 2502
Basic Multi-Line Control for B 2502 only
8-Channel Extension for B 2353
Voice Response Generator
Terminal Control for B 3350-2 CTU
Central Terminal Unit for B 3500
Single Line Control for B 3501
Basic Multi-Line Control for B 3501
8-Channel Extension for B 3353
Voice Response Generator
Audio Recording (Special)
Audio Recording (Library)
Audible Alarm for CTU

-

-

1

Control and Line AdaQters
B2650
B2651
B2652-1
B2652-2
B2653-1
B2653-2
B2653-3
B2653-4
B2654-1
B2654-2
B2655-1
B2655-2
B2656-1
B2657
B2659-1
B2659-2

B2500:
CTU Adapter for B 2350-1
Typewriter Inquiry Station
TWX/Remote Typewriter
TWX/Remote Typewriter with
Automatic Dial Out
B 2500/B 3500
B 2500/B 3500 w/ADO
B 300/B 500/B 5500 Modem 201 Type to 2400 bits/sec
B 300/B500/B 5500 Modem 201 Type to 2400 bits/sec w/ ADO
DCT 2000
DCT 2000 w/ ADO
IBM 1050
IBM 1050 w/ ADO
IBM 1030
Model 35 on 8A1 Selective Calling Service
Input and Display - Direct Connect
(Multi-Wire)
Input and Display - Data Set
(1200 bits/second)

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

5/69

210:221.109

BURROUGHS B2500/3500

IDENTITY OF UNIT

PRICES

CLASS
Model
Number
COMMUNICATIONS
(Contd. )

Feature
Number

Name

Monthly Purchase
Rental
$(1)
$(1)

Monthly
Maint.
$(1)

Control and Line AdaQters (Contd.)
B2659-3
B2659-4
B2659-5
B2659-8
B2659-9
B2659-10
B2659-11
B2659-12
B2659-13
B2659-14
B2662
B2663
B2664-1
B2664-2

Input and Display - Data Set
(2400 bits/second)
Input & Display - 1200 bps w/ADO
Input & Display - 2400 bps w/ADO
B9352 Series I & D - Direct Connect
B9352 Series I & D - Modem-202 Type
up to 1200 bits/sec
B9352 Series J & D - Modem-201 Type
up to 2400 bits/sec
B9352 Series I & D - Morovided for indexing
(using a single tndex register) and for indirect addreSSing (non-recursive).
The 160 has an instruction repertoire of 97 instructions, most of which are variations
of a few basic instructions. The basic arithmetic mode is fixed-point binary on singleword operands. There are add-to-storage and Boolean instructions, but no standard
facilities are provided for multiplication, division, direct comparisons, . radix conversions, or floating-point arithmetic. In the basic 160 system, these functions are
usually performed by standard subroutines. The optional 168-1 Auxiliary Arithmetic
Unit provides automatic facilities for double-precision addition and subtraction and for
single-precision multiplication and division, all in fixed-point mode. The 168-2 Auxiliary Arithmetic Unit speeds the execution of floating-point arithmetic operations.
Although the basic ll-bit add time is only 19.2 microseconds, double-precision addition
takes 225 microseconds using subroutines and 145 microseconds using the 168-1 Arithmetic
Unit. Floating-point add times are about 4,000 microseconds using subroutines and 1,000
microseconds using the 168-2 Arithmetic Unit.

C 1968 AUERBACH Corporation and AUERBACH Info, Inc.

4/68

242:011. 200

.2

CDC 160

HARDWARE (Contd.)
Simultaneous operations cannot be performed in Control pata 160 systems; the central
processor is interlocked during all input-output operations. There is no interrupt system.
The principal input-output devices used with Control Data 160 systems are the built-in
tape reader and punch and the follOwing optional units:
• An on-line typewriter, which provides the typewriter input-output facUities

that are lacking in the basic 160 system.
• A card reader, rated at either 250 or 1,200 cards per minute.
• A card punch, rated at either 100 or 200 cards per minute; the slower model
is an IBM 523.
• A line printer with a rated speed of either 150 or 1,000 lines per minute.
• Up to four magnetic tape units. Either the 163 Magnetic Tape Subsystem or
the newer 603 Magnetic Tape Units can be connected. Both use IBM 729compatible tape in either BCD or binary mode. The 163 subsystem consists
of one to four tape handlers and a control unit; its peak speed is 30, 000 characters per second at a recording density of 200 rows per inch. The 603 has
peak speeds of 15, 000 or 41, 667 characters per second at recording densities
of 200 and 556 rows per inch, respectively.

•3

A 160 computer can be connected on-line to a larger Control Data 1604, 1604-A, or 3000
Series computer. Data can be transferred between the coupled systems by direct coreto-core transfers or by way of shared magnetic tape units •
SOFTWARE
Because of the Control Data 160's short word length and machine-language programming
complexity, the software available for the system is of particular importance. A useful
assortment of programming systems and subroutines, developed by Control Data and by
160 users, is now available, although many significant programs that have been developed
for the 160-A cannot be run on 160 systems. The software facilities that are properly
documented and in general use are supplied through a well-organized, CDC-supported
users' group caJled SWAP.
OSAS is the basic symbolic assembly system for the 160 and 160-A., It provides no
facilities for macro-instructions, but library subroutines can be assembled along with
the user's source programs. Only about 250 symbolic labels can be accommodated by
the OSAS translator for the 160. The translator is available in different versions for
systems that use paper tape, magnetic tape, or punched card input-output.
A compiler is available for 160 FORTRAN, a restricted but useful version of the
FORTRAN II language. The restrictions are imposed mainly by the hardware limitations of the 160 itself. Fixed-point arithmetic is limited to Single precision (11 bits),
while each floating-point variable occupies three words of core storage. Mixed-mode
arithmetic is permitted. Object programs compiled by 160 FORTRAN are executed
interpretively.
INTERFOR is a floating-point interpretive system for the 160. Its repertoire of 22
instructions is a subset of the larger Control Data 1604's machine-language instruction repertoire; thus, INTERFOR makes it easy for 1604 programmers to write programs that can be executed interpretively on a 160. Each floating-point data value
occupies four words of 160 storage. standard INTERFOR subroutines handle input,
output, and mathematical functions. INTERFOR programs can be written directly
in octal format, or they can be written in a more convenient symbolic format and
assembled by FLAP, a special-purpose assembler.
Among the library subroutines available for the 160 are routines to perform singleprecision multiplication and division, multiple-precision fixed-point arithmetic,
decimal arithmetic, floating-point arithmetic, mathematical functions, radix conversions, ma'trix inversion, and data transcriptions (card-to-tape and tape-to-printer).

4/68

A

AUERBACH

•

242:221.101
CDC 160
PRICE DATA

CDC 160
PRICES

IDENTITY OF UNIT
CLASS

PROCESSOR

.-- ..
ATI A~'I1:\IENTS,
.\DAPTERS,
AND
CHANNELS

Model
Number

Feature
Number

*160
*168-1
*168-2
*3681G

Name

Monthly
Monthly
Rental Purchase Maint.

$
Processing Unit (includes 4,096 words of core
storage, paper tape reader, and paper tape
punch)
Auxiliary Arithmetic Unit (fixed point)
Auxiliary Arithmetic Unit (floating point)
Data Channel Converter (permits use of CDC
3000 series peripheral equipment)

$

$

1,600

63,000

200

410
475

12,500
14,000

125
130

290

11,500

20

580
530

25,000
21,000

120
105

1,000
1,550
2,100
2,650

40,500
62,500
84,000
105,500

235
400
565
730

420
420
105
310
121
350

16,500
23,500
5,000
19,000
5,100
14,500

140
65
10
60

720
1,950

31,500
77,000

325
400

465
275

16,000
11,000

110
95

220

7,400

80

Magnetic Tape

I~Pl'T-

OlTTPlTT
*603
*162-1
*163-1
*163-2
*163-3
*163-4

Magnetic Tape Unit
Tape Synchronizer (for 1 to 4603 tape units)
Magnetic Tape Subsystem:
One tape handler and control
Two tape handlers and control
Three tape handlers and control
Four tape handlers and control
Punched Card

*167-1
405
*177G
415
523
*170G

Card Reader (250 cards/min.)
Card Reader (1200 cards/min.)
Card Reader Controller (for 405)
Card Punch(200 cards/min.)
IDM Card Punch (100 cards/min.)
Card Punch Controller (for 415 or 523)

-

55

Printers
*166-2
*1612

Line Printer (150 lines/min)
Line Printer (1000 lines/min)
Typewriters

*161F
*161G

On- Line Input-OUtput Typewriter (Flexowriter)
On- Line Input-OUtput Typewriter
Plotter

*165-1

Incremental Plotter

;\OTES:
*;\0 longer in production.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

-A

243:011. 100
SI......

~EDP

AU(~

-

CDC 1604-A

.. PI ...

SUMMARY REPORT

SUMMARY REPORT: CONTROL DATA 1604-A
•1

AVAILABILITY
The Control Data 1604-A is a medium-to-Iarge-scale data. processing system that is primarily oriented toward scientific, simulation, or control applications in which extensive
computations are performed.
The 1604-A was introduced in 1962 to replace the 1604, Control Data's first generalpurpose digital computer system. The 1604-A features a more powerful interrupt system
and the ability to handle higher input-output data rates with lower central processor delays.
(A 1604-B computer was also manufactured; it provided the basic 1604 facilities plus increased input-output transfer capacities.) A total of approximately sixty 1604 and 1604-A
systems were manufactured before production facilities were shifted over to Control Data's
neyver 3000 Series systems in 1964. Used 1604 and 1604-A systems can be purchased or
rented from Control Data Corporation, and Control Data assures prospective users that it
will continue to provide full hardware maintenance and software support services. No
inventory of used 1604-A systems is maintained; they are offered on an "as available"
basis, and a turnaround time of about three months is required to refurbish each returned
system •

•2

HARDWARE
A Control Data 1604-A system can contain 8,192, 16,384, or 32,768 word locations of
core storage. Each 48-bit word can hold a fixed-point or floating-point number, a binary
data pattern, or two 24-bit instructions. The core storage is divided into two banks, each
with independent access facilities and a 6. 4-microsecond cycle time. The resulting capability to overlap core storage accesses leads to an effective cycle time of approximately
4.8 microseconds. No parity checking is performed on data transferred to or from core
storage.
The 1604-A Central Processor operates in the binary mode on 48-bit operands in either
fixed-point or floating-point form. Floating-point data values are represented by an 11bit exponent and a 36-bit-plus-sign fraction. A useful repertoire of arithmetic, logical
branching, storage search, and data transfer operations is provided, but there are no
direct facilities for radix conversion, format control, or multi-word data transfers.
Approximately 50,000 or 100,000 instructions can be executed per second in floatingpoint or fixed~point mode, ,respectively.
Each instruction consists· (!)f'24 bits: a 6-bit operation code, a 3-bit index designator, and
a 15-bit address portion. The index designator can specify either indexing, using one of
six index;-reg1sters, or indirect addressing, which may be recursive. The address portion
can speCify anoperand,ad'ystem includes thc following(l): 1n,750
Central Processing Unit
6 Peripheral Processing Units (includes 11aintenance Control Unit)
Operator station
Disc File station
Card Reader
Card Punch
2 Printers
:2 l\lagnetic Tape Drives

7602-1
7611-1
763t'-1
407-1
417-1
5517-1
617-1

Peripheral Processing Unit
Operator station
Disc File System
Card Reader
Card Punch
Printer
l\lagnetic Tape Drive

7601-1
7602-1

1,250
5,685
12,550
460
345
1,110
~) ~J fi

5- Year
Lease
155,475

1,125
5,115
11,295
415
310
1,000
895

Purchase

$

Monthly
Maint.

$

8,250,000

23,225

50,000
236,500
400,000
23,500
19,000
45,000
44,000

150
1,010
1,655
115
105
405
225

XOTES:
(1)

11w user must acquire the basic

s~'stem

package.

.\c1c1-on components are optional.

1969 AUERBACH Corporation and AUERBACH Info, Inc.

4/69

GENERAL ELECTRIC
COMPANY

(

)

AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

AUERBACH
~
Printed in U.S.A.

309:011.100

A.

AUERBACH

STANDARD

EDP

GE-I05

REPORTS

ADV ANCE REPORT

ADVANCE REPORT: GE/10S
.1

BACKGROUND
The GE-I05, announced in January 1969, further expands the system options available for users of
GE's small-scale business-oriented systems. The marketing effort is directed to users currently
employing manual or tab processing techniques as a means of "stepping-up" to the expanded data
processing possibilities that a general-purpose computer system can provide.
As was true of the earlier GE-115 systems, the GE-I05 will be produced by the General Electric
Information System of Italy (GEISI) in collaboration with General Electric (USA). GEISI, originally
Olivetti-GE, is now a wholly-owned subsidiary of GE, and GE has plans to market the new system
both in the U. S. and Europe.
The GE-I05 is upwards compatible within the GE-I00 series, permitting growth as the user's data
processing needs expand to the faster GE-115 and GE-130 systems, which have larger memory capacities and higher speed card processing equipment and printers, as well as magnetic tape drives, disc
units, and communications controllers. The larger systems also offer COBOL and FORTRAN compilers, and comprehensive tape and disc operating system support.
Software support for the GE-I05 is derived primarily from proven GE-115 routines, including an
assembly programming system (APS); a report program generator (LOGEL); punched card processing
routines (List and Summarize, Reproduce and Gangpunch); and a variety of utility routines providing
such functional aids as memory dumping, card loading, reading, punching, printing, etc.
The emphasis is on "simplicity of operation" to facilitate the transition from tab processing to computer operations. The system is compact, it can be installed anywhere in a user's installation, and
it is designed for quick user familiarization of programming and operational procedurcs.
Two models of the GE-I05 are expected to be available by July 1969. The Model A, including a card
reader, card punch, printer, and 4,096 core storage locations, leases for $1, 250 and sells for
$57,370. The Model B, with a combination card reader/punch and printer, and an 8,192 location
storage capacity, carries a monthly rental of $1,450 and a purchase price of $66,410. The user will
be offered the package configuration with each Model, and these peripheral components cannot be
ordered separately or selectively. The projected delivery schedule target is an interval of 6 months
following the receipt of an order .

.2

HARDWARE

.21

Data Structure
The basic unit of data storage is an eight-bit (plus 1 parity bit) byte, a standard GE-I00 series data
unit, which GE refers to as \ an "octet" and is equivalent to an IBM byte. Each octet is directly
addressable and can store an alphanumeric character, two packed decimal characters or one 8-bit
binary operand.
GE-I05 arithmetic instructions perform operations on two data fields, each of which can contain
from one to sixteen digits in length. The non-arithmetic instructions process fields up to 256 alphanumeric characters in length. The instruction'\lengths' are two, four, or six characters, and zero,
one, or two core storage addresses can be specified depending on the instruction function. There are
a total of 39 separate instructions including decimal and binary arithmetic, data transfer, packing
and unpacking, comparison, edit, translate, logical, search table, etc.
Note that there is no direct compatibility between the GE-I05 and the IBM System 1360 Model 20,
although both systems use 8-bit character codes .

. 22

System Configuration
The GE-I05 system is offered in two basic packages, both of which are dedicated to punched-card
processing, and, hence, as replacements for unit record accounting operations.
Package A, as specified below, differs from Package B in that A has a faster card reader and a
slower card punch and printer. The two packages are designed to accommodate the needs of users
who place greater emphasis on either input or output processing. Package A also provides the option
of reduced core storage capacity and reduced rental by eliminating the additional memory module (AMI05).
© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

5/69

ADVANCE REPORT: GE-I05

309:011.101

Typical Card System: Standard Configuration I (Gh"105 Package A)
Equipment

Rental

1 - GE-105 Model A04 Central Processor, with 4, 096 characters of core storage

$1,250

1 - AM105 Additional Memory, provides 4,096 additional characters of core storage

180

1 - GE-105A04 Card Reader; 350 cpm

NC

1 - GE-105A04 Card Punch; 60-200 cpm

NC
NC

1 - GE-105A04 Printer; 250 1pm
Total Rental

$1,430

Typical Card System: Standard Configuration I (GE-105 Package B)
Equipment

Rental

1 - GE-105 Model B08 Central Processor, with 8,192 characters of core storage

$1,450

1 - GE-105B08 Card Reader/Punch; reader, 300 cpm, punch, 300 cpm

NC

1 - GE-105B08 Printer, 300 cpm
Total Rental
.23

NC
$1,450

Central Processor
Both the GE-105A and GE-105B Control Processors are composed of four functional units: a Store
Unit, a Command and Control Unit, an Arithmetic Unit, and a Peripheral Control Unit. These units
are housed in a "U"-shaped cabinet comprising three wings, one of which functions as a printer controller.
The Command and Control Unit fetches the instructions from core storage, and then decodes and
controls the execution. The Arithmetic Control Unit performs the actual execution of all instructions
except the I/O operations.
The Peripheral Control Unit provides two I/O channels and three peripheral connectors for the control of the transmission of data between core storage and the peripherals. An overlap feature allows
the simultaneous operation of all the peripherals .

. 24

Internal Storage
The Central Processor Store Units in both systems operate at a machine cycle rate of 7.5 microseconds - slightly less than the 6.5 microsecond cycle time of the GE-115.
Package B provides 8,192 core storage locations, while package A offers either 4,096 or 8,192 locations. Each location can accommodate an eight-bit alphanumeric character or "octet" plus one parity
bit. Odd-parity checking is performed on all data stored in or read from core.
There are no index registers and, consequently, no address modifications are possible other than
direct modification of the instruction .

. 27

Peripheral Equipment
GE-105A Card Reader
This unit operates at a speed of 350 cards per minute, and can read standard 80-column cards punched
in either Hollerith or column-binary (core storage, two decimal digits per column), or it can read
them intermixed. The cards are read in a serial fashion, column-by-column, as they pass by twelve
photoelectric cells. A 51-column option is available.
The single input hopper and the primary output stacker can each hold a maximum of 2000 cards. One
auxiliary output stacker is available with a 500-card capacity, and stacker selection under program
control is permitted.
Error conditions requiring operator attention include malfunctioning of the photocells, columns misread, input hopper empty, output stacker full, end-of-file mark sensed, and card jam.

5/69

A

(Contd.)

AUERBACH
®

ADVANCE REPORT: GE-I05

309:011.102

An early termination feature permits program control of the number of columns read, and a computeroverlap feature allows the feed and eject portions of the read cycle to be overlapped by processor
activity.
GE-105A Card Punch
This unit punches the standard SO-column cards at 100 columns per second, or 60 to 200 cards pcr
minute depending on the number of columns punched in each card. Hollerith or column-binary cards
are fed in a serial manner. A 1500-card input hopper and output stacker are provided.
Error checking includes card synchronization, parity, chad box full, and echo checking while punching, in addition to the card feed, card jam, stacker full, and hopper empty checking that takes place
on the card reader.
The unit has a fully buffered controller providing'overlap of punching and compute functions, as well
as other peripheral operations.
GE-105A Card Printer
This is a 250 line per minute printer that prints a 120 character line (136 characters optional) at a
density of 10 characters per inch. Skipping control is by a program space command or a paper tape
control loop. All 64 alphanumeric characters can be printed. Error checking features include out
of paper, end of page, hammer failure, and memory transfer parity error.
GE-105B Card Reader/Punch
This unit both reads and punches separately or intermixed at a speed of 300 Hollerith or columnbinary cards per minute. The cards are fed in a parallel manner and the reading and punching are
performed row-by-row. The same card may be read and then punched.
The reader/punch controller is fully buffered permitting simultaneous compute and card operations.
A separate SO-character external buffer is also provided which is shared by both the reader and
punch.
System Package B has decreased the operational speed of the reader and increased the speed of the
punch over that of Package A, thus the emphasis on output as opposed to input processing capabilities.
GE-105B Printer
This unit is identical in all respects to the GE-105A Printer with the exception that the operational
speed has been increased to 300 from 250 lines-per-minute. This increased printer capability further emphasizes the output orientation of Package B .
.3

SOFTWARE
There has been no comprehensive operating system developed to date for the GE-105, as contrasted
with the Extended Tape and Disk Operating Systems provided with the larger GE-115 and GE-130
systems, primarily because of the punched-card orientation of the GE-I05. All of the card-oriented
GE-115 software may be used with the GE-I05 including an Assembly Programming System, an
extended report program generator called LOGEL, and an assortment of utility card processing
programs for listing, summarizing, reproducing, and gangpunching.
The Assembly Programming System (APS) is a basic card-oriented language processor which translates source programs written in assembly language into object programs. Mnemonics are provided
for all GE-105 instructions. In addition, supplementary mnemonics extend the range of the conditional jump instructions, and pseudo mnemonics control memory assignment and program listing
formats.
Data constants may be coded in decimal, alphanumeric, or hexadecimal form. APS provides listings
of both the source and object programs, a listing of all symbolic names used in the program, and
programming error diagnostics.
LOGEL is a logic-generating language offering the user a simple file-management card system including report generation. LOGEL is a medium-level compiler divided into five divisions; GENERAL,
which defines the input and output files; INPUT, wh~h defines the input records; DATA, which
specifies the data fields, constants, and core storage areas; CALCULATION, which contains the
instructions to perform the required operations; and FORMAT, which specifies the format for the
print lines and the output records.
The LOGEL program essentially generates sets of routines and the logical linking of these routines
in accordance with the source program statements. Assembler instructions may be intermixed with
the LOGEL statements.
© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

5/69

ADVANCE REPORT: GE-I05

309:011.103

The List and Summarize Program Generator creates reports from a punched-card file controlled by
"parameter cards", which also define the file format. The object program produced to generate the
reports is listed and punched into an object program deck.
Reproduce and Gangpunch operations are performed by a prior-developed program and a set of
specification statements for each given run. A variety of card processing functions can be performed:
the reproduction of a deck of cards from one to nine times for each run, the transfer or selection of selected fields, the gangpunching into detail cards of fields contained in master cards, and the calculation of a product or percentage from a field in each detail card and a field in a master card.
All of these card-oriented programs are currently available with the GE-1l5 system. They require
only 4,096 core storage locations, so that a GE-I05A04 configuration will suffice.
Several GE-1l5 business application packages have been developed for such applications as invoicing,
inventory control, payroll, purchasing, general accounting, production control, public utility services,
etc. Those application packages specifically oriented toward card processing will be available with
the GE-I05 including SIMTAB for the simulation of tab jobs, a 4, 096 core capacity payroll program,
PROCON-1l5 for production control, and Critical Path Method (CPM).
CPM is a management tool for planning, scheduling, and controlling business, industrial, and scientific
projects. A model is constructed through specification statements showing the critical path network
of all events contained in the project. Up to 350 nodes (events) can be handled in a minimum 4,096
character storage configuration. The end product are reports designed to simplify project evaluation
and review.

5/69

fA

AUERBACH
@

(Contd.)

-£

309:221.101
STAND'"

AEDP

AUERBAC~
e

-

GE-l05
PRICE DATA

REPORTS

PRICE DATA

IDENTITY OF UNIT
CLASS
Model
Number
BASIC
SYSTEMS

Feature
Number

Name

PRICES
Monthly
Rental

Purchase

Monthly
Maint.

$

$

$

105A04

GE-105A Systenl including: 4,096
octets* of core storage
Card Reader (300 cards/min)
Card Punch (60 to 200 cards/
min)
Printer (250 lines/min)

1,250

57,370

261

105A08

GE-105A System including: 8,192
octets* of core storage
Card Reader (300 cards/min)
Card Punch (60 to 200 cards/
min)
Printer (250 lines/min)

1,430

65,890

274

105B08

GE-105B System including: 8,192
octets* of core storage
Card Reader/Punch (300 cards/
min)
Printer (300 lines/min)

1,450

66,410

288

35

1,540

8

OPT105

16 additional print positions for
105A04, 105A08, or 105B08

Notes:
*Each octet consists of eight data bits plus one parity bit.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

5/69

310:000.002
GE-115
REPORT UPDATE

REPORT UPDATE
~

ADDITIONAL PERIPHERAL UNITS EXPAND GE-115 AND GE-130 SYSTEMS
General Electric has expanded the capabilities of the GE-115 and GE-130 systems by providing
another magnetic Tape Subsystem. This is in addition to the two magnetic tape units (MTH106
and MTH103) added previously. A paper tape punch was also added to the product line.
Magnetic Tape Units
The new tape subsystem consists of a MTC 114 Controller used with a MTH 117 Tape Handler.
This is a seven-track unit compatible with the GE-400 and the IBM 729 and 7330 units. '.The
MTH 117 operates at a speed of 37.5 inches per second, and records at 200 or 556 rows per inch
with a data transfer rate of either 7,500 or 21,000 bytes per second, respectively. The unit can
read backward, and has a rewind speed of 75 inches per second. The inter':'record gap is 0.75
inches, with a stop/start time of 21. 1 milliseconds for continuous read/write, 31.1 milliseconds when the tape stops moving, and 16.9 milliseconds when the Central Processor is busy.
Up to six tape units can be connected to each controller. Connection can be through a multiple
peripheral adapter, or directly on connector 3 on the GE-115, and connector 3 or 4 on the
GE-130.
The tape used is standard 1/2 inch, 1.5 mil iron oxide coated Mylar film tape on 1200 and
2400-foot reels.
The following chart can be used to compare the new Tape Subsystem with the MTH 103 and MTH
106 Magnetic Tape Handler previously reported.
Model No. of
Tracks
MTH
7
103 (uption)

9

MTH
7
106 (option)
9

Tape Tape
Speed Density
(IPS) (BPI)

Transfer
Rate
(Kc)

InterRecord
Gap
(Inches)

Stop/
Start
(ms)

Compatibility

37.5

200
556
800

7.5
21
30

0.75

21.1(a)
31.1(b)
16. 9(c)

GE-400
IBM 729
IBM 7330

37.5

800

30

0.60

16.0(a)
26.0(b)
11.8(c)

IBM 2400

75

200
556
800

15
42
60

0.75

75

800

60

0.60

10.8(a)
20.8(b)
8.8(c)
tl.3(a)
18.3(b)
6.3(c)

GE-400
IBM 729
IBM 7330

Read
ReBack- wind
ward (IPS)

Yes

75

Yes

150

IBM 2400

(a) Nominal inter-block gap time in continuous read/write.
(b) Stop/Start time when tape stops moving.
(c) Stop/Start time when Central Processor is busy.
Paper Tape Punch
The PTP 110 Paper Tape Punch for the GE-100 line has a speed of 60 characters per second.
Tape used can be the standard 1-inch, 7- or 8-channel tape, or 11/1S-inch, 5-channel tape, with
10 characters per inch. Each self-loading 8-inch reel holds 1,000 feet of tape, which can be
Mylar. Checking is performed at the end of the reel, with a parity check on transfer between
memory and the controller, and checking for broken or torn tape. Connection can be by a
standard interface on connector 3 or 4 or a multiple peripheral adapter on connector 3 for the
GE-1l5 and connector 3 or 4 for the GE-130.

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

12/68

310:000.003

A•

AUERBACH

SliMDnD

EDP

GE-IOO SERIES
REPORT UPDATE

REPGRTS

REPORT UPDATE
~ TWO NEW HIGH-DENSITY MAGNETIC TAPE SUBSYSTEMS INTRODUCED
GE recently announced the availability of two new nine-track, 1600 bit-per-inch magnetic tape units
for use on the GE-115 and GE-130. The MTS-163 and MTS-166 Magnetic tape subsystems both comprise a controller and one tape unit. Up to seven additional tape units can be connected to each
subsystem.
The MTH-163 Magnetic Tape Unit is IDM 2400 Series-compatible, operates at a tape speed of
IS. 75 inches per second, and has a maximum transfer rate of 30 KC. The MTH-166 Magnetic Tape
Unit, which is also IDM 2400 Series compatible, is twice as fast with a tape speed of 37.5 inches per
second and a maximum transfer rate of 60 KC. Both subsystems rewind at 300 inches per second
and can read forward or backward in continuous or start-stop mode. Seven-track, IBM 729-compatible options are available for both sUbsystems with lower recording densities of 200, 556, and SOO
bits per inch. These same recording densities are also available for the nine-track unit.
The subsystem prices are summarized as follows:
Monthly
Rental

Purchase
Price

Monthly
Maintenance

SOO

$36,600

$S3

MTH-163 Magnetic Tape Uandler*

300

13,520

45

OPT-1S3 Controller Option; 200,556, SOO bpi

200

9,910

16

OPT-193 Magnetic Tape Handler: 9-track; 200,556,
SOO bpi

25

1,130

4

OPT-173 Magnetic Tape Handler; 7-track; 200,556, SOO
bpi

25

1,130

4

1,200

53,910

122

MTH-166 Magnetic Tape Handler*

450

19,290

64

OPT-1S6 Controller Option; 200,556, SOO bpi

200

9,910

16

25

1,130

4

25

1,130

4

MTS-163 Magnetic Tape subsystem, includes one
controller and one MTH-163 Magnetic Tape Handler,
1600 bpi

MTS-166 Magnetic Tape subsystem, includes one
controller and one MTH-166 Magnetic Tape Handlel'
, 1600 bpi

OPT-196 Magnetic Tape Handler;
200,556, SOO bpi

9~track;

OPT-176 Magnetic Tape Handler; 7-track; 200,556, SOO
bpi

$

*Up to seven additional handlers can be added to the basic magnetic tape SUbsystems.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc,

5/69

310:011. 010

1&

AUERBACH

SUNDUD

EDP

GE-115

I[PORlS

SUMMARY REPORT

~

SUMMARY REPORT: GE 115
. 01

INTRODUCTION
The GE-115, announced in March 1965, represents the first joint development and
marketing effort by General Electric (USA), Bull-GE (France), and Olivetti-GE (Italy).
Designed by Olivetti-GE, the 115 was first marketed as a free-standing system in Europe
and, with limited peripheral device support, as a remote terminal system in the United
States market.
In May 1966, customers in the U. S. A. were offered the GE-115 as a full-scale computer
system and the peripheral device complement was correspondingly expanded. Initial
emphasis was placed upon using the GE-115 as a "first step" into data proceSSing for
punched-card tabulating installations. In early 1967, GE rounded out its U. S. marketed
small-scale system with the addition of magnetic tape units, random access disc storage
units, and an improved line of card readers and printers, all of which had previously been
offered in Europe.
With this increased capability, the GE-115 moved into direct competition with such systems
as the IBM System/360 Model 20, Burroughs B 100, Honeywell 120, RCA Spectra 70/15,
and the Univac 9200/9300 systems.
Within this highly competitive group, the GE-115 offers much in the way of high-speed
peripheral devices, very impressive software support, and an efficient central processor
with a good instruction repertOire. Although system performance using the AUERBACH
benchmark problems has yet to be measured, it is evident that the combination of the
above factors make the GE-115 a real contender in the small computer sweepstakes.
A GE-115 system with a card reader, printer, and communications adapter (a typical
remote terminal configuration) can be rented for as little as $1, 370 per month. Typical
card systems will range between $1, 655 and $2, 615 per month; typical tape systems will
range between $3, 640 and $4, 600; and typical disc systems will range between $3, 920 and
$6, 410. First system deliveries were made in early 1966, and the present delivery
schedule is three to six months for most components.

Figure 1. An expanded GE-115 configuration.

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

10/67

GE-1I5

310:011. 020

.01

INTRODUCTION (Contd.)
Significant features of the GE-115 include:
•

Up to 16. 384 eight-bit character positions of 6. 5-microsecond core storage
(Paragraph. 041).

•

Over 59 million alphameric characters of on-line random-access storage using the
DSS130 Disc Storage Unit (Paragraph. 042).

•

Card readers with speeds of 300 or 600 cards per-minute (Paragraphs . 071 and
.072).

•

Economical 300-card-per-minute photoelectric card reader/punch (Paragraph. 073).

•

Card punching speeds of up to 300 cards per minute (Paragraphs. 074 and .075).

•
•

Punched paper tape reading at 500 characters per second (Paragraph . 076).
Printing at speeds of 300. 600, or 780 lines per minute (Paragraphs . 081, • 082,
and .083).

•

Magnetic tape data transfer rates of 30KC and 60KC (Paragraphs. 091 and .092).

•

Adapters that permit communication with a remote computer system
(Paragraph. 101).

•

Ability to perform two unbuffered I/o data transfer operations simultaneously, plus
as many additional data transfers as there are buffered I/o devices connected to
the processor. Available buffered I/o devices include a printer, several card
punches, and a card reader/punch (Paragraph. 11).

•

.02

A wide range of software (most of it already delivered) that includes language
processors, service routines, a logic generating language (more extensive than
typical report program generators), input-output control systems, operating
systems, utility programs, applications packages, and data communications
packages. Claiming a high degree of software efficiency, GE states that an overwhelming majority of its card-oriented systems sold or leased to date are equipped
with only 4K-characters of core storage. Because of this, each card system configuration in Paragraph. 03 specifies prices at both 4K- and 8K-characters of core
storage. (The AUERBACH standard System Configuration for typical card systems
requires use of at least 8K characters of main memory. )
DATA STRUCTURE
The basic unit of data storage is an eight-bit (Plus 1 parity bit) byte, which GE prefers to
call an "octet". Each octet can contain one alphameric character. two decimal digits
(packed), a one-decimal-digit arithmetic operand, or an 8-bit binary operand. Decimal
arithmetic is performed on unsigned 4-bit BCD digits ~ digit per octet): the remaining
four bits of each octet are ignored. This mode of arithmetic may sacrifice efficiency in
the use of core storage, but, by avoiding repetitive packing and unpacking operations it can,
in many cases, result in a savings of processor time.
GE-115 arithmetic instructions can process operands from 1 to 16 digits in length. Most
others, including code translation and editing instructions, can operate on fields of up to
256 characters. GE-115 instructions are two, four, or six characters in length and
specify zero, one, or two core storage addresses, respectively.
Note that there is no direct compatibility between the GE-115 and the IBM System/360
Model 20, although both systems use 8-bit character codes •

• 03

SYSTEM CONFIGURATION
Every GE-115 computer sy~tem has a GE-115 Central Processor with a built-in console,
and one to four core memory modules of 4, 096 characters each, availing the user of from
4, 096 characters to 16, 384 characters of 6. 5-microsecond core storage.
One printer and one card reader, each having an integrated controller, can be connected
directly to a GE-115 Central Processor. Two other peripheral devices can be connected
through the GE-100 Standard Interface. Alternately, one of the two I/o connectors can
accommodate up to 64 peripheral devices operating in overlapped mode through Multiple
Peripheral Adapters (MPA) connected to the Standard Interface.
Peripheral devices available include line printers, card readers, card punches, a paper
tape reader, magnetic tape units, and a removable disc storage unit. The peripheral
devices are summarized in Table I with their rated speeds.
Table I also indicates the configuration possibilities of each peripheral device, and the
connector or connectors which must be used to service the device. These connectors
are diagrammed in Figure 1, Configuration Selector. Table II shows the relationship
between the number of peripheral units connected to the shared-channel and the corresponding requirement for Multiple Peripheral Adapter Units (MPA's).

10/67

IA
AUERBACH

II

(Contd.)

SUMMARY REPORT

310:011. 030

TABLE I.

GE-115 PERIPHERAL DEVICES
Uses
Connector(s)
No.

Maximum
Number
in System

Model Number

Device
Card Readers

CRZI00: 300 cpm
CRZI20: 600 cpm

1

2

B

Card Punches

CPZI0l: 60/200 cpm
CPZI03: 300 cpm

67(6)

3,4

C
D

Card Reader/Punch

CRPI00: 300 cpm

67(6)

3,4

E

Line Printers

PRTI00: 300 lpm
PRTllO: 600 lpm
PRTI20: 780 lpm

1
1

67(6)

1
1
3,4

F
G
H

4(2)

3

24(3)
24(3)

-

DSCI30: Controller
DSUI30: (77. 5KC ;
2.98 million chars.
per Disc Drive Unit)

4(4)
20(5)

-

J

Punched Paper Tape
Reader

PTRI00: 500 cps

67(6)

3,4

K

Data CommunicatiOl'
Controller

DATANET-I0: 2000
or 2400 bps

4

L

Magnetic Tape Units

MTCI03/MTCI06 :
Controller
MTHI03: 30 KC
MTHI06: 60 KC

Disc Storage Units

(1)
(2)
(3)
(4)
(5)
(6)

4 Controllers
(single buffer)

-

I

3

Only one peripheral unit can be serviced by Connector No.4.
Up to 4 controllers can be connected, using MPA115 Multiple Peripheral
Adapter units.
Up to 6 Tape Handlers can be serviced by one controller.
System! is limited to 4 Disc Controllers using MPA channel.
Each Disc Controller can handle 5 Disc Drive Units.
Theoretically limited to 67; practical system configilration considerations
will reduce this number considerably.

TABLE il: MULTIPLE PERIPHERAL ADAPTER (MPA) REQUIREMENTS
NUMBER OF MPA'S AS A FUNCTION OF NUMBER OF PERIPHERAI.
SUBSYSTEMS CONNECTED TO CONNECTOR NO.3
NO. OF
PERIPHERAL
SUBSYSTEMS

NO.
OF
MPA'S

NO. OF
PERIPHERAL
SUBSYSTEMS

NO.
OF
MPA'S

0-1

-

23 - 25

8

44 - 46

15

2 - 4

1

26 - 28

9

47 - 49

16

5 -

2

29 - 31

10

50 - 52

17

8 - 10

3

32 - 34

11

53 - 55

18

11 - 13

4

35 - 37

12

56 - 58

19

14 - 16

5

38 - 40

13

59 - 61

20

17 -19

6

41 - 43

14

62 - 64

21

20 - 22

7

7

NO. OF
PERIPHERAL
SUBSYSTEMS

C 1967 AUERBACH Corporation and AUERBACH Info, Inc.

NO.
OF
MPA'S

10/67

GE-tl5

310:011.031

(Choose one) (Chose one)

"'©

®~

(Ch&08e one)
Figure 2: Configuration Selector
• 03

SYSTEM CONFIGURATION (Contd.)

A typical configuration that could be used to replace a unit record accounting machine is
presented in Paragraph. 031 - AUERBACH standard Configuration I. :Paragraph. 032
illustrates a GE-1l5 configuration suitable for use as a remote terminal for a GE-400
or GE-600 Series computer system. Paragraphs. 033 through. 036 show GE-1l5
systems arranged according to AUERBACH Standard Configurations II, m, mR, and
IVR•
. 031 Typical Card System: Standard Configuration I
Rental
Equipment
$750
1 - GE-1l5 Model I Central Processor, with 8, 192 characters of
core storage
PRT120 Printer; 780 Ipm

945

1 - CRZ120 Card Reader; 600 cpm

315

1-

605

1 - CPZI03 Card Punch; 300 cpm
Total Rental:

$2,615

Note: the price for a GE-115 card system that is similar to Configuration I but which
uses a lower-speed CRZI00 Card Reader (300 cpm), PRTIOO printer (300. lpm) , and
CPZIOI Card Punch (65-200 cpm), rents for $1,655 per month. Use of the smallest
available core storage unit of 4, 096 characters reduces the system rental still further
to $1,475 per month .
. 032 Typical Remote Terminal System
Equipment

Rental

1 - GE-115 Central Processor with 4, 096 characters of
core storage
1 - PRTI00 Printer; 300 lpm
1 - CRZI00 Card Reader; 300 cpm
1 - Datanet-10 Communications Terminal*
Total Rental:

$570
450
140
210
$1,370

*Does not include cost of the necessary digital subset.
Note: A DSU130 Removable Disc Storage Unit can be added to provide 2, 980, 000
characters of on-line random-access storage. Total rental of the above system with
one DSU130 unit (one disc handler and controller) would be $1, 985 .
. 033 4-Tape Business System: Configuration II
Equipment

Rental

1 - GE-115 Central Processor, with 8, 192 characters of

1 1
1
1
1

-

core storage
PRTllO Printer; 600 Ipm
CRZ120 Card Reader; 600 cpm
CPZIOI Card Punch; 60-200 cpm
MCTI03 Magnetic Tape Controller
MTHI03 Magnetic Tape Handlers; 30KC

$750
650
315
315
475
1,360

Total Rental:

$3,865
(Contd.)

'10/67

.

AUERBACH

SUMMARY REPORT

310:011.034

.034 6-Tape Business System: Configuration

m

Same as Configuration n. except that 2 more MTHI03
Magnetic Tape Handlers are added.
Total Rental:

$4.545

.035 5-Million Byte RanrJomAccess System: Configuration IIIR
Equipment

Rental

] - GE-115 Central Processor. with 16. 384-characters of
core storage
1 - PRTll 0 Printer; 600 lpm
1 - CRZ120 Card Reader; 600 cpm
1 - CPZ10.L Card Punch; 60-200 cpm
1 - DSC130 Disc <.;ontroller
3 - DSU130 Disc Drive Units

$1.425

Total Rental:
.036 20-Million Byte Random Access System: Configuration IVR
Equipmpnt
1 - GE-1l5 Central Processor, with 16,384 characters of
core storage
1 - PRTllO Printer; 600 lpm
1 - CRZ120 Card Reader; 600 cpm
1 - CPZI01 Card Punch; 60-200 cpm
2 - DSC130 Disc Controller
10 - DSU130 Disc Drive Units
1 - MTC106 Magnetic Tape Controller
4 - MTHI06 Magnetic Tape Handlers; 60 kc
1 - MPA1l5 Multiple Peripheral Adapter
Total Rental:

650
315
315
315
900
$3.920

Rental
$1,425
650
315
315
630

3,000
475
1,980
75
$8,865

Note: A minimum system similar to Configuration IVR above, but with 8,192
characters of core storage and no tape units, rents for $5, 635 per month.
According to GE, this minimum configuration would be adequate to accommodate
their low-level disc-resident software •
. 04 INTF.nNAL STORAGE
. 041 Core Storage
Magnetic core storage for the GE-115 Processor is available in storage capacities of
4, 096. 8. 192, 12, 288, and 16, 384 alphameric characters. Core storage cycle time is
6.5 microseconds per character. Each character position is individually addressable
and consists of eight data bits and one parity bit. The maximum effective internal
transfer rate is 46, 090 characters (or 92, 180 packed decimal digits) per second, and
the peak nata transfer rate is 154, ooe characters per second•
. 042 nsS1!lo Disc Storage Subsystem
The OSS130 Disc Storage Subsystem recently added to the GE-115 product line consists
of from one to four DCS130 Disc Controllers, and from one to five DSU130 DIsc Drive
Units per controller for a total system capacity of 20 DIsc Drive UnIts.
The DSU130 Disc Drive Unit currently being marketed for use with the GE-115 system
is the Control Data 852 Disc Storage Drive. This random access device provides a
minimum on-line data capacity of 2 million alphameric characters, and access times
which vary from 30 to 145 milliseconds. The unit uses the same data recording mode
as the IBM 1311 Disc Storage Drive, and their respective Disc Packs are functionally
interchangeable.
The DSU130 Disc Drive Unit holds a single removable DIsc Pack that co~ 6 discs.
Ten of the 12 available disc surfaces are used for recording data. There are 100 data
tracks on each disc surface, yielding a total storage capacity of 2. 0 or 2.98 million
alphameric characters in the Sector and Track modes, respectively.
Each DSU130 Disc Drive Unit is serviced by a single comb-like access mechanism that
moves horizontally between the disc surfaces. Each of the ten access arms that make
up the access mechanism contains a single dual-gap read-write head to service all 100
tracks of one disc surface. The ten data tracks that can be accessed when the tenarmed access mechanism is in any given position are referred to as a "cylinder." The
total number of characters that can be stored per cylinder and accessed by electronic
switching (without access arm positiOning) is from 20,000 to 29,800 alphameric
characters.
C 1967 AUERBACH Corporation and AUERBACH Info, Inc.

10/67

GE-115

310:011. 042

.042 DSS 130 Disc Storage Subsystem (Contd.)
When access arm positioning is required to read or record on a selected track, the
access time ranges from 30 to 145 milliseconds. The rotational delay, or the time
required for the addressed record to pass under the read-write head once the proper
track has been selected, varies from 0 to 40 milliseconds. The total update cycle
time to read a randomly-addressed 100-character record, update it, and perform a
write-check operation, is 159 milliseconds.
The DSU130 Disc Drive unit can store up to 2 million 6-bit (Plus parity) characters
when recording is performed in the Sector Mode. In this mode, each track is divided
into 20 sectors, and each sector holds a 5-character address and up to 100 six-bit
alphameric characters of data. When data is recorded in the Full-Track Mode (i. e. ,
with each record occupying a full data track), each track can hold 2, 980 six-bit
characters, for a total drive capacity of 2. 98 million characters. The DSU130 Disk
Drive Unit records data in the NRZI (Non-Return to Zero) data recording mode. A
parity bit is generated and recorded with each character of data.
The DSU130 can also operate with an 8-bit character format. In this case, each
sector holds a five-character address and up to 75 8-bit alphameric characters, or
150 4-bit numeric digits. This feature is of particular value when handling numeric
data, since it increases the disc storage capacity by 50 percent. A summary of the
characteristics of the DSU130 Disc Drive is provided in Table m below.
The DSC130 Disc Controller can serve up to five DSU130 Disc Drive Units. A maximum GE-115 mass storage configuration can include four Disc Controllers and a
total of 20 Disc Drive Units. Thus, in the Track and Sector modes, respectively, an
on-line storage capacity of 40. 0 and 59.6 million alphameric characters is provided•
. 05

CENTRAL PROCESSOR
The GE-115 Central Processor is basically a character-oriented, variable-word-Iength,
two-address, sequential processor. All addressing is in the binary mode and is directi
1. e., no indexing or other automatic address modification facilities are provided.
The basic instruction format is:
Part:

F

C

IA

IB

Size in bits:

8

8

16

16

Reduced formats of two or four 8-bit characters are used for some instructions which
require no reference, or only one reference, to memory. The operation code is
represented by F; the high-order two bits of this word specify the format of the instruction. The C character can represent an operand length for logical instructions
(up to 256 characters), the length of two operands for arithmetic instructions (up to lEI
digits each), an 8-bit literal, an I/o device specification, or the complement of the
TABLE

m.

CHARACTERISTICS OF GE DSU130 DISC DRIVE UNITS
MODEL NUMBER

DSU130

Storage Capacity per Pack (millions
of characters)
Discs per Pack
Recording Surfaces per Pack
Tracks per Disc Surface
Sectors per Track
Characters per Sector
Characters Stored per Track*
Disc Rotation Speed (rpm)
Rotational Delay (msec)
Access Time with Direct Seek
(msec)
Data Rate (char/sec)

2. 0 (Sector Mode)
2.98 (Track Mode)
6
10
100
20
100
2,000 (Sector Mode)
2, 980(Track Mode)
1,500
o to 40
30 to 145
77,730

* In 6-bit character format.
10/67

A

(Contd.)

AUERBACH

co

310:011. 050

SUMMARY REPORT

.05

CENTRAL PROCESSOR (Contd.)
operation code. depending upon the particular instruction. The 16-bit fields IA and lB.
when present. represent the addresses of the operands.
A total of 39 instructions prvvide facilities for decimal addition and subtraction, binary
addition and subtraction, decimal and binary comparison, editing. branching based
upon the status of indicators set by compare operations, and the Boolean operations
Inclusive OR, AND. and Exclusive OR. Literal operands can be used only in a onecharacter store and a one-character compare operation.
Several interesting and potentially valuable instructions are included in the GE-115
repertoire. Among these are the Transcode instruction for translating between any
two 8-bit codes; the Pack and Unpack instructions for converting decimal data between
the two-digits-per-Iocatlon packed format and the one-digit-per-Iocation format required for arithmetic instructions; and search instructions for locating a speCified
character within a field. Note that all decimal arithmetic instructions operate on unsigned fields. A subroutine is required to obtain the conventional algebraic type of
arithmetic operations. The only interrupt facility is the capability for recognizing a
request from a DATANET-10 terminal.
Probable execution times for decimal arithmetic are as follows, where B represents
the operand length in 8-bit characters and D represents the operand length in decimal
digits. Note that these times are for unsigned fields; additional time must be allowed if
signed. algebraic-type operations are desired.
Time. Microseconds
For random addresses 78.0 + 27. 6D.
c = a +b: • • • • • • . • • • . . • . . . . • . . • . . • • . . • . . .
39.0 + 14. 6D.
b = a + b: • . . . • • • • • • . • . • . . • • • . . • . . . • • . • . .
(39. 0 + 14.6D) N.
Sum N Items: . . . . . . . . . . . . . . . . . . . . . . . . . . • .
c =ab: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
*
c = alb: ............................... .
For arrays of data 302 + 32D.
c. = a. + b.: . . . . . • . . . . . . . . . . . . . . . • . . . . . . .

..

1
1
J
b j = ai +bj :

............................ .

Sum N items: . . . . . . . . . . . . . . . . . . . . . . . . . . . .
c = c + a.b.: . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1

J

190 + 16D.
(176 + 16D) N.
'(328 + 32D) N.
39 + 13B.

Moving data: . . . . . . . . . . . . . . . . . . . . . . . . . . . .
*Subroutines are required for multiplication and diviSion;
typical subroutine times supplied by GE are as follows:
5 digit x 5 digit multiply - 4. 380 microseconds.
10 digit ... 5 digit division - 12. 500 microseconds .
. 051 Compatibility
There is no direct program compatibility between the GE-115 and any of GEls other
computer systems (the 200 Series. 400 Series and the 600 Series). Neither is there
any direct compatibility with the IBM System/360. However. through use of the
Transcode instruction. data files on punched cards and paper tape from almost any
other system can be utilized. At the source language level. GE states that GE-115
COBOL will be a compatible subset of the COBOL-61 language implemented in the
GE 400 Series, with many additional features specified by DOD COBOL 65 .
. 06 CONSOLE
A control panel built into the central processor cabinet provides the switches. keys,
and lights required for manual control of the system. No provision for keyboard
inPut or console typewriter output has been announced to date .
• 07 PUNCHED CARD AND PAPER TAPE INPUT-OUTPUT

J

.071 CRZI00 Card Reader
Developed by GE (USA). the CRZ100 is a low-cost device that can read standard 80column cards punched in either Hollerith code or Column Binary code (core image. two
decimal digits per column); intermixed reading of the two codes is also allowed. Card
reading. which takes place column-by-column in a serial fashion. occurs at a 300-cpm
rate. as the cards are fed past a bank of twelve photoelectric cells.
The single inPut hopper smd single output stacker each hold 500 cards. and can be
loaded and unloaded during normal machine operation.
The reader can operate simultaneously with other peripheral devices when in the batch
mode. or with data transmission when in the remote terminal mode. Demand imposed
upon the processor by the CRZ100 at peak reading speed is 55 percent.
Error f.i1dications include malfunctioning of the photocells. column misread. input hopper
empty. output stacker full. end-of-file mark sensed. and card jam

o

1967 AUERBACH Corporation and AUERBACH Info. Inc.

10/67

310:011.072

GE-lIS

· 072 CRZ120 Card Reader
This unit, developed by Olivetti-GE, reads standard SO-column (or, optionally, 51column) punched cards at a peak speed of 600 cards per minute. The effective speed will
normally be very close to the peak speed bec,ause the unit has an infinite clutch, so a
complete cycle is not lost when the processitlg time exceeds the time available between
cards.
A 2000-card input hopper and primary output stacker are provided, along with an
auxiliary output stacker having a 500-card capacity. All can be loaded and unloaded
during normal machine operation. Selection of either the primary or the auxiliary
stackers is under program control. Column-binary or Hollerith code can be read, and
may be intermixed within a single run. Error conditions sensed are identical to those
in the CRZ100 Card Reader.
The card reader can operate concurrently with any other peripheral device connected to
the other data channel. However, the only time available for internal processing is the
time between card cycles. This time varies with the number of characters read from
the card and is a minimum of 27 milliseconds per card when all SO columns are being
read at the peak 600-cpm speed.
· 073 CRP100 Card Reader/Punch
With all reading and punching operations done in a row-by-row manner, the fullybuffered CRP100 performs card reading, punching, or intermixed reading and punchi~
operations at the constant rate of 300 cards per minute upon standard Hollerith or
column-binary coded SO-column cards.
Cards to be read and/or-punched are loaded into a 300-card input hopper. Up to SOO
blank cards are loaded into a secondary input hopper for punching. Three output
stackers are provided - a 3000-card normal stacker, an SOO-card select stacker, and
a 600-card reject stacker. Routing of cards to either the normal or reject stackers is
program-controlled. All hoppers and/or stackers are accessible for loading or unloading purposes during normal machine operations.
Checking features include read after punching (echo check), read brushes functioning,
input hopper(s) empty, card jam, memory-to-buffer and buffer-to-memory parity
errors, output stackers full, end-of-file, and off-line malfunction.
A fully-buffered controller permits complete overlapping of the reader/punch's
operations with central processing and with other I/O operations .
. 074 CPZIOI Card Punch
This unit, developed by Bull-GE, punches standard SO-column cards serially by column
at 100 columns per second. The peak punching speed varies from 60 cards per minute
when punching SO columns per card to a maximum of 200 cards per minute when punching
a restricted number of card columns. The CPZI01 is equipped with one 1500-card hopper
and one 1500-card stacker. Cards can be loaded or removed at any time while the punch
is operating. Cards can be punched in either Hollerith or card-image code format. A
fully-buffered controller overlaps the card punch operations with internal computing and
with other I/o operations.
Checks are made to ensure correct card feed and synchronization, proper parity, and
accurate punching (echo check). Also sensed are card jam, stacker full, chad box full,
and hopper empty conditions.
· 075 CPZ103 Card Punch
This is a 300-cpm constant speed device that punches SO-column cards row-by-row in
either Hollerith or 12-column binary format. The 1200-card hopper and stacker are
supplemented by an auxiliary program-selectable 100-card output stacker. Cards can
be loaded or removed at any time while the punch is in operation. A fully buffered
controller provides overlap of the punching operations with computing and with other
I/O operations.
Detectp.ble error conditions include card feed, card synchronization, parity, card jam,
stacker full, chad box full, and hopper empty. An echo check is used to insure accurate
punching .
. 076 PTRIOO Paper Tape Reader
This unit, _developed by Olivetti-GE, is capable of reading up to 500 characters-persecond. If the unit must stop after each character and await a command to proceed to
the next character (asynchronous), the reading rate falls to a maximum of 200 charactersper-second.
Rewinding proceeds at 775 characters-per-second under console control, or 500 charactersper-second under program control. Thus, the S-inch, lOOO-foot reel can be reviewed in
either 2.7 minutes (console oontrol) or 4.0 minutes (program control). Reels are selfloading for convenience.
10/67

A

AUERBACH

•

(Contd.)

SUMMARY REPORT

310:011. 076

.076 PTRIOO Paper Tape Reader (Contd.)
standard I-inch, 7- or 8-channel tape, or Telex 1l/16-inch, 5-channel tape can be read
undirectionally or bi -directionally (with an optional feature).
Parity checking, using the parity check channel, and end-of-reel-checks are performed.
• 08 PRINTERS
.081 PRTIlO Line Printer
The PRTIlO Printer is an asynchronous line printer developed by OIivetti-GE. Skipping is
initiated immediately following the last printed character of a line. Some of the more
important characteristics of this printer are:
•
•
•
•
..

104, 120, or 136 printing positions.
10 characters-per-inch horizontal spaciDll:.
6 lines-per-inch vertical spacing.
64 printable characters (GE standard character set).
17.0 inches-per-second continuous skipping speed.

The PRTIlO will accept continuous forms from 3 to 22 inches in width. An optional..
feature permits forms skipping at 63 inches per second.
The maximum printing rate utilizing the full 64-character set is 600 single-spaced lines per
minute.
· OS':! PRT100 Line Printer
The PRTIOO Printer is a slowed-down version of the PRTIlO Printer described in the
previous paragraph. The maximum printing rate of the PRTlOO is 300 single-spaced
!ines per minute when using the full 64-character set. Other characteristics of the
PRT100 Printer are similar to those of the PRTIlO, except that the high-speed skip
option is not available for the PRT100.
· 083 PRT120 Buffered Printer
The newly-announced PRT120 Printer is a fully-buffered device capable of printing up to
780 lines per minute using a maximum of 48 of the 64 available print characters. "The
peak speed drops to 620 lines per minute when the full character set is used. Multipleline paper advances can take place at 63 inches per second using a fast-skip optional
feature. The PRT120 Printer with 120 print pOSitions rents for $945 per month; addition
of the l36-print position feature adds $105 to the basic monthly rental.
The PRT120 will accept continuous forms from 3.5 to 22 inches in width. Up to five
carbon copies can be made, and ledger and card stock can also be handled .
. 09

MAGNETIC TAPE

.091 MTH103 and MTH106 Magnetic Tape Handlers
Two tape units are available for use with GE-115 systems. The two units are designated
the MTHI03 and MTH106 Magnetic Tape Handlers. Both are nine-track units (with seventrack options), fully compatible with the IBM 2400 Series Magnetic Tape Units used with
the IBM System/360. The MTHI03 unit operates at 37. 5 inches per second, records at
either 200, 556 or 800 rows per inch, and has a peak data transfer rate of 30, 000 bytes
per second. The basically similar MTHI06 unit operates at 75 inches per second,
developing a peak data transfer rate of 60, 000 bytes per second. Both units can read in
the forward and backward directions. Table IV summarizes the important characteristics
of each tape unit.
TABLE IV"

Model

MTH103

MTHI06

CHARACTERISTICS OF THE MTHI03 AND MTHI06 MAGNETIC TAPE HANDLERS

No. of
Tracks

Tape
Speed,
inches
per sec

Tape
Transfer
Density,
Rate,
rows per chars per
sec
sec

Jnter- Record
Gap
Compatibility
Inches

msec

Rewind
Time,
inches
per sec

7
(option)

37.5

200
556
800

7,500
21,000
30,000

0.75

20

GE-400
IBM 729
IBM 7330

75

9

37.5

800

30,000

0.60

16

IBM 2400

75

7
(option)

75

200
556
800

15,000
42,000
60,000

0.75

10

GE-400
IBM 729
IBM 7330

150

9

75

800

60,000

0.60

8

IBM 2400

150

C 1967 AUERBACH Corporation and AUERBACH Info, Inc.

10/67

GE-115

310:011.091

.091 MTH103 and MTH106 Magnetic Tape Handlers (Contd.)
Up to six tape units can be connected to an MTCI03 or MTCI06 Magnetic Tape Controller,
and up to four Controllers can be connected to a GE-115 central processor, providing
a maximum of 24 magnetic tape units per system.
· 10

OTHER INPUT-OUTPUT EQUIPMENT

.101 SLC100 DATANET-10 Communications Controller
This device enables a GE-115 system to be connected to a remote DATANET 30, GE-400
Series, GE-600 Series, another GE-115, or any other manufacturer's computer equipment that uses USASCII data codes via a second DATANET (any model) at the remote site
and a dial-up or private-line communication circuit.
The DATANET-10 allows the GE-1l5 to be connected to the Bell System DSS 201A on a
2, OOO-baud circuit. Use of a Bell System DSS 201B data set allows communications via
2, 400-baud private-line circuits.
Typical transmission rates between a GE-400 or GE-600 Series computer system and the
GE-115 are shown in Table V. These rates are based on record lengths of 80 characters
per card or 120 characters per print line. Reduced record lengths can increase the
transmission rates up to the peak rates of the individual peripheral devices.
TABLE V.

TYPICAL GE-115 REMOTE TERMINAL PERFORMANCE

Peripheral Device

Rate

Card Reader (any model)
CPZ101 Card Punch
Printer (any model)

125 cards/min
85 cards/min
95 lines/min

The SLCIOO receives and transmits 7-bit characters (plus odd parity bit) in a serial
synchronous mode. Parity is checked as each character is received or transmitted, and
a longitudinal parity check is performed upon each message block. Standby status and
automatic disconnect features are provided which can be initiated if data is not received
in the SLC100 buffer within a specified time interval. Synchronization of data transmission timing with processor timing is controlled by a buffer within the controller.
· 11

SIMULTANEOUS OPERATIONS
The GE-115 Central Processor has two data channels and four outlets for connecting
peripheral devices (see Figure 1 earlier in the report). Under program control, the
data channels can be switched to service different outlets. Data Channel 1 can service
outlets 1 and 2, Data Channel 2 can service outlets 2, 3, and 4. Only a printer can be
connected to outlet 1; only a card reader can be connected to outlet 2. One peripheral
device with controller can be connected to outlet 3 and one to outlet 4 through the GE-IOO
Standard Interface. Alternatively, a communications device can be connected directly to
outlet 4, and a total of up to 64 peripheral devices with controllers can be attached to outlet
3 via Multiple Peripheral Adapters.
Data transfers on both channels can take place concurrently through time-sharing of the
core storage accesses required by each peripheral device. The processor, however, is
locked out during every peripheral operation from the initiation of the data transfer until
all data for that operation has been transferred. Thus, the time between card columns
is not available for internal processing, but the time between successive cards and
successive print lines is available. In general, the processor delay is dependent upon the
number of characters transferred in a peripheral operation (see Table II).
The MPA115 Multiple Peripheral Adapter (or "Channel Expander") enables four peripheral
controllers to be connected to one outlet. Each outlet of the Adapter can be similarly
expanded, and up to 21 levels of Adapters can be cascaded in this manner. Thus, up
to 64 peripheral controllers can be connected to outlet 3. Each controller is addressed
individually. Two peripheral devices connected to the same outlet via Adapters cannot
transfer data simultaneously.
Table VI summarizes the delays imposed upon central processor operations by most of
the GE-115 input-output devices.

· 15

SOFTWARE
The GE-115 system's software consists of language processors, program generators.
input-output control systems, operating systems, utility programs, application
packages, and data communications software support. There are three levels of
support: basic card support, tape system software, and disc-oriented software. Table
VII summarizes the major software packages offered and indicates the amount of core
storage required to use each system.

10/67

fA.
AUERBACH

(Contd. )

SUMMARY REPORT

310:011. 151

TABLE VI.

PROCESSOR DELAYS DURING I/O OPERATIONS·

Function

Device

Peak
Speed

Cycle Time,
msec

Maximum
Processor
Delay, msec

Card Reading

CRZ100
CRZ120
CRP100

300 cpm
600 cpm
300 cpm

200
100
200

108
80
3.6

Card Punching

CPZ101
CPZ103
CRP100

100 col/sec
300 cpm
300 cpm

300(min.) t
200
200

2.4
2.4
2.4

Printing

PRT100
PRT110
PRT120

300lpm
600 lpm
780lpm

200
100
77

160
80
1

Paper Tape
Reading

PTR100

500 cps

t

?

t Varies with number of characters read, punched, or written.
• Processor delays during magnetic tape and disc storage operations have not
been specified by GE to date.
TABLE VII. CORE STORAGE REQUIREMENTS FOR GE-115 SOFTWARE

Core
storage
Size

4K

16K
.151 Card-Oriented Software
The GE-1l5 boasts an impressive array of card system software elements. With the
exception of the matrix inversion routine, all operate in a minimum 4K character core
storage environment. Included in the card-level support are utility routines, such as
LIST AND SUMMARIZE, REPRODUCE AND GANG PUNCH, MEMORY DUMP, etc.; a
basic assembler (BAPS); powerful file-management program generators (LOGEL I and
LOGEL 2); an input-output control system (BIOS); a card management system; application packages such as CPM; and data communications programs (GERTS/115, 115/115,
DATANET 30/115, etc.). Brief descriptions of each component follow.
List and Summarize Program Generator - a very basic language processor designed to
create reports from a file of punched cards. It also offers the option of punching the
speCific report-generating program, prepared during any run, for future use as an
independent program. List and Summarize is controlled by using up to five types of
"parameter cards, " which direct the operations and also define the file format. The
program. which can be executed in a single pass. reads and prints up to nine fields;
summarizes up to six of these fields on up to three levels; maintains page overflow
control; provides page numbering; prints variable numbers on report headings, page
headings, control headings, and detail and control totals; algebraically accumulates
@

1967 AUERBACH Corporation and AUERBACH Info, Inc.

10/67

GE-115

310:011. 151

.151 Card-Oriented Software (Contd.)
input field values conditionally or unconditionally to one or more summary fields;
provides page subtotals on any or all summary fields; carries page subtotals to the
following page; punches summary cards; prepares a user-specified number of reports
from a single input file; and lists and punches the object program.
Machine requirements at the minimum level include: 4K characters of core storage,
one card reader and one printer. A card punch is required if object program or summary
and punching is desired.
Reproduce and Gang Punch - a set of specification statements and a program developed
for card-oriented systems. This package analyzes the specifications coded by the use·r
and processes an input card file in order to: reproduce a deck of punched cards; provide
a transfer or suppression of selected fields; reproduce detail cards with gang punching
of fields contained in master cards; calculate a product or percentage from a field in
each detail card and a field in a master card, and punch the result; gangpunch a serial
number or a constant into some or all cards; and produce from one to nine copies of
each card punched.
Required hardware includes: at least 4K characters of core memory, one card reader,
one card punch, and one printer.
APS Basic Card Assembler - APS, which stands for Assembly Programming System,
is a basic card-oriented language processor (sometimes referred to as BAPS) which is
capable of translating source programs written in assembly language into object
programs. An Assembly language equivalent is provided for all GE-115 program
instructions. In addition, supplementary mnemonics are included to extend the range
of conditional jump instructions, and pseudo mnemonics are provided to control core
storage allocation and program listing formats.
Symbolic names can be used to represent all core storage addresses and data constants.
Constants can be coded in decimal, alphameric, or hexadecimal form. BAPS accepts
all Basic Input/Output subroutines (BIOS), and provides programming error diagnostics,
source and object program listings, and a listing of all symbolic names used in the
program.
Basic APS requires the use of at least 4K characters of core storage, one card reader,
one card punch, and one printer.
LOGEL 1 and LOGEL 2 - logic-generating language processors that provide the user
with a simple file-management card system that is adaptable to many data processing
applications, including (but not limited to) report generation. In general, LOqEL is
a medium-level language and compiler that provides the flexibility and power of an
assembly with the convenience and sophistication of a higher-level language processor,
like COBOL.
LOGEL 1 provides a basic program-generating capability for use with a card system.
LOGEL 2 is very similar but has added features which permit faster and easier
programming.
LOGEL is divided into five divisions; GENERAL, which defines the input and output
files; INPUT, which defines the input records; DATA, which specifies the data fields,
constants, and core storage areas; CALCULATION, which contains the instructions
to perform the required operations; and FORMAT, which specifies the format for the
print lines and the output records.
As an example of the convenience and ease of programming using LOGEL, consider
the following multiplication of Ax B = C, where A and B are signed numbers of 5 and
3 digits, respectively.
APS Assembler
LOGEL
MVQ 251 (16), A
MULN ABC
MVC 251 (1), A + 4
MVQ 219 (16), B
MVC 219 (1), B + 2
SUB YMULS
MVQ C, 234 (8)
MVC C, 7 (1), 234
LOGEL also permits the inclusion of assembler instructions intermixed with LOGEL
statements, enabling the programmer to take advantage of the basic features of the
assembler program. Another feature of LOGEL is that it can reduce the need for
flowcharting the problem program.
The information on the five LOGEL coding forms is punched into source cards and
translated by the LOGEL compiler into an intermediate level program in basic APS
language. This deck is then translated. by the basic APS assembler into an executable
machine language object program.
10/67

A...

AUERBACH

(Contd.)

SUMMARY REPORT

310:011. 151

.151 Card-Oriented Software (Contd.)
The LOGEL program generates sets of routines and establishes the logical linking of
these routines in accordance with the data and instructions contained in the source
program. Each programming job that is to be produced by LOGEL is structured within
the five LOGEL divisions, regardless of the type of application.
The logical structure of LOGEL 1 and LOGEL 2 is the same. The main differences are
that LOGEL 1 can accept one input file and produce two output files, while LOGEL 2
can accept two input files and produce three output files. LOGEL 1 contains 28
instructions and LOGEL 2 has 58 instructions. LOGEL 2 provides arithmetic conversion
and operations with constants, whereas LOGEL 1 does not. Programs written in
LOGEL 1 can be compiled by a LOGEL 2 compiler, with only minor changes required
prior to compilation.
The minimum system configurations required to compile a LOGEL lor LOGEL 2
program are shown in Table VIII below. Note that these are minimum compilation
requirements and are not intended to indicate restrictions on system size or unit types
which may be required for actual object program execution.
Basic In ut/Out ut Subroutines BIOS - a set of subroutines developed to control basic
input output operations using card readers, printers, card punches, card reader/
punches, and paper tape readers. All subroutines are supplied on cards. They are
utilized by including appropriate calling sequences in the source program. The cards
containing the called subroutines are placed at the end of the source program deck.
When the source program is assembled, the subroutines are integrated into the object
program by the Basic APS processor.
Card Management System (CM) - a set of utility routines and subroutines that perform
functions usually associated with operating systems, such as card program loading,
memory to printer dump, memory to card dump, program listing. card reproduction,
and program patch tracing.
Critical Path Method Program (CPM) - a powerful management technique for the
planning, scheduling, and controlling of business, industrial. and scientific projects,
based on a model showing the critical path network of all events contained in the project.
The CPM applications package consists of a set of specification statements and a CPM
program. The program analyzes the specifications coded by the user and produces
reports designed to simplify project evaluation and review. Up to 350 nodes (events)
can be handled in a minimum 4K character core storage configuration. Use of 8K
character core memory allows up to 1,350 nodes to be processed. One card reader.
one card punch, and one printer are also required for use of the CPM program.
Matrix Inversion Program (MIP) - allows the user to invert a square matrix of Nth
order, in which each operand is expressed in floating-point format. using the GaussJordan method. Systems with at least 8, 192 positions of core storage are required to
use MIP.
Data Communications Program (GERTS) - a package that allows a GE-115 serving as
a remote terminal to transmit user programs and data to the central computer for
processing, and to print or punch results received from the central computer's
processing runs.
A GE-115 to GE-115 communication package is also available that allows one GE-115
serving as a remote terminal to read and transmit punched card data to a second
GE-115 processor, and to print or punch data received from the other GE-1l5
,processor. A Datanet-lO Communication Controller, a card reader. and printer are
, required to use this system.
T ABLE VIII:

MINIMUM LOGEL MACHINE CONFIGURATIONS

Equipment
GE -115 Core Storage (chars)

LOGEL 1
4K

LOGEL 2
8K

Card Reader

CRZI00 or
CRZ120

CRZIOO or
CRZ120 or
CRPIOO

Card Punch

CPZIOlor
CPZI03

CPZIOlor
CPZI03 or
CRPIOO

Printer

PRTIOO or
PRT110

PRTIOO or
PRTllO

© 1967 AUERBACH Corporation and AUERBACH Info. Inc.

10/67

GE-115

310:011. 152

. 152 Tape/Disc-Oriented Software
Magnetic tape and disc oriented software support for the GE-115 includes a small
Tape Operating System (TOS), a Disc Operating System (DOS), a COBOL compiler,
an extended Automatic Programming System (APS) assembler, a Tape Sort/Merge
Program, magnetic tape IOCS routines, an improved version of the card-oriented
LOGEL (Logic Generating Language) compiler, and programs for performing CPM,
matrix inversion, and linear programming operations.
All magnetic tape-oriented software except the COBOL compiler can function with a
minimum hardware configuration that includes a GE-115 processor with 8K characters
of core storage, three magnetic tape handlers, and one card reader and printer. Use
of the COBOL compiler requires a minimum system of 12K characters of core storage,
four magnetic tape units (or two disc storage units), and one card reader and printer.
All software can utilize the full 16K characters of core storage available in the
largest GE-115 processor.
Tape Operating System (TOS) - scheduled for delivery in December 1967, TOS will
permit staCked-job processing using system control cards. No form of TOS multiprogramming has been announced to date. TOS can load user programs, perform
the processing operations, terminate the jobs, and call in the next job.
In addition, TOS will provide memory and' tape dumps, generate data samples, prepare
system tapes, perform tape prints, update system library tapes, and generate
debugging routines.
Disc Operating System (DOS.)., -also S,cheduled .for delivery-lin becember 1967, the
Disc Operating System for the GE-115 is functionally identical to TOS. OOS requires
the use of one disc subsystem with two disc drive units in place of the three tape
handlers required by TOS.
COBOL Compiler - functioning under control of TOS or DOS, the GE-115 eOBOL
compiler language is a subset of COBOL 65, and includes arithmetic, logical. and
decision functions, input-output statements. editing capabilities, report writer statements, program segmentation provisions, the ability to include APS routines within
the COBOL source program, and a provision for requesting object program listings,
Scheduled availability of the GE-115 TOS COBOL compiler is March 1968. OOS COBOL
is scheduled to be released one month later.
Extended APS - an expanded version of Basic APS, and designed to run under
supervision of the Tape Operating System (TOS). The additional features provided by
Extended APS include: a complete set of signed and unsigned arithmetic macroinstructions, a larger set of input-output macros, and the ability to subdivide programs
into segments and to overlay these segments.
The arithmetic and the input-output macro-instructions allow faster and more efficient
coding than does Basic APS. The ability to segment a program and to overlay segments
permits the programmer to use considerably less storage than he normally would require.
Source programs written in Basic APS can be assembled using the Extended APS
language processor.
LOGEL 3 - due for completion in February 1968. LOGEL 3 is an extension of LOGEL
1 and 2 previously described for the card-oriented systems. It is run under supervision of either the Tape Operating System or the Disc Operating System and permits
the user to write file management programs that utilize up to three input files 'and
three output files, four consultation tables, and one printed report.
Tape and Disc Sort/Merge Generator - analyzes specification statements completed
by the user in order to generate a program capable of sorting and/or merging the
user's data files. Input data files can be read from either punched cards, magnetic
tape, or discs (under DOS). The sorted and merged output files produced by the
program are on magnetic tape (TOS) or discs (OOS). The sort method used is the
polyphase technique, and exits are provided within the program that allow the user to
branch to routines he has written during execution of the program. All tape sort/
merge operations are performed in ascending order, using from three to eight tape
units.
The time required to sort 10, 000 100-character records, using 8, 192 characters of
core storage and three 30KC tape handlers, is 37 minutes. Substituting 60KC tape
units lowers the processing time to 27 minutes.
Extended Input-Output Subroutines (EIOS) - a set of macro-instructions developed for
use with either the Tape or Disc Operating System. These macros are inserted in
source programs written in the extended assembly language. Upon encountering such
a macro-instruction in a source program, the extended assembly language processor
generates the object code necessary to perform the following tape or disc input-output
operations, such as: read a record from tape or disc (GET), write records (PUT) , seek a
disc record (SEEK), update a disc record (PUTX), delete a disc record (DEL), skip all
records in the current block of records (RLS), write a checkpoint on tape (CHKP),
write a checkpOint on disc (DCKP), open a tape or disc file (OPEN), close a disc or
tape file (CLOSE), and change a tape or disc volume (NEXT).
10/67

A.

AUERBACH

310:221.101

.u.....
BDP
.......

GE·115
PRICE DATA

GE·115
IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

Name

PRICES
MO:1thly
Monthly
Rental Purchase Maint.

$

$

$

Processing Unit (includes core storage)

PROCESSOR

GE-115 Central Processor (4K characters)
GE-115 Central Processor (8K characters)
GE-115 Central Processor (12K characters)
GE-1l5 Central Processor (16K characters)
Multiple Peripheral Adapter

115A04
115A08
115A12
115A16
MPA115

570
750
1,000
1,300
75

25,440
33,320
46,160
63,600
3,320

40
56
80
100
6

300

11,910

52

315
20
590
450
50

13,780
560
25,510
19,872
2,200

320
450
480
450

12,390
20,070
18,200
20,070

82
57
116
57

55

2,500

3

140
230
50
315
40
20
315
605
590

5,390
9,180
2,016
12,240
1,540
680
11,670
22,420
21,510

30
45
60
8
6
72
138
142

120

4,520

26

450

16,800

100

40
75
650

1,540
2,790
24,240

8
17
145

50
95
75
945

1,880
3,510
2,790
35,000

11
22
17
216

Disc Storage

MASS
STORAGE
DSU130
DSC130
DCT160
DSU160
DSC161
OPT005

Removable Disc storage Unit
(2 million characters)
Removable Disc Storage Controller
Removable Disc Cartridge for DSUlSO
Removable Disc Storage Unit
Removable Disc Storage Controller
115 Accelerator

28

-

70
36
5

Magnetic Tape

INPUTOUTPUT
MTHI0S
MTCI03
MTHI06
MTC106
OPT007

Magnetic Tape Handler (9-track, 30kc)
Magnetic Tape Controller (7/9-track, SOkc)
Magnetic Tape Handler (9-track, 60kc)
Magnetic Tape Controller (single channel,
7/9-track)
7 Track Option for MTH103 or
MTHI06
Punched Card

CRZIOO
CRZl11
OPT024
CRZ120
OPT025
OPT026
CPZIOI
CPZI03
CRPIOO

Card Reader (300 cards/min)
Card Reader (400 cards/ min)
51-Column Card Option
Card Reader (600 cards/min, 2 stackers)
51-Column Card Option
Transcoder Bypass Option
Card Punch (60-200 cards/min.)
Card Punch (300 cards/min, 2 stackers)
Card Reader/Punch (300 cards/min.)

8

Paper Tape
PTRIOO

Paper Tape Reader (500 char/sec)
Printers

PRT100
OPT075
OPT076
PRTllO
OPT077
OPT078
OPT079
PRT120

Printer and Control (300 lines/min,
104 columns)
120 Column Option
136 Column Option
Printer and Control (600 lines/min,
104 columns)
120 Column Option
136 Column Option
Fast Skip Option
Printer and Control (780 lines/min.
120 columns)

© 1969 AUERBACH Corporation and AUERBACH Info,lnc.

2/69

GE·llS

310:221,102

PRICES

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

INPUTOUTPUT

Name

Monthly
Monthly
Rental Purchase Maint.

$

$

$

PI'intl'l':" (Conte\.)

(Contd. )

OPT085
OPT086

136 Column Option
Fast Skip Option

105
75

3,890
2,790

24
17

210

9,320

16

300

12,000

48

Channels

C 0 l\1l\111I\'l-

CATIONS

8LCI00
(CLIlOO)
8LCI02

2/69

DAT ANET-I0 Communications
Controller, 8)nchronous (2000 or 2400
bits/sec)
DATANET-12 Single Line Communications
Controller (19,200,40,800, or 50,000
bits/sec)

fA...

AUERBACH

311 :000.000

£

'i ..... '

~EDP

AUERBAC~

GE-130

REPORTS

REPORT UPDATE

~

REPORT UPDATE
~GE

ANNOUNCES NEW ADDITION TO GE-I00 SERIES

General Electric recently announced the GE-120 system, the fourth member of the expanding
GE-I00 family of small-to-medium scale computers.
The GE-120 is designed with monolithic integrated circuitry (as is true of all members of the
GE -100 family) and has a memory cycle time of four microseconds and core storage capacities
of 12,288, 16,384, or 24,576 bytes. This places the GE -120 between the smaller, slower GE115 (with a memory cycle time of 6. 5 microseconds and a maximum memory capacity of 16, 384)
and the larger, faster GE-130 (with a memory cycle time of two microseconds and a maximum
storage of 32, 768 bytes). All of the GE-I00 Series computers are upwards compatible in that
the data structure, instruction repertoires, and supporting software are either common or a
subset of the larger-scale system.
The GE-120 is available in both magnetic tape and disc configurations. A typical tape configuration rents for $2, 980 and sells for $128,600: a disc system carries a monthly lease of
$3,310 and a purchase price of $143, 702. Estimated delivery time following the receipt of an
order is six months.
The GE-120 has a repertoire of 63 instructions, eight index registers, the capability of overlapping processing with I/O operations, a program interrupt and communication capabilities. A
full line of peripherals are available including card readers and punches, magnetic and paper tape
handlers, line printers, disc storage units, and communication controllers and terminal devices.
A new Report Program Generator, GE-I00 RPG, is now available with all members of the GE100 Series, facilitating compatibility with competitive equipment. In addition to the RPG, the GE120 system provides extensive software support including tape and disc operating systems, an assembly programming system, COBOL 65, FORTRAN IV, and sort-merge generators.

@

1969 AUERBACH CorporatIon and AUERBACH Info, Inc.

8/69

£.

311:011.100
II......

~EDP

AUER8AC~

GE-130

IE'••TS

•

ADVANCE REPORT

ADVANCE REPORT: GE-130
•1

INTRODUCTION
The GE-130, announced in May 1968, represents a significant enhancement of the earlier
GE-115 (Report 310) and the establishment of a GE-100 line of compatible systems. A
GE-115 system can be upgraded to a GE-130 by simply changing the central processor. The
GE-130 offers a larger memory capacity, faster storage cycle time, index registers, a
larger instruction repertoire, greater overlapping of input-output operations with internal
processing, more efficient software, a wider range of peripheral units, and a one-level
interrupt capability for data communications.
Developed by General Electric Information System of Italy (GEISl) in collaboration with
General Electric (USA), the joint marketing effort for the GE-130 is directed toward upgrading GE-1l5 installations, as well as toward offering users of other systems a more
sophisticated replacement for second-generation equipment. General Electric regards the
new GE-130 system as in direct competition with the mM System/360 Model 25/30, the
Honeywell 120/125, the RCA Spectra 70/25, and the UNIVAC 9300 systems. The GE-1l5
will still be offered as a lower-priced system to satisfy the needs of users with smaller
data processing requirements.
The primary objective in designing the GE-130 has been to maximize performance in typical
business data processing applications while maintaining competitive price levels. The available operating systems currently do not support punched-card systems. Typical magnetic
tape systems range between $4, 685 and $9, 070 in monthly rental, while typical disc systems
range between $4, 270 and $9, 055. Paper tape readers and data communications controllers
are also available for expanded processing requirements. Core storage capacities range
from 16,384 to 32,768 eight-bit locations.
Two operating systems, oriented toward magnetic tape and disc systems, provide executive
monitoring, input-output control, and debugging routines to control the system's operations
and reduce the need for operator intervention. Language processors include an Assembly
Programming System (APS), COBOL 65, and FORTRAN IV. In addition to the basic software, a number of application programs and mathematical and statistical subroutines are
available. All of the software that is operational for the GE-115 can be used with suitablyequipped GE-130 systems.

•2

The first customer deliveries of GE-130 systems are scheduled for April 1969; the first
DSS161 Disc Storage Subsystem will be delivered in October 1969. The new CRZ111 Card
Reader is currently available, as are all of the peripheral devices used with the GE-115
system. GE is currently promising deliveries 8 months following the receipt of an order •
HARDWARE

• 21

Data Structure

• 22

The basic unit of data storage in the GE-130 is an "octet" consisting of eight bits plus a
parity bit. Thus, a GE octet is equivalent to an IBM byte. Each octet is addressable and
can hold data represented in alphanumeric, packed numeric, unpacked numeric, or binary
form. Maximum arithmetic operand lengths are 16 digits in the unpacked numeric mode,
31 digits in the packed numeric mode, and 16 octets in the binary mode.
Unpacked numeric digits contain the numeric value in four low-order bits of each octet
with a constant binary configuration in the four high-order "zone" bits. In an unpacked
numeric field, the binary value of the zone quartet of the right-most octet represents the
sign.
In the packed numeric mode, each octet contains two 4-bit digits, leading to more efficient
allocation of data in core storage and on magnetic tape and disc storage, as well as increased transfer rates and a reduction in computing time. The sign of a signed packed
numeric field is located in the low-order 4 bits of the rightmost octet •
System Configuration
Every GE-130 system includes a central processor with a built-in Operator Control Panel
and 16,384, 24,576, or 32,768 octets of 2-microsecond core storage.
The central processor has three I/O channels and four peripheral connectors. The appropriate connections between the central processor and all peripheral subsystems are established under program control, within certain limitations, via the three channels and four
connectors.
@ 1968 AUERBACH Corporation and AUERBACH Info, Inc.

6/68

311:011.220

GE-130

Figure 1. A typical GE-130 configuration•
• 22

System Configuration (Contd.)
Integrated controllers for a 300-lpm unbuffered printer and a card reader are permanently
associated with Connectors 1 and 2, respectively. Connectors 3 and 4 can be linked to
either single or multiple peripheral subsystems. The maximum number of peripheral
subsystems that can be used in a GE-130 system is 34.
Table I lists the peripheral subsystems available for the GE-130 and shows which of the
connectors and channels each subsystem can utilize •

• 221 GE-130 6-Tape Business System; Configuration m
Equipment

Rental

1111161-

$2,000
230
315
650
450
1,920
75
$5,640

130A16 Processor with 16, 384 octets of core storage
CRZll1 Card Reader; 400 cpm
CPZ101 Card Punch; 60-200 cpm
PRTll0 Printer; 600 lpm
MTC103 Magnetic Tape Controller
MTH103 Magnetic Tape Handlers; 9-track, 60KC
MPA130 Multiple Peripheral Adapter
Total Rental:

.222 GE-130 6-Tape Auxiliary Storage System; Configuration V

6/68

Equipment

Rental

1 - 130A16 Processor with 16,384 octets of core storage
1 - CRZ111 Card Reader; 400 cpm
1 - CPZ101 Card Punch; 60-200 cpm
1 - PRT100 Printer; 300 lpm
1 - MTC103 Magnetic Tape Controller
6 - MTH103 Magnetic Tape Handlers; 9-track, 60KC
1 - DSC161 Disc Storage Controller
3 - DSU160 Disc Storage Units; 7.6 million char.
1 - MPA130 Multiple Peripheral Adapter

$2,000
230
315
450
450
1,920
450
1,770
75
$7,660

A•

A.UERBACH

Total Rental:

(Contd.)

ADVANCE REPORT

311 JOI1. 230

TABLE I: GE-130 PERIPHERAL SUBSYSTEMS
Description

Model No.

Connectors

Channels

CARD READERS
CRZ111
CRZ120

400 cards per minute
600 cards per lI}.inute

2
2

1-2-3
1-2-3

3-4

1-3

3-4

1-3

1
1
3-4

2
2
1-3

3-4

1-3

CARD PUNCH
CPZ101

100 columns per second
CARD READER-PUNCH

CRP100

300 cards per minute
PRINTERS

PRT100
PRTll 0
PRT120

300 lines per minute
600 lines per minute
780 lines per minute
PAPER TAPE READER

PRTI00

500 characters per second
MAGNETIC TAPE SUBSYSTEMS

MTC103

Controller for up to six MTHI03 7/9-track
tape drives, up to 30 KC

3-4

1-3

MTCI06

Controller for up to six MTHI06 7/9-track
tape drives, up to 60 KC

3-4

1-3

3-4

1-3

3-4
3-4

1-3
1-3

MAGNETIC DISC SUBSYSTEMS
O8S161

DSC161 controller for up to eight DSU160
disc drives
LINE CONTROLLERS

SLC100
SLC102
• 23

DATANET-10: 2,000/2,400 BPS
DATANET-12: 19,200/40,800/50,000 BPS

Internal Storage

.231 Core Storage
The working storage in the GE-130 consists of 9-plane ferrite core stacks. Each grouping
of nine bits (including parity bit) constitutes one octet location. Storage capacities of 16,384,
24,576, and 32,768 octets are available. An odd-parity check is performed on all data stored
in or read from core. The maximum internal data transfer rate is 500,000 octets per second,
reflecting the cycle time of 2 microseconds - a significant improvement over the 6.5-microsecond cycle time of the GE-115.
Each storage location is directly addressable and holds an octet consisting of eight data bits
plus one parity bit. Address modification is provided by index registers. No indirect addressing facility is available. Special-purpose areas are provided in lower memory for peripheral unit referencing, a multiplication and division work area, eight index registers, and program interrupt flags .
• 232 Disc Units
The DSS161 Removable' Disc Storage Subsystem, consisting of a DSC161 Disc Controller and
from one to eight DSU160 Disc Units, provides a flexible, removable-cartridge mass storage
system that is suitable for both batch and random-access processing.
The DSU160 Disc Units allow removable disc packs to be mounted and dismounted, thereby
providing external data storage with an enormously expanded storage capacity. The DCT160
Removable Disc Cartridge (disc pack) has six interchangeable discs with ten recording
surfaces. The GE disc pack is IBM 1316-compatible, although the GE recording format
differs from IBM's. The disc pack, with a diameter of 14 inches and a height of 4 inches,
© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

6/68

311:011.232

.232

GE 139

Disc Units (Contd.)
has a total of 2000 working tracks divided into 20,000 individually addressable sectors.
Each track has 10 sectors capable of storing 576 numeric characters (288 octets or 384
alphanumeric characters). Table II specifies the capacities of the disc storage organizational
subdivisions •
The DSU160 Disc Unit contains five positiOning arms with ten "floating" read-write heads,
all of which move in unison. The average positioning time to access one cylinder is 75
milliseconds, with an average latency time of 12.5 milliseconds.
Data, recorded in either 6-bit alphanumeric or 4-bit packed numeric form, is transferred
at a rate of 312,000 numeric or 208,000 alphanumeric characters per second. A parity
check is performed on each character transferred.
TABLE II: DSU160 DISC STORAGE CAPACITIES
Location
Sector
10-Sector Track
10-Track Cylinder
200-Track SUrface
2000-Track Disc Pack
8-Disc-Pack SUbsystem

.24

Numeric
Characters

Alphanumeric
Characters

576
5,760
57,600
1,152,000
11,520,000
92,160,000

384
3,840
38,400
768,000
7,680,000
61,440,000

Octets
288
,2,880
28,800
576,000
5,760,000
46,080,000

Central Progessor
The GE-130 Central Processor is functionally organized in the same manner as the GE-115,
with character orientation, variable word length, and sequential processing. The component
parts include: a magnetic core memory for program and data storage, a command and control unit for program execution, an arithmetic control unit for computation and initiation of
input-output operations, a peripheral control unit for channel switching and direct control
of the card reader and line printer, an operator control panel, and connectors for peripheral
subsystems •

• 241 Instruction Format
The basic GE-130 instruction format includes lengths of 2, 4 and 6 octets. Zero, one or
two memory addresses may be specified. The three types are represented as follows:

I Two-Address
lOne-Address

I
I

OC
OC

I CC
I cc

EIA

E

IA or I

.1..

IB---:03

=+j

I No-Address
Where OC
CC

CC
OC
I-octet operation code, specifying the arithmetic, logical,
data movement, input-output, etc., operation to be executed.

I-octet complementary code, modifying the basic operation code to
define the specific conditions of execution. Complementary codes
can define effective field lengths (minus 1), program transfer conditions, constant values, peripheral unit names, or index register
references.

lA' IB = 2-octet addresses of first and second operands.
I

2-octet address of the next instruction to be executed.

Address fields, in binary form, always occupy two octets. The high-order bit specifies
absolute or relative addressing. When relative addressing is indicated, the binary value
of the index registcr to be used is located in the next three high-order bits. This leaves
the 12 low-order bits to express the relative address, with a binary addressing capacity
of 4096 octets. The effectiYe storage address is calculated by adding thc relative address
contained in the instruction to the base address in the specified index register. Relative
addressing facilitates the repeated use of the same instruction, without altering the instruction itSelf, through changing the contents of the index register.
.242 Arithmetic Control Unit
The arithmE1tic control unit executes the fundamental operations of the GE-130 system. The
unit's functions in,clude: arithmetic operations on signed packed numbers, unsigned unpacked
numbers, and binary fields; Boolean operations; data comparisons; table searches; internal
data transfers; packing and unpacking of numeric data; editing of output data, and data transfers between the central processor and peripheral devices.
6/68

A

AUfRBACf<

(Contd.)

ADVANCE REPORT

311:011.242

.242 Arithmetic Control Unit (Contd.)
The arithmetic control unit can also enter qualitative results obtained from arithmetic or
logical operations into the status indicators (underflow/overflow and zero/non-zero). The
values in these indicators can subsequently be tested by conditional transfer instr1,lctions •
• 243 Peripheral Control Unit
The peripheral control unit determines, in response to program signals, the appropriate
connections between the peripheral subsystems and the central processor. Three channels
and four connectors are provided for this purpose. The peripheral control unit also contains
integrated controllers for a 300-lpm or 600-lpm unbuffered printer and a card reader.
The input-output channels receive orders from the central processor, check the availability
of the peripheral units, transfer signals and data between the central processor and peripheral units, and regulate accesses to the central processor from the peripheral units.
Data is transferred in parallel, one octet (eight bits plus parity bit) at a time.
An interleaving technique is employed to overlap the core storage accesses of several
concurrently-operating periph\lral units. Core storage cycles are divided among the three
channels, servicing requests for central processor time in the order in which they are received and giving priority to Channell, which is also used by the arithmeti9 control unit
for internal processing.
The maximum total I/O data rate of the GE-130 system is 380,000 octets per second.
Channels I and 3 can each accommodate data transfer rates of up to 220,000 octets per
second, while Channel 2 is limited to the data rate of the printer or card reader connected
to the integrated control units.
Maximum simultaneity in a GE-130 system will normally be achieved in one of two situations:
•

Overlapping of three I/O operations, two of which may involve magnetic
tape and/or discs, without internal processing; or
•
Overlapping of two I/O operations with internal processing.
In either case, additional I/O operations on buffered peripheral units can overlap other
peripheral operations and internal processing. Buffered I/O units available for the GE-130
include the card punches, the CRPIOO Card Reader/Punch, and the 780-lpm PRTl20 Printer •
• 244 Program Status Register
The Program Status Register (PSR) in the central processor occupies 4 octets and contains
a "picture" of the current program status. The PSR contains three elem£lnts: the current
instruction address, an interrupt mask, and the status indicators.
The instruction address increases progressively as a program is executed and is changed
in response to a program jump or a DATANET (data communications) interrupt. The
interrupt mask allows the programmer to specify whether interrupts shall be accepted or
rejected. The status indicators reflect the qualitiative results of certain arithmetic or
logical operations: overflow/underflow and zero/non-zero conditions •
• 245 Command and Control Unit
The Command and Control Unit directs program execution by performing the following
functions:
•
Updates the Program Status Register.
•
Extracts the program instructions from memory and transfers them to the
Instruction Register.
•
•

Analyzes each instruction for the type of operation involved.
Processes instruction operands and addresses, as well as status indicators
and interrupt signals.

•

Controls the actual execution of the instructions by the arithmetic control
unit.
•
Communicates the program status to the operator by means of console lights,
and permits operator intervention•
• 246 Interrupt Processing
Unless interrupt servicing is inhibited by the interrupt mask in the PSR, interrupt signals
from DATANET line controllers have immediate access to the central processor and the
system allows real-time control of the remote terminals.
The follOwing elements are part of the interrupt facility:
•
Detection of interrupt signals on connectors 3 & 4.
•
Modification of the Program Status Register (PSR) in the command and
control unit.
© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

6/68

311:011.246

GE-I30

.246 Interrupt Processing (Contd.)
•

Provision of two reserved storage areas: the Old Program Status Register
(OPSR) stores the status of an interrupted prdgram to permit future reinitiation, while the Interrupt Program Status Register (IPSR) contains the
data for servicing the interrupt condition.

At installation time, connectors 3 & 4 are enabled to receive interrupt sigJJals.
The programmer can disable this facility by changing the Progr~ Status Register interrupt
mask. The command and control unit tests for the presence of an interrupt signal. When
detected, and the connectors are enabled, the current program status is saved and the interrupt program status is activated. When the interrupt request has been satisfied, the original
program status is re-established .
• 247 Processing Facilities
The GE-130 system has a repertoire of sixty-three instructions, as listed in Table m.
Facilities are provided for: decimal and binary arithmetic operations; comparisons and
data movement; editing of output data; packing and unpacking of numeric fields; transl~tion
between any two B-bit codes; memory searches; conditional branches on internal status
indicators and external sense switches; Boolean operations for bit manipulation and testing;"
subroutine linkage; console-light alert control; and input-output operations.
Addition and subtraction operations can be performed in the packed decimal, unpacked
decimal, and binary modes. Multiplication and division operations, which are available
only through software routines in the GE-115, are standard hardware instructions in the
GE-130, though only in the packed decimal mode.
The input-output instructions transfer data to and from core storage and the p~ripheral units,
perform auxiliary peripheral operations such as line spacing and tape rewinding, testing for
parity errors, and testing the availability of channels, connectors, or peripheral units.
Control fields- specifying the type of operation and the execution procedure are established
in core memory and referenced by the I/O instructions.
Table IV summarizes the performance of the GE-130 Central Processor •
• 24B Operator Control Panel

Manual control of the GE-130 system is provided by an Operator Control Panel built into
the Central Processor. Switches, keys, and lights allow the operator to set external program switches, execute the program one step at a time, clear the system to initial conditio~,
display the operation code and addresses of the next instruction to be executed, and display
the contents of the interrupt mask and the status indicators •
• 25

Input-Output Equipment
General Electric offers a fairly wide choice of peripheral subsystems to support the GE-130
as listed in Table I. All of the devices available for the GE-130 system can also be used
with the smaller GE-115.
The new Disc Storage Subsystem is described in Paragraph. 232 of this Advance Report,
while the new CRZ111 Card Reader and the augmented data communications facilities are
described in the following paragraphs. For details on all the other peripheral subsystems,
please turn to Paragraphs. 071 through. 091 of the GE-115 Summary, Report 310 .

• 251 CRZll1 Card Reader
The CRZ111 Card Reader is a new 400-cpm photoelectric reader designed especially for
the GE-100 system. It is identical to the 600-cpm CRZ120 Card Reader in all respects
except for its lower speed. Standard BO-column, 12-row punched cards are read in columnby-column fashion by 12 photocells. Data in either Hollerith or column-binary (two BCD
characters per column) code can be read, and the two codes can be intermixed within a
single run. The input hopper holds ~WOO cards. The primary 2000-card-capacity output
stacker and the auxiliary 500-card output stacker can be unloaded while the reader is running. Output stackers can be selected under program control.
.252 Data Communications
The DATANET-10 and DATANET-12 Communications Controllers enable the GE-130 system
to be connected to a remote DATANET-30, a GE-115, a GE-400 or GE-600 series system,
another GE-130, or another manufacturer's computer system. Data in 7-bit (plus odd
parity bit) character format is transmitted and received in a serial synchronous mode. A
"trans coding" table in the Central Processor translates the 7-bit codes. The DATANET-10
allows the GE-130 to be connected to a Bell System data set through a half-duplex 2,000 or
2,400 bps circuit. The DATANET-12 operates over a faster TELPAK 19,200, 40, BOO, or
50,000 bps circuit.

6/68

A

(Contd. )

AUERBACH
@

311:011. 252

ADVANCE REPORT

TABLE

m:

INSTRUCTION LIST FOR THE GE-130

Arithmetic Operations

Data Transfers

Add decimal
Subtract decimal
Add decimal packed
Subtract decimal packed
Multiply decimal packed
Divide decimal packed
Add binary
Subtract binary
Add memory to register
Subtract memory from register

Move complete octets
Move right quartets
Move packed
Move immediate octet to store
load register
Store register
Load address
load program status register

Comparisons
Compare complete octets
Compare right quartets
Compare packed
Compare memory to immediate
Compare register to memory
Search to the right
Search to the left
Logical Operations
A}lTD on complete octets
OR on complete octets
Exclusive OR on complete octets
AND on immediate
OR on immediate
Exclusive OR on immediate
Test under mask
Branch Instructions
9 conditional jumps
Jump and return
Jump if Switch 1 set
Jump if Switch 2 set

Reformatting Instructions
Pack
Unpack
Pack with sign
Unpack with sign
Translate to octets
Edit
Input-Output General Instructions
Call peripheral
Call peripheral indirect
Transfer peripheral
Command peripheral
Examine peripheral
Set peripheral
Test peripheral
Miscellaneous Instructions
Halt system operation
No operation
Turn alert light on
Turn alert light off
Inhibit single-step
Enable single-step

.252 Data Communications (Contd.)
Each character transmitted or received is checked for parity, and each message block
undergoes a longitudinal parity check. Interrupt signals can be transmitted to the GE-130
for program interruption if the interrupt mask in the Program Status Register is enabled.
A four-word buffer synchronizes data transmission timing with processor timing. The
Communications Controllers are automatically disconnected if no data is received in the
DATANET-I0 or DATANET-12 buffer within 60 seconds •
•3

SOFTWARE
Upward software compatibility from the GE-115 to the GE-130 perrr.its the utilization of all
programs written for the lower-level corrputer. Two operatine; systems designed specifically for the GE-130 system, EXTENDED TAPE OPERATING SYSTEM (ETOS) and EXTENDED DISC OPERATING SYSTEM (EOOS), are oriented toward magnetic tape and disc systems,
respectively. In addition, language processors, program generators, input-output control
systems, utility programs, application packages, and data comrr.unications software are
provided.
General Electric states that the aims of its operating systems include the full exploitation of
the operative capacities of the electronic system and the reduction of system start-up and
management costs. Programming is done on the logical input-output level, without dependence on the type of peripheral units in the system. Compiling and test execution, as well
as the system production programs, are automatically linked.
The minimum configuration requirements common to both the Extended Tape Operating
System (ETOS) and the Extended Disc Operating System (EOOS) are 16K octets of core
storage, any GE-I00 series card reader, and any GE-I00 series line printer. ETOS also
requires an MTCI03 Magnetic Tape Controller with four MTHI03 or MTHI06 Magnetic Tape
Handlers. A DSC161 Disc Controller with two DSU160 Disc Units completes the minimum
configuration for EOOS.
@

1968 AUERBACH Corporation and AUERBACH Info. Inc.

6/68

GE-I30

311;011.310
TABLE IV: PERFORMANCE OF THE GE-130 CENTRAL PROCESSOR
Execution Time
in Microseconds

Task
Unpacked Decimal Mode {I)
c =a +b
b =a +b
c =a xb
c =a+ b
Move a to b
Compare a to b

78
44
11=
11=

34
44

Packed Decimal Mode {2)
c -a +b
b =a +b
c=axb
c =a+b

62
34

548
864

Binary Mode (3)
c -a +b
b =a +b
c =a xb
c =a+ b

11=

68
38
11=
11=

Hardware facility not available.

(1) Based on unsigned 5-octet fields.
(2) Based on signed 5-digit (i. e .• 3-ootet) fields.
(3) Based on 32-bit (i. e .• 4-octet) binary fields .
• 31

System Operation
The System Tape (ETOS) or Disc (EOOS) is used for compiling, testing, and debugging of
the user's programs. It includes the entire software system used to prepare the Library
and Master Tapes or discs. The Library Tape (ETOS) or Disc (EOOS) contains all of the
user's 'programs which are already debugged and operational. The Master Tape (ETOS) or
Disc (EOOS) relates tb a given work period (i. e., one day's work) and contains all of the
system and user programs required to accomplish a given task.

• 32

The normal operational scheme involves off-line updating of the System Tape or Disc from
the Library Tape or Disc, and then the_ updating of the Master Tape or Disc from both the
System and Library Tape or Disc. Under a production run, it is the Master Tape or Disc
that contains the linked program flow for orderly system task execution •
Supervisor
The ETOS and EDOS Supervisor is a central management program that coordinates all of the
software elements for the automatic control of a complex of generator, service, compiler,
and debugging programs. The Supervisor initializes the system and loads all of the system
programs from magnetic tape or disc •

• 321 Extended Input-Output System
The Extended Input-Output System (EIOS) controls, on a logical level, the handling of inputoutput files. Such functions as file blocking and deblocking, I/O error analysis, end-of-file,
and multi-volume processing are performed. The Supervisor, in turn, initiates and controls
the input-output activity on the physical level •
• 322 Service Routines
The Supervisor is supported in program control by service routines that provide for automatic
linkage of pz:ograms for a production run, program segmentation, and generation and management of the program library. Other service routines handle operator-machine communication
control •
• 323 Debugging Routines
A package of debugging routines enables the operating systems to prepare the user's programs
_for execution through memory-to-print and memory-to-card dumps, program listings, and
patch tracing.

6/68

A

(Contd.)

AUERB~CH

'"

ADVANCE REPORT

311 :011. 324

.324 Lanugage Processors
The principal language processors for the GE-130 are the machine-oriented Assembly Programming System (APS), a COBOL compiler, and a FORTRAN IV compiler.
The Assembly Programming System (APS) is a card-oriented language processor that translates source programs written in assembly language into object programs ready for testing
or execution. APS requires a 16K GE-130 Central Processor with at least one card reader,
one card punch, and one printer. Mnemonic operation codes are provided for the entire
GE-130 instruction repertoire, with extended mnemonics for the conditional jump instructions and pseudo-mnemonics to control core storage allocation and program listing formats.
EIOS macros and other standard routines can be called by APS for insertion into the assembled object deck.
For magnetic tape and disc-oriented systems, tape-extended APS and disc-extended APS
versions are available. The extended APS versions provide the following additional features:
a complete set of signed and unsigned arithmetic instructions, a larger set of input-output
macros, and the ability to subdivide programs into segments and to overlay these segments.
GE-130 COBOL is a subset of COBOL-65, including arithmetic, logical, and decision functions, EIOS routines, editing capabilities, report writer statements, program segmentation
provisions, the ability to include APS routines within the COBOL source program, and a
provision for requesting object-program listings.
The GE-130 FORTRAN IV Compiler requires a minimum system configuration of one card
reader, one printer, 16K octets of core storage, and either four magnetic tape units (ETOS)
or two disc drives (EDOS).
The GE-130 Sort Program Generator (SPG) is designed to facilitate sort and merge applications
on magnetic tape (ETOS) and discs (EDOS). User-coded sequences in APS can be added. All
sort and merge operations are perfomed in ascending order using the polyphase technique.

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

6/68

311:221.101

A

11111'"

EDP

"""'IACH

GE-130

IU"l$

PRICE DATA

GE·130
IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

PROCESSOR

Name

PRICES

Monthly
Monthly
Rental Purchase Maint.
$
$
$

Processing Unit (includes core storage)
130A16

GE-130 Central prffiessor (16,384 octets
of core storage)
GE-130 Central Processor (24,576 octets
of core storage)(l)
GE-130 Central Processor (32,768 octets
of core storage)(1)
Multiple Peripheral Adaptor (MPA)

130A24
130A32
MPA130
MASS
STORAGE

2,000

88,000

150

2,500

111,120

185

3,000

133,440

220

75

3,312

6

450
590
315
300

19,872
25,510
13,780
11,910

36
70
28
52

20

560

0

320
450
480
450
55

12,390
20,070
18,200
20,070
2,500

82
57
116
57
3

230
315
40
20
315
605
590

9,180
12,420
1,540
680
11,680
22,420
21,510

45
60
8
6
138
142

120

4,520

26

450

16,800

100

35
70
605

1,540
2,790
24,240

8
17
145

45
90
70
900

1,880
3,510
2,790
35,000

11
22
17
216

100
70

3,890
2,790

24
17

Disc Storage
DSC161
DSU160
DSC130
DSUl30

Disc Storage Controller
Disc Storage Unit
Disc Storage Controller
Removable Disc Storage Unit (2 mUlion
characters)
Removable Disc Cartridge (for DSU130
or DSU160)

DCT160
I NPUTOUTPUT

Magnetic Tape
MTH103
MTC103
MTHl06
MTCI06
OPT007

Magnetic Tape Handler (9-track, 30 kc)
Magnetic Tape Controller (7/9-track, 30 kc)
Magnetic Tape Handler (9-track, 60 kc)
Magnetic Tape Controller (7/9-track, 60 kc)
7-Track Option for MTH103 or MTH106
Punched Card

CRZ111
CRZ120
OPT025
OPT026
CPZlOl
CPZ103
CRP100

Card Reader (400 cards/min)
Card Reader (SOO cards/min)
51-Column Card Option
Transcoder Bypass Option
Card Punch (60-200 cardS/min)
Card Punch (300 cardS/min)
Card Reader!PWlch (300 cards/min)

72

Paper Tape
PTR100

Paper Tape Reader (500 chari sec)
Printers

PRTlOO
OPT075
OPT076
PRTllO
OPT077
OPT078
OPT079
PRTl20
OPT085
OPTOS6

Printer (300 lines/min, 104 print
positions)
120 Column Option
136 Column Option
Printer (600 lines/min. 104 print
positions)
120 Column Option
136 Column Option
Fast Skip Option
Printer (780 lllles/min, 120 print
positions)
136 ColUmn Option
Fast Skip Option

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

GE-130

311:221.102

IDENTITY OF UNIT
CLASS

COMMUNICATIONS

Model
Number
SLCIOO
SLCI02

Feature
Number

Name
DAT ANET-IO Communications
Controller (2000 or 2400 bits/sec)
DAT ANET-12 Communications
Controller (19,200, 40,800 or 50,000
bits/sec)

NOTES:
(1)

2/69

Each ,,('tpt c()nsists ()f l'i!!-ht oata bits and one parity bit.

fA

AUERBACH

'"

PRICES
Monthly
Monthly
Rental Purchase Maint.
$
$
$
210

9,230

16

300

12,100

48

1.

320;011. 100
sa .....

AEDP

AUERBAC~e

GE-200 SERIES
SUMMARY

REPORTS

SUMMARY
.1

INTRODUCTION
The GE-200 Series consists of three compatible, small-to-medium-scale, secondgeneration computers. Designated the GE-215, GE-225, and GE-235, the three machines differ primarily in their processor speeds. All have identical instruction repertoires, and there is a common set of input-output units, as well as a common set of
software packages for all. The GE-200 computers are binary, word-oriented systems,
and they offer both floating-point and decimal arithmetic options. Each word consists of
20 bits and may contain a one-address instruction, a data word, or 3 alphameric characters. Mon.thly system rentals can vary from approximately $2,600 to over $40,000,
with typical configurations renting in the $5,000 to $18,000 range. Several hundred
GE-200 Series systems have been installed.
The initial member of the GE-200 Series was the GE-225, first delivered in March 1961.
The "family" was actually created in early 1963, when the GE-215 (a slowed-down
GE-225) and the GE-235 (a faster, re-engineered GE-225) were introduced. The GE-215
possesses reduced input-output capability, with fewer data channels than the GE-225.
The GE-235, with a basic cycle three times as fast as the GE-225, is further distinguished by its much faster floating-point module.
The essential differences among the three GE-200 Series systems are summarized in
Table 1.
The GE-215 is no longer being produced, but used machines can be delivered as available.
The GE-225 is officially out of production, but it is still being marketed, tied to a
DATANET-30 communications processor, as the GE-255 Time-Sharing System. The
GE-235 is still being actively produced and marketed under its own number and as the
GE-265, a time-sharing system combining a GE-235 processor with a DATANET-30.
The following sections describe the data structure, hardware characteristics, software
offerings, and intra-line compatibility of the GE-200 Series •

.2

DATA STRUCTURE
The GE-200 Series is basically word-oriented. A GE-200 Series word consists of 20 bits
plus a parity bit. Each 20-bit word can hold a one-address instruction, a 20-bit binary
data word, or 3 alphameric characters in 6-bit BCD representation. A floating-point
number is stored in two consecutive GE-200 Series words, using 30 bits plus sign for the
fraction and 8 bits plus sign for the exponent .

•3

HARDWARE

.31

Central Processors
The GE-215, GE-225, and GE-235 systems all utilize essentially the same central processor. The GE-215 processor is a slowed-down GE-225 unit; the GE-235 processor,
while re-engineered for additional speed, retains all the features of the earlier GE-225
model. All three models use identical instruction sets. The essential differences among
them are the core storage cycle times, the number of input-output channels, and the maximum I/O channel transfer rates, as listed in Table I.
TABLE I. DISTINGUISHING CHARACTERISTICS OF THE GE-200 SERIES SYSTEMS

System

GE-215
GE-225
GE-235

Core Storage
Cycle Time,
Microseconds

Max. No. of I/O
ControJlers (other
!han card and
paper tape)

36

3

18 (14 wi!h
option)
6

Maximum
Mag. Tape
Speed,
ChartSec.

Maximum
Printer
Speed,
LineS/Min.

1-0
Typewriter
Facil!ties

27,800

15,000

450

Output std.
Input opt.

8

55,600

60,000

900

Output std.
Input opt.

7

111,000

60,000

900

Output std.
Input std.

Maximum Gross

Channel Transfer
Rate, Words/Sec.

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

9/68

320:011. 310

.31

GE-200 SERIES

Central Proc..essors (Contd.)
The GE-200 Series central processors provide complete arithmetic facilities for single
word-length binary operands. Loading, storing, addition, and subtraction of doublelength binary data items can also be performed. An optional feature permits addition and
subtraction (but not multiplication or division) of single- or double-length data items in
BCD form. This feature can significantly reduce the number of time-consuming radix
conversions required in business data processing, but will seldom eliminate the problem
completely.
Three index registers and a fourth location that serves as a convenient counter register
are standard. An optional feature makes 31 additional 4-word groups in core storage
available as index registers or counters. Only one group, selected by a special instruction, can be active at a time. Other optional features for the central processor are a
Move command (which expedites internal block transfer operations). Three-Way Compare. Automatic Priority Interrupt. and a Real-Time Clock. In typical routines. instructions are executed at the rate of about 10.000 per second in the GE-215. about
20.000 in the GE-225. and about 75.000 per second in the GE-235.
An Auxiliary Arithmetic Unit can perform double-length arithmetic in either fixed or
floating-point mode under control of the central processor. This optional unit greatly
increases internal processing speeds on scientific problems .

• 32

Internal Storage
Working storage is provided by a magnetic core store. which can contain 4,096, 8, 192, or
16,384 word locations. Core storage cycle times are listed in Table I. A parity check is
performed upon all internal transfer operations.
A Disc Storage Unit (DSU) is also available. This unit consists of 4, 8, 12, or 16
permanently-mounted discs, and provides disc storage for a maximum of 18.87 million
alphameric characters in 98,304 fixed record locations of 64 words (or 192 characters)
each. The average total waiting time for access to a randomly-placed record is 225 milliseconds. Up to 294,912 characters per disc unit can be transferred without repositioning
any of the 16 access arms. A maximum of four disc file units can be connected to each
disc controller. No more than one disc controller may be connected to a GE-215 system;
the GE-225 and GE-235 systems permit eight and seven controllers, respectively •

. 33

Input-Output Devices
All peripheral devices except the console typewriter and the card and paper tape readers
and punches are connected to the central processor through a multiplexing device called
the Controller Selector. This device provides capabilities for simultaneous operations that
rival those of far more costly systems. Controllers for magnetic tape units, disc storage
units, printers, magnetic document handlers, data communication equipment, and the Auxiliary Arithmetic Unit can be connected to the Controller Selector: three on the GE-215,
eight on the GE-225, and seven on the GE-235. In addition, if the Auxiliary Arithmetic
Unit is included, it occupies one "hub" of the Controller Selector (except in the GE-235,
where it is connected directly to the Central Processor). One peripheral unit on each controller can operate simultaneously with internal processing and card or paper tape reading
and punching. Accesses to core storage are allocated automatically.
Standard 80-column punched cards can be read at 400 or 1,000 cards per minute and
punched at 100 or 300 cards per minute. Paper tape can be read at 250 or 1,000 characters per second and punched at 110 characters per second. A console typewriter provides
typed output; input via the console typewriter is standard on the GE-235 and optional on the
GE-215 and -225.
The high-speed printer available with the GE-215 has a rated speed of 450 alphameric lines
per minute. The GE-225 and -235 systems may include a 900-line-per-minute model. Both
printers have skipping speeds of 25 inches per second, and the controllers for both provide
automatic editing and format control.
Three magnetic tape handler models are available. One, the only model available with the
GE-215, has a peak data transfer rate of 15,000 characters per second at a recording density of 200 rows per inch. A second model offers a choice of 200 or 556 rows per inch,
with corresponding peak speeds of 15,000 or 42,000 characters per second. The remaining
tape handler offers a choice of 200, 556, or 800 rows per inch, with corresponding speeds
of 15,000, 42,000, or 60,000 characters per second. The 60KC model is a single-unit
handler, mounted one to a cabinet; the two slower models are mounted two to a cabinet, one
above the other. Up to eight tape handlers can be connected to each tape controller. No
more than two 42KC or 60KC tape read or write operations may occur at a time, but the
number of simultaneous 15KC tape operations is limited only by the number of tape controllers in the system.

9/68

A ..

AUERBACH

(Contd.)

SUMMARY

.33

320:011. 330

Input-Output Devices (Contd.)
Magnetically encoded paper documents can be read and sorted at a peak speed of 1, 200
documents per minute. Two document handlers can be connected to each controller, providing a peak sorting speed of 2,400 documents per minute.
GElS MOSE (Modification of Standard Equipment) group offers a variety of special-purpose
hardware for use with GE-200 systems, such as peripheral device switching controllers,
printer plotting option, plotter interface units, etc •

. 34

Data Communications Equipment
The DATANET-15 controls the transmission and reception of digital data over telephone and
telegraph lines and two-wire cables at speeds ranging from 60 to 2,400 bits per second. Up
to 15 data transmission lines and a paper tape reader and punch can be connected to a
DATANET-15, but it can control only one data transfer operation at a time.
GElS line of data communications equipment available for use with the GE-200 Series also
includes:
•
•

The DATANET-30 programmed data communication system.
The DATANET-90 magnetic-tape-to-computer terminal.

•

The DATANET-91 off-line magnetic-tape-to-magnetic-tape terminal.

•

A variety of special digital input-output terminal devices .

.4

SOFTWARE

.41

Process Oriented Languages
GECOM is offered as an all-purpose process oriented language. The basic language structure is similar to that of COBOL-61 but is not compatible with it. GECOM also handles
algebraic expressions and mathematical functions, and includes a report writer and
TABSOL, a system that permits decision logic to be expressed in a concise tabular format.
At least four magnetic tape handlers and 8,192 core storage locations are required for
GECOM compilations.
COBOL-61 is also offered for the 200 Series, implemented by a COBOL-to-GECOM translator. The translator converts COBOL source language to a source program in GECOM.
The GECOM compiler then processes this source language to produce an objectlanguage program.
FORTRAN is available in two versions: one for card-oriented systems, the other requiring four magnetic tape units. Both versions require at least 8,192 core storage locations,
and both provide several useful extens ions of USASI Bas ic FORTRAN.
WIZ is a one-pass algebraic compiler for use on punched card or paper tape systems with
at least 8,192 core storage locations. WIZ is less powerful than the FORTRAN language,
but it is easy to learn and provides high compilation speeds .

• 42

Machine Oriented Languages
The General Assembly Program (GAP) is the basic symbolic assembly system for the
GE-200 Series. It permits full utilization of the hardware facilities, is relatively easy to
learn and use, but provides few refinements. GAP-coded programs can be assembled on
GE-200 Series systems with punched card, paper tape, or magnetic tape input-output
equipment.
ZOOM is a "macro assembly system" designed to facilitate machine-oriented programming
bj reducing the amount of detailed coding required while retaining high object program
efficiencies. The ZOOM programmer uses a coll'.bination of pseudo-English statements,
algebraic expressions, and GAP symbolic statements. These are translated into an allGAP program which is then assembled in the normal manner. Magnetic tape is not required, but can be utilized to facilitate the translation process .

• 43

Problem Oriented Facilities
BRIDGE IT is a tape file maintenance and run sequencing program whose functions are
directed by control cards. FORWARD is a generalized sort/merge generator. Simulation
programs are available for simulating the operations of mM650 and Control Data
LGP-30 computers on GE-200 systems. The Card Program Generator Simplifies the
programming of existing punched card tabulator and calculator runs. An adequate library
of generalized input-output, diagnostic, and mathematical routines is available, as are
special-purpose packages for the banking and electric utility industries, numerical tool
control, inventory management, assembly line balancing, critical path method (CPM), and
information retrieval.

@

1968 AUERBACH Corporation and AUERBACH Info, Inc.

9/68

GE-200 SERIES

.320;011.500

.5

COMPATIBIIlTY WITHIN THE SERIES

Since the GE-215, GE-225, and GE-235 all use the same instruction set and essentially the
same central processor, there is, in effect, complete interchangeability of programs
among the three systems. The obvious restrictions are the following:

•
•
•

9/68

The systems must be equipped with comparable input-output units.
A program cannot call on a feature not present on the system in
question.

Since the input-output and core cycle times are different for each
of the systems, programmed timing loops must be modified or
avoided.

A

AUERBACH

•

320:221.101
GE·200 SERIES
PRICE DATA

GE·200 SERIES
IDENTITY OF UNIT
CLASS

Model
Number

F,"ature
Number

Name

PRICES
Monthly
Monthly
Rental PurchasE Maint.

$

$

$

1,200
1,300
2,500

57,600
64,900
120,000

130
150
200

~,200

2,500
3,900

72,500
87,500
132,300

130
150
200

1,900
2,500
3,900

72,500
87,500
132,300

130
150
200

2,900
3,500
4,900

72,500
87,500
132,300

130
150
200

1,930
2,530
3,930
600

116,000
140,000
184,000
24,000

130
150
200
20

650

26,000

43

225
200

4,400
3,530

7
11

75
75
75
1,400
175

3,040
2,640
2,880
44,800
8,080

5
5
7
50
13

450

10,500

31

550

12,500

37

100

2,000

6

50

1,000

3

Processing Unit (includes core storage)

PROCESSOR

GE-205 Central Processor; including console,
output typewriter, single channel adapter and:
4,096 word magnetic core memory
8,192 word magnetic core memory
16,384 word magnetic core memory
GE-215 Central Processor, including console,
output typewriter, 3-channel controller
selector. and:
4,096 word magnetic core memory
8,192 word magnetic core memory
16,384 word magnetic core memory
GE-225 Central Processor, including console,
output typewriter (no controller selector),
and:
4,096 word magnetic core memory
8,192 word magnetic core memory
16,384 word magnetic core memory
GE-225 Central Processor, including console,
output typewriter. 8 channel controller
selector, and:
4,096 word magnetic core memory
8,192 word magnetic core memory
16, 384 word magnetic core memory
GE-225 Central Processor, including console,
output typewriter, printer adapter and:
4,096 word magnetic core memory
8,192 word magnetic core memory
16,384 word magnetic core memory
Magnetic Core Memory (4,096 words)

*CA205A
*CB205A
*CC205A

*CA215A
*CB215A
*CC215A

*CA225C
*CB225D
*CC225B

*CA225B
*CB225C
*CC225A
*CA225D
*CB225E
*CC225C
*CB225L

Features for GE-205/215/225
X225A
JI04A
J225A
J225B
J225C
J225E
J225F
OPT036

Auxiliary Arithmetic Units:
Auxiliary Arithmetic Unit (for floating point,
double preCision arithmetic)
N-RegiBter Interface
Decimal Package conSisting of three way compare, additional index words, decimal add
and subtract
AutomatiC Interrupt
Move Command
Real Time Clock
Additional 8k Memory
225 Time Sharing Option
Speed-up OptiOns for GE-225

EPN209
EPN216
EPU216
EPN200

Speed-up Option (14-microsecond memory) for
GE-225 Processors with 8k memory
Speed-up Option (l4-microsecond memory) for
GE-225 Processors with 16k memory
Speed-up Kit to Upgrade from 8k to 16k
processor
*GA{i51G
*OPTfj52

Tap!' Punch (110 char/sec) with ::lpooler
1': lIlt'l' Tap(' Hcader' (1000 char/::lec) a.nd Paper
Tape 1'1I11dl (110 ('h:lI'/sec) without spooler
1'; Ill"!' 'l'ap<' I«'adcr , Punch, :md Spooler (5-,
li-, 7 , and 1l-1ev ell
i': l(ler Tape lu.'ader :md Punch with 7-8 (5-, 6-,
7-, :md H-i('vel)
P: IP('I' 'rallC Header and Spooler (1000 char/sec)
p :!pt'r Tap(' Header (1000 chari sec; without
,.;pooler)
p,ap!)l- 'rape ll.eader and Spooler (5-, fi-, 7-,
an,1 H-level)
P: llX'l' Tape Readel' (5-, 6-, 7-, and 8-level)
1'; Ipcr 'rape HeadcI' (1000 char/sec) and Paper
T,tpe Punch (11.0 char/sec) with spooler
(i)-channel)
P: Iper Tape Header (1000 char/sec) and Paper
Tape Punch (110 char/sec) without spooler
(5-ch:mnel)
p..tper Tape Punch ( 110 char/sec) G-, 7-, or
R-channel
1': Ip!'r Tape Punch (110 char/sec) 5-channel
p: tpel' Tapt' Reader with spooler (1000 chari
::lec; 5-ehannel)
Exp;u\::llOn to ij-channel code capabIlity

490

17, (;00

88

440

15,840

79

585

20, (;80

103

535

18,920

95

300
260

12,000
10,400

52

395

15,080

75

355
490

13,480
17,600

{,8
88

440

15,840

80

190

7,200

3G

190
745

7,200
28,550

36
103

99

3,080

3

795

48,3GO

246

775
1,275

48,000
49,200

246
246

1,295

49,560

246

80

3,(;00

18

80

3,600

18

550

33,000

246

600
2,950

36,000
109,800

24G
300

3,500

126,000

392

1,700

65,280

335

540
680

-

8,640
10,880
24,500

29
36
4G5

1,850

29,440

4G5

(W

p nnwr::>

-

P215C

1'215E
P225A
1'225C

PDH.225
P])H22G
*1'225D
*P22513
*PAG!lOA

*PAG9013

*l\ISLG90

P 1'lI1tOl' (120 columns, 450 lines/mm. includes
controller with F ORTRAN character
eompatllnl1ty)
l'l'lIltel' (120 ('ol'nnl IS; 450 lines/mIll)
I >rmtcr (12\l colwnru::>; !lOO lines/min; includes
controlle r)
l' l' ll1tCl', \ )n-llflc (12 o Column 900 lines/mill.)
(inelud('~ ContL'o Her with FORTRAN
charaeLcr COli tp:ttibility)
Stand:!l'rllltel' (300 lInes/!111n)
( )ff/On-Imp Printer (120 Column; 900 llnes/
mill; Printel' Re:! del' handies magnetic tape
of 200 hIts/inch density)
()IT/OIl-lllJe PI' IllLe I' (120 Column; 900 llnes/
!lun; Printer Re: lder handles magnetic tape
of 200/S55. 5 hitS/inch density)
1\1 ult q)le Tapo L l::ltp.1' and Control (900 lines/
mill)
Documcnt liandler S'ubsystem for GE-225/235

S,\225A
Sll225A
S1213
S121>

SI nglP Ch,ll1l1('\ :-'OL'W, I' :\,lapteJ'
Dual ('II,lllIld SOlil'1, Adapt.cr
(; E l\JlCIl I )O':ltnH'nt lteader/Sortcr (12 pocket,
1200 dOc'/I11I1l)
G E i\lll'l{ j)O(,UIlIPlit Hpader/Sorter (12-pocket,
1200 doe/ Hun)
-

2/69

-

A

,.

;\[01 f

(Contd. )

SUMMARY

330:011. 500

.4 ' PERIPHERAL EQUIPMENT (Contd.)
The DATANET-30 Data Communications Processor can be connected to a GE-400 Series
system to provide access to a communications network and handle simultaneous inputoutput from many remote stations. In addition, the DATANET-20 and DATANET-21
Single Line Transmission Controllers can handle data communications on a single-line
basis. In this case, an operator can dial any remote station using a digital subset, or an
Automatic Calling Unit can enable the computer to initiate the call, send data, and terminate the call.
The DATANET-760 Display Terminal (see Section 6321 of AUERBACH Data Communications
Reports) is also available for GE-400 Series systems. This device provides local or remote alphanumeric or graphical displays.
GE-115 computer systems (Report 310) can be connected, via communications links, to a GE-400
Series system. The GE-115 is a small-scale, 8-bit-character-oriented system manufactured by Olivetti-GE in Italy. Peripheral equipment available for the GE-115 includes card
readers, card punches, printers, paper tape readers and punches, and a small-capacity,
removable-cartridge disc storage unit •
•5

SOFTWARE
Four programming systems are offered for use with the GE-400 Series. A GE programming system consists of an operating system, language translators, and miscellaneous
service routines suitable for a particular purpose and generally requiring a specified hardware configuration for efficient operation.
MTPS, the Magnetic Tape Programming System, includes a single-task Basic Operating
System (BOS/MT), an Extended Operating System (EOS/MT) capable of handling up to six
additional tasks, BAL, MAP, COBOL, and two versions of FORTRAN. (The characteristics
of the GE-400 Series language processors are summarized below.) MTPS requires a
minimum of 8K words of core storage. The Card Operating System for the GE-400 Series
is loaded via cards rather than a system tape.
DPS, the Disc Programming System, includes a Basic Operating System (BOS/Disc), an
Extended Operating System (EOS/Disc) capable of controlling three additional tasks, BAL,
MAP, COBOL, RPG, and ABA FORTRAN. DPS requires a minimum of 16K words of core
storage.
TSPS, the Time-Sharing Programming System, includes TSOS (Time-Sharing Operating
System), BASIC, and ABA FORTRAN. COBOL is not currently available in the timesharing system. A GE-415 with 32K words of core storage, a DATANET-30, and the
DAP930 Direct Access Option is required for a time-sharing system. Extended Memory
modules are at present not recognized by the operating system.
In addition, the following programs and programming systems are provided by GE. Except where otherwise noted, the software is currently available:
•

The Macro Assembly Program (MAP) is the basic symbolic programming
system for the GE-400 Series. It consists of the Basic Assembly Language
(BAL), which is machine-oriented and supplies assembly-control pseudooperations, and the Macro Assembly Program language, which is fieldoriented and uses COBOL-like data descriptions and sequencing. The Macro
Assembly Program language supplies macro-instructions for communication
with the Basic and Extended Input-Output Systems, which facilitate the coding
of input and output operations. Macro-instructions for arithmetic, data movement, and procedure control operations help to minimize the amount of tedious
hand coding that must be done and reduce coding errors. At least 8,192 words
of core storage, 4 magnetic tape units, card reader, punch, and printer are
required for MAP assemblies.

•

The GE-400 Series COBOL compiler can translate source programs that
use all of Required COBOL-61 and selected elements of Elective COBOL.
Equipment required for COBOL compilation is the same as for the Macro
Assembly Program.

•

A Basic FORTRAN IV Compiler, first available in June 1965, facilitates the
programming of scientific applications. The principal restriction upon
this version of FORTRAN IV is the lack of capabilities for handling complex, logical. and double-precision operations. Equipment requirements
are the same as the Macro Assembly Program.

•

An ABA FORTRAN Compiler is fast (approximately 500 average statements compiled per minute on a GE-435) and capable of handling real,
integer, logical complex and double-precision operations. ABA FORTAN
is coded for both tape and disc systems. The Time-Sharing ASA FORTRAN
has all the capabilities of the tape and straight disc versions mentioned above
except double-precision operations. Moreover, it offers free-field I/O.
C 1968 AUERBACH Corporation and AUERBACH Info, Inc.

7/68

330:011.501

.5

GE-400 SERIES

SOFTWARE (Contd.)
•

The IBM 1401 Simulator Program and the IBM 1401 Compatibility Option
routine enable a GE-400 Series computer system to run IBM 1401 object
programs. See Paragraph. 2, "Compatibility," for descriptions of these
two simulation methods.

•

The Tape Operating System is an integrated set of three routines: the
Program Monitor, Loader, and I/O Supervisor. The Prograrr. Monitor
can speed up run-to-run changeovers in both debugging and production
operations. In the debugging function, the System Tape, which contains
all language processors and debugging aids, is used as the operating tape,
providing a "compile and run" capability. The Program Monitor can also
use a library tape of production programs as the operating tape.

•

The Card Operating System provides all the facilities of the Tape Operating
System except the System Tape; language translations cannot be performed
while using this version of the Operating System.

•

The Report Program Generator (RPG) provides for the preparation of
reports or records from files on punched cards, punched tape, or magnetic
tape. Output may be assigned to magnetic tape, printer, and/or card
punch.

•

The GE-400 Series Sort and Merge Generators produce programs for
efficient sorting and merging of magnetic tape files. User-coding options
permit pre-sort and post-sort editing.

•

Service routines for debugging, program library maintenance, media conversion, recovery for reruns, program loading, and other utility functions
are available.

One of the important strengths of the GE-400 Series is its rich selection of packaged
application programs, some of which are listed below:
IDS - Integrated Data Store
IDS/COBOL for Index Sequential Management Systems
IDS - Batch Sort
IDS - File Restructuring System
MLRD - Multiple Linear Regression
ANOV - Analysis of Variance
FACS - Flexible Accounting Control System
Transportation Package
Linear Programming Packages
Numerically-controlled Machine Tool Packages
CPM - Critical Path Method
CPM/Monitor
ASTRA - Automatic Scheduling Program
SIGMA - Status Analysis of Elastic Structural Systems
Traverse Analysis - Survey Traverses
Curved Bridge Geometry
Slope Stability Analysis System
Retaining Wall Design
Composite Beam Analysis
Horizontal Geometry
Earthwork Analysis
Transient Electrical Stability Programs
B-matrix Loss Formulas
Load Flow Programs
Electrical Fault Programs
Expandable Machine Accounting Simulator
Simeon-Scientific Inventory Management & Control
GE Parts Explosion System
Automated Costing and Estimating
Payroll Packages
Accounts Receivable
Purchasing, Receiving & Accounts Payable
Document Handler Lister
Magnetic Tape Reporting for IRS
Mortgage & Installment Loan Accounting
Bond Portfolio Analysis
Savings Account Transactions
Corrugator Scheduler Program
GE CAST - Time Series Forecasting

7/68

A

AUERBACH

'"

......

330:221.101

1. "'". .
fA,

.

AUERBAC~

EDP

GE-400 SERIES
PRICE DATA

REPORTS

GE·400 SERIES

PRICES

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

Name

Monthly
Monthly
Rental PurchasE Maint.

$

$

$

Processing Units and Core Storage(l)

PROCESSOR
*CPU405
*AMM405
*AMM405A
CPU415
AMM415
AMM416
CPU425
AMM425
AMM426
CPU427

CPU435
AMM436
CPU437

MSM416
ASM416

CPU430
CPU440

GE-405:
GE405 Central Processor (4,096-word
memory)
Additional Memory (4,096 words)
GE-415:
Additional Memory (4,096
words; second addition)
Central Processor (8, 192-word memory)
Additional Memory (8,192 words;
first addition)
Additional Memory (16,384 words;
second addition)
GE:-425:
central Processor (8,192-word memory)
Additional Memory (8,192 words;
first addition)
Additional Memory (16,384 words;
second addition)
Central Processor (for e}'F,anded-memory
GE-425 configuration)( )
GE-435:
Central Processor (16,384 word- memory)
Additional Memory (16,384 words)
Central Processor (for e~anded-memory
GE-435 configuration)( )
Main Memory for CPU427, CPU437, and
CPU440:
Main Storage Module (16,384 words;
maximum of two per CPU427, CPU437
or CPU440)
Additional Storage
Module (16,384 words; maximum of
three per MSM416)
GE-430:
GE-430 Time-Sharing Central
Processor (8,192-word memory)(3)
GE-440:
GE-440 Time-Sharing Central
Processor( 4)

1,800

75,540

158

800

34,400

52

250

9,700

48

2,135
1,250

94,620
55,400

158
92

1,250

55,400

92

3,120
1,250

138,500
55,400

231
92

1,460

64,600

107

2,990

132,900

222

5,111
1,875
5,715

230,000
73,440
254,000

420
150
424

1,870

83,120

139

4,200

186,700

312

7,000

311,200

519

200
115

8,150
4,930

23
8

315

13,850

23

Central Processor Features

*CMP405
*TDC405
CM6051

GE-405:
mM 1401 Compatibility
Time of Day Clock
Other 400 Series Processors:
1401 Compatibility Module

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

4/69

PRICE QATA

330'221.102

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

Name

PRICES
Monthly
Monthly
Rental IPurchase Maint.

$

$

$

Central Processor Features (Contd.)
PROCESSOR
(Contd. )

TC6012
DAP930

SM6010
FP6015
FP6025
FP6035

*CEF405
CE6010
HC6012
MASSSTORAGE AND
INPUTOUTPUT

SLC401

SLC403

Character, word, and doubleword channels
are available as needed without extra
charge except for channel expansion
feature for more than eight channels.
A special channel for the MICR Reader!
Sorter is available at no extra charge.
The High Speed channels are priced
separately.
Channel Expansion (GE-405)
Channel Expansion (other GE-400
series systems~
High Speed Channe (for all except
GE-405)

DATANET-20 Single Line Controller
(asynchronous)
DATANET-21 Single Line Controller
(synchronous)
DATANET-21 Single Line Controller
(synchronous)

NOTES:

(1)

5,080
13,850

8
23

68
365
470

3,000
16,160
20,770

5
27
35

575

25,390

42

80
130

3,410
5,770

6
10

260

11,540

19

290

12,230

41

315

13,340

45

390

16,670

55

The GE-400 Series computer system share
a common set of peripheral devices
with the GE-600 Series systems; these
are shown in the GE-600 Series Price
Data. The GE-405 cannot accommodate
the following peripherals:
DSU270 File Storage Unit
MDU200 Magnetic Drum Unit
MSU388 Mass storage Unit
DSU160 Disc Storage Subsystem

SLC402

*

115
315

I/o Channels

ATTACHMENTS,
ADAPTERS,
AND
CHANNELS

COMMUNICATIONS

Time of Day Clock
Direct Access Package (Includes memory
protect, interval timer, symbol
controlled move, and provision for
channel expansion)
Symbol Controlled Move
Floating Point Hardware for CPU415
Floating Point Hardware for CPU425,
CPU427 and CPU430
Floating Point Hardware for CPU435,
CPU437 and CPU440

\

No longer in production.
Field changes from one memory size to another or from one processor model to
another are normally ordered from a list of CPC options.

(2) CPU427 and CPU437 require DAP930 Direct Access Package. They use MSM416
and ASM416 memory modules. Minimum memory allowable for these two
processors is 49,152 words.
(3) GE-430 requires 32,768 words of memory. Additional memory modules for the
GE-430 are the same as those for the GE-425; i. e., AMM425 and AMM426.
(4) GE-440 require 65,536 words of memory. The CPU440 uses the MSM416
and ASM416 memory modules.

\

4/69

fA..,
AUERBACH

-1.

340:000.000
ST .... "

..-EDP

AUE~

GS-600 SERIES
REPORT UPDATE

REPlan

•

REPORT UPDATE
~

GE INTRODUCES NEW ADDITION TO GE-600 SERIES FAMILY
The recently-announced GE-615 represents GE'S latest entry in the medium-scale computer
market, expanding the options of users of the GE-600 Series. The new system is essentially a
slower and less-expensive GE-625 system.
All of the peripheral equipment available to the larger GE-600 Series systems can be used on
the GE-615. Compatability is emphasized. Like the GE 625/635 systems, the GE-615 operates
in three concurrent modes: multiprogrammed, local and remote batch proceli/sing, and timesharing. Overall system control is offered by the General Comprehensive Operating System
(G ECOS III).
The GE-615 is available in memory sizes of 65,536 to 262, 144 36-bit words in 32,768 word
increments. The memory cycle time is 2 microseconds per access of one 36-bit word, as contrasted with the 2-microsecond cycle time for a two-word fetch on the GE-625 and with the 1microsecond cycle time for a ~-word fetch on the GE-635.
A system configuration conSisting of a 65,536 word memory, 30 million characters of disk
storage, a card reader and punch, eight magnetic tapes, a 1200-line-per-minute printer, and
a console typewriter rents for approximately $30,000 per month.
The first deliveries of the GE-615 are expected in June, 1969.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

4/69

340:000.001
GE-600 SERIES
REPORT UPDATE

REPORT UPDATE

~

MULTIPLE-DRIVE DISK SUBSYSTEMS INTRODUCED
GE has recently introduced two new mass auxiliary storage subsystems for use with the GE-600
Series systems. The DSS-167 and the DSS-170 Disk Storage Subsystems are similar in design to the
IBM 2314 in that the subsystems contain multiple, removable storage disc drives. The disc packs
used with both systems have eleven-high disc stacks as contrasted with the six-high stacks in the
DSS-160 packs used with the GE-100 and GE-400 Series.
The DSS-167 provides 90 million six-bit characters of storage, expandable to 120 million; the DSS170 has a 200 million character storage capacity. Access time for both systems is 75 milliseconds,
and there is simultaneous seek overlays. The DSS-167 transfer rate is 208,000 characters per second, \which is doubled with the 416,000 character per second transfer rate of the DDS-170.
Deliveries of the new units are scheduled to begin by July 1970. The following prices will be icharged
for the two subsystems:
Monthly
Rates

Purchase
Price

Monthly
Maintenance

$3,185

$136,080

$375

185

8,230

14

DEP-167 Fata File Protect

52

2,320

4

STC-167 Stack Command

15

720

1

NC

NC

675

28,800

80

5,675

256,400

705

DSS-167 Removable Disc Storage Subsystem
90 million characters
ADC-167 Additional Data Channel (nonsimultaneous)

DCA-167 Disc Controller Adapter
ADU-167 Additional Disc Unit, 30 million characters
DSS-170 Removable Disc Storage Subsystem 200
million characters

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

5/69

......

340:011. 100

A;;~p

AUERBACH

GE-600 SERIES
SUMMARY

IIEPORTS

......-_-'~""'""--_..J

SUMMARY: GE-600 SERIES
.1

SUMMARY
The GE-600 Series represents the General Electric Company's current entry in the largescale computer field. Operating under the control of a sophisticated operating system, the
General Comprehensive Operating Supervisor (GECOS III) , the GE-625 and GE-635 systems
offer formidable competition for such systems as the IBM System/360 Models 50 and 65,
the Control Data 6400, and the UNIVAC 1108.
A "three-dimensional" approach, designed to facilitate concurrent local and remote batch
processing with time-sharing in a multiprogramming and multiprocessing environment,
underscores GE's determination to satisfy the needs of virtually any large-scale computer
application and to improve its competitive position in the industry.
The GE-625 and GE-635, announced in May 1964, share all components except core storage,
and their performance and prices are quite similar. The GE-625 operates at a cycle time
of 2 microseconds per pair of 36-bit words, while the GE-635 has a I-microsecond core
storage cycle time. A typical configuration, operating under GECOS III and including 20
magnetic tape units and random-access drum storage carries a monthly rental of $54, 275
for the GE-625 and $55,400 for the same configuration with GE-635 memory modules. Customer deliveries of the GE-600 Series systems began in April 1965.
Specialized GE-600 Series models designed for military procurement include a militarized
version of the GE-625 (the M-625) and the M-605, which is similar to the M-625 but lacks
floating-point and double-precision hardware.
In December 1965, GE announced the GE-645, a large-scale computer system specifically
designed for multiple-console, time-shared operation. The formal announcement of the
GE-645's commercial availability had been anticipated since GE's prior announcement,
months earlier, of the receipt of orders for prototype time-sharing systems from Massachusetts Institute of Technology, Bell Telephone Laboratories, and Ohio State University.
The GE-645 is still in the research and development stage and is not currently being
marketed.
The characteristics of the GE-645 hardware and the associated MULTICS (Multiplexed
Information and Qomputing .§.ervice) programming system have been the subject of considerable interest throughout the computer industry. The design of the GE-645 and its
supporting software has been largely a joint effort among the GE Computer Division, MIT,
and Bell Telephone Laboratories. The GE-645's design is based on that of the GE-635,
with several improvements to facilitate time-shared operation. The most significant
change is a different form of addressing logic that incorporates segments and pages to permit efficient dynamic reallocation of memory. Other improvements include changes in the
interrupt logic, a redesigned input-output control unit, and 2- to 4-way interleaving of
memory accesses.

Figure 1. A typical GE-625 computer system configuration.
© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

8/68

340:011. 200

.1

GE-600 SERIES

SUMMARY (Contd.)
The GE-645 uses core storage that can range in size from 32,768 to 1,048,576 36-bit words.
Core storage cycle time is 1 microsecond per pair of 36-bit words. Monthly rental prices
for GE-645 systems range from about $40,000 to $250,000, with typical systems renting
for about $90,000 per month.
An analysis of the GE-645 will be published in AUERBACH Standard EDP Reports as soon
as possible after detailed specifications are made available by GE.
The remainder of this report is devoted exclusively to the two commercially available
members of the series, the GE-625 and GE-635 .

•2

HARDWARE
A GE-625 or GE-635 computer system includes four major types of components: Processor
Module(s), Memory Module(s), Input/Output Controller Module(s), and peripheral devices.
These components, and the configuration rules for combining them, are described in the
paragraphs that follow •

• 21

Processor Module
The GE-600 Series Processor Module uses a single-address instruction format and has a
wide range of address modification capabilities, including various combinations of indexing
and indirect addressing. There are two basic modes of processor operation: master mode
and slave mode. Control programs will normally be executed in the master mode, and the
user's object programs in the slave mode. Programs running in the master mode have
access to the entire core memory, can initiate peripheral and internal control functions,
and do not have base address relocation applied. Programs running in the slave mode have
access to a limited portion of the memory (as specified by the Base Address Register),
cennot initiate peripheral control instructions, and have the contents of the Base Address
Register added to all relative memory addresses of the object program. The processor is
automatically put into the master mode of operation when the Master Mode Indicator is set
or when any interrupt is recognized. In a system having multiple Processor Modules, one is
designated the control processor. Only the control processor, operating in the master
mode, can respond to I/O interrupts.
Instructions are fetched in pairs - an even word and the successive odd word. Address
modification, operand fetching, instruction execution, and fetching of the next pair of instructions are overlapped to increase processor performance wherever possible. Indexing
does not increase the instruction execution times, but indirect addressing does.
Processor registers include a timer register, eight index registers, an indicator register,
an instruction counter, a 72-bit accumulator (which can also be used as two independent
accumulators or four independent index registers), an exponent register for floating point
operations, and the Base Address Register mentioned above.
A total of 175 basic instructions are available, most of which will be familiar to programmers of other large-scale binary computers. The instruction repertoire includes comparisons (logical, algebraic, magnitude, masked, and between limits), loading, storing,
Boolean operations, branching, and shifting instructions. Provision is made for the use
of half-word, single-word, or double-word operands in many operations.
Floating-point operations include single or double precision loading, storing, comparison,
addition, subtraction, multiplication, and division. Floating-point numbers are represented by a mantissa of 28 bits (single precision) or 64 bits (double precision) and a binary
exponent of 8 bits. Both the exponent and mantissa are represented in two's complement
notation. Single precision is equivalent to about 8 decimal digits, and double precision to
19 decimal digits.
Several special instructions can reduce programming effort and increase efficiency by
facilitating the processing of lists of data and the coding of routines that require multiword precision. There are, however, no editing instructions, no code translation instructions other than Gray to binary, and no radix conversion instructions other than a onedigit-at-a-time binary to BCD instruction.
Interrupt processing is handled by GECOSm, the standard supervisory routine. Machine
or program error conditions are regarded as system "faults", and when these situations
occur, a GECOS m master-mode control program can swap out the current job and call
in another one in order to maintain continuous on-line operation. Remedial action is
always attempted. Certain faults, such as arithmetic overflow or memory parity error,
can be masked to allow processing to continue. Other faults, such as time runout, memory
address violation, illegal operation code, machine lockup, etc., are non-maskable and
will halt processing on the current job. The status of the location counter and indicators
are stored in a fixed control supervisor area before transferring to the applicable fault
routine.

8/68

fA

AUERBACH

'"

(Contd. )

340:011. 220

SUMMARY

.22

Memory Module
The two general-purpose members of the GE-600 Series, the GE-625 and the GE-635. differ
only in the speed of their core storage units. The GE-625 uses a core memory with a cycle
time of 2 microseconds; the GE-635 uses a unit with a 1-microsecond cycle time. Each
access, in both systems, is for a word-pair (two 36-bit-plus-parity words).
Up to 262,144 36-bit words of core storage can be incorporated in a single-processor GE-600
Series system, in modules of 32,768 words.
The Memory Module is the heart of every GE-600 Series system. Each Memory Module is
composed of a System Controller and one to four 32K modules of core storage, and is an
independent unit capable of being accessed simultaneously with other Memory Modules.
The System Controller performs many of the priority and control functions in a GE-600
Series system. Among these functions are:
•

Control of communication between memory and the central processor and
between memory and the I/O Controller.

•

Control of input-output interrupts for multiprocessor systems, system
programs, and peripheral devices.

•

Switching of control signals, addresses, and data to and from the Memory
Module.
Each System Controller has eight "memory ports" (channels) for connection to Processor
Modules, I/O Controller Modules, or non-standard peripheral devices .
• 23

Input/Output Controller Module
The I/O Controller is a small processor containing the necessary logic circuits for independent handling of all I/O operations once a connection to a Memory Module has been established. The I/O Controller uses information from the supervisory area of core memory
to indicate the input or output area of memory. It also performs an address check to prevent an I/O operation from either reading or writing in an area outside the proper program
area. An I/O Controller can have up to 16 input-output channels: 10 standard-speed (up to
25,000 characters per second) and 6 high-speed (up to400, 000 characters per second). Each
I/O Controller can access up to four Memory Modules, and each Memory Module can be
connected to up to four I/O Controllers, providing the capability for connecting a large
number of peripherals on-line to a GE-600 Series computer system.

TABLE I: GE-600 SERIES PERIPHERAL SUBSYSTEMS
Required I/O Channels
No.

Minimum
Channel Rate

1

25KC
25KC
25KC
25KC

1
1
1

25KC
25KC
25KC

1

200KC

2

200KC

lor 2

400KC

1

200KC

1

400KC

1

200KC

1

25KC
25KC

1
1
1

1

Subsystem

CRZ201 Card Reader - 900 cpm
CPZ100 Card Punch - 100 cpm
CPZ200/201 Card Punch - 300 cpm
PRT201 Printer - 1200 Ipm
PTP200 Perforated Tape Punch - 100 char./sec.
PTR200 Perforated Tape Reader - 500 char. /sec.
PTS200 Perforated Tape Reader/Punch
Single.,.channel Magnetic Tape Subsystem (1 to 8
magnetic tape units) - 7,500 to 160,000 char. /sec.
Dual-channel Magnetic Tape Subsystem (1 to 16
magnetic tape units) - 7.500 to 160,000 char. /sec.
DSU270 Disc Storage Unit (15.3 mUlion chars.) 26 msec average access time
DSU200 Disc Storage Unit (23. 6 million chars.) 225 msec average acooss time
MDU200 Magnetic Drum Unit (4.66 or 9.32 million
chars.) - 17 msec average access time
MSS388 Mass Storage Subsystem (680.0 mUlion
chars.) - 430 msec average access time
DATANET-30 Data Communication Processor
Console with Typewriter

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

8/68

340:011.240

• 24

GE-600 SERIES

Peripheral Equipment
The peripheral devices currently available for the GE-600 Series computers are listed in
Table I, along with the number of high-speed or standard-speed input-output channels required for each subsystem •

. 25

System Configuration
Configuration rules for the GE-600 Series Components can be summarized as follows:

•3

•
•

Each Processor Module can be connected to 1 to 4 Memory Modules.
Each Memory Module can be connected to a total of up to 4 Processor
Modules and 4 I/O Controller Modules. Up to 262,144 words of core
storage can be incorporated in a single-processor system.

•

Each I/O Controller Module can be connected to 1 to 4 Memory
Modules and can have from 3 to 6 high-speed and from 5 to 10
low-speed input-output channels •

Software
General Electric is providing a well-integrated line of software for the 600 Series that
includes:
•

General Comprehensive Operating Supervisor (GECOS lID - This
is a master control routine, and all activities of a GE-600 Series
computer system are normally carried out under its control.
GECOS III has provisions for receiving job programs from a card
reader or from a program library, scheduling, allocation of
peripherals and memory, and communication with the operator.
It can control the execution of up to 63 programs concurrently in
a multiprogramming mode. Scheduling is based on priority and
peripheral availability. Communication with GECOS III is handled
through control cards or the console typewriter.
Introduced in November 1967, GECOS III includes all the facilities
of its predecessors, GECOS I and II, plus numerous improvements.
An on-line, tree-structured, hierarchial file system is provided,
with file protection and access control. Serial magnetic tape processing can be simulated on disc or drum. An elaborate accounting
system accounts for all of the time spent by each processor and
peripheral device. All input-output requests, as well as activities
and jobs, are queued according to a priority scheme. An on-line
peripheral test system runs concurrently with other system programs. Core storage is allocated dynamically with memory protection control.
Jobs may be entered from remote terminals, with concurrent local
and remote batch processing. The Time-Sharing Executive (a
specialized control routine of GECOS Ill) supervises and controls
from one to three DATANET-30 Data Communications Processors.

8/68

•

General File/Record Control (GEFRC) - This is the control routine
that will usually be used by programmers specifying input-output
operations. It permits all input-output data to be regarded by the
programmer in terms of files, and frees the programmer from tedious
coding of input-output operations. File Specifications in the user's
programs specify record sizes, blocking, and other information.
(They are produced automatically by the COBOL and FORTRAN compilers.) The device assigned to each file at execution time depends
upon the content of the File Control Card submitted at load time, providing a degree of freedom from the need for specific types of peripheral devices.

•

General Loader - The functions of the General Loader include:
(1) loading programs from the magnetic drum (or disc) into core
storage when they have been scheduled to run; (2) relocating subprograms into a contiguous area of memory and setting the required
linkages; and (3) loading overlay segments and setting up the required
linkages. The General Loader can also cause debugging facilities to
be incorporated at load time.

•

General Remote Terminal S ervisor GERT - GERTS is the control
program or handling jobs from remote terminals. It accepts jobs,
stores them on the magnetic drum (or disc), and submits them to
GECOS for execution based on a priority transmitted with the job.

A

AUERBACH
-",

(Contd. )

SUMMARY

.3

•4

340:011.300
Software (Contd.)
•

MACRO Assembler (GMAP) - GMAP is the symbolic assembly language
for the GE-600 Series. The prime feature of GMAP is its extensive
macro capabilities.

•

~ - GE-600 Series COBOL incorporates all of Required COBOL-61,
most of optional COBOL-61, and the SORT and Report Writer facilities
of Extended COBOL-61. The implemented features of Elective COBOL-61
include the CORRESPONDING option of the MOVE verb and the COMPUTE,
ENTER, and USE verbs.

•

FORTRAN - This is a standard implementation of the IBM 7090/94
FORTRAN IV language, with a few extensions. Capabilities for debugging and variable-field input and output are featured.

•

SORT/MERGE - The GE-600 Series Sort/Merge routine accepts input
from magnetic drum, disc, or tape and will produce output to any of
the same devices. Sorts can be performed on numeric or alphanumeric
keys, with the individual fields of a key in either ascending or descending
order.

•

Bulk Media Conversion - The Bulk Media Conversion routine is contained
in the system library and can be called by control cards. Conversion capabilities include punched card to magnetic tape or disc; perforated tape to
disc; magnetic tape to printer or punched card; and disc to punched card
or magnetic tape.

•

Mathematical Routines - An extensive library of mathematical routines
includes trigonometric, exponential, and logarithmic function evaluation,
matrix manipulation, curve fitting, and polynomial root determination.

•

Service Routines - An integrated set of service routines is provided for
file maintenance, software maintenance (updating of system or user's
compilers or programs), and diagnostics.

•

Integrated Data Store Q-D-S) - This routine provides the capability for
organizing files on a disc storage unit in a non-sequential manner.
Individual detail records are linked together to form chains. A record
can belong to more than one chain, effectively eliminating the need to
store duplicate information. Macro operations are provided for. obtaining
a record to be processed, for storing and linking a processed record,
and for deleting a record. I-D-S can be used to provide mass storage
facilities for COBOL or assembly-language programs for any GE-600
Series computer system that includes a disc storage unit.

•

GE-225 Simulator - This routine simulates the operations of GE-225
systems on a GE-625 or GE-635.

•

IBM 1401 Simulator - This routine simulates the operations of IBM 1401
systems on a GE-625/635. Provision for IBM 1401 simulation was desirable because of the number of IBM 7090/7094 installations that use
IBM 1401 systems as satellites for performing data transcription and
edit/print operatiOns. IBM 7090/7094 installations are likely prospects
for GE-625/635 equipment because of the hardware/software simulation
offering described in the following Compatibility section •

COMPATIBIUTY
General Electric has a hardware/software simulation system that enables IBM 7090/7094
object programs to be run with little, if any, alteration on a GE-625 or GE-635 computer
system. The "7094 Simulator Aid" (or 9SA) is a hardware device that serves as an
auxiliary central processor unit, containing the same number and types of accumulators
and registers as the simulated IBM 7090 Series processor. The 9SA feature can directly
execute many 7090 instructions. GE provides a series of subroutines that will perform
all requests for multiply, divide, floating-point, and input-output operations. The software subroutines that handle I/O operations act as an interface mechanism with the General
File Record Control (GE FRC) I/O system that operates under control of GECOS III. GE
states that the IBM 7090 Series object programs run on a GE-600 Series computer system
will produce results identical to those obtained on the simulated computer.
The IBM 7044 can also be simulated on a GE-625/635 computer system with the aid of the
same hardware facility - 9SA. IBM 7044 instructions that cannot be directly executed
by 9SA are performed by software subroutines.
There is no direct machine-language program compatibility between the GE-600 Series
and the smaller GE-100, 200, or 400 Series computer systems, all of which have different
word lengths and instruction repertoires.
© 1968 AUERBACH Corporation and AUERBACH Info. Inc.

8/68

340:221.101

-A

,'UUII

BDP

GE-600 SERIES
PRICE DATA

.u,.n

GE·600 SERIES
,---

PRICES

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

PROCESSORS
CP8030
CP8030
OPT809

I

SA8030

MM8038

AMM601
MM8030

*MM8032
AMMGOO
AM8030
OPT802
OPT815
*MM8033
*OPT800

CP8031
CPP600

Name

Monthly
Monthly
Rental Purchase Maint.

$
Central Processini Unit !GE-625/635!
Central Processor Unit (includes 1 CPU Port)
Central Processor Unit, Dual (includes 1 CPU
Port for each processor)
CPU Port (maximum of 3 additional per
processor)
7090/7094 Simulator Aid
Main Storage (GE-625/635)
GE-625:
32,768 Words with Controller
(2-microsecond cycle time; includes 2
Memory Ports)
GE-625:
Additional 32,768 Word Memory
(2-mtcrosecond cycle time)
GE-635:
32,768 Words with Controller
(I-microsecond cycle time; includes 2
Memory Ports)
GE-625:
40K Memory with Controller
(2-microsecond cycle time)
GE-635:
Additional Memory Module, 32,768
words (l-microsecom cycle time)
Auxilliary Memory, 32,768 words
(I-microsecond cycle time)
Memory Port (up to six additional)
Execute Interrupt
GE-635:
49K Memory with Controller
(I-microsecond cycle time)
Additional 24K Memory (I-microsecond
cycle time)
Central Processing Unit (GE-645)
645 Central Processor with one
Memory Port Pair
Processor Port Pairs for CP8031

$

$

10,400
18,720

480,000
864,000

680
1,224

73

3,360

5

4,160

192,000

272

6,085

280,800

398

4,025

185,800

264

6,760

312,000

442

7,020

324,000

460

4,475

206,400

292

4,475

206,400

292

84
105

3,840
4,800

5
7

7,800

360,000

510

5,620

259,200

368

14,560

672,000

952

260

12,000

17

6,760

312,000

442

85
260

3,840
12,000

6
17

4,475

206,400

292

4,475

206,400

292

84
105
760

3,840
4,800
35,000

6
7
50

Main storage (GE-645)
MM8040
REC600
CAB601
AUM600
AMM600
OPT802
OPT815
Clli600

32,768 Words with GE-645 System Controller
(I-microsecond cyole time)
Reoonflguration Option for MM8040
Auxtliary Cabinet for 32/64K Memory
extension and clook/ switch options
32,768 Words (I-microsecond cycle time;
first memory module installed in
CAB60l)
Additional Memory Module (32,768 words
I-microsecond cycle time; second
memory; module installed in CAB601)
Memory Port (up to 6 additional)
Additional Interrupt Cells
Optional GE-645 System Clock with time
of day and alarm interrupt features

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

GE-600 SERIES

340221.102

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

I' noc ESS( lHS
(('ontl!, )

Name

PRICES
Monthly
Monthly
Rental !Purchase Maint.

$

$

$

50
12,430
NC

2,250
531,200
NC

3
1,328
NC

NC
260

NC
12,000

420
390
31

19,200
18,000
1,440

28
2il
2

600

27,700

139

5,620

259,200

368

260

12,000

17

315

14,400

20

575

26,400

37

84
52

3,840
2,400

5
3

105

4,800

7

5,620

259,200

368

260
65
990
NC
600
35

12,000
2,880
45,600
NC
27,600
1,440

17
3
G5
NC
39
2

680
195
85
195
85
130
35
420

31,200
8,800
3,840
8,800
3,840
6,000
1,440
19,200

46
10
5
10
5
9
2
29

125
65
680
2,325

5,760
2,880
31,200
55,000

9
2
52
92

Main Stur'a,," (G 1-:-/;4[;) (ContJ,)
MSW/;oO
E MlI:!02
E 1\[C':102
J\1C1I302
.\ 1'1'302

Memory :->wlteh
Extlmdod Memory I Tnit (24 million characters)
Controllt~r' for EMlT302 (includes 2 memory
ports)
Coolin" Unit for EMU302
Additional Port pairs for EMC302

NC
:lO

Processing Unit Options (GE-625/635)
C08030
C08031
ST8030

Master Console
Auxiliary Console
Console Storage
Processing Unit Options (GE-645)

S('(,600

,\T'[..\CI["rENTS,
,\D,\PTI-:HS,

System Control Console (permits semiautomatic
system reconfiguration)
Input-Output Channels {GE-625/635l

DC8032

:\(\'O

CIIAl'\l'\ELS

CII0500
('110030
('110[;30
J\JlI'GOI
OPT806
01'T811

Input-Output Controller (IOC) (includes one
Memory Interface Port)
Additional Peripheral Channel Package (five
25kc channels)
Additional Peripheral Channel Package (three
400kc channels)
Additional Peripheral Channel Package (five
25ke and three 400ke ehannels)(l)
Additional Memory Interface Port
Additional Peripheral Channel, 25kc
(maximum of 5 extra)
Additional Peripheral Channel, 200ke
(maximum of 3 extra)
Input-Output Channels (GE-645)

DC8031

J\fl PGOO

PLP600
C ,\TI600
AMA600
11',\600
IP(,GOO
II P(,GOO
(,AA600
CAC600
('SAGOO
CS(,600
D(;A600
DGC600
TTA600

TTC600
TTL600
DDA600
('DA600

2/69

645 C'nmeralized I/O Controller (GlOC) includes
one memory Port pair, four status channels,
two priority level groups, one HPC600, one
IPA with six IPC, and one CAB600,
GlOC Port Pairs
Priority level modules, 12 level
Adapter Cabinet
GIOC Adapter Maintenance Aid
Indirect Peripheral Adapter
Indirect Common Peripheral Channel for
IPA600 (maximum of 6 per IPA600)
Direct Common Peripheral for GIOC
Character Asynchronous Adapter for GlOC
Character Asynchronous Channel for CAA600
Character Synchronous Adapter for GIOC
Character Synchronous Channel for CSA600
Dialing Adapter for GIOC
Dialing Channel for DGA600
Teletypewriter Adapter (includes Teletype speed
options of 45.5, 50, 56.9, 74.2, 110, 133.2,
150, and 165 bits/sec)
Teletypewriter Channel Group for TT A600
Teletypewriter Channel Extensions
Direct Disc Adapter (for DSC11F)
Custom Di rect Adapter

A

AllfRBACH
~

340:221.1 03

IIRICE DATA

,

,,

PRICES

IDENTITY OF UNIT
CLASS

ATTACHMENTS,
ADAPTERS,

'NO
CHANNELS
(Contd. )

Model
Number

Feature
Number

Monthly
rMonthly
Rental IPurchase Maint.
$
$
$

Name
Motor Generators

MG80S0
MG80Sl
MG8032
MG8033
MG8034
OPT825
OPT826

M.G. Set, Sl. S KVA,
M.G. Set, 62.6 KVA,
M.G. Set, 62.6 KVA,
M.G. Set, 62.6 KVA,
M.G. Set, 62.6 KVA,
input
Power Sequencer, 60
Power Sequencer, 50

60
60
60
50
60

Cycle,
Cycle,
Cycle,
Cycle,
Cycle,

220/440 V.
440 V.
480 V.
380 V.
208 v. a. c.

Cycle
Cycle

so

265
S20
320
340
320

12,190
14,590
14,590
15,600
14,590

S6
36
39
36

17
21

770
960

1

175

7,620

13

42
520

1,850
23,080

8
38

540
185

24,000
8,230

14

52
590
20
1,535
1,170
210
420
625
315
420
1,200
400
650
850

2,320
25,510
560
68,080
53,000
8,000
16,000
23,000
15,000
20,000
53,340
17,780
31,200
44,000

(3)
113
350
NC
NC
NC
NC
NC
89
30
52
121

3,435

146,700

367

2,090

89,580

224

270

11,560

29

95

4,080

10

835
3,020
3,435

36,930
136,000
146,700

62
650
367

940
940

41,540
41,540

69
69

1

Peripheral SWitches
Manual Peripheral Switch Console (includes
one OPT510)
Manual Peripheral Switch Unit
Programmable Peripheral Switch (not available
for GE-600 Series)

PSC200
OPT510
PS60l0

Disc Storage

MASS
STORAGE
DSCl60
ADCl60
DFPl60
DSU160
DCT160
DSC200
DSU204
OPT201
OPT202
OPT203
OPT204
OPT205
DSC270
ADC270
DFE270
DSU270

Disc Storage Controller, Single Channel (2)
Additional Data Channel (Nonsimultaneous)
Data File Protect
Removable Disc Storage Unit
Disc Cartridge
Disc storage Controller, Single Channel
Disc Storage Unit, 4 Discs
4 additional discs(4)
8 additional discs(4)
12 additional dlscs(4)
Fast Access I (4 dlscs)(4)(5)
Fast Access II (8 discs)(4)
Single Channel Controller
Additional Channel
File Electronics Unit
File Storage Unit

40
4
70

Drum Storage(6)
MDC201

ADS201
I
I

ADC201
MGS200

Magnetic Drum Controller for MDU200 and
ADS201 Includes MDU200 Magnetic Drum
Unit with 4.7 million-character storage
capacity
Additional Drum storage Unit (4.7 million
characters)
Additional (non-simultaneous) Data
Channel for MDC201(7)
Motor Generator Set, 2 KVA, for MDC201
(208 volt, 60 cycle, 3 phase to 120/208 volt,
60 cycle, 3 phase)
Magnetic Strip Storage

MSC388
MSU388

~MDS200

Mass Storage Controller
Mass Storage Unit
Magnetic Drum Storage Unit and Controller
(4.66 million characters)
Magnetic Tape

INPUTOUTPUT
MTC330
MTC331

Controllers, Single Channel: (8)
All Speeds, Address Select
All Speeds, No Address Select

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

GE·600 SERIES

340221.104

. -J01

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

Name

:\1'1'(':1:14
~IT(,3:Hi

1\1'1'(,404
:\11'('40(;

I

1\ IT 112 00
:\1'1'11201
:\1"1'11:100
:\ 1 r11301
:\I"I'H402
i'I1TI!403
I\TTlI404
i\IT1!405
l\IT1!492
i\TTJl4fl3
l\lTII372
l\IT1!373

$

$

Tape (Contd. )

n-('hannel. All Speeds Address Select
!1-Channel, All Speeds, No Address Select
ControllerH. Dual Channel:(9)
All Sp(!eds. Address Select
All Speeds. No Address Select
fl Channel. All Speeds, Address Select
fl Channel, All Speeds, No Address Select
Magnctic Tape Handlers:
7.5/21kc, Single. 7-Channel
15/42kc. Single. 7-Channel
7.5/21/30kc, Single, 7 Channel
15/42/60kc, Single, 7 Channel
10/28kc, Single 9 Channel ASA
10/28/40kc, Single, 9 Channel ASA
20/60kc, Single, 9 Channel ASA
20/60/60kc, Single, 9 Channel ASA
40/11kc, 200/556 bits/inch
40/111/160kc, 200/556/800 bits/inch
30/83kc, 200/556 bits/inch
30/83/120kc, 200/556/800 bits/inch

I\ITC400
!\TT(' 401

Monthly
Monthly
Rental Purchase Maint.

$
M:l~netk

INPllTOllTP\'T
(Contd. )

PRICES

1,010
1,010

44,770
44,770

75
75

1.435
1.435
L 545
1,545

63,700
63,700
68,540
68,540

lOr;
10(;
114
114

305
505
420
615
305
420
505
615
780
895
780
895

12,890
21,560
17,780
26,230
12,890
17,780
21,560
26,230
34,620
39,500
34,620
39,500

43
72
59
88
43
59
72
88
113
130
113
130

680
520
860

26,000
20,000
33,000

121
93
154

520
585
990

20,000
22,400
38,000

100
112
190

1,460

56,000

280

Punched Card
Card Reader and Control (900 cards/min)
Card Punch and Control (100 cards/min)
Card Punch and Control (300 cards/min)

(' HZ201
(,I'Z100
C I'Z201

Paper Tape
Perforated
Perforated
Perforated
PTR200

PTH200
PT 1'200
I'TS200

Tape Reader (500 char/sec)
Tape Punch (150 char/sec)
Tape Subsystem (Includes both
and PTP200)

Printers
PRT201
PDR200
PDR20l
PDR202
DRD200
I\lSM200
PDR210
PDR203
PCW20Q

Printer and Control (136 column: 1200 lines/
min). (Choice of PDR201, PDR202 or
PDR203. Drum at no extra charge. )(10)
Standard ASCII Print Drum and ASCII Code
Wheel standard FORTRAN Print
Drum and ADCn Code Wheel (0.103" characters)
Standard FORTRAN Print Drum and ASCII
Code Wheel (0,091" characters)
COC-5 Document Reader (2 stackers, 1200
doc/min)
Mark Sense Module for DRD200
Standard Print Drum with COC-5 characters
for PRT200, PRT201
Standard OCR Farrington (12F Self-Check)
Print Drum with ASCII Code Wheel
Standard FORTRAN Code Wheel (not required
for PDR201 and 202) to print in FOHTRAN
sequence

-

-

-

-

1,200

46,460

233

100
84

3,880
3,600

20
18

-

-

-

MICR Reader/Sorter
:\THS200
OPT311

2/69

MICR Document Reader/Sorter (12-pocket;
1200 doc/min)
MOD II Check Option

fA

AUERBACH
~

2,080

80,000

400

105

4,500

15

340: 221.105

PRICE DATA

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

Name

PRICES
IMonthly
Rental

IPurchase

Monthly
Maint.

$

$

$

MICR Reader/Sorter (Contd.)

INPllTOUTPUT
(Contd. )

OPT312
BFR200

Endorser Stamp Option
COC-5 Bar Font Reader Option

105
520

4,500
20,000

23
100

1,200

46,460

233

100

3,880

20

1,250
1,560
2,080
235
315
260
360
210

70,000
95,000
130,000
10,390
13,500
11,200
17,340
9,000

292
292
292
17
35
35
58
20

55
350
110
105
125
50
10
15
20
25

2,020
14,000
4,600
4,200
5,000
2,100
375
550
735
900

10
24

170

6,250

20

190

7.950

26

15
5
5
5
30

550
185
185
185
1.250

3
1
1
1
5

25
10
15
20
(11)

5
2
3
4

150

1,000
375
550
735
(11)
(11)
7,000

1,100
1,120
1,160
1.230
1,350
290

48,200
49,100
50,900
54,090
59,450
8.250

Optical Reader
DRD200
MSM200
COMMUNICATIONS

COC-5 Document Reader (2 stackers, 1200
doc/min)
Mark-Sense Module Option for DRD200
DATANET-30 Processors

DCP930
DCP931
DCP932
AMC930
CSU931
CPC930
CPC931
PlU930

Processor (4,096 word memory)
Processor (8, 192 word memory)
Processor (16,384 word memory)
Additional Module Cabinet
Controller Selector Unit
Common Peripheral Channel
Common Peripheral Channel, High Speed
Processor Interrupt Unit
Remote Terminals

DTU760
DCU760
TMU760
DLC760
DLC765
PPC760
EMC760
FKG760
FKG765
DMU765
DLC761
DLC766
LRU760
KVA760
PLJ760
PLM760
DMU761
EKB761
EMC761
FKG761
FKG766
SEXXXX

SEXXXY
*TIM224
******PTT600

Datanet 760:
Display Terminal Unit
Display Controller Unit
Terminal Memory Unit
Data Line Controller (1200 bits/sec)
Data Line Controller (2400 bits/sec)
Page Print Controller
Entry Marker Control
Function Key Group (8 Key Group)
Function Key Group (16 Key Group)
Display Monitor Unit (23-Inch Read-Only
Monitor)
Data Line Controller (1,200 bits/sec full
duplex, asynchronous)
Data Line Controller (2,400 bits/sec full
duplex. synchronous)
Line Repeater Unit
Keyboard Video Amplifier
Party Line Junction
Party Line Monitor
Display Monitor Unit (14 inch Read Only-or
for use with EKB761)
Electronic Keyboard
Entry Marker Control for EKB761
Function Key Group for EKB761
Function Key Group for EKB761
Closed Circuit TV Compatibfiity
Video Switching and Monitoring
248kc Digital Data Modem
DATANET-91 Magnetic Tape Terminal:
with 256 character Buffer
with 512 Character Buffer
with 1,024 Character Buffer
with 2,048 Character Buffer
with 4, 096 Character Buffer
DATANET-600 Paper Tape Terminal

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

(11)

8

14
16
4
2
3
4
5

(11)
(11)

23
160
164
170
180
197
35

2/69

GE·600 SERIES

340: 221. 106

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

Name

PRICES
Monthly
Monthly
Rental Purchase Maint.
$
$
$

NOTES:
,.. No
(1)

lon~r

in production.

CH0530 should not be used when CH0500 and/or CH0030 are ordered.

P) Price includes onc high-spccd channel (HC()012) when the DSC160 is ordered with the GE-400 Series System.
(:l)

Time and l1l:lfl'I'ials.

(4)

No charge for f:letory instillation.

(5)

Not available wIth OPT203.

(Ii)

Not available for GE-405 or GE-415 bystem.

(7)

ADC201 IS available only on GE-400 line (not GE-405 or GE-415); it allows two GE-400 line systems to share one
Magnetic Dnlln Subsystem.

(8)

Controls up to 8 single Magnetic Tape Handlers.

(9)

One time charge of $625 for field installation.

Controls up to 16 single Magnetic Tape Handlers.

(10)

b'pccial code wheels can be ordered using special (SEXXXX) number; prices will be quoted upon approval of
special.

(11)

Negotiated separately.

2/69

fA

AUERBACH

"

J

INTERNATIONAL
BUSINESS MACHINES
CORPORATION

(

AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

AUERBACH

@)

Printed in U.S.A.

-1.
-

401:011. 100

"'m"

/4",- EDP
AUERBAC~
. RfPOITS

IBM 1401
INTRODUCTION

SUMMARY: IBM 1401
.1

BACKGROUND
The IBM 1401 is a small-scale data processing system, oriented toward business applications, that features a wide range of peripheral devices and supporting software. Monthly
rentals for typical card configurations range from $1,300 to $4,300. In expanded configurations, including magnetic tape and disk storage, a 1401 system can rent for upwards
of $12,000 per month.
The IBM 1401 was originally announced in 1959 as a system specifically designed to facilitate the transition from punched card unit-record equipment to faster, larger-scale data
processing. The first 1401 system was installed in September 1960, and during the ensuing
years it has been regarded as the workhorse of the data processing industry. This reputation has evolved as a result of the wide acceptance that the 1401 has received, not only as
a small-scale business data processing system, but as an off-line input-output processor
for larger tape-oriented systems such as the IBM 7070 and 7090 Series. At this writing,
approximately 9300 1401 systems are still in use.
The use of transistorized circuits in the 1401 components resulted in a relatively low-cost
system with increased reliability and decreased maintenance and space requirements as·
compared to earlier vacuum-tube computers. Those components requiring operator attention are easily accessible. The controls and arithmetic components are consolidated into
a single set of modular cabinets.
The 1401 system is well suited to the processing of large volumes of card documents. The
magnetic tape configurations provide more compact record handling and storage for higherspeed data processing. The disk storage configurations permit rapid access to large
volumes of data without the necessity of processing large card volumes or sorting tape
records.
The 1401 Model G was introduced in 1964 for those installations not large enough to justify
the purchase or rental of higher-performance 1401 models. The 1401-G, with approximately
1600 installations to date, provides an economical punched-card system without tape or
random-access processing capabilities. The most important selling point of the 1401-G is
price. The average monthly rental of the minimum 1401 card system (1,400 core storage
positions, card reader-punch, and printer) has been reduced from $2,565 for a 1401 Model A
to $2,260 for the Model G.
Although the 1401 Model G has the same instruction set and processing speed as the earlier
1401 models, there is a significant reduction in system throughput as a result of the use
of slower models of the card reader-punch and printer.

Figure 1. A basic IBM 1401 card system, consisting of a 1401 Processing Unit (center),
1402 Card Read-Punch (left), and 1403 Printer.
@

1968 AUERBACH Corporation and AUERBACH Info, Inc.

5/68

- - - - ---.-.-" -

,

401:011. 101

.1

IBM 1401

BACKGROUND (Contd. )
The 1401 Model H, introduced in May 1967, represents a further reduction of the average
monthly rental to $1,300. Like its predecessor, the 1401-G, it was designed to replace
punched-card equipment. However, the 1401-H has a slower core storage cycle time than
that of the other 1401 models, which, when combined with its slower input-output devices,
further reduces the throughput rate.
The basic characteristics of the various models of the 1401 system are summarized in
Table I.
The 1401 was the first member of the IBM 1400 series of data processing systems, and it
is now the second smallest in price and throughput. The IBM 1440 (Report 414), the
smallest member of the series, is program-compatibile with (and slightly faster than)
thE' 1401 with respect to internal processing, but the 1440 uses slower input-output units
and different instructions to control them. The IBM 1460 (Report 415) uses the same set
of stored-program instructions as the 1401, so programs coded for a 1401 can, in general,
be run without alteration on a 1460 with the same (or expanded) complement of inputoutput units and optional features. The 1460 is nearly twice as fast internally as the 1401
and uses most of the same peripheral devices. The faster, more expensive IBM 1410
(Report 402) uses a different addressing method and instruction set, but can execute many
1401 programs without alteration through the use of built-in 1401 compatibility circuits.
The 1401 constitutes the principal "second generation" computer system from the leading
computer manufacturer. It was regarded, for a number of years, as the primary standard
of comparison for ~sers with small to medium-size data processing requirements.
The 1964 introduction of the "third-generation" IBM System/360 series, with its 1401
emulation facility and improved performance/cost ratio, made it apparent that the 1401
would gradually be replaced. However, the proven reliability of the 1401 hardware and
supporting software, the relatively low cost, the widespread availability of programmers
and analysts trained in 1401 processing, and the numerous available application programs
and special-purpose peripheral devices ensure that the 1401 will continue to have considerable practical utility for many computer users. As a result, the 1401 is virtually
assured a substantial position in the industry for some years to come .

.2

HARDWARE

.21

Processing Unit
The 1401 Processing Unit contains the logic, arithmetic, and control circuItry for the
entire system, together with a magnetic core memory with a capacity of 1,400, 2,000,
or 4,000 alphanumeric character pOSitions. Additional core storage is available in
4, OOO-position modules, providing system memory capacities of 8,000, 12,000, or 16,000
positions.
Each position is individually addressable and contains an 8-bit character consisting of six
data bits, a parity check bit, and a word mark bit for field definition. The character set
includes 10 numeric, 26 alphabetic, and 12 special characters.
Data and instructions are stored in variable word-length form. A word is a single character or group of characters that represents a complete unit of information, defined by a
word mark bit in the high-order position. Instructions range from one to eight characters
in length; the basic two-address instruction format consists of a one-character operation
code and two three-character operand addresses. Instructions are normally stored in
sequential locations and executed in that order except when a branch instruction is executed.

TABLE I: CHARACTERISTICS OF THE IBM 1401 MODELS
1401 Frace8SIDg Unit Model
Core storage capacity:
Maximum positions
Minimum positions
Co.-e storage cycle, "sec
Card reader speed. cpm

5/68

Model A

Model B

Model C

ModelD

Model E

Model F

ModelG

16.000
1,400

16.000
1,400

16.000
1,400

16,000
1,400

16.000
1,400

16,000
4,000

4,000
1,400

11.5
800

11.5
800

11.5

11.5
None

800

11.5
800

11.5

11.5

800

450

Model H

4,000
4,000
19.3
450

Card punch speed, cpm

250

250

250

None

250

250

250

250

Printer speed, lpm

600

600

600

600

600

600

465 or
340

340

729 MagnetIC Tape Units

None

None

Up to 6

Up to 6

None

Up to 6

None

None

7330 Magnetic Tape Units

None

None

Up to 6

Up to 6

Up to 6

None

None

1311 Disk storage Drives

None

Up to 5

Up to 5

Up to 6
None

Up to 5

None

None

None

1405 Disk storage Units

None

None

None

None

None

1 unit

None

None

A

(Contd.)

AUERBACH
~

INTRODUCTION

.21

401;011. 210

Processing Unit (Contd.)
Core storage cycle time is 1l.5 microseconds (19.3 microseconds in Model H), compared
to 11.1 microbeconds in the IBM 1440, 6.0 in the 1460, and 4.5 in the 1410. Instructions
are executed at the rate of about 4,000 per second in typical 1401 routines.
All data transfers are automatically checked by an odd-bit parity check code.
Every character read into core storage from the card reader is checked for validity.
Address validity checking is performed to insure that all addresses used in a program
are within the core-storage capacity of the system.
An instruction is addressed in its high-order position and scanned from left to right until
the word mark associated with the next sequential instruction is sensed. The final instruction must have a word mark set at the right of its low-order position. A data field is
read from right to left until a word mark is sensed.
Any of the possible positions in core storage is addressable by a three-character address.
The addresses are basically decimal, but zone bits over the hundreds and units positions
are used for all addresses larger than 0999. Table II outlines the 1401 addressing scheme.
The Advanced Progranlming feature provides an indexing facility for address modification;
zone bit over the tens position of the address indicates which of three 3-character index
registers shall be used to modify it.
There are four address registers in the 1401: one controls the program sequence, wo
control data transfers from one storage location to another, and one specifies the storage
location that is active during a particular storage cycle. An operatiorr code register and
two character registers store data during the execution of an instruction.
A "chaining" capability permits a series of basic operations, such as arithmetic functions
and data movements, to be performed efficiently on fields arranged in consecutive storage
locations, thereby saving storage space and execution time. A powerful editing operation
provides zero suppreSSion, insertion of identifying symbols, and punctuation of print
output. Four areas of storage are reserved for input-output buffering of card readers,
punches, and line printers.
The flexible editing capability is standard, but multiplication, division, indexing, threeway comparisons, sense switches, and multi-word internal transfers are all extra-cost
options. WitllOut these optional features, the processing capabilities of the 1401 are severely
limited. In fact, the Advanced Programming Feature (which provides three index registers,
instructions to store the address register contents, and the "move record" instruction) is
nearly indispensable if the user hopes to take advantage of the much-heralded variable
field-length capabilities of the 1401.

TABLE II: IBM 1401 ADDRESSING SCHEME

Actual Addresses

Zone Bits Over
Hundreds Position

Zone Bits Over
Units Position

0000
1000
2000
3000

to
to
to
to

0999
1999
2999
3999

No zone bits
A-bit (Zero-Zone)
B-bit (ll-Zone)
AB-bits (12-Zone)

No
No
No
No

4000
5000
6000
7000

to
to
to

4999
5999
6999
7999

No zone bits
Z-bit (Zero-Zone)
B-bit (l1-Zone)
ZB-bits (12-Zone)

A-bit
A-bit
A-bit
A-bit

(Zero-Zone)
(Zero-Zone)
(Zero-Zone)
(Zero-Zone)

8000
9000
10000
11000

to
to
to
to

8999
9999
10999
11999

No zone bits
A-bit (Zero-Zone)
B-bit (ll-Zone)
AB-bits (12-Zone)

B-bit
B-bit
B-bit
B-bit

(ll-Zone)
(ll-Zone)
(ll-Zone)
(ll-Zone)

12000
13000
14000
15000

to

12999
13999
14999
15999

No zone bits
A-bit (Zero-Zone)
B-bit (ll-Zone)
AB-bits (12-Zone)

AB-bits
AB-bits
AB-bits
AB-bits

to

to
to
to

zone
zone
zone
zone

bits
bits
bits
bits

(12-Zone)
(12-Zone)
(12-Zone)
(12-Zone)

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

3-Character Addresses

to
to
to
to

999
Z99
R99
199

1O;i
?O;i

to
to
to
to

99Z
Z9Z
R9Z
19Z

OO!
;iO!
!O!
?O!

to
to
to
to

99R
Z9R
R9R
I9R

OO?

to
to
to
to

991
Z91
R91
191

000
;ioo

100
?OO
oo;i

;io;i

;iO?
!O?
?O?

5/68

401:011.211

.21

IBM 1401

Processing Unit (Contd.)
The use of dynamically variable field lengths (i. e. , fields whose lengths vary from record
to record within the same file) deserves very serious consideration. The main advantage of
variable field lengths is that the required input-output time is reduced, and this is a valid
consideration when the input-output time is the limiting factor ~n overall processing time.
The additional data manipulation required to utilize these fields of varying length and varying
location within a record, however, can significantly increase the central processor time
(sometimes to the point where it exceeds tape input-output time) and the programming complexity.
An alternative method of reducing total time requirements for a problem is the use of a
variable record-length technique employing combinations of variable and fixed-length fields.
All fields (usually numeric) that require considerable manipulation are assigned fixed lengths
and fixed locations in the record, while any fields (usually alphabetic) that require very little
manipulation form the variable portion (usually the end) of the record. This method effectively reduces total time requirements for most applications without unduly complicating the
programming.

System operation is baSically serial ~ nature (i. e. , one operation at a time). Little overlapping of input-output operations ~h one another or with internal processing is possible
unless optional features such as Print storage, Processing Overlap, and Read Punch Release
are added. Use of these features increases the system's capability for simultaneous operations, but also increases programming complexity and input-output area storage requirements .
. 22

Peripheral Equipment
The 1401 Card Read-Punch reads standard 80-column cards at a peak speed of 800 cards
per minute (450 cpm for Models 4, 5, and 6) and punches them at 250 cards per minute in
all models. The 1403 Printer prints up to 600 alphameric lines per minute (465 lpm on
Models 4 and 5, 3401pm on Model 6). It features a horizontal-chain printing mechanism
that produces high-quality printing and permits interchangeable character sets. The 1404
Printer is a combination unit capable of processing either card forms at up to 800 lines per
minute or continuous paper forms at up to 600 lines per minute.
Up to six 729 and/or 7330 Magnetic Tape Units can be connected to a 1401 system. Peak
data transfer rates range from 7,200 to 62.500 characters per second. Only one tape read
or write operation at a time is possible. The central processor is interlocked during tape
read and write operations unless the Processing Overlap feature is added. With Processing
Overlap, internal processing can be overlapped with tape start-stop times and (at transfer
rates of 20,016 characters per second or below) with character transfers to or from a tape
unit.
Up to four 7340 Model 2 Hypertape Drives can be connected to a 1401 through the Serial
Input/Output Adapter. These magnetic tape units are cartridge-loaded, have peak data
transfer rates of 34,000 characters or 68,000 decimal digits per second, and are compatible
with the faster Hypertape Drives used on mM 7074, 7080, 7090, and 7094 systems (but not
with the 729 or 7330 tape units).
Up to five 1311 Disk Storage Drives can be used in a 1401 system. Each drive holds one
replaceable Disk Pack at a time, providing random-access storage for 2,000,000 alphameric characters in addressable sectors of 100 characters each. With the optional Track
Record feature, a single 2, 980-character record can be recorded on each track, increasing
the capacity of a single Disk Pack to 2,980,000 characters. Up to 20, 000 characters can
be read or recorded without movement of the comb-like access mechanism, so the system
is suitable for sequential as well as random processing. Total waiting time for access to
a randomly-placed record averages 270 milliseconds; with the optional Direct Seek feature,
the figure is reduced to 170 milliseconds.
The older 1405 RAMAC Disk storage Unit provides 10,000,000 or 20,000.000 character
positions of non-replaceable storage, in 200-character blocks. Average random access
time is about 600 milliseconds.
The capabilities of the 1401 can be expanded through the attachment of numerous other units,
including:
•

-5/68

mM 1009 Data Transmission Unit

•

mM 1026 Transmission Control Unit

•
•

IBM 7710 Data Communication Unit
IBM 7740 Communication Control System

•

IBM 7770 Audio Response Unit

•

IBM 1011 Paper Tape Reader

A

(Contd. )

AUERBACH

'"

401:011.220

INTRODUCTION

.22

Peripheral Equipment (Contd.)
•

mM 1012 Paper Tape Punch

•
•
•

mM 1445 Printer
IBM 1412 and 1419 Magnetic Character Readers
mM 1231 Optioal Mark Page Reader

•
IBM 1285 Optical Reader
•
mM 1418 Optical Character Readers
•
IBM 1428 Alphameric Optical Reader
The Serial Input/Output Adapter permits connection of anyone of the following devices at a
time: a paper tape reader or punch, a magnetic or optical character reader, a data transmission terminal, or a direct system-to-system link with an mM 1440, 1460, or another
1401.
.3

SOFTWARE

.31

Assemblers
The Symbolic Programming System (SPS) is basically a one-for-one assembly system in
which one symbolic statement is written for each instruction in the object program. SPS-l
requires a minimum core storage capacity of 1,400 positions, while SPS-2 requires 4,000
positions. The language of SPS-2 is essentially the same as that of SPS-l; a fixed-form
coding sheet is used, and there are no facilities for literals or macro-instructions. The
output deck contains a self-loading routine that loads the object program.
Basic Autocoder 2K is a symbolic assembler designed speCifically for card systems with
2,000 core storage positions. A processor program assembles the source program into
machine-language form ready for execution.
Autocoder (on Tape) is more flexible because the processor program resides on and operates
from magnetic tape. This version offers a more powerful language that includes macroinstruction facilities, free-form coding, and literals. The macro routines, provided in the
Autocoder Library, can relieve programmers of much repetitive work. Autocoder (on Tape)
requires a 4K Processing Unit and four magnetic tape units for assembly.
Autocoder (on Disk) requires no magnetic tape units for assembly. The assembly process
is made automatic through the use of 1311 Disk Storage. The Autocoder Library in this
version has expanded capabilities and can be relocated to any area in disk storage. Multiple Autocoder libraries can be built. Jobs can be stacked on the disk unit for consecutive
or selective execution. Alternatively, the object programs can be loaded from punched
cards •

. 32

Compilers
FORTRAN is a symbolic language that closely resembles algebra. Three FORTRAN compilers are available for the 1401. One operates on an 8K card system and uses a restricted language that is essentially FORTRAN I. The other two compilers accept the
FORTRAN IV language and require a 12K Processing Unit; one uses at least four magnetic
tape units,. while the other uses a 1311 Disk storage Drive for system residence.
COBOL is a common business-oriented language, similar to English, that was designed
to facilitate commercial data processing. The 1401 COBOL compiler (tape version) translates COBOL source statements from punched cards into Autocoder-language statements
which are then assembled into machine langu~e. The compiler resides on magnetic tape;
it requires a 4K Processing Unit and at least four tape units.

. 33

COBOL (on Disk) is similar in design to the tape version. The major difference is that
the disk-oriented compiler resides on a file-protected area of 1311 Disk Storage and operates under a system control program. The control program or monitor permits stacking
of tasks and user assignment of input-output devices for defined record files. A 4K Processing Unit and one 1311 Disk Storage Drive are required for compilation .
Report Program Generators
The 2K Report Program Generator (RPG) produces programs that write reports based upon
data from card input files. The RPG user writes a set of report specifications instead of
a detailed program for each report to be produced. The generator is designed for use on a
card system with 2,000 core storage positions. The output may be punched cards and/or a
printed report.
Larger versions of the 1401 Report Program Generator are available for card, tape, or
disk-oriented systems with at least 4,000 core s~orage positions. They provide more
flexibility in input-output media, an edit listing vf the source program, and an error analysis
of the specification cards.

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

5/68

401:011. 330

· 33

IBM 1401

Report Program Generators (Contd.)
FARGO, a ''load-and-go'' system, produces IBM 407-type reports in a 4K, card-oriented
1401 system. Report definition statements are read into the 1401 together with the data
cards. The FARGO program, which uses these specifications to modify itself, is ready
for immediate execution without any assembly procedure.

· 34

Input-Output Control Systems
10CS (on Tape) is designed to eliminate much of the detailed programming of input-output
operations. It consists of a set of library routines that supplement the Autocoder program
on tape. There are routines for blocking and deblocking of records, I/O error correction,
file labeling, and standardized macro-instructions for reading and writing such as GET,
PUT, OPEN, and CLOSE.
10CS (on Disk) differs from the tape version in that it makes available macro-instructions
that process disk I/o operations.
Communications 10CS (1026/DDC and 1448-7740/DDC) provides routines that free programmers using a data communications system from most of the detailed coding required
to transfer data to and from remote terminals connected to an IBM 1026 Transmission
Control Unit, or between a 1401 system and a 7740 Communication Control System.

· 35

Testing Aids
Autotest is an effective aid in the testing of Autocoder, SPS, and FARGO programs. The
programs may be stacked, and documentation is produced to evaluate the tested programs.
standard features include an automatic printout of core storage and magnetic tapes, and the
generation of operating instructions.
The Autotest program can use any of the following features in any combination for more
efficient testing: the ability to patch the object program without reassembling or manually
calculating patching addresses, to trace and print the flow of data during program execution, to snapshot core storage during program execution, and to produce an 80/80 listing
of any punched card output. The program can accept input only from cards.

· 36

.37

Utility Routines
IBM has developed a total of 11 generalized routines to handle sorting and merging operations
on 1401 systems, using magnetic tape, disk storage, or both. The Multiple Utility Program, designed primarily for 1401 systems serving as input-output processors for larger
IBM computers, can control card-to-tape, tape-to-card, and tape-to-printer data transcription operations simultaneously. Other utility routines perform functions such as:
•
•

System initialization.
Core storage dumps on cards or printer.

•
•

Multiplication and division (in systems that lack the hardware feature).
Control of paper tape punching.

•

Data transcriptions between cards, magnetic tape, disk storage, and printer.

•
Creation and maintenance of disk storage files.
•
Control of data transmission via a 1009 or 7710 controller.
Application Programs
Packaged programs have been developed for a wide variety of 1401 applications by IBM,
by users, and by independent software firms. Among the application programs currently
available from IBM are: Demand Deposit, Auto Rating for Fire and Casualty Companies,
Wholesale IMPACT (inventory control), Autoprops II (numerical control), Bank Management
Simulation, Management Decision-Making Laboratory, Data Analysis and Reduction, KWIC
(Key Word In Context) Indexing, Financial Analysis (common stocks), Bond Trade Analysis,
Decision Logic Translator (decision tables to FORTRAN), Selective Dissemination of
Information, Autoplotter, Portfolio Selection, Engineering Scheduling, Homeowners Rating
(insuranceh Linear Programming, and Allocation of Resources for Savings and Loan
Associations.

5/68

A ..

AUERBACH

401 :221.101

IBM 1401
PRICE DATA

IBM 1401
IDENTITY OF UNIT
CLASS

PROCESSOR

Model
Number

I Monthly

Feature
Number

Name

I

1406

Purchase

Monthly
Maint.

$

$

$

1,430
1,180
1,070
500
1,580
2,680
2,175
2,655
2,080
1,610
2,110
2,590

87,700
59,750
59,350
35,000
97.600
131,000
130,000
130,000
126,400
98,950
127,750
127,750

55.00
58.00
54.00
50.00
60.00
91.50
79.00
81.50
80.00
62.00
82.00
84.50

-230

-4,000

-1.50

-130

-3,200

-1. 50

0

0

0

+50
+75
+75

+2,150
+2,550
+2,550

+2.00
+3.50
+3.50

575
1,075
1,575

24,500
45,800
67,100

15.00
17.50
23.50

105
75
250
20
100
325
15
35
35
75

3,935
2,800
15,000
800
3,600
11,700
550
1,300
1,575
4,250

1.00
1. 75
15.25
0.50
2.50
9.00
0.50
3.25
0
1. 25

975

43,500

26.75

500

24,900

28.00

675

30,375

29.50

Main Storage

1060
4575
5730
1470
1990
5275
7600
2210
3580
5591
1414

Rental

Processing Unit

1401

Processing Unit (includes 4,000 positlons of
core storage):
Card System - Model A3
Card System - Model G3
Card System - Model G13
Card System - Model H3
Expanded Card System - Model B3
729 Tape/Card System - Model C3
7330 Tape System - Model 013
729 Tape ~stem - Model 03
7330 Tape Card System - Model E3
RAMAC/Card System - Model F3
RAMAC/Card/7330 Tape - Model F23
RAMAC/Card/729 Tape - Model F13
Corrections to above prices for other
storTo sizes:
I, 00 characters (includes core
storage)
2,000 characters (includes core
storage)
4, 000 characters (includes core
storage)
8,000 characters(l)
12,000 characters(l)
16,000 characters(l)

ATTACHMENTS,
ADAPTERS
AND
CHANNELS

PRICES

Core Storage Module:
Model 1 - 4, 000 characters
Model 2 - 8,000 characters
Model 3 - 12, 000 characters
Optional Features:
Advanced Programming
High-Low-Equal Compare
Processing Overlap
Bit Test
Column Binary
Multiply- Divide
Sense Switches
Compressed Tape (1060 required)
800 char/inch Tape Adapter
Read-Compare (for 5590 feature on
1404 Printer)
Attachments
Input/Output Synchronizer
Model 1 (controls up to 10 729 Magnetic
Tape units)
Model 2 (controls up to 10 7330 Magnetic
Tape units)
Model 3 (controls punched card devices)

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

IBM 1401

401 :221.102

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

Name

PRICES
Monthly
Monthly
Rental jPurchase Maint.

$

$

$
Attachnwnts (Contd.)

ATT ACIIMENTS.
AnAPTl-:l~S,

AND
CHANNELS
«('ontd. )
6025
7864
7871
7875
70S0

71GG

7814

7804
7805
7806
7807
7808
:lTASS
STOHAGE

3281
6396
6400
8011
1405

1008
1009
3327
5620

7576

3470

2/69

700

39,900

:12.75

1,125

35,700

23.75

600

49,500

27.50

17b

30,500
8,750

20.75
2.00

500

30,500

10.25

110
125
100

6,750
7,750
:3,750

3.25
3.25
1.50

30
50
65
80
45

1,050
1,550
2,050
2,600

6.50
12.25
lS.50
24.50
NC

0
45
45
45

0
0
2,250
2,250
2,250

0
0
0
0
0

360

16,510

45.50

385

17,610

46.50

50
35

2,400
1,680

3.25
.50

40
40

1,950
1.920

1. 75
.50

965
1,515

36,000
48,500

157.00
170.00

400
425
355
40
400

14,750
15,500
10,850
1,725
13,950

56.75
56.75
42.50
1. 25
41. 25

15

600

1. 00

325

13,000

8.25

a

Disk Storage

1311

3326

Model 4 (controls pillched card and
com munications units)
Model 5 (controls commlllications
devices)
Model 7 (controls up to 10 729 Magnetic
Tape units)
Model 8 (for 1403 only)
Read, Punch Column Binary (on
Model 4 only)
Telegraph I/O Feature (for 1414
Models 4 and 5)
Telegraph Input Feature (requires 78(4)
Telcgraph Output Feature (requires 78(4)
Serial I/o Adapter (required for 1009; 1011,
1012, 1231, 1285, 1445, 1412, 1418,
1410, or 7641; only one of thesc can
be connected to a system)
Switch Control Console:
Model 1 - For up to 2 tape units
Model 2 - For up to 4 tape units
Model 3 - For up to 6 tape Illits
Model 4 - For up to 8 tape units
Tape intermix (on 1414 Modell only; to
intermix 729 Irs, 729 IV's and 7330's
in any combination).
Tape intermix Units:
729 IVs with 729 II/Vs
729 II/Vs with 729 IVs
73309 with 729 II/Vs
7330s with 729 IVs
7330s with 729 IIs/IVs/Vs (require
7804 or 7805)

Disk Storage Drive Model 2 (additional on
system)
Model 4 (first on system, includes
3339 adapter)
Direct Seek (Model 4 only)
Scan Disk (Model 4 only; 4575 feature
required on 1401)
Seek Overlap
Track Hecord (Model 4 only)
Disk Storage Unit:
Modell (10,000,000 characters)
Model 2 (20,000,000 characters)
Additional Access Arm:
For 1405 Modell
For 1405 Model 2
Disk Storage Control
Priority Feature
Disk Storage Control (on first 1405
on each channel)
Successive Disk Storage (required
for each 1405 Disk unit after the
first one each channel)
Dual Synchronizer Adapter

fA

AUfRBACH
~

"01:221.103

IRICK DATA

PRICES

IDENTITY OF UNIT

CLASS

Model
Number

IMonthly

Feature
Number

Name

Rental iPurchasE

Monthly
Maint.

$

$

$

700
900
750
950
450
1,050
2,500

36,000
41,250
37,200
42 450
22:000
55,000
125,000

103.00
114.00
108."
119.00
58.50
98.00
65.59

550

30,000

120.00

400

28,000

109.00

380

27,000

107.00

10
60
80
25

215
3,810
2,985
950

0
31.75
5.75
0.50

500
465

20,200
17,950

56.00
60.00

725
775
550
600
400

32,900
34,000
31,400
32,500
29,000

166.00
177.00
134.00
142.00
131.00

1,550
75
20
375
175
75

75,000
3,125
750
12,600
9,750
60

372.00
0
0.50
24.25
23.00
0

2,000
15
250
50

91,400
600
9,750
2,250

171.00
2.25
17.75
2.00

40
65

1,950
3,000

1.75
2.75

2,275
250
50

110,500
9,750
2,250

229.00
17.75
2.00

20
40

1,225
1,950

0.75
1.75

Magnetic Tape

INPUTOUTPUT

Magnetic Tape Unit
Model n (15/4:1.7 kc)
Model IV (22.5/12.5 kc)
Model V (15/41.7/60 ItC)
Model VI (22.5/62.5/90 kc)
Magnetic Tape Unit
Model 2 Hypenape Drive
Hypertape Contn>l Unit

729

7330
7340
7641

Punched Card
1402

3550
4150,1013
5890,5895
6040

Card Read/Punch:
Modell (Read 800 cards/min; punch 250
cards/min)
Model 4,6 (Read 450 cards/min; punch 250
cards/min)
Model 5 (Read 450 cards/min; punch 250
cards/min)
Early Card Read
51-Column Feed
Punch Feed Read
Read Punch Release
Paper Tape
Paper Tape Reader (7080 adapter required)
Tape Punch

1011
1012

Printers
1403

1404
3740
3835
5585
5990
7246

Printer:
Modell (600 lines/min, 100 positions)
Model 2 (600 lines/min, 132 positions)
Model 4 (465 lines/min, 100 positions)
Model 5 (465 llnes/mtn, 132 positions)
Model 6 (340 lines/min, 120 positions;
Includes 5540 control)
Printer (Includes 5539 and 5563 contr()ls)(2)
Interchangeable Chain
Expanded Print Edit
Print Storage
Read-Compare
Space Suppression
QEtlcal and Mal'eetic Character Readers

1412
2385
3610
5215
7061
7062
1419
3610
5201
7440
7061
7062

Magnetic Character Reader
Document Counter
Electronic Accumulation (5215 required)
Multiple Column Select
Self-Cbecking Numbers:
Modulus 10
Modulus 11
Magnetic Character Reader
Electronic Accumulation (requires 5201)
Multiply Colmnn Control
Split Field
Self-Checking Numbers:
Modulus 10
Modulus 11

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

IBM 1401

401 :221.104

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number
3791

INPUTOUTPUT
(Cont. The annotation characters are variable in size and orientation;
the characters available are 0-9, A-Z, and - +. , * / = ( ) $ '.
The subroutines are
written in Assembler Language and can be called from either an Assembler-coded or FORTRANcoded program.
1130 Scientific Subroutine Package (SSP /1130)
The set of scientific subroutines called Mathpack has been replaced with a more comprehensive
set called the 1130 Scientific Subroutines Package (SSP/1130). The SSP/1130 contains all of the
computational subroutines of SSP/360, a total of 121. All subroutines are free of input-output
statements and are coded in FORTRAN. These subroutines require an 1130 system with a
minimum of 8,192 words of core storage. Some of the subroutines require the overlay facilities
of the 1130 Disk Monitor. All of the others can be used on either a card- or disk-oriented system.

e> 1967 AUERBACH Corporation and AUERBACH Info. Inc.

4/67

418:011. 100

1. "'.....

~'EDP

AUEA8AC~

....-

IBM 1130
INTRODUCTION

""'"

INTRODUCTION
.1

SUMMARY
The IBM 1130 is a desk-size, word-oriented computer intended primarily for smallscale scientific applications. It can also serve as a low-cost processor for certain
business applications that do not require high input-output speeds. The 1130 system is
designed to be easy for scientists and engineers to program (in FORTRAN) and operate,
so that it will be suitable for use instead of, or as an adjunct to, a larger centralized
scientific computing facility. System rentals vary from approximately $680 per month
with minimum storage and input-output equipment to over $1, 800 per month with disc
storage and a full complement of input-output equipment. IBM announced the 1130
system in February 1965, and initial customer deliveries were made in November 1965 .

•2

HARDWARE

The IBM 1130 system offers a choice of four processor models. The models differ only
in respect to core and disc storage capacities, as described below:
•

1131 Model lA Central Processor - 4,096 core memory locations.

•

1131 Model IB Central Processor - 8,192 core memory locations.

•

1131 Model 2A Central Processor - 4, 096 core memory locations
and 512,000 disc storage locations.

•

1131 Model 2B Central Processor - 8, 192 core memory locations
and 512,000 disc storage locations.

Core storage and the Disk File drive (in Models 2A and 2B) are housed in the desk console
which is an integral part of the 1131 Central Processing Unit. Core memory access time
is 3.6 microseconds for each access of one 16-bit word.
The basic word length is 16 bits. The high-order bit position of the word is considered a
sign bit during arithmetic operations. All processor models are essentially single-address,
fixed word-length, binary processors. The instruction repertoire consists of 29 instructions, most of which permit indirect addressing and indexing. Fixed-point binary addition
and subtraction in both single-word and double-word precision are provided. Singleprecision, fixed-point multiply and divide are also standard instructions. No hardware
facilities are provided for decimal or floating-point arithmetic. but extensive floatingpoint operations are possible through standard subroutines.
Excluding shift operations, all instructions can be written in either a "short" (one-word)
or "long" (two-word) instruction format. The short instruction format allows indexing but
not indirect addressing, and has an 8-bit address field which is added to one of three index
registers or the instruction counter to produce a final memory address. The long format
has a 16-bit address field, which permits direct addressing of the full range of core storage.
The Central Processor provides six interrupt levels. Interrupts are generated only by
peripheral devices, and an interrupt is generally initiated for each character transferred
between the device and the processor. The Disk File, however, generates an interrupt
only upon completion of an entire operation.
The removable-cartridge Disk File included in the Model 2A and 2B Ceptral Processing
Units stores up to 512,000 words on a single disc, in sectors of 320 words each. Up to
one full sector can be transferred in one read or write operation. The sectors are numbered sequentially, and up to 8 sectors (2, 560 words, or one "cylinder") can be accessed
without repositioning the access arm. The average total access time to randomly-placed
data is 790 milliseconds. Only one Disk File drive can be included in an 1130 system, but
unlimited off-line storage is possible with the use of additional disc cartridges.
The input-output devices that can be connected to an 1131 Central Processing Unit are
limited in range and number. Only one of each of the following devices can be connected
to a central processor (any model) :
•

1134 Paper Tape Reader - reads at 60 characters per second.

•

1055 Paper Tape Punch - punches at 14.5 characters per second.
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.

5/66

IBM 1130

418:011. 200

.2

HARDWARE (Contd.)
•

1442 Card Read Punch Model 6: reads a maximum of 300 cards per minute;
punches a maximum of 80 columns per second.
Model 7: reads a maximum of 400 cards per minute;
punches a maximum of 160 columns per second.

•

1132 Printer - prints at up to 82 lines per minute for alphanumeric output and up to 110 lines per minute for numeric
output.
1627 PiAtter Modell: plots at a maximum rate of 18,000 steps per minute.
Model 2: plots at a maximum rate of 12,000 steps per
minute; accepts paper of greater width than Modell.

• •

The various peripheral devices use several different data codes, and no automatic translation is performed during either input or output operations. IBM provides an extensive
array of subroutines to perform conversions between the various data codes and between
decimal and binary radices.
The interrupt structure and I/O logic of the 1130 permit overlapping of internal processing
with one or more I/O data transfers. The standard I/O control subroutines provided for
the 1130 Assembler allow the user to take advantage of this capability for simultaneous
operations. The FORTRAN I/O subroutines, however, permit no overlapping of non-I/O
processing with I/O operations, nor do they permit simultaneous operation of two peripheral devices. Since most programming for the 1130 will probably be done in FORTRAN,
the 1130 will appear to most users as a sequential system capable of only one operation
at a time.
One potentially interesting use of an IBM 1130 is as a remote terminal connected by a
communications link to a larger. central computer facility. IBM states that an adapter
which allows an 1130 to be connected to a communications line is available on an RPQ
basis, but complete details have not been released as yet .
•3

SOFTWARE
Three software packages are provided by IBM for use with the IBM 1130 system. Two are
quite similar and are intended for use in systems with punched card or punched tape inputoutput but without the Disk File. The other software package is the 1130 Monitor System
for use with disc-oriented systems. The Monitor System reduces the need for operator
intervention by providing automatic handling of run-to-run supervision. The punched card
and tape program packages became available in March 1966; the Monitor System, in
April 1966.
The punched card and tape program packages include the following facilities: an 1130
Assembler; a FORTRAN compiler; a set of utility routines which provide facilities for data
transcriptions, memory dumps, and loading programs; and an 1130 Subroutine Library,
including subroutines for extensive floating-point arithmetic procedures, code conversions,
and input-output control. The 1130 Monitor System includes a supervisor routine and a disc
utility program in addition to the facilities offered by the punched card and tape packages.
The IBM 1130 Assembler is a straightforward one-to-one assembler that provides for
symbolic representation of the central processor instructions. The assembler includes
pseudo-instructions for calling IBM or user-coded input-output subroutines and for
reserving core memory areas. Source programs can be assembled in either absolute
(core image) or relocatable format.
The 1130 FORTRAN language is a subset of the IBM System/360 Basic Programming
Support FORTRAN IV language. No facilities are provided for COMPLEX; DOUBLE
PRECISION, or LOGICAL operations, but the available language facilities will be
adequate for most small-scale scientific applications.
In addition, IBM provides applications packages to aid in solving petroleum exploration
and engineering problems, civil engineering problems (COGO), and to aid in type
composition. More general routines include statistical and numerical surface routines
and additional FORTRAN subroutines (Mathpak). In general, these programs require
a disc-oriented system.

5/66

A

AUERBACH

'"

418:221. 101

~

AUERBACH

STANDARD

EDP

IBM 1130

REPORTS

PRICE DATA

PRICE DATA

PRICES

IDENTITY OF UNIT
CLASS
Name

No.

Monthly
Rental

Monthly
Maintenance

$
CENTRAL
PROCESSOR
AND
INTERNAL
STORAGE

1131

Central Processing Unit, including
Console Typewriter/Printer, core
storage, and Disk Storage (if any):

1A

4, 096 words of core storage; no
Disk storage
8, 192 words of core storage; no
Disk storage
4, 096 words of core storage;
512,000 words of Disk Storage
(removable cartridge)
8, 192 words of core storage;
512,000 words of Disk Storage
(removable cartridge)

1B
2A
2B

2315

Disk Cartridge
PeriEheral Unit Attachments
Storage Access Channel
1132 Printer Attachment
Printer Expansion Adapter
1442 Attachment
1627 Plotter Attachment
1134 Paper Tape Reader Attachment
1134 Loader
1055 Paper Tape Punch Attachment

7490
3616
3854
4454
7187
3623
3624
7923
INPUTOUTPUT

1132
1442

Printer (requires 3616 and 3854)
Card Read Punch (requires 4454):
Model 6
Model 7
Paper Tape Reader (requires 3623):
Modell
Model 2 (includes supply and
take-up reels)
Paper Tape Punch, Modell
(requires 7923)
1055 Edge-Punching Feature
Take-Up Reel
Plotter (requires 7187):
Modell
Model 2

1134

1055
3571
6121
1627

$

Purchase

$

580

50.50

26,680

780

52.00

35,080

780

68.50

35,680

980

70.00

44,080

-

-

90

25
10
5
35
15
10
10
20

0.50
2.00
NC
3.00
0.50
1. 75
NC
1.50

1,125
450
225
1,575
675
450
450
900

260

25.00

11,700

265
380

40.00
50.00

14,575
15,725

35
60

10.00
10.00

1,310
2,260

40

6.50

2,025

5
3

0.25
0.25

245
120

36.00
38.50

4,700
8,150

-

NC - No Charge

© 1966 AUERBACH Corporation and AUERBACH Info, Inc.

5/66

418:221. 102

IBM 1130

IDENTITY OF UNIT
CLASS

No.

PRICES
Monthly
Rental

Name

Monthly
Maintenance

$
CENTRAL
PROCESSOR
AND
INTERNAL
STORAGE
(Contd. )

INPUTOUTPUT

$

25
40

0.50

1,125
1,800

165
45
40
45
40
450
500
25

10.00
2.00
3.00
2.00
3.00
30.00
35.00
0.50

7,425
2,025
1,800
2,025
1,800
20,250
22,500
1,125

232

11. 25

9,825

255

49.00

12,750

265

51. 00

14,575

380

61. 00

15,725

Peripheral Unit Attachments (Contd. )
7490
1133
1865
3201
3202
3203
3204
4424
4425
7492
7690
1442

3630
2501

3630
1231
1264
5045
1134

1055
3571
6121

Storage Access Channel
Multiplexor Channel Enclosure
(not with Processor Model 1A or
1B; requires 7490)
Channel Multiplexor
Disk Control 1
Disk Control 2
Disk Control 3
Disk Control 4
1403 Model 6 Attachment
1403 Model 7 Attachment
Storage Access Channel II
Synchronous Communications
Adapter
Card Read/Punch Model 5 (requires 4449 and 3630):
punches 160 col/sec
Model 6 (requires 4454): reads 300
cards/min; punches 80 col/sec
Model 7 (requires 4454): reads 400
cards/min; punches 160 col/sec
1130/1442 Model 5 Coupling
Card Reader (requires 8042, 3854,
and 3630) Model AI: reads 600 cards/min
Model A2: reads 1,000 cards/min
1130/2501 Coupling
Optical Mark Page Reader
(requires 8034, 3854, and 1264)
Asynchronous Mode
Master Mark

1627

Plotter Modell (requires 7187): plots up to
300 increments/sec
Model 2 (requires 7189): plots up to
200 increments/sec

1248
5858

Printer Model 6 (requires 1133 and 4424):
prints 340 lines/min
Model 7 (requires 1133 and 4425):
prints 600 lines/min
Display Unit Model 4 (requires
7490 or 7492)
Alphameric Keyboard
Program Function Keyboard

A

AUERBACH

-

-

225

201
263
5

33.25
45.75

430

36.25

23,100

40
50

4.00
1. 50

1,800
2,500

36
62

10.00
10.50

1,270
2,190

40

6.50

2,025

5
3

0.25
0.25

245
120

268

25.00

11,350

-

36.00

4,700

-

38.50

8,150

400

131. 00

29,000

650

153.00

32,700

2,400

110.00

115,200

75
150

2.00
2.50

3,600
7,200

Paper Tape Punch (requires 7923):
punches 14.8 char/sec
Edge-Punching
Take-up Reel
Printer (requires 3616 and 3854):
prints 80 to 110 lines/min

2250

5

Paper Tape Reader (requires 3623) Modell: reads 60 char/sec
Model 2 (includes supply and take-up
reels): reads 60 char/sec

1132

1403

10/67

Purchase

$

-

11,350
11,590
150

-~

420:000.000
STU"IO

~EDP

IBM SYSTEM/360

AUERBACH

MODEL 195
REPORT UPDATE

-

REPUIrS

8

REPORT UPDATE
•

mM ANNOUNCES SUPER-SCALE MODEL 195
mM introduced a new addition to the top end of its System/360 family line on August 20. The
super, large-scale Model 195, designed to compete directly with the recently-announced CDC
7600 computer system, replaces the discontinued Model 91 and 95 and offers the following
features:
•

High speed instruction execution - 18-nanosecond floating point multiply-divide cycle.

•

High speed 54-nanosecond buffer - 32, 768 bytes of storage.

•

Eight to sixteen way interleaving in main storage.

•

One to four million bytes of 756-nanosecond main storage.

•

104-microsecond timer

•

Universal instruction set, direct control, extended-precision, floating-point arithmetic.

•

Graphic display console

The Model 195 is essentially a faster Model 85 and it can utilize all of the peripheral devices
available in a Model 85 configuration with three exceptions; the 2520 Model Bl Card Read/Punch,
and the 7770 Model 3 and 7772 Model 3 Audio Response Units.
The Model 195 has five functionally separate CPU units: processor storage; storage bus control;
instruction processor; fixed point processor, and floating point processor. This processor organization permits up to seven different operations to proceed simultaneously.
Operating support is provided by OS MVT, available without charge. In a multiprogrammed environment, up to fifteen programs can be run concurrently. A maximum of six selector channels and one multiplexor channel can be used with the Model 195.
The following Model 195 prices are available at this time:
Monthly
Rental
2191 Processing Unit
Model J (1, 048, 576 bytes
of main memory)
Model K (2,097, 152 bytes
of main memory)
Model L (4,194,304 bytes
of main memory)

Purchase
Price

Monthly
Maintenance

$ 96,800

$ 4,266, 000

$12,735

127,800

5,766,000

13,485

167,800

7,766,000

14,985

7,200

308,000

450

2180 Power Unit

850

37,400

15

2185 Power Distribution Unit

700

30,800

10

2186 Coolant Distribution Unit

750

33,000

10

2160 Systems Console
(graphic displll¥ and light
pen included)

Monthly system rentals for the Model 195 range between $165, 000 and $300, 000 with purchase
prices from $7 million to $12 million, depending on the configuration. Initial delivery of the
Model 195 is scheduled for the first quarter of 1971.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

9/69

-&.

420:000.001

su,,'"

/A\\EDP

-

AUERBACH

s

IBM SYSTEM/360
REPORT UPDATE

UPDns

REPORT UPDATE
~ IBM UNVEILS KEYBOARD-TO-TAPE INSCRIBER AND TAPE CARTRIDGE READER
Companion units that enable information to be keyed directly onto magnetic tape and then read
into a System/360 computer were announced by IBM on April 25. The new units are the IBM
50 Magnetic Data Inscriber and the 2495 Tape Cartridge Reader.
IBM states that the new data entry system is "designed primarily for users who must continually
update records already stored in a computer in random order. In many other data processing
applications, the preparation of a punched card - a physical unit record of information - will
continue to be advantageous." Thus, IBM is responding to the Mohawk and Honeywell keyboardto-tape inscribers.
Magnetic Tape Cartridges
The recording medium used in the two new units is 16-millimeter (0. 64-inch-wide) magnetic
tape, housed in cartridges that can hold a maximum of 23, 000 characters of information.
The cartridges are identical with the ones that have been used for several years in the IBM
Magnetic Tape Selectric Typewriter (MT/ST). Information is recorded in nine tracks across
the width of the tape, using System/360-compatible code.
It is important to note that the tape cartridges recorded by an IBM 50 Magnetic Data Inscriber
can, at present, be read only into a System/360 computer and only by an IBM 2495 Tape
Cartridge Reader. This is the most significant difference between the IBM inscriber and the
competitive units manufactured by Mohawk, Honeywell, and Sangamo, which record data on
standard o. 5-inch, 7- or 9-track tape that can be read by any computer equipped with "IBMcompatible" magnetic tape units.
IBM 50 Magnetic Data Inscriber
The IBM inscriber features a variable-record-length capability that eliminates the 80-character
limitation that punched cards impose on unit record lengths. The operator can preset the
arrangement of data to be keyed in and indicate where repetitive data shall be entered automatically. Up to eight different formats can be established, and subsequently selected, via the
keyboard. Automatic data entries are recorded at the rate of 100 characters per second.
The 50 Magnetic Data Inscriber has a keypunch-style keyboard and can handle both numeric
and alphabetic information. An automatic "left-zero insertion" feature allows the operator to
enter only~ignificant digits; preceding zeros are inserted automatically. The console provides
a charac(er display, and there are special provisions to facilitate the correction of keying errors.
IBM 2495 Tape Cartridge Reader
The 2495 reader feeds information recol'cied on tape cartridges into a System/360 computer at
the rate of 900 characters per second. An entire cartridge, holding 23, 000 characters, can be
read in less than 30 seconds. A dozen tape cartridges can be loaded into the reader at the
same time, whereupon they are read, rewouni:4 and stacked automatically.
The 2495 Tape Cartridge Reader can be connected to the Multiplexor Channel of a System/360
Model 25, 30, 40, or 50 computer system. It can read tape cartridges recorded by an IBM
Magnetic Tape Selectric Typewriter, as well as those recorded by IBM 50 Magnetic Data
Inscribers. The MTST is especially well-suited for entry of unformatted text data into a computer. A single 2495 reader can, of course, serve multiple data inscribers and MTST's.
Prices and Availability
The IBM 50 Magnetic Data Inscriber rents for $180 per month and can be purchased for $9, 900.
The 2495 Tape Cartridge Reader has a basic monthly :r\ental of $350 and a purchase price of
$19,250. Customer deliveries of both units are scheduled to begin in the first quarter of 1969.
Each tape cartridge costs from $12 to $20, depending upon the quantity purchased.

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

5/68

420:000.002
IBM SYSTEM/360
REPORT UPDATING

Sill .....

EDP
UPiITS

REPORT UPDATE
~ IBl\l ADDS NEW OPTIC:\ I RECOG~ITION DEVICE TO ITS PIW\H1CT UNE

The 1288 Optical Page Ih'ader, announced by IB:ll on .July 13, reads data from eIther formatted
or unformatted documents. Typed or machine-printed data can bl' I'cad and stOl'l'd at a maximum
speed of 2,750,000 characters per hour, or the equl\'alent of ~l(iO tully-typed pages. The 12~~ is
an on-line deviee that ean be attached to System/360 :\Iodels :.!3, :lO, 40, or 50 and iH supportt'd
by both the 360 Disk Operating System and the 360 Tape Operating System.
The input documents may range in size from 3" x G. 5" to!l" x J.1". An input stacker accomodates up to ten inches of documents and two output stackers cach have a 4.5 inch capacity.
Data can be read both vertically and horizontally in the formatted mode. Format control allows
reading variable length fields in any sequence. Thl' stol'l'd prog-ram directs the clectronic flying
spot optical scanner beam, a field at a time, to the data to IJl' read. For unformatted data, the
1288 locates the first hne, reads and transfers it, and then s('arches for eaeh subsequent line.
Features
Handprinted numeric digits 0-9 and alphabetIc characters C, S, T, X, and Z may be read with
the Numeric Handwriting special feature. Character shapes and spacing must conform to the
basic rules of handwriting as specified by IB:lI, such as the requirements for large Simple block
printing and connecting lines and closing loops, and the restriction against Imking characters.

A special feature allows marks, that are either handmarked or machine pnntl'd, to be read
vertically or with a slanted orientation at an angle of 43°.
Price and Availability
The 1288 Optical Page Reader rents for ;i:4, 900 per month and can be purchased for $230,300.
Special features =# 5370 :\umeric Handwrltl11g and =# 5479 Optical :\lark Reading carry monthly
rentals of $1,000 and $100 and purchase prIces of $34,000 and $3,300 respectiwly. The first
customer deliveries are scheduled for January 1970.

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

11/68

-

~ "'

420:000.003
IBM SYSTEM/360
REPORT UPDATE

.....

I~EDP

-

AUER8AC~
~

","ts

REPORT UPDATE
~ lBl\1 MAKES AVAILABLE ADDITIONAL conE STonACa: CHOICES
Two new core sizes for the IBl\1 System/360 Models 40 and 50 were announced on November 4.
The Model 40 processor stora!-(e capacities range from }{;, :IH4 to 524, 2HH bytes and the Model 50
spans a 65,5:16 to 1,048,576 ran!-(e.
The prices and core

stora~e

Processor Model
Numbers
40 G F (196,608 bytes)
50 JIG (393,216 bytes)

positions of the two new processor models art' aH follows:
l\lonthly
Hental

$ H,750
17, GSO

Monthly
Maintenance Char~es
$ :.!50
550

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

Purchase
Price
$:381,610
756,BAO

11/68

......

£

420:000.004
IBM SYSTEM/360

Ill .....

~EDP

AUI~

•

REPORT UPDATE

I'"ITS

REPORT UPDATE
~ IBM ADDS ADDITIONAL FEATURES TO TIlE SYSTEM/:l60 MODEL 25
In November 1968, IBM made available fOllr new features that expand the capabilities of the Model
25. The Model 25 price and performance falls between the System/360 Models 20 and 30. Unlike
the Model 20, the Model 25 is program-compatible with the larger general-purpose System/360
processors, and when equipped with the appropriate compatibility features, it can execute programs
written for the IBM 1401, 1440, and 1460 systems. The introduction of the following features enabled
IBM to increase the compatibility attributes and add other peripheral devices to the Model 25:

•

The 1401/1440/1460 DOS Corn atibilit # 4470 - requires the prior attachment of the
1400 Series Compatibility Feature ( 4440) which allows the Model 25 to execute 1401/
1440/1460 instructions in 16K of control storage on the processor Model 2025. The
~4470 feature specifically permits 1401 and System/360 programs to be run in a single
intermixed jobstream in a multiprogramming environment. There is no additional
charge for the # 4470 feature.

•

The 2540 Emulation Control # 7800 - requires the prior attachment of the Integrated
2560 Attachment ( 4596) which, in conjunction with the # 9725 Model 25 Adapter,
permits the attachment of the 2560 l\1ulti-function Card Machine to the System/360 Model
25 system. Through the =# 7800 feature attachment, the I\lodel 25 user can emulate
a 2540 Card Read/Punch via the 2560 Card Machine. The # 7800 feature carries a
monthly rental of $60 and a purchase price of $3,048.

•

The 1100 LPl\1 Printer Adapter (-# 3615) - permits the attachment of the 1403 Printer
Model N1 to the 2025 processor via the Integrated 1403 Attachment (# 4590). The
# 3615 feature rents for $60 a month and sells for $3,048.
The Multiple Character Set (# 5111) - is used on the 1403 Printer N1 attached to the 2025
processor via the Integrated 1403 Attachment (# 4590) and the # 3615 Adapter. This feature
rents for $10 per month and sells for $430. The -j 5111 feature cannot be installed with
the Universal Character Set Feature (# 8641).

In addition to the above features, IBM has announced that the maximum data transfer rate for the
Selector Channel (# 6960), which previously did not permit the attachment of devices that exceed a
30KB rate, has been increased to 60KB. Thus, It IS now possible to attach the 2401 Tape Drive
Models 2 and 4.

420:000.005

A;;;;'p

AUERBAC~

•

IBM SYSTEM/360
REPORT UPDATE

IIEPntTS

REPORT UPDATE
~ IBM

ADDS NEW MAGNETIC TAPE UNIT TO ITS SYSTEM/3BO PRODUCT LINE

IBM has extended the available peripheral equipment options for users of the System/360 Modes
50, 65, 75 or 85, with the introduction of the 2420 Magnetic Tape Unit, Model 7.
The 9-track 2420-7 is now the fastest tape unit in IBM's product line, operating at a peak speed
of 320,000 bytes per second .•• twice the speed of the 2401-6 tape unit.
The 2420-7 features a 1,600 bytes per second recording density, a rewind speed of 500 inches
per second, and the ability to accelerate from a dead start to 200 inches per second in 1.4
milliseconds and stop within a 1 millisecond time frame.
The unit is currently available at a $1,050 per month lease rate and a $54,600 purchase price.
The user can expect a 4 to 5 month delivery following the placing of an order.

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

12/68

420:000.006
IBM SYSTEM/360
REPORT UPDATE

REPORT UPDATE
• IBM Provides New Disc Storage Models
Two new models of the IBM 2314 Direct Access storage Facility, the 2314-Al and the 2314-A2,
were announced in January 1969 to replace the original Model 2314-1, a changeable - cartridge
disc storage unit containing eight active disc drives, one spare drive, and internal control circuitry. Conversion and operations are facilitated by the fact that all of the supporting software is
compatible.
The two new models differ from the model 2314-1 prlmarllv in the faster average access time
--- a decrease from 75 milliseconds to 60 milliseconds. Model Al has eightindependent drives
on-line and one spare drive for use in case one of the other eight becomes inoperable. Total
on-line capacity is 233.4 mlllion bytes. Model A2 contains five independent drives with an
on-line capacity of 145.9 m1l110n bytes. As was true with the old model, the new models require
no external control unit.
The 2314 Model Al leases for $5,675 and sells for $256,400: the Model A2 carries a monthly
rental of $3,875 and a purchase price of $175,075.
The first customer shipments for new orders are expected to commence In September 1969, however the present orders for the 2314-1 wUI be fUled with the 2314-Al, startlng in June 1969.

C 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

422:011. 100

A.

AUERBACH

SUIOARD

EDP

IBM SYSTEM/360
MODEL 20
SUMMARY

REPORrs

SUMMARY
The Model 20 is the smallest currently-announced member of the IBM System/360 computer family. It was introd'lced in November 1964 as a card-oriented computer system designed
primarily to serve as a first step upward from punched-card tabulating equipment. Later, IBM's
addition of magnetic tape capability to the Model 20 expanded the range of its practical applications.
In August 1966, two new models of the IBM 2311 Disk Storage Drive were announced for use with
the Model 20, adding still another dimension to its processing capability.
In March 1968, IBM announced slower, lower-cost versions of the Model 20 system that
make its use practical in smaller businesses and low'er-volume applications. The new "Submodels
3 and 4" of the Model 20 Processor are fully program-compatible with, and somewhat slower than,
the earlier "Submodels 1 and 2." Configuration possibilities for Submodels 3 and 4, however, are
quite restricted; the only allowable peripheral devices are the new, slower models of the 2560
Multi-Function Card Machine and 2203 Printer, plus the 2311 Model 12 Disk Storage Drive, 2152
Printer-Keyboard, and 2074 Binary Synchronous Communications Adapter. Both prices and
throughput of Submodel 3 or 4 systems average about 70 per cent of those of equivalent Submodel
1 or 2 configurations.
In June 1968, IBM further expanded the Model 20 line by introducing five new "Submodel
5" Processors that provide substantially improved performance at rental increases of only $100 per
month over the corresponding Submodel 2 Processors. The Submodel 5 systems offer:

•

Internal processing speeds apprOximately three times as fast as Submodels 1
and 2 on typical instruction mixes.

•

Core storage capacities of 8K, 12K, 16K, 24K, or 32K bytes (compared with
a maximum capacity of 16K bytes in the earlier models).

•

Provision for connecting up to four 2311 Disk Storage Drives (compared with
a maximum of two drives in the earlier models).

•

Capability to overlap magnetic tape and/or disk reading or writing with other
I/O operations and with internal processing.

~5=•

~~

,.

.

"t"

Figure 1. The card-oriented Model 20 system in the foreground, with 2560 Multi-Function Card
Machine at right, is designed to perform all the functions of the conventional tabulating
machines in the background.
@

1968 AUERBACH Corporation and AUERBACH Info, Inc.

7/68

422:011.101

IBM SYSTEM/360 - MODEL 20

SUMMARY (Contd. )
•

Improved serviceability through a new diagnostic aid called "System Log-Out,"
which operates at the micro-code level and provides for automatic storage
of count and status data regarding intermittent errors.

•

Full upward compatibility with the slower Model 20 Processors.

A Model 20 system that includes a Multi-Function Card Machine and printer can be
rented for as little as $1,359 per month, though monthly rentals for most Model 20 systems will
range from about $1,500 to $4,000. Deliveries of Model 20 systems began in the first quarter of
1966. Initial deliveries of Submodel 3 and 4 systems are scheduled for November 1968, while
customer shipments of Submodel 5 systems will start in May 1969.
HARDWARE
The Model 20 uses the same basic data and instruction formats as the larger System/
360 models. The instruction repertoire is a compatible subset of the full System/360 repertoire,
except that the input-output instructions and some control instructions are unique to the Model 20.
Decimal arithmetic (including multiply and divide), editing, and code translation instructions are
standard. Floating-point arithmetic is not available. The scatter-read, gather-write, and extensive interrupt facilities of larger System/360 models are not implemented in the Model 20: interrupts occur only upon completion of peripheral data transfer operations.
A Model 20 Processor can contain from 4,096 to 32,768 bytes of core storage. Table I
summarizes the model designations, core storage capacities, and representative speeds of the 21
different Model 20 Processors that are now offered. The core cycle time is 3.6 microseconds
per one-byte access in Submodels 1 through 4 and 2.0 microseconds per two-byte access in Submodel 5. Every Model 20 Processor has eight 16-bit general registers (compared with sixteen
32-bit registers in the larger System/360 models), which are usable as fixed-point accumulators
or as index registers.
The use of decimal arithmetic is being emphasized
upon 4-bit BCD digits packed two to a byte, with a sign in the
order byte of each data field. Decimal operands can be up to
length; the length of each field is specified in the instructions
a word-mark in the field itself.

in the Model 20. It is performed
rightmost four bits of the low16 bytes (31 digits and sign) in
that reference it, rather than by

Because of the instructions which are unique to the Model 20 and certain other programming differences, programs written for a Model 20 system cannot be directly executed on
a larger System/360 model. A System/360 Model 25 Processor, however, can be equipped with
a Model 20 Compatibility Feature (announced in June 1968) that enables the Model 25 to execute
Model 20 programs at approximately four times the internal speed of the original Model 20
TABLE I: SYSTEM/360 MODEL 20 PROCESSOR CHARACTERISTICS
Core ~'orage
Capacity, nytes

Processing Unit Designation
&ibmodel1

Submodel2

&ibmodel3

&ibmodel4

&ibmodel5

4,096
8,192
12,288

B1
C1
BC1

16,384
24,576
32,768

D1

B2
C2
BC2
02

B3
C3
BC3

B4
C4
BC4

-

D3

D4

-

C5
BC5

-

-

-

-

05
DC5
E5

Cycle Time, Microseconds

3.6

3.6

3.6

3.6

2.0

Bytes Accessed per Cycle

1

1

1

1

2

Cycle Time per Byte,
Microseconds

3.6

3.6

3.6

3.6

1.0

675
7,000
10,810
185
489

675
7,000
10,810
185
489

1,207
7,530
11,340
451
755

1,207
7,530
11,340
451
755

212
1,730
2,130
68
206

Processor Speeds,
Microseconds'
c=a+b
c" a x b
c a .;- b
Move a to b
Compare a to b
.0

'Basis: Signed, packed, 5-digit decimal operands.

7/68

A

(Contd.)

AUERBACH

'"

SUMMARY

422:011.102

HARDWARE (Contd.)
systems (or about 30 per cent faster than the Submodel 5 Processors). This feature also permits
the 2560 Multi-Function Card Machine to be used with the Model 25 System.
A control panel, built into the top of the Model 20 Processor cabinet, provides the
switches, keys, and lights required for manual control of the system. The optional 2152 PrinterKeyboard, announced in January 1968, provides keyboard input and typewriter output facilities.
Peripheral devices are connected to a Model 20 system by means of special attachments. In most cases a separate attachment is required for each peripheral device. Thus, a
Model 20 processor model can be differentiated from other System/360 processors not only by its
core storage size and limited instruction set, but also by its method of connecting I/O devices.
A System/360 Model 20 system can include up to three punched-card input-output units,
one printer, one magnetic character reader, one printer-keyboard, one communications adapter,
one 2415 Magnetic Tape Unit and Control (containing two, four, or six tape drives), and four
2311 Disk Storage Drives. Table II summarizes the principal characteristics of the peripheral
devices available for use in Model 20 systems and shows the processor submodels to which each
device can be connected. The following additional configuration restrictions apply:
•

A system can have a maximum of three card read stations, two card punch
stations, and one card print station.

•

A 2560 Multi-Function Card Machine and a 2520 Card Punch or Card Read
Punch cannot be used in the same system.

•

A 2073 Communications Adapter can be connected only to a Submodel 2
rrocessor.

•

A 2074 Binary Synchronous Communications Adapter can be used only with
a Submodel 2, 4, or 5 Processor having at least 8K bytes of core storage.

•

Disk storage drives can be connected as follows:
Submodel2 Processor - up to two 2311 Drives, Model 11 or 12.
Submodel4 Processor - up to two 2311 Drives, Model 12 only.
Submodel 5 Processor - up to four 2311 Drives, Models 11 and 12
(may be intermixed).

Through interleaving of the core storage accesses required by I/O devices and by the
central processor, one of each of the following functions can occur simultaneously in a Model 20
system: card reading, punching, printing, typing and computing. Magnetic tape or disk read
or write operations, however, occur in the "burst" mode in Submodels 1 through 4, precluding
simultaneous execution of any other function except printing on the buffered 1403 Printer, magnetic tape rewinding, or disk seek operations. In Submodel 5 systems, one tape and one disk
read or write operation can be overlapped with other I/O operations and with computing.
The 2560 Multi-Function Card Machine (MFCM) is a unique punched-card input-output
unit developed especially for the System/360 Model 20. Equipped with two l,200-card input
hoppers, a reading station, a punching station, an optional printing station, and five l,300-card
radial stackers, the 2560 MFCM combines many of the facilities of a card reader, card punch,
collator, interpreter, and card document printer in a single unit under stored-program control.
Cards can be fed independently from either the primary or secondary hopper of the
2560 MFCM; they follow separate paths through pre-read, read, and pre-punch stations. The
cards are read serially (column by column) by means of solar cells, at a maximum speed of 500
cpm in Model Al and 310 cpm in Model A2. Upon leaving the separate primary and secondary
pre-punch stations, the cards merge into a single feed path through the punch, pre-print, and
print stations. Then the cards pass on into any of the five stackers, as selected by the program. The rated punching speed is 160 columns per second in Model Al and 120 columns per
second in Model A2. The effective speed depends upon the position of the last column punched
in each card (as in the mM 1442); when all 80 columns are punched, the punching rate is 91
cpm in Model Al and 65 cpm in Model A2. The optional Card Print feature (for Model Al only)
provides a printing unit that can print two, four, or six lines of information on any or all cards
passing through the MFCM at a rated printing speed of 140 character positions per second.
The IBM 2311 Disk Storage Drive has been adapted for use with the System/360 Model 20.
The new units, called the 2311 Models 11 and 12, are similar to each other in every respect except
storage capacity and access time. Model 11 provides 5.4 million bytes of auxiliary "Disk Pack"
storage with 75-millisecond average pOSitioning time; Model 12 provides 2.7 million bytes of storage with 60-millisecond average positioning time. Both models have an average rotational delay
of 12.5 milliseconds, a data transfer rate of 156,000 bytes per second, and a fixed sector length
of 270 bytes, with 10 sectors in each track.
The 2073 Communications Adapter provides the Model 20 (Submodel 2 only) with limited
data communications facilities. With this adapter, a Model 20 can function as a Single-line,
point-to-point processor terminal that can communicate with another Model 20, a larger System/
@

1968 AUERBACH Corporation and AUERBACH Info. Inc.

7/68

,422:011.103

IBM SYSTEM/360 - MODEL 20

TABLE II: SYSTEM/360 MODEL 20 PERIPHERAL EQUIPMENT

I

Device
2560 Multi-Function Card Machine:
Model Al

Usable with
Processor Submode1s

Characteristics
Reads 500 cards/minute;
Punches 160 columns/second;
Prints 140 char/second on cards (optional)

1.2. 5

Reads 310 cards/minute;
Punches 120 columns/second

3.4

2501 Card Reader:
Model Al
Model A2

Reads 600 cards/minute
Reads 1000 cards/minute

1. 2, 5
I, 2, 5

2520 Model Al Card Read Punch

Reads 500 cards/minute;
Punches 500 cards minute

I, 2, 5

2520 Card Punch:
Model A2
Model A3

Punches 500 cards/minute
Punches 300 cards/minute

I, 2, 5
I, 2, 5

1442 Model 5 Card Punch

Punches 160 columns/second

I, 2, 5

1403 Printer:
Model 2
Model 7
Model N1

Prints 600 lines/minute; 132 positions
Prints 600 lines/minute; 120 positions
Prints 1100 lines/minute; 132 positions

I, 2. 5
I, 2, 5
I, 2, 5

Prints
Prints
Prints
Prints

750
425
350
300

lpm
lpm
lpm
lpm

with
with
with
with

13-character
39-character
52-character
63-character

set;
set;
set;
set

I, 2, 5

Prints
Prints
Prints
Prints

600
300
260
230

lpm
lpm
lpm
lpm

with
with
with
with

13-character
39-character
52-character
63-character

set;
set;
set;
set

3.4

Model A2

2203 Printer:
Model Al

Model A2

1259 Magnetic Character Reader

Reads 600 6-inch documents/minute;
11 sorting pockets

2.5

1419 Magnetic Character Reader

Reads 1600 6-inch documents/minute;
13 sorting pockets

2, 5

Stores 5.4 m1l1ion bytes on-line;
87.5 msec average random access time

2. 5

Stores 2.7 m1l1ion bytes on-line;
72.5 msec average random access time

2,4,5

15,000 char/sec; 800 bpi
15.000 char/sec; 800 bpi
15.000 char/sec; 800 bpi

2.5
2.5
2,5

30,000 char/sec; 1600 bpi
30.000 char/sec; 1600 bpi
30.000 char/sec; 1600 bpi

2,5
2,5
2,5

Selectric Typewriter; prints 15.5 char/sec

2,4.5

2311 Disk Storage Drive:
Model 11
Model 12
2415 Magnetic Tape Unit:
Model 1 (two tape drives)
Model 2 (four tape drives)
Model 3 (six tape drives)
Model 4 (two tape drives)
Model 5 (four tape drives)
Model 6 (six tape drives)
2152 Printer-Keyboard

HARDWARE (Contd. )
360 processor, an IBM 1009 Data Transmission Unit, a 1013 Card Transmission Terminal, a 7701
or 7702 Magnetic Tape Transmission Terminal, or a 7710 or 7711 Data Communications Unit.
Data is transmitted and received in half-duplex synchronous mode over appropriate communications facUities at speeds r~ing from 75 to 600 characters per second. Multiple remote terminals
can be addressed via common-carrier switched telephone networks. Automatic connections and
disconnections can be made without operator intervention. Data is transmitted or received Under
control of the stored program. One message at a time can be transmitted or received. The Communications Adapter shares central processor time with data processing and input-output operations.
As an alternative to the 2073 Adapter, IBM offers the 2074 Binary Synchronous Communications Adapter (BSCA) as an optional feature for the SUbmodel 2, 4, or 5 Processor. This
unit permits a Model 20 system to function as a processor terminal on either a switch~d or
leased communications network, communicating with another appropriately-equipped System/360

7/68

A..

AUERBACH

IContd.)

SUMMARY

422:011.104

HARDWARE (Contd.)
processor in ffiM's Binary Synchronous Communications (BSC) mode. With the 2074 Adapter,
transmission speeds ranging from 75 to 6250 characters per second are possible, depending upon
the communications facilities. Data can be transmitted in EBCDIC, USASCII, or binary form.
Transmission is in half-duplex mode over privately-owned, leased, or switched telephone lines.
SOFTWARE
Because of the Model 20's restricted instruction repertoire, different I/o control methods, and limited core storage capacity, Model 20 systems cannot use the extensive array of software that ffiM is providing for the larger System/360 models. Software support for Model 20
systems is provided at three different levels, for card-, tape-, and disk-oriented systems.
The card-oriented software requires a 4K Model 20 Processor with punched-card I/o
equipment and a printer. The software components at this level include a basic assembler, report program generator, input-output control system (IOCS), and a group of utility routines.
Among the latter are four programs that use the 2560 MFCM to simulate many of the functions of
conventional tabulating equipment: Collate, Gangpunch-Reproduce, List-Summary Punch, and
Merge-Sort. (A card reader and punch can be substituted for the 2560 MFCM in the GangpunchReproduce and List-Summary Punch programs.) No operating system is available at the cardoriented level; operator intervention is required between programs.
The tape-oriented software provides expanded versions of the assembler, report program generator, IOCS, and utility routines, plus a Sort/Merge program. Moreover, a group of
three System Control programs (Initial Program Loader, Basic Monitor, and Job Control) provide
increased operational efficiency through automatic job-to-job transition, overlay control, and
use of a tape library. A Model 20 Processor with 8K bytes of core storage, four tape drives,
card I/O equipment, and a printer is required to take full advantage of these software facilities.
The disk-oriented software includes components and capabilities which are similar to
those of the tape-oriented software. Minimum equipment requirements include a Model 20 Processor with 12K bytes of core storage, one 2311 Disk Storage Drive, card I/O equipment, and a
printer.
ffiM is emphasizing the use of the Report Program Generator (RPG) as the primary
programming system for most Model 20 installations. The Model 20 RPG is a generalized program designed to generate coding to perform most of the routine business data processing functions. Input to the RPG consists of specifications written by the user in a format that is considerably easier to learn and use than assembly-language coding. Separate preprinted specification
sheets are used to describe the input to be provided, the calculations to be performed, and the
output to be produced. Most programs written in the Model 20 RPG language can be generated
and run on the larger System/360 models, provided that an equivalent configuration of I/o equipment is available (but note that no unit equivalent to the 2560 MFCM is currently available for the
larger System/360 models). No COBOL, FORTRAN, or PL/I compiler has been developed for
the Model 20 to date.
Other software available for Model 20 systems includes specialized input-output control systems for the communications adapters and magnetic character readers, and a group of
ffiM-developed programs for specific applications such as hospital billing, telephone revenue
accounting, wholesale inventory management, and bill-of-material processing.

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

7/68

.22:221.101
IBM SYSTEM/3eO
MODEL 20

IBM SVSTEM/360 MODEL 20
PRICES

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

Name

Monthly
Monthly
Rental Purchase Maint.
$

$

$

515
720
980
1,235
595
800
1,060
1,315
365
465
625
785
405
505
665
825
900
1,160
1,415
1,765
2,115

23,570
32,590
44,230
55,100
27,060
36,080
47,720
58,590
16,600
20,900
28,100
34,900
18,500
22,700
29,800
36,700
45,600
58,000
72,800
91,700
107,800

37.00
42.00
48.00
52.00
40.00
45.00
51.00
55.00
37.00
42.00
48.00
52.00
40.00
45.00
51.00
55.00
90.00
100.00
105.00
115.00
130.00

232
206
283
31
90
57
20
51
25
77
25
10
57
15
103
155
155

10,910
10,670
11,400
1,450
4,500
2,670
960
2.420
1,250
3,640
1,200
500
2,560
720
5,140
7,270
5,820

22.50
22.50
22.50
2.50
4.50
5.00
2.00
6.00
3.00
4.75
2.25
1.50
2.00
3.50
6.50
4.75
11.00

425

17,000

16.50

232
177
232

10.910
8,160
10,910

5.00

Processing Unit (includes core storage)

PROCESSOR

Processing Unit:
Model B1 - 4,096 bytes
Model C1 - 8,192 bytes
Model BC1 - 12,288 bytes
Model D1 - 16,384 bytes
Model B2 - 4,096 bytes
Model C2 - 8,192 bytes
Model BC2 - 12,288 bytes
Model D2 - 16,384 bytes
Model B3 - 4,096 bytes
Model C3 - 8,192 bytes
Model BC3 - 12,288 bytes
Model D3 - 16,384 bytes
Model B4 - 4,096 bytes
Model C4 - 8,192 bytes
Model BC4 - 12,288 bytes
Model D4 - 16,384 bytes
Model C5 - 8,192 bytes
Model BC5 - 12,288 bytes
Model D5 - 16,384 bytes
Model DC5 - 24,576 bytes
Model E5 - 32,768 bytes

2020

Peripheral Adapters (on 2020)

ATTACHMENTS,
ADAPTERS.
AND
CHANNELS

1580
3480
1;)575
8637
7081
4658
2073
2074
7495
7496
7497

1403 Model 2 Attachment
1403 Model 7 Attachment
1403 Model N1 Attachment
1442 Model 5 Attachment
2152 Attachment
2203 Attachment
2501 Attachment
2520 Model Al Attachment
2520 Model A2 or A3 Attachment
2560 Attachment
Card Print Control
Dual Feed Carriage Control
Printer Features Control
Universal Character Set Adapter
Serial I/o Channel
Input/Output Channel
Communications Adapter
Binary Synchronous Communications
Adapter
Disk Storage Controls:
For Submodel 2 CPU
For Submodel 4 CPU
For Submodel 5 CPU

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

422:221.102

IBM SYSTEM/JIG MODEL 20

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

MASS
STORAGE (2)

Name

PRICES
Montbly
Montbly
Rental !PurchasE Maint.

$

$

$

Q!!k
2311

Disk storage Drive (DIsk storage' Control
req'd.) Model 11 (5.4 million byte capacity)
Model 12 (2.7 m11110n byte capacity)
Disk Pack for 2311

1316
INPUTOUTPUT(2)

590
360
15

25,510
22,310
490

55.00
35.00

255

12,750

49.00

201
263
720

11,350
11,590
32,490

33.25
45.75
91.00

645
465

29,100
28.R10

87.00
67.00

590

27,890

90.00

460

20,500

90.00

129
129
129

6,060
6,060
6,060

13.00
13.00
13.00

-

Punched Card
1442
2051
2520
2520
2560

1575
1576
1577

Card Punch(l)
Card Reader(l)
Model Al - 600 cards/ml~
Model A2 - 1,000 carda/min.
Card Read Punch, Model Al(l)
Card Punch(l)
Model A2 - 500 carda/min.
Model A3 - 300 carda/min.
Multi-Function Card Machlne(l)
Model A1 (reads 500 cards/min., punches
160 col/sec)
Model A2 (reads 310 cards/min., punches
120 col/sec)
Card Print Feature (Mdt A1 only):
First 2 Lines (#1580 req'd.)
Second 2 Lines (#1575 req'd.)
Third 2 Lines (#1576 req'd. )
Printers

1403
1403
1376
4740
5381
6411
8640
8641
1416
2203

5558
3475
7815

Printer (600 lines/min. )(1)
Model 2
Model 7
Printer, Model N1 (1100 lines/min. )(1)
On 1403 Model 2 or 7 Auxiliary Ribbon Feediug Feature
Interchaugeable Chain Cartridll;e Adapter
On 1403 Model 2 only (#5575 req' d.) _
Numerical Print Feature
On 1403 Model 2 or Nl Selecttve Tape Listing Feature
Universal Character Set Feature
(#8637 req'd.) For Model Nl
For Model 2
On 1403 Model Nl only Interchaugeable Chain Cartridge
Printer
Model Al (300-750 lines/min.)
Model A2 (230-600 lines/min. )
On 2203 Prlnter24 Additional Print Positions
Dual Feed Carriage (#3480 req'd.)
6 Additional Tape ChanDels <#3475 req'd.)

775
650
900

34,000 177.00
32,700 153.00
41,200 183.00

75
75

3,075
3,125

17.75

225

9,050

9.00

190

8,100

12.00

10
10

450
450

1.75
1.75

100

3,000

-

525
400

22,310
17,000

71.50
71.50

46
103
10

2,405
4,850
400

4.00
8.50
1.00

-

Magnetic Tape
2415

~/II.Q

M~tic T~e Unit
15]>00 chU sec (Input/Ouiput ChanDeI req'd. ) Modell (two tape drtves and single-channel
controller, 800 BPI)
Model 2 (four tape drives and siDgle-channel
controller, 800 BPI)

A

775

35,650 100.00

1,240

57,040 180.00

422:221.103

PRICE DATA

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

INPUTOUTPUT(2)
(C'ontd. )

3228
7125
7127
7135
5320

Name

PRICES
Monthly
Monthly
Rental Purchase Maint.

$
Model 3 (six tape drives and single-channel
controller, 800 BPI)
30,000 char/sec (Input/Output Channel req'd.)Model 4 (two tape drives and single-channel
controller, 1600 BPI)
Model 5 (four tape drives and single-channel
controller, 1600 BPI)
Model 6 (six tape drives and single-channel
controller, 1600 BPI)
Data Conversion (#7125, 7127, or 7135
req'd. )
Seven-Track Compatibility Models 1-3
Seven-Track Compatibility Models 4-6
Seven- or Nine-Track Compatibility Models 4-6
Nine-Track Compatibility Models 4-6

$

1,705

$

78,430 260.00

935

43,160

115.00

1,500

69,350

205.00

2,065

95,540

295.00

46

2,100

1.00

51

2,330

1.25

98

4,420

3.50

160

7,220

13.00

139

6,290

10.00

2,275

110,500

224.00

1,100
135

49,500
5,600

250.00
61.00

Magnetic Character Reader
1419
1259
2152

Magnetic Character Reader
(Serial I/o Channel req'd.)
Magnetic Character Reader, Modell
(Serial I/O Channel req'd.)
Printer-Keyboard(l)

NOTES:
(1) Requires appropriate Peripheral Adapter on the 2020 Processing Unit.
(2) Most of the Model 20 peripheral devices can be connected only to certain submodels of the 2020 Processing Unit.
(3) The indicated monthly maintenance charges are those in effect for the first 36 months after installation.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

427:011. 100

1. "'.....

I~EDP

AUERB~

rBM SYSTEM/360
MODEL 67
SUMMARY

I£POlrs

~

SUMMARY
Model 67 is a large-scale member of the IBM System/360 family that is specifically
oriented toward time-sharing operation in scientific and educational applications. Its principal
design objective is to furnish continuous computing service to many users simultaneously, while
providing rapid responses to each of the users. The goal is to give each user the impression
that all the facilities of a large computing system are at his disposal and keep him unaware of
the fact that he is actually competing with numerous other users for the use of these facilities.
Model 67 was announced in April 1965 as a replacement for the slower time-sharing
Models 64 and 66, which had been announced only six weeks earlier. Model 67 was originally
offered as a non-standard model available only through special proposals and individual negotiations. In August 1965, Model 67 became a member of the standard IBM product line. In January
1967, however, technical difficulties encountered in the development of the specialized software
support for Model 67 caused IBM to revert to a "controlled marketing" basis, under which all
Model 67 selling efforts and proposals must be approved by IBM's regional management.
Initial customer deliveries of the Model 67 hardware were made ill the fourth quarter
of 1966. but Release 1 of the Time-Sharing System (TSS) software was not officially delivered to
Model 67 users until October 1967. Performance of early Model 67 systems has been far below
expectations, though steady progress is being made in solving the problems. Typical monthly
rentals range from $45,000 to $60.000 for a single-processor Model 67 system and from $100,000
to $200.000 for a two-processor system.
This subreport concentrates upon the specialized characteristics, performance, and
pricing of Model 67 systems and the supporting software. All general characteristics of the
System/360 hardware are described in Computer System Report 420: IBM System/360 - General.
The processors and core storage units used in Model 67 systems are essentially Model
65 units modified to provide effective performance in time-sharing, multiprocessor environments.
The Model 67 hardware is characterized by:
•

The use of virtual addresses rather than physical addresses in all programs to
facilitate dynamic reallocation of storage.

•

An eight-register associative memory t.hat speeds translations between virtual and
physical addresses.

•

A main core storage cycle time of 0.75 microsecond, with eight bytes being accessed per cycle; storage accessing is two-way interleaved.

Figure 1. Panoramic view of a large Model 67 time-sharing system,
with dual Processing Units in left background.
@

1968 AUERBACH Corporation and AUERBACH Info, Inc.

4/68

IBM SYSTEM/360

427:011. 101

•

Main core storage capacities ranging from 262.144 to 2.097.152 bytes. in independent modules of 262.144 bytes.

•

One or two 2067 Processing Units and up to two 2846 Channel Controllers per system.

•

Up to seven I/O channels (six Selector and one Multiplexor) per Channel Controller.
(In systems that do not use a 2846 Channel Controller. up to six Selector Channels
and one Multiplexor Channel can be connected directly to the 2067 Processing Unit. )

•

Ability to connect most (but not all) standard System/360 peripheral devices.

•

Ability to "partition" the system. by setting manual switches on the 2167 Configuration Unit. to make certain components "unavailable" for use by certain other
components.

•

Availability of a Compatibility Feature that permits emulation of IBM 709. 7040.
7044. 7090. and 7094 systems.

The standard complement of System/360 software is not usable for time-sharing operations in Model 67 systems. so IBM is providing a specialized software system called the TimeSharing System (TSS). Release 1 of TSS includes: a time-sharing Supervisor with time-slicing
capabilities. a Command Language for communication between the system and its users. a
FORTRAN IV compiler. an Assembler. and a number of related service routines. No additional
TSS facilities are currently scheduled; IBM withdrew a conversational PL/I compiler. a nonconversational COBOL compiler. a Sort/Merge routine. and other previously-announced TSS
facilities in January 1967.
The TSS Supervisor and its associated service routines control the execution of all jobs
entering a Model 67 system and the environment in which they operate. Operating in time-slicing
fashion. the Supervisor handles the simultaneous execution of conversational-mode programs for
multiple users at remote terminals. Meanwhile. programs that do not require user intervention
can be executed as non-conversational background tasks.
The TSS Assembler and FORTRAN compiler use source languages which are similar to
the Operating System/360 versions of these languages. though there are some differences because
of the specialized hardware features and operating environment of the Model 67. Both the Assembler and the FORTRAN compiler can be used in either the conventional batch mode or in a conversational mode. In the conversational mode. the syntax of each source-language statement entered
by the user is analyzed and. if necessary. corrected in a conversational interchange between
system and user before the assembly or compilation is performed.
The performance of Model 67 systems is difficult to predict accurately. Central processor execution times are somewhat slower (about 12 percent overall) than the Model 65's processor execution times because of the extra time required to transform each virtual operand
address into the appropriate physical address before execution in Model 67. Other factors that
will tend to degrade the performance of a Model 67 system include the Time-Sharing Monitor's
overhead. memory access conflicts. and cable-length delays. It can be assumed that in most
cases the maximum potential throughput of multi-processor Model 67 systems will be substantially
lower than that of an equal number of Model 65 processors in single-processor configurations.
However. where multiple users must be served simultaneously. the overall quality of the computing service provided to these users may well be more important than the maximization of total
throughput.

4/68

A.,

AUERBACH

6). . . .

.27:221.10'

n

BD.

2:1

..

11M SYSTEM/SaO
PRICE DATA

....n

~

IBM SYSTEM/360 MODEL 67

I

PRICES

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

Name

Monthly
Monthly
Rental Purchase Maint.

$

$

$

17,585
17,795

711,490
719,840

525.00
545.00

258
361

9,020
13,680

3.00
4.00

3,350
25
88

135,410
1,100
3,515

100.00
NC
2.00

620

24,250

19.00

232
144

10,545
5,840

9.00
4.00

9,530

397,700

575.00

9,710

405,965

584.00

103

4,700

18.00

103

4,700

18.00

103

4,700

18.00

1,340

55,550

92.00

1,340

55,550

92.00

1,340

55,550

92.00

1,520

63,020

102.00

Processing Unit

PROCESSOR

Processing Unit (max. of 1)
Processing Unit (max. of 2)

2067-1
2067-2

Processing Unit Options
3274
3800
3862
4434
5495
7119
7920
1102

Direct Control (2067-1 only)
Extended Direct Control (required on each
2067-2 in a two-processor system)
Extended Dynamic Address Translation
floating Storage Addressing (2067-1 only)
Partttion SenSing (required on each 2067-2
in a two-processor system)
709/7040/7044/7090/7094/7094 IT
Compatibfilty
1052 Adapter
Additional 2846 Attachment (required on
2067-2 in a system with two 2846s)
Main Storage

2365-2
2365-12

Processor Storage for 2067-1 Processing
Unit (262,144 bytes; max. of 4)
Processor Storage for 2067-2 Processing
Unit (262,144 bytes; max. of 8)
Options for 2365-12

8036
8088
8091

ATTACHl'.IENTS,
ADAPTERS,
AND
CHANNELS

2067 Switching Feature (required on each
2365-12 in a two-processor system)
2846 Switching Feature (required on each
2365-12)
2846 Switching Feature (required, in addition
to #8088, on each 2365-12 In a system
with two 2846s)
Attachments

2167

Configuration Unit:
Modell, for systems with up to:
2067-2s, two 2365-12s, two 2846B,
and 16 I/o control units
Model 2, for systems with up to:
two 2067-2s, three 2365-12s, two
2846s, and 16 rIo control units
Model 3, for systems with up to:
two 2067-2s, four 2365-12s, two
2846s, and 16 I/O control units
Model 4, for systems with up to:
two 2067-2s, eifot 2365-12s, two
2846s, and 32 I 0 control units

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

IBM SVSTEM/HO MODEL .7

427:121.102

PRICES

IDENTITY OF UNIT

CLASS

Model
Number

Feature
Number

IMonthly
Rental PurchasE
$
$

Monthly
Maint.

$

Attachments (Contd.)

ATTACH-

MEm's,

ADAPTERS,
M"D

Name

Channel Controller (max. of 2)
Additional Addressmg I (required on each
2846 m systems with five or more

2846
1086

CHANNEJ.3

(Contd.)

2,010
93

77,600
3,840

117.00
2.00

93

3,840

2.00

2,165
3,090
4,015
41

100,880
143,750
186,720
1,750

55.00
90.00
125.00
2.00

2,265

106,700

97.00

410
258
258
258
118

17,940
10,910
10,910
10,910
5,015

15.00
10.00
10.00
10.00
3.00

206
103
103
103
144
103

10,180
4,070
4,070
4,070
5,770
4,710

8.00
4.00
4.00
4.00
2.50
2.00

77

3,490

3.50

65

2,725

17.00

530

24,440

15.00

36
51

1,500
2,140

NC
NC

36

1,500
2,140

NC
NC

2365-12&)

Additional Addressmg IT (required on each
2846, in addition to fl086, in systems
with seven or eight 2365-12s)

1087

Channels
2860

1095
2870
6990
6991
6992
6993
1095
8100
8100
8100
8100
8170
81'70
8110
INPUTOUTPUT(l)

Selector Channel Modell: one chaDDsl
Model 2: two channels
Model 3: tb.ree channels
Address Preflxlng Feature (required on
each Selector Channel in a two-processor
system)
Multiplexor Channel
Selector Subchannels for 2870 First
Second
Third
Fourth
Address Prefixlng Feature (required on each
2870 in a two-processor system)
Two-Channel Switch for:
2821 Control Unit
2803 Tape Control Unit
2403 Tape Control Unit
2841 storage Control
2314 DIrect Access storage FacUlty
2820 storage Control
Two-Processor Switch for:
9,702 Transmission Control
Console

1052-7
2150
5475
5485
5476
5486

PrInter-Keyboard (connects to 2067
Processing Unit via either a 7920
Adapter or a 2150 Console)
Console
Operator Control Panel, FirstFor 2067-1
For 2067-2
Operator Control Panel. SecondFor 2067-1
For 2067-2

NOTES:
(1) For peripheral devices see IBM System/360 Price Data Sheet.

?/fI!I

A

51

-A
IA'-

AUERBAC~
.

430:011. 100

"ANO'"

IBM SYSTEM/360
MODEL 85
SUMMARY

EDP
REPDRTS

SUMMARY
The System/360 Model 85, announced in January 1968, is the most powerful computer now
offered by IBM. Though slower than the "limited-edition" Model 91 (which is no longer being
marketed), Model 85 can process data up to three times as fast as Model 75 while maintaining
program compatibility with the smaller general-purpose System/360 processors. Rental prices
for Model 85 systems will range from about $85, 000 to $220,000 per month, and deliveries are
scheduled to begin in the third quarter of 1969.
Key elements that contribute to the Model 85's high performance include:
•

Up to 4 million bytes of main core storage with a cycle time of 960 or
1040 nanoseconds per 16-byte access.

•

An 80-nanosecond buffer memory, ranging from 16,384 to 32,768 bytes
in size, that provides fast-access intermediate storage between the
processing unit and main storage.

•

Overlapped internal operations that enable the Model 85 Processing Unit
to fetch, decode, calculate memory addresses, and obtain the operands
for several instructions at the same time it is executing an earlier instruction.

•

Use of monolithic integrated circuits in conjunction with the hybrid SLT
circuits used in earlier System/360 models. Monolithics are used in the
80-nanosecond buffer storage and in arithmetic and logic circuits.
HARDWARE
Model 85 systems are available with four different core storage capacities, as summarized in
Table 1. Error-checking facilities provide for automatic correction of all single-bit storage
errors and detection of all double-bit errors.

Figure 1: The Model 85 Processing Unit (the double-H-shaped unit at left) has
a console with a CRT display and two microfiche viewers, used primarily for maintenance work.
© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

5/68

IBM SYSTEM/36Q-MODEL 85

430:011. 101

TABLE I: MODEL 85 CORE STORAGE CHARACTERISTICS
K

Processing Unit Model

I

J

Storage Capacity, bytes

524,288

1,048,576

Cycle Time, microseconds
Bytes Accessed per Cycle
Storage Interleaving

1. 04

1. 04

2.097,152
0.96

L
4,194,304
0.96

16

16

16

16

2-way

4-way

4-way

4-way

HARDW ARE (Contd.)
Though the Model 85 Processing Unit is program-compatible with the other general-purpose
System/360 models, the implementation of many processor functions is significantly different
in the Model 85. Principal functional components of the Model 85 Processing Unit include an
instruction unit, an execution unit, local storage, a storage control unit, buffer storage, and
control storage.
The instruction unit can concurrently prepare several instructions for execution. Using multiple
registers, it handles the fetching and decoding of instructions, calculation of addresses, fetching of required operands, and issuance of instructions to the execution unit.
The execution unit performs arithmetic and logical operations in a 64-bit parallel adder, an
8-bit serial adder, a 32-bit logical unit, and a 64-bit shifter. Controlled primarily by microprograms, the execution unit processes one instruction at a time, in program sequence.
Local storage contains the standard System/360 complement of 16 four-byte general registers
and 4 eight-byte floating-point registers. These registers are used by both the instruction unit
and the execution unit, and they are designed to permit data to be fetched from four general registers and stored into a fifth during a single 80-nanosecond machine cycle.
The storage control unit interfaces and controls all data transfers between main core storage
and the processing unit, buffer storage, and I/O channels.
Buffer storage has a standard capacity of 16,384 bytes, which can optionally be expanded to
24,576 or 32,768 bytes. It is used to hold the currently-active portions of main storage for
rapid reference, thereby greatly reducing the effective storage access time in most applications.
From the programmer's viewpoint, the buffer storage is completely "transparent"; its operations are carried out automatically, and the programmer cannot address it directly.
Both main and buffer storage are partitioned into 1024-byte pages, and each page is subdivided
into 16 blocks of 64 bytes each. For each fetch operation, the buffer storage control determines
whether there is a page in buffer storage that corresponds with the addressed page in main storage. If so, the data is fetched from buffer storage. If not, thc data is fetched from main storage, and one of the buffer pages is automatically assigned to hold the addressed main storage
page so that subsequent references to the same page can be handled efficiently. Store operations
cause both main and buffer storage to be updated at the same time.
To minimize processor delays when a new buffer page is assigned, the particular 64-byte block
that was referenced by the fetch operation is always the first block loaded into buffer storage.
Subsequently, as fetch operations referenCing other blocks in the same page are encountered.
those blocks are also loaded into the appropriate buffer storage locations.
The Model 85 Processing Unit automatically keeps track of the last time each page in buffer
storage was referenced. Whenever a new page is loaded into buffer storage, it replaces the
page that has gone unused for the longest period of time. As a result of this technique, IBM
estimates that the required operands will be present in buffer storage for about 95 per cent of
all fetch operations in typical programs.
Control storage holds the microprograms that control most operations of the Model 85 Processing Unit. It consists of approximately 2000 control-word locations of read-only storage (ROS)
and 500 locations of writable control storage (WCS), each with an 80-nanosecond cycle time.
The Model 85 Processing Unit provides all the facilities of the System/360 Universal Instruction
Set, including the Store and Fetch Protection, Direct Control, and Interval Timer features.
Other standard features - unique to the Model 85 - are extended-precision floating-point arithmetic, which uses 16-byte operands and offers 112-bit precision, and a byte-oriented operand
feature, which enables programmers to ignore many of the boundary constraints that normally
complicate System/360 programming (though extensive use of the feature leads to a significant
reduction in execution speed). Also standard is an instruction retry capability; when the processor detects an error during the execution of an instruction, it will in many cases automatically
retry the instruction.
Optional features available for the Model 85 include: (1) Buffer Expansion, which extends the
capacIty of buffer storage from 16K to either 24K or 32K bytes; (2) High-Speed Multiply, which
enables fixed-point and long-format floating-point multiply instructions to be executed in less

5/68

A

AUERBACH

(Contd.)

SUMMARY

430:011. 102

HARDWARE (Contd.)
than 450 and 600 nanoseconds, respectively; and (3) 709/7040/7044/7090/7094/7094 II Compatibility, which enables a Model 85 system to emulate the operations of the indicated secondgeneration IBM computer systems. IBM expects the internal performance of a Model 85 in the
emulation mode to be up to twice that of a 7094 II.
A system control console on the Model 85 Processing Unit provides, in addition to a typical
complement of switches and lights, a CRT display and two microfiche viewers. The CRT is
used to display the contents of data and address registers, while the microfiche viewers are
used mainly to facUitate system maintenance. Various optional features and peripheral units
can be added to provide extended or duplicate console I/O facilities.
The Model 85 requires a supply of chilled water, for system cooling, and a 400-cycle motorgenerator set, which can be located in a separate room.
I/O devices are connected to a Model 85 system via Multiplexor or Selector Channels. One
Multiplexor Channel and/or up to six Selector Channels can be used in a Model 85 system - the
same maximum channel complement as in the slower Model 65 and 75 systems. Each of the
2860 Selector Channels has 8 control unit positions and can control one I/O operation at a time,
at a data rate of up to 1,300,000 bytes per second. The 2870 Multiplexor Channel has 192 subchannels and can handle overlapped I/O operations with an aggregate data rate of up to 110, 000
bytes per second. The 2870 can optionally be equipped with up to four Selector Subchannels,
each capable of handling one I/O operation at a time.
Most of the I/O and mass storage units in the System/360 line can be used in a Model 85 system,
though magnetic character readers, optical readers, and the low-speed 2415 Magnetic Tape Unit
cannot be connected. Devices that can be connected only to a Selector Channel include 2301 and
2303 Drum Storage, 2311 Disk Storage, 2314 Direct Access Storage, and the high-performance
2420 Magnetic Tape Unit. The Multiplexor Channel is required for connection of a 2702 or 2703
Transmission Control or a 7770 or 7772 Audio Response Unit.
SOFTWARE
The Model 85 maintains program compatibility With the smaller general-purpose System/360
processing units even though the manner in which many of the processor functions are implemented is significantly different in the Model 85. As a result, Model 85 systems can take full
advantage of the software already developed for the other System/360 models.
Most Model 85 installations are expected to operate in a multiprogramming mode and to utilize
the MVT (Multiprogramming with a Variable number of Tasks) version of the Operating System/
360. The other versions of 08/360 - MFT (multiprogramming with a Fixed number of Tasks)
and PCP (Primary Control Program) - can also be used.
Additional software support to be provided specifically for the Model 85 includes:
•

Support of the extended-precision floating-point feature in OS/360
Assembler F.

•

Support of the Operator's Console feature under OS/360.

•

The Recovery Management Program (RMP) , an optional extension of the
MVT and MFT versions of OS/360 that will support the extensive errorchecking and diagnostic facilities of the Model 85 hardware.

PRICE DATA
Please refer to the general System/360 Price Data section, beginning on page 420:221.101, for
prices of the 2085 Processing Unit and the optional features, channels, and I/O units that can be
used with it.

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

5/68

-.£.

432:011.100
mimi

~EDP

AU~

-

•

IBM SYSTEM/360
MODEL. 25
SUMMARY

I£ ... r$

SUMMARY
The IBM System/360 Model 25, announced in January 1968, is intermediate in price and performance between the earlier Models 20 and 30. Rental prices for typical Model 25 systems will
range from approximately $3, 500 (for a small card system) to $10, 000 (for a disk/tape system)
per month, and deliveries will begin in January 1969.
Unlike the smaller Model 20, the Model 25 is program-compatible with the larger general-purpose
System/360 processors and can use most of the standard System/360 software. Moreover, the
Model 25 can be equipped with compatibility features that enable it to execute most programs
written for IBM 1401, 1440, or 1460 systems. Instruction execution speeds of the Model 25
Processing Unit are approximately five to ten times as fast as the Model 20 and one-third to
two-thirds as fast as the Model 30, depending upon the type of program.
Thus, the Model 25 is much more closely related to the Model 30 than to the Model 20; it can
reasonably be characterized as an updated version of the Model 30 that has been modestly reduced
in both price and performance. Some restrictions on Model 25 configurations, however, will
force users who need fast magnetic tape units or large-capacity mass storage to move up to the
more expensive Model 30.
HARDWARE
The Model 25 Processing Unit, unlike other System/360 models, has a core storage unit that is
divided into three distinct areas called program storage, control storage, and auxiliary storage.
•

Program storage corresponds to the main storage in other System/360 models and
holds the instructions and data for all programs written by users. Processing units
with the following program storage capacities are available:
Model
Model
Model
Model

D
DC
E
ED

-

16, 384
24, 576
32, 768
49, 152

bytes
bytes
bytes
bytes.

Figure 1. The Model 25 Processing Unit Features a Simplified Control Panel
and Uses the 1052 Printer-Keyboard (Left) for Many Console I/o Functions.

o

1968 AUERBACH Corporation and AUERBACH Info, Inc.

5/68

SUMMARY

432:011. 101

HARDWARE (Contd.)
•

Control storage holds the microprograms which interpret the instructions and control
This 16, 384-byte area is reserved for the control
programs and may not be directly accessed by users' routines. The control storage
area is reloadable, and a different control program must be loaded to switch from the
normal System/360 mode of operation to the 1400 Series compatibility mode.

•

Auxiliary storage holds the 16 general registers, the 4 floating-point registers, I/O
control words, and temporary work areas for various processor functions. Its
capacity ranges from 2, 048 to 4, 096 bytes, depending upon the processor model.

all system functions in a Model 25.

The cycle time for all three areas of core storage is 900 nanoseconds, but two cycles (1. 8 microseconds) are required to read or write either a byte or a halfword (two bytes) in either program
or auxiliary storage. The first cycle of each pair obtains information needed to control the dataaccess cycle that follows. In processing the microprogram steps in control storage, however,
only one 900-nanosecond cycle is required to read out each 2-byte control word.
The Model 25 Processing Unit also contains 64 bytes of "local storage, " a 180-nanosecond,
monolithic-circuit, "scratchpad" memory unit. Local storage holds the data being operated on
by the current microprogram step, as well as addresses and other information required for
internal processing and I/O operations.
Fixed-point binary and decimal arithmetic instructions (i. e., the System/360 Standard and
Commercial Instruction Sets) are standard features of the Model 25. The instruction set can be
extended through inclusion of the optional Floating-Point Arithmetic, Direct Control, and Storage
Protection features. Other available options include the Interval Timer, External Interrupt,
Emergency Power-Off Control, and 1400 Series Compatibility.
When operating in the 1400 Series compatibility mode, a Model 25 Processing Unit will execute
a typical mix of commercial instructions somewhat faster than the original 1401 or ]440 system.
Overall job performance will naturally vary with the type of program and the complement of I/O
devices used. A 16K Model 25 system equipped with appropriate I/O devices and optional features
can emulate a 1401, 1440, or 1460 system with up to 16,000 storage positions (the maximum 1400
Series storage capacity), all standard instructions, most optional features, and the following I/O
devices:
1402
1403
1407
1311
729,

or 1442 Card Read-Punch
or 1443 Printer
or 1447 Console Inquiry Station
Disk Storage Drives
7330, or 7335 Magnetic Tape Units.

Error-detection circuits in the Model 25 Processing Unit check the validity of core storage
addresses, the parity of data transferred within the processor and to and from I/O devices, and
other possible malfunctions. The microprograms include error recovery routines which attempt
to overcome the error condition when possible.
A vertical control panel on the front of the proceSSing unit contains the switches, keys, and lights
needed to operate a Model 25 system. A 1052 Model 7 Printer-Keyboard, connected via an
Integrated 1052 Attachment Feature, is required for console I/O use. Many functions that were
keyed in by separate buttons in earlier System/360 consoles are entered through the keyboard in
the Model 25.
Among the most distinctive aspects of the Model 25 are its system configuration capabilities and
limitations. Any or all of the following I/O units can be connected directly to a Model 25 Processing Unit - without their usual control units or channels - via the appropriate Integrated
Attachment Features:
•

One 1403 Printer, Model 2 (600 1pm, 132 print positions) or Model 7 (600 1pm, 120
print positions).

•

One 2540 Card Read-Punch (reads 1000 cpm, punches 300 cpm).

•

Up to four 2311 Modell Disk Storage Drives (7.25 million bytes each).

Other I/O devices can be connected to a Model 25 system via either an optional Multiplexor
Channel or Selector Channel, only one of which may be used in a system. Both types of channels
have 8 control unit positions, and the Multiplexor Channel has 32 subchannels. Most System/360
I/O units can be used on either type of channel, but there are some significant restrictions:

5/68

•

Magnetic tape units can only be connected to the Selector Channel.

•

The 1259 and 1412 Magnetic Character Readers, 2702 and 2703 Transmission Controls,
and 7770 and 7772 Audio Response Units can only be connected to the Multiplexor
Channel.

•

No mass storage units other than the 2311 Disk Drives can be used in a Model 25
system.

fA

AUERBACH

'"

(Contd. )

IBM SYSTEM/360

432:011.102

HARDWARE (Contd.)
•
Although the "integrated" 2311 Disk Drives have a data transfer rate of 156. 000 bytes
per second. no other I/O unit with a data rate exceeding 30. 000 bytes per second can
be used in a Model 25 system.
The "integrated" 1403 Printer. 2540 Card Read-Punch. and 1052 Printer-Keyboard are programmed and controlled as if they were connected to the Multiplexor Channel. though they do
not require the presence of this channel. Similarly. the 2311 Disk Drives are programmed and
controlled as if they were connected to the Selector Channel. though they do not require its
presence.
I/O operations on the various subchannels of the Multiplexor Channel can be overlapped with one
another. with operations of the "integrated" I/O devices. and with internal processing - though
there are aggregate data rate limitations which restrict the number of devices that can operate
simultaneously. The Selector Channel can control only one I/O operation at a time. but internal
processing and operations of the "integrated" I/O devices other than 2311 Disk Drives can be
overlapped with a Selector Channel operation.
SOFTWARE
The Model 25 maintains program compatibility with the larger general-purpose System/360 processing units even though the manner in which many of the processor functions are implemented
is significantly different in the Model 25. As a result. Model 25 systems can make use of most
of the software developed for the larger System/360 models.
Model 25 systems with adequate core storage capacities and I/O equipment will be able to use the
extensive software facilities that IBM provides within the Disk Operating System (OOS). Tape
Operating System (TOS). Basic Operating System (BOS). and Basic Programming SUpport (BPS).
For a list of these facilities. see Table V in the System/360 SUmmary, Section 420:011. Because
of the 49K-byte maximum storage capacity. it will not be practical for Model 25 systems to use
the large-scale Operating System/360 software.
PRICE DATA
Please refer to the general System/360 Price Data section, beginning on page 420:221. 101, for
prices of the 2025 Processing Unit and the optional features and I/O units that can be used with it.

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

5 / 68

435:011. 100

£. ........

~EDP
l"lns

-

IBM SYSTEM/360
MODEL 44
SUMMARY

AUE~

•

SUMMARY
The Model 44 is a special-purpose scientific data processing system that was officially
added to the mM System/360 family in August 1965. The Model 44 brings to the System/360 line
a computer that has been custom-tailored for performing high-speed binary arithmetic. operations
in scientifically-oriented applications. Capabilities are included in the Model 44 that permit its
use in process control, data acquisition, and real-time operations. IBM advertises that the
System/360 Model 44 offers the internal speed of the IBM 7094 at substantially less cost.
A basic Model 44 system with card read-punch and printer rents for $5,330 per month,
but typical system rentals will be in the $7,000 to $11,000 range. A Model 44 processor can
perform internal computations faster than a System/360 Model 50 processor, and has a monthly
rental that is less than half that of the Model 50. First deliveries of the Model 44 occurred
during the third quarter of 1966.
The Model 44 uses the same basic data and instruction format as the System/360
Models 30 through 75. The 32-bit (4-byte) binary word is the basic unit of internal data manipulation. The standard Model '44 instruction repertoire is a compatible subset of the full System/360
repertoire, though certain optional features add instructions which are unique to the Model 44.
There are no instructions for performing decimal arithmetic, editing, code translation, or radix
conversion operations. Floating-point arithmetic can be provided, but only as an optional feature
- though it will certainly be included in most Model 44 installations. Extensive interrupt facilities
are also available, but only as optional features. Tee Commercial Feature, announced in March
1968, promises compatibility with other System/360 models through trapping and emulation of
the instructions which are not included in Model 44's standard repertoire.
The Model 44 Processor can contain 32,768, 65,536, 131,072, or 262,144 bytes (8K
to 64K 32-bit words) of core storage that is addressable to the individual byte. Core storage
cycle time is 1 microsecond per 4-byte word. The 16 general-purpose registers are normally
implemented in an extension of the I-microsecond core storage. However, if the optional HighSpeed General Registers feature is installed, the standard registers are replaced by 16 registers
implemented in "solid logic technology" circuitry that has a cycle time of 250 nanoseconds per
word. Use of this feature substantially reduces the address generation time of all instructions
and the basic execution time of all fixed-point instructions.
Included in the Model 44 Processor as standard features are a console printer-keyboard and a single-disk storage drive housed in the processor cabinet. The disk drive uses a
replaceable single-disk mM 2315 Disk Cartridge that can store up to 1,171,200 bytes of data.

Figure 1. The Model 44 Processor, designed for scientific applications, features
an integrated disk storage drive that uses interchangeable .
single-disk cartridges.
@

1968 AUERBACH Corporation and AUERBACH Info, Inc.

4/68

IBM SYSTEM/36D-MODEL 44

435:011.101

The drive has an average random access time of 70 milliseconds. A second single-disk storage
drive can be installed as an optional feature, doubling the on-line storage capacity. Seek overlap
is possible when both drives are installed, but simultaneous reading and writing cannot be performed by the two drives. The primary function of the Model 44's built-in single-disk storage
drive is to provide residence for the system's software support.
The Model 44 Processor is optimized to perform high-speed fixed- and floating-point
binary arithmetic. Fixed-point arithmetic uses the 32-bit binary word (1 sign and 31 integer
bits) as its basic operand. Fixed-point binary half-word operations can also be performed.
When the optional high-speed general registers are installed, the basic fixed-point arithmetic
instruction execution times can be reduced by more than 50 percent. For example, the basic
register-to-register add time is 3.75 microseconds; the same operation performed with the
high-speed registers takes only 1. 75 microseconds.
The optional Floating-Point Arithmetic feature provides Single-length (24-bit precision)
and double-length (56-bit precision) arithmetic and comparison operations. Use of the 16 highspeed general registers will reduce by 0.75 microsecond the instruction execution time of each
floating-point instruction that references core storage. When the full 56-bit precision is not
required in specific problems, use of a rotary switch on the processor's control panel can adjust
the precision of the floating-point fraction down to 48, 40, or 32 bits. Each progressively lower
setting considerably reduces the time required to execute all double-length floating-point instructions. The 56-bit fraction length must be used if programs are expected to be run on other
System/360 models and produce identical results.
Peripheral devices are connected to a Model 44 system by means of either a standard
Multiplexor Channel and/or one or two High-Speed Multiplexor Channels. These channels share
many of the functional characteristics of similarly-designated I/o channels used with the System/360 Models 30 through 75. The standard Multiplexor Channel provides up to 64subchannels,
and each of the High-Speed Multiplexor Channels. provides up to 4 subchannels, setting the theoreticallimit of simultaneous input-output data transfer operations at 72. Selector Channels are not
available for Model 44 systems, but the High-Speed Multiplexor Channels can provide equivalent
capabilities.
Most of the standard System/360 peripheral devices can be connected to a Model 44
system, subject in most cases to the general configuration rules of the System/360 computer
family. One important difference in configuration possibilities lies in the fact that the Model 44
permits only the IBM 2311 Modell Disk Storage Drive - in addition to the built-in Single Disk
Storage Drive - for use as auxiliary, random-access storage. Other System/360 models can
use up to seven different random-access storage devices, offering a wide range of storage
capacities and access times. The Model 44 cannot use the high-performance 2420 Magnetic Tape
Unit or 7340 Hypertape Drive, but it can use both the 800 and 1,600 bpi models of the IBM 2401
Magnetic Tape Units. Display devices (Models 2Z50 and 2260) and data communications devices
(2101 and 2702 Transmission Control Units) can also be connected to a Model 44 system, but no
provisions have been included for use of System/360 optical readers, MICR readers, or audio
response units.
The peripheral device flexibility of the Model 44 is increased by its capability to be
connected to an IBM 1800 Data Acquisition and Control System via the standard I/o channels.
The IBM 1800 brings to the Model 44 the specialized facilities required to perform process control and high-speed data acquisition tasks. An optional Direct Data Channel feature in the Model
44 system permits exchange of data with an external device or system - such as the IBM 1800 at speeds up to 4,000,000 bytes per second.
Because of its restricted instruction repertoire, and because of its specialized built-in
disk drive, the Model 44 uses a custom-designed software package - integrated through use of
the system disk - that is generally not compatible with the extensive array of standard System/360
software. Designated the Model 44 Programming System, the software centers around a Supervisor program that controls the sequential execution of batched jobs in a non-multiprogramming
environment. Input-output device control routines are also included within the Supervisor. A
disk-resident USA standard FORTRAN compiler with extensive capabilities, an ar.sembler, and
an array of disk-resident utility routines are also provided. The FORTRAN language is fully
compatible with the FORTRAN H-levellanguage used Wtth the IBM Operating System/360; the
Model 44 assembly language is a generally compatible subset of the assembly language used with
the Operating System/360.
The Model 44 Programming System, released in June 1967, can be used by any
Model 44 system that includes at least 65,536 bytes of core storage, one Multiplexor Channel,
one card reader, card punch, and line printer, and the standard Single Disk storage Drive.
Users of Model 44 systems with only 32,768 bytes of core storage can use the
restricted software of the Model 44 Basic Programming Support (BPS). The BPS package is
basically card- and tape-oriented and includes a USA Standard Basic FORTRAN compiler, a
basic assembler, a group of input-output subroutines, and a set of basic utility programs.
Released in October 1966, the BPS software does not utilize the Model 44's Single Disk Storage
Drive and is far less efficient than the integrated Model 44 Programming System described above.

4/68

A.

AUERBACH

435:221.101
" ••,,11

IBM SYSTEM/360

BDP

I."'''

PRICE DATA

IBM SYSTEM/360 MODEL 44
IDENTITY OF UNIT
CLASS

Model
Number
2044

PROCESSOR

3895
4427
4583
5248
4598
4560
4561
4562
4599
4565
4566
4567
3275
3288
3621
4555
9509
5625
8501
7531
7531,7532
7531-7533
7531-7534
2251
2252
7500
7501
6415
2315

Feature
Number

Name
Processing Unit (includes core storage):
Model E - 32,768 bytes
Model F - 65,536 bytes
Model G - 131,072 bytes
Model H - 262, 144 bytes
Processing Unit Standard Features:
Single Disk Storage Drive
(1,171,200 bytes)
Console Printer-Keyboard
Processing Unit Optional Features:
External Interrupt.
Floating Point Arithmetic
High-Speed General Registers
Multiplexor Channel
High-Speed Multiplexor Channel (first)
Additional High-Speed Multiplexor
Subchannels:
First
Second
Third
High-Speed Multiplexor Channel (second)
Additional High-Speed Multiplexor
Subchannels:
First
Second
Third
Direct Data Channel
Direct Word
Emergency Power-Off Control
High Resolution Interval Timer
Pin Feed Platen
Priority Interrupt
Commercial Feature
Store and Fetch Protection, for:
Model E
Model F
Model G
Model H
Console Printer-Keyboard Multiplexor
Channel Attachment
Console Printer-Keyboard High-Speed
Multiplexor Channel Attachment
Single Disk Storage Drive Multiplexor
Channel Attachment
Single Disk Storage Drive High-Speed
Multiplexor Channel Attachment
Second Single Disk Storage Drive
Disk Cartridge

,

PRICES
Monthly
Monthly
Rental Purchase Maint.

$

$

$

3,570
4,805
6,455
10,370

119,890
174,310
249,970
428,450

200.00
220. 00
250. 00
350.00

-

-

-

31
283
720
360
670

1,220
11,200
28,520
14,240
26,480

1.00
11.00
28.00
17.75
29.25

129
129
129
370

5,090
5,090
5,090
14,650

6.00
6.00
6.00
18.00

129
129
129
825
283
NC
103
NC
410
620

5,090
5,090
5,090
32,590
11,200
NC
4,070
NC
16,300
19,590

6.00
_ 6. 00
6.00
30. 00
3.00
NC
2. 00
NC
10.00

232
283
334
411

9,170
11,210
13,250
16,310

4.00
4.75
5.50
7.00

?

NC

NC

NC

NC

NC

NC

NC

NC

NC

NC
237

NC
9,485
90

NC
40.00

-

-

NOTES:
(1) For peripheral devices, see System/360 Price Data Sheet.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

AmEDP
.....
lA,

AUERIA.C~

-

450:011.100
IBM SYSTEM/3
SUMMARY

REPlIlS

~

IBM SYSTEM/3
.1

SUMMARY
The IBM System/3 announced on July 30, 1969, represents IBM's serious entry into the small-scale
business-oriented computer market. The new system is designed to accommodate the needs of firsttime, computer users who more than likely are currently operating with an EAM installation. The
System/3 is not directed at the replacement market, nor is it intended as a terminal system. Large
companies wishing to decentralize their peripheral operations represent another prospective market.
The System/3 is offered in two basic Configurations; a card system starting at a monthly rental of
$945, and a disk system for $1,345 per month. A full complement of peripheral devices are offered;
on-line equipment including a multi-function card unit, a printer and keyboard-printer attachment,
and a keyboard-entry station; off-line equipment including a data recorder and acard sorter.
uf greatest interest to the industry at this time is the introduction of the new System/3 punched card.
The card is approximately one-third the size of the old standard SO-column card, holds 20 percent
more information (96 characters). uses 6-bit BCD codes rather than the standard Hollerith code,
and provides four lines of engraved printing at the top of the card. The new card lessens the space
requirements for off-line storage and permits more compact card handling equipment.
The System/3 is relatively fast-operates at a cycle time of 1. 52 microseconds. A good example of
its processing speed may be made by comparing the ability of the System/3 to add two five decimal
digits in 31. 9 microseconds as opposed to the same operation performed on an IBM System/360
Model 40 in 39 microseconds.
IBM offers an optional Application Customizer Service to support the System/3 user. The user defines his application speCifications on a questionnaire, and then IBM prepares at one of its Basic
System Centers a tailored application package of sample reports, flowcharts, cross-listings, record
layouts, etc., which enables the user to proceed to program with a fully-defined application analysis.
Extensive software support includes a new RPG II language, a disk assembler, and an assortment
of card and disk, sort and utility routines. All of these programs operate under a system control
program with executive and job control facilities.
In addition, IBM offers system engineering support for assistance in the preparation of the user
programs. System/3 customer training courses are also available. All of the above-mentioned
support (SE services, education, software except for the control program, and the application
customizer service) is separately priced. The program products are licensed to a specific CPU
number and may not be used with any other system.
Likely competition for the System/3 will fall in the range of small-scale systems including the
UNIVAC 9200, the NCR Century Series 100, the GE 100 Series, and the Honeywell 110.
The first delivery of the System/3 card configuration is scheduled for January, 1970, and the expected first installation of a System/3 disk configuration will be in September, 1970. IBM anticipates a period of no more than six months for delivery after the receipt of an order.
This report was prepared immediately after the public announcement of the System/3. It will be
refined and enhanced in a subsequent issue .

.2
. 21

HARDWARE
Data structure
The basic unit of data storage in the System/3 is the eight-bit (plus one parity bit) byte. This data
structure is COmmon in many competitive systems following the lead of the IBM System/360. A
byte can represent one alphanumeric character or a portion of a binary field.
Bytes can be addressed and manipulated individually, or consecutive memory locations can be grouped
to form variable length binary fields up to 256 bytes in length. The internal code is EBCDIC (Extended Binary Coded Decimal Interchange Code). Data may be conveniently represented by hexidecimal notation. Decimal data consists of one digit per byte. The zone portion of the byte or the
leftmost four bits contains the sign of the field when the byte is located in the units position of the
field. Decimal operations are performed on fields of up to 16 digits in length wi"th the resultant
field up to 31 digits long.
Instructions are either 3, 4, 5, or 6 bytes long and specifiy one or two memory addresses.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

8/69

450:011.220
.22

IBM SYSTEM/3

System/3 Punched Card
The System/3 punched card represents the most important innovation introduced with the new system.
The new card
features 96 characters of punched data (20 per cent more data storage) arranged in three tiers of
32 columns each of round-holed 6-bit BCD character codes. The card measures 2 3/4 by 11/4
inches. Four lines of engraved printing at the top of the card interpret the BCD code content.

It is approximately one-third the size of the standard 80-column hollerith code card.

Besides the advantage of more unit record storage and interpreting, the new card requires less
off-line storage space, is easier to handle by operations, and allows a more compact card handling
device (the Multi-Function Card Unit). In addition, processing is enhanced by lessening the hardware overhead conversion to internal processor codes. That is, BCD to EBCDIC code conversion
is faster and easier to accomplish than hollerith to EBCDIC .
. 231

System/3 Card System; Configuration I
Rental

Equipment
1- 5410 Processor with 8,192 bytes of memory
1- 5424 Multi-function Card Unit, Model 2:
reads 500 cpm, punches 120 cpm, prints
1201pm.
1- 5203 Printer, Model 2; 200lpm
1- 5496 Data Recorder; punches 60 cols, per sec.
1- 5486 Card Sorter; 1,000 cpm
1- 5701RGI RPG II Compiler: Card System
1- 5701 UTI Card Sort and Utilities

$310
405
280
155
85
35
10
Total Rental:

.232

System/3 5-Million-Byte Random Access System; Configuration III R
Equipment
1- 5410 Processor with 12,288 bytes of memory
1- 5424 Multi-Function Card Unit; Model 2:
reads 500 cpm, punches 120 cpm, prints 120 lpm
1- 5203 Printer, Model 2; 200 Ipm
1- 5444 Disk Storage Drive; Model 2: 4.9 million
characters
1- 5702 ROI II Compiler: Disk System
1- 5702 SMI Disk Sort
1- 5702 UTI Disk Utilities

· 24

Rental
$535
405
280
255

Total Rental:
· 241

$1,280

45
10
10
$1,540

Internal Storage
Core Memory
Working Storage for the System/3 is provided by magnetic core memory. The processor can access
one 8-bit byte per cycle of 1. 52 mi~roseconds. A system/3 configuration can have any of the following
capacities: 8,192; 12,288; 16,384; 24,396; or 32,768 bytes.
Each memory location is individually addressable and can hold one byte, consisting of eight bits plus
a parity bit. Each byte can represent one alphan\lUleric character or an 8-bit portion of a binary
field.
A parity bit is generated wherever a byte is stored and checked for validity when the byte is read. A
check is also made for memory addresses larger than the physical memory size. All of the core
storage is available for user programs or system software. No hardware protected or dedicated
areas are locked out.

· 242

Disk Unit
The 5444 Disk Storage comprises both removable and non-removable storage. The unit is an integral
part of the standard disk configuration and may not be acquired separate and apart from the package
configuration. Two disk units can be on-line, both of which are housed in a compact compartment
located directly below the 5424 Multi-Function Card Unit. In a non-disk configuration, drawer space
is provided in its place.
Three Models of the 5444 Unit offer storage capacities ranging up to 4.9 million bytes or 9.8 million
packed decimal digits. Data stored in packed-decimal form (two 4-bit BCD digits per byte) are
transferred to and from main memory intact and RPG II subroutines are available to process this
data organization due to the absence of standard decimal arithmetic operations.
The 5440 Disk Cartridge, unlike most of the other recently-announced disk packs, is not compatible
with the IBM 2316 Disc Pack. The 5440 consist of one 14-inch disc platter with two metallic recording surfaces, both of which are used for storage.

8/69

A"

AUERBACH

(Contd.)

SUMMARY

450:011.250

.242 Disk Unit (Contd.)
Each 5444 Disk Storage Unit has a standard non-removable pack offered without charge: the user
must purchase the removable storage. Model I provides 100 cylinders of removable storage and
100 cylinders of non-removable storage (2 tracks per cylinder) with a storage capacity of 2.45
million bytes or 4.9 million packed-decimal digits. Model 2 provides twice the storage, with 200
cylinders each of removable and non-removable storage and a capacity of 4. 9 million bytes or 9.8
million packed-decimal digits. Model 3 makes available the same storage as Model I, concentrating
all the data on the removable storage.
The System/3 disk configuration incorporates the I/o control for the 5444 within the processor,
thereby avoiding a separate disk controller. Data is transferred at a rate of 199,000 bytes per
second. Access time ranges from 39 to 395 milliseconds averaging 153 milliseconds for the Model
I, and 39 to 750 milliseconds for Models 2 and 3. The disks rotate at a speed of 1, 500 RPM.
A one-disk-drive configuration can include either Modell or Model 2: a two-disk-drive configuration
combines either Model 2 and Model 3, or two Model 2's. Consequently, the maximum disk storage
on-line is 9.8 million bytes or 19.6 million packed-decimal digits.
Parity checking is performed for each byte during read and write operations. All detected error
conditions are handled by the RPG software .
. 25

Central Processor
The design of the System/3 processor is conventional in its organization with the exception that the
control of the standard I/O devices (Multi-Function Card Unit, Printer, and Disc Storage Device in
a disc configuration) is incorporated into the processor hardware. In addition to the I/O device control unit, the processor consists of three basic units:

. 251

•

An arithmetic/logic unit (ALU) which contains the circuitry for computations and logical decisions.

•

A CPU I/O control unit which supervises the transfer of data between the peripheral equipment
and main memory.

•

A magnetic core memory unit, as described in Paragraph. 231, Core Memory .

Processing
There are three types of instruction formats which are represented as follows:
Command format
One-address formats
Two-address formats

OP
OP
OP
OP
OP
OP
OP

Q
Q
Q
Q
Q
Q
Q

I

B
B

1

B

1

B
A

B
B

I
B

A

A

I
I
A

I

Where OP=l-byte operation code; indicating the addressing mode of both operands and the type of
operation to be performed
Q = I-byte utility byte; indicating explicit lengths of one or more of the operands, immediate data,
a bit mask, register address, branching conditions, etc.; and in I/O commands, specifies device
involved.
A, B = 1 or 2-byte address operands.
There are two methods of addressing: direct and indexing. Direct addressing involves a two-byte
instruction address field in which the real address is extracted directly from the instruction and
loaded into the address register. Indexing takes place with a one byte address field. The OP code
incorporates the indicators that specifies whether or not indexing is to take place, and which one of
the two index registers are to be used.
The instruction repertoire consists of 28 separate instructions providing arithmetic, logical, and
device command capabilities. Because data may be stored on disk in packed-decimal iormaL (two
4-bit BCD characters per byte), pack and unpack routines are provided in the RPG software to process the packed-decimal operations .
. 26

Input-Output EqUipment
The IBM System/3 offers a broad range of peripheral devices to accomplish the needs of card-and
disk-oriented systems. No magnetic tape or paper tape devices are offered. The card-handling
device and printer are integral parts of the basic hardware configurations. Up to two disk units
can be incorporated in the disk configuration. These three devices cannot be acquired separately,
nor can additional units be attached. The logical control for all of the System/3 devices is in the
processor.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc

8/69

IBM SYSTEM/3

450:011.261
. 261

.£arl!.J~~ipment

The 5224 Multi-Function Card Unit covers the entire gamut of card handling operations including
reading, punching, interpreting, sorting and collating. Two models are offered: Model Al reads
250 cards per minute, punches 60 cards per minute, and prints at a speed of 60 lines per minute;
Model A2 operates at double these reading, punching, and printing speeds.
The 5224 provides four 2, OOO-card stackers. Other features include a photoelectric read, punch
verify controls, error checking and engraved printing (raised characters) .
. 262

Printer
The 5203 Printer is available in two Models: Model I prints 100 lines per minute and Model 2, twice
the speed or 200 lines per minute. The standard print line contains 96 columns, which can optionally be expanded to 108, 120, or 132 columns. A universal character set of 64 characters can also
optionally replace the standard 48 character set.
There is no paper tape loop control and all paper advance including spacing, skipping and page
eject is under program control. The 5203 is also capable of optionally printing two forms simultaneously.
The 5741 Printer-Keyboard functions both as an inquiry device and as a second printer. Inquiry
statements may be keyed directly into core storage. Printing takes place under program control
at a speed of 15.45 characters per second on a 12 1/2 inch writing line with a density of 10 characters per inch •

. 263

On-Line Data Recording Equipment
The 5475 Data-Entry Keyboard provides on-line data recording and verification. It functions in a
manner similar to the off-line 5496 Data Recorder with the exception that it operates under the
processor I/O control.
'

.264

Off-Line Equipment
The 5496 Data Recorder operates off-line and is, in effect, a combination key-punch and verifier.
The two operations can take place simultaneously because of a 96 position buffer that holds the
previous card for verification and punching while the data from the current card is keyed in.
A self-checking feature permits a check digit to be added to the basic code number for internal
hardware verification. The check digit is always in the unit position of the field and more than one
self-checking field can be checked per card.
The 5496 has a 64 character keyboard and operates at a speed of 60 columns per second.
The 5486 Card Sorter permits off-line sorting of the new IBM System/3 punch cards. The Sorter
is available in two models: Modell reads the new cards at 1,000 cards per minute and Model 2 is
fifty per cent faster at 1,500 cards per minute. The card hopper can accommodate 2,000 cards
and there are six 600-card stackers .

.3

SOFTWARE
The System/3 software support is designed to simplify the user programming tasks. It includes the
following elements, all operating under the supervision of the System/3 Control Program:
•

Card Programming System: RPG II Card Compiler, Card Utility Package.

•

Disk Programming System: RPG II Disk Compiler, Disk Assembler, Disk Sort Program,
Disk Utility Package.

The System Control Program includes such functions as supervisor, executive, and job control
routines. The Control Program is available with the hardware without charge. All of the other
programs are priced separately in line with IBM's new separate priCing policy•
• 31

RPG II
The Report Program Generators for both card and disk systems are designed to provide the user
with an easy method of writing programs. The new RPG compilers have 37 major functional improvements over previous RPG's including such features as internal debug facilities, more flexible
tables, a look-ahead facility, communications areas, program and file control, automatic overlays, square root subroutines, sort and collate routines, etc. For example, the disk sort is
capable of sorting indexed file by key. There is a roll in-roll out capability to service inquiry
messages.
The fixed core requirements for the disk RPG II is 3K memory with 4K required in a multiprogramming environment. Card RPG II requires 3K memory.

8/69

fA

AUERBACH

'"

(Contd.)

SUMMARY

.32

450:011.320

Disk Assembler
The Disk Assembler is a standard assembler package permitting the user, who does not wish to
utilize the RPG facilities, to write directly in either machine or symbolic language. The Assembler translates the assembly language instructions into machine instructions, assigns storage locations, and performs other functions that eventually result in an executable, machine-language program.
The minimum system requirements for the Disk Assembler consist of 12,288 bytes of main storage
in the 5410 Processing Unit, one 5203 printer with universal character set and interchangeable
chain cartridge features, one 5424 Multi-Function Card Unit, and one 5444 Disk Drive .

• 33

Sort and Utilities
A comprehensive package of sort and utility software is offered for both the card and disk systems.
In the card system, all of these functions are concentrated in one program. For a disk system,
the functions are broken up 'into separate programs.
The card utility package includes five basic programs: reproduce-interpret, sort, collate, a 96
column list, data recording and data verifying .

.4

APPLICATION CUSTOMIZER SERVICE
The IBM Application Customizer Service was developed specifically for System/3 users. The user
fills out a questionnaire relating to a particular application. He specifies the content and layout of
the records and reports, identifies the calculations required, and then chooses among the variety of
processing procedures. Related jobs can be linked, into an integrated family. The user then prepares punched cards from the questionnaire and IBM feeds the job specifications into an IBM/360
Model 20 at one of its Basic System Centers. The user must either have additional equipment to
punch the 80-column input cards for the 360 program or may engage the services of IBM to do the
job.
The specifications are checked for accuracy and completeness and the following tailored application
aids are produced for the user:
•

Edit Listings indicating the accuracy and consistency of the specifications.

•

Flowcharts defining the processing and clerical steps involved in each application.

•

A Data Dictionary defining the terms used in the particular application.

•

A Record Listing showing record formats.

•

A File Cross-Reference Listing referencing each record and the pertinent application.

•

Program Descriptions specifying detailed input specifications, calculation logic, and
output printing requirements for the programs to be ev~ntually coded in RPG or
assembler language.

•

Sample Reports showing the format of the final report according to the original
specifications.

The user then has the materials necessary to prepare the computer programs at his installation.
The preparation of these programs, the compiling and assembling, and the actual running of the
jobs is done at the user site. The customizer service is prepared at the Basic System Center.
Six major business application areas are provided through the customizer service (all of which
are separately priced):
•

Order Writing and Invoicing; pre-billing or post-billing, automatic backordering, automatic selection of item prices or discounts.

•

Accounts Receivable; open item or balance forward method

•

Inventory Accounting: stock status reports

•

Sales Analysis: reports classified by item, product class, customer or salesman.

•

Payroll: registers, paychecks, earning statement, etc.

•

General Ledger Accounting: internal or client accounting.

© 1969 Al!ERBACH Corporation and AUERBACH Info. Inc,

8/69

IBM SYSTEM/3

450:011.400

.4

APPLICATION CUSTOMIZER SERVICE (Contd.)
Besides the Application Customizer Service, IBM also offers comprehensive customer training
courses and system engineering services to accomodate the needs of the IBM System/3 users.
All of these services are separately priced. The education courses provide a good grounding in
basic System/3 concepts, application deSign, disk system design, RPG n programming
fundamentals, etc. System engineers are ready to do the job in its entirety for the user who does
not wish to acquire a programming staff or who do not wish to assign System/3 programming
tasks to his current programming staff. In addition, system engineers are available for
assisting the user in the preparation of the reports.

8/69

fA..
AUERBACH

450:221.101

S".'I"
EDP

IBM SYSTEM/3
PRICE DATA

Il,..n

IBM SYSTEM/3
PRICES

IDENTITY OF UNIT
CLASS

Model
Number

Feature

Name

Monthly Purchase Monthly
Maint.
Rental
$
$

$

PROCESSOR
Processing Unit* (Card Configuration):

5410
A2
A3
A4
A5
A6
5410

INPUTOUTPUT

8K bytes of core memory
12K bytes of core memory
16K bytes of core memory
24K bytes of core memory
32K bytes of core memory

310
410
525
755
985

15,200
20,100
22,575
37,000
48,275

27
29
29
30
30

435
535
650
880
1,110

21,325
26,225
28,700
43,125
54,400

73
75
75
76
76

155
255
155

8,075
9,700
8,075

47
47
47

Processing Unit (Disk Configuration):
A12
A13
A14
A15
A16

MASS
STORAGE

With
With
With
With
With

5444

With
With
With
With
With

8K bytes of core memory
12K bytes of core memory
16K bytes of core memory
24K bytes of core memory
32K bytes of core memory

Disk Storage Drive*
Modell: 2.45 million bytes
Model 2: 4.90 million bytes
Model 3: 2.45 million bytes

5440

Disk Cartridge

5424

Multi-Function Card Unit*:
Model A1: reads 250 cpm, punches 60 cpm,
prints 60 lpm
Model A2: reads 500 cpm, punches 120 cpm,
prints 120 lpm

5203

Printer*:
--

Modell: prints 100 lpm
Model 2: prints 200 lpm

-

171)

-

270

9,450

140

405

12,575

200

230
280

10,600
11.775

67
76

75
75
25
50
75
10

3,675
3,675
750
1,500
2.250
300

20
1

155

7,600

54

30
30

900
900

1
1

85
165

4,425
5,075

38
58

Options:
3475
4730
5559
5558
5560
8639
5496

Dual feed carriage
Interchangeable chain carriage
12 additional print poSitions
24 additional print poSitions
36 additional print positions
Universal character set
Data Recorder: punches 60 col. per sec.

-2
2
1

Options:
7061
7062
5486

Modulus 10 self check first digit
Modulus 11 self check first digit
Card Sorter:
Modell: reads 1, 000 cpm
Model 2: reads 1,500 cpm

* The 5410, 5444, 5424 and 5475 units are standard components of IBM System/3 card and disk configurations, and
these units cannot be ordered separately. Additional units also cannot be ordered.

o

1969 AUERBACH Corporation and AUERBACH Info, Inc.

8/69

450;221. 102

IBM SYSTEM/3

PRICES

IDENTITY OF UNIT
CLASS

Model
Number

INPUT
OUTPUT
(Contd. )

Feature

Name

Monthly
Monthly
Rental Purchase Maint.
$

$

$

7
10
10

210
490
490

1
3
1

100

4.700

32

40

2.250

7

Options:

1275
2370
7245
5471

Alpha sort
AUxiliary card counter
Sort suppress/digit select
Printer-Ke~board:

prints 15.45 characters

per secon .

5475
SYSTEM
SOFTWARE

Data Entry Keyboard
Card Programming System:

5701
5701

RGI
UTI

5702
5702
5702
5702

RGI
ASI
SMI
UTI

RPG II Card Compiler
Card Utility Program

35
10

Disk Programming System:

APPLICATION
PROGRAMS

8/69

RPG II Disk Compiler
Disk Assembler
Disk Sort Program
Disk Utility Program
Customized Application Packages:
Order Writing and Invoicing
Accounts Receivable
Inventory Accounting
Sales Analysis
Payroll
General Ledger Reporting

fA..
AUERBACH

45
75
10
10

250
225
225
205
265
180

-

-

-

--

---

-

-

-

-

/

HONEYWELL~

INC.

AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

AUERBACH
®
Print.A in II

c:.

A

501:011.100
•

_

STANDARD

EDP

Honeywell 400
Introduction

R[PORTS

I NTRODUCTI ON
§

OIl.

The H-400 is a small to medium scale business-oriented computer. It has a fair
range of conventional input-output and auxiliary storage units. Only one real option
(Multiply/Divide) exists so far as the central processor is concerned, so the computing
power of the unit is the same for most configurations. The H-4oo was first delivered in
1961 and is mainly used as an independent computer rather than as a supporting satellite for
larger systems. The system can be used to support the larger H-800 but such an application is comparatively unusual. Monthly rentals range from $5,000 to $14,000 and typical
systems are approximately $8,000.
o>mpatibility
The H-400 is the smallest of the Honeywell computers. The larger Honeywell
systems are the H-800 I and II (502:), the H-1400 (505:) and the H-1800 (503:).
There is complete programming compatibility between the H-400 and H-I400 systems,
which also share the same peripheral units, but there is no direct programming compatability between the H-400 and the H-SOO/1S00 systems. However, an H-400 simulator is
optionally available for use with the H-800 to permit H-4oo programs to be run on the H-SOO.
Hardware
The basic system, with no optional facilities, operates almost entirely serially
(Le., computation, input, and output are handled one process at a time and do not ove-rlap).
Simultaneous tape read and tape write operation is the only exception. Optionally, the
printer can be buffered so that the central processor can operate while the printer is
operating.
The processor, which has optional multiply/divide capabilities, uses binary or decimal arithmetic. Three address instructions ("ADD A, B, C" means ADD (A) to (B) and
place the result in C) are used and operands are in fixed word lengths (12 decimal characters
including sign, or 48 binary bits). The instruction repertoire is comprehensive and includes
especially good editing commands for translation of the 6-bit alphanumeric codes to and
from their decimal and binary equivalents. There is a powerful move command which allows
n words to be moved at a time. "n" can be of any size up to 4, 095.

" No variable length operations are possible. The processor also serves as the inputoutput controller. The system requires no additional controllers or buffers (beyond the
printer buffer) for this reason.
The core storage is available with 1,024, 2, 048, 3, 072, or 4, 096 4S- bit words.
Each 24-bit half of a word has a parity bit which is checked whenever the data is moved.
The store can accept words with incorrect parity from input-output devices. The processor
is made aware of this condition by a forced transfer of control to a fixed location. A paritychecking instruction is provided to find the incorrect word and correct its parity. Other
instructions are provided to implement techniques to correct the incorrect data. They are
part of an internal program -executed system called Orthotronic Control.
Up to eight magnetic tape units can be connected. The three magnetic tape unit
models available operate at 32, 000 characters or 4S, 000 digits per second, 64, 000 characters or 96, 000 digits per second, and SS, 666 characters or 133, 000 digits per second.
These units have pneumatic drives which handle tape more gently than mechanical drives-. A
feature of the H-400 (Orthotronic Control) enables it to ignore a taulty track when reading a

© 1963

by Auerbach Carporation and BNA Incarporated

Revised

7/63

501 :011.101

HONEYWELL 400
INTRODUCTION (Contd.)

§ OIl.

tape and to regenerate the correct data. Orthotronic Control is an error correction system
designed particularly to catch errors caused by tape skew. In contrast to read-aiter-write
error detection systems, Orthotronic Control has the advantage that it can cover errors
occurring during or after recording, either in storage or during reading. On the other hand,
it does not notice recording errors until later reading.
The printer operates at 900 lines per minute. A print storage option is available for
this unit that frees the processor for 98 per cent of the printing time. The IDM 1402 reader/
punch is now the card equipment normally used with the H-400, although some older installations are still using the converted version of the IDM 088 collator. The 1402 reads BOO cards
per minute and punches 250 cards per minute.
Punched tape equipment is also available; the reader operates at 500 or 1,000 characters per second, the punch at 110 characters per second.
Software
A number of programming aids are available for the H-400 system. These include:
(1) EASY I, a basic symbolic assembler for systems with I, 024-word stores.
(2) EASY II, a more complete assembler for systems with stores of 2,048 or more
words. This includes an input-output macro which is also used in other software systems, such as AUTOMATH and COBOL.
(3) A Sorting Generator and Merging Generator Routine. These are based on the
polyphase method, which has been pioneered by Honeywell.
(4) Disc File Programs which are presently under development.
(5) A COBOL-61 compiler for the H-400, which has just been released. This compiles on a 2K machine with a minimum of four tape units. The compilation time
is approximately one-half hour, which is good for a machine of this size. The
language facilities are fairly complete. The object programs are reported to
require approximately the same running time as those produced using normal
(EASY II) techniques.
(6) FORTRAN II (called AUTOMATH 400), a FORTRAN II compiler which has also
just been released. It includes a non-FORTRAN statement, OVERLAY, which
helps to overcome some of the limitations of systems with small storage (like
the H-400). It does a small amount of analysis of the coding and its context before creating the machine language and thereby improves the object time speed
of the program s. Subscripts are only allowed to two levels and error control of
the running program is not as strong as would be liked. Compilation times are
very good, approximately one hundred statements per minute. Object running
times are fllowed down by the need to simulate the floating point arithmetic.

501 :221.101
It ......

EDP

HONEYWELL H·400

It"'"

PRICE DATA

HONEYWELL H·400

i:ASS
I

IDENTITY OF UNIT
Model
Number

Feature
Number

PI\()C ESSl) H

PRICES
Monthly
Monthly
Rental Purchase Maint.

Name

$

$

$

Processing Unit
401-1
401A-1
401A-2

402-1
402-2
402-3
*413-3
*413-4
451
:'.IASS
STOHACE

Central Processor (includes 1,024 words
core storage, console, and power
unit (1) (2)
Central Processor (includes 1,024 words
core storage, console and power unit;
accepts 404-1 or 404-3 tape units) (2)
Central Processor (includes 1, 024 words
core storage, console and power unit;
accepts 404-2 tape units) (1)
Additional Memory Module (1,024 words)
Additional Memory Module (2,048 words)
Additional Memory Module (3,072 words)
Optional features:
Elapsed Time Clock
Real Time Clock
Multiply-Divide Option

of

3,975

178,875

347

of

4,215

189,675

347

of

5,215

234,675

44B

(3)

650
1,300
1,850

29,250
58,500
83,250

36
73
104

35
155
250

1,575
6,975
11,250

4
14
13

1,990

89,550

239

2,490

112,050

299

900

43,200

176

900

43,200

176

450

20,250

112

1,380
75

62,100
3,600

152
5

325

14,700

91

15
119
540
550
560

675
6,585
20,575
30,000
30,215

52
48
132
132

15

675

1

540

24,300

60

540

24,300

60

Disc Storage
-\60-0

Random Access Storage and Control (12.5
million characters) (1)
Random Access Storage and Control (25 million
characters) (1)

460-1
I!,(Pl"Tm"TPl"T

Magnetic Tape (4)
404-1

Magnetic Tape Unit (maximum number of tape
units per 400 system is 8)
High-Density Magnetic Tape Umt (maximum
number of tape units per 400 system is 8)
Economy Magnetic Tape Unit (maximum number
of tape units per 400 system is 8)
Tape Control Unit (1)
MagnetiC Tape Switching Unit

40-1-2
404-3
436-1
405

Punched Card
42:l-2
*423-2A

424-1
424-2
427
427-1
-427 -2A

High-Speed Card Reader (650 cards/min) (IBM
088 Model 3)
Pocket Selection Feature for the Model 423-2
Standard-Speed Card Punch (1)
High-Speed Card Punch (1)
Card Reader/Card Punch (1)
Card Reader/Card Punch (reads 800 cards/min;
punches 250 cards/min; includes Early Card
Read Feature) (5)
Pocket Selection Feature

1

Paper Tape
109
'410

Paper Tape Reader and Control (1,000 rows/
sec)
Paper Tape Punch and Control (110 rows/sec;
Modell accommodates 11/16-inch tape;
Model 2 accommodates 7/8- or I-inch tape)

(01969 AUERBACH Corporation and AUERBACH Info, Inc"

2/69

HONEYWELL H-400

501 :221.102

IDENTITY OF UNIT
CLASS

IN PUTOUTPUT
(contd. )

Model

Feature

Number

Number

Name

PRICES

Monthly
Rental

$

Monthly
PurchasE Matnt.

$

$

Printers
*418
422-3
422-3A
*422-3B
422-4
422-4A
*422-4B
450

Off- Line Printer Control (to be used with one
Model 404-1, 404-2 or 404-3 and one Model
422-3 or 422-4)
High-Speed Printer (900 lines/min; 120 out of
160 printing positions)
Vertical Spacing Option for the Model
422-3 (Allows spacing of six lines per
inch or eight lines per inch) (6)
Two-Speed Printing Option for the Model
422-3 (allows printing of 600 or 900
lines/min) (6)
High-Speed Printer (900 lines/minj 120 fixed
printing positions)
Vertical SpaCing Option for Model 422-4
(Allows spacing of six Hnes per inch
or eight lines per inch) (6)
Two-Speed Printing Option for the Model
422-4 (allows printing of 600 or 900
lines/min) (6)
Print Storage Option (for full simultaneous
operation of on-line printer with any other
Honeywell 400 operation)

1,550

69,750

120

1,550

74,400

347

100

4,800

20

40

1,920

8

1,050

47,250

236

100

4,800

20

40

1,920

8

390

17,550

20

Optical Reader

COMMUNICATIONS

440

O~tica1 Scanning Unit and Control (1)

2,530

121,440

426

481
484-3

Single-Channel Communication Control (1)
Multi-Channel Communication Control
(4 Bays) (1)
1
Communication Adapter \~
Communication Adapter (
Communication Adapter (1)

:JOO
1,210

13,650
54,450

24
121

25
25
30

1,125
1,125
1,350

3

485-IR
485-IT
485-IH

NOTES:
*No longer in production
(1) Not available on new orders.

(2) Not more than one card reader, one printer, one card punch, and/or one random access
unit may be attached to a Honeywell 400 system.
(3) Maximum memory capacity available on the Honeywell 400 is 4,096 words.
(4) 404-1, 404-2, and 404-3 Magnetic Tape Units cannot be intermixed on a single central
processor.
(5) The 427-1 is an IBM 1402 and will be supplied only if available.
(6) An in.ltallation charge of $250 will be made if this feature is field installed.

2/69

A ..

AUERBACH

3
3

502:011. 100
Honeywell 800
Introduction

I NTRODUCTI ON
§ 011.

The H- SOO is a medium to large computer system designed to process more than one
program at a time". This is an attempt to reduce the inefficiencies of individual programs,
which are usually input-output or central processor limited. In any installation, the degree
of success of multi-program operations depends upon how well the programs selected balance the sum of the demands on the central processor with the demands on the peripheral
units. In practice, installations with time- sharing programs operate an average of two programs at a time, with peaks of five or six. (The hardware is capable of sharing the central
processor time among up to eight programs.)
The H-SOO rents for between $lS, 000 and $40,000 a month, depending on the configuration, and size, is intermediate between the H-400 and the new H-1400 on the one hand, and the
H-lSOO on the other. The H-SOO uses the same data-codes as the H-400 and the H-1400, and
thus, magnetic tapes can be interchanged between these systems. The H-lSOO can run H-SOO
programs, as the H-SOO order code is a subset of the H-lSOO code.
The multi-running* feature of the H-SOO is particularly valuable where large volume
input-output files are processed with either relatively little or peaked internal processing.
Typical applications of this character are found in the insurance and utility fields. This
approach also permits a program mix which includes a series of scientific (low volume inputoutput) computation programs.
The manufacturer has undertaken the development of software which should encourage
more use of multi-running. A package has been released for controlling up to seven simultaneous conversions between cards, paper tape, magnetiC tape, and hard copy. The
elimination of the separate "program testing" executive system has been proposed because
many installations tend to retain it after testing has been completed rather than convert to
the different operating requirements of the standard production executive system.
The H- SOO has the capacity to execute 30,000 three-address instructions per second.
The computer uses a 4S- bit word, either as 44 bits plus sign character, 11 decimal digits
plus sign character, or 12 unSigned decimal digits. Alphameric characters can be stored
eight to a word, but cannot be used in arithmetic.
Decimal and binary computing facilities are available in the H- SOO, as are multiword transfers. which allow economical programming. However, the computer has no
facility for easy conversion of external data codes to internal code, or vice versa. All
shifts are right end-around shifts, so that editing is costly. An edit generator and several
standard routines are available, but most routines appear to be written for indiVidual cases.
The H- SOO storage is divided into two parts, a Control Memory with eight "program
groups," and a Main Memory which is divided into banks of 2, 04S 4S-bit words. The basic
H-SOO has two of these banks; larger units can contain up to 14. The eight program groups
are included in all cases, Each of these eight groups can control a separate program. A
total of 64 (eight per group) index registers are prOVided. The addressing structure is such
that while any program can reach or use any location in storage, it is necessary to use one
of a number of special addressing methods when referring to addresses in other program
groups or other blJnks. The index registers have restricted utility in that any base address
can be modified by no more than 256 positions.
·"Multi-running" is used in these Reports to describe the operation of a computer that is
simultaneously processing two or more independent programs. "Parallel programming" and
other terms are currently used to describe the same concept.

© 1963

by Auerbach Corporation and BNA Incorporated

2/63

HONEYWELL SPO

502:011.101
§

011.

INTRODUCTION (Contd.)

There are 8 input and 8 output channels, all of which can operate concurrently with
each other and the central processor. The peripheral units of the H- 800 can be arranged as
the user requires, with little restriction as to type or quantity. Honeywell-manufactured
units include a 900 line per minute printer (which is very similar to the Anelex Printer) and
magnetic tape units with character rates varying from 32,000 to 133, 000 alphameric
characters per second. Data communication units are being designed but no specifications
have been released.
The Honeywell printer is unusual in that it has 160 printing positions, any 120 of
which can be used at a time. Character and format selection is by plugboard. A paper tape
loop located in the printer provides paper feeding control. These features enable printing
two forms side by side and, where appropriate, the use of standardized print routines.
The Honeywell magnetic tape system is designed to allow the recovery of data lost
during writing, storage, or upon re-reading. This recovery is effected by forming and
checking "orthotronic control words" which are appended to each record on tape. The overheads involved in forming these words place an additional load on the central processor
when writing is in process. The size of this additional load varies from 2! per cent to 10
per cent, depending on the tape unit in use. No additional load is present during reading
operations.
Other available peripheral equipment includes: paper tape equipment capable of reading 1,000 characters per second and punching 110 characters per second; card readers
which operate at 250, 650, or 800 cards per minute; card punches which punch either 100 or
250 cards per minute, and mass-storage discs with capacities of up to 800 million alphameric characters.
The software provided with the H-800 includes an assembly language (ARGUS), a
FORTRAN II translator (AUTOMATH-800), and a business compiler (FACT). A COBOL
compiler and FORTRAN IV translator (AUTOMATH-1800) have been announced for 1963. All
of these are described in the language and translator sections.
The FACT compiler can handle files arranged as individual items, similar to COBOL
files; or files with "hierarchical" structure. This arrangement saves tape space by recording identical data in a number of consecutive items only once instead of a number of times.
A Sort package, using the cascade sorting method, is available for the H-800. Cascade sorting merges strings from all except one of the tape units available. thereby providing faster sorting.
An executive system able to control the operation of all program translators and production programs is provided.
The executive system presently in use is designed for batch processing through
assembly, and then running either serially under program testing methods, or in parallel in
production. The ordering and control during a production run is controlled by a schedule
which is created by a special run, but which relies very considerably upon the human skills
of the scheduler who sets up the basic data. The things to be considered vary conSiderably
from one installation to another, and the return which can be obtained from multi-running
depends in no small measure on the ability of the scheduler.
Running under the executive system causes no actual loss of time during production
running. as the executive program is not operating at this time. However, preparatory
runs consume approximately 15 minutes of running time to set up the schedule and program
tapes.

2/63

502:221.101

HONEYWELL H-800

PRICE DATA

HONEYWELL H·800
PRICES

IDENTITY OF UNIT
CLASS

Model
Number

Monthly

Feature

Name

Number

Monthly
Rental Purchase Maint.
$
$
$

Processing Unit

PROCESSOR

800
801
801-A
801-B
813-3
813-4
802

Central Processor (includes 4,096 words of
core storage, console, and power unit)
Floating-Point Option (1)
Floating-Point Option
Elapsed Time Clock
Real Time Clock
Additional Memory Module (available in units
of 4,096 words; maximum of seven 802
units per system. )

~
BOO-II

801-B
802

800-m

410,400
60,000
100,800
1,600
7,500

476
112
112
3
9

1,600

76,BOO

90

10,500
2,100
1,600

498,000
100,800
76,800

696
H2
90

10,000

475,650

602

6,100

275,000

730

8,100

365,000

920

900
900
450
75
2.000
3,100
2,000

43,200
43,200
20,250
3,600
96,000
148,800
96,000

176
176
112
5
105
163
105

1,950

93,600

215

(1)

Data processing system (consists of
801 Central Processor and Input-output
Control Center (lOCC), which includes
201, 202-1, 203A-1 or B-1, 207, 208, 212
and control for the 822-3 printer)
Floating-Point Option
Additional Memory Module
800-m

B,550
1,250
2,100
35
155

(1)

Data processing system (consists of 801
Central Processor, console, and
Power Unit; 212 On-Line Adapter;
201 Central Processor)
Disc Storage

MASS
STORAGE
860-1
860-2

Random Access Storage and Control (50
million characters)(l)
Random Access Storage an~ Control
(100 million characters) 1)
Magnetic Tape

INPUTOUTPUT
804-1
804-2
804-3
*805
803-1
803-2
803-3
* 836

Magnetic Tape Unit
.High-Density Magnetic Tape Unit
Economy Magnetic Tape Unit
Magnetic Tape Switching Unit
Tape Control
High-Density Tape Control
Economy Tape Control
Tape Control Unit (Controls one IBM 729n
Tape Unit)

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

HONEYWELL H-IOO

101:221,102

IDENTITY OF UNIT
CLASS

INPUTOUTPUT
(Contd. )

Model
Number

Feature
Number

Name

PRICES
Monthly
Monthly
Rental PurchasE Maint.
$
$
$

Punched Cards
Standard-Speed Card Reader (240
cards/min•• IBM 085)(1)
High-Speed Card Reader (650 cards/min. ,
IBM 088 Model 3)(1)
Standard-Speed Card Punch(l)
High-Speed Card Punch(l)
Card Reader-Card Punch(l)
Card Reader-Card Punch
(reads 800 cards/min. , punches 250 cards~
min. , includes Early Card Read Feature)( )
Card Reader Control(l)
Card Reader Control(l)
Card Reader Control (for 827-1)
Card Punch Control(l)
Card Punch Control(l)
Card Punch Control (for 827-1)(1)

823-1
823-2
824-1
824-2
827
827-1
807-1
807-2
807-3
808-1
808-2
808-3

125

6,900

33

325
154
490
550
560

14,700
7,335
20,575
30,000
30,215

91
68
48
132
132

950
1,100
1,100
1,050
1,150
1,150

45,600
52,800
52,800
50,400
55,200
55,200

50
60
60
60
65
65

690

33,120

76

690

33,120

76

1,050
1,250
1,450
1,550

50,400
60,000
69,600
74,400

55
100
115
347

100

4,800

20

40

1,920

8

1,700

81,600

85

1,950

93,600

150

1,700

81,60l)

85

1,950

93,600

150

Paper Tape
Paper Tape Reader and Control (1,000
rows/sec)
Paper Tape Punch and Control (110 rows/sec)
Modell accommodates 1l/16-inch tape;
Model 2 accommodates 7/8- or I-inch tape)

809
*810

~
806-1
806-2
806-3
*822-3
822-SA

*822-3B
811-1
811-3
811-4
811-6
*822-1

Printer Control(l)
Printer Control(l)
Printer Control (for 822-3)
High-Speed Printer (900 lines/min)
Vertical Spacing Option for the Model
822-3 (allows spacing of six lines per
inch or eight lines per inch)
~o-Speed Printing Option (600 or 900
lines/min)
Printer - Card Reader - Card Punch
Control(l)
Printer - Card Reader - Card Punch
Control(l)
Printer - Card Reader - Card Punch
Control(l)
Printer - Card Reader - Card Punch
Control(l)
(for use witli 822-3, 827-1)
Standard-Speed Printer (150 lines/min;
IBM 407)(1)

(3)

(3)

164

M!eetic Link Character Reader
833

Magnetic Link Character Sorter-Reader
Input Control Unit(l)

1,300

62,400

97

700
950

33,600
45,600

50
70

1,250

60,000

63

Controls
815
817

Real-Time Controls

COMMUNlCATIONS
812-1

2/69

Off-Line Output Auxiliary Control (1)
Off-Line Input-Qutput Auxiliary Control(l)

Real-Time Control Unit (non-simultaneous
input-output) (1)

A ..

AUERBACH

502:221.103

PRICE DATA

IDENTITY OF UNIT
CLASS

COMMUNICATIONS
(Contd. )

Model
Number

Feature
Number

Name

PRICES

Monthly
Monthly
Rental Purchase Maiot.
$
$
$

Real-Time Controls (Contd.)
812-2

Real-Time Control Unit (simultaneous
input-output) (1)

1,800

86,400

90

NOTES:

*No longer in production.
(1)Not available on new orders.
(2)The 827-1 is an IBM 1402 and will be supplied only if available.
(3)HoneyweU provides maintenance only.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

II

503:011.100
STANDARD

EDP

•

Honeywell 1800
Introducti on

REPORIS

INTRODUCTION
§

011.

The Honeywell 1800 is a large scale solid-state computer system designed to process
more than one program at a time. Based on the fast H -1801 central processing unit, the
1800 is program-compatible with the Honeywell 800 (Computer System Report 502:) and uses
the same peripheral devices and software systems. The principal differences between the
two systems are that the basic central processing unit is three times as fast on the Honeywell 1800 as on the 800, has twice as much core storage (8,192 words), and has been incre~sed in price by 90 per cent. The increased internal processing capacity (some 90, 000
three-address instructions per second) will be particularly useful when the system is simultaneously processing two or more independent programs.
The optional floating point hardware has been redesigned to use significantly fewer mem0ry cycles than its equivalent on the Honeywell 800. This places the Honeywell 1800 in the category of very fast scientific processors, on a par with the IBM 7094 Model II. New instructions
are available for conversions between fixed point decimal and floating point binary formats.
The Honeywell 1800 rents for between $30,000 and $60,000 per month, depending upon
the system configuration and size. The 1800 uses the same data codes as the smaller Honeywell 400 and 1400 systems, so magnetic tapes can be interchanged between them. There is,
however, no program compatibility between the 1800 and the 400 and 1400. The Honeywell
1800 can run Honeywell 800 programs without alteration, because the 800's instruction repertoire is the same as the 1800's repertoire.
An optional central processor, the Honeywell 1800-II, permits four magnetic tape units,
a card reader/punch, and a printer to be connected directly to the central processor without
intermediate adapters. These units can be used for off-line or on-line transcription. More
details on the 1BOO-II are given in the Summary Analysis which follows this Introduction.
The multi-running* feature of the Honeywell 1800 is an attempt to reduce the inefficiencies of individual programs, which are usually input-output or central processor limited,
by processing more than one program at a time. In any installation, the degree of success
of multi-program operations depends upon how well the programs selected balance the sum
of the demands on the central processor with the demands on the peripheral units. The hardware is capable of sharing the central processor time among up to eight programs. In practice, Honeywell BOO installations with time-sharing programs operate an average of two programs at a time, with peaks of five or six.
The multi-running capabilities are particularly valuable where large volume inputoutput files are processed with either relatively little or peaked (1. e., unevenly distributed)
internal processing. Typical applications of this character are found in the insurance and
utility fields. Multi-running also permits efficient processing of a program mix which includes a series of scientific (low volume input-output) computation programs.
The manufacturer has undertaken the development of software which should encourage
more use of multi-running. A package has been released for controlling up to seven simultaneous conversions between cards, paper tape, magnetic tape, and hard copy. The elimination of the separate "program testing" executive system has been proposed because many installations tend to retain it after testing has been completed rather than convert to the different operating requirements of the standard production executive system.
The Honeywell 1800 uses a 48-bit word, either as 44 bits plus sign character, 11 decimal digits plus sign character, or 12 unsigned decimal digits. Alphameric characters can
be stored eight to a word, but cannot be used in arithmetic.

*

"Multi-running" is used in these reports to describe the operation of a computer that is
simultaneously processing two or more independent programs. "Parallel programming"
and other terms are currently used to describe the same concept.

© 1963

Auerbach Corporation and Info, Inc.

Reprinted

9/63

503:011.101
§

OIl.

HONEYWELL 1800
INTRODUCTION (Contd.)

Decimal and binary arithmetic facilities and multi-word transfers allow economical
programming. However, the computer has no facility for easy conversion of external data
codes to internal code, or vice versa. All shifts are right end-around shifts, so that editing
is costly. An edit generator and several standard routines are available, but most routines
appear to be written for individual cases. Floating point arithmetic hardware is optional.
The H-800 storage is divided into two parts, a Control Memory with eight "program
groups, " and a Main Memory which is divided into banks of 2,048 48-bit words. The basic
Honeywell 1800 has 4 of these banks; larger units can contain up to 32. The eight program
groups are included in all cases. Each of these eight groups can control a separate program.
A total of 64 index registers (8 per group) are provided. The addressing structure is such
that while any program can reach or use any location in storage. it is necessary to use one
of a number of special addressing methods when referring to addresses in other program
groups or other banks. The index registers have restricted utility in that any base address
can be modified by no more than 256 positions.
There are eight input and eight output channels, all of which can operate concurrently
with each other and the central processor. The peripheral units can be arranged as the user
requires, with few restrictions as to type or quantity. Honeywell-manufactured units include
a 900 line per minute printer (which is very similar to the Anelex Printer) and magnetic tape
units with character rates varying from 32,000 to 133,000 alphameric characters per second.
Data communication units are being designed, but no specifications have been released.
The Honeywell printer is unusual in that it has 160 printing positions, any 120 of
which can be used at a time. Character and format selection is by plugboard. A paper tape
loop provides paper feeding control. These features facilitate printing of two forms side by
side and, where appropriate, the use of standardized print routines.
The Honeywell magnetic tape system is designed to allow the recovery of data lost
during writing, storage, or upon re-reading. This recovery is effected by forming and
checking "Orthotronic control words" which are appended to each record on tape. The overheads involved in forming these words place an additional load on the central processor when
writing is in process. The size of this additional load varies up to 3 per cent, depending
upon the tape unit in use. No additional load is present during reading operations.
Other available peripheral equipment includes: paper tape equipment capable of reading 1,000 characters per second and punching 110 characters per second; card readers which
operate at 250 or 800 cards per minute; card punches which punch either 100 or 250 cards
per minute; and mass-storage discs with capacities of up to 800 million alphameric characters.
The software provided with the Honeywell 800 can also be used with the Honeywell
1800. It includes an assembly language (ARGUS), a FORTRAN II translator (AUTOMATH800), a FORTRAN IV translator (AUTOMATH-1800), and a business compiler (FACT). A
COBOL-61 compiler has been announced for 1963. Software is described-in the Honeywell
800 report, Sections 502:161 through 502:191.
The FACT compiler can handle files arranged as individual items, similar to COBOL
files, or files with ''hierarchical'' structure. This arrangement saves tape space by recording identical data in a number of consecutive items only once instead of a number of times.
A Sort package, using the cascade sorting method, is available for the Honeywell
1800. Cascade sorting merges strings from all except one of the available tape units, thereby providing faster sorting. However, even with this increased sorting speed and the fast
tape units, it would be very inefficient to sort large files on the Honeywell-IBOO unless other
programs were proceeding in parallel.
An executive system able to control the operation of all program translators and production programs is provided.
The executive system presently in use is designed for batch processing through assembly, and then running either serially under program testing methods, or in parallel in

~::on: The oroeriog and ~trol

d=m"A

9/63
AUERBACH
®

OCtiOn

= I. controlled

by a .chOOnle

HONEYWELL 1800
§

503:011.102
INTRODUCTION (Contd.)

OIl.

which is created by a special run, but which relies considerably upon the human skills of the
scheduler who sets up the basic data. The things to be considered vary considerably from
one installation to another, and the return which can be obtained from multi-running depends
in no small measure on the ability of the scheduler.
Running under the executive system causes no actual loss of time during production
running, because the executive program is not operating at this time. However, preparatory
runs consume approximately 15 minutes of running time to set up the schedule and program
tapes.

© 1963

Auerbach Carporalion and Info, Inc.

8/63

•

503:221.101

1I1I,".
BDP

HONEYWELL H-1800

I(Plns

PRICE DATA

HONEYWELL H·1800
PRICES

IDENTITY OF UNIT

CLASS

Model
Number

Feature
Jl.,j"umber

PROCESSOR

Name

Monthly
Monthly
Rental Purchase Maint.

$

$

$

Processing Unit (includes core storage)
1800

--

1801
*1801B
*019
*1813-3
*1813-4
1802
1802-1

Central Processor (includes 8,192 words of
core storage, console, and power unit)
F1oatlng-Point Option
Memory Barricade
Elapsed Time Clock
Real Time Clock
Additional Memory Module (available in units of
8,192 words; maximum of three 1802 units
per system.)
Additional Memory Module (available in units of
16,384 words; maximum of two units per
system; basic requirement of three 1802
units)

16,050

770,400

920

4,300
250
35
155
3,200

206,400
12,000
1,600
7,500
153,600

241
14
3
9
180

7,500

360,000

421

17,500

835,650

,046

1800-III (2)
1800-III

Data Processing System (includes 1801 Central
Processor with console, and power unit; 212
On-Line Adapter; and 201 Cettral Processor

NOTES:
*No longer in production.
(1) For peripheral equipment see Honeywell 800.
(2) Not available on new orders.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

505:011.100

~

~

Honeywell 1400
Introduction

INTRODUCTION

Ii 011.

The Honeywell 1400 is a medium scale computer system oriented primarily toward
business data processing applications. Recent hardware developments make it possible to
adapt the system to a variety of real-time applications. Monthly rentals for H-1400 systems
range from about $9,000 to $18,000 and average around $13,000. Initial customer deliveries
were made in January, 1964.
Compatibility
The H-1400's throughput capacity places it in the middle of Honeywell's expanding
line of computers. The larger Honeywell systems are the H-800 (Computer System Report 502)
and the H-1800 (Report 503). The smaller systems are the H-200 (Report 507) and the H-400
(Report 501).
The Honeywell 400 and 1400 are fully program-compatible and, with a few exceptions,
offer the same range of peripheral units. There is no direct program compatibility between
H-400/1400 systems and either H-200 or H-800/1800 systems, though a simulation routine permits H-1400 programs to be run on an H-800 or H-1800.
All Honeywell computers can communicate with one another (but not with most competitive equipment) by means of a line of magnetic tape units using three-quarter inch tape.
Hardware
The central processor has facilities for both binary and decimal arithmetic. Both
multiply-divide instructions and floating point arithmetic are optional facilities, and all floating
point arithmetic is performed in the decimal mode. Three-address instructions are used (e. g. ,
"ADD A, B, C" means "add the contents of A to the contents of B and place the result in
location C"). The instruction repertoire is comprehensive and includes especially good editing
commands for translation of the 6-bit alphanumeric codes to and from their decimal and binary
equivalents. Except for the editing instructions, operand lengths are fixed at one 48-bit word.
A Honeywell 1400 word can hold one instruction, eight 6-bit alphanumeric characters, twelve
4-bit decimal digits (or eleven digits plus sign), sixteen octal digits, or a single 48-bit binary
data item. A powerful "move" command permits the contents of up to 4,095 word locations to
be moved by one instruction.
The effective core storage cycle time, 13 microseconds per 48-bit word, is 30 percent faster than that of the Honeywell 400, providing an internal processing capacity for 10,000
to 12,000 typical three-address instructions per second. Other improvements over the H-400
include an increase in core storage capacity from the H-400 maximum of 4,096 words to the
H-1400 maximum of 32,768 words, addition of the Floating Point and Card Storage options, and
increases in the number of printers (now 2, previously 1) and magnetic tape units (now 16, previously 8) that can be connected.
The core storage is available in multiples of 4,096 48-bit word locations; maximum
size is 32,768 words. Each 24-bit half of a word has a parity bit which is checked whenever the
data is moved. The store accepts words with incorrect parity from input-output devices. The
processor is made aware of this condition by a forced transfer of control to a fixed location. A
parity-checking instruction is provided to find the incorrect word and correct its parity. Other
instructions are provided to implement techniques to correct the incorrect data. They are part
of a system called Orthotronic Control, which is used primarily with magnetic tape units and
disc files.
The central processor serves as the main input-output controller in H-1400 systems,
thereby minimizing the need for additional controllers or buffers. A special central processor
model, however, must be used with the fastest magnetic tape units (88,666 six-bit characters
per second).

@1964 Auerbach Corporation and Info,lnc.

3/64

HONEYWELL 1400

505:011.1 01
INTRODUCTION (Conld.)

§ OIl.

The basic H-1400 system without optional facilities has very limitcd capabilities for
simultaneous operations. Except for a simultaneous tape reading and writing operation, computation, input, and output are handled one at a time and do not overlap. Optional buffer
features called Print Storage and Card Storage permit internal processing to be overlapped with
printing and/or card reading or punching.
Up to 16 magnetic tape units can be connected. The three available magnetic tape
unit models operate at 32,000 characters (or 48,000 digits) per second, 64,000 characters (or
96,000 digits) per second, and 88,666 characters (or 133,000 digits) per second. These units
have pneumatic drives which handle the tape more gently than mechanical drives. The
Orthotronic Control feature enables the H-1400 to ignore a faulty track when reading a tape and
to regenerate the correct data. In contrast to read-ruter-write error detection systems,
Orthotronic Control can correct errors occurring during recording, in storage, or during
reading. On the other hand, it does not notice recording errors until a later reading.
The printer operates at 900 lines per minute. The Print Storage option frees the
processor for 98 percent of the printing time. The IBM 1402 Card Read Punch is the card
equipment normally used with the H-1400. It reads 800 cards per minute and punches 250 cards
per minute. The Card Storage option allows card reading or card punching (but not both) to be
overlapped with processing.
Up to 5 input and 4 output general-purpose peripheral trunks are available for connecting any of the following devices:
•

Magnetic disc files (random access storage for up to
100 million alphameric characters per file unit).

•

Communication controls (process messages to or
from remote equipment).
.

•

Paper tape reader (500 or 1,000 characters per
second).

•

Paper tape punch (110 characters per second).

•

Optical scanner (196 to 312 documents per minute).

Real-Time Processing
The basic Honeywell 1400 system is designed primarily for standard batch processing
applications. Through the addition of communication controls and magnetic disc files, the
H-1400 can handle inquiry, data collection, and management control functions as well. Batchtype production programs can be interrupted as necessary to process incoming messages and
tra.'1smit the replies.
Three types of communication control units are available. Up to five such control
units, in any combination, can be connected to an H-1400. The 484 multi-channel control can
accommodate up to 56 communication channels and handle several messages simultaneously to
or from remote devices with speeds of up to 300 characters per second. The 481 single-channel
control is designed for lower message volumes and handles only one channel. The 480 control
handles the transfer of data between an H-1400 and another computer or a high-speed remote
device. The central processor's interrupt facility is used to initiate a transfer of data between
core storage and a buffer in the communication control whenever the buffer has been filled
(dvring input) or emptied (during output). Priorities can be established so that some routines
will be interrupted freely, other routines will be interrupted only to handle selected functions
of higher priority, and still other routines will never be interrupted.
A wide variety of remote input-output devices can be used in Honeywell 1400 realtime systems. Virtually any business data transmitter that can be connected to a telephone or
teleprinter circuit can be used. The remote equipment can be connected to the computer either
through a standard switched telephone network or through leased lines.
Software
Software for the H-1400 is the same as for the program-compatible H-400, with minor
modifications. Programs and programming systems available from Honeywell include:
•

3/64

EASY II, a standard assembler with symbolic addreSSing
and reJ ocatable output. It includes an input-output macro
facility which is also used in other systems, such as
COBOL-61 and AUTOMATH.

505:011.1 02

INTRODUCTION

§

INTRODUCTION (Conld.)

011.

•

A COBOL-61 compiler which can be used on any
H-1400 system with a minimum of four tape units.
The compilation time for typical programs is approximately one-half hour, which is good for a
machine of this size. The language facilities are
fairly complete. The object programs are reported
to require approximately the same running time as
those produced using normal (EASY II) symbolic
coding techniques.

•

A FORTRAN II compiler (called AUTOMATH 400) that
includes a non-FORTRAN statement, OVERLAY,
which helps to overcome some of the limitations of
systems with limited internal storage (like the H-400).
The compiler does a small amount of analysis of the
coding and its context and thereby improves the execution speed of the object programs. Only two levels
of subscripting are allowed, and the facilities for
detecting and handling errors at execution time are
limited. Compilation speed is high: approximately
one hundred statements per minute. Object program
execution times are slowed down by the need to simulate the floating point arithmetic on all H-400 machines,
but should be much improved when the Floating Point
option is available on H-1400 computers.

•

Sort Generator and Merge Generator Routines. These
are based on the polyphase method, which has been
pioneered by Honeywell.

•

Disc File Programs, which are currently under development to facilitate the programming of disc file
operations.

•

THOR (Tape Handling Option Routine), a general
routine for locating, copying, comparing, editing, and
correcting information on magnetic tape.

•

TABSIM, a "load-and-go" program that simulates the
functions of conventional punched card tabulating equipment, using a source language that is compatible with
IBM 1401 FARGO.

•

Mathematical and statistical routines, which handle
functions, conversions, programmed multiply-divide
and floating point arithmetic, and curve fitting.

•

PERT and Linear Programming Packages.

© 1964 Auerbach Corporotion and Info, Inc.

3/64

505:221.101
HONEYWELL H-1400

PRICE DATA

HONEYWELL H-1400
PRICES

IDENTITY OF UNIT

CLASS

Model
Number

Feature
Number

Name

Monthly
Monthly
Rental !Purchase Maint.

$

$

$

Processing Unit (includes core storage)

PROCESSOR
1401-1
1401-2
1401-3
1401-4
1402-0

1402

*1401-B
1451
1403
*1406
*1411
1423-2A
*1427-2A
1450

Central Processor (includes 4,096 words of
core storage, console, power unit; accepts
404-1 or 404-3 magnetic tape units) (1)
Central Processor (includes 4,096 words of
core storage, console, and power unit;
accepts 404-2 magnetic tape units)
Central Processor (includes 2,048 words of
core storage, console, and power unit;
accepts 404-1 or 404-3 magnetic tape units)
Central Processor (includes 2,048 words of
core storage, console, and power unit;
accepts 404-2 magnetic tape units)
Additional Memory Module (2,048 words;
must be first module added to 1401-3 and
1401-4; not more than one 1402-0 may be
added)
Additional Memory Module (4,096 words;
(1402-0 is a prerequisite on 1401-3,
1401-4); maximum of 7 modules per
system)
Optional Features:
Floating-Point Option (requires 1451
Option)
Mul tiply /Divide Option
Optional features for peripheral devices: (3)
Extended Tape Control (required to control
the 9th through 16th tape units) (2)
Storage & Control for Eecond Printer (for
on-line operation; requires 1450 option on the first printer).
Card Storage Option (to be used with one
Model 423-2 or one Model 427)
Pocket Selection Feature for the Model 423
Pocket Selection Feature for the Model 427-1
Print Storage Option

7,350

330,750

659

8,100

364,500

726

6,550

294,750

5R7

7,300

328,500

654

800

36,000

45

1,600

72,000

90

150

6,750

14

250

11,250

13

100

4,500

9

625

28,125

50

490

22,050

39

15
15
390

675
675
17,550

1
1
20

NOTES:
*No longer in production.
(1) A 1400 System must include a central processor and one magnetic tape unit.
(2) Up to 16 magnetic tape units may be controlled by a 1401 Central Processor that
includes a Model 1403 Extended Tape Control.
(3) These features are used in place of, or in addition to corresponding features for
the Honeywell 400 peripherals. The 1400 can accommodate all Honeywell 400 peripherals;
see the Honeywell 400 Price Data Sheet.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

-£. ........

510:011.100

~EDP

.

AUlIIAC~

HONEYWELL SERIES 200
SUMMARY

IU,IU

SUMMARY
.1

BACKGROUND
The Honeywell Series 200 line of computers consists of nine program-compatible central
processors - Models 110, 120, 125, 200, 1200, 1250, 2200, and 4200. The ninth processor in the
Series, the large-scale Honeywell 8200, offers compatibility not only with other members of the
Series 200, but also with the earlier Honeywell 800 and 1800 systems.
The Series 200 family of computer systems - with the exception of the Model 8200 - is
based upon an improved version of the original Honeywell 200 system. first delivered in July 1964.
The Honeywell 2200 system was the second entry in what has since become the Series 200 family.
Announced in 1964, the 2200 was first delivered in December 1965. The Honeywell 120, 1200, and
4200 systems were announced in February 1965. The 120 was first delivered in February 1966 and
the 1200 in January 1966. Delivery of the first 4200 system was made in March 1968.
Three additional models, fitting into the small-to-medium-scale ranges of the Series 200,
were anno\Ulced in October 1967. The new Model 125 was delivered first in December 1967, and
both the 110 and 1250 in August 1968.
The Honeywell 8200 formally joined the Series in June 1965. It provides compatibility
with the H-800 and H-1800 systems through use of a 48-bit word processing subsystem. The word
processor in the Model 8200 provides the hardware capability to r\Ul up to eight independent user
programs concurrenUy. A second 8200 subsystem, the variable-length field (VLF) processing
subsystem, provides compatibility with other members of the Series 200. First delivery of the
Honcywell 8200 is scheduled for 2nd quarter, 1969. Throughout the remainder of this Computer
System Report, all gen~Lal statements concerning the Series 200 refer to the H-8200's VLF processor only. Separate paragraphs are devoted to descriptions of the word processor and to the
overall performance of the Honeywell 8200.
The Honeywell 200, and the computer family that grew from it, had as a major marketing
goal the replacement of the slower, "second generation" IBM 1400 Series systems. With such a
goal, certain advances in computer system design. such as 8-bit character codes and extensive
multiprogramming facilities, were not seen as necessary inclusions in the line. To a large extent,
the instruction complement of the 1400 Series was incorporated in the Honeywell 200 Series, and
software routines were developed to resolve the minor incompatibilities between the instruction sets
of the two series. In order to keep abreast of its competition. Honeywell has since substantially
expanded its hardware and software product line and currenUy competes favorably with most third
generation small-to-medium-scale computers manufactured by its competitors.
The key software package released with the original Honeywell 200 centered around a
program called Bridge. the "Liberator" for 1400 Series users. This program accepted IBM 1401
object programs as input and generated Series 200 object programs after a fairly straightforward
translation process. Linkages to simulation subroutines were generated to resolve most discrepancies between the two machines. This program is still supported by Honeywell. to the extent it
is used.
Honeywell currenUy stresses a symbolic assembly language translator program called
Easytran as an alternative to the Bridge translator approach to conversion of IBM 1400 Series
programs. Through the use of Easytran. almost 100 per cent of the 1400 Series source language
statements can be correcUy translated to Honeywell's Easycoder assembly language. which can be
readily modified as part of normal program maintenance. Honeywell maintains that programs so
translated from IBM 1400 Series assembly languages will operate on Honeywell Series 200 systems
at least 80 per cent as effiCiently as programs origlnally written for execution on Series 200 systems. Honeywell's current Easytran translator converts IBM 1401 and 1460 programs for use with
any Series 200 system. A similar translator. Easyauto Symbolic Translator translates IBM 1410
and 7010 assembly language programs into source-language programs which can be assembled to
r\Ul either on an IBM 1410/7010 computer system or on a Model 1200, 1250, 2200. or 4200 computer system.
The Disk Liberation System enables users of IBM 1401, 1440. and 1460 disk systems to
convert to Honeywell Series 200 disk systems. Disk Liberatio~ includes routines for both translation of 1401/1440/1460 programs to Honeywell Type 258, 259. or 259A Disk Pack Drives.

C 1969 AUERBACH Corporation and AUERBACH Info, Inc.

Z/69

510;011.101

HONEYWELL SERIES 200

A user of IBM 1400 Series equipment who wants to "trade up" to new equipment is faced
with m:U\y important considerations when comparing offerings by Honeywell in its Series 200 and by
llll\! in its System/3GO. Among these considerations are the following:
•

Decimal arithmetic in Honeywell Series 200 processors is in many cases faster
than that in comparable processor models of the IBM System/3GO.

•

Conversion to Honeywell Series 200 computer systems can be accomplished with
little reprogramming via the program translation process - without sacrificing
many processing facilities in the new system. Conversion to IBM System/360
computers can involve either total reprogramming or "emulation" of the 1400
Series object programs. With the emulation technique, the full potential of the
emulating system cannot be utilized (although it is paid for), and the 1400 Series
programs must be maintained in their original languages.

•

The equipment delivery period for a Series 200 is generally shorter than for a
System/360.

•

The retraining of personnel familiar with 1400 Series equipment will be minimal
when converting to Series 200 equipment, since the processors within this series
use the same data structure and largely the same instruction sets as the IBM
1400 Series processors. Use of the System/3GO will require extensive retraining
of personnel

In this Introduction, a number of important topics are discussed.
dent, and can be read separately if desired. The topiCS are:

Each topic is indepen-

· 1 Background.
· 2 Central Processors.
·3

Peripheral Units.

·4

Software.

·5

Compatibility with the IBM 1400 Series.

. 6 Compatibility within the Honeywell Series 200 and with the Honeywell 800 and
1800.
·7
.2

Pricing Policy.

CENTRAL PROCESSORS
Nine central processors currently form the nucleus of the Honeywell Series 200. Honeywell considers that these processors - Models 110, 120, 125, 200, 1200, 1250, 2200, 4200 and 8200 - span
a range equivalent to that spanned by the IBM System/3GO Models 20 through 65. Listed in Table I
are certain central processor tasks and the times required to perform these tasks for each Seriel?
200 processor. Comparable execution times for the System/3GO processors can be found in Table I
of the IBM System/3GO report, Section 420:011.
All of the character-oriented Series 200 central processors use add-to-storage logic. There is no
addressable accumulator. Both instructions and operands can be of variable length. Operand
le'lgths are not specified in Series 200 instr1l.ctions; instead, most operations are terminated when
the processor senses a word mark, item mark, or record mark in the operand field. Table n
summarizes the principal distinguishing characteristics of the nine central processors of the
Series 200 .

. 21

Model 110
The Honeywell Model 110 is a card, tape or disk ~ oriented computer system, which can handle one
or two input-output operations simultaneously with computing. It has 6 index registers; the basic
core storage capacity is 4,096 characters, and is expandable to 32,768 characters. Core storage
cycle time is 4 microseconds per character. The Model 110 Disk system, consisting of one Disk
Pack Control and one or two Disk Pack Drives, provides 9. 2 million characters of on-line storage,

2/69

fA

AUERBACH
~

(Contd.)

SUMMARY

510:011.210

and can transfer data at the rate of 147KC. Peripheral equipment which can be connected to the
Model llO Processor consists of a printer. card equipment, and up to four tape units with a tape
control unit. The console, which is a control and communication center for use with the Model 110
system. includes a typewriter which transmits and records instructions and data between the operator and the computer. The Model 110 Processor can be connected to another Series 200 computer, or to a data communications network. The rental for standard Model 110 systems ranges
from about $2,405 per month for a 8K card system to about $4,520 per month for a 16K, 2-disk system. Deliveries of the Model 110 Processor started in August, 1968.

· 22

Model 120
The Honeywell Model 120 is primarily a tape- or disk-oriented computer system with the ability to
control two or three input-output operations concurrently with processing. Automatic processor
interrupt facilities are also provided. The Model 120 has 6 index registers and a core storage capacity of 2,048 to 32,768 characters. Core storage cycle time is 3 microseconds per charactcr. The
Model 120 Processor can be connected to any of the Series 200 peripheral deVices, to another Serics
200 computer, or to a data communications network.
The Model 120 is a general-purpose data processing system, able to operate either as an independcnt,
stand-alone system or as a satellite in an integrated operation. The rental for typical Model 120
systems ranges from about $3,800 per month for a 8K card system to about $3,400 per month for a
BK, 4-tape system. Deliveries of the Model 120 Processor started in February 1966.
The Model 120 contains built-in peripheral device control units to regulate the operations of a 450Ime-per-minute printer, a 400-card-per-minute reader, and a 100 to 400 card-per-minute card
punch. A built-in magnetic tape control unit is optionally available to control up to four 13. 3KC
magnetic tape units. In addition to the control units already mentioned, either of two optional features
permits the connection of up to six more standard Series 200 peripheral device control units.

· 23

l\lodel 125
The Model 125 is a faster version of the Model 120. It has a 2. 5 microsecond cycle time and can
control up to four input-output operations concurrently with processing. The Model 125 has up to six
index registers and a core storage capacity of 4,096 to 32,768 characters.
Like the Model 120, the Model 125 is a general-purpose data processing system, able to operate
either as an independent stand-along system or as a satellite in an integrated operation. Monthly
rental for a Model 125 Processor is about $600 higher than that of the corresponding Model 120
Processor. Delivery of the Model 125 is 90 days from the date of order.
The Model 125 contains integrated controls for printers and card devices. In addition, both magnetic
tape units and disk devices are available with this processor. Among the available magnetic tape
units are the new Honeywell 204B-15 and 204B-16 magnetic tape units, which transfer data at a rate
of 26,700 characters per second.

· 24

Model 200
The Model 200 is a card-, tape- or disk-oriented computer system with the ability to control either
three of four input-output operations concurrently with processing. It has 15 index registers and a
core storage capacity of 4,096 to 65,536 characters. Core storage cycle time is 2 microseconds per
character. The Model 200 Processor can be connected to any of the Series 200 peripheral devices,
to another Series 200 computer, or to a data communications network.
The Model 200 is suitable for use either as an independent, stand-alone system or as a satellite system in an integrated operation. The rental for typical Model 200 systems ranges from about $4,000
per month for a 8K card system to $14,600 per month for a 32K, B-tape system. Deliveries of the

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

HONEYWELL SERIES 200

510:01' .211

TABLE I: ARITHMETIC EXECUTION TIMES FOR THE SERIES 200 PROCESSORS
TASK (Times expressed
in mioroseoonds)

CENTRAL PROCESSOR MODEL
1250 2200
1200
200

4200

8200**

110

120

125

156

123

103

84

63

63

51

12 to 16

*'*'

*'*'

*''*

'*
'"

*''*

'*
'*

*'*

'*

156

123
3,100 (s)
3,700 (s)

103
2,580 (s)
3,080 (s)

360
900

63
360
900

51
244
600

'*
'#*

*'

84*
120*
149*

84*
120*
149*

'*'*

*'

60*
60*

60*
60*

Fixed Point Binary
c=a+b
c=axb
c=a/b
Fixed Point Decimal
o=a+b
c=axb
c=a/b

*'
*'

84
480
1,148

63

*

3.12 to 4.62
7.37 to 8.00
17.12 to 18.62

12 to 16 3.12 to 0.62
82 to 66 7.37 to 8.00
59 to 63 18. 12 to 18. 75

Floating Point Binary
'if
'if

c=a+b
c=axb
c=a/b

'*

#

*'

*'
'*

#

'**

'*

56* 9 to 13
4. 0 to 11.75
81* 13 Ix> 20 8.75 to 9.37
99* 18 to 22 16. 50 to 18. 00

Radix Conversion
Decimal to Binary
Binary to Decimal

'*
'*

40*
40*

15.75 to 17.25
11. 50 to 13. 00

14
9

Subroutine times; hardware facility not available.
Hardware facility not available; subroutine times not provided.
With optional feature.
Times are for 8200 Word Processor Subsystem; range of times reflects the use of maximum memory
bank interleaving to the use of no interleaving.
All decimal operands are considered to be five digits in length.

(s)

It

*
**
Note:

TABLE II: SUMMARY OF SERIES 200 PROCESSOR CHARACTERISTICS
Extended
:\tnin

"Iemol'\
Processor
'Iodel

110

~lp""d

Ic\cle

tIme)

4

Memory
Capacity

(thousands of
charactel'S}

Maximum
Number of

I/o Q>era-

Peripheral
Controllers
Accepted

CompuUng

Maxe No. of
Hone: Slmult.aneOU8 with

Advanced

Programmil'li'
Instructions

:l mlcroseconds per
character

200

seconds per
character
2 microseconds per
character

1250

seconds per
character
1.5 mtcroseconds per
character

-

-

-

·
·
·

·
·
·

-

-

-

-

-

-

-

-

Standard

-

-

-

standard

Standard

·

Standard

Standard

Standard

•

.
.

8

Standard

Standard

Standard

96

16

Standard

Standard

Standard

96

34

-

-

Standard

2 to 32

"

3

4 to 32

9

4

4 to 65

16

4

16 to 131

16

4

Standard

32 to 262

32

6

16 to 262

32

131 to 524

262 to 1, 04B

1 rnlcro2200

second per
character
750 rmno-

4200

8200 word
processor

seconds per
4 cbaracters
750 nanoseconds per
B characters

Memory
Protect
Facility

-

2

1. 5 micro1200

Scientific
Proces8ing
Instructions

·

1

2.5 mler\)125

Multiply
and Divide
Instructions

·

4 to 32

mlcroseconds
120

Financial
Edll
Instruction

Multlprogramming
and 8-BIt
Transfer
Capab!litv

·

•

•

•

·

Standard

·
·
•

- Feature not available on this model.
Feature optionally available.

(I:

2/69

fA..

AUERBACH

(Contd.)

510:011.240

SUMMARY

new Model 200 Processor began in November 1965. The noteworthy chaDgel between the original
Model 200 Processor and this version are the inclusion of automatic interrupt facilities, an 8-bit
compatibility feature, and multiply/divide ,instructions as standard equipment.
. 25

Model 1200
The Model 1200 is a tape- or disk-oriented computer system with the ability to control four inputoutput operations concurrently with processing. It has 15 or 30 index registers, an automatic interrupt system, and a core storage capacity of 16,3S4 to 131,072 characters. Core storage cycle time
is 1. 5 microseconds per character. A floating-point arithmetic option i8 available. The Model 1200
Processor can be connected to any of the Series 200 peripheral devices, to another Series 200 computer, or to a data communications network.
The Model 1200 is a general-purpose data processing system, able to operate either as a stand-alone
system or as part of a larger, integrated operation. The rental for typical Model 1200 systems falls
between $5,000 per month for a 16K, card system and $16,000 per month for a 9SK, 10-tape system.
Deliveries of the Model 1200 Processor began in January 1966. Compared to the Model 200 Processor, the Model 1200 offers increased core storage speed and capacity, the optional availability of
floating-point instructions for scientific applications, optional table look-up facilities that permit
IBM 1410 "Liberation," a Memory Protect facility that provides 15 additional index registers, and
an extended multi-programming and S-bit transfer hardware feature.

· 26

Model 1250
The Model 1250 adds to the capabilities of the Model 1200 with a main memory capacity ranging from
32,768 to 262,144 characters. Expanded I/O capabilities enable the 1250 to perform up to six inputoutput operations concurrently with processing. Core storage cycle time is 1. 5 microseconds. In
addition, the Model 1250 includes a 16- to 37- register control memory with a cycle time of 500
nanoseconds. Monthly rental for a Model 1250 Processor is about $200 higher than that of the
corresponding Model 1200 Processor. Deliveries of the Model 1250 began in July 1968.

· 27

Model 2200
The revised Model 2200 is primarily a tape- or disk-oriented computer system with the ability to
control either four or eight input-output operations concurrently with processing. It has either 15
or 30 index registers, an automatic interrupt system, and a core storage capacity of 16,384 to
262,144 characters. Core storage cycle time is 1 microsecond per character. All options currenUy
available with the Model 1200 are also available with Model 2200. The Model 2200 Processor can
be connected to any of the Series 200 peripheral devices, to another Series 200 computer, or to a
data communications network.
The Model 2200 is a general-purpose system, able to operate either as a stand-alone system or as
part of a larger, integrated operation. The rental for typical Model 2200 systems ranges from
about $8,900 per month for a 16K, 6-tape system to about $17,700 per month for a 70K, 10-tape
system. Deliveries of the Model 2200 Processor started in December 1965.

· 28

Model 4200
The Honeywell Model 4200 is a medium-scale disk-oriented computer system with the ability to control either 8 or 16 input-output operations Simultaneously with processing. It has either 15 or 30
index registers, automatic interrupt capabilities, and a core storage capacity of 131,072 to 525,288
characters.
The core storage cycle time is O. 75 microsecond, and 4 characters are accessed in parallel All
options currently available with the Models 1200, 1250 and 2000 are also available with the Model
4200, although the table lookup instructions are included as standard equipment with the 4200.
Additional features in the Model 4200 include an electrOnically alterable read-only memory with an
access time of 0.125 microseconds; interleaving of addresses across main memory modules; overlapping of main memory accesses; and true simultaneous memory access by both the central processor and the input-output controller. A separate maintenance processor monitors peripheral
operations and automatically assumes control of malfunctioning input-output equipment without
affecting the rest of the system.
The Model 4200 is designed as a general-purpose system, able to operate either as a stand-alone
system or as the master system in an integrated operation. The rental for typical Model 4200 systems ranges from approximately $25,000 to $36,000 per month. Deliveries of the Model 4200
Processor started in September 1968.

C 1969 AUERBACH Corporation and AUERBACH Info. Inc.

2/69

HONEYWELL SERIES 200

510:011.290

. 29

Model tl200
The 1\201 Central Processor consists of five independent functional units: a 48-bit word-oriented
processor, a variable-length-field (VLF) processor, a memory subsystem, an input/output controller, and a hardware/ software master control facility that coordinates the activity of the other units.
Communication between units is effected by means of program interrupts and control instructions.
The word-oriented processor utilizes a three-address instruction similar to that used with the older
Honeywell 800 and 1800 processors. Each instruction can consist of either one 48-bit word (normal
mode) or one 96-bit double word (extended mode). depending on whether 12-bit or 24-bit operands
have been specified. Performance of the 8200 is substantially higher than that achieved by the 800
or 1800, because of the faster main memory of the 8200 (750 nanoseconds as compared to 6 microseconds and 2 microseconds) and because of interleaved memory accesses among up to eight independent memory modules.
Eight independent groups of control registers, with 32 registers per group, permit up to eight independent user programs to be run concurrently, with minimal switching time required when
transferring, control between programs. Such transfer of control is performed entirely by hardware, with no central processor delay imposed. The 24-bit length of all control registers enables
explicit referencing of all main memory locations (from 262,144 to 1,048,576 characters), and
also provides facilities for indexed and/or indirect addressing.
The master control facility of the Honeywell 8200 is the ninth group of word processor control
registers. Master control consists of independent, specialized control registers that, together
with the master control program, coordinate the overall activities of the system. The master
control facility controls and monitors the interactions of the word and VLF processing subsystems and the input-ou1put controller. This facility also sets memory partitions so as to
allocate blocks of memory (512 word or 4,096 characters) to individual program control groups
as part of an effective memory protect scheme. In addition, the master control facility supervises the issuing of peripheral commands and device assignments, diagnoses program and
memory usage violations, and maintains identification information regarding protected memory
areas.
The variable-length-field (VLF) processor within the 8201 Central Processor qualifies the 8200
as a multiprocessor system whose facilities are shared by the two main (word and character)
processors. The VLF processor is identical in most respects with the Honeywell 4200 Processor
(see. 27 above) and provides the only real relationship with the remainder of the Honeywell Series
200. Model 4200 programs that are not time-dependent can be executed directly by the Model
8200 VLF processor without reassembly or recompilation. The VLF processor contains a single
group of 45 program control registers, and permits the concurrent operation of multiple
character-oriented programs through interrupt-controlled multiprogramming. The VLF processor can serve the 8200 word processor by controlling all input-output operations in a data
communications system in the manner of a powerful data communications controller. Except for
this case, Honeywell does not expect the Model 8200 word and VLF processors to work in constant communication while executing a single program. However, it is entirely possible and
highly desirable for all slow-speed input-ou1put data transfers and associated charactermanipulation operations to be performed by the VLF processor. The powerful processing capabilities of the word processor can then be better utilized by handling input-ou1put data batched
on high-speed magnetic tape or mass storage devices.
Like the Model 4200, the Model 8200 includes a separate maintenance processor to allow maintenance and repair of peripheral devices without interference with central processor and other
system operations.
One of Honeywell's design goals for the Model 8200 is to provide a powerful "third generation"
system for users of its earlier 800 and 1800 computer systems. However, its powerful multiprogramming facilities and the concurrent, independent operations of its major system component have already attracted new Honeywell customers. The rental prices for typical Model 8200
systems are expected to parallel those of the IBM System/360 Model 65 (i. e. , between $39,000
and $51,000 per month). Deliveries of the Model 8200 system will start in the second quarter
of 1969.

.3

PERIPHERAL UNITS
With the exeeption of the Model 110, the main peripheral units aVailable for use in Honeywell
Series 200 systems are listed in Table m. Any of the peripheral devices can be used with any
of the Series 200 processors, except Types 111, 113, and 114. Each peripheral device or controller requires one or two unit power loads and one or two a4dress assignments: one address
assignment for input or output only; two address assignments for two-way datatraIUlfers. Table 1I
shows the maximum number of peripheral controllers accepted by each processor model

2/69

A ..

AUERBACH

(Contd.)

510:011.300

SUMMARY

Four different series of magnetic tape units are available for Series 200 systems. The 204A
Series units are compatible with the 3/4-inch tape used in Honeywe11400, 1400, 800, and 1800
systems, while the 204B Series units provide compatibility with the 1/2-inch, 7-track tape used
in IBM 1400 and 7000 Series systems. The 204B-17 and 18 Tape Drives are for use with the
Model 110 only. The 204C Series units are 1/2-inch 9-track tape units that are compatible with
the IBM 2400 Series magnetic tape units used with the System/36G. The 204D Series consists of
1600 BPI, 9-channel tape units. The 204D-3 Magnetic Tape Unit is designed for use on the entire
Series 200 line; the 204D-5 is for use with the Model 4200 and 8200 systems, while the 204D-1 can
be used either on small, self-contained systems or on satellite systems of larger computers.
Mass Storage facilities are provided by the large Model 261/262 Disk Files, Models 155, 257,
258, 259, 273, and 275 Disk Pack Drives, Model 270A Random Access Drum, and Models 265,
266 and 267 High-speed Drums. The disk devices provide between 3. 68 and 300 million characters of storage with average access times ranging between 62. 5 to 117.5 milliseconds. The
Model 270A Drum Storage Unit has an average access time of 31 milliseconds and a data transfer rate of 111,000 characters per second.
The Model 288 Data Station consists of a remote terminal that can provide console typewriter input
and output, paper tape equipment, punched card and optical code readers, and a line printer. Its
function is to connect remote locations directly to the computer room and, potentially, to provide
direct connections between the computer itself and programmers or operators at the remote locations. Such a data communications link, which permits the processing of considerable volumes
of data without involving the costs of a full computer system at the remote location, clearly can
be used in many ways and can lead to major changes in a firm's data processing operations. The
implications of this unit in any particular situation require special consideration on a systems
level
In addition to the Model 235 Optical Journal Reader Control, the Model 238 Control Unit for the
IBM 1287 Optical Reader, and the Honeywell Teller Terminal Equipment, recent significant
additions to the Honeywell Series 200 line of peripheral devices include the Model 232 MICR
Reader/Sorter and Control, Model 233-2 MICR Control Unit for the Burroughs B103 Sorter/
Reader, the Model 234 Control for the Calcomp Series 500 Plotter, the Model 237 Bill Feed
Control for the IBM 1404 Alphanumeric Printer, and the 289 Series of Data Communication
Station units.

The Model 235 Optical Journal Reader Control permits a National Cash Register Model 420-1
or 420-2 Optical Journal Reader to operate on-line with any Honeywell Series 200 processor.
The NCR Model 420 Readers read printed journal tapes from cash registers, accounting machines, and adding machines at a maximum speed of 26 printed journal tape lines per second for
the Model 420-1 and 52 lines per second for the Model 420-2. Data is transferred from the
Journal Reader memory to the Series 200 processor by the Model 235 Control, which has a maximum data transfer rate of 83,300 characters per second.
The Model 238 Control Unit allows either Modell or Model 2 of the IBM 1287 Optical Reader to
be operated on-line with any Honeywell Series 200 processor. The Model 1287, when equipped
with appropriate optional feature., can read intermixed handwritten, machine-printed, imprinted,
and optical-mark flelds. It is described in detail in Paragraph 420:105. 127 of the IBM System/360
Report.
The Model 232 MICR Reader/Sorter and Control reads and sorts documents printed with magnetic
ink in standard E-13B font; the reading speed is 600 documents/minute.
The Honeywell Teller Terminal EiHi1ment for banking applications operates on-line with any
consists of one or two Model 375 Junction Control Units
Honeywell Seriell 200 processor,
which can connect up to 6 or 10 Model 370 Teller Registers to a transceiver that relays the information to the central processor in lO-bit modified ASCn code. Transmission speed at distances of up to 200 feet is 120 characters per second. In addition, a remote connection can be
used for asynchronous transmission at up to 1200 bits per second over half-duplex communication
lines. The Teller registers optionally can operate off-line to perform cross-footing and to
maintain teller transaction accumulations.
The display equipment available for the Series 200 is manufactured by Bunker-Ramo Corporation
and marketed by Honeywell. The line consists of five Display Stations. Models 303, 304, 311,
and 312 feature keyboard input and cathode-ray tube alphameric data display capabilities. The
fifth Display Station, the Model 317, does not have a keyboard and cannot be used to generate
inquiries. Editing feature!! permit a non-destructive cursor or entry marker to be moved to any
character position on the display screen for character Qorrection or deletion purposes. From 32
to 768 characters can be dililplayed at one time, with the exact display size determined by unit
model number and display arrangement. Each 4isplayed character is regenerated on the screen
more than 40 ttmes per lIecond.
C 1969 AUERBACH Corporation and AUERBACH Info. Inc.

2/69

510:011.301

HONEYWELL SERIES 200

The data communications equipment available for the Series 200 includes sihgle-line and multiline communication controls compatible with TELEX, 'tWX, and voice-grade lines. Although
the Series 200 uses 6-bit characters as its basic data format, a special pair of instructions in all
processors allows for the use of ASCD (7-bit) codes or other codes with up to 12 bits per character.
Instruction facilities are also available for dedRing message control characters at the time the
message is being transmitted. This permits simple matching of codes with most currently available data communications equipment.
.4

SOF1WARE
Software for the Honeywell Series 200 is currently organized in five principal levels of support,
which are designated in order of increasing power and comprehensiveness: Basic Programming
System, Operating System - Extended Mod I, Operating System - Mod 2 Extended, Operating
System - Mod 4, and Operating System - Mod 8. Within each of these five levels of software,
several system control routines, language processors. and service routines are included. Moreover, there are in many cases several versions of the same basic software program within each
major leveL One reason for the provision of multiple versions of a basic software program within
each operating system is the use of two-, three-. or four-character addresses within Honeywell
Series 200 instructions, depending upon the number of core storage locations that must be accessed
in a particular system. As a result, the same assembly language, for example, can have one or
more assembler programs associated with it, depending on the size of the instruction addresses to
be generated. Users of small-scale systems can choose to use exclusively software programs
that generate two-address instructions, and thus they can conserve core storage space by always
using instructions of relatively short length.
In addition, two levels of the Operating System - Extended Mod 1 and the Operating System Mod 2 Extended are available: one to operate on tape-oriented systems and another for systems
which include mass storage devices.
Most of the components of the Basic Programming System and the magnetic tape version of Operating System - Mod 1 have been available and in steady use with Series 200 systems for three
years or more. In view of new and powerful software facilities offered by its competitors,
Honeywell. has gone on and developed versions of the Operating System - Mod 1 and Operating
System - Mod 2 Extended to t81.t> advantage of the software flexibility and power inherent in mass
storage devices. Also, a message-mode facility for data communications applications has been
developed to enhance the real-time processing capabilities of both the Mod 1 and Mod 2 Extended
Operating Systems by permitting the transmission of entire message sequences, as opposed to
character-by-character transmission.
Operating System - Mod I, Operating System - Extended Mod I, and Operating System - Mod 2
Extended are available and in use. The new Operating System - Mod 4 provides II. multiprogramming system for Honeywell's large scale computers. while the Operating System - Mod 8
supports the Honeywell Model 8200 system.

2/69

fA..

AUERBACH

(Contd. )

510:011.410

SUMMARY

• 41

Basic Programming System
Programs within the Basic Programming System are designed to operate within from 4K to 12K
characters of core storage; these programs can uUllze up to 32K characters of core storage.
Versions of the component programs are provided in either the two- or three-character addressIng mode. The Basic Programming System is designed for punched card-oriented Series 200
users. Operator intervention is generally required to effect transition from one program to the

next.
The principal programs offered with the BasIc Programming System are listed below.
• A software system called Easytab permits efficient transition from off-line tabulating
equipment to a Series 200 computer system. Easytab consists of a number of uUllty
routines that perform common functions of unit record equipment, and a "built-in"
compact COBOL compiler that operates in an 8K-character storage environment..
• The Bridge and Easytran program translators, described in Paragraph. 5, provide
program compatibility with the IBM 1401.
•

The Easycoder Assemblers provide close source-language compatibility with the
assembly programs offered with the larger Series 200 operating systems. Eventual
transition to the larger operating systems is therefore simplified.

• The Simultaneous Media Conversion A program permits the concurrent operation of
up to three data transcription routines. Use of this program increases the effiCiency
of small Series 200 systems that serve as satellites to larger computer systems.
File-to-file transcription facilities are included for the following devices: both 3/4inch and 1/2-inch magnetic tape units, line printers, punched card units, and paper
tape equipment..
• A Report Program Generator (RPG) provides report-writing capabilities that are
similar to those provided with IBM 1401 Report Generators. The report format
specification sheetti are also similar to those used with the 1401 system•
. 42

Operating System - Mod 1
The Honeywell Series 200 Operating System - Mod 1 functions within from 12K to 65K characters
of core storage. Honeywell currently provides two Significantly different versions of this software package.
The magnetic tape-oriented version was announced with the original H-200 system as a package
called PLUS (program Loading, Updating, and Selection System). The larger tape-oriented
version of Operating System - Mod 1 contains many more facilities than the origiDal PLUS package. Independent programs are included within this new software to control automatic job sequencing, program retrieval and loading, overlay handling, and program library maintenance.
Some programs within the Operating System - Mod 1 are supplied in two or three versions. This
variety is provided because of the three- and four-character addressing options, the presence or
absence of floating-point hardware (for FORTRAN processors), and the desire to make available
a choice of language facilities at various program design levels. COBOL D, for example, operates
in a 16K-character memory environment and offers 270 language elements; COBOL H, by contrast,
operates in a 32K environment and offers 346 language elements.
The second and newer version of Operating System - Mod 1 makes extensive use of an auxiliary
mass storage device in its centralized software control system. The core-resident portion of the
Mass Storage Resident Operating System - Mod 1 requires only 1,500 characters of core storage.
Approximately 2.9 million characters of random-access auxiliary storage are also required to
utilize this system.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

510;011.420

HONEYWELL SERIES 200

TABLE III: PRINCIPAL SERIES 200 PERIPHERAL UNITS
PERIPHERAL TYPE
Random Access
Storage

Punched Card
Equipment

Punched Paper
Tape Equipment

2/69

MODEL
NO.

NAME

CHARACTERISTICS

155

Disk Pack Drive

3.68 million characters of storage;
117.5 msec.average access time.

257

Disk Pack Drive

4. 6 million characters of storage;
77.5 msec average access time.

261

Disk Files

262

Disk Files

150 million characters of storage;
78 msec average access time.
300 million characters of storage;
78 msec average access time.

258

Disk Pack Drive

4.6 million characters of storage;
92 msec average access time.

259

Disk Pack Drive

259A

Disk Pack Dri ve

9. 2 million characters of storage;
92 msec average access time.
9. 2 million characters of storage;
97 msec average access time.

273

Disk Pack Drive

275

Disk Pack Drive

270A

Random Access Drum

265

High-speed Drum

266

High-speed Drum

267

High-speed Drum

123
123-2

Card Reader
Card Reader

reads 400 cpm.
reads 600 cpm.

227

Card Reader/Punch

reads 800 cpm;
punches 250 cpm.

223

Card Reader

reads 800 cpm.

223-2

Card Reader

reads 1,050 cpm.

224-1

Card Reader/Punch

reads 300 cpm;
punches 50 to 270 cpm.

224-2

Card Reader/Punch

reads 400 cpm;
punches 91 to 360 cpm.

214-1

Card Punch

punches 100 to 400 cpm.

214-2

Card Reader/Punch

reads 400 cpm; punches
100 to 400 cpm.

209-2

Paper Tape Reader

reads 600 char/sec.

210

Paper Tape Punch

punches 110 char/sec.

A

18.4 million characters of storage;
62. 5 msec average access time.
18.4 million characters of storage;
62.5 msec average access time.
2. 6 to 20.4 million characters;
25 msec average access time.
2. 1 million characters of storage;
8.6 msec average access time.
4. 2 million characters of storage;
8.6 msec average access time.
4. 2 million characters of storage;
8.6 msec average access time.

(Contd. )

AUERBACH


510:0".421

SUMMARY

TABLE III: PRINCIPAL SERIES 200 PERIPHERAL UNITS (Contd.)
PERIPHERAL TYPE
Printers

Magnetic Tape
Units

Data
Communication

MODEL
NO.

CHARACTERISTICS

122
Series

Printers

print 300, 4501pm.

222
Series

Printers

print 450, 650, 950 or 1100 lpm.

229

Printer

prints 400 Ipm; available to
Educational market only.

203A
Series

Magnetic Tape Control
Units

for use with 204A-series tape.

204A
Series

Magnetic Tape Units

use 3/4-inch tape; compatible
with H-400/1400 and H-800/
1800 systems; transfer rates
of 32, 64, and 88 KC.

204B
Series

Magnetic Tape Units

use 1/2-inch tape; compatible
with IBM 729 series; transfer
rates of 4. 8 to 144 KC.

204C
Series

Magnetic Tape Units

use 1/2-inch tape; compatible
with IBM 2400 series; transfer
rate of 28. 9 KC.

2040
Series

Magnetic Tape Units

9 channel-transfer rates of 77,
154, and 224 KC.

281

Single-Channel
Communication Control

controls 1 half-duplex line at
up to 5,100 cbar/sec.

286

Multi-Cbannel
Communication
Control

controls up to 63 lines, at up to
300 char/sec per line; maximum aggregate data rate is
7,000 char/sec.

288-1

Central Control Unit

288-3

Central Control Unit

transmil!sion speed, 120
characters/ second •
transmission speed, 250 or 300
characters/second.

289-2

Printer and Keyboard

289-2A

Central Control Unit
Keyboard
Printer

289-3

289-4
289-5
289-7
other

NAME

220
Series

Paper Tape Reader
and Spooler
Paper Tape Puncb
and Spooler
Card Reader
COMole wUh
Typewriter

transmission speed, 10
cbaracters/second (100 wpm).
transmission speed, manual typing
speed input; no output.
transmission speed, 40 characters/
second (400 wpm) Input; manual
typing speed output.
transmission speed, 120 cps max.
transmission speed, 120 cps max.
transmission speed, approximately
100 cards/min max.
10 cps typing rate; 64 characters/
line.

C 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

HONEYWELl. SERIES 200

510:011.422

T ABLE III: PRINCIPAL SERIES 200 PERIPHERAL UNITS (Contd.)
MODEL
NO.

PERIPHERAL TYPE

NAME

233-2

MICR Reader/Sorter
and Control
MICR Control

234

Plotter Control

235

Optical Journal Reader
Control

237

Bm Feed-Printer
Control

238

Optical Reader Control

232

CHARACTERISTICS
reading speed, 600 documents/
minute
allows Burroughs BI03 Sorter/
Reader to operate with Series
200 Central Processor, Up to
1,560 documents/minute.
enables Colcomp series 500
Plotters to operate with Series
200 Central Processor.
directs operation of NCR 420
Optical Journal Reader with
Series 200 Central Processor.
enables IBM 1404 alphanumeric
printer to operate with Series
200 Central Processor.
controls the IBM 1287 Optical
Scanner when used with the
Series 200 system.

Provided with the Mass Storage Resident version of Honeywell's "Mod 1" software are assemblers,
COBOL and FORTRAN language processors, and general utility programs that are comparable to
offerings in the Tape Resident version of the same operating system. The chief difference between
the two versions lies in the fact that the Mass Storage Resident operating system provides integrated
system control routines. Among the functions performed by these routines are the foHowing:
loadin~.

•

Program

•

Automatic job stackIng.

•

Data management of sequential, indexed sequential, and direct access rnes.

•

Generating of common output from aU assemblers and compilers.

All versions of the Q:>erating System - Mod 1 are currently in use with Honeywell Series
200 systems .
• 43

Q:>erating System - Extended Mod 1
The Extended Mod 1 (MSR) Operating System is designed to provide foreground/background multiprogramming for Honeywell Models 1200, 1250, and 2200. When used in the multiprogramming
mode, program protection is provided through the hardware. The system is completely compatible with existing Mod 1 (MSR) and Mod 1 (TR) programs. The Mod 1 (MSR) systems programs
operate in the background; only 8K characters of memory are required by the Extended Mod 1
(MSR) Supervisor.
Both FORTRAN and COBOL Compilers are available for the Extended Mod 1 (MSR) Operating
System. The FORTRAN Compiler is an improved version of the FORTRAN Compiler H in the
Mod 1 (MSR) Q:>erating System; it compiles as a background program, and has object-time disk
I/O. The COBOL Compiler is an improved version of the Mod 1 (MSR) COBOL Compiler I; the
source and library update functions and compilation are executed as background programs, while
the object code generated by the compiler can be a background or foreground program. The
addition of the USASI standard CALL and CANCEL verbs eliminate the need for link loading under
the Extended Mod 1 Q:>erating System.
System generation is facilitated by the fact that two tapes are provIded with each system, a
Binary Run Tape (BRT), and a source tape. By specifying the system functions deSired, a user
can generate his particular system at system load time; the appropriate executable modules
(pre-assembled by Honeywell), are prepared for transcription to the disk and bootstrapped into
memory. Updating or modifying is accomplished by updating the source code tape, assembling
the appropriate modules and placing them on the BRT. Specific system tailoring for different
configurations is achieved easily by generating memory resident linkage tables; lengthy assemblers
are not necessary.

2/69

fA..
AUERBACH

(Contd. )

510:011 .....0

SUMMARY

.44

Operating System - Mod 2 Extended
The Mod 2 Extended medium-scale software system operates with Models 1200, 1250, 2200, and
4200 computer systems that have at least 49K and 64K characters of core storage, respectively,
and Honeywell's Optional Instruction Package.
Mod 2 Extended operates in a combined tape and mass storage environment, USing a minimum of
one disk device and three magnetic tape units. It can supervise the concurrent processing of
two independent stacked job streams.
FORTRAN and COBOL compilers and an assembler are provided in improved versions with Mod 2
Extended, offering additional language faciUties and efficiencies that are possible only with large
core memories. The output of all system programs is produced in common-format relocatable
program blocks. An Easyauto translator program converts mM 1410 and 7010 assembly language
programs into source-language programs which can be assembled and run either on an mM 1410/
7010 system or a Honeywell Series 200, Model 1200, 1250, 2200, or 4200 system. In addition,
Operating System - Mod 2 Extended provides a high-performance sort routine for use on magnetic
tape or mass storage devices, and several utility routines to facilitate the use of mass storage
devices.
The components of Qlerating System - Mod 2 Extended are divided into those of a supervisory
nature and those of a processing nature. The supervisory components include a Resident Monitor,
Transitional Monitor, Input/Output Control System, Communications Supervisor, and Data
Transcription Supervisor. Processing components comprise Language Processors for Assembler
L, Easycoder L, COBOL CompiIer L, FORTRAN Compiler L, and Report Generator L. The
important aspects of the system are stacked-job processing, program modularity, mass storage
support, and multiprogramming. Three methods of file access are available: sequential, direct,
and indexed sequential.

• 45

The Operating System - Mod 4
The Mod 4 C\>erating System is an integrated, modular, multiprogramming system which can
execute up to 20 independent programs concurrently. The system maintains dependent program
and data compatibility with the Mod 2 Extended Operating System. Programs written in COBOL,
FORTRAN, Easycoder, Assembler and RPG for the Mod 2 (Extended) can be run under Mod 4
without modification. The Mod 4 job control language and messages are compatible with those of
Mod 2 (Extended). Jobs can be entered from a remote terminal into any user partition, A basic
accounting routine, which the user incorporates into his system, is provided as a standard feature.
Four kinds of restarts are offered: system, job, job step, and checkpoint. Dynamic resource
allocation is performed on a job baSis. The user can allocate time to partltlons on a linear,
round-robin, or combination basis.
Memory in the Mod 4 Operating System is divided into two principal areas - one for the user,
one for the system. The system resident area can be as small as 40K.
The job control function in the Mod 4 Operating System is performed by four nonresident system
components, in conjunction with the reSident components. These nonreSident components are:
Input Reader, Scheduler, Transitional Monitor, and Output Writer. The Input Reader can run in
any available partition of 12K or more characters. When a job is initiated in a partition by the
Scheduler, the Transitional Monitor is brought into the partition to prepare the job for the
execution of its first job step. The Oltput Writer transcribes the contents of the Job Output
Files onto the device specified by the user.
The Mod 4 Input/Output File Control System consists of program modules which supervise all
input/output processing in the system and provide the overlapping of I/o operations necessary in
a multiprogramming environment.
The standard access method of Mod 4 services three types of file organizations: sequential,
indexed sequential, and direct.

C 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

510:011,450

.45

HONEYWELL SERIES 200

Operating System - Mod 8
The fifth level of software available for Honeywell Series 200 systems is designated the Operating
System - Mod 8. Mod 8 is specially designed for use with the hybrid Model 8200 system.
Facilities included within the Mod 8 control system will supervise the operations of both the word
processor of the Model 8200 and the variable-length-field (VLF) character processor (which bears
close resemblance to a Model 4200 processor). Mod 8 will function primarily as a mass storage
resident system, requiring at least two Type 259 Disk Pack Drives. The system also requires the
use of about 8,192 48-blt words of internal core storage for permanent residence. Other
peripheral devices required for operation of Mod 8 include at least six l/2-inch magnetic tape
units, one card reader. and one printer.
Among the more important functions of the Operating System - Mod 8 are the following:

. 46

•

Multiprogramming control for up to nine user programs.

•

Dynamic scheduling of computer usage according to job priorities and equipment availability.

•

Automatic allocation of memory; special register groups, and peripheral
equipment to the scheduled programs.

•

Handling remote batch processing and data communications operations.

•

Loading and relation of program segments •

Applications Software
Honeywell offers users of Series 200 computer systems an exceptionally large selection of
applications packages for use in the manufacturing, education, hospital, banking, utilities,
insurance, retailing, petroleum, and transportation industries. Honeywell estimates that approximately 70 such programs are currently available. The core storage requirements for use of
these programs generally range between 8K and 32K characters, although some programs
performing very large and complex tasks can require as much as 65K characters of core storage.
Use of these application programs is not restricted to anyone operating system; in general each
application will run under any Series 200 operating system provided the program's minimum core
storage and configuration requirements are met.
Many of Honeywell's application programs constitute completely integrated management information
and control systems composed of modular subroutines which can be selectively chosen according to
the specific needs of each user.
Some examples of Honeywell-supplied applications software include the following:
Manufacturing FACTOR: an integrated Factory Management System which includes subroutines for performance
accounting, inventory control, sales forecasting, management planning, production scheduling
and control, and on-line data collection,
General Distribution SALE: a package developed to direct the even flow of merchandise from warehouse to retail
store.
TIP and LIMIS: a comprehensive system of programs for proceSSing life insurance transactions. LIMIS has random access and data communication capabilities.
HOSPITAL INFORMATION SYSTEM: will include approximately 100 subprograms to perform business functions common to most hospitals.

2/69

A ..

AUERBACH

(Contd.)

SUMMARY

510:011.470

TRUMP: a teller-register unit monitoring program controUing more than 80 on-Une banking
operations and peripheral devices,
GENPAY: an application package consisting of 13 programs which handles an entire multipayroll operation in one cycle •
.5

COMPATIBILITY WITH THE IBM 1400 COMPUTER SERIES
The IBM 1400 Series of computers, consisting of the 1401, 1410, 1440, and 1460 processors, is
still the most widely-used computer family in the world. All of the 1400 Series processors use a
similar data format and instruction set, although each system has certain pecularities designed
to make it more suitable for particular functions; e. g., the larger memory size and overlapping
operations of the 1410, the orientation toward removable "Disk Pack" cartridges of the 1440, and
the higher internal speed of the 1460 relative to the basic 1401.
In the Series 200, Honeywell uses a basic instruction set that is largely identical with the basic
IBM 1400 Series instruction set. Honeywell has also adopted the 6-bit character structure used
in the 1400 Series, but uses eight bits plus a parity bit to store each 6-bit character. The added
bit provides improved punctuation facilities (record marks and item marks in addition to the
1400 Series word marks).
Execution of an IBM 1400 Series program on a Honeywell Series 200 processor can be performed
by means of a machine-aided conversion of the program and a subsequent manual checking
operation. Normally, the program can then be run on any equivalent Series 200 system that has
at least 4,096 extra core storage locations beyond those used by the original IBM 1400 Series
program. Production programs, compilers, assemblers, and industry packages can all be converted in this manner. Sorts, data transcription operations, report programs compiled by means
of Report Program Generators, and COBOL and FORTRAN programs are more commonly converted by using the original source programs or control cards to derive the necessary input to
the equivalent Honeywell software routines; this allows for more efficient use of the capabilities
of the Honeywell Series 200 hardware.
Two major types of machine-aided program conversion routines are available from Honeywell:
The Bridge and Easytran systems. Bridge conversions can be performed only upon machinelanguage 1401 or 1460 programs. The program is loaded into the Series 200 computer's core
storage in the normal way and is then processed by the special Bridge program. This results in
the production of a new program input deck and supporting documentation. The new deck replaces
the old program deck when the program is run on the Honeywell System, and otherwise operation
continues as before.
By contrast, Easytran conversions are performed upon assembly-language programs. The IBM
assembly input deck is processed and converted into a Honeywell assembly-language (Easycoder)
deck, and a supporting diagnosis of possible incompatibilities and other problems is produced.
After these potential problems have been investigated and the necessary actions have been taken,
the amended assembly deck is assembled by the standard Easycoder assembler. The resulting
Series 200 program can then be run in the normal manner. Honeywell currently recommends the
Easytran conversion process rather than the Bridge machine-language translation technique.
Both Bridge and Easytran conversion routines have already been widely and successfully used for
converting IBM 1401 and 1460 programs into Series 200 programs. Easytran routines that enable
IBM 1410 and 7010 programs to be run on Honeywe111200, 1250, 2200, and 4200 systems are
available. An Easyauto routine also converts IBM 1410 and 7010 programs to Mod 2 Extended
assembly language, but also enables the converted program to be assembled and run on either the
IBM 1410 or 7010 processors. Disk Sort C Is a key sort which sorts items from a mass storage
file. Extracted fields of the input item can be sorted, stored, and retrieved with the key fields.
An RPG-to-COBOL translator produces Series 200 COBOL source programs from 360-20 cards.
These programs are capable of execution within a tape or disk environment•

.6

COMPATIBILITY WITHIN THE HONEYWELL SERIES 200
With the exception of the 110, all the character-oriented computer systems in the Series 200 use the
same data format, the same instructions, the same perlptutral units, and the same software. As a
result, there is a fairly high degree of upwardrdownward program compatibility, subject to the

@

1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

510:011.800

HONEYWELL SERIES 200

usual limitations such as the availability of sufficient memory, peripheral equipment, and
peripheral address assignments and the degree of time-dependency within the programs.
Three areas of potential incompatibility do exist:
•

The addressing system. Different Series 200 processors use two-, three-,
or four-character addressing. This pr.ovides economies in both storage space
and execution time when addressing memory locations with short absolute
addresses. However, in moving a program from a small processor to a
larger one - particularly when the operating system requires the program to
be relocatable - aU address sizes and all address constants may need to be
changed by hand if the user wishes to make use of the extra storage space for
larger tables or other purposes.

•

The index registers. Different Series 200 processors may have none, 6, 15, or
30 index registers, depending on the model, the size of core memory, and the
inclusion of certain optional features. No provision has been made for simulating
the operation of index registers not present in the hardware, so this factor may
lead to incompatibilities in moving programs from a larger system to a smaller
one.

•

The instruction repertoire. A number of instructions are unavailable in the
small systems, optional in the medium-price systems and standard In the
larger systems. Any program written for a system that includes these
instructions may be unable to run on a Series 200 system that either cannot
or does not have all the necessary instructions provided in the hardware. No
proviSion bas been announced for automatically "trapping" these instructions
and using software routines to perform their functions.

The Honeywell Model 8200 computAr system is compatible with the rest of the Series 200 systems
to the extent that its VLF character-oriented processor can execute directly most non-timedependent programs that were originally written for execution on a Model 4200 system. The 8200
VLF processor does not have hardware for floating-point arithmetic operations. Therefore, when
Model 4200 programs attempt to perform floating-point arithmetic on a Model 8200 system, the
instructions involved will be trapped and their operations simulated by software.
In addition, the Model 8200 provides direct machine-language program compatibility with the
earlier Honeywell 800 and 1800 computer systems. All non-I/O instructions in H-800/1800
programs can be executed directly by the word processor of the Model 8200; all H-800/1800
input-output instructions are trapped and interpreted by the master control fac11tty of the Model
8200 and reissued in a format acceptable to the 8200 •

•7

PRICING
In October 1966, Honeywell announced a major revision of its pricing policy for Series 200
equipment. Notable changes in the new price schedule include a 2 to 4 per cent increase in
monthly rental charges, the establishment of uniform equipment purchase prices regardless of
when purchased, and an increase in the purchase price for some equipment. Monthly maintenance charges were not increased.

The single purchase price replaces Honeywell's former pollcyof giving discounts for the outright purchase of equipment, or for purchase during the first year of rental. The new purchase
prices, however, remain below those previously charged for purchase of equipment after one year
of rental. Moreover, a credit of a percentage of rentals paid to date can still be applied toward
the purchase of equipment.

2/89

A•

AUERBACH

510:221.101
HONEYWELL SERIES 200
PRICE DATA

HONEYWELL SERIES 200
IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

Name

PRICES

Monthly Rental $

Monthly
Purchase Maint.

1-Year

5-Year

910
1,055
1,310
1,575
1,780
1,990
2,195
2,400

810
940
1,170
1,405
1,590
1,775
1,960
2,145

39,600
45,900
57,150
68,625
77,625
86,625
95,625
104,625

93
105
118
130
140
150
160
170

75

75

3,240

6

50
225

50
200

2,160
9,900

4
15

1,085
1,345
1,605
1,815
2,020
2,225
2,430

970
1,200
1,435
1,620
1,805
1,990
2,175

47,250
58,500
69,975
78,975
87,975
96,975
105,975

190
203
215
225
235
245
255

1,385
1,640
1,905
2,110
2,320
2,525
2,730

1,235
1,465
1,705
1,885
2,070
2,255
2,440

60,300
71,550
83,025
92,025
101,025
110,025
119,025

190
203
215
225
235
245
255

880
1,000
1,270
1,520
1,780
1,980
2,190
2,390
2,600

795
910
1,140
1,375
1,605
1,790
1,975
2,160
2,345

37,155
42,55553,355
64,155
74,955
83,595
92,235
100,875
109,515

86
93
105
118
130
140
150
160
170

75

75

3,240

6

50
25
160
315

50
25
140
280

2,160
1,080
6,480
12,960

4
2
15
30

830
950
1,220
1. 470
1,730
1 930

745
860
1,090
1,325
1,555
1 740

34,905
40,305
51,105
61,905
72,705
81 345

86
93
105
118
130
140

$

$

Processing Untts(l)
PROCESSOR
111-1
111-2
111-3
111-4
111-5
111-6
111-7
111-8
1111
1113
1119
114-2
114-3
114-4
114-5
114-6
114-7
114-8
113-2
113-3
113-4
113-5
113-6
113-7
113-8
121-1
121-2
121-3
121-4
121-5
121-6
121-7
121-8
121-9
1011
1013
1014
1015
1016
121-0-1
121-0-2
1::!1-0-3
121-0-4
121-0-5
121-0-6

Model 110:
4, 096 Characters
8,192 Characters
12,288 Characters
16,384 Characters
20,480 Characters
24,576 Characters
28,672 Characters
32,768 Characters
Features for III Processors:
Advanced Programming
Instructions
Editing Instructions
Simultaneity for Second R/W
Channel
Model 110-2:
8, 192 Characters
12,288 Characters
16,384 Characters
20,480 Characters
24,576 Characters
28,672 Characters
32,768 Characters
Model 110-3:
8,192 Characters
12,288 Characters
16,384 Characters
20,480 Characters
24,576 Characters
28,672 Characters
32,768 Characters
Model 120:
2,048 Characters
4,096 Characters
8,192 Characters
12,288 Characters
16,384 Characters
20,480 Characters
24,576 Characters
28,672 Characters
32,768 Characters
Features for 121 Processors:
Advanced Programming
Instructions
Editing Instructions
8- Bit Code Handling Instruction
Sertes 200 Control Unit Adapter
Series 200 Control Unit Adapter
and R/W Channel
Model 120-0:
2,048 Characters
4,096 Characters
8,192 Characters
12,288 Characters
16,384 Characters
20 480 Characters

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

510:221. 102

PRICES

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

Name

Monthly Rental $
I-Year

5-Yeu

.,.,. rcilasE

$

Monthly
Maint •

Processing Units (Contd.)
PROCESSOR
(Contd. )

121-0-7
121-0-8
121-0-9
011-0
013-0
1015-0
1016-0
121A-3
121A-4
121A-5
121A-6
121A-7
121A-8
121A-9
1015-3
1016-3
126-1
126-2
126-3
126-4
126-5
126-6
126-7
126-8
126-2-9
126-2-10
126-2-11
126-2-12
1011
1013
1014
1017
1018
127-2
127-3
127-4
127-5
127-6
127-7
127-8
1017
1018
201
202-1

2/69

Model 120-0: (Contci. )
24,576 Characters
28,672 Characters
32.768 Characters
Features for 120-0 Processors:
Advanced Programming
Instructions
Editing Instructions
Series 200 Control Unit Adapter
Series 200 Control Unit Adapter
and R/W Channel
Model 120-3:
8, 192 Characters
12,288 Characters
16,384 Characters
20,480 Characters
24,576 Characters
28,672 Characters
32,768 Characters
Features for 121A Processors:
Series 200 Control Unit Adapter
Series 200 Control Unit Adapter
with R/W Channel
Model 125(2):
4,096 Characters
8,192 Characters
12, 288 Characters
16,384 Characters
20,480 Characters
24,576 Characters
28,672 Characters
32,768 Characters
40,960 Characters
49, 152 Characters
57,344 Characters
65,536 Characters
Features for 126 Processors:
Advanced Programming
Instructions
Editing Instructions
8- Bit Code Handling Instructions
Simultaneity for Third R/W
Channel
Auxiliary R/W Channel (Requires
Feature 1017)
Model 125-3:
8,192 Characters
12,288 Characters
16,384 Characters
20,480 Characters
24,576 Characters
28,672 Characters
32,768 Characters
Features for 127 Processors:
Simultaneity for Third R/W
Channel
Auxiliary R/W Channel (Requires
1017)
Model 200:
2, 048 Characters
2, 048 Character Memory Module.
(must be first module added to
201; not more than one 202-1
may be added)

A•

AUERBACH

2,295

89,985
98,625
107,265

150
160
170

75

75

3,240

6

50
160
315

50
140
280

2,160
6,480
12,960

15

2,140
2,340
2,550

1,925
2.110

1,595
1,850
2,110
2,315
2,520
2,725
2,930

1,900
2,085
2.270

1,440
1,670

4
30
205

2,455
2,640

66,865
77,615
88,365
96,965
105,565
114,165
122,765

160
315

140
280

6,480
12,960

J5
30

1,270
1,540
1,790
2,050
2,305
2,515
2,715
2,925
3,405
3,870
4,190
4,505

1,150
1,380
1,G15
1,845
2,080
2,265
2,450
2,635
3,045
3,460
3,745
4,025

53.535
64,335
75,135
85,935
96,775
105,415
114,055
122,695
138,600
157,500
170,520
183,330

130
143
155
165
175
185
195
210
225
240
255

75

75

3,240

6

50
25
160

50
25
140

2,160
1,080
6,480

4
2
15

60

50

2,160

4

1,945
2,200
2,455
2,720
2,925
3,130
3,335

1,755
.1,985
2,215
2,455
2,640
2,825
3,010

81,485
92,235
102,985
113,950
122,550
131,150
139,750

230
243
255
265
275
285
295

160

140

6,480

15

60

50

2,160

4

1,000
130

895
120

41,040
5,400

76
6

218
230
240
250
260
270

lIB

510:221.103

PRICE DATA

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

Name

PRICES
Monthly Rental $
I-Year

5-Year

Monthly
urchasl! Matot.

$

$

Processing Units (Contd.)
PROCESSOR
(Contd. )

202-2

201-1
202-3

202-4

202-5

201-2-1
201-2-2
201-2-3
201-2-4
201-2-5
201-2-6
201-2-7
201-2-8
201-2-9
201-2-10
201-2-11
201-2-12
011
012
013
015
016
010
013
015
016
1201-1
1201-2
1201-3
1201-4
1201-5
1201-6
1201-7
1201-8
0191
1100
1100A
1114

Model 200: (Contd. )
4,096 Character Memory Module
(202-1 1s a prerequisite; maximum of seven 202-2 Memory
Modules per system)
2,048 Characters (multiply/
divide capability)
2, 048 Chs.racter Memory Module
(must be first module added to
the 201-1; not more than one
202-3 may be added)
4,096 Character Memory Module
(202-3 i.s a prerequisite maximum of seven 202-4 Memory
Modules per system)
8, 192 Character Memory Module
(seven 202-4 Modules are a
prerequisite; maximum of four
202-5 Memory Modules per
system)
4,096 Characters
8,192 Characters
12,288 Characters
16,384 Characters
20,480 Characters
24,576 Characters
28,672 Characters
32,768 Characters
40,960 Characters
49,152 Characters
57,344 Characters
65,536 Characters
Features for 201 and 201-1
Processors:
Advanced Programming
Instructions
Program Interrupt (standard on
201-1)
Editing Instructions
Eight Additional Unit Loads of
Power
One Auxiliary Read/Write Channel
Features for 201-2 Processors:
Advanced Programming
Editing Instructions
Eight Additional Unit Loads of
Power
One Auxiliary Read/Write Channel
Model 1200:
16,384 Characters
32,768 Characters
49,152 Characters
65,536 Characters
81,920 Characters
98,304 Characters
114,688 Characters
131,072 Characters
Features for 1201 Processors:
Optional Instruction Package
Scientific Unit(3)
Scientific Unit
Storage Protection

265

235

10,800

13

1,265

1,130

51,840

96

130

120

5,400

6

265

235

10,800

13

320

285

12,960

15

1,410
1,680
1,940
2,205
2,470
2,735
3,005
3,270
3,585
3,905
4,225
4,540

1,255
1,500
1,735
1,970
2,205
2,440
2,680
2,920
3,200
3,485
3,775
4,055

57,240
68,040
78.840
89,640
100,440
111,240
122,040
132,840
145,800
158,760
171,720
184,680

106
119
131
144
156
169
181
194
209
224
239
254

115

105

4,320

8

60

50

2,160

4

105
160

95
145

3,890
6,480

7
12

60

50

2,160

4

115
105
160

105
95
145

4,320
3,890
6,480

8
7
12

60

50

2,160

4

2,775
3,745
4,495
5,305
5,780
6,210
6,585
6,960

2,480
3,345
4,015
4,735
5,160
5,545
5,880
6,210

112,320
146,880
177,120
205,200
224,640
241,920
257,040
272,160

196
236
271
304
326
346
369
386

60
460
580
60

55
410
530
55

2,160
19,350
24,750
2,160

4
40
40
4

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

HONEYWELL .RID 200

11 0:111.' Got

PRICES

mENTlTY OF UNrl'
CLASS

Model
Number

Feature
Number

Name

Monthly Rental $
I-Year

5-Year

~..rchasE

$

Monthly

Maint.
$

Processing Units (Contd.)
PROCESSOR
(Contd. )
1120

1251-1
1251-2
1251-3
1251-4
1251-5
1251-6
1251-7
1251-8
1251-9
1251-10
1251-11
0191
1100
1100A
1114
1120

2201-1
2201-2
2201-3
2201-4
2201-5
2201-6
2201-7
2201-8
2201-9
2201-10
2201-11
2201-12
0191
1100
1100A
1115

1117
1121

4201-3
4201-4
4201-5
4201-5A
4201-6
4201-7
4201-8
4201-9

2/69

Features for 1201 Processors:
(Contd. )
Extended Multiprogramming and
8-Bit 'Il'ansfer (Requires Feature
1114)
Model 1250:
32,768 Characters
49,152 Characters
65, 536 Characters
81,920 Characters
98,304 Characters
114,688 Characters
131,072 Characters
163,840 Characters
196,608 Characters
229,376 Characters
262,144 Characters
Features for 1251 Processors:
Optional Instruction Package
Scientific Untt(3)
Scientific Unit
storage Protection
Extended Multiprogramming and
8-Bit Transfer (Requires
Feature 1114)
Model 2200:
16,384 Characters
32,768 Characters
49,152 Characters
65,536 Characters
81,920 Characters
98,304 Characters
114,688 Characters
131,072 Characters
163,840 Characters
196,608 Characters
229,376 Characters
262,144 Characters
Features for 2201 Processors:
Optional Instruction Package
Scientific Unit(3)
Scientific Unit
Second Input/Output Sector
(additional four R/W Channels,
16 I/o Address Assignments,
and 16 Unit Loads of Power)
storage Protection
Extended Multiprogramming and
8- Bit Transfer (requires
Feature 1117)
Model 4200:
131,072 Characters
196,608 Characters
262,144 Characters
262,144 Characters
Two-way Interleaving
327, 680 Characters
393,216 Characters
458,752 Characters
524,288 Characters
Four-way Interleaving

A•

AUERBACH

225

200

9,900

35

3,975
4,725
5,535
6,010
6,440
6,815
7.190
8,000
8,640
9,205
9,725

3,545
4,215
4,935
5,360
5,745
6,080
6,410
7,135
7,705
8,210
8,675

156,880
187,120
215,200
234,640
251,920
267,040
282,160
319,400
354,400
390,000
424,000

252
287
320
342
362
385
402
442
482
526
566

60
460
580
60
225

55
410
530
55
200

2,160
19,350
24,750
2,160
9,900

4
40
40
4
35

3,775
4,835
5,785
6,695
7,545
8,285
8,820
9,345
10,355
11,155
11,680
12,215

3,370
4,315
5,170
5,975
6,735
7,395
7,875
8,345
9,245
9,960
10,430
10,910

153,360
196,560
235,440
272,160
306,720
336,960
358,560
380,160
421,200
453,600
475,200
496,800

284
334
379
422
462
497
527
552
599
637
662
687

60
460
580
115

55
410
530
105

2,160
19,350
24,750
4,320

4
40
40
8

60
115

55
100

2,160
5,000

4
17

13,285
15,410
17,535
17,975

12,300
14,270
16,240
16,650

567,005
657,725
748,445
767,160

590
680
770
790

19,660
21,785
23,915
26,040

18,205
20,175
22,140
24,110

839,165
860
929,885
950
1,020,605 1,040
1,111,325 1,130

510:221.105

PRICE DATA

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

PRICES
Monthly Rental $

Name

I-Year

5-Year

Monthly
Purchase Maint.
$

$

Processing Units (Contd.)
PROCESSOR
(Contd. )

1101
1116

1118
4214A
4214B
4215
8201-1
8201-2
8201-3
8201-4
8205-1
8205-2
8201-B
8214

8215
8272
212
212-1
213-3
071
213-4
215
220-2
220~3

008

Features for 4201 Processors:
Scientific Unit
Third Input/Output Sector
(additional 8 R/W Channels,
16 I/O Address Assignments,
and 16 Unit Loads of Power)
Extended Multiprogramming and
8-bit Transfer
Two Buffered Input/Output
Sectors (requires 1116)
Two Additional Buffered Input/
Output Sectors (requires 1116
and 4214A)
High Speed Third Input/Output
Sector (requires 1116)
Model 8200:
262,144 Characters
524,288 Characters
786,432 Characters
1,048,576 Characters
Features for 8200 Processors:
Adapter for one 3/4-lnch Tape
Control (803-1, -2, or -3)
Adapter for two 3/4-lnch Tape
Controls (803-1, -2, or -3)
Scientific Unit
Expanded Input/Output Capability
(additional 18 Simultaneous
R/W Channels, 48 I/O Address
Assignments, and 48 Unit
Loads of Power)
High-Speed Third Input/Output
Sector
Standby Console Typewriter
Features for all Models except 110:
On-line Adapter (for connection
of Series 200 processor and
801 or 1801)
Central Processor Adapter (for
connection of any two Central
Processors in the Series 200)
Interval Times
Interval Selector
Time-of-Day Clock
Communication Switching Unit
Console (replaces control panel)(5)
Console (replaces control panel;
also for Model 110
Pin- Feed Drive

,

535
535

495
495

22,685
22,685

40
40

155

145

6,810

12

405

375

17,625

19

355

325

15,275

17

155

140

6,580

18

26,520
35,465
44,725
53,670

24,555
32,835
41,415
49,695

1,161,220
1,552,900
1,958,400
2,350,080

1,920
2,560
3,230
3,880

360

335

12,000

30

670

620

20,000

55

790
1,580

730
1,465

34,560
69,120

60
120

205

190

8,930

24

325

290

13,500

30

530

475

22,500

50

425

380

18,000

40

95
60
215
90
325
325

85
55
190
80
290
290

3,600
2,250
9,000
3,375
13,500
13,500

8
5
20
8
30
30

25

25

1,125

5

370

330

14,910

55

610

545

24,925

60

410

365

17,100

60

610

545

24,925

60

670

595

26,800

75

Disk Storage
MASS
STORAGE

155
157B
258B
257
257-1

Disk Pack Drive (3.6 million
char. )(6)
Disk Control (for up to two 258B
Disk Pack Drives)(6)
Disk Pack ~rive (4.6 million
char. )(6
Contl'Ol for 258 and 259 Disk Pack
Drtves
COntrQI for 258 and 259 Disk Pack
Drives (6- and 8-bit transfer
modes)

cD 1969 AUERBACH Corporation and AUERBACH Info,lnc.

2/69

HONIYWILL ....D 200

110:221.101

PRICES

IDENTITY 011' UN1T
'CLASS

MASS
STORAGE
(Contd. )

, Feature
Model
Num. . , Number
!

Name

Monthly
- reba8e Malnt.
s-Year
$
$

Montbly Reatal •

I-Year

Disk Storage (Contd.)

*Z5'lA
257B
257~1

258
074
079
259
074
079
*259A
259B
074
079
M4005
260
077
261
078
262
078
273
275
278

Control for 259A Disk Pack
Drive
Control for 259B Disk Pack
Drive
Control for 259B Disk Pack
Drives (6- and 8-bit tranafer
modes)
Disk Pack Drive (4.6
million characters)
Write Protect Switch
Central Processor Finished
Disk Pack Drive (9.2
million characters)
Write Protect Switch
Central Processor Finished
Disk Pack Drive (9. 2 (3)
mtl1lon characters)
Disk Pack Drive (9.2
million characters)
Write Protect Switch
Central Processor Finished
Disk Pack for use in 258, 259,
259A, and 259B Disk Pack
Drives
Control for 258 and 259 Disk Pack
Drives and for 261 and 262
Disk Files
a-bit Transfer Mode
Disk File (150 million
characters)
Heat Exchanger
Disk File (300 million
characters)
Heat Exchanger
Disk Pack Drive (18.4 wilton
char.)
Disk Pack Drive and Control
(147.2 mUlion char.)
Disk Pack Drive and Control
(280 million char.)

610

545

24,925

60

610

545

24,925

60

670

595

26,800

75

410

365

17,100

60

20
30
575

20
30
515

900
1,350
24,000

2.50
4.00
75

20
30

20
3fl

900
1,350

2.50
4.00

5'15
575

515
515

24,000
24,000

75
75

20
30
15

20
30
15

900
1,350
490

2.50
4.00

675

600

2&,100

90

6'0
4,270

50
3,825

2,475
166,000

10
610

130
7,260

115
6,500

5,500
297,000

15
760

130
775

115
695

5,500
31,500

15
70

4,235

3,780

172,200

615

5,370

4,795

218,400

615

1,340

1.195

54,600

156

2,270

2,025

92,400

264

2,425

2,165

98,700

282

825
205

735
180

33,600
8,400

96
24

1,340
310
1,150
30
985

1,195
275
1,030
25
880

54,600
12,600
50,070
1,125
41,625

156
36
166
4
138

1,665

1,490

70,425

233

--

Drum
-265
266
267
260-1
072
260-2
073
270
075
*270-1
*270-2

2/69

High-Speed Drum (2.1 million
characters)
High-Speed Drum (4.2 mtUlon
characters)
High-Speed Drum (4.2 million
characters)
Control for 265 and 266
High-Speed Drums
Angular Position Indicator
Control for 267 High-Speed
Drums
Angular Position Indicator
High-Speed Drum
Track Protect (3)
Random Access Drum Storage
and Control (2.6 m1llion characters)
Random Access Drum Storage
and Control (5.2 million characters)

A•

AUERBACH

5t 0:221.1 07

PRICE DATA

PRICES

IDENTITY OF UNIT
CLASS

MASS
STORAGE
(Contd. )

Model
Number

Feature
Number

Name
Drum (Contd.)

*270-3
075
270A-1
270A-2
270A-3
075A

Monthly Rental
l-Year

$

5-Year

",rchase

$

Monthly
Maint.

$

2,345

2,095

99,225

328

30
1.150

25
1.030

1,125
50,070

5
166

J5.

1,950

1,740

84,630

280

J7.

2,745

2,450

119,400

394

35

30

1,370

5

310

275

13,500

60

25
60

25
50

1,050
2,160

2
5

20
450

15
450

670
20,250

2
100

900

900

43,200

155

900

900

43,200

155

375

335

15,120

74

325

290

12,960

63

530

475

21,600

105

480

425

19,440

95

750

670

30,240

147

425

380

17,280

84

50

50

2,100

5

640

570

25,920

126

850

760

34,560

168

25

25

1,300

2

295

265

12,375

58

235

210

10,125

48

20

15

670

2

Random Access Drum Storage and
Control (7. 8 million characters)
Track Protection(3)
Random Access Drum Storage and
Control (2.6 million characters)(7)
Random Access Drum storage and
Control
2 million characters)( )
Random Access Drum Storage and
Control
8 miUion characters)( )
Track Protection
Magnetic Tape

INPUTOUTPUT
103D
056
1056

Tape Control (for up to four 204B-17
and 204B-18 Tape ~nits; includes
one 204B-17 Unit)( )
Dynamic Tape Addressing
IDM Magnetic Tape
Compatibi1i~

1059
204A-1
204A-2
204A-3
204B-1
204B-2
204B-3
204B-4
204B-5
204B-7
055
204B-8
204B-9
054
204B-ll
204B-12
1059

DenSity Switch( )
Magnetic Tape Unit (3/4-inch tape;
31,760 char.lsec.)
Magnetic Tape Unit (3/4-inch tape;
63,520 char./sec. )
Magnetic Tape Unit (3/4-inch tape;
88,800 char./sec.)
Magnetic Tape Unit (1/2-inch tape;
primary unit; 200/556 bitS/inch;
7,200/19,980 char./sec.)
Magnetic Tape Unit (1/2-inch tape;
secondary unit; 200/556 bits/inch;
7,200/19,980 char./eec. )
Magnetic Tape Unit (1!2-inch tape,
primary unit) (200/556 bits/inch)
(16,000/44,400 char./sec.)
Magnetic Tape Unit (1/2-inch tape,
secondary unit) (200/556 bits!
inch) (16,000/44,400 char./sec. )
Magnetic Tape Unit (1!2-inch tape)
(200/556 bitS/inch) (24,000/66,600
char./sec. )
Magnetic Tape Unit (l/2-inch tape)
(556/800 bitS~inCh) (19,980/28,800
char./see.)( )
1200 Bits/inch Recording
Density
Magnetic Tape Unit (1/2-inch tape)
(556/800 bitS/inch) ~44,400/
64,000 char./sec.) 8)
Magnetic Tape Unit (l!2-inch tape)
(556/800/1200 bitJ/inch) (668600/
96,000/144,000 char./sec)()
1800 Bits/inch Recording
Density
Magnetic Tape Unit (1/2-inch tape,
primary unit) (200/556 bitS/inch)
(4,800/13,320 char/sec)(3)
Magnetic Tape Unit (1/2-inch tape,
secondary unit) (200/556 bits/
inch) (4,800/13,320 char/sec)(3)
Density Switch(9)

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

HONEYWELL .RID 100

ItO:UI.'.

mENTlTY OF UNIT
CLASS

Model
Number

Feature
Number

INPUTOUTPUT

(Contd. )

Monthly Rental $
l-Year

5-Year

Magnetic Tape Unit (1/2-iDCh tape:
secondary unit)(6)
Magnetic Tape Unit (1/2-1Dch : r i
secondary unit; 556/800 bits
1Dchi 13,300/19,200 char./

175

Magnetic Tape Unit (l/2-1Dch tape;
secondary unit; 200/556 bits/
1Dchi 96600/26,700 char./
sec. )(1 )
Magnetic Tape Unit (1/2-1Dch tape:
niue-channel, primary 1mlt; 800
bits/tDCh: 28, 800 eight-bit
char. /sec. )
Magnetic Tape Unit (1/2-1Dch tape;
niue-channel, secondary unit;
800 bits/inch: 28,800 eight-bit
char./sec.)
Magnetic Tape Unit (l/2-illCh tape:
9 channel: 800/1600 bits/bwh:
38,400/76,800 char/sec)
Magnetic Tape Unit (l/2-iDCh tape:
9 channel: 800/1600 b1ts/1Dch;
38,400/76,800/153,600 char/sec)
Magnetic Tape Unit (1/2-inch tape;
9 channel; 800/1600 bits/bIOh;
38,400/112,000/224,000 char/sec)
Tape Control (for up to four 204A-1
units)
Tape Contr~l (for up to four 204A-2
units)
Tape Control (for up to four 204A-3
units)
Tape Control (for up to eight 204B-1
ana 204B-2 or 204B-3 and
204B-4 units)
mM Format Feature (provides
end of file recognition)
mM Code CompatiblUty
Feature (BCD code
translation)
Dynamic Tape Addressing
Tape Control (for up to eight 204B-5
units; no interrupt)
mM Format Feature (provides
end of file recognition)
mM Code CompaUbiUty
Feature (BCD code
translation)
Dynamic Tape Addressing
Tape Control (for up to eight 204B-5
units)
Dynamic Tape Addressing
Tape Control (for up to eight 204B-7
or 204B-8 units)
mM Format Feature (provides
end of file recognition)
mM Code Compatibility
Feature (BCD code
translation)
Dynamic Tape Addressing

Monthly
I..._rchase Matnt.

$

$

155

7,650

45

280

255

12,175

58

320

295

14,400

67

480

425

20,250

95

480

425

20,250

95

395

350

15,960

76

570

510

23,100

110

825

740

33,600

160

295

265

12,375

28

295

265

12,375

28

425

380

18,000

40

455

405

18,360

42

60

55

2,250

5

60

55

2,250

5

25
455

25
405

1,050
18,360

2
42

60

55

2,250

5

60

55

2,250

5

25
455

25
405

1,050
18,360

2
42

25
455

25
405

1,050
18,360

2
42

60

55

2,250

5

60

55

2,250

5

25

25

1,050

2

Magnetic Tape (Contd. )
204B-18
204B-14

204B-16

204C-13

204C-14

204D-1
204D-3
204D-5
203A-1
203A-2
203A-3
203B-1
050
051
056
203B-2
050
051
056
203B-2A
056
203B-4
050
051
056

2/69

Name

PRICES

fA•

AUERBACH

510:221.101

PRICE DATA

PRICES

IDENTITY OF UNIT
CLASS

INPtTTOtTTPUT
(Contd. )

Model
Number

Feature
Number

Name

Monthly
PUrchasE Maint.
5-Year
$
$

Monthly Rental $
I-Year

Magnetic Tape (Contd. )
203B-5
057
059
056
203B-6
050
051
056
203C-7
2030-1
2030-3
2030-5
103
103A
103B
059
1055
056

Tape Control (for up to four 204B-11
and 204B-12 units)
mM Magnetlc Tape Compatibility
Density Switch
Dynamic Tape Addressing
Tape Control (for up to eight 204B-9
units)
mM Format Feature (provides
end of file recognition)
mM Code Compatibility
Feature (BCD code
translation)
Dynamic Tape Addressing
Tape Control (for one 204C-13 unit
and one 204C-14 unit)
Tape Control (for up to eight 2040-1
units)
Tape Control (for up to eight 2040-3
units)
Tape Control (for up to eight 2040-5
units)
Tape Control (for up to four 204B-11
and 204B-12 Tape Units; includes
one 204B-11 Unit)(10)
Tape Control (for up to four 204B-13
and 204B-14 Tape Units; includes
one 204B-13 Unit)(10)
Tape Control (for up to four 204B-15
and 204B-16 Tape Units; includes
one 204B-15 Unit)(10)
Density Switch
mM Magnetic Tape
Compatibility
Dynamic Tape Addressing

325

290

12,960

30

60

55

2,250

5

20
25
455

15
25
405

670
1,050
18,360

2
2
42

60

55

2,250

5

60

55

2,250

5

25
390

25
350

1,050
15,750

2
35

725

645

29,400

70

825

740

33,600

80

930

830

37,800

90

470

420

19,440

63

515

465

21,490

73

555

505

24,300

75

20
60

15
50

670
2,160

2
5

25

25

1,050

2

210
30
210
265
30
255

185
25
185
235
25
225

9,000
1,125
9,000
11,475
1,125
10,800

50
6
50
65
6
24

50
40
(12)

50
40

2,280
1,800

5
3

225

200

9,450

21

40
(12)

40

11,800

3

50
125
35

50
125
35

2,280
5,690
1,350

5
13
5

100

]00

4,565

10

Punched Card
123
1043
123
123-2
1043
*207
*040
*017
*017-1
*208
*017
*017-1
*060
*061
*062
*062

Card Reader (400 cards/min)(6)
51-Column Adapter
Card Reader (400 cards/min)(10)
Card Reader (600 cards/min)(10)
51-Column Adapter
Card Reader Control (for read side
of 227)
Direct Transcription
Stacker Select
Three-Stacker Select (017
prerequisite)
Card Punch Control (for punch side
of 227)
Stacker Se lect
Three-Stacker Select (017
prerequisite)
Direct Transcription
Hole-Count Checking
Punch-Feed Read (modification
of the punch)
Punch- Feed Read (modification
of the Gontrol unit)

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

HONEYWELL 8RIEI ZOO

510:221.110

IDENTITY OF UNIT
CLASS

INPUTOUTPUT
(Contd. )

Model
Number

Feature
Number

Name

PRICES
Monthly

Monthly Rental $
I-Year

-_chase Malm.
6-Yelll'
$
$

Punched Card (Contd. )
208-1
064
066
208-2
064
066
214-1
214-2
223
043
044
223-2
045(13)
044
*224-1
*065
*224-2
*227

Card Punch Control (for 224-1 or
224-2 or 214-1)
Direct Transcription
High-Speed Skip (214 only)
Card Read/Punch Control (for 224-1
or 224-2 or 214-2)
Direct Transcription
High-Speed fldp (214 only)
Card Punch, 100-400 oards/min
Card Reader/Punch, 400 cards/min
100-400 cards/mm
Card Reader and Control, 800
cards/min
51-Column Adapter
Direct Transcription
Card Reader and Control, 1050
cards/min
90-Column Card Reading
CapabiUty (requires
Feature cards)
Direct Transcription
Card Reader/Punch (1442-1), 300
cards/min/50-270 cards/min
Reject Stacker
Cai'd Reader/Punch (1~42-2). 400
cards/min!91-360 cards/min
Card Reader-Card Punch (800
cards/min/250 cards/min) with
Early Card Read Feature

160

145

6,750

15

30
30
245

25

1,125
1,125
10,125

3

25
215

3

22

30
30
325
375

25

1,125

25
290
335

1,125
13,500
15,750

'14

325

290

13,500

74

45
30
445

45
25

4

395

1,800
1,125
18,000

90

125

125

5,025

15

30
320

25
320

1,125
19,900

:I
34

25
440

440

1,100
21,050

1
41

660

660

35,610

140

295

265

12,375

34

350

315

14,625

34

245

215

10,125

28

465
61>

415
50

20,250
2,250

120
12

520
60

465
50

22,500
2,250

124
12

160
75

145
70

6,'150
2,925

15
16

30

25

1,125

3

770

695

33,750

186

850

760

36,000

198

100

100

4,500

25

25

:3
3
8'7

3

Paper Tape
*209

Paper Tape Reader and Control
(600 frames per second)
Paper Tape Reader and Control
(600 frames per second)
Paper Tape Punch and Control
(120 frames per second)

209-2
210

Printers
122-1
1035
122
1034
206A

031
033
*206

I
I

222~1

032

2/69

Printer (300 lines/min. )
Extension of PriDt Positiou from
120 to 132
Models'120, 120-0, and 125:
Printer (450 lines/min, )
Extension of Print Positions from
120 to 132
Printer Control for the 822-3
Extension of Printing Positions
from 120 to 132
Vertical Spacing Feature (6 or 8
Unes per inch)
Higb-Speed Printer: aDd Control,
900' LPM, 120 posItions
650- LPM Printer and Control (96
priJrtposttlons)
Extension of Print Pos1tions
from 120 to 132 for the
222-3, -4, -6.

A ..

AUERBACH

510:221.111

PRICE DATA

PRICES

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

Montbly Rental $

Name

I-Yeai'

' ' . . ___ .I_'. . . .,. ' ' ' .__

£i-Year'

, ...rchasE

Monthly
Maint.

$

$

G,6Z5

12

Printers (Contd.)

INPUTOUTPUT
(Contd. )

034
035
036
222-2
222-3
222-4
222-5
1034
1036
222-6
229(14)
031
033

l~ti

Numerk 1'\'11'11 1.'V'the 222·-1,

--2, . S
Numertc ::-'·7~~~'':.

r J~-'~

',,~," .~'7:~-4

Prult £luffer for UtI'1 2:l2-2., -,t,
-5, -6
650-LPM Pr!nter 8.ud Control {lOB
print posi'i;ionli}
650-LPM Printer and Control (120
print positions)
950-LPM Printer and Control (120
print pOf'itionsj
450-LPM PrintAr and Control (120
print pOHitiOllS)
EAtauslon of Print Positions
from 120 to 132
Eight-Channel Fo:rmat Tape
1l00-LPM Printer and Control (120
print positions)
400- LPM Printer 2nd Control (120
print POSltiOllS)
Extension of Printing Positions
from 120 to 132
Vertical Spacing Feature (6 or
d lines per inch)

1
1i

SO

120

,l

i

!

1, 1~··,5

3

S,l()O

28

38,250

211

40,500

223

57,375

316

29,250

161

2,250

12

1,285

,1,125
tlO,!:'75

3
325

405

360

15,500

150

75

70

2,925

16

ilO

25

1,125

3

1,290
415
160
345

1,150
370
145
310

56,250
17,550
6,500
12,960

260
39
15
30

1,140
900

1,015
800

45,450
39,000

80
50

MIA

2 100(15)
2: 31 0(15)

90,000
99,000

600(15)
660(15)

N/A

1,815

81,000

400

MIA

810

34,000

250

21.5
325
25

190
290
25

9,000
13,500
1,125

20
30
5

160
160
160
160
160
160

145
145
145
145
145
145

6,750
6,750
6,750
6,750
6,750
6,750

22
22
22
22
22
22

160
160

145
145

6,750
6,750

22
22

leO

146

6,750

22

~~5

:i;2f,

90S
960

200

II

1,355

30
1,440

!l05

855

I

1,210

699
60

"

615
I

50

i

I
I

I

25

I
~

I

I

Optical and MagnetiC Ink Character
Reaaers
232
233-2
234
235

MICR Reader-Sorter and Control
MICR Control (BI03)
Plotter Control (Calcomp Plotter)
Optical Journal Reader Control
(NCR 420)
Bill Feed Printer Control (IBM 14(4)
Optical Readei' Control (IBM 1287.
Modell or 2)
Burroughs BI03 MICR Reader-Sorter
Burroughs BI03 MICR ReaderSorter with Endorser
Burroughs B332 Mastel' Lister and
Control
Burroughs B333 Slave Lister (2
maximum)

237
238

"t'J/A

Console
220-1
220-3
008
COMMUNICATIONS

Console
(6
Console (Replaces Control Panel) )
Pin-Feed Drive
Single-Channel Controls

---

For the following remote terminals:
281-lA
281-lB
281-lC
281-10
281-lE
281-lH
281-lK
281-1KTP
281-1KTS

W. U, Telex

TWX CE; 8-Level Teletypewriter
5-Level Teletypewriter
8-Level Teletypewritl.:lr
TWX CE; mM 1050
Voice Lines for use with
Data Speed 2
W. U, 180-Baud; ffiM 1050
KEYTAPEI Communicator
(private lineI')
KEYTAPE/Communicator
(switched network)

___ .1--.._..
© 1969 AUERBACH CorporatIon and AUEHBACH Info, 1m;,

2/69

HONEYWELL SERIES 200

510:221.t 12

PRICES

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

Monthly Rental $

Name

l-Year

IPurchasE
5-Year
$

Monthly
Maint.
$

Single-Channel Controls (Contd.)

COMMUNICATIONS
(Contd. )

281-lM
281-1R
281-15

I

281-2A
281-2B
281-2D
281-2F
281-2M
281-2R
281-28

281-137P
281-1378

087

287-1

Data Station (288-1 Central
Control Unit)
VIP Series Displays, Asynchronous
(private lines, switched network,
or direct connection)
VIP Series Displays, Asynchronous
(direct connection only)
Voice-band Hnes (mM 7701, 1013)
Voice-band lines
Voice-band lines (mM 7702, 1013)
Telpak A
Data Station (288-3 Central
Control Unit)
VIP Series Displays, Synchronous,
(private lines, switched network,
or direct connection)
VIP Series Displays, Synchronous,
(direct connection only)
150 bits/sec (8-level Teletypewriter)
Voice-band lines (8-level
Teletypewriter)
Feature for the 281:
Long Check (available only on
certain models)
Autodin Communication Control
(USASC II Code)

160

145

6,750

22

215

190

8,930

30

230

205

9,620

32

245
195
245
245
195

215
175
215
215
175

10,125
8,100
10,125
10,125
8,100

34
27
34
34
27

260

230

10,765

36

265

235

10,995

36

60

55

2,250

7

160
160

145
145

6,750
6,750

22
22

625

(16)

27,000

73

225
340
425
850

200
305
380
760

9,450
14,400
18,000
34,000

32
48
60
120

1,225

1,095

51,750

180

60

55

2,250

7

60

55

2,250

7

45
45

45
45

1,800
1,800

6

35
35
45
35
45
45
45
45
45

30
30
45
30
45
45
45
45
45

1,350
1,350
1,800
1,350
1,800
1,800
1,800
1,800
1,800

5
5
6
5
6

45

45

1,800

6

115
45

100
45

4,500
1,800

15

45

45

1,800

6

100

90

3,930

13

Multi-Channel Controls
286-1
286-2
286-3
286-4
286-5
086
087
285-137P
285-137S

\I

285-T
285-lA
285-1B
285-1C
285-lD
285-lE
285-1H
285-1K
285-lKTP
285-1KTS
285-1L
285-lM
285-lN
285-1PD

2/69

MCCC (for 2-3 Lines)
MCCC (for 4-15 Lines)
MCCC (for 16-23 Lines)
Message-Mode MCCC (for 2-32
half-duplex lines)
Me ",sage-Mode MCCC (for 33-63
half-duplex lines)
Features for the 286-1, -2, or -3:
Parity Check on Reception &
Parity Generation on
Transmission
Long Check
Adapters for MCCC:
150 bitskec (8-1eve1 teleprinter)
Voice-band lines (8-level
teletypewriter)
Communication Interval Timer
W. U. Telex
TWX CE; 8-Level Teletypewriter
5 - Level Teletypewriter
8- Level Teletypewriter
TWX CE; IDM 1050
Voice lines for use wi th Data Speed 2
W. U. ISO-Baud/ IDM 1050
KEYTAPE/Communlcator
(private lines)
KEYTAPE/Communicator
(switched network)
Friden Collectadata 30
Data Station (288-1 central
control unit)
100 words/min. USASCII TWX
Service
Communication Adapter (Direct)

fA

AUERBACH
It

6

6
6

6
6

6

510:221.113

PRICE DATA

IDENTITY OF UNIT
CLASS

PRICES

I
Nam e
Number
Numbi"r
~-C-(-)M--M-l-IN--I_--~----------~'~. ~1Uli~_.c.~.~nnel c~ntrola (Contd.)
Model

Monthly Rental $

Feature

I-Year

5-Year

"-"

CATIONS
(Contd. )

Communieation AdapteI' (Modem)
VIP Series Display s, Asynchronous
(pnvate lines, S witched network,
or direct connec tiofl)
VIP Series Display s, Asynchronous
I
(direct connectio n only)
,
Voice Lines mrvr 7'101, 1013
. Voice Lines
Voice Unes IBM '77 02, 1013
Data Station (288-3 central control
unU)
Synchronous 2000/2 400 bUs/sec,
remote and loca1 (50 feet max. )
Synchronous 2000/:"?4000 bits/sec,
local (200 feet m ax. )
Voice Lines-DATA SPEED 5
Voice Lines-DATA SPEED 5
Switched circuits, Automatic Dialing
Multi-Adapter Unit • FuH Drawer
(8 Units for D. C · telegraph lines
Mod, 285-1A and IC, in any
combination)
Multi-Adapter Unit • Full Drawer
(8 Units for star t-stop transmiSSion; Mod. 2 85-lB, lD, lH,
lK, 1M in any c ombination)
Multi-Adapter Unit , Fun Drawer
(8 Units for Sertes 200 Communication Mod. 285 -2B)
Multi-Adapter Unit , Full Drawer
(8 Units. Automatic Dialing Mod.
285-5A)
Multi-Adapter Unit , Fun Drawer
(15 Units for D. C. telegraph lines
Mod. 285-1A, an d lC, in any
combination)

2B5-1PM
285-1R

I

285-18

I

285-2A
285-2B
285-2D
285-2M
285-2R
285-28
285-3A
285-4A
285-5A
285-7A

285-7B

285-7C

285-7D

285-7E

Monthly

n.uTchase Malnt.
$

$

80
100

75
90

3,130
3,895

11
13

115

105

4,580

15

115
70
115
70

105
65
105
65

4,500
2,700
4,500
2,700

15
9
15
9

135

120

5,270

18

140

125

5,500

18

95
95
45
225

85
85
45
200

~,600

3,600
1,800
9,450

12
12
6
32

295

265

12,375

42

440

390

18,450

62

295

265

12,375

42

405

360

17,100

57

420

375

16,875

40

95

85

3,825

8

620
35
815
45
1,290
50

550
30
725
40
1,150
45

24,750
1,350
32,625
1,800
51,750
2,050

57
3
73
3
117
3

20

15

675

2

20

15

675

2

-

-

2,050

-

-

-

2,550

-

-

-

3,200

-

Audio Response Sys tern
285-8
285-8A
285-8C
285-8D
285-8F
285-BG
285-8J
285-8K
082-1
082-2
083-1

083-2

083-3

Basic Touch Tone Adapter
Module (6 lines)
Touch Tone Adapte r Module
Expansion (2 lines)
Audio Unit (31 Elem ents; 6 lines)
2,-Line Expansion (31 Elements)
Audio Unit (63 Elem cuts; 6 lines)
2- Line Expansion {63 Elements)
Audio Unit (189 Ele menta; 6 lines)
2- Line Expansion (189 Elements)
Features for the 28 5-8
Tone Answer Back (:pHon
(2 lines)
Voice Answer Back Option
(2 Lines)
Voice CyUnde rs (31 Elements,
phrases on1y; original and
spare)
Voice Cylinde rs (31 Elements,
words and phrases or words
only; origi nal and spare)
Voice Cylinde rs (63 Elements,
phrases on1y; original and
spare)

L -_ _ _ _--L_ _ _ _- L_ _ _ _. l -_ _ _ _ _ _ _ .

(01969 AI.IFRRAf:H

rN!'')~~''a''

3r.::

!"'~!EP'9/\CH

Infv, Inc

2/69

HONEYWELL SERIEI 200

lSI 0:221.1 14

PRICES

IDENTITY OF UNIT

, CLASS

Model
Number

COMMUNICATIONS
(Contd. )

Monthly Rental •

Feature
Number

Name

I-Year

5-Year

rchase

Monthly
Maint.

$

$

Audio Res20nse System (Contd.)
083-4
083-5

Voioe Cylinders - (63 Elements
words and phrases or words
only: original and spare)
Voice Cylinders - (189 Elements words only: orlginal
and spare)

-

-

3,600

-

-

-

3,600

30

190
95
45
285
95
45
90

170
85
40
255
85
40
80

6,750
3,150
1,350
9,900
3,150
1,350
2,925

62
30
17
92
30
17
30

45
225

40
200

1,350
7,875

17
72

85

75

2,700

29

120

105

3,825

39

95

85

3,150

30

315

280

11,250

103

910

810

40,500

125

60

50

2,250

12

15

15

675

2

81

74

2,600

20

92

84

3,020

21

60

55

1,840

16

65

60

1,980

18

60

55

1,980

10

78

71

2,500

14

278

255

10,820

60

92

84

3,879

10

121

109

5,255

12

Data Station
288-1
088-1
088-2
288-3
088-3
088-4
289-2
289-2A
289-3
289-4
289-5
289-7
289-8
289-9
1034
089-2

Central Control Unit
Buffer
Extended Operation
Central Control Unit
Buffer
Extended Operation
Page Printer (10 char/sec)
and Keyboard
Central Control Unit Keyboard
Page Printer (40 char/sec)
and Keyboard
Paper Tape Reader (120 char /
sec) and Spooler
Paper Tape Punch (120 chari
sec) and Spooler
Card Reader (143 char/sec)
(Feature 088-1 required)
~tical Bar Code Reader
(Feature 088-1 required)
Remote Line Printer (120
print positions; required
288-3)
Extension of Print Positions
from 120 to 132
Multi-Line Block (Requires
Feature 088-3)

.

Data Entry and Display Equipment
303
304
311
312
317
318
323
331
332

Display Station (alpha-numeric
teletype keyboard; 7- by 5-inch
display area)
Displav Station (alpha-numeric
Navcor keyboard; 7- by 5-fnch
display area)
Display Station (numeric keyboard
5- by 4-inch display area)
Display Station (numeric-block
alpha keyboard; 5- by 4-inch
display area)
Display Station (alpha-numeric
without keyboard, 7- by 5-inch
display area)
Display Station Monitor
(without keyboard, 18- by 13inch display area)
Universal Control Unit (one to
eight expansion modules)
Communication Interface
(asynchronous: 1200 bits/sec)
Communications Interface
(synchronous, 2000/2400 bits/
sec)

/

2/69

A..
AUERBACH

510:221.115

PRICE DATA

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

Name

PRICES
Monthly Rental $
I-Year

5-Year

138

Monthly
PurchasE Matnt.

$

$

126

5,590

36

12

11

410

75

70

2,810

10

118

109

4,475

15

28

25

940

5

7

6

260

--

Data Entry and Display Equipment
(Contd. )

COMMUNICATIONS
(Contd. )
335
339
341
332
343
351
352
353
353A
355
360
361
362
370
371
372
373
374
375
376
377
386
386-1
2306
2317A
2317B
2317C
2317D
2317E
2322A
2322B
2322C
2322D
2322E
2331
2339
2353A
2353B
2353C
2353D
2353E

High Speed Interface
(42,000 char/sec)
Clock Generator (modem
by-pass)
Expansion Module (numbers
I, 2, 4, 5, 7, and 8)
Expansion Module (numbers
3 and 6)
Expansion Module (for nondisplay I/o devices only)
Logic Module for Message
Editing
Logic Module for Multi-Message
Transactions
Receive Only Printer Control
(numbers 1, 4 and 7)
Receive Only Printer Control
(numbers 2, 3, 5, 6, 8 and 9)
Polling Option for Type 332
Communications Interface
RO Printer (friction feed:
Teletype Model 33)
RO Printer (friction feed,
Teletype Model 35 AT)
RO Printer Sprocket Feed
(Teletype Model 35 AU)
Teller Terminal
Transceiver
Transceiver (off-line
adaption)
Polling Option
Crossfoot Accumulator
Junction Control
Transaction Accumulator (2
accumulators)
Transaction Accumulator (2
accumulators)
High Speed Control Unit
Channel Adapter
2300 Series:
Alphanumeric Keyboard
960-Character Display Station
888-Character Display Station
480- Character Display Station
444-Character Display Station
222-Character Display Station
Multi-Station Control:
960-Character
888-Char
480-Char
444-Char
222-Char
Communications Interface
Modem Bypass Clock
RO Printer Control:
960-Character
880 Char
480 Char
444 Char
222 Char

---

Available at No Extra Cost
60

55

2,030

8

28

25

990

3

17

16

555

2

35

31

550

21

75

67

1,600

36

81

72

1,820

38

234
229
245

215
210
225

8,363
8,570
9,225

40
52
52

16
32
47
37

16
31
46
36

280
995
1,975
1,420

6
8
8
10

37

36

1,420

10

460
140

410
125

20,025
5,500

30
18

31
83
83
73
73
52

28
74
74
65
65
47

1,260
3,360
3,360
2,940
2,940
2,100

15
16
16
14
14
10

403
403
367
367
295
47
16

360
360
328
328
263
42
14

16,380
16,380
14,910
14,910
11,970
1,890
630

78
78
71
71
57
9
3

109
109
78
78
67

97
97
70
70
60

4,410
4,410
3,150
3,150
2,730

21
21
15
15
13

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

4/69

HONEYWELL SERIES 200

510:221,116

PRICES

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

Displa~

Data Entry and
(Contd. )

COMMUNICATIONS
(Contd. )
2358A
2358B
2358C
2358D
2358E
2323
2354
2343
2390
2391
2392

Monthly Rental $

Name

1-Year

5-Year

114
114
83
83
73
258

Monthly
Purchase Maint.

$

$

102
102
74
74
65
230

4,620
4,620
3,360
3,360
2,940
10,500

22
22
16
16
14
65

11

10

420

2

16

14

630

4

52

47

1,100

25

78

70

2,100

37

83

74

2,250

40

Eguipment

Local RO Printer Control:
960- Character
880-Char
480-Char
444-Char
222-Char
960-Character Single
Display Station & Control
Printer Control for Single
Display Station
Print Buffer for Single
Display Station
Model 33 RO Printer
(friction feed)
Model 35 RO Printer
(friction feed)
Model 35 RO Printer
(sprocket feed)

NOTES:

*

No longer in production.

(1)

Central Processor Units include memory, control panel, and power supply.

(2)

Field upgrade of memory from processor models 126-1 through 126-8 to models 126-2-9 through 126-2-12
requires exchange of processors.

(3)

Not available on new orders.

(5) Available only with 201 and 201-1 processors.
(6)

For Model 110 only.

(7)

Prices for larger drum configurations will be quoted on request.

(8)

Can be changed to read/write 200/800 bpi.

(9)

Two hundred bits/inch density available only with feature 059 or 1059.

(10) Models 120, 120-0, and 125 only.

(Ii) Density of 556 bpi available only with feature 059.
(12) One time charge of $500 if factory or field installed.
(13) Minimum installation and rental charge of $100 is made for six months or less usage, more than six
months is at $125/month rate.
(14) Restricted to educational institutions.
(15) Plus maintenance surcharge of $150 if located outside a Honeywell Service area.
(16) Quoted on request.

4/69

A

AUERBACH

'"

(Contd. )

531 :011.100
Monrobot XI
Introduction

INTRODUCTION
§

OIl.

The Monrobot XI is a compact, solid-state data processing system that is suitable
for a variety of fairly complex but low-volume business and scientific applications. It is
also being used in small instrumentation and process control systems. The basic system,
consisting of computer, input-output typewriter, and paper tape reader and punch, can be
purchased for $24,500 or leased for $700 per month. This makes it one of the lowest
priced internally programmed data processing systems currently available.
The central processor is housed in a desk-size cabinet and weighs only 375 pounds.
Most of the peripheral devices are housed in matching cabinet modules of desk height that
can be arranged in a number of ways for maximum operating efficiency. There are no
special power or air conditioning requirements.
A magnetic drum provides 1,024 word locations of working storage; a 2,048-word
drum is a recently-announced option. Each 32-bit location can hold two single-address
instructions, one binary data word, or from four to six alphameric characters. Seven of
the addressable storage locations are Fast Access Registers with a constant access time
of 0.73 milliseconds. Average access time for all other storage locations is 5.85
milliseconds.
The small but convenient instruction repertoire includes addition, subtraction, .and
multiplication of single word-length, fixed point binary data. Division can only be accomplished by subroutines. Binary and decimal shifts and a repetitive subtraction ("detract")
instruction facilitate the programmed radix conversions that usually must be performed
upon input and output data. Neither index registers nor indirect addressing are provided,
so a large proportion of the instructions in many programs will be devoted to "housekeeping'
operations. Program execution speed will usually average 60 to 80 instructions per second.
Somewhat higher speeds can be achieved if operand addresses are optimized where possible,
but the increase in speed will seldom justify the extra coding time.
Up to three separate input devices and three output devices can be connected to the
Monrobot XI and selected under program control. Each input or output instruction initiates
the transfer of a single character between the processor and the addressed peripheral device. Overlapping of input-output operations and internal processing is possible.
Paper tape or verge-punched cards with 5- or 8-level codes can be punched and read
mechanically at a peak speed of 20 characters per second. A photoelectric reader provides
maximum input speeds of 40 to 50 characters per second. IBM 024 or 026 Card Punches can
be connected through special couplers and used for on-line punched card input, output, or
both. Standard 80-column cards are read and punched at 16 columns per second.
Printed output can be produced at up to 10 characters per second by either a modified
IBM electric typE.'-Vriter or a Teletype printer; either unit can also be used for manual entry
of data. A 16-key keyboard is useful for rapid entry of all-numeric data.
The Monroe-Card Processor reads and records information on magnetizable cards.
Up to 1,566 decimal digits or 1,044 alphameric characters can be stored on each card.
Monroe-Cards will be useful for master file storage in a variety of data processing
applications.
The Monrobot XI software Situation, when viewed by potential users with a strong
desire to minimize programming time and effort, leaves much to be desired. Routines
currently available from the manufacturer are limited to general utility routines, a userdeveloped symbolic assembly system, and a group of scientific routines (floating point

©

1962 by Auerbach Corporation and BNA Incorporated

10/62

MONROBOT XI

531:011.101

INTRODUCTION-Contd.
§

OIl.

arithmetic, functions, matrix inversion, etc.). No compiler systems, interpretive
systems, or report generators are available or under development.
Most coding is done in machine language; the coder writes four hexadecimal digits
per instruction, or eight per word. The hexadecimal addressing scheme is easy to master,
but the operation codes have no mnemonic relationship to their effects. Generalized subroutines are available to handle division, loop control, address modification, and inputoutput with radix conversions, but the manufacturer encourages the use of individuaUytailored, user-coded routines for greater efficiency.
A Monrobot XI users' group is now being formed, under Monroe's auspices. to
encourage and control the publication, standardization, and distribution of routines developed
by users and by the manufacterer.

•

10/62

531:221.101
Monrobot XI
Price Data

PRICE DATA
§

221.
IDENTITY OF UNIT
CLASS
No.

Central
Processor

Name

Monrobot XI - Basic System
Includes the following units:
Computer and control unit
Input-Output Typewriter
Paper Tape Reader
Paper Tape Punch
Knee-hole Desk
2 Input-Output Buffers
Optional Features
Oscilloscope View Box
Input-Output Buffer
Cabinet (2 legs)
Table (2 legs)
2, 04H-word Drum
Note: For punched card inputoutput, 24 and 26 Couplers
can be substituted for Paper
Tape Reader and Punch on a
one-for-one basis; 024 or
026 Card Punches must be
ordered from IBM.

InputOutput

Paper Tape Reader (includes
Cabinet)
Paper Tape Punch (includes
Cabinet)
Paper Tape Reader and Punch,
in single Cabinet
Edge-PunchedCdrd Reader
Edge - Punched Card Punch
Optional Features
5-S Channel Switch
Paper Tape Unwind Reel
24
26

Coupler (for punched card input)
Coupler (for punched card output)
Input-Output Typewriter
Output Typewriter
Optional Features
20-inch Carriage
20-inch Pinfeed Platen
16-inch Pinfeed Platen
Form Aligner (tracter feed)
Form Stand (paper tray)
Special keys (each)

PRICES
Monthly
Rental

Annual
Maintenance

Purchase

~

$

$

700.00

1,200.00

24,500.00

NA
20.00
12.50
NA
IS5.00

5.25
30.00
0
0
?

105.00
600.00
400.00
60.00
5,250.00

60.00

S2.50

1,650.00

33.00

55.00

1,100.00

S1.OO
70.00
43.00

125.00
97.50
70.00

2.500.00
1,950.00
1,400.00

5.00
NA

4.50
0

90.00
20.00

25.00
25.00

40.00
40.00

SOO.OO
SOO.OO

120.00
SO.OO

165.00
123.75

3,300.00
2,475.00

0
0
0
0
0
0

100.00
100.00
100.00
100.00
65.00
75.00

NA
NA
NA
NA
NA
NA

Teletype Printer:
In lieu of basic Typewriter
As additional output unit

50.00
120.00

50.00
165.00

1,000.00
3,300.00

Monroe-Card Processor:
96 words/card
174 words/card

230.00
290.00

?
?

6,500.00
S,500.00

©

1962 by Auerbach Corporation and BNA Incorporated

10/62

MONROBOT XI

531:221.102

PRICE DAT A-Contd.

§

221.
PRICES

IDENTITY OF UNIT
CLASS
No.

Monthly
Rental

Name

$

Annual
Maintenance

Purchase·

$

$

fjPututput

16-Key Keyboard

12.50

20.00

400.00

Data
Orig!nation

Synchro-Monroe Punch Tape Adding
Machine:
One-register model
Two-register model

88.00
98.00

110.00
120.00

1,950.00
2,175.00

Notes: NA in rental column means unit or feature is available for purchase only.
Maintenance charges apply only to purchased equipment.
Prices do not include Manufacturers' Excise Tax of 6 percent on purchase or
10 percent on rental.

10/62

I AUERBACH I $:I

(

"

NATIONAL
CASH REGISTER CO.

", /

AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

AUERBACH

(!)

Print.-d in IL~.A.

601:011.100

£ ""''''
IA'-

AUERBAC~

-

NCR 315

EDP
REPDRTS

SUMMARY

~

SUMMARY
.1

SUMMARY
The NCR 315 is a small to medium scale, solid-state computer system oriented toward
business data processing applications. A library of floating point subroutines equips the
315 to handle modest scientific computational loads as well. System rentals range from
$2,850 to over $25,000 per month, with most installations falling within the $5, 000 to
$15,000 range. First customer deliveries of the NCR 315 were made in the Fall of 1961,
and more than 500 systems have been installed to date .

.2

COMP ATIBILITY
The NCR computer line was expanded in July 1963 I)y the announcement of the 315-100
series (Report 602:), in July 1964 by the announcement of the 315 RMC (Report 603:), and
again in September 1966 by the announcement of a new Multiprogramming Processor for
the RMC system (Report 603:).
The 315-100 is essentially an economy version of the 315 which uses the same processor,
except that many of the features (such as multiply/divide and the capability to connect
magnetic tape units) are optional. A line of low-performance, low-cost peripherals
originally announced solely for the 315-100 is now available for the 315. This greatly
reduces the effective differences between the two systems.
The NCR 315 RMC (Rod Memory Computer), on the other hand, uses a completely new
central processor and internal storage. The 315 RMC uses the first commerciallyavailable complete thin-film memory and performs internal operations from 3 to 10 times
faster than the original NCR 315. The instruction repertoire includes all the instructions
of the 315 and features several extensions - primarily floating-point arithmetic hardware.
The new RMC Multiprogramming Processor provides additional hardware, including
separate operating modes and registers for the Executive program and the users' programs, for facilitating the concurrent execution of several programs in a multiprogramming fashion. Except for minor differences (and the added commands of the 315 RMC), all
of the processors in the 315 series are program-compatible and utilize the same software.
There is no program compatibility between the 315 line and NCR's other second-gtmeration
computers - the NCR 304 (a medium-scale system which is no longer in production), the
NCR 310 (an adaptation of the Control Data 160 oriented toward MICR sorter-reader operations), and the NCR 390 and 500 Series (small-scale data pz:ocessing systems built
around the concept of magnetic ledger cards). The third-generation NCR Century Series,
however, offers a 315 Emulation Unit as an optional feature for the Century 200 Processor.
The NCR 315 is tape-compatible with the IBM 729 series and other "ffiM-compatible"
magnetic tape handlers. Because code translation for most NCR 315 peripheral devices
is performed by the stored program, a wide variety of data codes can be accommodated
on punched tape and cards .

.3

HARDWARE

. 31

Central Processors and Main Memory
The basic addressable unit of internal storage in NCR 315 systems is the "slab, " which
consists of 12 data bits and 1 parity bit. Each slab can hold two 6-bit alphameric characters or three 4-bit decimal digits. Instructions are provided to convert information
from the alphameric to the decimal mode, and vice versa. All arithmetic operations are
performed upon data stored in the 4-bit decimal mode. Arithmetic operands can be from
1 to 8 slabs (or 3 to 24 digits) in length, as specified in the instruction. A minus sign
requires one digit position, whereas a plus sign does not. The results of most arithmetic
operations are developed in a variable-length accumulator.
Instructions occupy either two or four slabs each; most are of the one-address type, but
others function as two-address instructions. The repertoire of approximately 150 instructions plus variations includes fixed point multiplication and division, add-to-storage,
binary addition, three-way comparison, shifting, and block transfer facilities. Literal
operands up to three digits in length can be specified in many instructions. Edit, Suppress,
and Scan instructions facilitate format control and character manipulation. Internal instructions are executed at the rate of about 16,000 per second in typical NCR 315 routines.

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

6/68

601:011. 310

· 31

NCR 315

Central Processors and Main Memory (Contd. )
Interrupt facilities aid in achieving efficient utilization of the NCR 315's input-output capabilities by informing the central processor when a peripheral device is ready to deliver or
receive information. When the master Demand Permit Flag is on, any peripheral unit
whose individual Unit Demand Flag is also on will generate an interrupt signal whenever it
is ready to accept another input or output instruction. When the central processor receives
the interrupt signal, it completes execution of the current instruction and then jumps to a
special routine. This routine tests all active peripheral units to determine which one
caused the interrupt, and then initiates the appropriate action.
Core storage is available in module sizes of 5,000, 10,000, or 15,000 slabs. Up to four
of the 10, OOO-siab modules can be used in a system, providing a maximum core storage
capacity of 40,000 slabs, 80,000 characters, or 120,000 decimal digits. Cycle time is 6
microseconds for each access to one 12-bit slab. A parity check is performed upon all
internal data transfers.
An auxiliary core storage bank, which functions independently of the main core store,
holds 32 index registers, 32 jump registers, the accumulator, and a number of programtestable "flags" which indicate the result of a comparison, an arithmetic overflow, or an
interrupt condition. Because only 1,000 storage locations can be directly addressed by
the 3-digit instruction address, nearly every NCR 315 instruction utilizes index register
modification. (Indexing requires no additional execution time.) The 32 jump registers
are used primarily to store "jump tables, " which transfer control to specified locations
when specific conditions (errors, end-of-tape marks, etc.) arise in the execution of certain instructions.

· 32

Auxiliary Storage
CRAM (Card Random Access Memory) is a key feature of the NCR 315 system that combines many of the advantages of magnetic tape and disc storage units. The CRAM storage
medium is a deck of flexible magnetic cards. A cartridge containing up to 384 cards can
be quickly removed from the CRAM Unit, replaced by another cartridge, stored off-line,
and reinserted when necessary, in the same manner as a reel of magnetic tape. Four
models of CRAM are available, differing in storage capacity and recording density. From
5.5 million (Model 353-1) to 82.9 million (Model 353-5) characters can be stored in a
single CRAM cartridge.
One selected CRAM card at a time is dropped from the on-line cartridge and wrapped
around a revolving drum; this takes from 120 to 235 milliseconds, depending on the model.
Then any or all of the data bands (7, 56, or 144, depending on the model) can be read
and/or written sequentially. The 353-5 uses a movable head mechanism to record four
groups of 36 bands (144 bands total) on each card. Data is transferred at a peak rate of
100,000 characters per second (Model 353-1), 50,000 characters per second (Model 353-5).
or 38,000 characters per second (Models 353-2 and 353-3).
Up to 16 CRAM units can be connected to an NCR 315 system, and different models can
be intermixed if desired. Card dropping time can be overlapped, but only one CRAM read
or write operation can be performed at a time. Both lateral and longitudinal read-afterwrite parity checks are performed when writing a CRAM reco])d.
The only non-CRAM random access storage device available for 315 systems is the C365
Disc Unit, which offers fast access to relatively small volumes of data. The C365 features
a head-per-track design with four fixed discs per unit. Average access time is 16.7
milliseconds, peak data transfer rate is 120,000 characters per second, and maximum
data capacity is 2, 000, 896 characters per unit.

· 33

Peripheral Equipment
NCR offers a large array of peripheral equipment for 315 systems, but complex configuration rules limit the selection of components for a particular installation. Several of the
peripheral devices previously offered by NCR have been discontinued from production.
Some of these units are still available; others are not available at all. However, NCR
states that it will continue to support all devices in existing installations.
Magnetic tape units are available with peak transfer rates ranging from 12,000 to 120,000
characters per second (tape speeds of 60 to 150 inches per second). NCR has discontinued
the 333-bits-per-inch recording density and now uses the IBM 729-compatible densities of
200, 556, and 800 bits per inch. In all models, block length is variable, and a read-afterwrite parity check is performed upon recording.
Up to eight magnetic tape handlers can be connected directly to an NCR 315 central processor, in which case no overlapping of magnetic tape reading or writing with computation is
possible. Alternatively, Magnetic Tape Simultaneity Controllers can be used to provide
either read-compute and write-compute overlapping (with one controller) or full read-writecompute simultaneity (with two controllers), through time-sharing of accesses to core
memory. Up to eight tape handlers can be connected to each controller. A mix of up to 16
tape handlers can be connected directly and via a controller. Tape handlers of different
tape speeds can be intermixed in a 315 system.

A.

6/68

AUERBACH

(Contd.)

SUMMARY

.33

601:011. 330

Peripheral Equipment (Contd.)
Two card readers (400 or 2,000 cards per minute), two card punches (both with completely
buffered operation at 100 or 250 cards per minute), and two card read-punch units (which
read at 300 or 400 cards per minute) provide punched card input-output. The 100-cpm
punch and the card read-punches are the IBM 523, 1442 Modell, and 1442 Model 2 units,
respectively. The 250-cpm card punch is an adaptation of a Control Data unit. All these
devices handle standard SO-column cards, and the slower card reader is also available in
a 90-column version. A maximum of one card reader or two card read-punch units can be
connected on-line to an NCR 315 system. A total of four card punches and printers, in any
combination, can be connected.
Two paper tape readers (600 and 1,000 characters per second) and two paper tape punches
(120 and 110 characters per second) provide paper tape input-output. All models are unbuffered. Only one paper tape reader and one punch can be connected on-line at a time.
A number of printers are available for the NCR 315. including several recently-announced
units. Buffered and unbuffered units are available with peak printing rates of up to 1,000
lines per minute for numeric or alphanumeric data, and with 120 or 132 print positions.
One recently-announced printer can optionally be equipped to function as a lister for numeric
data at up to 2, 000 lines per minute. A total of four printers Rnd card punches, in any
combination, can be connected to an NCR 315.
Documents encoded in magnetic ink can be read and sorted at the rate of 750 or 1,200 documents per minute. An optical character reader can read journal tapes produced by cash
registers, adding machines, and accounting machines at up to 1,664 characters per second.
Up to four buffered MICR and optical readers, in any combination, can be connected to an
NCR 315.

.4

The Model 321 Central Communication Controller enables an NCR 315 computer system to
handle up to 99 narrow-, voice-, or broad-band lines in any combination. The 321 Controller supersedes the various models of the 356 Inquiry Buffer previously used with
NCR 315 computers. The Controller accesses memory directly, sharing memory cycles
with the processor. Various adaptors are provided to accomodate low-, medium-, and
high-speed terminals using start/stop synchronization and to accommodate NCR's banking
terminals. NCR provides a set of specialized communications subroutines which can be
tailored to fit individual applications. These subroutines provide control of the message
flow in a data communications system and are assembled into suitable Supervisory programs
by individual installations.
The 795 Data Display System, a modified version of the Sanders 720 system, adds important remote data entry and retrieval capabilities to the NCR 315 line. The CRT display unit provides a 7. 5-by-9. 5-inch image area and can display 256, 512, or 1024 characters. A keyboard permits data to be entered into the computer system from the display
terminal locations .
SOFTWARE
The NEAT Compiler is an advanced symbolic assembly system designed for use in NCR
315 systems with at least 10,000 slabs of core storage, a punched tape or card reader, a
printer, and either 1 CRAM unit or 4 magnetic tape handlers. (A special NEAT Compiler
is available for systems with 5, 000 slabs of core storage and 4 magnetic tape units.) References to an extensive library of macro instructions cause the insertion of in-line and/or
closed subroutines in the object program. User-defined macros can be added to the library.
The data to be processed by the object program is defined in terms of its hierarchical
structure of files, records, groups, and fields, using COBOL-like level indicators. standard forms are provided for tape or CRAM file specifications and compiler control. All
object programs produced by the NEAT Compiler are compatible with the STEP and PACE
operating systems described below.
The NEAT Assembler is a basic symbolic assembly system designed for small NCR 315
installations. It requires only 5,000 slabs of core storage, punched tape or card inputoutput, a printer, and 1 magnetic tape or CRAM unit. The coding format is fixed, and
none of the macro instructions or data definition facilities of the NEAT Compiler are available. All the facilities of the target c.omputer can be utilized.
BEST (BuSiness EDP Systems Technique) is a technique developed by NCR to speed the
programming and debugging of programs to perform routine business data processing functions. A job is defined in terms of BEST functions (41 ~e currently provided), and a series
of parameter sheets is fHled out. Cards, key-punched from the parameter sheets, are
input to the BEST program generator, where the calls for BEST functions are replaced with
subroutines coded in symbolic language (NEAT). The NEAT compiler is then used to produce
a machine-language program. Facilities provided by the currently-offered set of BEST functions include such operations as input-output, file control, arithmetic, paper tape code translation, report writing, and sorting. The minimum configurations required to utilize the BEST
program generator is 10,000 slabs of memory and either five magnetic tape units or two
CRAM units (any model).
C 1968 AUERBACH Corporation and AUERBACH Info, Inc.

6/68

601:011. 400

.4

NCR 315

SOFTWARE (Conte!.)
The NEAT COBOL Compiler accepts nearly all of Required COBOL-61 (there are minor
exceptions) and most of the COBOL-61 Electives. The compiler requires at least 10,000
slabs of core storage and either 2 CRAM units or 5 magnetic tape units. COBOL source
statements are translated into NCR 315 machine language object programs at an average
rate of 10 to 20 statements per minute. A useful, non-standard addition to the COBOL language is the LOCATE'verb, which enables the COBOL programmer to utilize CRAM units for
file storage. Object program efficiency is strongly influenced by the data arrangements in
core storage and in the files. These data arrangements are prescribed by the COBOL programmer, and guidelines a1'e available which help him to maximize efficiency by arranging
the data in accordance with the NCR 315's internal structure. NCR states that a version of
COBOL 65 will be available in November 1968.
STEP (Standard Tape Executive System) is an input-output control and supervisory
routine for NCR 315 magnetic tape systems; PACE (Packaged CRAM Executive) is
its counterpart for systems that utilize CRAM memory. Both systems are capable
of controlling run-to-run changeovers, program loading, restarts, and overlays, as
well as all routine tape and CRAM input-output operations. The Librarian routine
creates and maintains a program library tape or CRAM deck in which each program includes all the information required by STEP or PACE.
The Tape and CRAM Sort Generators utilize parameters specified in control cards to
generate sorting routines that use from 4 to 8 tape units or 1 to 4 CRAM units, l,"espectively. Either fixed- or variable-length records can be sorted according to either
fixed- or variable-length keys. The user can insert his own coding to add, delete, or
edit selected records during the first and/or last pass. Restart points are established
at the end of each merge pass.
Other available software for the NCR 315 includes a well-planned library of Scientific and
Engineering Subroutines; a FORTRAN II compiler; a FORTRAN IV compiler; FAST (a
load-and-go algebraic compiler); a variety of diagnostic and printout routines; and a set of
"canned" programs for specific applications such as demand deposit accounting, on-line
savings, accounts payable, inventory management, and PERT.

6/68

A

AUERBACH
~

A

601 :221.101

SU.I.II

EDP
NCR 315
PRICE DATA

tI"lts

AUlnAe"

NCR 315
IDENTITY OF UNIT
CLASS

Model
Number

PROCESSOR

Feature
Number

Name

PRICES
Monthly
Monthly
Rental Purchase Maint.

$

$

$

1,400
1,500
1,6flO
75
100

90,000
95,000
100,000
3,250
5,000

180
190
200
10
5

1,100
1,600
1,800
2,800

55,000
75,000
80,000
132,000

20
20
20
40

Processing Unit
315-3
315-4
315-35

File Processor
Bank File Processor

File Inquiry Processor
Unbuffered Inquiry Adapter
Automatic Recovery Option
Main Memory

316-2
316-301
316-302
316-4

PERIPHERAL
DEVICES

Memory
Memory
Memory
Memory

(10,000 characters)
(20,000 characters)
(additional 20,000 characters)
(30, 000 characters)

All Units: See NCR 315-100 (1)

NOTES:
(1) The 315 can accommodate all 315-100 peripheral devices except the unbuffered 402-4 MICR Sorter Reader.
The Processor model determines which of the peripheral devices can be connected. No additional adapters
are required.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

A.

602:011. 101
m .....

~EDP

_~

-

NCR 31~IOO
INTRODUCTION

IIPlIIS

INTRODUCTION
The NCR 315-100 was announced in July, 1963, as an economy version of the NCR
315 computer system using essentially the same central processor and core memory. The
multiply-divide facility and various input-output control features of the 315 were made optional,
and a low-cost, low-performance line of peripheral devices was made available for the 315-100.
The purpose of this, of course, was to reduce the cost of obtaining an installation's first computer system from NCR. Recently, the full line of NCR 315 peripheral equipment was made
available for 315-100 systems. This has greatly reduced the effective differences between the
315 and 315-100.
The performance of the 315-100 central processor and core storage is essentially
identical with the performance of the corresponding components of the original NCR 315 system,
Thus, programs can be freely interchanged between a 315 system and a 315-100 system having
equivalent facilities, peripheral equipment, and core memory. All of the software available for
the 315 can be used with 315-100 systems.
To emphasize the close relationship between the NCR 315 and 315-100 computer
systems, this report presents only the information that pertains specifically to the 315-100.
See the NCR 315 report (page 601:011.100) for descriptions of the facilities and characteristics
of the various hardware and software components of the 315 line.

© 1967 AUERBACH Corporation and AUERBACH Info. Inc.

3/67

602:221.101

A

nU'11i

EDP

AUIR ....CH

1t ••

NCR 315·100
PRICE DATA

I"

NCR 315·100
PRICES

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

Name

Monthly
Monthly
Rental Purchase Maint.

$

$

$

Processing Unit and Main Storage

l' H OC F. SSO Tl

:115-101

:l1H-:102

Central Processor, including Memory Unit:
316-102 (10,000 characters)
316-103 (20,000 characters)
:116-104 (30,000 characters)
Additional 20,OOO-Character Memory
Automatic Recovery Option
Low-Speed File Adapter(l)
High-Speed File Ada~ter(2)
CRAM-File Adapter )
Simultaneity Adapter(4)
Multiply-Divide
MICR Buffer Adapter(5)
Inquiry Buffer Adapter(6)
Unbuffered Inquiry Adapter(7)

1,600
2,200
3,000
l,AOO
100
100
400
100
100
200
100
250
75

A2,600
104,400
144,100
AO,OOO
5,000
5,000
20,000
5,000
5,000
9,200
5,000
10,000
3,250

190
190
265
20
5
15
25
10
10
20
10
10
10

1,340
2,140
2,940
3,740

57,500
92,000
12H,500
IHl,OOO

140
240
340
440

950
700
A25
1,350
:10

19,900
30,600
:15,500
H:l,OOO
1,500

158
120
150
120
15

300

12,400

70

225
400

9,100
Hi,OOO

fiO
80

300
975

12,500
40,000

70
140

fi95

32,000

58

450

15,000

50

450

20,000

83

900

35,000

133

Disc Storage

'I1.\SS
STOH.\CE

365-101
:165-102
3H5-103
:165-10.f

Disc
Disc
Disc
Disc

Controller
Controller
Controller
Controller

with
with
with
with

one Disc Unit
two Disc Units
three Disc Units
four Disc Units

Magnetic Card storage
*353-1
35:~-2
:~5:3-:I

:15:1-5

100KC CRAM (5.5 Million Characters)
:J8KC CRAM (8.0 Million Characters)
3AKC CRAM (IH. 1 Million Characters)
38KC CRAM (82 Million Characters)
Automatic Recovery Option
Magnetic Tape

r;.;Pl"TOl"TPl"T
33.f-l01
:3:1.f-102
33.f-1:11
:I:l.f-132
333-101
:3:1:1-102
:l2.f-l

Magnetic Tape Unit (with controller; 12KC;
controls up to four 3:14-102' s)
Magnetic Tape Unit (without controller; 12KC)
Magnetic Tape Unit (with controller; 33KC:
controls itself and up to four 3:14-132' s)
Magnetic Tape Unit (without controller; 33KC)
Magnetic Tape Unit (120KC)
Magnetic Tape Unit (8:1.4KC)
Magnetic Tape Simultaneity Controller
Punched Card and Punched Tape

*472-1
*472-2
*-172-3

Input/Output Console (includes 1,000 chari
sec paper tape reader and 110 char/sec
paper tape punch)
Input/Output Console (includes 400 card/min
card reader)
Input/Output Console (includes paper tape
reader, paper tape punch, and card
reader)
90-column adapter for console reader

© 1969 AUERBACH Corporation and AUERBACH Info, Inc"

-

200

-

2/69

NCR 315-100

602:221.102

PRICES

IDENTITY OF UNIT

CLASS

Model
Number

Feature
Number

Monthly
Rental

Monthly

Purchase Maint.

$

$

$

250
250
375

9,800
9,800
23,900

50
50
21.50

500

25,200

27. 50

150
750
35
125

6,600
35,000
1,400

10
125
5
16.25

400

22,500

90

450

25,000

30

650
1,350

-

90
125

750
1,450

30,000
55,000
350
35,000
60,000

150

8,000

24

1,700

45,000

458

1,700

45,000

458

450

23,500

30

2,100

95,000

417

250

15,000

30

1,450
1,950

60,000
80,000

184
265

190

6,650

18

80
80

7

40

2,800
2,800
2,800
1,400

40

1.400

2

100
20
22
150
175

3,500
700
770
5,250
6,125
245

14
2

Punched Card and Punched Tape (Contd. )

INPlIT-

OllTPUT
«('ontd. )

Name

Paper Tape Reader (600 char/sec)
Paper Tape Punch (120 char/sec)
Card Read Punch (300/50 cards/min;
requires 354-601 controHer)
Card Read Punch (400/88 cards/min:
requires 354-601 controller)
Card Read Punch Controller
Card Reader (2000 cards/min)
IBM Translator feature
Card Punch (100 cards/min;
reqires 354-101 buffer)
Card Punch (250 cards/min;
requires 354-101 buffer)
Card Punch Buffer

361-201
371-201
376-7
376-8
354-601
380-:1
*376-2
376-101
354-101

-

Printers
Printer - Unbuffered (805 lines/min)
Printer (1000 lines/min; includes buffer)
Special characters, per character
Printer - Unbuffered; (620 lines/min)
Printer - Buffered (100 lines/min)
Printer - Buffered with Lister Adapter
(100 lines/min)
Lister Attachment

340-503
340-601
340-532
340-632
340-644
644-1

-

110
150

MICR Sorter Readers

402-3

MICR Sorter Reader (750 documents/min;
requires 355-1 Buffer)
MICR Sorter Reader (750 documents/min;
unbuffered)
MICR Sorter Buffer (controls up to four
402-3 Sorter-Readers)
MICR Sorter-Reader (1200 documents/min;
requires 355-3 Buffer)
MICR Sorter Buffer (controls up to 4
407-101 Sorter-Readers)

402-4
355-1
407-101
355-3

Optical Readers

*420-1
420-2

Optical Journal Reader
Optical Journal Reader
Display Unit

795-100
795-151
795-152
795-153
795-201
795-202
795-300
795-400
795-500
795-510
795-521
795-401

2/69

Display Controller (includes 795-101 Edit
Module and 795-102 I/O Module)
Memory Module (lx1024 characters)
Memory Module (2x512 characters)
Memory Module (4x256 characters)
Display Controller Cabinet (for one
controller)
Display Controller Cabinet (for up to 3
controllers)
Display Screen
Keyboard
Split Screen
315 Interface
Multi I/O Channel (on~ per controller)
Keyboard Adapter (each keyboard)

A ..

AUERBACH

80

7

7
7

2

3

3
12
2

602:221 103

PRICE DATA

-PRICES

IDENTITY OF UNIT
CLASS

INPllTOl1TPllT
(eontd. )

Model
Number

Feature
Number

Name

Mnnthly
Monthly
Rental IPurchas~ Mamt.

$

$

$

other
435-201
435-202
435-203

Universal Interconnecting Device (1 module)
Universal Interconnecting Device (2 modules)
Universal Interconnecting Device (3 modules)

125
190
250

5,000
7,500
10,000

!l
l:l

321-1

Central Communication Controller (includes
space for 3 Adapter Cages)
Auxiliary Cabinet (contains space for 4
Adapter Cages)
Adapter Cage (contains space for 10 Adapters)
Asynchronous Character Adapter
Bit Adapter
Bank Adapter
Central Inquiry Buffer (1 character;
alphanumeric; controls up to 8 adapters)
Central Inquiry Buffer (17 digits;
numeric; controls up to 8 adapters)
Communication Line Adapter -Teletype
Communication Line Adapter -Monitor
Auxiliary Cabinet
Window Machine Controller
Branch Controller for:
1 to 8 window machines
9 to 16 window machines
Single Window Controller

850

36,000

100

250

10,000

:30

30
60
15
60
675

1,200
2,500
600
2,500
28,200

fi

3
10
60

975

40,500

fi

175
130
160
25

7,100
5,500
6,450
1,000

19
20
10

220
260
180

9,500
11,500
7,300

40
40
32

IR
--

COMMUNICATIONS

359-302
359-303
359.-304
356-1
356-3
*359-3
*359-4
358-3
428-3
438-318
438-320
438-428

:3

(j

NOTES:

* No longer in production.
(l)Required for 334 Series Magnetio Tape Units.
(2)Req\lired for 333 Series Magnetic Tape Units and 353-1 CRAM unit.
(3)Req\lired for all CRAM units and for 365 Disc units.
(4)Required for 324-1 and 324-2 Magnetic Tape Simultaneity Controllers.
(5)Required for 355-1 MICR Sorter-Reader Buffer.
(6)Required for 356-1 and 356-3 Central Inquiry Buffers and 321-1 Central Communications Controller.
(7)Required for connection of unbuffered Teletype devices.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

2/69

603:011 100

1&

AUERBACH

STANDARD

EDP

NCR 315 RMC

REPons

INTRODUCTION

~

INTRODUCTION
.1

SUMMARY
The NCR 315 Rod Memory Computer was announced in July, 1964, as the first commercially-available, general-purpose computer system utilizing a thin-film storage medium
for the entire working memory. The higher speed of this memory makes the 315 RMC
about 7.5 times as fast internally as the older NCR 315 and 315-100 central processors,
with which it is program-compatible. The 315 RMC also offers additional processing
capabilities beyond those of the other two systems in the 315 line and is a logical candidate for replacement of an NCR 315 system when the needs of the installation outgrow
the capabilities of the 315 central processor. A new Processor, the 315-502 Multiprogramming Processor was announced on September 22, 1966; this processor includes
hardware provisions to facilitate multiprogramming operations. Systems rentals for the
NCR 315 RMC range from approximately $8, 000 to over $20,000 per month.
To emphasize the similarities between the NCR 315 and the 315 RMC, only the information
that pertains uniquely to the 315 RMC is presented in this report. Material common to
both systems is presented in the NCR 315 report (Section 601:011) .

.2

COMPATIBILITY
The 315 RMC is the third in the NCR 315 line of program-compatible computer systems.
Non-time-dependent programs originally written for a 315 or 315-100 system can be run
by a 315 RMC system having equivalent memory and peripheral equipment. Programs
written for the 315 RMC which make use of its added hardware capabilities will need
modification before they can be run on either a 315 or 315-100 .

.3

HARDWARE
The Rod Memory is composed of beryllium-copper "rods", O. 015 inch in diameter, coated
with an iron-nickel substance and wound with solenoids at periodic intervals along the rod.
Each memory location is called a "slab" and is composed of 12 data bits and 1 parity bit the same arrangement as in the NCR 315. Each slab can hold two 6-bit characters or three
4-bit decimal digits. Cycle time for each memory access of one slab is 800 nanoseconds
(0.8 microsecond), making NCR's Rod Memory one of the fastest units currently available
in its price range. Each Rod Memory unit has a storage capacity of 20, 000 slabs. Up
to four of these units can be used in a 315 RMC system, for a maximum storage capacity
of 80, 000 slabs (160, 000 characters or 240, 000 digits).
Two models of the 315 RMC Central Processor are offered, and each contains control
logic for all peripheral devices. The control and process ing functions, including interrupt facilities, for the standard 315 RMC Processor have been implemented in the
same manner as in the 315. The auxiliary memory containing the accumulator, index
registers, jump registers, and processor flags is of the same thin-film type as the main
working storage and can be accessed simultaneously with the main memory.
The new Multiprogramming Processor is identical to the standard RMC Processor except
for the hardware provided to facilitate multiprogramming operations. These provisions
include:
•
•
•
•
•
•
•

Three processor operating modes - Executive, SERF (Special Executive Routine
Functions), and User;
Indirect addressing;
Separate sets of index and jump registers for the Executive and User operating
modes;
Memory protection in the User Mode via limit registers;
Processor polling and control of peripheral interrupts;
Restriction of input-output and certain other operations to the Executive Mode; and
.Internal interval timer with interrupt.

A confi!S\lration incorporating the 315-502 Multiprogramming Processor must have at
least 40, 000 slabs of memory and at least one CRAM unit (any model).
The instruction repertoire is composed of the original NCR 315 instruction repertoire
plus some additional facilities. The added instructions include data movement instructions
that aid in handling data communications input and output, floating-point arithmetic operations, and several control instructions. The data movement instructions provide automatic conversion between the one-character-per-slab or one-digit-per-slab format of
Teletype input and output and the internal format of the 315 RMC. The floating-point
© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

3/67

603:011. 300
•3

NCR 315 RMC

HARDWARE (Contd.)
operations include add, subtract, multiply, divide, and normalize. All floating-point
results can be automatically rounded. Special instructions are included in the Multiprogramming Processor to control the registers reserved for the Executive Mode.
The instruction format of the 315 RMC is identical with that of the 315. Addresses in
the instructions themselves can be no larger than 999; the index registers permit addressing up to 80,000 locations.
Two special instructions, Memory Expand and Memory Protect, are provided for compatibility with the NCR 315 and to facilitate the dual operation of one batch program and
one inquiry program. The operator can set a manual switch to select either the Compatibility Mode or the Real Time Mode. In the Compatibility Mode, the two special
instructions are used to permit addressing the full memory or to restrict addressing
to the lower 40,000 locations - the maximum memory size of the original NCR 315. In
the Real Time Mode, these two instructions allow the separation of an inquiry program
located in the upper half of memory from a batch program in the lower half of memory.
The program in upper memory is protected from interference by the program in lower
memory.
Other control instructions provide facilities for automatically storing the contents of the
accumulator and the status of the processor flags in a specified 14-slab area, and for
loading the accumulator and setting the processor flags from the contents of a specified
14-slab area.
Other capabilities of the 315 RMC are the same as those of the NCR 315. A brief description of the basic characteristics of the 315 is contained in the NCR 315 report
(Section 601:011) .

•4

PERIPHERAL EQUIPMENT
All of the peripheral devices available for the NCR 315, except the Input/Output Consoles,
are also available for 315 RMC computer systems. The configuration rules for attaching
peripheral devices to the 315 RMC are the same as for the 315. The NCR 315 report
(Section 601:011) describes the available equipment .

.5

SOFTWARE
All of the software described in Section 601:011 of the NCR 315 report is also available
for the 315 RMC. In addition, NCR has developed a supervisory program, Executive,
that enables a 315 RMC system with the Multiprogramming Processor to run several
programs simultaneously in a multiprogramming mode; detailed specifications are not
available to date.

3/67

A

AUERBACH

'"

603:221.101

51UIIiI

EDP

NCR 315-RMC
PRICE DATA

Innn.

NCR 315·RMC
IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

Name

PRICES
Monthly
Monthly
Rental Purchase Maint.

$

$

4,700(1)

211,500

210

5,800(1)

270,000

220

10,000(1)

460,000

270

$

Processing Unit

PROCESSOR
315-501

315-502

Central Processor (includes console, I/O
Typewriter) and:
One 316-501 20,000-character Rod
Memory Unit
One 316-502 40, OOO-character Rod
Memory Unit
Multiprogramming Processor; includes two
316-502 40, OOO-character Hod Memory
Units, console, and I/O Typewriter
Optional Features
Automatic Recovery Option (2)
Mixed File Adapter

100
75

5,000
3,000

-

5

Main Memory

PERIPHERAL
DEVICES

316-504

Additional Rod Memory (40,000 characters;
a maximum of three 316-504 units can be
added to the basic 316-501 unit and a
maximum of two to the basic 316-502 unit)

324-3
333-501
333-502
354-602
321-3

Magnetic Tape Simultaneity Controller
Magnetic Tape Unit (120 KC)
Magnetic Tape Unit (83.4 KC)
Card Read-Punch Controller
Central Communication Controller
(includes space for 3 Adapter Cages) (3)
Other units: see NCR 315-100(3)

2,200(1)

100,000

40

695
800
750
300
850

32,000
36,000
33,750
13,500
:36,000

5R
140
140
12
100

NOTES:
(1) Extended term rental agreements are available for these components for a two- or a three-year period.
The reduction in monthly rental for the 315-501 Processor with 20,000 characters of memory is 8r;; for the
two-year arrangement and 15tJ for the three-year plan. The reduction for the other components is 10' r for
the two-year agreement and 15% for the three-year plan.
(2) Required on all "On-Line" bank systems; the feature is required on the processor and each CRAM unit.
(3) The 315-RMC can accommodate all NCR-100 peripheral devices except the 333-101 and 333-102 Magnetic
Tape Cnits, the 354-601 Card Punch Controller (replaced bv the 354-602), the 402-4 unbuffered MICR
Sorter Reader, the 321-1 Communications Controller (replaced bv the 321-3), and the 472 series punched
card and punched tape console units. All communications adapters for the 321-1 can be used with the
321-3.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

610:011. 100

r---

A

STA ..' "

~,EDP

AUERBAC~

NCR SMALL COMPUTERS
INTRODUCTION

REPIiIS

~

INTRODUCTION
.1

•2

•3

GENERAL
The Class 395, 400, and 500 Electronic Accounting Machines are the top models of NCR's
extensive line of adding and calculating machines. These three systems bridge the gap
between completely manual data handling and full-power electronic data processing. They
are oriented primarily toward accounting and bookkeeping applications, and within this realm,
sales have fared quite well, with more than 10,000 machines installed to date.
All three systems are related in their incorporation of the proven NCR Class 33 Accounting
Machine* as a primary system component. Supplementary to this device, each machine offers increasing relief from the necessity of performing manual operations, beginning with
the Model 395; progressing thru the Model 400, which widens the applications possibilities;
to the Model 500, with its solid-state logic, stored-program feature, and automatic inputoutput units. None, however, offers quite as much general-purpose processing power,
performance, or flexibility as the best-known small-scale data processing systems such as
the GE-115, Honeywell 120, IBM 360/20, RCA Spectra 70/15, or UNIVAC 9200 and 9300.
Application possibilities range from simple invoicing and statement preparation to the
production of incentive payrolls and summary reporting. To this end, magnetic-stripe
ledger card reading and writing facilities play an important role in the Model 400 and 500
systems. All systems can use punched card input and provide punched card output. In
accommodating large data files, therefore, the external storage medium for the threE' systems is ei ther punched cards or magnetic ledger cards .
HARDWARE AND PROGRAMMING
All of the models can handle alphanumeric input-output data, but only the NCR 500 can
internally process alphabetic data. Each machine uses a similarly-structured data word;
word length is 14, 13, and 12 BCD digits for the 395, 400, and 500, respectively. The
395 and 400 can punch paper tape; the 500 can read and punch paper tape; and all can handle
punched cards. Magnetic-stripe ledger cards can be written and read by Models 400 and
500 but cannot be used with Model 395. Printout (including continuous forms) can be printed
by the accounting machine itself, or, in the case of the 500 system, by using an optional
125 -lpm line printer.
Magnetic disc storage that contains from 40 to 200 13- or 14-digit words is available for
use with Model 395 or 400. Up to 20 words can be stored on each disc track; average
access time is 19 milliseconds.
Core memory with a 22. 5-microsecond cycle time and a capacity of 200 or 400 12-digit
(BCD) words is provided with the Model 500. Processing and input/output characteristics
of the three systems are summarized in Table I.
From 12 to 43 basic operations (e. g., add, clear memory, check-digit verify) can be programmed on the various models. NCR supplies the programs and assembles the stop
control bars for customer applications. Complete operating instructions are also provided •
OPERA TION
In general use, variable data is entered on the Class 33 type keyboard, and varying amounts
of constant data, such as item prices and descriptions, can be read by the input units.
Carriage movement and initiation of processing activities are controlled by a combination
of stops located on a control bar on the accounting machine carriage, by the wiring of a
rear program panel, or by a punched paper tape program control. In addition, the Model
500 allows stored-program control of activities.
In most updating operations, the operator need only pull the proper ledger card from a file,

insert it in the carriage, key variable data in the correct columns, and refile the card.
Carriage movement and arithmetic operations (e. g., total charges and credits, develop
new balance, test for limits) are controlled automatically. Ledgers, posted at any desired
frequency, provide individual audit trails that cannot readily be produced by tab equipment
or a computer system. If the ledgers are updated frequently, they provide a visual, upto-date activity record that is immediately available.

* Detailed reports on the NCR Adding Machines and mechanical Accounting Machines are included
in AUERBACH Data Handling Reports, another looseleaf reference service published by
AUERBACH Info., Inc. AUERBACH Data Handling Reports is a comprehensive, two-volume
guide to selecting and applying the wide range of support equipment and supplies used in conjunction with computer systems.
© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

5/68

610:011.300

NCR SMALL COMPUTERS

TABLE I: CHARACTERISTICS OF THE NCR SMALL COMPUTERS
COMPUTER SYSTEM
DATA
STRUCTURE

NCR 395

---

---

14

13

Memory capacity, words

20, 40, 100, or 200

Type of storage
Access time

Disc
19 msec

40, 80, 120, 160,
or 200
Disc
19 msec

Timing (fixed-point), msec:
c=a+b
c = ab
c = alb

22
265
238

22
265
238

161
189

PROGRAMMING
FEATURES

Instruction repertoire
No. of addresses
Indexing

12
5
Yes

42
1
Yes

75
4
No

INPUT

Console
Punched cards
Punched paper tape
Magnetic ledger cards

Yes
100 cpm

Yes
100 cpm

Yes
100 cpm
650 cps
37 cpm

Printed copy
Line printer
Punched cards
Punched paper tape
Magnetic ledger cards

Yes

Yes

100 cpm
30 cps

---

100 cpm
30 cps
Yes

Yes
125lpm
100 cpm
120 cps
Yes

MONTHLY
RENTAL

Minimum system
Average system
Expanded system

$290
$355
$830

$450
$600
$795

$ 795
$1,250
$2,000

.3

OPERATION (Contd.)

PROCESSOR
PERFORMANCE

OUTPUT

---

No

---

4

NCR 500

Numeric digit size, bits
Alphabetic char. size, bits
Word length, BCD digits

STORAGE

4

NCR 400

--37 cpm

---

4
8
12
200 or 400
Core
22.5 Ilsec
11

From two to four magnetic stripes, functionally similar to magnetic tape, can be stamped
on the ledger cards processed by the NCR 400 and 500. Posting line indicators are stored
on one stripe. Stripes read by the 400 can contain up to 20 numeric words. Automatic
zero suppression allows the 500 to record up to 324 digits per stripe. The magnetic-stripe
data is automatically read and the card is automatically pOSitioned for posting when the
card is aligned between the front-feed forms guides. The magnetic-stripe data is rewritten
as the card is ejected after posting. The magnetic-striped cards are a form of external
storage that provides a basis for master file updating systems. Earnings, interest paid,
sales, and other totals can be computed and stored on the stripes for later processing.
The processing speeds of the 395, 400, and 500 are limited by the manual operations that
are required, and most of the models are not well suited for applications that require a
high volume of printing. One model of the 500 includes a medium-speed printer and enough
punched card input and output units to qualify it as a basic card system. One or two punched
card readers, one or two punched tape readers (only one tape reader if the optical journal
tape reader is used), and a magnetic-stripe ledger reader can be connected as input units;
and a card punch, a paper tape punch, and a line printer can be connected as output units.

5/68

A ..

AUERBACH

612:011. 100

1&

STANDARD

EDP

NCR 395
SUMMARY

REPORTS

AUERBACH

"

SUMMARY: NCR 395
.1

BACKGROUND
The NCR 395 Electronic Accounting Machine was announced in May 1964. It combines the
features of the NCR Class 33 Accounting Machine with an electronic computing unit. More
than 7,000 NCR 395 Systems have been installed to date.
The NCR 395 features console and punched card input, and punched card, punched paper
tape, and console printer output. A magnetic disc storage device provides either 20, 40,
80, or 200 14-digit word locations of storage for use by the user's program. Programming
is performed through a rear program panel and a front stop control bar which are installed
on the 395 carriage. Programs are "loaded" for execution by exchanging program panels
and bars. A flexible command rcpcrtoire permits twelve discrete functional operations to
be programmed, including add, subtract, multiply, divide, compare, and copy.
All computation in the NCR 395 is performed though transistorized logic. There are no
mechanical memory units. The only mechanical elements within the console are those required for printing and car riage control, and a mechanical buffer.
NCR 395 Electronic Accounting Machines were first delivered in July 1964; deliveries are
normally made from three to five months after placement of order .

.2

HARDWARE

.21

Data Structure
Each NCR 395 data word consists of 14 digits, stored as four binary bits per digit. Full
alphabetic information can also be handled, but only in input-output operations. Access to
any storage location normally obtains a full 14-digit word, but partial words can also be
manipulated by programming .

. 22

Main Storage
All data entering the NCR 395 system passes through a single-word mechanical buffer, from
which it is directed either to the electronics portion of the machine or to an output device.
The principal storage device in the 395 is a built-in single-disc drive used to store all
amounts or totals. Housed in a work-level cabinet to the left of the 395 Console, the disc
drive contains a cobalt-coated aluminum disc that is nine inches in diameter and 7/16 inch
in thickness. The disc rotates in a horizontal plane at 1750 revolutions per minute. Each
storage track is serviced by its own read/write head, permitting average access times as
low as 19 milliseconds.

Figure 1.

The NCR Accounting Machine with a 382 Card Reader.

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

1/68

612:011. 220

. 22

NCR 395

Main Storage (Conal. )
The NCR 395 Electronic Accounting Machine is available in four models: those with 20, 40,
80, and 200-word disc memories. Up to twenty words of data can be stored on each disc
track. One disc track is reserved as an "Electronic Buffer Channel" to retain in temporary
storage the last amount transacted .

• 23

Processing Unit
The central control and processing unit in NCR 395 systems is the 395 Console. In outward
appearance the Console resembles the NCR Class 31 and 33 Accounting Machines. Externally, the Console features a split-carriage, multi-form printer, a 72-character twin-shift
typewriter, a 14-row adding-machine-style keyboard, a rear control panel, a front program
bar, and various control keys and indicators. Internally, the 395 Console contains the
solid-state logic necessary to perform all electronic calculations and to initiate input-output
operations.
The Console printer contains 260 carriage positions, and all processing is performed on a
single-stop basis; up to five discrete operations (I/O, arithmetic, compare, etc.) can be
performed at each carriage stop. All processing at a given carriage stop can be performed
in a single machine cycle (or within 400 milliseconds). As directed by the rear program
panel (described in Paragraph. 3, Programming), the Console uses a "five-address" command structure to perform its operations at each stop.
The operations that can be performed (many within the same machine cycle) include add and
subtract, multiply and divide (rounded or unrounded), selective stores, copy shift, compare
for equality, test comparison indicators, and table lookup. Positive and negative values
determined by the Test command cause forward or reverse carriage movement to programmed stops on the front program bar of the carriage. The new stop position then initiates
the next machine cycle and its five-part machine instruction.
All internal processing is performed directly on values in storage; e. g., an Add instruction
adds the value of an amount, just entered into the system and stored in the mechanical
buffer, to another value in storage where the sum is accumulated. Information enters the
mechanical buffer either through the Console keyboard or an on-line card reader. The
Multiply command takes the value in the mechanical buffer, multiplies it by the contents of
a specified storage location, and stores the product in another specified storage location
(c = a x b). As examples of the processing speed of the NCR 395 Console, a 14-digit multiply
is completed in 265 milliseconds, and a 14-digit divide is completed in 238 milliseconds.
Optional features in the NCR 395 include the Divide command, a special "IVS" package that
is standard on the Model 395-303, and the new 80- and 200-word models. The IVS package
includes a "step rate" feature that automatically performs a table lookup to obtain appropriate tax rates, depending on salary, for payroll applications. The same feature is also
useful in billing applications. The IVS package also includes an automatic check-digit
verifier for modulo-l0 account numbers, and a card search feature that causes consecutive
card reading until a specified account number or other control field is discovered.

.24

Input-Output
The basic input device in NCR 395 systems is the 395 Console keyboard, as shown in Figure
2. Each keyboard entry can represent a variable amount to be added or subtracted (charges,
credits, etc.) or descriptive information not to be included in the totals (salesman number,
sales terms, etc.). An entry is recorded by depressing keys representing the data and a
motor bar. The middle motor bar is used for recording normal entries (e. g., add from the
charge column, multiply times price from the quantity column), and the upper and lower
motor bars are used for exception entries.
The amount keyboard (in the style of a full-keyboard adding machine) provides 14 amount
rows with two implied decimal points. Accuracy is provided to three decimal places. Numeric information can be keyed into the amount keyboard for use as an immediate operand
or as a stored constant, depending on the program. Alphabetic information can be entered
in upper case only via the electric typewriter keyboard where it is printed and, optionally,
sent to the card punch or paper tape punch as alphanumeric output. The 48-key twin-shift
keyboard has a 72-character set.
As an option, NCR 395 machines can also be equipped with a Model 382-1 Card Reader. The
382-1 is a compact, 25-pound, desk-top unit that can read punched cards at a peak rate of
100 cards per minute (reading 14 columns per card). Selected fields within each card can
be read under the direction of a user-prepared program disc within the card reader's cabinet.
Hollerith-coded alphanumeric information is read by star wheels, converted to binary code,
and transmitted to the system's main storage.
Constant data such as employee name and wage data can be punched into cards from the keyboard. Payroll checks and a payroll journal (including year-to-date and deduction information) can then be prepared by reading constant data from the cards and entering variable
data from the keyboard. Updated cards are punched as the checks and journal are printed.

A

AUERBAC~

(Contd. )

SUMMARY

612:011. 240

2

6

5

Figure 2. The NCR 395 Console.
024

Input-Output (Contd.)
The standard output device of the NCR 395 Electronic Accounting Machine is the dual printing
mechanism of the Console's adding machine and typewriter. The adding machine printer consists of a block of 18 print bars, 14 of which can print the numeric character set plus three
special symbols. Three print bars in the block are used for printing of commas, and a fourth
bar for the decimal point. The typewriter's stationary printing mechanism is located about
five inches to the left of the numeric print block. Both printers can print at any of the 260
print locations along the 26-inch carriage on the Console. However, for any given application only 65 stops can be programmed because of the minimum tabbing distance of four
spaces. Removable stops in the control bar regulate the horizontal movement of the carriage at each print position and control forms spacing. The control bars are exchangeable
to fit the specific application.
The print platen is split into 10-inch and 16-inch sections, and each is capable of independent
operation using different ledger cards, statement forms, journal forms, checks, etc., as the
printing media. Individual carbon-backed or NCR paper forms (such as customer statements)
can be inserted in front of forms already in the carriage, or data can be printed on a form in
the left side of the carriage and all or part of it can be automatically repeated on a form in the
right side of the carriage.
A statement, ledger, invoice, and journal printed by the 395 are shown in Figure 3.
The Forms Feeder option permits use of fan-fold continuous forms with a maximum width of
26 inches. With this feature, vertical spacing is performed under control of user-prepared
carriage-control tapes. The 395 Console also provides control keys for horizontal carriage
control and vertical paper spacing.
Punched card output is available as an option with the NCR 395. The NCR 376-6 card punch
(an IBM 26 Printing Card Punch) can be connected to the NCR 395 system through couplers.
It typically punches cards at about 18 columns per second. By front program-bar programming and the use of a wired selection board, numeric data from the 395's main storage can
be punched. Alphabetic information originating from the Console typewriter or the card
reader can also be transferred to the on-line punch for punched card output.
Punched card output can also be obtained through use of the NCR C-465-202 Alpha-Numeric
Keypunch, manufactured by Bull-GE, which connects directly to an NCR 395 and punches at
25 columns per second.
The only other available output device for NCR 395 systems is the NCR 462-1 Punched Paper
Tape Recorder. The 462-1 can punch data in 5-, 7-, or 8-level code at a peak rate of 30
characters per second. Tape reels can hold up to 1,000 feet of paper tape. Control of paper
tape punching is programmed by a wired NCR 395 selection board in conjunction with preset
stops on the front program bar of the Console's carriage. Punching of alphanumeric information requires use of a 434-2 Alpha Tape Coupler and a 434-1 Universal Switch Bank (to
generate alphanumeric impulses from the typewriter keys).
© 1968 AUERBACH Corporation and AUERBACH Info. Inc.

612:011.241

NCR 395

SALES JOUI_

......

THE ROIERTS COWAN'
444 ""UCE STIllEtT
DAYTON 2t OHIO

==

==

= ":"

ci:..
~ ~S fR£1I1tiT I='~
~~
_moo
F+-=;;;:'+~r~-+=+=+-;';;"+=+~
,

&l5t.JO

7.7. !to.OO '67• .0

, ... , :I.:S'U.O 1.20'.20

.00

:.:::

.IU-

2.,,,'''0
:uS".

"Ul
leU 01

7..,•.0,

12..... '

1.'''',10

1.'lIIt.10
1."0"0
:U"O
~oou.

70010
•• 060
860.60
7.0",0

2.2,"'20
2.731.20
1,1IUO
,"OU.l1

stATEMINT

THE ROBERTS cOtoP",,"
444 SPRUCE STREET

DAYTON 2' OHIO

I.LLllIa JOURNAL

,SOLD TO

r

SHlml tal

r

Tt£ ROeERTS CQW'ANY
444 SPRuct STREET

Tl1E fIOKRH OOY'ANY
9054 PARK AvtNJE

RICHFIELD 10 CALlf'ORNIA

DAYTON 29 OHIO

L

""fto~ ..

=~'J

•

TO

"

1.....1
\':362

11_ " .

lo;~T

MOTOR FREIGHT

I
ASSEt.eLY NO 16M3 SPEC
IliACHINE YODEL 4032
C(UI(INENT PARTS

......

r~

,
"

6•

4 6-

MU

.-~

12463

... ...

10.00

117"0

A2
A2
A2

tOO

-."...... ....
,

TUell

~f:m...
"'.00
1.11225.
.... 0

In,

JUt, 1.260.00

I ' Ia..... ~

Figure 3. Printed output from the NCR 395 .
. 24

Input-Output (Contd. )
Monthly statements, posted at any desired frequency, provide individual audit trails that cannot readily be produced by tab equipment or computer systems. If the ledgers are updated
frequently, they provide a visual, up-to-date record that is immediately available. A master
file on ledger cards is particularly well suited for low-activity updating applications. However, the processing speed of the 395 is limited by the manual operations that are required,
and the 395 is not well suited for applications that require a high volume of printing .

.3

PROGRAMMING
Programming the NCR 395 is performed primarily through use of an interchangeable control
panel on the rear of the Console. All programmed operations are performed relative to
individual positions or stops along the front carriage of the printer (i. e., at each carriage
stop, several programmed operations can be performed). Carriage movement to the stops
is initiated by the depression of a motor bar or control key.
The rear control panel consists of 260 programmable stop positions that correspond to the
260 print positions along the 'carriage of the printer. Each programmable position consist::
of a row of 57 slots, into which plastic selector plates can be inserted to activate up to 57
sensing switches. Each activated sensing switch initiat"ls a discrete operation or specifies
an operand address in disc storage.
The 57 switches that can be programmed at each carriage stop of the printer are logically
grouped into six segments: one master command segment and five "address" segments.

A.

1/68

.,

AUERBACH

(Contd.)

612:011. 300

SUMMARY

.3

PROGRAMMING (Contd.)
The command segment provides ten switches, among which are those that specify multiply,
divide, copy, table lookup, and test operations. A positive or negative result of a test operation sends the printer's carriage in forward or reverse direction to the next carriage
stop preset on the front program bar; when the carriage arrives at the preset stop, another
set of programmed operations is performed.
The five remaining segments in a given stop position on the rear control panel can be specified to perform such operations as keyboard and card reader input, storage clear, and
add/subtract. A segment consists of from 4 to 11 programmable switches. Each segment
can be programmed to perform either one simple and discrete operation or to combine with
other segments to perform a more complex operation. Since a single operation can be
performed by each of five segments at a given stop position, NCR refers to the machine as
a computer with a "five-address" instruction.
For example, an amount entered into the keyboard (via Segment I) can be added to another
amount at an address in disc storage that is specified in Segment II; the same keyboardentered amount can be added to a second, third, and fourth accumulator by so programming
Segments ill, IV, and V. By contrast, a mUltiply operation utilizes several segments to
perform the single operation: the Command segment specifies the multiply operation; Segment I specifies the location of the multiplicand; Segment II specifies the location of the
multiplier; and Segment III specifies the location of the product.
Four of the segments can directly address any of the 120 words of disc storage by means of
a modified binary addressing scheme. Eight switch positions within each segment are
weighted as follows: 80, 40, 20, 10, 8, 4, 2 and 1. Any operand can be directly addressed
by inserting plastic selector plates in the particular combination of switch positions whose
total of weighted values equal the operand address. For example, disc storage address 96
can be specified by inserting selector plates in switch positions with weights 80, 10, 4 and
2.
Operations programmed via the rear control panel are often performed in conjunction with
carriage control and input-output operations programmed via the front program bar. The
interchangeable front program (or control) bars are similar to those used with NCR Class
31, 32, 33 and 35 Accounting Machines. The bar is programmed by setting removable
,metal plates or "tab stops" at selected print positions. The front program bar controls
formatting of printed output, forms spacing, date printing, etc. Inserting special plates at
specific locations along the bar can also control input card reading and output card punching
after these operations are initiated by the rear control panel.
Because of the combination of mechanical and electronic functions and the dual function of
many of the control keys, programming the NCR 395 is rather complex. Carriage movement must be coordinated with the arithmetic and input-output operations. However, many
users of this equipment will not concern themselves with programming considerations,
since NCR supplies its 395 machines already programmed to the specifications of the users'
principal applications .

.4

PRICE DATA

Component or Feature
NCR Electronic Accounting Machine:
Model 395-100
Model 395-200
Model 395-300

Monthly Rental
(I-Year Lease)
$

Purchase
Price

460 to 560
495 to 595
310 to 395 (2)

15.900 to 18,900
17 .. 00 to 20,400
10,900 to 12,500

$

Monthly
Maintenance
$ (1)
71

92
45

Options
465 Alpha-Numeric Keypunch
376 Card Punch
Numeric Card Coupler
Alpha Output Switch Bank
Alpha Card Coupler
rvs Feature

150
87
40
25
15
40 (2)

6,990
5,150
1,240
1,000
240
600

27
26
12
4
7
6

NOTE: The machine prices do not include the optional input-output units. IBM Card Punch prices
range from $2,000 to $4,500.
(1) Slightly-higher in rural areas.
(2) Also available on a three-year lease at lower rates.

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

1/68

6f3:0ft. tOO

A

SUI.AlD

EDP

NCR 400
SUMMARY

REPllrs

AUERBACH

SUMMARY: NCR 400
.1

BACKGROUND
Announced in January 1967, the NCR 400 Electronic Accounting Machine adds magneticstripe reading and writing ability to the capabilities of the NCR 395, as described in
Report 612. In addition, programs and constants can be read from punched tape, allowing
greater flexibility.
The NCR 400 features keyboard and punched card input; punched card, punched tape, and
console printer output; and magnetic-stripe read and write heads on the console carriage.
The front stop control bar controls carriage movement and printing. From 40 to 200
13-digit words of disc storage are available in increments of 40 words. The NCR 400
Electronic Accounting machines were first delivered in December 1967, and over 100
systems were delivered during that month. Deliveries are normally made from five to
seven months after an order is placed .

.2

HARDWARE
All NCR 400 hardware is identical to that of the 395 (Report 612), except for a wider choice
of internal storage sizes (40, 80, 120, 160, or 200 words of disc storage). two 16-digit
arithmetic registers, a smaller word and keyboard size (13 digits), a magnetic-stripe
capability, and the ability to read punched program tapes. The principal characteristics of
the NCR 400 are compared with those of the NCR 395 and 500 systems on page 610:011. 100 .

. 21

Magnetic Ledger Card Input-Output
The dimensions and capacities of ledger cards that can be processed through the magneticstripe carriage are shown in Table I and Figure 2. Two magnetic stripes are stamped on
the back of the cards; one stripe is for data storage and the other is for line-finding
indicators. Front-feed forms guides are included on the carriage so that individual magnetic
and non-magnetic forms can be inserted in front of a continuous journal. The cards are
aligned between the guides, the magnetic stripe data is read, and the cards are automatically positioned for posting the next line. Data is written on the stripes in a 4-bit binary
code at 20 numeric digits per inch. When the last posting line is printed, an indicator
is lit. The required heading and cumulative data is automatically printed and encoded on a
new ledger card when it is inserted in the carriage.

Figure 1. The NCR 400 Electronic Accounting Machine with the 465 Keypunch (left) ana
382 Card Reader (rjght).

o

1968 AUERBACH Corporation and AUERBACH Info. Inc.

1/68

613:011.210

NCR 400

TABLE I: CAPACITIES OF NCR MAGNETIC-STRIPE LEDGERS
Magnetic-stripe ledger length (inches)
Capacity
10

11

12

13

14

15

15

17

19

20

20

20

33 to 22

39 to 27

45 to 33

52 to 40

58 to 45

64 to 51

13-digit magneticstripe words
Posting lines
(with 1/2- to 3inch headings)

-\

1....- - - - - - - - - - - - 8" to 16.5"

1. 0".

to
3.0"

....ME

DONALD

-

-;£~=_ -=---~- =-~~~

!BQ9

... ,,£>'11:&&
<;'V........ DSU'!'!:

1563

IIOC'A.. SECU",TY NO

__...

___=

_______________

pHONE

444-2529

~~~~;A:;rH ~t~JllQ

_

_ ,9TI1I''''

~

_________

__

x vn

[) ..o

X ....... J

"'-"1- O. 9"

0 I'E"''''''£

ea:r~~PL_Qng

-------".Af.l2!!!.~t.'"'

.L
10"

to
15"

Dec 31,6-

615

325

1250

3

200

10000

f

3.2"
P.I ........... , 6' ....."."

NOTE: l;ata cannot be printed in the bottom or right margin.
Figure 2. Ledger card printed by the NCR 400 .
. 21

Magnetic Ledger Card Input-Output (Contd.)
Magnetic-stripe ledger cards are a form of external storage that provides a basis for an
elementary master-file updating system. Earnings, interest paid, sales, and other totals
can be computed and stored on the magnetic stripes for later updating. Continuous singleor multiple-part forms, up to 26 inches in Width, including 1/2 inch on either side for
pinfeed holes, can be fed through the console carriage by the optional Continuous Forms
Feeder (CFF). The forms can be from 3 1/2 to 12 inches in length. The dimensions
of forms that can be processed through the 26-inch carriage are shown in Figure 3. A

~
. " 'I:r

JOURNAL(25'.") ~

,."

9.8"---j

14.8"

.95"

II+-STATEMENT - . . j . - a . 5 "
NON-MAGNETIC---I f..--MAGNETIC/NON-MAGNETIC
FORM
LEDGER
L.EFT PLATEN

RIGHT PLATEN

FIXED F

GUIDE

NOTE: An 8"/18" split is also available.
Figure 3. Specifications of the NCR 400 split platen.

1/68

A•

AUERBACH

(Contd. )

SUMMARY

.21

613:011.220

Magnetic Ledger Card Input-Output (Contd.)
3-channel carriage tape and two console keys control forms spacing and skipping when the
CFF is used. Depression of the Field Position Switch causes the carriage to open and space
to the next printing line. Depression of the Home Position Switch causes the carriage to
open and to skip to the first line to be printed on the next continuous form .

. 22

Punched Tape Input-Output
The optional 462-1 Punch Paper Tape Recorder provides paper tape output punched with any
5-, 6-, 7-, or 8-level code. It provides up to six punching formats. The formats can be
modified by wiring through switches actuated by an add, subtract, or non-add condition.
The character set is limited to 10 digits and 12 special characters. With the zero suppression option, non-significant zeros are not punched and a character is emitted to identify the
starting position of the entry in the data field.
The paper tape may be 7/8, 11/16, or 1 inch in width. Paper tape reels contain 900 feet
of blank tape, and the take-up reels hold 350 fflet of punched tape.

.3

A tape reader is provided to read the aluHlillum-coated Mylar program tape .
PROGRAMMING
The NCR 400 is controlled by programs read from punched tape and by a front stop bar.
Commands such as add, subtract, multiply, divide, test, clear, and print are read from the
program tape and executed. Constant data can also be punched into and read from the
program tape. A sample section of a program tape is shown in Figure 4. Representative
commands and their execution times are shown in Table II. The machine cycle time is
approximately 400 milliseconds. Computing and program tape reading can occur concurrently during the last 135 milliseconds of a cycle. The program tape is read and
searched photoelectrically at 118 characters per second. The front stop bar controls
carriage movement and printing. Carriage movements of up to O. 4 inch can be performed
during a cycle.
TABLE II: NCR 400 COMMAND TIMES
Command Name

Execution Time in Milliseconds

ADDRESS
ADD
ADD PUT AWAY
ALPHA (Read from punch card)
CLEAR MEMORY
CLEAR MEMORY STRING

19 (average)
13.6 to 41 (average, 22)
24 to 56 (average, 38)
100 per typed character, 7 per skipped character
19
27 + 1. 7 per address cleared

DELETE
DIVIDE
EJECT (Read punch card)
EXTRACT
KEYBOARD ENTRY
LEDGER FULL (TEST)
MODIFY ADDRESS ¢
MODIFY ADDRESS, POSITIVE
MULTIPLY

8.5
238 (quotient = 55555) average
17 not time-shared, 7 per column time-shared
22 to 90, (average = 48)
275, not time-shared.
25 + 8.5 per step skipped
35
35
265 (multiplier = 55555) average

READ LEDGER - NEW
CHECK DIGIT VERIFY
READ PUNCH CARD (NUM)
READ PROG. TAPE - Start
READ PROG. TAPE - End
PAPER FEED
PRINT
PRINT AUTOMATIC
READ LEDGER
SEARCH PROG. TAPE

Approximately 1. 5 seconds for an 11" card
265 (same as multiply)
17 + 7 per digit read
17 + 8.5 per digit read
17 + 8.5 per digit read
Approximately 150
275 (not time-shared)
275 (not time-shared)
2 seconds average (11" card), plus operator time
Forward = 8.5 per step skipped
Reverse = 8. 5 + 8. 5 per step skipped

SELECT ADDRESS
SUBTRACT
SUBTRACT PUT AWAY
COMPARE FOR ZERO OR POSITIVE
TRANSFER
READ LEDGER - EJECT
WRITE LEDGER

19 (average)
13. 6 to 41 (average, 22)
24 to 56 (average, 38)
25 + 8.5 per step skipped if branch
8.5
2.5 seconds average, plus operator time
2 seconds (average)

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

1/68

613:011. 300

NCR 400

Figure 4. Sample section of an NCR 400 program tape .

•3

PROGRAMMING (Contd.)
The NCR 400 console includes: on-off switches, peripheral control switches and indicators,
error-correction control keys, and keys for manually controlling most of the programmable
functions. Memory words can be addressed through the use of the Load Address key, the
middle motor bar, keys in columns 1, 2, and 3, and the Z key (for the second arithmetic
register) .

.4

PRICE DATA
Monthly
Rental
$ (1)

Component or Feature

Purchase
Price
$

Monthly
Maintenance
$

NCR 400 Electronic Accounting Machine:
Model 400-100;
40-word memory

430 to 480

15,200 to?

Model 400-200;
SO-word memory

525 to 630

17,500 to 21,800

83. 33 to 101. 67

Model 400-300;
I20-word memory

575 to 680

19,500 to 23,800

87. 50 to 105. S3

Model 400-400;
I60-word memory

625 to 730

21,500 to 25,800 91. 66 to 110.00

Model 400-500;
200-word memory

675 to 780

23 , 500 to 27 , 800 95.83 to 114.17

(2)

NOTE: The above prices include the basic accounting machine and specified memory.
Prices for optional input-output units are shown in Paragraph. 4 of Report 612.
(1) Prices shown are for one-year leases; monthly rates for three-year leases range
from $375 to $730.
(2) Not currently available.

1/68

fA..
AUERBACH

614:011. 100

£

-

STAND ...

AEDP

AUERBAC~

NCR 500

REPORTS

SUMMARY

~

SUMMARY: NCR 500
.1

BACKGROUND
The NCR 500 Computing System allows the gradual implementation of automatic accounting
and bookkeeping techniques within small business environments. Companies whose gross
volume is in the $8,000,000 to $65,000,000 range are primary market prospects for this
small-scale data processing system. NCR delivered the first 500 system in October 1965
and has installed about 1600 of them to date. Although the 500 has the capability for automatic processing, over half of the present installations maintain a mix of manual and
automatic operation; this mix continues because the magnetic-stripe ledger cards are
popular with NCR customers.
The NCR 500 system is primarily used for billing, payroll, cost accounting, inventory,
banking, and sales analysis applications.
Monthly rentals can range from $795 for the minimum configuration to $2,620 for the full
system with all peripherals. Average rental for all installed systems is about $1, 500 per
month.
The NCR 500 Computing System utilizes a C-517-1 Processor using discrete transistor and
diode circuitry, 200 or 400 12-digit (BCD) words of magnetic core memory, one of three
console models, and peripheral devices that can include paper tape reader and punch,
punched card reader and punch, a magnetic-stripe ledger card reader, an optical journal
tape reader, and a line printer. The system operates under the control of a program
stored in memory. The instruction repertoire consists of 19 basic instructions that can
be modified to produce over 100 individual instructions. Some of these instructions can be
used with mechanical devices, paper program tapes, and program discs on the peripher,u
devices to produce a wide range of input and output data formats.
Each instruction occupies one word in memory and consists of a 2-digit command code, a
I-digit modifier, a I-digit register code, and four 2-digit addresses. The four addresses
normally specify the memory locations of the first operand (A), the second operand (B),
the result (C), and the next instruction (NI). ill order to address all 400 memory locations
using only two BCD digits, the register code selects digits of an address register to find
the plane in which the address is located. (The memory consists of 2 or 4 planes, with
each plane containing 100 12-digit words. )
The three console models available for use with the C-517 Processor are: the relatively
simple C-521-1 numeric console with journal tape printing; the C-590-2 alphanumeric
console, with a 26-inch carriage that can perform a wide range of functions controlled by
programmed instructions, keys on the console, the block assembly, and mechanical inserts;
and the C-590-1 console, which can read and write magnetic-stripe ledger cards in
addition to the functions of the C-590-2.

Figure 1. A typical NCR 500 Computing System. Shown from left to right: C-562 Strip Tape
Reader, C-571 Tape Punch, C-561 Tape Reader, C-517 Processor, C-521 Console, and
C-541 Line Printer.
© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

1/68

614:011. lOT

.1

NCR 500

BACKGROUND (Contd.)
One useful, but complex, feature of the C-590 consoles is the ability to override carriage
tab stops with programmed instructions from the processor. This feature eliminates the
need for mechanical changes in the carriage controls between jobs if all job formats are
planned when the carriage assembly is made up. This overriding feature is also used with
the C-541-1 Line Printer to allow mixed alphabetic and numeric printing across fields.
Although these features indicate that varied formats are possible, data in a fixed format
can be processed more efficiently; editing routines are costly in both time and memory
space.
The variety of functions that can be performed on the consoles is an attractive feature of
the NCR 500 systems. But, in order to fully utilize the processor and the stored program,
the console should be used sparingly and primarily to initiate the internally stored program that obtains- input data automatically from tape, punched cards, or magnetic-stripe
ledger cards and produces output automatically on the tape punch, card punch, and/or line
printer. Since magnetic-stripe ledger cards can be updated only in the carriage of the
console, at comparatively low speeds, their use should be carefully thought out when including them in the system.
The NCR Series 500 Computing System is well suited for small-volume accounting applications. The variety of tasks it can perform is large, but the overall speed of the system
(see Table 1) is significantly lower than those of most stored-program computers.
The programming of the processor is relatively easy, but the organization of processing
runs, formatting of data, and making up of the carriage assembly requires well-trained
personnel. Although NCR has more than a hundred preprogrammed routines covering
TABLE I: EXECUTION TIMES FOR REPRESENTATIVE INSTRUCTIONS
Explanation of Operation

Instruction

Time, msec

Add

Two 12-digit numbers and store sum

11. 29

Multiply

12-digit multiplicand by 5-digit multiplier,
store rounded 12-digit product

32.30

Divide

12-digit dividend by 5-digit divisor,
store unrounded 12-digit quotient

160.57

Read Punch Card

All 80 columns, numeric

583.45

Punch Card - Serial

All 80 columns, numeric

3595.24

Punch Card - Parallel
(uses C-551 Buffer)

All 80 columns numeric, or 48 columns,
alphanumeric
Processor time
Available buffer time

600.00

Read Paper Tape

One 12-digit word
Field of five 12-digit words

50.04
130.20

Punch Paper Tape

One 12-digit word
Field of five 12-digit words

112.69
512.53

Line Print (uses
C-551 Buffer)

One line of 96 numeric digits or
48 alphanumeric characters

480.00

Read Ledger Card

10-inch-Iong ledger card
Processor time for five 12-digit words

1400.00
448.00

Manual Ledger Card
Handling*

Per ledger, any allowable size

2000.00

Read Console

12-digit word

Operator
speed

Print on Console

12-digit word

Tab time
+540.00

Tab Time

0.5
1. 5
4.0
20.0

120.00
260.00
500.00
2030.00

inch
inches
inches
inches

77.40
37.00

*Depends on operator's speed.

1/68

A

(Contd. )

AUERBACH

co

SUMMARY

.1

614:011. 200

BACKGROUND (Contd.)
a wide range of small business applications, NCR recommends that each installation have
at least one operator-programmer who can change these routines to fit the specific
applications. The help NCR provides in this respect varies from area to area; in large
metropolitan areas, NCR has programming staffs available to set up the entire installation;
in rural areas, a salesman may not have the expertise to provide this help. NCR feels
that the best results are generally obtained in installations in which the customer's staff
assumes at least 50 percent of the programming responsibility.
The first NCR 500 Computing System was delivered in October 1965. They are now available for immediate delivery .

.2

HARDWARE
Each NCR 500 Computing System consists of a C-517-1 Processor, a Console, a C-551
Buffer (if a C-541-1 Line Printer or a C-577-1 Parallel Punch is used), and the selected
peripheral devices. Figure 1 shows a typical installation.
Input to the C-517-1 Processor can come from:
•

One or two C-561-1 or -2, C-562-1, or C-563-1 Paper Tape Readers;

•

One C-420-2 Optical Journal Tape Reader (instead of one C-562-1 Reader);

•

One or two C-582-1 Punched Card Readers;

•

One C-586-1 Magnetic Ledger Reader; and

•

One C-521-1, C-590-1, or C-590-2 Console.

Paper tape can be punched off-line with a C-464-2 Alpha-Numeric Printing Tape Punch or
with a Class 31, 32, 33, or 35 Accounting Machine coupled, via a C-434-5 Alpha Tape
Coupler, to a C-462 Punched Paper Tape Reader. Punched cards can be produced offline with any keypunch using the Hollerith code. Magnetic-stripe ledger cards cannot be
produced off-line.
Output from the C-517-1 Processor can be produced on:

. 21

•

One C-576-1 Serial Card Punch or one C-577-1 Parallel Card Punch
(requires the C-551 Buffer);

•

One C-541-1 Line Printer (requires the C-551 Buffer);

•

One C-571-1 or C-472-1 Paper Tape Punch; and

•

One C-590-1 (with magnetic stripe ledger card reader and writer)
or C-590-2 Console .

Data Structure
The C-517 Processor uses a 12-digit (BCD) word as the basic unit of data. The word does
not include a sign; negative numbers are expressed as nine's complements. Each alphanumeric character is expressed by 2 BCD digits; thus, each memory location can hold
six alphanumeric characters.
When words of less than 12 digits are entered into the memory, an end-of-word signal is
generated by the input device to fill out the balance of the word with zeros. The output
device can suppress insignificant zeros transferred from memory. A negative symbol
will fill out the word with nines when negative numbers of less than 12 digits are read
into memory.
The core memory consists of 2 or 4 planes of 100 words each; the planes are labeled "0",
"2", "4" and "6". Memory location 99 of plane 0 is adjacent to location 00 of plane 2;
location 99 of plane 6 is adjacent to location 00 of plane 0; but, in the 200-word memory,
location 99 of plane 2 is not adjacent to location 00 of plane O. Most instructions that use
consecutive memory locations can operate across adjacent planes.
Each instruction consists of one word, decoded in the following manner:

Command

Addresses

R

l:Y

Code Mod. Sel.

12

In

10

9

A
8

I

B
7

6

I

NI

C
5

4

I

3

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

2

I

1

Digit

1/68

614:011. 210

.21

NCR 500

Data Structure (Contd.)
Digits 11 and 12 designate the code of the basic instruction. Since there are only nineteen
basic instructions, the 12th digit is a "0" or "1" for all instructions. To utilize further
the 12th digit, it also designates the plane in which the address of the next instruction is
located.
The tenth digit modifies the basic instruction to select the peripheral device in input and
output instructions, to override mechanical stops in the console instructions, or to qualify
the instruction in some way.
The ninth digit is a register select code that selects digits of a memory register to identify
the plane in which the A, B, and C addresses are located. The memory register must be
set by an instruction each time any of the addresses refers to a new plane other than "0";
addresses in the "0" plane can be referenced without using the memory register.
Digits 3 through 8 specify the memory locations of the A, B, and C addresses within the
plane specified by the R Sel code and the memory register. Digits 3 and 4 are used as
an alternate exit for comparison instructions and as a control code in some instructions
not requiring a C address.
Digits 1 and 2 select the address, within the plane specified by the 12th digit, of the next
instruction.
Data is transferred to and from the peripheral devices in a 4-bit BCD code; see Table II.
Numerics use one BCD digit, and alphanumerics use two. The binary numbers 10, 11,
13, 14, and 15 in each BCD digit are used as control codes:
10 = AI - alternate instruction. When read, the next instruction is not specified
by NI but by the C address. Used by the operator at the console to reach a
recovery routine for error correction.
11 = EOT - end-of-tape. When read, the next instruction is specified by the C
address and not by NI.
TABLE II: NCR 500 ALPHANUMERIC DATA CODE
MEMORY
DIGIT
CODES

10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
33
40
41
42
43
44
45
46
47
48
49
53
54
61
62
63
64
65
66
67
68
69
73
74

1/68

PUNCHED CARO
LINE
PRINTER

TVPEWRITER

CHAR-

FORMAT

0
1
2
3
4
5
6
7
8
9

0
1
2
3
4
5
6
7
8
9
12
12-1
12-2
12-3
12-4
12-5
12-6
12-7
12-8
12-9
. blank
12-3-8
11-0
11-1
11-2
11-3
11-4
11-5
11-6
11-7
11-8
11-9
11-3-8
11-4-8
0-1
0-2
0-3
0-4
0-5
0-6
0-7
0-8
0-9
0-3-8
0-4-8

0
1
2
3
4
5
6
7
8
9

0
1
2
3
4
5
6
7
8
9
&

&

A

A

A

B
C

B
C
D

B
C
D

•

D

E

E

E

F
G

F
G

F
G

H

H

H

I

I

I

space
space
$
J

space
(.)

none
(.)

space
J

J

K
L
M
N
0

K

K

P

L
M
N
0
P

L
M
N

0
P

~

~

space

,

tab/EOR
tab
space

is

is

is

T

T

T

U
V

U
V

U
V

W
X
Y

W
X
Y
Z
ret/EOR
ret

W
X
Y
Z

~

space

Z

space
space

fA..

AUERBACH

PUNCH

ACTER

$

*/

%

(Contd. )

SUMMARY

.21

614:011. 220

Data Structure (Contd.)
13 = EOW - end-of-word.

Signals the processor to zero-file the rest of a word.

14 = EOR - end-of-record. Used with instructions that read from punched
paper tape, punched cards, or the console to signal the end of a variablelength record.
15

=

DEL - delete. Used for error correction.
processor.

This code is ignored by the

The alphanumeric character set includes the numerals 0-9. upper-case letters, control
codes, and special characters. The special characters are listed below:
Memory Digit Code
20
30
33
40
53
54
61
73
74
.22

Punched
Card

Line
Printer

Typewriter

*space

&

&

space

none

space

(. )

(. )

$

space
tab/EOR
Tab
space
ret/EOR
ret

space
space

(,)
space
space

$

*1
(.)

%

Punch
Format
12
blank
12-3-8
11-0
11-3-8
11-4-8
0-1
0-3-8
0-4-8

Console
The C-521-1 Console has a numeric keyboard with journal tape printing and control keys
for operating the computing system. Its main function is to access the memory for initiating, debugging, and correcting the program. Instructions or numeric data can be entered
into the processor from the keyboard. Normally, this keyboard is used in an application
where all of the input and output data is on punched paper tape or punched cards and the
printing is done on the C-541-1 Line Printer.
The C-590 consoles have a 26-inch carriage with a split platen, a continuous forms feed,
and an alphanumeric keyboard with processor controls. The C-590-1 has a magneticstripe ledger card reader and writer built into the carriage; the read/write heads have the
automatic line-find feature. The ledger cards can be read and updated by a program stored
in memory. An indicator on the console tells the operator when the processor is on a
magnetic ledger card read instruction; once the processor is finished with the ledger
card, the program can automatically eject it. On new C-590 Models, the typewriter
keyboard can enter alphabetic data into memory.
The stop block combined with plates inserted in the carriage assembly provides a wide
range of formats that can be printed by programmed instructions. The modifier (10th digit)
of the console print instruction can override some of the plates in the carriage assembly
to effect a change in the pre-set format; e. g. , a carriage return to a #1 insert can be
overridden by a "4" modifier to skip to a #2 block.

. 23

Continuous fanfold forms can be printed on the C-590 Consoles by means of a Continuous
Forms Feeder controlled by a disc. The disc size is determined by the number of lines
printed per fold. Since the spacing on the carriage is 6 lines to the inch, the length of all
forms must be divisible by 1/6 inch .
Punched Paper Tape Input and OUtput
The paper tape readers and punches use I-inch paper tape with eight channels plus a sprocket channel. Of the eight channels, only five are used: four for the BCD code and one for the
odd parity check bit.
Numeric data is stored one digit per row; alphanumeric data requires two rows per character. Each word ends with an EOW (13 or 1101) code. When the tape is read for internal processing, the EOW code signals the processor to complete the word with zeros,
and the EOR code signals the processor to fill the record with zeros. If the EOW code
does not follow every 12th digit, the processor halts and indicates an error. If the tape
read instruction tries to terminate before the EOR code is received, meaning the record
is too long for the allotted space in memory, the processor halts and indicates an error.
Tape on the 561-1 or 561-2 can be rewound by a programmed instruction; all tape units
except the C-562-1 have a manual rewind control.
Paper tape input can be prepared off-line by using an NCR C-464-2 Alpha-Numeric Printing Tape Punch or by coupling, via a 434-5 Alpha Tape Coupler, any NCR Class 31, 32,
33 or 35 adding machine, with a 48- or 72- character typewriter, to a 462 Punched Paper
Tape Recorder.
© 1968 AUERBACH Corporation and AUERBACH Info. Inc.

1/68

614:011. 230

.23

NCR 500

Punched Paper Tape Input and Output (Contd.)
The C-420-2 Optical Journal Tape Reader can be connected to the processor by using
the circuitry normally used for the C-562-1 Reader.
See Table III for the characteristics of the paper tape readers and punches.
TABLE III: CHARACTERISTICS OF PAPER TAPE INPUT AND OUTPUT UNITS
Rate,
Char/sec

Device

Max. Length
of Tape, Feet

C-561 Paper Tape Reader:
Model C-561-1
Model C-561-2

400
600

1000
1000

C-562-1 Paper Tape Strip Reader*

650

1000

C-563-1 Paper Tape Reader

50

350

C-571-1 Paper Tape Punch

120

1000

C-572-1 Paper Tape Punch

30

1000

* No provision for rewind, backspace, or take-up .
. 24

1/68

Punched Card Input and Output
The punched card input and output equipment uses the standard Hollerith code. One or two
C-582-1 Readers can be connected to the C-517-1 Processor, via a C-581-1 Controller.
A program disc for each reader supplies the EOW signal to the processor, thus fixing
field lengths. Since alphabetic data requires two BCD digits and numeric data only one,
the program disc must be punched after every six columns in fields that may contain
either numeric or alphabetic data. This limits the numeric data fields to six digits or
requires an editing routine to reformat the fields. Consequently, two punched card readers
are required for efficient processing in any application that uses more than one card format.
An ll-punch in any column of a numeric field will cause the number to be stored in
memory as a 12-digit nine's complement. If the negative number is punched in the card
as a nine's complement, as is the case when the card was punched as output from the 500
system, and if the number is less than 12 digits, the programmed instructions must fill
out the word with nines to form the true complement.
One C-577-1 Parallel Punch or one C-576-1 Serial Card Punch can be connected as output
from the C-517-1 Processor.
The C-577 -1 Parallel Punch uses the C-551 buffer; thus, the processor can continue with
the next instruction once the data has been transferred from the memory to the buffer.
One card is punched for each punch instruction. The data must be stored in adjacent
memory cells; all digits are punched; and negative numbers are punched as nine's complements.
Since each punch instruction specifies whether the data is alphanumeric or numeric, mixed
data must be in the alphanumeric code. This means that the numeric data must be edited
prior to punching. Editing is facilitated by an Expand instruction that converts one word
of numeric data into two words in the alphanumeric code. Alternatively, the alphabetic
data could be punched on one card and the numeric data on the following card.
The C-576-1 Serial Card Punch is controlled by a drum card and by instructions stored
in memory. The drum card controls the punching or suppression of insignificant zeros in
specified fields, non-printing in specified fields, and the skip stop. The instructions that
control punching on the C-576-1 are: Punch Numeric, Punch Alphanumeric, Skip, and
Release.
All 80 columns of a card can be punched from adjacent memory locations with one instruction if all the data is numeric and the fields are of the same length. If all the data is in
the alphanumeric code, Dnly 48 columns can be punched with one instruction.
The Punch Numeric instruction specifies the field length; thus, insignificant zeros are not
transferred from the memory to the buffer. If the field length varies from field to field,
several Punch Numeric instructions are required to punch one card, with one instruction
required for each change in length of adjacent fields. For example, if a record has fields
of 12, 12, 10, 10, 5, 5, 6, 10, 4, and 6 digits, seven Punch Numeric instructions are
required to punch the card. On the other hand, if the fields are rearranged into the format
12, 12, 10, 10, 10, 5, 5, 6, 6, and 4, five instructions are needed to punch the card.
When all 80 columns are not punched in the card, a skip or a release instruction must be
used to move the card from the punching station. Skip instructions are controlled by a
12-punch in the drum ca:l.'d; when the C-576-1 receives a Skip instruction, it skips all
succeeding columns to the column preceded by a 12-punch in the drum card.
Table IV summarizes the characteristics of the punched card equipment.

A

(Contd. )

AUERBACH
1<366-1
*366-2
*366-3
*3R6-4
:l97 Series
*317-1
*317-2
*391

Record File
Disc File
Disc File
Disc File
Disc File
Disc File
Disc File
Disc File
Disc File and Control
Disc File and Control
Disc File and Control
Disc File and Control
Disc File Controls (any model)
Record File Control
Record File Control
Record File Mode Control

Magnetic Card storage
348R-1
3488-2

Random Access Computer Equipment
File Expansion Assembly

INPl'TOl'TPl'T

Magnetic Tape
381
381-3
381-4
382-:l
130
:l82-4

DO
382-6
130
581
582
*681
3484
3485
:l18
83
31!)
83

2/69

Hi-Data Tape Group (6 tapes)
Hi-Data Tape Group (3 tapes)
Hi-Data Tape Group (4 tapes)
Hi-Data Tape Group (3 tapes)
382 Simo Feature
Hi-Data Tape Group (4 tapes)
382 Simo Feature
Hi-Data Tape Group (6 tapes)
382 Simo Feature
Tape Station
Tape Station
Tape Station
Tape station
Tape Station
Hi-Data Tape Group Control
CMC Feature
Hi-Data Tape Group Control
CMC Feature

A•

AUEPBACH

701 :221.101

PRICE DATA

IDENTITY OF UNIT

CLASS

INPllTOUTPUT
(Contd. )

Model
Number

Feature
Number

Name

PRICES
y

Reita!

MOnthly

iPurChase
$

Malnt.

$

Magnetic Tape (Contd.)
399-1
189
:l99-2
189
347

Hi-Data Tape Group Control
CMC Feature
Hi-Data Tape Group Control
CMC Feature
Tape Switching Units

560
36

560
36
265

25,300
1,600
25,300
1,600
11,900

51.00

15,850
23,000

51.50

3.25

51.00
3.25

24.50

Punched Card
*323
*329-1
*329-2
400
330
78
79
80
*334
*336
*313-1
*314-lR
*315
*358-1
1A8
245
369-1
54
70
86

Card Reader
Card Reader
Card Reader
51-Column Read Feature
Card Reader/Punch
Early Card Read
51-Column Read Feed
Punch Feed Read
Card Punch
Card Punch
Card Punch Control
Card Reader Control
Card Punch Control
Card Reader Control
51-Column Read Feature
CMC Feature
Card Reader/Punch Control
Punch Feed Read
51-Column Read Feed
CMC Feature

355

440
710
36
550
10
50
25
204
484
410
133
280
204
15
52

29,800
1,500
30,000
215
3,175
935
8,900
21,100
19,800
6,900
13,750

91.50
119.25
5.75
120.00

-31.75

98
41
52

2,300
30,200
4,400
1,800
2,300

4.00
29.00
70.00
41.00
13.00
25.50
19.00
1.00
5.25
54.00
9.00
3.50
5.25

173
108
82
25
52
98
170
355
52
82
159
82

7,800
4,400
3,360
1,100
2,100
4,000
7,600
14,500
2,100
3,600
7,150
3,360

29.50
19.75
14.50
3.50
9.25
16.50
28.00
44.50
9.25
10.25
28.00
14.50

25
82
675
175
46
123
31
(2)
31
15
(1)
142
57
46
111
153

1,100
3,400
28,500
7,300
1,700
5,900
1,400
(2)
1,400
700
(1)
6,900
2,600
2,100
4,500
6,200

3.50
14.00
194.00
49.25
12.75
11.75
3,00
(2)
3.00
1.00
(1)
13.00
5.50
4.00
22.00
28.50

590

9,950
690

Paper Tape
321
24
25
95
97
137
152
322
96
103
331
25
95
136
332
63
104
311
64
71
132
1:~5

259
312
105
106
325
326

co

Paper Tape Reader/Punch
5- and 7-Level Tape Read Feature
5- and 7-Level Tape Punch Feature
Advanced Sprocket 6-Level Punch
Advanced Sprocket 6-Level Read
Extended Cable Feature
5-7-8 Level Read-Punch Feature
Paper Tape Reader
Advanced Sprocket 6-Level Read
Extended Cable Feature
Paper Tape Punch
5- and 7-Level Tape Punch Feature
(Also requires Feature 71 in Model
311)
Advanced Sprocket 6-Level Punch
Extended Cable Feature
Paper Tape Punch
Advanced Sprocket 6-Level Punch
Extended Cable Feature
Paper Tape Reader/Punch Control
CMC Feature
Gapless Paper Tape Feature
5-7-8 Level Read-Punch Feature
Extended Cable Feature
Allotter Control Feature
Paper Tape Reader/Punch Control
Extended Cable Feature
Extended Cable Feature
Tapewriter
Tapewriter-Verifier

1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

RCA 301

701:22U04

PRICES

IDENTITY OF UNIT
CLASS

INPllTOUTPUT
(Contct. )

Model
Number

Feature
Number

Name

!Monthly
Monthly
Rental Purchase Maint.
$
$
$

Printers
:1:1:1

39
42
;l:15
39
42
338
:n6-1
85
316-2
85
308
:145-1
119
345-2
119
346-1
119
346-2
119
396-1
294
396-2
294

On- Line Printer
Variable 6 or 8 lines/inch Spacing Feature
Variable 6 or 10 lines/inch Spacing
Feature
On-Line Printer
Variable 6 or 8 lines/inch Spacing Feature
Variable 6 or 10 lines/inch Spacing
Feature
Monitor Printer
On-Line Printer Control
CMC Feature
On-Line Printer Control
CMC Feature
Monitor Printer Control
Printer Buffer and Control
CMC Feature
Printer Buffer and Control
CMC Feature
Printer Buffer and Control
CMC Feature
Printer Buffer and Control
CMC Feature
On-Line Printer Control
CMC Feature
On-Line Printer Control
CMC Feature

715
36
36

32,200
1,480
1,480

149.00
7.25
7.25

1,140
36
36

51,500
1,480
1,480

234.00
7.25
7.25

194
153
36
345
36
168
660
36
660
36
740
36
740
36
260
36
590
36

7,300
7,850
1,600
15,500
1,600
8,100
29,900
1,600
29,900
1,600
33,400
1,600
33,400
1,600
13,400
1,600
26,500
1,600

30.75
14.25
3.25
29.50
3.25
15,75
58.50
3.25
58.50
3.25
69.50
3.25
69.50
3.25
24.00
3.25
53.00
3.25

1,630
3,260

85,000
170,000

270.00
544.00

249
239

9,960
11,500

36,00
22.25

325
615

15,700
28,000

28.50
58.00

3,555
113
113
385
385

145,900
4,700
4,700
15,900
15,900

5S8.00
18.50
18.50
lS.50
lS.50

340
340
360

15,200
15,200
16,200

31.00
31.00
32.25

515
565
615
670
565
615

23,000
25,300
27,600
29,900
25,300
27,600

47.00
50.25
53,50
61.50
50.25
53.50

Listers
Multiple Tape Lister and Control
Multiple Tape Lister and Control

340-6
340-12

Typewriter
328
398-1

Inten ugating Typewriter
Interrogating Typewriter Control
Magnetic Ink Character Reader

371
373

MICR Sorter/Reader Control
MICR Mode Control
Optical Character Reader

*5820
102-10
102-11
102-15
102-16
C Ol\I:\ll'KICATIO;-';S

Controls
376-11
376-12
376-34
378
378-21
378-41
378-61
37S-81
378-22
378-42

2/6'3

Video Scan Document Reader
Mark Read-IO Row-Vertical
Mark Read-IO Row-Slant
Multiple Mark Read-IO Row-Vertical
Multiple Mark Read-l0 Row-Slant

Communications Control
Communications Control
Communications Control
Communications Mode Controls: (3)
Single Scan (20-line)
Single Scan (40-line)
Single Scan (60-line)
Single Scan (SO-line)
Dual Scan (20-Une)
Dual Scan (40-line)

A.

AUERBACH

701 :221.105

PRICE DATA

IDENTITY OF UNIT
CLASS

COMMUNICATIONS
(('ontd. )

Model
Number

Feature
Number

Name

PRICES
Monthly
IMonthly
Rental Purchase Maint.
$
$
$

Controls (Contd.)
378-62
378-82
144
250

Dual Scan (60-Hne)
Dual Scan (SO-line)
Features for Dual Scan:
Unshift Feature
NNNN Terminate Fea1l1re

670
720

29,900
32,200

(;1.50
64.75

20
77

900
3,450

1.25
(;.75

NOTES:

*

No longer in production.

(1)

Furnished on request, when required, at no extra charge.

(2)

Furnished on request with Feature 25, at no extra charge.

(3)

For prices of the various 6000 series communications buffers see the RCA 3301 Price List.

© 1969 AUE RBACH Corporation and AUERBACH Info, Inc.

2/69

703:011.100
RCA 3301
Introduction

I NTRODUCTI ON

Ii 011.

The RCA 3301 REALCOM is a medium-scale general purpose computing system. It
can be used as a data processor, as a real-time processor, or as a switching center in a message
switching system, depending upon the equipment complement selected. Hardware and software
facilities are being provided that enable these functions to be combined as needed, to allow for
more economic operations. This flexibility will be particularly advantageous when functional
processing requirements (such as real-time operations) are being phased in or phased out.
Monthly rentals for the RCA 3301, as a conventional data processor, range from about
$11,000 to $30,000 per month, with a median rental of about $15,000. When real-time or communications facilities are added, the minimum system rental is about $14,000 per month and the
median rental is around $20,000. Initial customer deliveries were made in July, 1964.
As a data processor, the 3301 has adequate input-output control capabilities to serve
a complement of peripheral devices chosen from among the following:
•

1 or 2 high-speed print~rs, rated at 800 or 1,000 lines per minute,
depending upon the size of the character set used.

•

lor 2 80-column card readers, rated at 900 or 1,470 cards per
minute.

•

1 or 2 card punches, rated at 300 cards per minute.

•

1 or 2 paper tape readers, rated at 100 or 1,000 characters
per second.

•

lor 2 paper tape punches, rated at 100 characters per
second.

•

Up to 24 magnetic tape stations, described on the next page.

•

Up to 8 Model 3488 Random Access Computer Equipment units,
each with a maximum capacity of 681 million characters and
an average access time of about 300 milliseconds.

•

1 Model 3465 Data Drum Memory, with a maximum capacity of
2. 6 million characters and an average access time of 8. 6
milliseconds.

These peripheral devices are serviced by two (or at most three) data channels that provide for time-sharing of High Speed Memory (the main core storage). Except for the printers and
card punches, which are buffered, each of these units monopolizes a data channel throughout an
input or output operation.
In addition, the RCA 3301 has available hardware and software capabilities to accept
and transmit information via up to 160 telegraph or telephone lines. It is expected that these
facilities will be used to serve real-time processing requirements. while most of the peripheral
units will remain available for conventional batched processing.
The CMC (Communications Mode Control) connects the RCA 3301 system to these communications lines, scanning and servicing them as often as required. Two models are available:
the Single Scan CMC, which scans all lines with equal frequency; and the Dual Scan CMC, in
which some of the lines are 'scanned more frequently than the others. Internally, the CMC transmits the data from each line to a separate 100-character block in High Speed Memory, called a
"line slot".
Periodic peaks of activity occur in most real-time applications, and to satisfy them a
certain volume of processing power must be instantly available. Since the peak loads are so much
higher than the normal usage, it is often impossible to justify the exclusive use of the full equipment complement by the real-time process. In such cases, a system that can process a normal
data proceSSing installation workload, can be interrupted with small cost, and can operate both
real-time programs and "production programs" (RCA's term) simultaneously is highly desirable.

© 1964 Auerbach Corporation and Info, Inc.

8/64

103:011.101

RCA 3301

iii 011.

The RCA 3301, following these ideas, incorporates two levels of interrupts (real-time
interrupts being separated from all others, such as normal I/O terminations); two sets of operating
registers (held in the 214-nanosecond Micro Magnetic "scratch-pad" or control memory); and a
single operating system which handles either real-time programs or production programs, or
both types concurrently. The difficulties that arise (mainly in arranging smooth change-overs
between programs) have been simplified by making this operating system solely responsible for
handling all functions involved in these change-overs.
In addition to the real-time remote networks, the RCA 3301 can be connected to:
•
lor 2 other adjacent RCA 301 or 3301 computers (by means of
the Data Exchange Control).
•
lor 2 300 or 5,100 character-per-second lines via Dataphone
or equivalent equipment (by means of the Communications
Control).
Either of two central processors can be used in an RCA 3301 system: the basic Model
3303 or the Model 3304, which features a parallel adder for 8-character fields. The 3301 system
and both its central processors are developments of, and largely upward-compatible with, the
RCA 301 system and its character-oriented central processor (see Section 701:051). In the basic
Model 3303 Processor, the serial character-by-character processing is continued as the normal
mode of operation, so that each comparison or arithmetic operation takes a fixtld number of memory cycles for each character. A decade (10-character field) transfer instruction is included to
facilitate data manipulation operations. In the more powerful Model 3304 Processor, a fixed,
10-character word format is generally adopted in order to take full advantage of the high speed
parallel adder.
The basic Model 3303 Processor has an approximate capacity of 22,000 two-address
arithmetic instructions per second. This capacity can be increased to approximately 100,000
instructions per second by use of the High Speed Arithmetic Unit in the Model 3304 Processor.
Floating point arithmetic operations are handled automatically in the Model 3304
Processor. No simulator of the Model 3304 is presently available for the Model 3303 Processor.
The basic 40,000 character pOSitions of High Speed Memory can be expanded to a maximum of 160,000 characters in increments of 20,000 characters. One, two, or ten characters car
be read from or written into High Speed Memory during each 1.5 or 1. 9 microsecond cycle, depending upon the specific function being performed.
Three index registers and indirect addressing are available in both processors. Tally
facilities (used for loop control in the RCA 303 Processor, which has no index registers) are retained in the 3301 system, so an unusually wide choice of control methods is available.
The editing and code translation functions both make effective use of the character
orientation of the processors and yield very favorable times for the standard performance measures
that use them. Code translation is particularly efficient where various five- or six-level codes are
in use; codes with more than six data levels cannot be directly handled by the code translation
instruction.
A "time-of-day" clock that reads to one second on a 24-hour basis is available. While
this clock does not appear to be fully adequate for recording the time used by the real-time programs
for cost accounting purposes, it is very useful for general purposes.
Regarding the peripheral units available for RCA 3301 systems, the major innovation is
the introduction of magnetic tape units which are tape-compatible with the IBM 729 series and with
equivalent units produced by other manufacturers. The Model 3485 provides peak data transfer
rates of 30,000, 83,400, or 120,000 characters per second, and the Model 3487 (with a tape speed
of 75 inches per second compared to 150 inches per second for the 3485) has peak rates of 15,000,
41,700 and 60,000 characters per second.
Three other magnetic tape stations are available. The Model 581, 582, and 681 Tape
Stations were used in earlier RCA data proceSSing systems and have peak data transfer rates of
33,333, 66,667, and 120,000 characters per second, respectively. Each of these three models
has its own recording system, density, and gap length; only Models 582 and 681 can produce
mutually compatible tapes, through the use of special recording modes.
The other peripheral units are mainly products designed or manufactured by outside
suppliers to RCA's specifications. Two come from Europe - the ICT 900-card-per-minute card
reader and the Bull 300-card-per-minute punch. Others are the RCA adaptations of the Allelex
printer and the Uptime card reader. RCA-manufactured paper tape equipment is available for
five- through eight-channel tapes, and for the advanced-sprocket tapes found in some applications.
No optical or magnetic character readers are currently offered.
8/64

703:011.102

INTRODUCTION
§

011.

The software for the RCA 3301, when used solely as a data processor, is organized
exactly as if it were to be used as a combined data processing, real-time processing, and message
switching system. There is only one comprehensive operating system, and individual installations
(or occasions) use only those parts of the system which are applicable.
The needs of the full system are naturally complex, anal'these needs have been met by
the introduction of a new concept of writing programs. The writing of the actual coding for different parts of programs has been separated from the interconnections between them, and the
control of all input and output functions has been placed solely in the operating system.
In this new method, all coding, in the form of separate routines, is assembled and
placed on tape. Input-output instructions in the form of macros are used in the routines. A
series of "task descriptions" is prepared after assembly, which lays out the logical relationships
between these routines (which together comprise all of the coding). When a program is executed,
this complete subdivision of the program into logical units is used by the operating system to allocate the available storage space in the most advantageous way, conSidering the other tasks that
are running in the system at the same time.
Under this system, several programs (tasks) can be independently run, with each task
receiving storage space and processing ability according to the possibilities of the moment. It
makes no difference (except in the allocation of priorities) whether the particular task is a realtime or batch process.
Three properties of this operating system are of particular interest:
(1)

It appears to be practical. Using the special hardware facilities,

the change-over from one program to another is expected to take
between O. 1 and 1. 0 millisecond, which is relatively fast.
(2)

It appears to reduce the need for reprogramming due to changing
circumstances. If a processing method is to be used which differs

from the one originally implemented, then rewriting of the "task
descriptions" is usually all that will be required. Changing over
to real-time operations, for permanent or experimental purposes,
would likewise require no more than reforming of the task
descriptions.
(3)

It appears to allow economical interruptions of normal batch
processing to handle priority work.

The special software used for real-time and communication functions (scheduling,
message compilation, etc.) is incorporated into the operating system, together with routine
functions such as checking for errors.
No specialized functions, such as separate accounting or totally reliable inter-program
protection, are included in the operating system. The clock, which works in units of one
second, and the lack of stopper registers which positively prevent one program from overwriting
another are hardware factors which would make it difficult to include such functions effectively.
The current pricing structure, which is based on continuous full use of the equipment, does not
reflect the potential use of the system on a demand basis (e. g., to handle infrequent real-time
requests outside the normal business hours).
The question of compatibility between the RCA 3301 and its earlier, less powerful
predecessor, the RCA 30 I, has two important facets:
(1)

The operating programs of an RCA 301 user can be run on a 3301 in an
interpretive mode. This means, however, that the greatly improved
input-output facilities of the 3301 will not normally be employed. A
number of specific hardware configurations are not directly compatible,
but most 301 configurations which are in the field can be simulated in
this manner. In particular, there is no compatibility between the
301 Scientific Processor and any RCA 3301 system.

(2)

RCA 301 programs and programming systems are being used to back
up the 3301 system. These include the 301 FORmAN II and COBOL-61
compilers, which run under the 301 compatibility program for b»th
compilation and execution. In this mode of operation, the compiler
user may have to tolerate considerable inefficiencies in his object
program input-output.

@1 964 Auerbach Corparatian and Info, Inc.

8/64

703:011.103
II 011.

The execution of FORTRAN object programs in the 301 compatibility mode will be
especially inefficient; the RCA FORTRAN program which runs on the 301 system includes a
simulation of the RCA Scientific 304 Processor, so that at some points a double level simulation
may be in operation.
Software that has been announced for the RCA 3301 includes the REALCOM Assembly
System, a report program generator, sort/merge programs for magnetic tapes and for the
Model 3488 Random Access Computer Equipment, and a COBOL-61 Extended compiler. A
FORTRAN IV compiler has been announced for the RCA 3304 system, which has automatic
floating point instructions, but this will not be available on the basic 3301 system.
In summary, the RCA 3301 is an advanced, powerful, and flexible system that can effectively handle the problems of real-time processing concurrently with those of normal batched
processing. A top limit for message switching appears to be in the region of 1,000 40-character
messages per second, and this function would use up approximately 25 per cent of the system's
computational capacity. Similarly, processing 3 real-time transactions per second, including
a reference to a file held in the Random Access Computer Equipment, would probably demand the
use of one of the two (or at most three) data channels and some 5 to 10 per cent of the computational capacity. This is probably well beyond most actual anticipated loadings. However, a
number of difficult problems have been solved (notably the ability to make use in other programs
of the time spent in waiting for access to mass-storage and magnetic tape devices), so that such
operational magnitudes can be seriously contemplated.

8/64

RCA 3301

703:221.101

~ "......
~EDP
.....~
....."

RCA 3301
PRICE DATA

~

RCA 3301
PRICES

IDENTITY OF UNIT

CLASS

Model
Number

Feature
Number

Monthly
Monthly
Rental Purchase Maint.

Name

$

$

5,150
(1)
(1)

250,000
(1)
(1)

300.00
(1)
(1)

6,570
310
(1)
(1)
415
6
103

320,000
15,600
(1)
20,000
250
4,600

386.00
19.00
(1)
(1)
27.00
0.50
9.50

to

1,030

50,000

60.00

to

1,545

75,000

90.00

to

2,165

105,000

128.00

to

2,780

135,000

161.00

to

3,400

165,000

199.00

to

4,120

200,000

240.00

310
57
88
175

15,600
2,750
3,900
7,800

19.00
7.00
8.50
16.00

415
465
815
875
815
875
645

24,000
27,000
39,500
42,500
39,500
42,500
32,500

35.00
37.50
64.50
67.50
64.50
67.50
51. 25

1,080
1,405
2,055
2,625
2,915
3,490
325

49,800
64,600
94,200
120,300
133,500
159,900
19,900

205.00
259.00
367.00
460.00
509.00
603.00
26.00

$
Processing Unit (includes core storage

PROCESSOR

3303
422
584-1
3304
164
422
584-1
3313-2
6793
6080

Processor (40,000 characters)
NNNN Terminal Feature
System Automatic Recovery
Processor (40,000 characters: high-speed
arithmetic unit)
Simultaneous Mode #3
NNNN Terminal Feature
System Automatic Recovery
Supplemental Power Supply
Power Supply
Power Supply

{I)

Main Storage

3361-2
3361-3
3361-4
3361-5
3361-6
3361-7
ATTACH-

High-Speed Storage (expands
60,000 characters)
High-Speed Storage (expands
80,000 characters)
High-Speed Storage (expands
100,000 characters)
High-Speed Storage (expands
120,000 characters)
High-Speed Storage (expands
140,000 characters)
High-Speed Storage (expands
160,000 characters)
Attachments

MENTS.

ADAPTERS,
AND
CHANNELS

164
3416
3446-1
3446-2

Simultaneous Mode #3
Digital Clock
Peripheral Switching Unit
Per~pheral Switching Unit
Channels

3383-6
3383-12
3384-6
3384-12
3385-6
3385-12
3388-4

Dual Tape
Dual Tape
Dual Tape
Dual Tape
Dual Tape
Dual Tape
3488 Channel

Channel
Channel
Channel
Channel
Channel
Channel

Drum Storage

MASS

STORAGE

3464-1
3464-2
3464-3
3464-4
3464-5
3464-6
3364

Data
Data
Data
Data
Data
Data
Data

Drum
Drum
Drum
Drum
Drum
Drum
Drum

Memory
Memory
Memory
Memory
Memory
Memory
Control

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

RCA 3301

703:221.102

IDENTITY OF UNIT

, CLASS

Model
Number

MASS
STORAGE
(Contd. )

Feature
Number

Name

PRICES
Monthly
[Monthly
Rental Purchase Maint.

$

$

$

Magnetic Card Storage
3488-1
3488-2
581
582
*681
3484
3485

Random Access Computer Equipment
File Expansion Assembly
Tape Station
Tape Station
Tape Station
Tape Station
Tape Station

INPUTOUTPUT

3,000
1,625
865
950
565
800

127,500
65,000
26,400
36,750
39,850
26,400
37,200

440
710
515
490
206
1,370
1,370

23,000
29,800
24,000
23,750
10,000
57,550
57,550

91. 50
119.25
105.00
38.75
16.00
159.00
159.00

Reader
Punch
Reader-Punch
Control

355
159
173
196

14,500
7,150
7,BOO
9,500

44.50
2B.OO
29.50
15.50

00-Line Printer
00-Line Printer
Printer Buffer and Control
Printer Buffer and Control

715
1,140
490
805

32,200
51,500
23,750
39,780

149.00
234.00
38.75
64.00

3,555
113
113
385
385
129
129
410
410
635

145,900
4,700
4,700
15,900
15,900
5,300
5,300
17,000
17,000
26,400

588.00
18.50
18.50
62 50
62.50
21.50
21.50
66.00
66.00
106.00

283
283
283
258
258
258
46
46
46

11,600
11,600
11,600
10,600
10,600
10,600
1,900
1,900
1,900

46.50
46.50
46.50
43.00
43.00
43.00
7.25
7.25
7.25

415
415
415

20,000
20,000
20,000

35.00
35.00
35.00

544

---

Punched Card
329-1
329-2
*3436
*3336
3329
3330-1
3330-2

Card
Card
Card
Card
Card
Card
Card

Reader
Reader
Punch
Punch Buffer and Control
Reader Control
Reader/Punch Buffer and Control
Reader/Punch Buffer and Control

528.00
314.00
163.00
163.00
171. 00
114.00
164.00

Paper Tape
322
331
321
3321

Paper
Paper
Paper
Paper

Tape
Tape
Tape
Tape

Printer

---

333
335
3333
3335

Optical Reader
5820
102-10
102-11
102-15
102-16
102-20
102-21
102-25
102-26
102-259

Videoscan Document Reader(l)
Mark Read-10 Row-Vertical
Mark Read-l0 Row-Slant
Multiple Mark Read-l0 Row-Vertical
Multiple Mark Read-l0 Row-Slant
Mark Read-12 Row-Vertical
Mark Read-12 Row-Slant
Multiple Mark Read-12 Row-Vertical
Multiple Mark Read-12 Row-Slant
Mark Read Feature

0

Display Units
*6050-11
*6050-12
*6050-13
*6050-21
*6050-22
*6050-23
*6051-1
*6051-2
*6051-3
COMMUNICATIONS

Data
Data
Data
Data
Data
Data
Data
Data
Data

Terminal
Terminal
Terminal
Terminal
Terminal
Terminal
Interrogator
Interrogator
Interrogator

Buffers and Controls
3376-11
3376-12
3376-34

2/69 .

Video
Video
Video
Video
Video
Video
Video
Video
Video

Communications Control
Communications Control
Communications Control

A

AUERBACH
,.

~ICE

DATA

703:221.103

IDENTITY OF UNIT

CLASS

COMMUNICATIONS
(Contd. )

Model
Number

Feature
Number

PRICES

Monthly
Rental iPurchase Maint.
$
$
$

IMonthly

Name
Buffers and Controls (Contd.)

3378
3378-21
33'18-41
3378-61
3378-81
3378-101
3378-121
3378-141
3378-161
3378-22
3378-42
3378-62
3378-82
3378-102
3378-122
3378-142
3378-162
6071
6072
249
493
584-2
6002-11
6002-12
6002-21
6003
6013
6015
6016
3377
6009-1
6009-2
6009-3
6010-21
6010-22
6010-23
6011-10
6011-11
6011-12
6012-11
6012-12
6012-21
6012-22
6020-11
6020-12
6020-13
6020-14
6020-15
6092
6025-10
6025-15
6025-20
6025-25
6027
6041
6042

Communications Mode Controls:
Single Scan (20 line)
Single Scan (40 line)
Single Scan (60 line)
Single Scan (80 line)
Single Scan (100 line)
Single Scan (120 line)
Single Scan (140 line)
Single Scan (160 line)
Dual Scan (20 line)
Dual Scan (40 line)
Dual Scan (60 line)
Dual Scan (80 line)
Dual Scan (100 line)
Dual Scan (120 line)
Dual Scan (140 line)
Dual Scan (160 line)
Communications Rack
Buffer Frame
Features for Model 3378 Series:
NNNN Terminate Feature
Expanded Character Recognition
System Automatic Recovery
Buffers for 3378 Series:
Telegraph Buffer
Telegraph Buffer
Telegraph Buffer
Telegraph Buffer
Telegraph Buffer
Telegraph Buffer
Parallel Buffer
Data Exchange Control
Communication Buffer
Communication Buffer
Communication Buffer
Communication Buffer
Communication Buffer
Communication Buffer
Communication Buffer
Communication Buffer
Communication Buffer
Communication Buffer
Communication Buffer
Communication Buffer
Communication Buffer
Communication Buffer
Communication Buffer
Communication Buffer
Communication Buffer
Communication Buffer
Communication Buffer Switch
Buffer Interface Unit
Buffer Interface Unit
Buffer Interface Unit
Buffer Interface Unit
Line Termination Assembly
Time Generator
Code Translator

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

445
465
485
505
535
555
575
595
470
485
500
515
555
570
585
605
62
31

21,500
22,500
23,500
24,500
26,000
27,000
28,000
29,000
22,750
23,500
24,250
25,000
27,000
27,750
28,500
29,250
2,800
1,400

82
82
(1)

4,000
2,500
(1)

15
20
15
82
82
82
62
400
1,133
1,133
1,133
77
118
201
196
206
227
180
180
180
180
72
129
221
155
258
77
62
72
72
82
46
82
57

700
900
700
3,650
3,650
3,650
2,800
17,900
50,900
50,900
50,900
3,400
5,300
8,900
8,500
8,800
9,900
8,100
8,100
8,100
8,100
3,200
5,800
9,900
6,900
11,500
3,500
2,800
3,200
3,200
3,700
2,100
3,700
2,500

36.50
37.50
38.50
39.50
41.00
42.00
43.00
44.00
37.75
38.50
39.25
40.00
42.00
42.75
43.50
49.25
5.75
3.00
6.00

(1)

1.00
1.25
1.00
7.00
7.00
7.00
5.75
29.50
103.50
103.50
103.50
7.00
10.50
18.75
18.25
19.00
21. 25
16.25
16.25
16.25
16.25
6.50
12.25
20.00
14.75
24.25
7.00
5.75
6.50
6.50
7.00
4.00
7.00
5.50

2/69

103:221.1o.t

RCA 3301

PRICES

IDENTITY OF UNIT

CLASS

Model
Number

COMMUNICATIONS
(Contd. )

Feature
Number

Name

Re$tal

Pur$haSE

Monthly
Maint.

$
Terminals

6220-3
6228-3
6235-3
6240-:1
*5936-1
445

*5940-1
242

565
*5941-1
242
~'.c,

565
*5942-1
403
565
6740-11
563
6740-21
563
564
6741-11
563
6741-21
563
564
6742-11
563
6742-21
563
564
6077

EDGE Data Input Station
EDGE Auxiliary Card Reader
EDGE Line Concentrator
EDGE Central Recorder
Teletypewriter (KSR)
Special Typewhee! Feature
First Typewheel
Additional Typewheels
Addtiional Keycaps
Teletypewriter (ASR)
Special Typewheel!Keycap Feature
First Typewheel
Additional Typewheels
Keycap
Data Set Coupler
Teletypewriter (KSR)
Special 'Typewheel/Keycap Feature
First'Typewheel
Additional Typewheels
Keycap
Data Set Coupler
Teletypewriter (RO)
Special 'Typewheel/Keycap Feature
First Typewheel
Additional Typewheels
Data Set Coupler
Teletypewriter (ASR, Friction Feed)
Data Set Coupler
Teletypewriter (ASR, Sprocket Feed)
Data Set Coupler
Wide Carriage
Teletypewriter (KSR, Friction Feed)
Data Set Coupler
Teletypewriter (KSR, Sprocket Feed)
Data Set Coupler
Wide Carriage
Teletypewriter (RO, Friction Feed)
Data Set Coupler
Teletypewriter (RO, Sprocket Feed)
Data Set Coupler
Wide Carriage
Interrogator Control Terminal

NOTES:
*No longer in production
(1)

Purnished on reQuest, when required, at no extra charge.

(2)

To be quoted on request for each specific system and location.

(3)

Single use charge.

2/69

Monthly

fA

AUERBACH

'"

69
35
215
400
72

(2)
(2)
(2)
(2)
14.50

2,600
1,500
7,440
14,800
1,500

200(3)
75(3)
0.90(3)
82

200
75
0.90
1,800

-

200(3)
75(3)
0.90(3)
8
67

200
75
0.90
350
1,300

-

200(3}
75(3)
0.90(3)
8
57

200
0.90
350
1,100

200(3)
75(3)
8
140
45
145
45
45
90
45
95
45
45
80
45
85
45
45
1,195

200
75
350
5,650
350
5,850
350
250
3,650
350
3,850
350
250
3,250
350
3,450
350
250
48,700

'/5

16.50

1.75
12.75

-- 1.75
12.00

-

- 1. 75
25.25
1. 50
26.00
1. 50
1. 00
16.25
1. 50
17.00
1.50
1. 00
14.50
1. 50
15.25
1. 50
1. 00
197.00

-£. "'.....

710:000.001
RCA SPECTRA 70
REPORT UPDATE

~EDP

-

AUE~

•

IEPlI"

REPORT UPDATE
~

RCA ANNOUNCES NEW DISC SYSTEM
RCA is going to be manufacturing its own magnetic disc memory units. The 70/590-8 Direct
Access storage Systems represents storage as it will be available to the user only as a pack
consisting of the Direct-Access Controller and nine Direct Access Disc Units, one of which is
a spare. Each Drive Unit holds more than 29 milllon bytes, a total of over 233 million bytes for
the eight on-line units.
Each Unit uses the RCA-built 70/593 disc pack which is interchangeable with the 70/590 or mM
2314 units. The mM 2316 can be used in place of the 70/593. Files employ1'ng Index Sequential
are not interchangeable, however, ~tween 70/590 and mM 2314.
' .
The 70/590-8 system is offered in a "one-by" configuration. This allows one selector channel
to: Read, Write, or Search on any ~ of the eight on-line units. The multichannel Switch
Feature, 5518-2, permits either of two selector Channels to share in operating the basic "oneby" configuration. The Selector Channels can be on the same or different processors.
Software support is provided at the physical and logical levels for the Tape-Disc Operating
System (TDOS) and at the logicallev:el for the Time Sharing Operating System (TSOS). This
includes systems residency, Indexed Sequential, and Disc Sort.
The characteristics of the 70/590-8 system can be summarized as follows:
Capacity
Track-7, 294 bytes (one record per track)
Cylinder-20 tracks or 145,88 bytes
Unit-200 cylinders or 29,176,000 bytes
System-8 units on-line, 233,408,000 bytes
Access time
Minimum-25 ms (one position more)
Average-75 ms
Maximum-l35 ms
Rotational latency
Minimum-O.O ms
Average-12.5 ms
Maximum-25.0 ms.
The rental costs of the 70/590-8 system will be $5410 per month for the system with an
additional $144 per month for the Multichannel Switch Feature, 5519-2. Purchase prices will
be $249,4000 for the 70/590-8 and $5,770 for the 5519-2. Monthly maintenance charges have
been announced as $615 per month and $11. 50 per month respectively. RCA includes maintenance charges in its monthly rental figures. Initial deliveries are scheduled for December 1969.

'4f,!>

,;~::;:

..>:-.:: '

~

";'.;,,,:

.

~'>'

710:000.002

.l. "...."
/4.'-

AUERBAC~

EDP

RCA SPECTRA 70
REPORT UPDATE

.....-_-----1
R[PORTS

~

REPORT UPDATE
• RCA ANNOUNCES 70/61 TIME-SHARING SYSTEM
RCA announced the 70/61 time-sharing system on September 8, 1969. The 70/61 is intended to
compete in the large-scale time-sharing market, although the facilities for background batch processing concurrent with time sharing will also be emphasized. The new computer bears the
same relationship to the 70/60 as the earlier 70/46 system has to the 70/45.
The 70/61 has a main memory of from 524,288 to 1,048,576 bytes, with a cycle time of 765 nanoseconds for a 4-byte access. The instruction execution rate will be more than three times that
of the 70/46. Up to 128 concurrent time-sharing users can be supported. Among other improvements in the 70/61 over the 70/46 are a larger virtual memory storage capacity (up to 3200 pages)
and enhanced failure-tolerant and diagnostic facilities, including the use of switchable main memory banks for fail-soft operation.
The 70/61 is fully upward compatible with the 70/46. User tasks written for the 70/46 can be
transferred to the 70/61 without change provided no shared code is used. 70/46 tasks using
shared code will need recompilation before they can be used on the 70/61 since the details of the
implementation of shared code are different in the two systems; the 70/61 implementation of
shared code removes the restriction on the amount of shared code, which is inherent in the 70/46
implementation. RCA is developing the OS61 operating system for use with the 70/61. OS61 is
an enlarged and enhanced version of the Time-Sharing Operating System (TSOS) used with the
70/46, and will contain all the facilities of TSOS as a subset while taking advantage of the improved hardware features of the 70/61.
Detailed price information for the 70/61 will be available shortly - a typical system including
524,288 byte of main storage, 233 million bytes of Direct Access Storage, 2 Drum Memories
holding 800 pages each, and line buffers for 64 remote terminals rents for about $51,000 per
month. The first customer deliveries of the Spectra 70/61 and OS61 are scheduled for the
first quarter of 1971.
• RCA ANNOUNCES 'UNBUNDLING' OF RENTAL AND MAINTENANCE PRICES
A new price schedule for RCA computer equipment became effective on September 1, 1969. The
standard monthly rental prices now quoted by RCA do not include maintenance. While the purchase
prices for the equipment are unchanged, the new price schedules indicate that, when the rental
and maintenance prices are combined, most of the equipment will be slightly more expensive to
rent (with maintenance by RCA) than under the previous schedule of rental prices which included
maintenance. The rental price information in the published Reports will be updated to reflect the
new pricing as soon as complete information on the new schedules is available. Maintenance
prices for 70/60 Processors are also included for the first time and range from $585 per month
for a 70/60-F Processor with 131, 072 bytes of main storage to $1,770 per month for a 70/60-N
Processor with 1,048,576 bytes of storage.
• NEW PRINTER FOR SPECTRA 70 COMPUTERS
RCA has introduced the 70/246 Train Printer for use with Spectra 70 computer systems. This
train printer has a horizontal line of 132 characters spaced at 10 characters per inch; the vertical
line spacing is 6 or 8 lines per inch and is set by the operator. The printing speed varies with
the number of characters in the character set as shown in the table below. Printing at higher
rates than those given in the table can be sustained for short periods - for example, a printing
speed of 2400 lines per minute can be maintained for up to 5 minutes when a 16-character set is
used. The RCA 64-character set is standard, but 10 different character sets are available including FORTRAN-COBL, PL/1, Text Printing, and a Hi-Speed Alphanumeric set. Each character set contains 288 pOSitions, and up to 254 different characters can be used in special-purpose
character sets. The unit is fully compatible with the existing Spectra 70 printers.
The purchase price for the 70/246 Train Printer is $77,600. Two optional features are available - the 5262-5 Document Processing Option, which allows the interconnection of the mM 9364
Document Roll Input Unit and the mM 9361 Document Converter, and the 5278-40 Special Format
Control, allowing 1401 type format control.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

9/69

710:000.002

RCA SPECTRA 70

TABLE SHOWING SPEEDS OF RCA 70/246 TRAIN PRINTER

Character Set
Size, Characters
240
144
96
64
48
24
16

9/69

Speed,
Lines/Minute
264
470
670
860
1100
1500
1500

A .,

AUERBACH

r--..

1.
lA,

AUER8AC~

-

710:011.100

STU""
EDP

RCA SPECTRA 70
SUMMARY

REPORTS

~

SUMMARY
.1

SUMMARY
Spectra 70 is the "brand name" for RCA's third-generation family of central processors,
peripheral devices, and supporting software. Noteworthy characteristics of the Spectra 70
include:
•

The high degree of program compatibility, both upward and downward, among six
of the eight Spectra 70 processor models. Compatibility is also achieved with the
IBM System/360 processors through similar hardware design and compatible source
languages.

•

The wide range of input-output and storage devices.

•

The numerous arithmetic modes and data forms, and the resulting complexity of
machine-language coding.

•

The emphasis upon software support through several levels of integrated operating
systems.

•

The use of true monolithic integrated circuits in the 70/35 and higher-numbered
processors.

•

The availability of optional features that enable certain Spectra 70 processor models
to emulate a number of second-generation RCA and IBM computers.

Figure 1. A typical RCA Spectra 70/45 System Configuration
This Summary is divided into five independent sections, each of which describes and
(where pertinent) analyzes some particular facet of the Spectra 70 series. Each section is
independent and can be read as your needs and interests warrant. The five sections are:
.1
.2
.3
.4
.5
.2

Summary
Data Structure
Hardware
Software
Compatibility

DATA STRUCTURE
Spectra 70's data structure is identical in all respects with that of the IBM System/360.
The basic unit of data storage is the "byte," which consists of eight data bits plus (in most
system components) one parity bit. The eight data bits in a byte can represent one alphameric character, two decimal digits, or a portion of a binary field.
Bytes can be handled individually or grouped together into fields. A "halfword" is defined
as a group of two consecutive bytes, or 16 bits. A "word" in the Spectra 70 is a group of
four consecutive bytes, or 32 bits. A "double word" consists of two consecutive words,
or 64 bits. The location of any field or group of bytes is specified by the address of its
leftmost byte.
© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

9/69

710:011.200

.2

RCA SPECTRA 70

DATA STRUCTURE (Contd.)
Every fixed-length field (halfword, word, or double word) must be located in main storage
on an "integral boundary"; i. e., the storage address of the field must be a multiple of the
length of the field in bytes. This restriction is particularly important for effiCient operation of the Spectra 70/55, 70/60, and 70/61 Processors, which access up to four bytes in
parallel, and the same restriction has been applied to the smaller processors in order to
maintain compatibility. Variable-length (decimal) fields are processed serially by byte
in all models, so they may start at any byte location.
At the low end of the Spectra 70 line of processors, the 70/15 and 70/25 Processors can
perform arithmetic operations on two basic types of operands: fixed-point binary and
variable-length decimal. The larger Spectra 70 processor models can perform arithmetic
operations on four basic types of operands. In addition to fixed-point binary and variablelength decimal, these models can also perform arithmetic operations on two sizes of
floating-point binary operands. The basic arithmetic operand size is the 32-bit fixed-point
binary word. Most fixed-point instructions can alternatively specify the use of 16-bit
halfword operands.
Floating-point numbers can be represented in either a "short" (32-bit) or "long" (64-bit)
format. The fractional part occupies 24 bits in the short format and 56 bits in the long
format. The hexadecimal character occupies 7 bits in both formats and permits representation of numbers ranging from 10-78 to 1075.
Decimal arithmetic is performed upon 4-bit BCD digits packed two to a byte, with a sign
in the rightmost four bits of the low-order byte. Decimal operands can be up to 16 bytes
(31 digits and sign) in length
The 8-bit byte structure has certain basia advantages over the 6-bit data format: decimal
digits can be packed more conveniently, the standard 7 -bit ASCII code and the Extended
BCD Interchange Code can be used, and today's familiar character sets can be conveniently
expanded .

.3

HARDWARE

.31

Central Processors
Eight processor models currently form the nucleus of the Spectra 70 Series. Six of the
processor models are program-compatible for a broad range of business and scientific
applications: 70/35, 70/45, 70/46, 70/55, 70/60, and 70/61. The 70/15, with its
restricted instruction repertoire, may be of primary interest as a satellite or remote
terminal system for the larger Spectra 70 processors. The 70/25 also has a limited
instruction repertoire, but its expanded throughput capability for magnetic tape-oriented
applications makes it suitable for certain single-processor installations. The 70/25 is no
longer manufactured by RCA but is marketed as available. During 1969, the original
version of the 70/45 was replaced in the Spectra 70 product line by the 70/45 Type II. The
70/45 Type II differs from the Type I in having slightly enhanced Selector Channel capabilities and one additional instruction (Test and Set). Other minor differences include the
availability of certain sizes of main memory, change in power supply buffering, and the
availability of Memory Protect on the 70/45 Type II. All information presented here for
the Spectra 70/45 applies to both types unless indicated otherwise.
The 70/46 and 70/61 Processors are time-sharing versions of the 70/45 and 70/60 Processors, respectively.
Comparative arithmetic execution times for the various Spectra 70 processors are
illustrated in Table 1. Table II shows the various core storage capacities that can be
obtained with each of the basic processor models.
The Spectra 70 processors which are larger than the 70/25 Processor are designed to
facilitate achieving program compatibility with the IBM System/360 computers. The
remainder of this discussion of Central Processors concerns itself exclusively with the
processors associated with the Spectra 70/35 and higher-numbered systems. These processors offer the full System/360 instruction repertoire except for the "privileged" instructions, which are normally reserved for operating system use and are not permitted
in users' programs. Thus, RCA expects to achieve two-way program compatibility - to
a limited extent at the machine-language level and to a much greater extent at the assembly,
COBOL, and FORTRAN language levels.

9/69

A

(Contd. )

AUERBACH

"

SUMMARY

. 31

710:011.310

Central Processors (Contd.)
The Spectra 70 processors contain facilities for addressing main storage, for fetching and
storing information, for executing stored-program instructions in the desired order, for
arithmetic and logical processing of data, and for initiating all communication between
main storage and peripheral devices. Each program uses sixteen 32-bit general registers
and four 64-bit floating-point registers. The general registers can be used as fixed-point
accumulators or as index registers. In the 70/45 and higher-numbered processors, these
registers are contained in a scratchpad memory, whose cycle time is 300 nanoseconds per
four-byte word for the 70/45, 46, and 55 Processors. The scratchpad memory used in
the 70/60 and 70/61 Processors has an access time of 85 nanoseconds. In lieu of scratchpad memory, the Spectra 70/35 Processor provides 128 words of additional core storage
for use as general registers. Different parts of these memory units are used as the
operational registers depending upon which processor state is being used.
Instructions can be two, four, or six bytes in length. A 2-byte instruction causes no
reference to main core storage. A 4-byte instruction causes one reference to main
storage, while a 6-byte instruction causes two storage references.
Main storage addresses are formed by adding a 12-bit "displacement" contained in the instruction to a 24-bit "base address" contained in one of the 16 general registers. The
addresses in many instructions (including most binary arithmetic and logical instructions)
can be further modified by adding a 24-bit "index" contained in another general register;
this effectively provides a double indexing capability. The base-register technique of
address formation facilitates program relocation and segmentation, at the expense of increased programming complexity.
The basic arithmetic mode of these processors is fixed-point binary, using 32-bit
operands and two's-complement notation. Most instructions can alternatively specify the
use of 16-bit "halfword" operands to improve storage utilization. Most products and all
dividends are 64 bits long. Fixed-point arithmetic and comparison instructions specify
TABLE I: ARITHMETIC EXECUTION TIMES FOR THE RCA SPECTRA 70 PROCESSORS
Processor Model
70/15

70/25

70/35

70/45*

70/55

70/60*

62
#
#

23.3
#
#

51. 2
163.2
243.2

25.2
81. 9
111. 2

7.8
17.9
25.0

6.1
15.6
23.7

83
#
#

46.5
106.5
351.3

76.5
223.9
377.7

39.2
109.1
174.0

18.4
51.8
41.5

19.3
22.0
25.0

Fixed Point Binary
c=a+b
c = ab
c = alb
Fixed Point Decimal
c=a+b
c = ab
c = alb
Floating Point - Short
c=a+b
c = ab
c = alb

#
#
#

#
#
#

80.9
202.6
445.5

37.4
67.6
101. 2

13.2
23.0
28.4

7.7
14.6
20.7

#
#
#

#
#
#

115.9
536.4
1282.1

52.6
211.5
305.2

18.5
50.0
83.8

10.0
37.2
55.3

Floating Point - Long
c=a+b
c = ab
c = alb

NOTE: All times are expressed in microseconds. The fixed-point decimal times
are based on 5-digit (3-byte) decimal operands. The floating-point times
are based on both the short-form (32 bits) and the long-form (64 bits)
binary operands. The 70/15 and 70/25 do not require programmer-initiated
operand movement to a fixed accumulator register.

#
*

Facility not available,
Times for the Spectra 70/46 and Spectra 70/61 Processors are identical to
those shown for the Spectra 70/45 and Spectra 70/60 Processors, respectively.
© 1969 AUERBACH Corporation and AUERBACH Info. Inc,

9/69

RCA SPECTRA 70

710:011.311

.31

Central Processors (Contd. )
one operand in a general register and a second operand in either main storage or a
general register; these instructions are 4 bytes long when they specify an operand address
in main storage and 2 bytes long when both operands are in registers.
The System/360-compatible instruction set includes instructions which perform fixedpoint arithmetic, comparison, branching, moving, loading, storing, shifting, radix conversion, code translation, packing, unpacking, and Boolean operations. The radix conversion operations perform automatic conversions between signed, packed decimal fields
up to 15 digits in length and 32-bit signed binary integers. The code translation instruction uses a table to translate any 8-bit data code to any other 8-bit code. The packing and
unpacking instructions convert numeric BCD data between the one-character-per-byte
format used in most input-output devices and the two-digits-per-byte format used for
decimal arithmetic.
The decimal arithmetic facility provides additional instructions for addition, subtraction,
multiplication, diviSion, comparison, and editing of decimal numbers. Decimal arithmetic
is performed upon 4-bit BCD digits packed two to a byte, with a sign in the rightmost four
bits of the low-order byte. Decimal operands may be up to 16 bytes (31 digits and sign)
in length. The length of each decimal field is specified in the instruction referencing it.
Two-address (6-byte) instructions of the storage-to-storage type are used for all decimal
operations; the general and floating-point registers are not utilized.
The floating-point arithmetic facility provides additional instructions for addition, subtraction, multiplication, diviSion, loading, storing, and comparison of both "short"
(32-bit) and "long" (54-bit) floating-point numbers. Floating-point instructions specify
one operand in a floating-point register and a second operand in either main storage or
a floating-point register.
The optional Storage Protection feature can protect the contents of specified 2, 048-byte
blocks of core storage from being altered as a result of program errors or misguided
input data. This feature prevents overwriting by unauthorized programs, but it does
not guarantee privacy since any program can still read the contents of any desired
portion of core storage. 2, 048-byte blocks of core storage can be protected both from
reading and from writing with the Memory Protect feature, which is available for the
Spectra 70/45 Type II, 70/46, 70/60, and 70/61 Processors.
TABLE II: SPECTRA 70 MAIN CORE STORAGE CHARACTERISTICS
Core Storage
Capacity, Bytes
4,096
8,192
16,384
32,768
49,152
65,536
131,072
196,608
262,144
393,216
524,288
655,360
786,432
917,504
1,048,576

70/15

70/25

70/35

70/45

70/46

70/55

70/60 70/61

55-E
55-F

60-F

15-A
15-B
25-C
25-0
25-E

35-0
35-DC
35-E

45-C*
45-0*
45-E
45-F
45-FE**
45-G

46-G

55-G
55-H

60-G
60-H
60-J
60-K
60-L
60-M
60-N

61-G
61-H
61-J
61-K
61-L
61-M
61-N

Cycle Time, Ilsec

2.0

1.5

1.44

1.44

1.44

0.84

0.765 0.765

Bytes Accessed
per Cycle

1

4

2

2

2

4

4

4

Effective Cycle Time
per Byte, Ilsec

2.0

0.38

0.72

0.72

0.72

0.21

0.19

0.19

* 70/45 Type I only - no longer manufactured.'
** 70/45 Type II only.

9/69

A

(Contd. )

AUERBACH
®

SUMMARY

710:011.320

. 32

Internal Storage
Table II indicates the range of core storage sizes and speeds available with the various
Spectra 70 processor models.
Three different types of auxiliary storage devices are available in the form of magnetic
drums, discs, and cards. The storage capacity of these devices ranges from less than
0.8 million bytes for a drum unit to over 530 million bytes for the magnetic card mass
storage unit. Similarly, average access times can range from 8. 6 milliseconds to 488
milliseconds for the same two devices, respectively. The single controller used for all
three types- of storage devices allows an installation to tailor its complement of storage
devices according to specific capacity and access time requirements. Table In lists the
various Spectra 70 auxiliary storage devices with their principal functional characteristics.
None of these devices can be used with the small-scale 70/15 or 70/25 Processors.
TABLE ITI: SPECTRA 70 AUXILIARY STORAGE UNITS

Device

70/564 Disc Storage Unit*

7.25 to 58.0

70/568 Mass Storage Unit

536. 9 to 4,295

Average
Access
Time
(msec)
98
488

Data
Transfer
Rate
(bytes/sec)
156,000
70,000

70/565 Drum Memory Unit

0.8 to 3.2

8.6

210,000

70/567 Drum Memory Unit

4.1 to 16.4

8.6

333,000

*
.33

Capacity Range
(Millions of Bytes
per Control Unit)

This is RCA's designation for the mM 2311 Disk Storage Drive, which RCA is currently
marketing for use with Spectra 70 systems.
Time-Sharing Systems
RCA has entered the full-scale time-sharing market with the 70/46 and 70/61 Processors,
based on the 70/45 and 70/60 respectively. The systems include hardware paging facilities, a special drum unit, and a specialized software package.
The Spectra 70/46 Processor hardware is largely that of the Model 70/45, with several
improvements to facilitate time-shared operation. The 70/61 Processor is similarly
based on the 70/60. Significantly, the addition of hardware logic for program paging and
segmentation provides up to 2,097,152 bytes of virtual storage and assists in the dynamic
reallocation of memory. The 70/46 and 70/61 feature a fast associative memory for
translation of all virtual memory addresses to real core storage addresses. Other
improvements include changes in the interrupt logic and the use of read-only memory
to implement the address translation tables and to provide by hardware the special
instructions frequently used by the software system to manipulate the various translation tables. Program pages in 4, 096-byte blocks will be loaded either from disc or
from the drum system, which is capable of storing up to 16 (70/46) or 32 (70/61)
million bytes of data. The drum has an average access of time of 8.6 milliseconds
and a peak data transfer rate of 333,000 bytes per second.
A most significant fact is that the Spectra 70/46 and Spectra 70/61 are designed primarily for batch processing in a multiprogramming, random-access environment, with
the facility for time-sliced, remote time-sharing operations expected to consume less
than 20 per cent of the system's capacity. For this reason the specialized Spectra 70/46
Time Sharing Operating System (TSOS) provides not only dynamic interactive processing
capabilities, but also efficient multiprogramming capabilities for up to four concurrent
background jobs. TSOS is 100 per cent compatible with the Spectra 70 Tape Operating
System (TOS), permitting the user to progress easily from the tape-oriented system to
the more powerful TSOS random-access-oriented operating system. An enhanced version
of TSOS, 0861, will be available for Spectra 70/61.
All source program written for Spectra 70/35, 70/45, 70/55, and 70/60 systems will
operate on the 70/46 and 70/61 systems. Most object programs written for the nontime-sharing processors will also run on the time-sharing processors when the timesharing processors are run in the Direct (1. e., non-translate) mode in which hardware
paging is not used. Programs intended for use on more than one processor should,
however, be timing-independent and avoid .he use of hardware features peculiar to one
machine. Because of this intra-family c_ompatibility, software systems such as
© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

9/69

RCA SPECTRA 70

710:011.330

. 33

Time Sharing Systems (Contd.)
FORTRAN, COBOL, Assembler, Report Generator, and Sort/Merge, which are
originally developed for use with the Tape Operating System (TOS), can also be used
with the time-sharing processors.
The 70/61 has larger main memory and virtual memory storage capacities than the
70/46 and handles up to 128 remote users concurrently compared with 48 for the 70/46.
Full upward software compatibility between the systems will be maintained, the 70/61
software containing all the 70/46 software features as a subset. Shared code, however,
will be compatible only at the source level between the two systems; this restriction is
necessary to increase the amount of virtual memory available for shared code on the
70/61 •

• 34

Sequential Input-Output Units
RCA has announced a wide range of input-output units for the Spectra 70 computer family.
Some of the more significant units are:
•

9-track System/360-compatible magnetic tape units with transfer rates of up
to 120, 000 bytes per second.

•

A fast card reader (1,435 cards per minute) with optional mark-senSing capabilities
at a slower rate.

•

A triple-purpose optical scanner with optional mark-sensing and punched-hole
reading capabilities.

Recent additions to the product line include a MICR controller, a high-speed paper tape
reader, new printer models with 96-character drum sets, and 7 -track magnetic tape units
and controllers to facilitate compatibility with earlier RCA 301 and 501 systems. Table IV
summarizes the capabilities of the available sequential input-output units.
TABLE IV: SPECTRA 70 SEQUENTIAL INPUT-OUTPUT UNITS
Unit

Models Available

1,435 cpm
100 cpm
300 cpm

70/237 Card Reader
70/234 Card Punch
70/236 Card Punch
70/221 Paper Tape
Reader-Punch
70/224 Paper Tape
Reader
70/242 Printer
70/243 Printer

200 cps (reader)
100 cps (punch)
Bi-directional reading
132 or
160 columns
132 or
160 columns; 64or 96-character
print drums

7 -track Magnetic
Tape Units

1,000 cps
6251pm
1,2501pm

600lpm
400 cpm

70/248 Bill Feed
Printer
9-track Magnetic
Tape Units

Peak Speed

30, 60 or 120 KB
versions; seventrack adapters

120,000 bytes/sec

16.6 or 25KB versions

25,000 bytes/sec

70/251 Videoscan
Document Reader

1,600 doc/min.

~

.35

Display Equipment
Display devices are a means for presenting information either to a camera or directly
to man. They generally hold only a small amount of data for only a short time; their
value is in the variety and speed at which they can display the appropriate information.
RCA provides a display device called the 70/752 Video Data Terminal. This combination
entry and display device can be used for both local and remote operations. Up to 1,080
characters can be displayed on a 12-inch rectangular cathode-rio/-tube screen.

9/69

A

(Contd.)

AUERBACH

'"

SUMMARY

. 36

710:011.360

Data Communications Equipment
The RCA 70/668 Communications Controller - Multichannel (CCM) operates on the 70/35
and higher-numbered processors and terminates from 1 to 48 communications lines
serving a wide variety of remote terminals. Each of the 1 to 48 scan positions requires
a communications buffer, and in some cases a data set, to interface with the communications line. The CCM is connected to the Spectra computer by one trunk of the Multiplexor
Channel. Each scan position of a CCM uses one Multiplexor subchannel. The maximum
total communications data rate that one 70/668 CCM can handle is 6,000 characters per
second.
The RCA Communication Controls (Single Channel) permit remote half-duplex communications between an RCA Spectra 70 computer system and an RCA 301,3301, or Spectra 70
computer system that is equipped with the appropriate communications equipment. Different models of these Controls permit communication over the public switched telephone
network, a common-carrier leased voice-band line, or common-carrier leased broadband line. Some models offer facilities for programmed automatic dialing over the public
telephone network through use of a Bell System Automatic Calling Unit. '\

a

The RCA 70/510 Voice Response System is an on-line inquiry/response system that delivers recorded human-voice responses to an inquirer at a Bell System Touch-Tone telephone. A basic Voice Response Unit (VRU) can handle up to 10 communication lines;
optional features permit the maximum number of communication lines per VRU to be expanded to 50. Various models of the Voice Response Unit offer vocabularies that range in
capacity from 31 to 189 words, all of which can be user-specified.
The RCA 70/630 Data Gathering System (DGS) is an on-line data collection system designed
for gathering information at remote, point-of-transaction terminals and transmitting this
data to a central Spectra 70/35, 70/45, 70/46, or 70/55 system. DGS connects to the central system via a 70/725 DGS Buffer and the 70/668 Communication Controller - Multichannel. DGS input stations can consist of 70/6321 Badge Readers, 70/6331 Card Readers,
and 70/6341 Variable Data Readers with 10 decks of numeric pushbutton keys .
. 37

System Configuration
The Spectra 70 peripheral devices and their controllers are connected to the 70/25 and
larger systems through input-output channels of various types and capacities. A single
Multiplexor Channel is provided as standard equipment for the 70/35 and highernumbered processors and as optional equipment for the 70/25 Processor. The Multiplexor Channel of the Spectra 70/45 and larger systems can control up to 256 low-speed
devices. Selector Channels are provided as standard equipment for the 70/15 and 70/25,
and two Selector Channels are supplied as standard equipment with the 70/60 and 70/61
systems; Selector Channels for all the systems are available optionally. A Selector
Channel provides direct control of one high-speed input-output operation at a time.
Table V shows the various combinations and capacities of Multiplexor and Selector
Channels possible for all Spectra 70 systems, together with the maximum number of
simultaneous I/O operations per system .

. 38

Simultaneous Operations
An RCA Spectra 70 Central Processor (except for the small-scale Model 70/15) can concurrently execute:

• One machine instruction; and
• Up to eight high-speed input-output operations (one per Selector Channel); and
• Multiple slower input-output operations via a Multiplexor Channel.
Table V summarizes the mix possibilities and simultaneous operations capabilities of the
various Spectra 70 input-output channels.
In general, the relationships between RCA Spectra 70 peripheral devices and input-output
data channels are determined at installation time and cannot be altered under program
control except by the inclusion of special optional features. Since it is not normally possible to assign by program any free channel to any available peripheral device, the number
of input-output operations that can actually occur simultaneously can in many cases be
considerably fewer than the theoretical maximum. However, special features are available
to switch a limited number of devices to free data channels under program control.

.4

SOFTWARE
RCA's software systems for the Spectra 70 series, in general, closely parallel the structure and contents of the software supplied by IBM for its System/360 series. Since announcement of the Spectra 70, RCA has greatly improved its standard software by adding "thirdgeneration" software facilities such as disc-oriented control systems, disc file language
facilities, automatic on-line file management techniques, comprehensive data communications control routines, and time-sharing processing support. Multiprogramming control
© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

9/69

710:011.400

RCA SPECTRA 70

.4

SOFTWARE (Contd. )
for up to six jobs is provided for SpeCtra 70 systems that have a minimum of 65K bytes of
core storage. A small-scale multiprogramming system called Primary Communications
Oriented System is also available for basic two-level multiprogramming with systems as
small as 16K bytes.
The principal levels of RCA Spectra 70 software are designated Primary Operating System,
Tape Operating System, and Tape/Disc Operating System, in order of increasing complexity and capability. Software for the small-scale Spectra 70/15 system, however, is
a specially-designed, card-oriented set of routines that provides assembly language,
Report Program Generator, I/o control, and service routine facilities at 4K- and 8K-byte
core storage design levels. The system can also be supplied in a magnetic tape-oriented
version. The more powerful Spectra 70/25 system functions with an integrated operating
system, similar to RCA's Primary Operating System (see below), that provides an
assembler, RPG, Sort/Merge, and standard utility programs. The 70/25 can also use a
special stand-alone simulator program to simulate mM 1401 programs.
The Spectra 70/25 operating system offers basically the same supervised facilities as the
Primary Operating System (POS) for the larger Spectra 70 systems and functions with a
minimum hardware configuration of 16K bytes of core storage, four magnetic tape units,
console typewriter, printer, and card reader and punch. The principal limitation of 70/25
POS facilities in comparison to the POS facilities for the 70/35, 70/45, and 70/55 systems
is the omission of a COBOL language processor.: (Random-access mass storage devices
cannot be used with 70/25 systems, constituting a significant hardware difference between
the 70/25 and the larger members of the Spectra 70 family). The method of implementation
of POS programs for use with the Spectra 70/25 differs from that used with the larger
Spectra 70 systems due to the fact that the 70/25 Processor has a somewhat restricted
instruction set .

. 41

Primary Operating System (POS)
The Primary Operating System for use with the Spectra 70/35, 70/45, and 70/55 systems
is a magnetic tape-oriented software system that provides basic supervisory control for
the sequential execution of programs, interrupt control, and input-output control, as well
as a COBOL compiler, assembler, report program generator, and standard utility
routines. POS COBOL is a subset language of full COBOL 65 and requires a minimum of
32K bytes of core storage for compilations. The POS Assembler also requires use of
TABLE V: SPECTRA 70 INPUT-OUTPUT CHANNEL COMBINATIONS
Processor Model
Standard
Channel Complement

Selector Channels Trunks per channel
Number of simultaneous
data transfer operations
Multiplexor Channels Number of devices
Number of simultaneous
data transfer operations
Fully Expanded
Channel Complement
Selector Channels Trunks per channel
Number of simultaneous
data transfer operations
Multiplexor Channels Number of devices
Number of simultaneous
data transfer operations
Combined total of possible
simultaneous data transfer operations

9/69

70/15

70/25

70/35

1
6
3

4
1
4

-

-

-

--

1
256
8

1
256
8

0

0

70/45 70/46
0

0

70/55

70/60

70/61

-

0

2
1
1

2
1
1

1
256
8

1
248
16

1
248
16

70/55

70/60

70/61

-

-

0

1
192
7

70/15

70/25

70/35

1
6
3

8
1
8

2
2
2

3
2
3

4
2
4

6
4
6

6
3
6

6
3
6

-

0

1
115
8

1
192
7

1
256
8

1
256
8

1
256
8

1
248
16

1
248
16

3

16

9

11

12

14

22

22

A

70/45 70/46

(Contd.)

AUERBACH

'"

SUMMARY

.41

710:011.410

Primary Operating System (FOS)

(Contd. )

32K bytes of core storage. other POS facilities are designed to permit operation in a
minimum environment that includes 16K bytes of core storage and four magnetic tape units.
No FORTRAN or PL/I processors are provided under POS, nor are any routines supplied
for the automatic control of random-access devices, although the operation of these devices
can be programmed at the assembly-language level. The only forms of multiprogramming
supported by POS are the RCA-provided Peripheral Control Routine, which permits concurrent operation of up to three data transcription routines, and the Primary Communications Oriented System for basic two-level multiprogramming .
• 42

Tape Operating System (TOS)
The second major level of Spectra 70 software support designed for use with the 70/35
and higher-numbered systems is designated the Tape Operating System (TOS). TOS is a
magnetic tape-oriented integrated software package that provides supervisory control
programs, language processors, and utility programs for installations that have a minimum. hardware configuration of 65K bytes of core storage, five magnetic tape units, console typewriter, card reader, and line printer. The facility to control multiprogrammed
operation of up to six programs concurrently is the primary feature of TOS software. The
basic TOS Executive program requires a minimum of 16K bytes of core storage. The
Monitor program that coordinates the operations of stacked-Job processing requires an
additional 4K bytes, and the File Control Processor for input-output device and file control requires another 4K to 8K bytes of core storage. Although the theoretical maximum.
number of problem and control programs that can be processed concurrently is six, the
actualI1mit will frequently be fewer than six, limited by the amount of available core
storage and number of available peripheral devices. As many as five magnetic tape units
can be dedicated to system control and library functions when processing in a stacked-job,
multiprogramming environment.
In addition to a comprehensive assembly system, TOS offers a COBOL language similar
to mM's Operating System/360 COBOL F, as well as full-scale FORTRAN IV language
that includes all Operating System/360 FORTRAN IV facilities except random-access
device control statements. No PL/I language processor has been scheduled for implementation to date .

• 43

Tape/DiSC Operating System (TOOS)
RCA's Tape/DiSC Operating System (TDOS) is an improved and extended version of its
Tape Operating System (TOS). An enhanced version of TDOS, OS60, will be available
for the 70/60. In addition to all TOS software facilities, TOOS offers options that permit
system control routines, problem programs, and library subroutines to reside on either
the 70/564 Disc or 70/565 Drum units in order to improve the Spectra 70's throughput
capabilities. As a result, TOOS offers more efficient multiprogrammed operations than
does the tape-oriented Tape Operating System. Also, with TDOS the number of Job Control Language statements required to prepare and compile object programs is reduced,
increasing the efficiency of program preparation. Another significant addition to the
TDOS software package is a comprehensive set of input-output routines for control of
data communication devices. This communications package, called the Multichannel
Communication Program (MCP) , offers most of the same facilities as the IBM System/
360 Queued Telecommunications Access Method (QTAM) software for data communications control. The MCP system can accept remote messages either as they are entered
or as polled, in contrast to the polled-only acceptance technique of QTAM.
The minimum Spectra 70 core storage requirements for use of TDOS remains at 65K bytes,
of which 16K bytes are permanently reserved for the Executive. Both the Executive
Monitor and the optional data communications package require 4K bytes of storage. The
Monitor, however, is a transient routine and does not require permanent residence in
core; the data communications control routines are permanently resident. The principal
software components of the Tape and Tape/DiSC Operating Systems are listed in Table VI,
where the scheduled availability date for each element is also shown .

. 44

Time Sharing Operating system (TSOS)
The Time Sharing Operating System (TSOS) is a specialized software system designed for
control and support of the RCA Spectra 70/46 time-sharing processor. An enhanced
version of TSOS, 0861, will be available for the 70/61. The system is scheduled to provide advanced time-sharing capabilities, as well as improved facilities for handling batch
processing in a multiprogramming mode.

I

Provided within the TSOS Executive program are routines for handling task scheduling
(capable of using a time-slicing algorithm), memory management, device allocation,
phYSical-level input-output, and a combination command and job control language. Also

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

9/69

710:011.440

. 44

RCA SPECTRA 70

Time Sharing Operating System (TSOS)

(Contd. )

provided as a TSOS system program will be a File Control Processor (FCP) with extensive automatic data management capabilities. In addition to all of the language processors
and utility routines used with TOS and TDOS, TSOS also provides a full conversational
FORTRAN compiler, plus conversational text editor and desk calculator programs. Conversational syntax checking will also be provided for the FORTRAN, COBOL, and Assembler languages.
The minimum equipment required to use the Time Sharing Operating System includes a
Spectra 70/46 processor with at least 262K bytes of core storage, a 70/567 Drum Memory
Unit, two 70/564 Disc Units, two 9-track magnetic-tape units, one card reader, and one
printer. To support conversational users, the system must also include a 70/668 Communications Controller - Multichannel and from 1 to 48 (70/46) or 128 (70/61) remote
terminal units .
. 45

Basic Time-Sharing System (BTSS)
The Time-Sharing Operating System (TSOS) should not be confused with another Spectra 70
time-sharing system called the Basic Time-Sharing System (BTSS). BTSS is a much more
restricted software system than TSOS, and is designed to run on the general-purpose
Spectra 70/45 hardware. BTSS is a full operating system, designed to permit up to 16
concurrent users at remote terminals to use a conversational, interpretive FORTRAN
compiler, and to perform library maintenance functions on data and program files, which
can then be processed under another general-purpose Spectra 70 operating system.
The Basic Time Sharing System requires a Spectra 70/45 system with either 131K or
262K bytes of core storage, a minimum of one 70/564 Disc Storage Unit, one 70/668 Com~uni~ations 

.

5JfJ)
850
10
31

(~)0(3)
850
10
31

(~0(3
920
120

-

153.00
1.75
5.50
(2)

-

153.00
1.75
5.50
(2)

-

166.00
20.75

(Contd.)

PRICE DATA

710:221.105

IDENTITY OF UNIT
CLASS

INPUTOUTPUT
(Contd. )

Model
Number

Feature
Number

Name

PRICES
Monthly
Monthly
Rental iPurchase Maint.

$

$

$

Paper Tape (Contd.)
70/221-10
5219-10
5219-11
5256
70/221-11
5219-10
5219-11
5256
70/221-20
5219-10
5219-11
5256
70/221-21
5219-10
5219-11
5256
5292
5296
5297
5298
5299
70/224-10
5264
5273
70/224-11
5264
5273
5293

*70/242-10
5221
*70/242-20
5221
70/242-30
5221
70/242-40
5221
*70/243-10
*70/243-20
70/243-30
70/243-40
70/243-51
70/243-61
70/248-11
5216
70/249-11
5262-1
5262-2
5262-3
5274
5278-30

Paper Tape Reader/Punch
Advanced Sprocket 6- Level Read
Advanced Sprocket 6- Level Read
Long Block Indicator
Paper Tape Reader/Punch
Advanced Sprocket 6- Level Read
Advanced Sprocket 6- Level Read
Long Block Indicator
Paper Tape Reader/Punch
Advanced Sprocket 6- Level Read
Advanced Sprocket 6- Level Read
Long Block Indicator
Paper Tape Reader/Punch
Advanced Sprocket 6- Level Read
Advanced Sprocket 6- Level Read
Long Block Indicator
4N Terminate (10)
Read Kleinschmidt Format (10)
End of Tape (10)
Gapless Mode (10)
Punched Kleinschmidt Format (10)
Paper Tape Reader
Long Block Indicator
Supply Reel Reverse
Paper Tape Reader
Long Block Indicator
Supply Reel Reverse
4N Terminate
Printers(4)
Printer, Medium Speed
Dual Speed Form Advance
Printer, Medium Speed
Dual Speed Form Advance
Printer, Medium Speed
Dual Speed Form Advance
Printer, Medium Speed
Dual Speed Form Advance
Printer, Hi-Speed
Printer, Hi-Speed
Printer, Hi-Speed
Printer, Hi-Speed
Hi-Speed Printer
Hi-Speed Printer
Bill Feed Printer
Interchangeable Chain Cartridge
Bill Feed Printer Control
Document Processing Operation(5)
Document Processing Operation(6)
Document Processing Operation(7)
Line Counter(8)
Special Format Control(9)

515
46
46
20
540
46
46
20
570
46
46
20
595
46
46
20
40
40
10
10
35
565
20
20
590
20
20
49

720
103
1,030
103
720
103
1,030
103
1,030
1,340
1,080
1,390
1,290
1,595
1,725
75
1,065
25
25
25
15
45

24,250
2,250
2,250
950
25,500
2,250
2,250
950
26,950
2,250
2,250
950
28,250
2,250
2,250
950
1,900
1,900
475
475
1,650
26,700
950
950
27,950
950
950
1,900

92.75
8.25
8.25
3.50
97.25
8.25
8.25
3.50
103.00
8.25
8.25
'3.5Q·
107.00
8.25
8.25
3.50
8.00
8.00
2.00
2.00
7.00
102.00
3.50
3.50
106.00
3.50
3.50
8.00

33,950
4,850
48,500
4,850
46,100
4,850
60,650
4,850
48,500
63,050
50,950
65,500
60,650
75,200
84,750
3,125
50,250
1,200
1,200
1,200
700
2,100

126.00
17.00
180.00
17.00
172.00
17.00
226.00
17.00
180.00
235.00
189.00
243.00
226.00
279.00
395.00
85.25
3.50
3.50
3.50
2.75
4.75

16,050
8,800
50
50
1,200
1,450
4,050
4,300

20.50
33.25
-

Console
--70/97
70/216
5259
5260
5276-1
5276-2
5277-1
5277-2

Console
Input/Output Typewriter
Form Supply Box
Special Type Slug and Key Cap
Paper Tape Reader
Paper Tape Reader
Paper Tape Punch
Printing Paper Tape Punch

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

340
185
50(3)
50(3)
25
30
85
90

--

4.50
5.50
15.25
16.25

4/69

RCA SPECTRA 10

710:221. 106

IDENTITY OF UNIT
CLASS

INPUT
OUTPUT
(Contd.)

Model
Number

Feature
Number

Name

PRICES
Monthly
Monthly
Rental ~rchase Maint.

$

$

$

Magnetic Ink Character Reader
MICR Sorter-Reader Controller
MICR Sorter-Reader Controller
MICR Sorter-Reader Controller

70/272-10
70/272-20
70/272-30

620
620
620

29,100
29,100
29,100

49.50
49.50
49.50

Video Data Systems

70 '751-10
70/751-11
70/751-12
70/752
5707
5710
5711
5733-01
70/755
70/756-11
70/756-21
70/756-31
5716
5721
70/759-11
70/759-21
5715

Video Data Terminal
Video Data Terminal
Video Data Terminal
Video Data Terminal
Station Selection
Data Format
Printer Adapter
Special Symbol Set (Puerto Rican)
Video Data Switch
Video Data Generator
Video Data Generator
Video Data Generator
Data Format
Variable Start of Transmission
Video Data Controller
Video Data Controller
Station Selector

75(11)
75(11)
75(11)
190
20
20
40
Exchange
125
275(11)
225(11)
200(11)
10(11)
10(11)
600(11)
400(11)
35(11)

3,525
13.00(11)
3,525
13.00(11)
3,525
13.00(11)
8,325
23.50
850
2.50
850
2.50
1,700
5.00
basis at no charge
5,900
13.75
12,925
47.50(11)
10,575
39.00(11)
9,400
34.50(11)
470
1. 75(11)
470
1. 75(11)
28,200 104.00(11)
18,800
69.00(11)
1,175
4.50(11)

Voice Response Unit

,COMl\IUNICATIONS

70/510-11
5514-11
5518-10
5518-20
70/510-21

70/510-26

5514-21
5518-30
5518-40
5514-26
5518-50

Voice Response Unit
Line ExpanSion Feature
CUstom Vocabulary
Custom Vocabulary
Voice Response Unit
Line Expansion Feature
Custom Vocabulary
Custom Vocabulary
Voice Response Unit
Voice Line Expansion
CUstom Vocabulary

515
160
1,275(3)
1,650(3)
620
206
2,100
2,350
825
273
2,350(3)

24,250
7,550
1,200
1,550
29,100
9,700
2,000
2,225
38,800
12,900
2,225

51.50
16.00
1.25
1.50
62.00
20.50
2.00
2.25
82.50
27.25
2.25

465
103

21,900
4,850

37.25
8.25

103

4,850

8.25

310

14,550

24.75

310

14,550

24.75

310

14,550

24.75

440

20,700

46.00

25
(2)
(2)
1,100
1,370
1,450
1,450
720

1.175
(2)
(2)
57,625
59,400
63,000
63,000
33,950

2.50
(2)
(2)
200,00
238.00
250.00
250.00
86.50

Controls

---

70/627-10
*70/652-25
*70/652-26
70/653-25
70/653-26
70/653-34
70/656
5628
5630-1
5630-2
70/658-11
70/658-12
70/658-13
70/658-132
70/668-11

4/69

Data Exchange Control
Communications ControlSingle Channel
Communications ControlSingle Channel
Communication ControlSingle Channel
Communication ControlSingle Channel
Communication Control Single Channel
Communication ControllerSingle Channel
Auto-call Feature
Line Adapter
Line Adapter
Autodin Communications Controller
Autodin Communications Controller
Autodin Communications Controller
Autodin Communications Controller
Communications ControllerMultichannel

fA

AUERBACH

'"

(Contd. )

710:221.107

PRICE DATA

PRICES

IDENTITY OF UNIT

CLASS

Model
Number

COMMUNICATIONS
(Contd. )

Feature
Number

Name

Monthly
Monthly
Rental Purchase Maint.

$

$

$

50
20
15
10
10
15
(2)
20
925

2,350
940
700
470
470
700
(2)
940
43,650

5.00
2.00
1.50
1. 00
1. 00
1.50
(2)
2.00
111. 00

50
20
15
10
10
15
(2)
20
1,135

2,350
940
700
470
470
700
(2)
940
53,350

5.00
2.00
1.50
1. 00
1. 00
1.50
(2)
2.00
136.00

50

2,350

5.00

27
38
38
43
20
43
20
10
175
52
110
20
20
40
75
75
113

1,300
1,850
1,850
2,050
950
2,050
950
475
8,550
2,450
5,175
850
850
1,700
75
75
5,350·

2.75
2.75
3.75
4.25
2.00
4.25
2.00
l. 00
17.50
5.25
8.75
2.50
2.50
5.00

Controls (Contd.)
5617-1
5618
5620
5622
5623
5624
*5632
5635
70/668-21
5617-1
5618
5620
5622
5623
5624
*5632
5635
70/668-31
5617-1
70/710
70/712
70/715
70/720Ser
5705
70/721
5705
5714-2
*70/722
*70/724
70/725
5707
5710
5711
5713
5725
70/780

Telex Operation
USASCn Block Check Character
Timer Restart/Interval Selector
Message Separation
OW-Unshift
Timer Reset
STR Operation
Synchronous Full-Duplex Operation
Communications ControllerMultichannel
Telex Operation
USASCn Block Check Character
Timer Restart/Interval Selector
Message Separation
OW-Unshift
Timer Reset
STR Operation
Synchronous Full- Duplex Operation
Communications ControllerMultichannel
Telex Operation
Communications Adapters:
Telegraph Buffer
Telegraph Low Level RIffer
Parallel Buffer
ADS Buffer
Auto- Call Feature
SDS Buffer
Auto-Call Feature
Full Duplex Operation
STR Buffer
EDGE Demodulator Buffer
DGS Buffer
station Selection
Data Format
Printer Adapter
Keyboard Cable Extension
Data Set Cable Extension
Time Generator/Buffer

NOTES:
* No longer in production.
** Type n Processors not available until after June 30, 1969.
(1) 7-Channel Feature available at no extra cost.
(2) Furnished on request, when required, at no additional charge.
(3) Single use charge.
(4) Various options are available for altering the standard character set. These options are available on a
"one-time" usage charge or sale-only basis. Typical charges are $150. per special character, $1,940
for a drum tooling charge, and $2,000 or $2,230 for a print drum segment set.
(5) For all 242 series Printers.
(6) For 242-10, 242-20, 243-10, and 243-20 Printers.
(7) For 243-51 and 243-61 Printers.
(8) For 242-30, 242-40, 243-30, and 243-40 Printers.
(9) For 242-30, 242-40, 243-30, 243-51, and 243-61 Printers.
(10) For all 70/221 series models.
(11) Rental and monthly maintenance charges are higher in remote areas; charges stated are
for metropolitan areas.
(12) Charges have not yet been released.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

-

-

11.25

-

4/69

712:011. 100

~

II ......

~'\EDP
I(POnS

RCA SPECTRA 70/15
SUMMARY

AUERBACH

•

SUMMARY
The Spectra 70/15 is a small-scale general-purpose computer with a restricted instruction rcpertoire. Its primary thoul-\h not exclusive function is to serve either as a satellite
systl'm for large'r computers or as a remote communications terminal unit. Most peripheral
units available for the Spectra 70 series can be connected to the 70/15, with the exception of all
t'andom-access de'vices. (Please refer to the main RCA Spectra 70 report, behind tab 710,
for descriptions of the available' peripheral devices.) Rentals for typical Spectra 70/15 systems
fall bl'tll'een $2,500 and $5,000 per month.
The 70/15 was announced in December HhH. The first customer delivery was made
during the last quarter of 1965, and the Spectra 70/15 software package was also supplied at
this time. Descriptions and representative performance timings of the various software elements are included within this subre'port, in Section 712: 151.
Every 70/15 system includes a central processor and either 4,096 or 8,192 bytes of
core' storage with a cycle time of two microseconds per byte. Section 712:051 provides a
detailed description of the 70/15 Processor's capabilities.
The input-output facilities of the Spectra 70/15 computer system consist oLone
input-output channel with six subchannels. Each subchannel can control up to 16 peripheral
devices. Normal use of the I/O channel prevents operation of the central processor while
the channel is in use. However, an auxiliary mode of operation allows either a read or
write operation to occur in parallel with central processor operations. Such auxiliary read/
write operations are unsupervised by the processor and come to a halt only when the inputoutput data is exhausted or when the I/O device finishes its operation cycle. This mode can
be used to advantage by the unbuffered card reader and by the magnetic tape units. Section
712:111 provides details of the demands placed upon the Spectra 70/15 Processor during the
operation of the individual peripheral units. A total System Performance analysis is presented in Section 712:201.
The System Configuration section (712:031) shows two typical Spectra 70/15 equipment configurations, including monthly rental prices. Configurations shown are a typical
punched-card system and a four-tape business system, arranged according to the standard
rules set forth in the Users' Guide, page 4:030.120.
The Spectra 70/15 Software Package has been developed to function with a minimum complement of hardware, including 4,096 bytes of core storage, an on-line printer,
card reader and card punch. The use of magnetic tape units and 8,192 bytes of core storage
expands the power of each entry in the basic software package and reduces inter-job setup
time by making available a Program Library Tape. The entries within the software package
for the Spectra 70/15 are described in detail in Section 712:151. Included are a two-pass
assembly system, an input-output control system, a sort/merge generator, a report program generator, a group of utility routines, a Single-Channel Communications Control System, and a Program Binder that helps to alleviate the restrictions on program size imposed
by the 4K or 8K memory size. No compiler for COBOL, FORTRAN, or any other processoriented language has been announced for the Spectra 70/15.

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

11/68

-

713:011.100

A ;;~p
_-------J
AUER8AC~

RCA SPECTRA 70/25
SUMMARY

REPORTS

~

SUMMARY
The Spectra 70/25 is a sequential processor that uses parallel input-output channels to
obtain overlapped operations. It can be connected to any of the Spectra 70 peripheral units except
the random access storage units. The restricted machine instruction repertoire includes decimal
add, subtract, multiply, and divide operations; binary add and subtract operations; editing and
other data handling operations; logical instructions; 11 decision and control instructions; and 7
I/O instructions. (See the Instruction List, Section 710: 121, for details.) No automatic facilities
for conversion between binary and decimal radices or for any floating-point operations are included. The processor registers are stored as addressable parts of the main core storage. Interruption facilities are standard.
The 70/25 Processor can contain from 16,384 to 65,536 bytes of core storage. The
core cycle time is 1. 5 microseconds per four bytes for internal operations. Input-output operations take place one byte at a time, so the effective core cycle for input-output purposes is 1.5
microseconds per byte.
Although no longer in production, this system is still obtainable on an as-available
basis.
Rentals for typical Spectra 70/25 systems range from about $4,000 to $8, 000 per
month for unlimited use.
The Spectra 70/25 software includes an assembly language, a report program generator,
and various utility systems, including a communication control system for single-line data communication operations. An operating system based on the same principles as the Primary Operating System (POS) for the larger Spectra 70 computers is available, and concurrent data
transcription operations are possible where there are at least 32K bytes of core storage. No
COBOL, FORTRAN, or other compilers are available for the 70/25.
All of the Spectra 70/25 software is designed to work exclusively with the 70/25; it is
not possible, for instance, to compile a FORTHAN program on the Spectra 70/45 for operation on
the Spectra 70/25. In the opposite direction, a Compatibility Support Package is being provided
to assist in checking over 70/25 programs so that they can be safely run on the Spectra 70/45
or other larger Spectra 70 systems.
This report concentrates upon the characteristics and performance of the Spectra
70/25 system in particular. All general characteristics of the Spectra 70 hardware are
described in Computer System Heport 710: RCA Spectra 70 - General.
The System Configuration section which follows shows the Spectra 70/25 in the following
standard System Configurations:
II:
III:
IV:

4-Tape Business System
6-Tape Business System
12-Tape Business System.

These configurations were prepared according to the rules in the Users' Guide, page
4: 030.120, and any significant deviations from the standard specifications are listed. *
Section 713: 051 provides a detailed description of the central processor capabilities
and timings for the Spectra 70/25.
The input-output channel capabilities of the Spectra 70/25, and the demands upon
the processor during input-output operations, are described in Section 713:111.
The software that can be used with Spectra 70/25 systems is described in Sections
710:151, 710:171, and 710:191.
The overall performance of any Spectra 70 system is heavily dependent upon the processor model used. A full System Performance analysis of the 70/25 is provided in Section
713:201.
Four input-output Selector Channels and an Elapsed-Time Clock are standard features
of the Spectra 70/25. Optional processor features include an input-output Multiplexor Channel,
four more Selector Channels, and the option to use two High-Speed Selector Channels. A HighSpeed Selector Channel replaces two standard Selector Channels and has a maximum data rate
of 500,000 bytes per second, as compared with the 200, OOO-bytes-per-second rate of each of
the replaced channels.

* Spectra 70/25

systems do not permit use of random access storage devices, so our several
random access-oriented standard configurations cannot be shown.
© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

11/68

~ "'

714:011. 100

.....

~'- EDP

AUER8AC~'

•

RCA SPECTRA 70/35
SUMMARY'

1£'0115

SUMMARY
The Spectra 70/35 computer system was formally announced by RCA in September
1965, nine months after the original announcement of the Spectra 70 line. First delivery of
a 70/35 system occurred in February 1967. The 70/35 Processor is the least expensive
of the program-compatible Spectra 70/35, 70/45, and 70/55 processors. The performance
of the Spectra 70/35 system generally falls somewhere between that of IBM System/360
Models 30 and 40.
Internal storage capacity of Spectra 70/35 systems can range from 32,768 to 65,536
bytes of core storage. In addition, "Non-Addressable" core storage is supplied to provide
control registers for each input-output device. Non-Addressable core storage also provides
the processor's general registers. floating-point registers, and various other control registers. The core storage cycle time is 1.44 microseconds per two bytes.
The rental for typical Spectra 70/35 systems ranges betwecn $5,000 and $13,000 per
month. A Spectra 70/35 system arranged in AUERBACH's Standard Configuration III, with 32K
bytes of core storage, six 30KB magnetic tape units, printer, card reader and punch, rents for
$7.616 per month (see Report Section 714:031).
To hold the optional emulators that enable it to execute machine-language programs
written for IBM HOI/1460 or RCA 30] computer systems. the Spectra 70/35 utilizes a readonly memory unit similar to that used in the Spectra 70/45. Each 54-bit word of read-only
memory holds two processor "elementary operations, " twice the capacity of the Spectra 70/45's
read-only memory. As a result. the Spectra 70/35 emulators will require less memory and
will generally perform more efficiently than the emulators used with the Spectra 70/45 system.
Detailed descriptions of the emulators for the 70/35 can be found in Sections 710:131 and 710:134.
This subreport concentrates upon the characteristics and performance of the Spectra
70/33 system in particular. All general characteristics of the Spectra 70 hardware and software
are described in Computer System Rcport 710: RCA Spectra 70 - General.
The System Configl.lratlOn section which follows shows the Spectra 70/35 in the following standard System Configurations:
I:
II:
III:
IIIR:
IVR:
VI:
VIIA:

T~'pical

Card System

-! -Tape Business System

G-Tape Business System
;;-:'IIillion-Byte Random Access System
20-:\Iillion-Byte Random Access System
G-Tape Business/Scientific System
10-Tape General System (Integrated).

These configurations \\'ere prepared according to the rules in the Users' Guide,
Pages 4:030.120 and 4:200. GOO, and any sip;nificant deviations from the standard specifications
are listed.
Section 714:03] pro\'ides detailed central processor timings for the Spectra 70/35.
See Section 710:031 for all the other characteristics of the Spectra 70 processors.
The input-output channel capabilities of the Spectra 70/35, and the demands upon the
processor during input-output operations, are described in Section 71-1:111, Simultaneous
Operations. The Selector Channels used with the Spectra 70/35 have a 60 per cent greater
transmission-rate capability than those used with the faster 70/45 Processor. However, the
effecth'e 70/35 Multiplexor Channel transmission-rate capabilities are 50 per cent slower
than those possible with the 70/4;; :'IIuitiplcxor Channel.
The software that can be used with a given Spectra 70 system configuration depends
upon the amount of core storage and the number and type of peripheral devices that are available. A detailed description of the software that can.be used with the Spectra 70/35 and other
Spectra 70 systems can be found in Sections 710:]3] through 710:193.
The overall performance 01 any Spectra 70 system is heavily dependent upon the
processor model used. A full System Performance analysis of the 70/35 system is provided
in Section 7 ]4:201 of this subreport.
The Multiplexor Channel. with seven subchannels. is a standard feature of the RCA
Spectra 70/33 Processor. Memory Protect, an Elapsed-Time Clock, Direct Control, an RCA
3f)1 Emulator, an IBM 1401/14GO Emulator, and up to two Selector Channels are optional
features.
© 1968 AUERBACH Corporation and AUERBACH Info. Inc.

11/68

715:011.100

A

AUERBACH

STUDnl

EDP

RCA SPECTRA 70/45
SUMMARY

UPGRTS

SUMMARY
The Spectra 70/45 Processor can be connected to any of the Spectra 70 peripheral
units, can control a communications network, and can handle a read/write Direct Control channel.
Sharing of core memory between different processors is not currently possible in the Spectra
70/45, but memory-to-memory transfers can be made by means of the Data Exchange Control.
The 70/45 Processor contains from 16,384 to 262,144 bytes of core storage for program data, in addition to some non-addressable core storage used for input-output purposes.
The core cycle time is 1.44 microseconds per two bytes.
The rental for typical Spectra 70/15 systems will generally fall between $8, 500 and
$15,000 per month, with Standard Configuration III (16K core, six 30KC magnetic tape units,
reader, punch, and printer) renting at $8,717 per month for unlimited use.
The Spectra 70/45 uses a read-only memory as an internal control system, which
can be expanded to permit the "emulation" of other computer machine languages. The characteristics of the 70/45 as an emulator for RCA 301 and 501 programs, as well as for IBM 1401
and 1410 programs, are discussed in Sections 710:131 through 710:135.
This report concentrates upon the characteristics and performance of the Spectra
70/45 in particular. All general characteristics of the Spectra 70 hardware and software are
described in Computer System Report 710: Spectra 70 - General.
The System Configuration section which follows shows the Spectra 70/45 in the following
standard System Configurations:
III:
6-Tape Business System
IlIR:
5-l\Iillion-Byte Random Access System
IV:
12-Tape Business System
IVR:
20-:\1 illion -Byte Random Access System
VI:
6-Tape Business/Scientific System
VIlA: 10-Tape General System (Integrated)
VIIB: 10-Tape General System (Paired).
These configurations were prepared according to the rules in the Users' Guide, pages
-! :030:120 and 4 :200.600, and any signifIcant deviations from 'the standard specifications are

listed. As a matter of general interest, the rentals that'would be incurred if faster magnetic tape
units were installed are listed on the diagrams for Configurations VIlA and VIlB.
Section 715:051 provides detailed central processor timings for the Spectra 70/45.
See Section 710:051 for all the other general characteristics of the Spectra 70 processors.
The input-output channel capabilities of the Spectra 70/45, and the demands upon the
processor during input-output operations, :are described in Section 715:111.
Three integrated software systems are available for use with the Spectra 70/45: the
Primary Operating System, Tape-Tape/Disc Operating System and the Disc Operating System.
The Primary Operating System offers small-scale software (16K-byte design level) that includes
an assembler, report program generator, and COBOL compiler (requiring 32K bytes of core
storage) for use in a sequential processing environment. The Tape-Tape/Disc Operating System
is designed at a 65K-byte level, offering more extensive and more powerful software than the
Primary Operating System. The Tape-Tape/Disc Operating System features multiprogramming
control for up to six concurrently-operating programs. COBOL and FORTRAN compilers and
a full assembly system are also provided. The Disc Operating System (DOS), scheduled for
release on November 29, 1968, will enable the user of a totally random-access oriented system
to do away with the tape units heretofore required for the operation of the RCA Spectra Operating Systems. Among the components planned for release in the near future are a 32K COBOL'
Compiler, a 22K Assembler, and a 65K FORTRAN. The DOS system will offer multiprogramming
capabilities similar to those currently offered by TOS and TDOS, along with a Report Program
Generator and a full complement of utility programs. The Disc Operating System is designed
to operate in a minimum of 32K of storage. Please refer to Sections 710:151 through 710:193
for descriptions of the principal software elements supplied for use with the Spectra 70/45
system.

© 1968 AUERBACH Corporation and AUERBACH Info. Inc.

11/68

RCA SPECTRA 70/45

715:011.101

The overall performance of any Spectra 70 system is heavily dependent upon the
processor model used. A full System Performance analysis of the 70/45 is provided in
Section 715:201.
The Multiplexor Channel, with eight subchannels, is a standard feature of the RCA
Spectra 70/45 processor. Memory Protect, an Elapsed-Time Clock, Direct Control, and up
to three Selector Channels are optional features.
RCA has recently announced that the current model of the 70/45 Processor will be
replaced by the 70-45 Type II. The Type II will differ from its predecessor in the following
areas:
1.

Due to the incorporation of multiple byte buffering, the maximum selector
channel throughput rate will be increased, when more than one selector
channel is installed, from 520 KES to 1000 KES. The maximum rate for a
single channel remains at 465 KES.

2.

The maximum number of optional selector channels will be increased from
three to four.

3.

The Model FE, with a core-storage capacity of 196,608 bytes will be added
to the product line.

4.

The Memory Protect Feature will be expanded to include read/write protection as well as write protection.

5.

The Test and Set instruction will be added to the Spectra 70/45 instruction
list to maintain compatibility with the Spectra 70/46 and mM System/360.

6.

Power supply buffering will be improved to minimize power line transience.

7.

The 70/45-C (16K) and 70/45-D (32K) will no longer be manufactured.

Production cutover from the Type I to the Type II is scheduled for the early part of
the third quarter of 1969. After this date, the Type I will be marketed on an as available basis.

11/68

fA·
AUERBACH

716:011. 100

A

SIUDA"

EDP

RCA SPECTRA 70/55
INTRODUCTION

AUERBACH
!

INTRODUCTION
The Spectra 70/55 Processor can be connectcd to any of the Spectra 70 peripheral
units, can h;Uldle a read/write Direct Control ch'Ulnel, and can control a data communications
network. Communication between different computers can be via memory-to-memory transfers or via core memory modules shared with another Spectra 70/55 processor.
Processor control is not by read-only memory, as in the Spectra 70/35 and 70/45;
conventional wired circlllts :Lre used for control purposes. It is not possible to add read-only
memories to the 70/55 l'rocL'ssor, so there is no compatibility between previous HCA or IBM
systems and the Spectra 70/55 through the compatibility technique called "emulation. "
The 70/55 Processor contains from 65,536 to 524,288 bytes of core storage for
program data, in addition to some non-addressable core storage which is used for input-output
purposes. The core cycle time is 0.84 microseconds per [our bytes.
The rental for typical Spectra 70/55 systems will generally fall between $13,000 and
$30,000 per month, with st:Uldard Configuration VIIB (65K core, eight 60KC magnetic tape units,
and a satellite Spectra 70/15 Processor) renting at $19,620 per month for unlimited use.
This report concentrates upon the characteristics and performance of the Spectra
70/55 in particular. All general characteristics of the Spectra 70 hardware and software are
described in Computer ~'ystem Report 710: RCA Spectra 70 - General.
The System Configuration section which follows shows the Spectra 70/55 in the followirg
standard System Configurations:
III:
IIIR:
IV:
IVR:
VI:
VIlA:
VIlB:
VIIIB:

6-Tape Business System
5-l\Iillion-Byte Random Access System
12-Tape Business System
20-l\lillion-Byte Random Access System
6-Tape Business/Scientific System
IO-Tape General System (Integrated)
10-Tape General System (Paired)
20-Tape General System (Pai red).

These configurations were prepared according to the rules in the Users' Guide, pages
4: 030.120 and 4: 200.600, anc! any significant deviations from the standard specifications are
listed. As a matter of general interest, the rentals that would be incurred if faster magnetic
tape units were installed are listed on the diagrams for Configurations VIlA and VIlB.
Section 716:051 provides detailed central processor timings for the Spectra 70/55.
See Section 710:051 for all the other general characteristics of the Spectra 70 processors.
The input-output channel capabilities of the Spectra 70/55, and the demands upon the
processor during input-output operations, are described in Section 716:111.
Three integrated software systems are available for use with the Spectra 70/55: the
Primary Operating System, the Tape-Tape/Disc Operating System and the Disc Operating System.
The Primary Operating System offers small-scale software (16K-byte design level) that includes
an assembler, report program generator, and COBOL compiler (requiring 32K bytes of core
storage) for use in a sequential processing environment. The Tape-Tape/Disc Operating System is designed at a 65K-byte level, offering more extensive and more powerful software than
the Primary Operating System. The Tape-Tape/Disc Operating System features multiprogramming control for up to six concurrently-operating programs. COBOL and FORTRAN compilers
and a full assembly system are also provided. The Disc Operating System (DOS), scheduled for
release on November 29, 1968, will enable the user of a totally random-access oriented system
to do away with the tape units heretofore required for the operation of the RCA Spectra Operating
Systems. Among the components planned for release in the near future are a 32K COBOL Compiler, a 22K Assembler, and a 65K FORTRAN. The DOS system will offer multiprogramming
capabilities similar to those currently offered by TOS and TDOS, along with Report Program
Generator and a full complement of utility programs. The Disc Operating System is designed
to operate in a minimum of 32K bytes of storage. Please refer to Sections 710:151 through
710:193 for descriptions of the principal software elements supplied for use with the Spectra
70/55 system.

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

11/68

RCA SPECTRA 70/55

716:011.101

The overall performance of any Spectra 70 system is heavily dependent upon the processor model used. A full System Performance analysis of the 70/55 is provided in Section
716:201.

The Multiplexor Channel, with eight subchannels, is a standard feature of the RCA
Spectra 70/55 processor. Memory Protect, an Elapsed-Time Clock, Direct Control, and up to
six Selector Channels are optional features.

11/68

A

AUERBACH
~

717:011.100

A

AUERBACH

stU1Il1D

EDP

RCA SPECTRA 70/46
SUMMARY

UPOlrs

SUMMARY
INTRODUCTION
RCA's entry in the competition for the time-sharing market, the Spectra 70/46, was announced on May 4, 1967. This medium-scale system is based on three new components:
the 70/46 processor, largely a modification of the 70/45 processor; the 70/567 high-speed
magnetic drum unit; and a software system (TSOS) especially designed for time-sharing
applications. Both the hardware and the initial release of the software are scheduled for
delivery on January 31, 19(i9.
The 70/46 represents RCA's first entry into the full-scale commercial time-sharing computer market, joining such competition as the GE-625/635, IBM System/360 Model 67,
and SDS Sigma 7 computer systems. However, RCA has designed its time-sharing system
with more modest goals than those of GE and IBM. The apparent intention of RCA ib to
remain competitive in all areas of the small-to-medium class commercial computer
market, without enduring the frustrations of the more ambitious pioneers in the largescale commercial time-sharing business.
The monthly rental of an RCA Spectra 70/46 time-sharing processor with 262,144 bytes
of core storage is $14,125 (, ee Section 710: 221, Price Data). RCA estimates that typical
70/-l6 system configuration;, will rent for between $25,000 and $30,000 per month. Contributing to the relatively modest prices of planned Spectra 70/46 svstems are the facts:
(1) that the 262K-b~'te 70/ ..l(j Processor is basically an expanded version of the $11,125per-month Spectra 70/-15 2G2K-byte Processor; (2) that only single-processor systems
ha\'e been announced to elate: and (:1) that a maximum of 48 on-line remote terminal units
can be controlled by the s\ ~tem.
According to its design goals, the Spectra 70/-16 Time Sharing SYstem will be an efficient
batch proceSSing s\'stem with advanced multiprogramming capabilitIeS: remote, conversational time-shared operation, II1 time-sliced mode will be a powerful available facility
that ma\', in some installations, consume only a limited amount of the system's total
proceSSing capacit\'.
Noteworthy features of the 70/-lG Time Sharing System include:
•

Hardware logic within the processor to facilitate program paging and
segmentation, providing simultaneous system access for up to 48 users
at remote terminals.

•

Up to 2,097, 152

•

A fast associative memory for translation of all virtual memory
addresses to real core storage addresses.

•

The use of read-only memory to provide by hardware frequently-used
routines required to manipulate the contents of the various translation
tables.

•

A conversational FORTRAN IV compiler with interpretive execution
capability.

•

Full compatibility with the systems programs and language processors
of the RCA Spectra 70 general-purpose Tape Operating System (TOS),
Tape/DiSC Operating System (TDOS), and Disc Operating System (DOS).

b~,tes

of virtual memory available to programmers.

The initial release of the Time Sharing Operating System will consist of a cross section of
the programs necessary for the operation of the system. In addition to the Executive, the
following components will be available:
•

Assembler, version 1 of a modified TOS/DOS Assembler.

•

COBOL, version 1 of a modified TOS/DOS COBOL.

•

COBOL Syntax Checker.
© 1968 AUERBACH Corporation and AUERBACH Info, Inc,

11/68

717:011.101

RCA SPECTRA 70/46

•

FORTRAN, version 1 of a modified TOS/DOS FORTRAN.

•

Terminal Diagnostic Houtine for FORTRAN.

•

IDA (Interactive Dplmgging Aid).

•

Interactive BASH'.

•

Heport Program (;pnerator.

•

Sort/Merge.

•

File Editor.

•

\lata Management.

•

Linkage Editor.

•

Dynamic Linkal-\e Editor.

•

System

•

Hardware lliah'11ostic Houtines.

•

All TOS l1tilit~· Programs.

(~pnt'l·atOl·.

Fh'e subsequent releases, at approximately two month intervals, are planned by RCA,
These enhancements will make the full Time Sharing Operating System available during
the fourth quarter, 1!l6!l.
This subreport concentrates upon the speeialized characteristics of Spectra 70/46 Systems
and the supporting software. All general characteristics of the Spectra 70 hardware are
described in Computer System Report 710: RCA Spectra 70-General.

11/68

A.

AUERBACH

718:011. 100

1. "',,'"
fA'-

AUERBAC~

EDP

RCA SPECTRA 70/60
SUMMARY

I(PDITS

~

SUMMARY
With the announcement of the Spectra 70/60 in February 1969 the computing capabilities of the
top end of the Spectra 70 family of computers were substantially extended. The performance of the
Spectra 70/60 has been further enhanced since the original announcement. and recently. RCA announced the 70/61 time-sharing system. which bears much the same relationship to the 70/60 as
does the 70/46 to the 70/45. The 70/60 is a fully upward-compatible member of the Spectra 70 family of computers and uses the standard Spectra 70 peripherals - no new peripheral devices were
introduced with the 70/60.
The basic 70/60 processor includes 131, 072 bytes of main memory. which can be increased
in 131, 072-byte increments to a maximum size of 1, 048. 576 bytes. The main memory cycle time is
765 nanoseconds for a four-byte access. the fastest of any Spectra 70 computer. Besides the main
memory. the 70/60 includes a scratch-pad memory. with an access time of 85 nanoseconds per 32bit word. The non-addressable memory used in input-output operations on the 70/35 and highernumbered processors has its 70/60 analog in the "shaded memory"; shaded memory is provided with
the first 131. 072-byte module of each 262 •144-byte bank of main memory and is addressable by two
additional privileged instructions (Load Shaded Memory and Store Shaded Memory). which have been
provided as a part of the general policy of supplying improved diagnostic and maintenance facilities
on the 70/60 compared with the earlier Spectra 70 computers.
The read-only memory in the 70/60 provides an internal control system and consists of
3072 72-bit words with a cycle time of 255 nanoseconds for a 72-bit access. Space is provided for
one emulator; no information is currently available on which machine(s) will be emulated. The emulators already developed for the 70/35 and 70/45 processors cannot be used on the 70/60.
The 70/60 is fully upward compatible with the 70/35 and higher-numbered processors. It
includes the full Spectra 70 instruction set. together with the two extra instructions already mentioned
and the Test and Set instruction included in the 70/45 Type II. The 70/60 basic processor includes a
Multiplexor Channel with up to 16 trunks and two Selector Channels with three trunks each; four more
Selector Channels can be added in two groups of two. The maximum input-output transfer rate is
216.000 bytes per second for the Multiplexor Channels and 900.000 bytes per second for each Selector
Channel. with an overall maximum of 5.240. 000 bytes per second for the whole system. Among the
processor options available are two types of Memory Protect (both write only and read/write. as announced for the 70/45 Mark II). two Elapsed-Time Clocks (decrementing at 16-2/3 or I-millisecond
intervals), and the Direct Control feature.
This report concentrates upon the characteristics and performance of the Spectra 70/60 in
particular. All general characteristics of the Spectra 70 hardware and software are described in
Computer System Report 710: RCA Spectra 70 - General.
The System Configuration section which follows shows the Spectra 70/60 in standard System
Configurations:
III:

lIIR:
IV:
IVR:
VI:
VIIA:
VlIB:
VIIIB:

6-Tape Business System
5-Million-Byte Random Access System
12-Tape Business System
20-Million-Byte Random Access System
6-Tape Business/Scientific System
10-Tape General System (Integrated)
lO-Tape General System (Paired)
20-Tape General System (Paired).

These configurations were prepared according to the rules in the Users' Guide. pages
4:030.120 and 4:200.600. and any significant deviations from the standard specifications are listed.
As a matter of general interest, the rentals that would be incurred if faster magnetic tape units were
installed are listed on the diagrams for Configurations VIlA and VIIB.
Section 718:051 provides detailed central processor timings for the Spectra 70/60. See
Section 718:051 for all the other general characteristics of the Spectra 70 processors.
The input-output channel capabilities of the Spectra 70/60. and the demands upon the processor during input-output operations, are described in Section 718:111.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

9/69

RCA SPECTRA 70/60

718:011.101

Software for the 70/60 is provided by the operating systems already introduced for the
70/35-45-55 processors; the Primary Operating System (POS). the Tape-Tape/Disc Operating System (TDOS) and the Disc Operating System (DOS). The facilities of TOS and TDOS are fully discussed
in Sections 710:151 through 710:193; DOS is a small-scale random-access operating system mainly
intended for 70/35 users and will be of only passing interest to most 70/60 users. An enhanced version of TDOS. OS60. will be provided for the 70/60 but details of this are not yet available. RCA expects that most. if not all. 70/60 installations will use OS60. (TDOS). which offers COBOL. FORTRAN.
and assembly language programming capabilities with multiprogramming of up to six programs and extensive data communications capabilities.
The overall performance of the Spectra 70/60 system is roughly double that of the IBM
System/360 Model 50. The purchase price for the basic processor ranges from $549.900 for the
70/60-F processor with 131, 072 bytes of core storage to $1,663.800 for the 70/60-N processor with
1,048.576 bytes of core storage. The first customer delivery of the 70/60 is scheduled for the second
half of 1970.

9/69

fA

AUERBACH

'"

"

(
\

I

/

SCIENTIFIC
DATA SYSTEMS,. INC.

(

/

~.

/

\_/

AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

AUERBACH
(!)

-

740:000.001
SDS SIGMA 7
REPORT UPDATE

REPORT UPDATE
~ SDS ADDS FAST FIXED-HEAD DISC SYSTEM TO SIGMA LINE

Scientific Data Systems has announced a new Rapid Access Data (RAD) disc-file storage system
for use with its Sigma 5 and 7 computers. The basic configuration consists of a Model 7211
Controller and from one to four Model 7212 Storage Units.
storage Capacity
The Model 7212 contains 64 bands of storage. Each band is made up of eight tracks which are
read and recorded in parallel. The basic addressable unit of data is a sector of 1024 bytes.
Eighty-two sectors are recorded in each band. Each 7212 storage Unit has a capacity of 5.3
million 8-bit bytes.
Performance
The primary attribute of the new RAD system is its high-speed data transfer at a rate of up to
3 million bytes per second. Head positioning delays are eliminated by the use of a fixed read/
write head for each track. The average access time, or rotational delay, required to access
any disc location is 17 milliseconds (34 milliseconds maximum). The rotational delay can be
reduced by a program feature that senses the unit's current rotational position before initiating
a read or write operation, and then transfers to the next sector that can be accessed. This
feature allows a complete band of 83,968 bytes to be read or written in approximately 34 milliseconds, including access time.
'
Data Protect Features
Recorded data is protected by two standard features; manually-set write-protect switches, and
a power fail-safe feature that prevents data loss during periods of transient primary power
fluctuations or primary power loss.
Applications
This system is well suited for high-speed, high-volume applications such as telemetry data
processing, real-time control, data communications and message-switching, and time-sharing
systems. Problems requiring a large number of passes at the data, as in seismic data processing, are facilitated by the speed of RAD. Adequate storage is available for programs and
data files.
Price
The Model 7211 Controller sells for $18,000, and each Model 7212 storage Unit sells for
$50,000.
Initial deliveries are scheduled for the third quarter of 1968.

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

4/68

740:000.001

1&

AUERBACH

e

nUDnD

EDP

sDs SIGMA 7
REPORT UPDATE

UPIIlS

REPORT UPDATE
~SIGMA 7 GETS NEW CARD READER, PLOTTERS, AND DISPLAY UNITS

During the past few months SDS announced the availability of several new peripheral devices for
the Sigma computers with the apparent intention of expanding the applications flcxibility of these
systems.
Model 7140 Card Reader
The SDS 7140 Card Reader rElads standard 80-column punched cards in either EBCDIC or binary
code at a peak speed of 1,500 cards per minute. The 7140 Card Reader is a photoelectric,
serial-by-column reader that includes two program-selectable output stackers and a built-in
controller. The present 1,500-cpm version of the 7140 Card Reader replaces (at no increase in
price) the formerly available 800-cpm version.
The 7140 occupies one control unit position in the 8471 Multiplexor Input-Output Processor.
Data is transferred to the central processor one byte at a time. When reading in the EBCDIC
mode, the reader transfers one 8-bit byte for each 12-bit column read; each character is
automatically checked for validity prior to transmission. When reading is performed in binary
mode, the readers transmit three 8-bit bytes to the computer for every two 12-bit columns read.
Deliveries of the new 7140 Card Reader are scheduled to begin during the fourth quarter of 1967.
Models 7530 and 7531 Graph Plotters
The SDS Graph Plotters provide the capability to plot digital data in two axes. The plotters
operate on the digital incremental principle using decoded commands from the central processor.
The incremental ink-on-paper plotting steps can be performed in either direction along both axes.
The basic motions used in plotting are: movement of the pen horizontally across the surface of
the drum, drum rotation (with resulting paper motion), and pen up and down.
The Model 7530 Graph Plotter includes a 12-inch wide drum with a maximum plot size of 11
inches by 120 feet; incremental speed is 300 steps per second. The Model 7531 Graph Plotter
includes a 30-inch-wide drum with a maximum plot size of 29.5 inches by 120 feet; incremental
speed is 200 to 300 steps per second. Both models offer a choice of the following step sizes:
0.01 inch, 0.005 inch. or O. 1 millimeter.
Each incremental command from the Sigma 7 central processor (via the standard Multiplexor
or Selector Channel) is transmitted as one 8-bit byte. The byte configuration specifies paper
and/or pen movement and the direction of this movement. A single Start Input/Output (SIO)
command can specify that a virtually unlimited string of bytes be used as consecutive incremental
commands.
The SDS Graph Plotters are manufactured by Calcomp.
Models 7550 and 7555 Multipurpose Keyboard Display Units
The SDS Model 7550/7555 Multipurpose Keyboard Display Unit is a self-contained buffered I/O
device featuring keyboard input and alphanumeric CRT output capabilities for use at local or
remote locations. This device should prove invaluable to the Sigma computers in time-sharing,
inquiry/response, and text editing applications.
The Model 7550 is a low-speed keyboard display device capable of transmitting data at a 15character-per-second maximum rate; the Model 7555 is generally identical to the 7550 in
characteristics and functions, but, with an optional feature, can transmit data at speeds up to
180 characters per second.
The video unit has a 7- by ll-inch screen, organized in 32 lines of 86 characters each. Character display is refreshed at a rate of 50 frames per second to ensure flicker-free viewing. A
self-contained 2,048-character buffer enables a full page of text to be displayed and manipulated
without excessive demands on the central processor. Through use of versatile cursor control
and other standard features, the display device has the capability to replace, erase, or roll
text; insert or delete data from any point in the text; and to combine or separate lines of displayed data. A half-line spacing capability facilitates the use of subscripts, exponents, and
underscores.
The keyboard is compatible with the standard Teletype Model 37 KSR key set; it can generate
96 ASCII graphic and 32 ASCII control characters. Special keys permit generating frequently
used control codes, and an automatic-repeat function permits repetition of data being entered
(at a rate of approximately 15 characters/second).

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

10/67

740:000.002

SOS SIGMA 7

The standard Model 7550/7555 Display can I?e operated in full-duplex mode over standard data
communications networks, eventually connecting to a centrally located SDS 7611 CharacterOriented Communications Controller (COC). Up to I, 024 Keyboard Di"splay Units can operate
concurrently over lines controlled by the maximum complement of 16 Sigma 7 COC's - 64 lines
per COCo
As an optional feature, either model of the display unit can use feature 7551 - Message-Mode
Option - which enables the operator to prepare messages through use of the Transmit code.
Feature 7153, the Hard-Copy Option, is also available to either display tinit model. This feature
permits a Teletype "receive only" printer unit to be attached to the 7550/7555 Keyboard Display
to provide hard-copy records of the visually displayed data.
Feature 7556, the High-Speed Option, is available only for Model 7555 Keyboard Display Units.
This option permits data transmission at speeds up to 180 characters per second. Type 202
Data Sets provide the interface with remotely located Keyboard Display Units that are equipped
with the High-Speed transmission feature.
PRICES

Unit
7140

Card Reader, 1500 cpm

7530
7531

Monthly
Maintenance

Purchase

$665

$540

$150

$24,000

Graph Plotter
Graph Plotter

415
?

340
?

75
?

13,000
?

7550
7555
7551
7553
7556

Keyboard Display
Keyboard Display
Message Mode Feature
Hard-Copy Option
High-Speed Option

280
350
23
17
45

225
285
18
13
35

50
65
5
N/C
10

10,000
12,500
750
500
1,500

7611

Character-Oriented Communication
Controller (controls 8 fullduplex lines)
Line Interface Unit (up to 7 7613's
can be added to 7611 to control
64 full-duplex lines)

284

241

45

10,500

44

36

N/C

1,500

7613

10/67

Monthly
Rental
1-yearlease 4-yearlease

A•

AUERBACH

740:001. 010

A

$lUBAID

EDP

AUERBACH

SOS SIGMA 7
SUMMARY REPORT

IEPOR1S

SUMMARY REPORT:
SDS SIGMA 7
.01 INTRODUCTION
.011 Sigma 7
Scientific Data Systems announced a new computer system, Sigma 7, on March 15, 1966.
Sigma 7 was heralded as the first of a new family of Sigma computers from SDS that would provide "at least two times more computations per dollar than any other machine in the industry".
With the announcement of the general-purpose Sigma 7, SDS has begun a gradual expansion of
its marketing goals to include not only scientific-oriented computing (where SDS has earned a
fine reputation), but also business data processing.
Sigma 7 is a medium -scale computer that is compatible with the IBM System/360 in internal
data structure, external data codes, input-output media, and FORTRAN and PL/I languages.
Source language compatibility is facilitated by internal fixed-point and floating-point arithmetic
formats that are virtually identical with those used in the System/360 processors.
Sigma 7 makes extensive use of monolithic integrated circuitry in a central processor whose
internal design is radically different from that of the IBM System/360. Monthly rental prices
for typical Sigma 7 system configurations range from $5,000 to $25,000. Core storage
sizes range from 4, 096 to 131,072 32-bit words, with a cycle time of 1. 2 microseconds per
word.
Sigma 7 is a general-purpose, highly modular system designed to function in a wide variety of
application areas and in several different processing modes. The primary design goal is to
produce a fast, responsive real-time system that can provide the full services of the computer
to multiple user programs. In small-scale configurations, Sigma 7 can function as a relatively
inexpensive but powerful scientific processor that executes one program at a time. In somewhat larger configurations, Sigma 7 can serve as a medium-scale business/scientific system
capable of multiprogrammed processing of one "background" production program and one
"foreground" real-time inquiry program. Configurations with mass storage devices and at
least 12K words of main memory can provide full hardware/software control of the operating
environment and multiprogrammed operation for three concurrent programs. In large-scale
disc-oriented configurations, Sigma 7 can handle remote, interactive time-sharing operations
for up to 200 competing users, concurrently with processing background production programs.
Sigma 7 hardware and software also permit multiple central processors to share common
core storage and peripheral units.
The central components of every Sigma 7 system - central processor, core storage, and
I/O control system - feature flexibility, expandability, and capability for asynchronous
independent operations. The instruction set is large and powerful, and the input-output
system (which can include up to eight channel controllers of the selector and/or 32-subchannel
multiplexor variety) is comparable to the I/O systems in higher-priced, large-scale computers. Up to eight core storage modules are capable of independent operation, and up to
six of the modules can be accessed simultaneously. Sigma 7 core storage is large in capacity
(up to 524,288 bytes) and among the least expensive in the industry.
SDS currently offers a limited number and variety of input-output devices for use with Sigma 7,
although it is expected that interface units will be announced to permit connection of I/O
devices from other manufacturers. The Price Data section (page 740:221.101) lists the
current peripheral devices and their rated speeds. At present SDS offers one Sigma 7 mass
storage device: a!. 5 million-byte, fast-access unit of comparatively high price. However,
SDS has indicated that several low-cost, head-per-track disc files of various capacities and
speeds are under development and due for release in the near future. Apart fro,!D- manufacturing
its own disc files and magnetic tape units, SDS does not appear interested in competing at this time
in the development of a broad range of special-purpose peripheral devices.
Software for Sigma 7 is provided at four levels, all upward-compatible, and features real-time
multiprogramming and disc-oriented operating systems. Table I summarizes the software
facilities and their availability dates. FORTRAN IV and PL/I compilers will be supplied
in three different versions: debug, high-efficiency, and conversational. Assemblers will be
provided with the first systems delivered, beginning in the fourth quarter of 1966. Although
the availability of a COBOL compiler was not annQunced in the earliest Sigma 7 software

© 1966 AUERBACH Corporation and AUERBACH Info. Inc.

9/66

740:001. OtT

SOS SIGMA 7

.011 Sigma 7 (Contd.)
schedules, it is expected that SDS will shortly announce a COBOL, probably supplied on a lease
or purchase basis. An IBM 1401 Simulator program is also expected to be announced in the
near future. SDS states that the Sigma 7 software development effort began almost two years
ago and that all published delivery schedules are being adhered to. More than half of the
initially scheduled software systems will be written by outside software contractors.
TABLE I: SIGMA 7 SOFTWARE AVAILABILITY
A vailablhty

Name

Class

Bastc Control Momtor

Momtors

Batch Processmg Momtor
Universal Tlme-Sharmg Monitor"
Standard versIOn
Extended version

1st Qtr. 1967
2nd Qtr. 1967
4th Qtr. 1967
1st Qtr. 1968

BaslC FORTRAN IV
Debug FORTRAN IV
HIgh Eff,ciency FORTRAN IV
Conversational FORTRAN IV

4th Qtr.
2nd Qtr.
2nd Qtr.
4th Qtr.

1966
1966
1967
1967

Debug PL/I
H,gh Eff,ciency P L/I
Conversational P L/I

3rd Qtr. 1967
4th Qtr. 1967
4th Qtr. 1967

COBOL

to be announced

Assemblers

Bastc Symbol Assembler
Meta-Symbol Assembler

4th Qtr. 1966
2nd Qtr. 1967

ServlCes

ADAPT Application Package
Sort/Merge
MANAGE with RPG
IBM 1401 Simulator
ApplIcation Programs

2nd Qtr. 1967
to be announced
to be announced
to be announced

ComplIers

to be announced

Price/performance comparisons between Sigma 7 and the IBM System/360 Model 50 indicate
that in comparable central configurations (i. e., with equivalent central processors, core
storage, and I/O control facilities), Sigma 7 is approximately 10 to 20 per cent less expensive
than the Model 50 and the basic processing power of Sigma 7 is approximately 40 to 50 per cent
greater than that of the Model 50. There are indications that this advantage in basic processing
speed will increase still further, as SDS contemplates improving the Sigma 7 core storage unit
by reducing the cycle time from 1. 2 to 0.9 microsecond per 32-bit word .
. 012 Sigma 2
On August 1, 1966, Sigma 7 officially became a family member when SDS announced the smallscale Sigma 2 computer system. Sigma 2 is a low-cost computer system designed for SCientific,
engineering, and process control applications. Sigma 2 has good real-time processing capabilities
and hardware facilities that will permit multiprogrammed operation of a background production
program and a foreground real-time program. The purchase price of a basic Sigma 2 configuration.(consisting of a processor with four I/O channels, 4,096 words of core storage, and a keyboard/printer device with slow-speed paper tape reader and punch) is $26,000; the same
configuration under terms of the standard 4-year lease contract rents for $875 per month.
Deliveries of the Sigma 2 systems are expected to begin during the first quarter of 1967.
Sigma 2 contains an internal core storage unit that ranges in size from 4,096 to 16,384 16-bit
words. Core storage access time is 0.9 microsecond per word. The core storage capacity can
be increased to 65,536 words by the addition of Sigma 7 memory modules. Thus the Sigma 2 can
share core storage with the Sigma 7, permitting multiprocessing operations. Real-time processing is facilitated by an interrupt system than can service up to 148 different interrupt levels,
and memory protection is available to safeguard programs and data in core storage.
The instruction set of the Sigma 2 is limited to 35 standard 16-bit instructions, with multiply and
divide instructions optional. All arithmetic is performed in fixed-point binary format, and no
radix conversion nor code translation instructions are provided. Add, subtract, load, and store
instructions can be executed in 2.25 microseconds, and 16-bit binary multiply in 10.35 microseconds. Comprehemlive software that will be provided to utilize the Sigma 2 computation speeds
includes two monitor-controlled operating systems, a Basic FORTRAK and a FORTRAN IV compiler, two assemblers, and a number of library and utility programs. Consistent with the design
and scope of Sigma 2, no business data processing software will be provided. Sigma 2 can use all
of the peripheral units announced for use with Sigma 7 (and listed in the Price Data section). There
is no program compatibility between the two current Sigma systems.
The SDS Sigma family is expected to increase by the addition of still another computer system
within a few months. The new system will probably be smaller and less expensive than the
Sigma 7, but completely compatible.
9/66

fA

AUERBACH

'"

740:221. 101

J&•

AUERBACH

SUHOARD

ED]?

SDS SIGMA 7
PRICE DATA

REPORTS

PRICE DATA

IDENTITY OF UNIT
CLASS
Name

No.

PRICES
Monthly Rental
$
I-year lease

SIGMA 7
CENTRAL
SYSTEM

8401

8411
8413
8414
8415
8416
8418
8419
8421
8422
8495

Proces SOl' 0Etions
Two Additional Real-Time Clocks
Power Fail Safe
Memory Write Protect
Memory Map
AdditIonal Register Block
Floating Point Arithmetic
Decimal ArithmetIC
Interrupt Control Chassis
Priority Interrupt, 2 levels
System Supervisory Console

8451
8452
8456
8457

Core Storage
Memory Module: 4,096 words
Memory Increment: 4,096 words
Three-Way Access
Six-Way Access

OUTPUT
DEVICES

INPUTOUTPUT
DEVICES

Purchase

$

$

3,030

2,475

450

1l0,000

:30
30
140
665
85
835
1,000
60
10
695

20
20

1,000
1,000

540
65
675
810
50
8
565

5
5
20
80
10
100
120
10
NC
100

20,000
2,500
25,000
30,000
2,200
350
25,00U

495
130
335

900
400
115
270

160
40
20
50

40,000
17,500
5,000
10,000

I. 11 0

115

5, 000

f
555

450

80

20,000

8472
8481
8482

In[>ut-Out[>ut Processors
Multiplexor Input-Output Processor,
with 8 Multiplexor Channels
Additional 8 Multiplexor Channels
Selector Input-Output Processor
Additional Selector Channel

130
415
280

110
340
215

15
60
40

4,000
15,000
10,000

7120
7140

Card Reader, 400 cpm
Card Reader, 800 cpm

445
665

360
540

100
150

16,000
24,000

7061

200

610

30

7,000

7062
7064

Paper Tape Controller and Equipment
Cabinet
Paper Tape Reader, 300 cps
Paper Tape Spooler

55
45

45
35

15
10

2,000
1,500

7160

Card Punch, 300 cpm

890

720

210

32,000

7063

Paper Tape Punch, 120 cps

65

55

25

2,500

7440
7445

Buffered Line Printer, 600lpm
Buffered Line Printer, 1,0001pm

970
1, 110

790
900

230
255

35,000
40,000

7010
7020

Keyboard/Printer
Keyboard/Printer, with
Paper Tape Reader and Punch

165
220

135
180

35
50

6,000
7,500

7060

Paper Tape Reader (Model 7062),
with 7063 Paper Tape Punch,
7064 Spooler, and 7061 Controller

335

270

80

12,000

220
745

180
610

35
155

8,000
27,000

280
720

225
585

40
185

10,000
27,000

335
720

270
585

50
185

12, 000
27,000

8471

INPUT
DEVICES

Central Processor
Sigma 7 Central Processing Unit with
2 real-time clocks, control panel,
and power sllpplles

4-year lease

Monthly
Maintenance

7201
7205

Mass Storage Devices
RAD Controller
RAD Storage Unit, 1. 5 million bytes
Ma~etic

7321
7322
7371
7372

Tape Units
Magoetic Tape Controller
Magoetic Tape Unit, 9-channel,
800 bpi, 60KB
Seven-Channel Magoetic Tape Controller
Magoetic Tape Unit, 7-channel,
200/556/800 bPI, 60KC max.

© 1966 AUERBACH Corporation and AUERBACH Info, Inc.

9/66

\

/

UNIVAC DIVISION
SPERRY RAND CORP.

(
\

/

\_j

AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

AUERBACH

~

770:011.100
UNIVAC 1004
Introduction

INTRODUCTION

Ii 011.
The UNIVAC 1004 is a compact, plugboard-programmed computer. Its two basic models, the
1004 I and 1004 II, can process punched card input at speeds of about 340 and 600 cards per
minute, respectively, including the necessary allowances for a typical amount of computation
and for I/O interlocks. (Card reading and printing can proceed simultaneously, but cannot be
overlapped with computation. )
Rentals for the basic 1004 system (consisting of processor, card reader, and printer in a single
cabinet) range from $1,150 to $1,625 per month. Additional peripheral equipment that can be
connected includes a card punch, a second card reader, a card read/punch unit, paper tape
equipment, and data communication terminals. A special processor model, the 1004 ro, can
control one or two magnetic tape units in addition to the above equipment.
First deliveries of the UNIVAC 1004 I were made in January 1963, and over 1,300 systems
have been installed to date. The faster 1004 II and 1004 III were announced in March 1964.
The 1004 is most commonly used as an independent data processing system for small business
applications. As such, it is attractive to many organizations considering a step upward from
conventional tabulating installations because the 1004 requires less retraining of their staffs
than a stored-program computer system would require. Furthermore, the 1004 offers economic
advantages over stored-program systems for many applications whose processing and internal
storage requirements are relatively small.
The 1004 can serve as a satellite system for a larger computer, such as a UNIVAC 490 or 1107.
It is also suitable for use as a small computer in a branch office, communicating with a larger,

home-office computer either by means of direct communication lines or simply through physical
interchange of card decks or tape reels.
Where the 1004 is used as a complete data processing installation, there is no larger programcompatible computer system into which the installation can grow as its workload increases.
UNIVAC, however, has announced provisions for connecting a 1004 system to its larger 1050
series of computers (described in Computer System Report 777:). The 1050 can then be used
in conjunction with, and perhaps eventually replace, the 1004.
The UNIVAC 1004 can be used with more than one coding system. It normally operates with
either the standard UNIVAC XS-3 code or with the Remington Rand 90-column card code. Which
code is to be used is program-selectable, so it is possible to use both codes within a single program. This allows, for instance, reading a mixture of SO-column and 90-column cards, or
reading SO-column cards and punching 90-column cards.
Codes other than the XS-3 and 90-column codes can be automatically translated to either of these
codes by a special Translate Feature, provided that there are no more than 6 data bits per character in the original code. In particular, the IBM BCD codes used on the 1401 and other IBM systems can be translated, thus allOwing the 1004 to be used as a satellite to many non-UNIVAC
computer systems.
The UNIVAC 1004 has 961 alphameric character positions of core storage. Each core position
contains six data bits. Core storage cycle time is S. 0 microseconds in the UNIVAC 1004 I and
6.5 microseconds in the UNIVAC 1004 II and III.
The plugboard of the basic machine has a capacity for 31 program steps (expandable to 62).
Each step can specify two operand addresses, and multiple operations can be performed in a
single program step. Arithmetic operations include add and subtract (both algebraic and
absolute) and compare. Multiply and divide operations require the use of wired subroutines.
Seven types of transfer processes are provided, including several with editing facilities. Inputoutput areas are assigned fixed locations in core storage. Input-output commands can be
combined in the same step with other operations.
Operands can be of any length up to the capacity of core storage. Operand length is specified by
the operand addresses wired in each program step. Instructions are executed at the rate of
about 6,500 instructions per second in the 1004 I proCElssor and about S, 000 instructions per
second in the 1004 II and III.

© I 964 Auerbach Corporation and Info, Inc.

S/64

UNIVAC 1004

770:011.101
§ 011.

INTRODUCTION (Contd.)

The 1004 can read cards and print simultaneously, but neither of these operations can be overlapped with computation. Card punching can overlap either computing or other peripheral
operations. The optional peripheral devices may:
(1) be able to overlap both computing and card reading and/or
printing (e.g., the paper tape punch or the card read/punch
operations);
(2) be able to overlap computing but not card reading or printing
(e.g'., the auxiliary card reader or the paper tape reader); or
(3) be unable to overlap any other operation (e. g., the Data Line
Terminals).
The 1004 is available in 80-column, 90-column, or 80/90-column models. The basic system
consists of a card reader, central processor with plugboard control, and printer. All are
housed in a single compact cabinet. The card reader in the 1004 I Processor has a rated
speed of 300 cards per minute, and the printer has a rated speed of 300 lines per minute.
These rated speeds include an allowance for 35 milliseconds of computation per card or line,
which has been found to be quite conservative. In typical applications, computation time is
about 5 milliseconds per card, and reading and/or printing speeds of about 340 cards/lines
per minute are obtained.
In the 1004 II and III, the card reader operates at a speed of 615 cards per minute, and the

printer operates at 600 lines per minute; both these speeds are based on 5 milliseconds of
computation per record.
A card punch can be connected to the UNIVAC 1004. It punches at a speed of 200 cards per
minute. The card punch is available in a read/punch model which reads and/or punches
cards at a speed of 200 card~p r ,minute. The read/punch enables a 1004 system to read
data from and punch results i to the same card. A 400-card-per-minute Auxiliary Card
Reader can also be used wit the 1004 Processor.
Two different Data Line Terminals are available. The Data Line Terminal, Type I, can be
used to communicate with it UNIVAC 1050, 490, 1107, or another 1004. The Data Line Terminal, Type 2, permits communication with magnetic-j;ape terminals such as the Digitronics
Dial-O-Verter.
A 400-character-per-second'paper tape reader and a 1l0-character-per-second paper tape punch
can be used with the 1004.
One or two Uniservo magnetic tape units can be connected to the UNIVAC 1004 III processor only.
Three density levels - 200, 556, and 800 pulses per inch - provide speeds of 8,000, 23,000,
and 34,000 characters per second, -respectively. These magnetic tapes can be written in a mode
compatible with either UNIVAC or IBM standards, although programmed translation may be required.
The software available with the 1004 is naturally limited. It consists primarily of short subroutines for handling multiplication, division, and a number of common commercial problems.
These include suggested methods for handling reconciliations, deleting subtotals where there
has only been a single card to be totaled, handling missing numbers in a matching operation,
checking the sequence of alphanumeric identification numbers, and verifying check digits.
In addition, a number of complete programs are available.

These include standard listing and
transcription programs, and at least one General Purpose Program, which is a report writer
that can facilitate setting up the equipment for new reports. A start has been made on supplying
some scientific routines, such as sine-cosine and square root routines, and a Critical Path
Method routine has been announced.
Software routines are circulated by the UNIVAC Division to 1004 users.

8/64

6

l~~I~\

770:221.101
UNIVAC loeM
PRICE DATA

UNIVAC 1004
UNIVAC does not include monthly maintenance charges in their published prices; tllese charges have been
included in the Monthly Rental column in this Digest to permit convenient comparison with other systems.

PRICES

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

PROCESSOR

Name

Monthly
Monthly
Rental Purchase Maint.

$

$

$

Processing Unit

*2010-00
*2012-74
*2012-72

*2010-06
*2010-86
*2012-70
*2012-68
*F0743-00
*F0743-01
*F0743-02
*F0594-00
F0595-00
*F0624-00
*F0624-97
F0675-00

*2010-06
*2012-70
*2012-68
*F0743-00
*F0743-01
*F0743-02
*F0594-00
F0595-00
*F0624-00
F1)675-00

1004-I-Processors (include 400-line/min
printer, 400-card/min reader, and 8
IJsec core memory):
1004- I - Processor (31, 30, 10, 40, 80
45, 6)(1)
1004-1 - Processor (47, 45, 15, 65, 120,
63, 8)(1)
1004-1 - Processor (62,60,20,105,160,
80,10)(1)
1004-U Processors (include 600-line/min
printer, 615-card/min reader and 6.5
~sec core memory):
1004-U-Processor (31,30,10,40,80,45,
6)(1)
Card Processor Mod U C
1004-U-Processor(47,45,15,65,120,
63,8)(1)
1004-U-Processor (62,60,20,105,160,
80, 10)(1)
Features for 1004 I and n Processors:
Selectors (group of 5)
Program Selects (group of 5)
Collectors (group of 5)
ExpaIl@ion Set ~'iands 2010-00 to 2012-74
or 2010-06
012-70)
Expansion Set (Expands 2012-74 to 2012-72
or 2012-70 to 2012-68
Expansion Set (Expands 2010-00 to 2010-06,
2012-74 to 2012-70, or 2012-72 to 2012-68)
Expansion Set (Expands 1004-IC to 1004-U C
Memory Expansion (Expands memory to
I, 922 characters)
1004-m Processor (includes 600-line/min
printer, 615-card/min reader, and 6.5 Usec
core memory):
1004 - m Processor (31, 30, 10, 40, 80,
45, 6)(1)
1004 - m Processor (47, 45, 15, 65, 120,
63, 6)(1)
1004- m Processor (62, 60, 20, 105, 160,
80, 10)(1)
Features for 1004 m Process9rs:
Selectors (group of 5)
Program Selects (group of 5)
Collectors (group of 5)
Expansion Set (expands 2010-06 to 2012-70)
Expansion Set (expands 2012-70 to 2012-68)
Expansion Set (expands 2010-00 to 2010-06,
2012-74 to 2012-70, or 2012-72 to
2012-68)
Memory ExpanSion (expands memory to
I, 922 characters)

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

1,200

46,000

225

1,460

56,000

250

1,565

60,000

260

1,335

51,000

300

2,045
1,595

79,500
61,000

390
325

1,700

65,000

335

5
15
5
260

200
600
200
10,000

--

-25

105

4,000

10

135

5,000

75

125
70

6,200
2,600

65
5

1,335

51,000

300

1,595

61,000

325

1,700

65,000

335

5
15
5
270
105
135

200
600
200
10,000
4,000
5,000

10
75

70

2,600

5

-25

2/69

UNIVAC 1004

770:221.102

IDENTITY OF UNIT
CLA."S

Model
Number

Feature

INPUTOUTPUT

Monthly Purchase Monthly
Rental
Maint.

Name
Magnetic Tape (1004

*0858-04

PRICES

m Processors

only)

Mag Tape Unit and Control (Uniservo VI C;
7-Track 8,540, 23,741, or 34,160 char/sec)
Mag Tape Unit and Control (Uniservo A;
8-Track)
Mag Tape Unit (Second Uniservo VI C unit)
Mag Tape Unit (Second Unlservo A unit)

"'7915-02
0858-01
"'7915-04

515

17,350

115

500

17,700

115

310
300

10,470
10,680

70
70

160
260

6,000
10,000

20
35

190
27
25
315
470
300
15
28
5
155

7,200
1,000
960
12,000
18,000
11,760
600
1,000
200
6,000

40
7
10
100
150
90

35
35

1,200
1,200

5
5

210
300
25
210
290
210

8,000
11,600
1,000
8,000
11,600
8,000

60
60

-60

575
690
210
1,150
685
375
375
350
400

22,000
26,000
8,000
44,000
29,250
15,000
15,000
14,000
16,000

165
60
60
330
200
55
55
50
60

Paper Tape
"'0902-00
"'F0606-00

Paper Tape Reader (400 char/sec)
Paper Tape Punch (110 char/sec)
Punched Card

"'0704-00

F0591-00
F05S8-00
F0592-00
"'F0620-00

Card Reader (400 cards/min)
Code Image Read
Invalid Code Check
Card Punch (200 cards/min)
Card Read/Punch (200 cards/min)
Card Punch (90 Column, 200 cards/min)
Double Punch Detector
Code Image Punch
Punch Stacker Select
Punch Read Conversion

"'F0757-00
"'F0793-00

Other
Adapter (1004/Card Controller)
Adapter (DLT-3/Card Controller)

F0587-00
F0785-00
2009-00
"'2009-01
*2011-00

COMMUNICATIONS

"'F0585-00
F0753-00
"'F0753-01
"'F0611-00
"'F0765-00
F0792-00
"'F0555-00

"'F0555-01
"'F0556-01
"'F0863-00
*FOS63-01
"'F0863-02
*FOS63-03

Data Line Terminal (DLT-l)
Data Line Terminal (DLT-IB)
Data Line Terminal Option (low speed)
Data Line Terminal (DL T-2)
Data Line Terminal (DLT-2B)
Data Line Terminal (DL T-3)
Dat:1 Line Terminal
DLT 5(1004/XS-3, 80 character blocks)
DLT 6(1004/AUTODIN)
D L T 7 (1005/XS-3, continuous)
Dual DLT 5 (two DLT-5's)
Data Line Terminal (DLT-8)
Data Line Terminal (DLT-9)
Data Line Terminal (DLT-9)
Data Line Terminal (DLT-9A)
Data Line Terminal (DL T-9B)

NOTES:

'(1)
"

2/69

No longer in production.
The numbers in parentheses indicate the munber of Program Steps, Selectors, Program
Selects, Collections, Distributors, Address Combines, and Comparators, respectively.

A.

AUERBACH

-5
-

50
I
I

55
60

771 :011. 100

1&.

SIAMIIARO

EDP

UNIVAC 55 80/90 MODEL I
SUMMARY

REPORTS

AUERBACH

SUMMARY: UNIVAC SS 80/90 MODEL I
.1

BACKGROUND
The UNIVAC Solid-state 80/90 Model I system, introduced in 1958, uses a magnetic drum
main memory. As a result, its internal processing speeds are significantly lower than
those of most of the newer, third-generation computers. The Solid-State 80/90, however,
included two significant advances over earlier drum-oriented computers: solid-state circuitry, which reduced installation and maintenance cpsts, and a fully buffered card reader,
card punch, and line printer, which provided increased simultaneity.
The Solid-State 80 is designed to use standard SO-column cards, while the Solid-State 90
uses Remington-Rand 90-column cards. Therefore, there are differenc.es in the I/O
buffering and in the code translation instructions. In most other' respects, including price,
the two types of equipment are identical.
The Solid-State 80/90, as originally announced, was a punched-card-oriertted computer
system with a fixed drum storage capacity of 5000 ten-digit words. Four signi1ficant
improvements during the next few years were the introduction of:
•

MagnetiC tape input-output.

•

Large-capacity RANDEX drum storage.

•

Main memory drums of varying storage capacities, ranging from 2600 to 9200 words.

• The Solid-State 80/90 Model II Central Processor, which features a combination of
drum and core storage (see Report 772).
First deliveries of Solid-State 80/90 Model I 1;ystems were made in August 1958. The
faster (and more expensive) Solid-State 80/90 Model II Processor, described in the
following report, was initially delivered in June 1962. A total of more than 500 Solid-State
SO/90 systems were installed, and approximately 100 are still in use at this writing. Both
systems went out of production in 1965 .
.2

HARDWARE
The Solid-state 80/90 Model I Processor handles data in words of 10 digits plus a sign.
Each digit consists of four information bits plus an odd parity bit. Parity is checked during
all data movements to or from storage.
The central processor normally handles only numeric data, whereas the peripheral units
handle alphanumeric data in Hollerith or 90-column card code. This difference is resolved by
splitting each group of 10 alphanumeric characters into two computer words, with one word
holding the zone bits and the other word holding the numeric bits for all of the 10 characters.
The processor has three one-word arithmetic registers, designated A, X, and L. Register
A is the accumulator, and Registers A and X together form a double-length register that is
used in shifting and multiplication operations.
Programming for the Solid-State 80/90 Model I is similar to programming for the widelyused IBM 650. As in most drum-memory computers, coding is complicated by the
desirability of locating instructions and operands in "optimum" drum locations that will
minimize lost time due to latency delays.
The instruction format is "one plus one." Each instruction occupies one word and consists
of a two-digit operation code, two four-digit addresses, and a one-digit index register
designator. The first address normally specifies the location of an operand, while the
second specifies the location of the next instruction to be executed. Three index registers
can be used to modify the first address in an instruction; the second (or next-instrUction)
address cannot be modified by indexing.
There are 62 instructions, including fixed-point arithmetic, comparison, logical masking,
right and left shift, zero suppress, and automatic code translation instructions. All of the
instructions operate on full words. Character manipulation is performed by a combination
of shifts and logical AND and OR instructions. Multiplication and division instructions are
an extra-cost option on the modular STEP (Simple Transition to Economical Processing)
processor and a standard feature on the larger processors. Program Interrupt is an
optional feature for all processors.
Data is stored on the magnetic drum in 200-word "bands" of two types. Each so-called
"fast" band is served by one read/write head and has access times of 0 to 3.4 milliseconds;
the average for randomly-located data is 1. 7 milliseconds. Each "high-speed" band is
© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

12/67

771 :011. 200

.2

UNIVAC SS 80/90 MODEL 1

HARDWARE (Contd.)
served by four read/write heads spaced equidistantly around the drum's circumference, so
the access time ranges from 0 to 0.875 millisecond. These access times are much longer
than the basic instruction execution times; addition, for example, takes only O. 085 millisecond. Thus, there is a great deal to be gained through optimum allocation of the
instructions and operands so as to minimize drum access times. This optimization is
largely an automatic function of the UNIVAC-supplied assembly routines, X-6 and S-4.
Solid-state 80/90 Model I Central Processors are available with three basic capacities of
drum storage, as summarized in the following table.
Processor Model

"Fast" Storage

"High-Speed" storage

STEP
Standard
Expanded

2400 words*
4000 words
7600 words

200 words**
1000 words
1600 words

* Expandable in increments of 400 words; maximum of 4 increments.
** Expandable in increments of 200 words; maximum of 4 increments.
A central processor, a 600-cpm reader, a 150-cpm card read/punch, and a 600-1pm
printer make up the basic Solid-state 80/90 system. Other units that can be included are:
up to 10 magnetic tape units, up to 10 RANDEX drums, and a 500-cps paper tape reader
and 100-cps paper tape punch. The tape units and/or RANDEX drums are controlled by a
Tape- RAND EX Synchronizer.
Effective drum buffering of I/O operations enables the card reader, punch printer, and one
tape unit or RANDEX drum to operate simultaneously with computing. Simultaneous
magnetic tape reading and writing, however, is not possible.
The card reader operates at 600 cards per minute and is available in 80- and 90-column
versions. Each card is read at two separate read stations. The two readings are not
compared in the reader itself, but both images are transferred to the drum and compared
there for accuracy.
The card read/punch operates at 150 cards per minute, has a single card feed path, and is
available in 80- and 90-column models. The unit can include two read stations, one that
reads before and one that reads after punching. All checking must be performed by the
stored program.
The line printer, rated at 600 lines per minute, uses a conventional drum-type, on-the-fly
printing mechanism. There are 51 printable characters, and the printed line can be 100,
110, 120, or 130 characters in length. There is no paper tape loop, so all vertical format
control must be effected by the stored program.
Up to 10 Uniservo magnetic tape units can be connected, though only one unit can read or
write at a time. The tape units are rated at 25, 000 characters per second, but the fixed
block length (normally 1100 characters) leads to a maximum effective speed of 16,400
characters per second at the maximum density of 250 bpi. Alternative block lengths of 720 or
120 characters and/or a lower density of 125 bpi lead to even lower 'effective speeds. A row
parity check is performed upon reading operations.
Up to 10 RANDEX drum units can be used to add random-access storage to a Solid-state
80/90 system. Each RANDEX unit stores 24 million digits and consists of two large
drums, mounted one above the other and served by a single movable read/write head
assembly. Drum revolution time is 70 milliseconds, and head repositioning (when
required) takes from 125 to 540 milliseconds. Records composed of 48 ten-digit words
(one RANDEX sector) can be accessed within 600 milliseconds regardless of their position
on the drums .
.3

SOFTWARE
Development of software for the Solid-state 80/90 systems suffered from a number of false
starts. The Model I was originally advertised as being designed for programming in
UNIVAC's Flowmatic language (a forerunner of COBOL), but the Flowmatic compiler did not
live up to expectations. A COBOL-60 compiler was also announced for the system but was
later withdrawn. FORTRAN I and II compilers were made available to users in field-test form
but were never formally released.
UNIVAC-supplied software for the Solid-State 80/90 systems includes two assemblers (X-6 and
S-4) , a sort routine, about 25 I/O control routines, some program testing routines, a group of
mathematical functions and routines, a linear programming package, and a CPM system.
X-6 is a basic symbolic assembly system for the Model I, while S-4 is a somewhat more
advanced assembler that can be used on both models. X-6 is a card-oriented assembler,
while S-4 can use magnetic tape units, when available, to speed up the assembly process.
Neither assembler provides facilities for macro-instructions, though S-4 permits the
creation and use of a program library on magnetic tape. Programs coded in the X-6
symbolic language can, with minor revisions, be assembled by the S-4 translator.

A

12/67

(Contd.)

..

AUERBACH

771 :221.101
"'1.111

aDP
1,".11

UNIVAC 55 80/90 I
PRICE DATA

UNIVAC 55 80/90 I
Manufacture of this equipment has been discontinued. Orders will be accepted for available equipment on
hand. The list purchase prices below are shown for informational purposes only. Selling prices will be
quoted by UNIVAC upon request.

PRICES

IDENTITY OF UNIT
CLASS

Model
Number

Monthly
Monthly
Rental Purchase Maint.

Name

$(1)

$

$

Processing Unit (includes main memory)

PROCESSOR
7944
7934
7947
7937

7907
7909
7933
7913

7940
7930

MASS
STORAGE

Feature
Number

7965
7957
7966

Step Central Processors:
90-Column Card Only
80-Column Card Only
90-Column Card and Tape
80-Column Card and Tape
Standard Equipment:
2400 words of fast drum storage
200 words of high speed drum storage
3 Index Registers
Options:
Program Interrupt
Multiply and Divide
400 Additional Words of Fast
Drum Storage (1600 words max.)
200 Additional Words of High-Speed
Drum Storage (800 words max. )
Standard Central Processors:
90-Column Card Only
80-Column Card Only
90-Column Card and Tape
80-Column Card and Tape
Standard Equipment:
4000 words of fast drum storage
1000 words of high-speed drum storage
Multiply and Divide
3 Index Registers
Option:
Program Interrupt
Expanded Central Processors:
90-Column Card and Tape
80-Column Card and Tape
Standard Equipment:
7600 words of fast drum storage
1600 words of high-speed drum storage
Multiply and Divide
3 Index Registers
Option:
Program Interrupt

I

}

RANDEX II Drum Storage Systems:
First Storage Drum Cabinet;
12 million digits (2)
First Storage Drum Cabinet;
24 million digits (2)
Additional Storage Drum Cabinet;
24 million digits

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

1,735

110,000

336

60
400
400

3,000
12,000
12,500

0
70
25

275

10,250

20

4,835

213,000

576

60

3,000

0

5,635

248,200

620

60

3,000

0

1,900

125,000

565

1,850

140,000

650

1,705

85,000

195

2/69

UNIVAC IS 80/90 I

771 :221.102

IDENTITY OF UNIT
CLASS

Model
Number

INPUT/
OUTPUT

PRICES

IMonthly

Feature
Number

Name

Rental

$(1)

Monthly

~rchase Maint.
$
$

Magnetic Tape
7914
7915

Tape/RANDEX Synchronizer
Uniservo Magnetic Tape Unit

1,000
450

50,000
20,000

250
112

255

11,200

55

50
35
35

2,300
1,350
1,350

10
18
18

725

32,000

200

100
100
50

4,200
4,200
2,300

20
20
10

900

45,000

170

935
20
30
0

41,100
800
:1.,320
275

0

0

335
5
10
15
0

Punched Card
7935
7945

7936
7946

High Speed Reader (600 cards/min):
SO-Column
90-Column
Options:
stacker Select (for 7935 or 7945)
90-Column Read (for 7935)
80-Column Read (for 7945)
Read/Punch Unit (150 cards/min):
80-Column
}
90-Column
Options (for 7936 or 7946):
Pre-Read
Post-Read
Stacker Select

}

Paper Tape

-

Paper Tape System
(500 char/sec reader and 100-char/sec punch)
Printer

7912

High-Speed Printer (600 lines/min):
10 Additional Print Positions
20 Additional Print Positions
Variable Line Spacing (6 or 8 per inch)
Continuous Paper Feed

I
KOTES:
(1)

These rentals do not include monthly maintenance charges.

(2)

First MNDEX cabinet in a system can be either a 7965 or 7957;
prices for both include a Synchronizer and Power Control.

2/69

A.

AUERBACH

1. "'.....

772:011. 100

~\EDP

--

AUERBACH

UNIVAC 55 80/90 MODEL II
SUMMARY

REP.ITS

s

SUMMARY: UNIVAC SS 80/90 MODEL"
.1

BACKGROUND
The UNIVAC Solid-State 80/90 Model II system adds a 1280-word magnetic core memory
to the facilities of the Solid-State 80/90 Model I system described in Report 771. The core
memory, despite a fairly slow"cycle time of 17 microseconds, significantly increases the
Model II's internal processing speed and can eliminate many of the coding complexities involved in minimizing drum latency times in the Model I. The Model II central processor
can control two simultaneous magnetic tape or RANDEX drum read/write operations (versus one for the Model I), and its monthly rental is $1,500 higher than that of a comparablyequipped Model I. In most other respects - including the complement of peripheral equipment they can use - Models I and II are nearly identical.
Solid-State 80 systems are designed to read and punch standard 80-column cards, while
Solid-State 90 systems use Remington-Rand 90-column cards. Therefore, there are differences in the I/O buffering and in the code translation facilities. Otherwise, there are
few significant differences between the 80- and 90-column systems.
First deliveries of Solid-State 80/90 Model II systems were made in June 1962, nearly four
years after Model I systems began reaching users. A total of more than 500 Solid-State
80/90 systems were installed, and approximately 100 are still in use at this writing. Both
systems went out of production in 1965 •

•2

HARDWARE
The Solid-State 80/90 Model II Processor handles data in words of 10 digits plus a sign.
Each digit consists of four information bits plus an odd parity bit. Parity is checked during
all data movements to or from storage.
The central processor normally handles only numeric data, whereas the peripheral units
handle alphanumeric data in Hollerith or 90-column card code. This difference is resolved by
splitting each group of 10 alphanumeric characters into two computer words, with one word
holding the zone bits and the other word holding the numeric bits for all of the 10 characters.
The Model II Processor can have from 2600 to 8800 words of drum storage. In addition,
every Model II Processor contains 1280 words of 17-microsecond magnetic core storage
and 9 index registers (versus 3 in Model I) as standard eqUipment. Model II also has an
expanded instruction complement that permits block transfers of data between core and
drum memories, simplifies packing of data, and facilitates coding of alphanumeric operations. Otherwise, Model II uses the same instruction repertoire and the same "one-plusone" instruction format as Model I (see Report 771).
The recommended programming technique for the Solid-State 80/90 Model II is an extension
of the Model I technique. Instruction operands (i. e., data) are preferentially held in core
storage, as are any sequences of instructions that would operate inefficiently if they were
held on the drum. (Such inefficiencies could result from non-optimized coding or from use
of the multiplication or division instructions, whose execution times are variable and datadependent. )
Computation can be as much as 10 times faster than on the Model I (depending upon the
degree of program optimization), and is frequently faster than on an IBM 1401 system. In
fact, computational speeds of the Model II Processor are high enough (e. g., 51 microseconds for a 10-digit addition) so that the system's performance in typical business applications is usually limited by the speeds of the input-output devices, all of whose operations
can be overlapped with internal processing.
On the magnetic drum, data is stored in 200-word "bands" 011 two types. Each so-called
"fast" band is served by one read/write head and has access times of 0 to 3.4 milliseconds.
Each "high-speed" band is served by four read/write heads spaced equidistantly around the
drum's circumference, so the access time ranges from 0 to 0.875 millisecond.
Model II Processors are available in two versions. The Basic Processor has 2400 words
of "fast" drum storage' and 200 words of "high-speed" drum storage, which can be expanded
in increments of 400 words of fast storage and/or 200 words of high-speed storage. The
Expanded Processor has 7600 words of fast drum storage and 1200 words of high-speed
drum storage. Both versions have 1280 words of core storage. Multiply and divide instructions are optional in the Basic Processor and standard in the Expanded Processor. A program interrupt facility is optional for both versions. No hardware facilities for floatingpoint arithmetic are offered.
C 1967 AUERBACH Corporation and AUERBACH Info, Inc,

12/67

UNIVAC SS 80/90 MODEL II

772:011. 200

.2

HARDWARE (Contd.)
A Solid-State 80/90 Model II system can include the same peripheral devices as the Model I.
A central processor, a 600-cpm card reader, a 150-cpm card read/punch, and a 600-lpm
printer make up the basic system. Other units that can be included are: up to 20 magnetic
tape units, up to 20 RANDEX drums, and a 500-cps paper tape reader and 100-cps paper
tape punch. All of these devices are described in Report 771. Up to 10 tape units and/or
10 RANDEX drums can be controlled by a Tape-RANDEX Synchronizer, and a maximum of
2 synchronizers can be connected to a Model II Processor.
Effective drum buffering of I/O operations enables the card reader, punch, printer, and one
tape unit or RAND EX drum per synchronizer to operate simultaneously with computing.
Simultaneous tape reading and writing can be performed in systems that include two synchronizers .

.3

SOFTWARE
Development of software for the Solid-State 80/90 systems suffered from a number of false
starts. The Model I was originally advertised as being designed for programming in
UNIVAC's Flowmatic language (a forerunner of COBOL), but the Flowmatic compiler did not
live up to expectations. A COBOL-60 compiler was also announced for the system but was
later withdrawn. FORTRAN I and II compilers were made available to users in field-test
form but were never formally released.
UNIVAC-supplied software for the Solid-State 80/90 systems includes two assemblers (X-6
and 8-4), a sort routine, about 25 I/O control routines, some program testing routines, a
group of mathematical functions and routines, a linear programming package, and a CPM
system.
X-6 is a basic symbolic assembly system for the Model I, while S-4 is a somewhat more
advanced assembler that can be used on both models. X-6 is a card-oriented assembler,
while S-4 can use magnetic tape units, when available, to speed up the assembly process.
Neither assembler provides facilities for macro-instructions, though S-4 permits the
creation and use of a program library on magnetic tape. Programs coded in the X-6 symbolic language can, with minor revisions, be assembled by the S-4 translator.

12/67

A.

AUERBACH

A

772;221.101

.,HIIII

~BDP

__

IINlII

UNIVAC 51 '0/10 II
JllltICE DATA

UNIVAC 55 80/90 II
Manufacture of this equipment has been discontinued. Orders will be accepted for available equipment on hancl,
The list purchase prices below are shown for informational purposes only. SelUng prices wl11 be quoted by
UNIVAC upon request.

PRICES

IDENTITY OF UNIT
CLASS

Model
Number

Name

Rettal
$ 1)

PurchasE

Monthly
Maint.

$

$

Processing Unit (includes main memory)

PROCESSOR
7961
7962

7963
7964

MASS
STORAGE

IMonthly

Feature
Number

7965
7957
7966

Basic Central Processor
90-COlumn}
80-Column
standard Equipment:
2400 words of fast drum storage
200 words of high-speed drum storage
1280 words of core storage
9 Index Registers
Options:
Program Interrupt
Multiply and Divide
400 Additional Words of Fast Drum
Storage (1600 words max. )
200 Additional Words of High-Speed
Drum Storage (800 words max. )
Expanded Central Processors:
90-Column}
80-Column
Standard Equipment:
7600 words of fast drum storage
1200 words of high-speed drum storage
1280 words of core storage
9 Index Registers
Multiply and Divide
Option:
Program Interrupt
RANDEX II Drum Storage System:
First Storage Drum Cabinet:
12 million digits (2)
First Storage Drum Cabinet;
24 million digits (2)
Additional Storage Drum Cabinet;
24 million digits

3,235

177,500

560

60
400

3,000
12,000

0
70

400

12,500

25

275

10,250

20

7,135

315,700

797

60

3,000

0

1,900

125,000

624

1,850

140,000

650

1,705

85,000

195

1,000
450

50,000
20,000

240
108

255

11,200

53

50
35
35

2,300
1,350
1,350

10
18
18

Magnetic Tape
7914
7915
INPUTOUTPUT

Tape/RANDEX Synchronizer
Uniservo Magnetic Tape Unit
Punched Card
High-~eed

7935
7945

Card Reader (600 cards/min):
80-Column}
90-Column,
Options:
stacker Select (for 7935 or 7945)
90-Column Read (for 7935)
80-Column Read (for 7945)

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

772:221.102

UNIVAC SS 80/90 II

IDENTITY OF UNIT
CLASS

INPUTOUTPllT
(Contd. )

Model
Number

Feature
Number

7936
7946

Name

PRICES

Monthly
Monthly
Rental tpurchase Maint.
$(1)
$
$

Read/Punch Unit (150 cards/min):
BO-Column
90-Column
Options (for 7936 or 7946):
Pre-Read
Post-Read
Stacker Se lect

}

725

32,000

192

100
100
50

4,200
4,200
2,300

19
19
10

730

45,000

170

935
20
30
0
0

41,100
800
1',320
275

322
5
10
15

0

0

Paper Tape
Paper Tape System (500-char/sec reader and
100-char/sec punch)
Printer
--7912

High-~eed

Printer (600 lines/min):
10 Additional Print Positions
20 Additional Print Positions
Variable Line Spacing (6 or 8 per inch)
Continuous Paper Feed

NOTES:
(1) These rentals do not include monthly maintenance charges.
(2) First RANDEX cabinet in a system can be either a 7965 or 7957; prices for both include a

Synchronizer and Power Control.

2/69

A

AUERBACH

"

-

~

........

774:011.100

~EDP

...4

-

....,,'

UNIVAC III
Introduction

INTRODUCTION
1011.
The UNIVAC III is a large scale data processing system suitable for both business
and scientific applications. System rentals range from approximately $19.000 to $40.000
per month, and most installations wlll probably fall within the $25,000 to $35.000 range.
8y means of the software operating system, it is possible to utilize hardware facilities for
simultaneous processing of a number of independent programs. Hardware facUlties that
have been incorporated to help achieve this objective are:
• A series of interrupt levels which permit varied peripheral equipments to
make their demands on the central processor.
• Provision of scatter-read and gather-write facUlties throup the use of
function speclflcation words to specify address assignments.
• AvailabUlty of 9 or 15 index registers plus indirect addressing.
• Four one-word arithmetic registers, which may be used individually
or in combination in ascending order only.
•

Con~ol of all input-output operations by independent input-output channels.
Up to 13 channels can be connected. and all chaMels can operate simultaneously with each other and with the central processor.

The central processor can perform additions or subtractioll8 on binary or decimal
operands. These operands can be distributed over one to four words of '11 bits each. Each
UNIVAC nI word uses two bits for modulo 3 checking. The remaining 25 bits can contain
an instruction word, four 6-bit alphameric characters plus a sign bit. six 4-bit numeric
characters plus a Sign bit, or 24 binary data bits plus a Sign.
Multiplication and division can be performed on decimal data only. The UNIVAC m
can perform logical AND and inclusive OR functions and binary comparison operations.
Branching, alphameric-to-decimal and decimal-to-alphameric conversion. and zero
suppression capabillties ease data manipulation and program control; however. most edtting functions, floating point arithmetic, and conversion of data to floating point format must
be handled by subroutines. Scatter-read and gather-write facUlties provlde fast means of
assembling data into and disseminating data from core storage.
Core storage capacity ranges from 8,192 to 32,768 word locations in increments of
8, 192. Cycle time is 4 microseconds per word, but the majority of ill8tructions take 8
microseconds.
A wide range of input-output equipment is offered for the UNIVAC nI. A system
can include a maximum of 32 Uniservo UlA Magnetic Tape Units, 6 Uniservo UA Magnetic
Tape Units, and a total of 8 units of the following equipments in any combination: High
Speed Card Readers. High Speed Printers, Card Punch Units, Punched Paper Tape Units,
and Uniservo mc Magnetic Tape Synchronizers controlling from 2 to 8 tape units each.
Two models of both card readers and card punches are available. Cards can be
read at a peak rate of 800 cards per minute and punched at a peak rate of 300 cards per
minute. Punched paper tape can be read at 250, 500, or 1,500 characters per second.
and punched at 110 characters per second. The line printer has 128 print positions
and a set of 51 characters, and can print 700 alphameric or 922 numeric single-spaced
lines per minute.

I

1963 by Auerbach Corporation and IINA Incorporated

3/63

UNIVAC III

774:011.101

INTRODUCTION (Contd.)
1011.
Three types of Unlservo Magnetic Tape Units are available for the UNIVAC W
system. the Uniservo lIlA. IIA. and IIIC.
The Uniservo IlIA Magnetic Tape Unit operates at peak data transfer rates of
133.000 alphameric characters or 200.000 numeric digits per second with a density of
1. 000 frames (1.330 characters or 2.000 digits) per inch. Tape can be read forward or
backward. but data can be recorded in the forward direction only. Tape can be read or
written in either the Start-Stop mode or the Non-Stop mode. A read-after-write check is
made upon recording.
The Uniservo IIA Magnetic Tape Unit operates at peak data transfer rates of
25.000 or 12.500 characters per second at densities of 250 or 125 characters per Inch
respectively. When recording at 250 characters per inch. the format is compatible
with the UNIVAC U: the 125 character per inch recording makes the format compatible
WIth the UNIVAC I.
The Unlservo WC Magnetic Tape Unit operates at peak data transfer rates of 22.500
or 62.500 characters per second at densities of 200 and 556 characters per Inch. The
block lengths are variable and the format is IBM-compatible. Tape can be read or written
either with or without translation. When translation Is specified. the Six-bit IBM tape code
is converted to six-bit excess three code or vice versa. A read-after-wrlte check is made
upon recording.
Major emphasis has,been placed on development of software packages to achieve the
maximum throughput capabilities of the system and to simplify programming. These
packages provide complete Input-output control. the means of associating and running independently prepared programs simultaneously. the ability to call many routines and subroutines. and the ability to incorporate new routines or subroutines in the library. Program
testing aids such as SNAPshot. DUMP. and TRACE have also been incorporated in the software packages. Data sorting and merging are provided by a sort generator which generates
the instructions for the sort or merge from a set of parameters outlined by the user. The
original input and final output routines are the responsibll1ty of the user. Input-output
routines provided for the intermediate collati~g pass use any available tape in the system
(even the unused portion of data tapes).
Two complete machine oriented software packages are available for the UNIVAC W;
however. no compatibility exists between them. One paclcage consists of SALT. a machine
oriented language; DUTY. a library of routines and subroutines; and CHIEF, an executive
routine. The other consists of UTMOST. a machine-oriented language; SUPPORT W, a
library of routines and subroutines; and BOSS III. an executive routine. New developments
and innovations will be incorporated in the already more sophisticated UTMOST. SUPPORT
III. and BOSS W package; however. both packages will be maintained.
Both SALT and UTMOST provide an easily understandable mnemonic representation
of instructions. pseudo operations for directing the assembler. and the ability to perform~
operations to develop the operand address. UTMOST is more extensive than SALT in the
functions that it provides.
DUTY and SUPPORT III each provide the ability to update and maintain a library of
routines and subroutines. and an independent library of object programs for the system.
Both CHIEF and BOSS III are comprehensive operating systems that control the
scheduling. loading. and multi-running of programs; handle most errors; and permit twoway communication between the operator and the system. All functions of these executive
routines are initiated by and closely integrated with the hardware interrupt facilities.
Both COBOL-61 and FORTRAN IV have been implemented for the UNIVAC III. Object
programs produced by both the COBOL and FORTRAN compilers can be run under the
control of BOSS Ill.

3/63

I.------~
AUERBACH I WJ

INTRODUCTION

n4:011.102
INTRODUCTION (Contd.)

lOll.

UNIVAC ill COBOL is essentially Required COBOL-1961. Several useful electives
have been implemented, including segmentation of the object program and arithmetic
operands up to 18 digits in size. Extensions to COBOL-61 include a SORT facility,a
MONITOR verb that facilitates program testing, and the ability to add independently compiled COBOL subprograms to a main program at run time.
The UNIVAC ill FORTRAN language is largely compatible with the WM 7090/7094
implementation of FORTRAN IV. Most FORTRAN II statements will also be accepted and
correctly interpreted by the translator. Double precision and complex variables, however,
are not permitted.

© 1963

by Auerbach Corporotion and BNA Incorporated

3/63

A. ,.......

774:221.101

~EDP

_~

II ... "

UNIVAC III
PRICE DATA

....-

UNIVAC III
UNIVAC does not include monthly maintenance charges in their published prices; these charges
have been included in the Monthly Rental column in this Digest to permit convenient comparison
with other systems.

PRICES

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

Name

Monthly
Monthly
Rental Purchase Maint.
$
$
$

Processing Unit (includes core storage)

PROCESSOR
*4121
*4120
*4119
*4118
*F0622-00
*F0623-00

----- - -

Central Processor (8 K word memory)
Central Processor (16 K word memory)
Central Processor (24 K word memory)
Central Processor (32 K word memory)
Programmable Clock
Index Register Expansion

8,300
9,755
11,410
12,475
210
300

390,000
457,500
534,000
583,500
10,000
15,000

575
680
750
805
45
15

2,005

92,500

175

3,015
2,125

145,000
102,500

275
125

1,040

48,000

60

3,010

145,000

230

2,125

102,500

125

780
790
475
530
350
365
365
365
365
350

28,360
36,500
20,000
24,000
17,500
17,500
17,500
17,500
17,500
17,500

70
175
110
150
30
35
35
35
35
30

1,665

80,000

165

795
750
905
850
850
35
60

35,000
35,000
40,000
40,000
40,000
1,350
3,000

175
150
325
285
285
20
20

1,740
1,650

79,000
79,000

400
350

Magnetic Tape

INPUTOUTPUT
5302-00
4135
*5303-00
*0950-00
*5300-00
*5303-00
*0851-00
*0850-00
0854-00
*4126
*1350-00
*1350-02
*1350-03
*1350-04
*1350-05
*4123-00

Mag Tape Control (controls 1 to 6 Uniservo IIA
Tape Units)
Mag Tape Control (Uniservo III Tape Control)
Mag Tape Control (controls 1 to 8 Uniservo mc
Magnetic Tape Unit)
Tape Adapter Cabinet (used with Uniservo mc
Tape units to provide NRZ Tape format)
Mag Tape Control (controls 1 to 16 Uniservo
IlIA Tape Units)
Mag Tape Control (controls 1 to 8 Uniservo
mc Tape Units)
Mag Tape Unit (Uniservo mC)
Mag Tape Unit (Uniservo rnA)
Mag Tape Unit (Uniservo IIA)
Mag Tape Unit (Uniservo ill)
Power Supply
*Power Supply
*Power Supply
*Power Supply
*Power Supply
*Uniservo Power Supply
Paper Tape

*0901-00

Paper Tape Subsystem (reads 250/500/1500
char/sec; punches 110 char/sec)
Punched Card

*0703-00
*4133-00
*0652-00
*4127-00
*4183-00

Card Reader (700 cards/min)
High Speed Reader (TOO cards/min; 80 column)
Card Punch (300 cards/min)
Card Punch Unit (300 cards/min. ; 80 column)
Card Punch Unit (300 cards/min.; 90 column)
90 Column Read Feature
Stub Card Feature

~
*0752-00
*4152-00

Printer (700 lines/min)
High Speed Printer (700 lines/min)

NOTES:
*No longer in production.

© 1969 AUERBACH Corporation and AUERBACH Info,lnc.

2/69

777:001.001
UNIVAC 1050
Contents

CONTENTS

§ 001.

1.
2.
3.

4.

6.
7.

8.
9. .

10.
11.
12.
13.
14.

15.
16.
17.

18.
19.

20.
21.

Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Data Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
System Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
I
Typical Card System . . . . . . . . . . . . . . . . . . . . . . . . .
II
4-Tape Business System . . . . . . . . . . . . . . . . . . . . . . .
III
6-Tape Business System (Model III Processor) . . . . . . . .
6-Tape Auxiliary Storage System . . . . . . . . . . . . . . . . .
V
Typical Real-Time System . . . . . . . . . . . . . . . . . . . . •
Typical On-Line Card Processing System (UNIVAC 1004).
Internal Storage
Model III Core Storage . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . .
Fastrand II Mass Storage . . . . . . . . . . . . . . . . . . . • . . . . . . . . .

777: 011
777:021
777:031
777:031.100
777:031. 200
777:031. 300
777:031. 400
777:031. 500
777:031. 600

Central Processor . . . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . . . .
Console . . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Input-CAltput; Punched Tape and Card
Card Readers (600 and 800/900 cpm) . . . . . . . . . . . . . . . . . . . . .
Card Punches (200 and 300 cpm) . . . . . . . . . . . . . . . . . . . . . . . . .
Punched Tape Equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Input-CAltput; Printers
Printers (600/750 and 700/922 lpm) . . . . . . . . . . . . . . . . . . . . . .
Input-CAltput; Magnetic Tape
Uniservo III A Magnetic Tape Handler . . . . . . . . . . . . . . . . . . . . .
Uniservo IV C Magnetic Tape Handler . . . . . . . . . . . . . . . . . . . . .
Uniservo VI C Magnetic Tape Handler . . . . . . . . . . . . . . . . . . . . .
Input-Output; Other
Standard Communications Subsystem . . . . . . . . . . . . . . . . . . . . . .
UNIVAC 1004 Adapter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Simultaneous Operations . . . . . . . . . • . . . . . . . . . . . . . . . . . . . . . . . .
Instruction List . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Data Code Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Problem Oriented Facilities
UNIVAC 1050 Tape Sort . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
REGENT (report program generators) . . . . . . . . . . . . . . . . . . . .
Data Transcription . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
File Maintenance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Magnetic Tape Diagnostic Routines . . . . . . . . . . . . . . . . . . . . . . .
Process Oriented Language
COBOL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
FORTRAN IV • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Machine Oriented Languages
PAL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Program Translators
PAL TAPE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
PAL JR . . . . . . . . . • • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
PAL CARD. . . . . . . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Operating Environment: OPR . . . . . . • . . . . . . • . . . . . . . • . . . . . . . .
System Performance
Comments . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . . . . . . .
Worksheet Data Table . . . . . . . . . . • • . . . . . . . . . . . . . . . . . . .
Generalized File Processing . . . . . . . . . • • • • • . . . . . . . . . . . . .
Sorting . . . . . . . . . . • . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . .
Physical Characteristics . . . . . . . . . . . • • . . . . . . . . . . . . . . . . . . . . .
Price Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

777:051
777 :061

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

777:041
777:042

777:071
777:072
777:073
777:081
777:091
777:092
777:093
777:101
777: 102
777:111
777:121
777: 141
777:151.13
777:151.14
777: 151. 15
777:151.16
777:151.17
777:161
777:162
777:171
777:181
777:182
777: 183
777:191
777 :201. 001
777:201. 011
777:201.100
777:201. 200
777 :211
777:221

1/69

1.

lA,

AUERBAC~

777:011. 100
sum ..

UNIVAC 1050
SUMMARY

EDP
."DIIS

~

SUMMARY
The UNIVAC 1050 is a small to medium scale, solid-state computer that is oriented
toward business data processing applications. The Model Ill, with internal processing rates
of about 8,500 instructions per second is the only model available. The Model IV, which would
have been about 3.5 times as fast, was never produced.
The UNIVAC 1050 was conceived several years ago as a contender for the IBM 1401class market. For marketing reasons, it was first announced in May 1962 as an off-line inputoutput processor, with a maximum of two magnetic tape units, for larger UNIVAC systems. An
expanded version, the UNIVAC 1050-ll, was developed which included communications and
Fastrand mass storage equipment. This system was eminently successful - 152 systems were
ordered by the U. S. Air Force in November 1963. Finally, in March 1964, the UNIVAC 1050
Model ill and Model IV were announced as general-purpose, commercial EDP systems, with a
complete line of peripheral equipment and software. The first delivery of the Model ill Processor
was made in 1963.
While the UNIVAC 1050 offers many useful features and is a relatively easy machine to
program, its earlier origin is apparent in some performance comparisons of the Model TIl against
other recently-announced computer systems in the same price class.
.
A wide range of peripheral devices is offered for the UNIVAC 1050, permitting effective use of the system as an off-line input-output processor for larger, tape-oriented computers
and as a control center for a data communications network. A novel feature (the 1050-1004
Adapter) permits the UNIVAC 1050 to be used as an "expansion package" to increase the processing capacity of UNIVAC 1004 installations. UNIVAC 1050 system rentals range from approximately $1, 400 to over $20, 000 per month, but most installations probably fall within the $5, 000
to $12, 000 rental range.
Hardware
UNIVAC 1050 systems can have from 4, 096 (included with the processor) to 32,768
character positions of core storage in 4, 096-character modules. Each core storage location
contains six data bits and one parity bit. Cycle time is 4.5 microseconds per character.
The 1050 processor has typical single-address logic, with the addition of add-to-storage
capabilities. Two 16-character accumulators and 7 index registers are provided. Facilities for
indirect addressing are not provided. Arithmetic is basically decimal, but binary add and subtract instructions are provided to facilitate address modification and binary arithmetic. (The
core storage addressing mode is pure binary.) All operations are performed serially by character.
The basic instruction format is a 30-bit "binary word" (5 consecutive characters) which
usually contains a 5-bit operation code, a 3-bit index register specification, a 16-bit address,
and a 6-bit modifier. The External Function (input-output) instruction uses the 30 bits in a
slightly different manner. Operand length is variable and can range from 1 to 16 characters for
most instructions, as specified in the instruction. All instructions except the multiply and divide
instructions and the input-output instructions can be indexed.
Extensive editing facilities, including character insertion, floating dollar, and check
protection, are provided. A block transfer instruction can cause the transfer of up to 1, 024
characters from one location in core storage to another. A translate instruction can autoIl).atically
translate up to 64 different characters from one 6-bit code to any other 6-bit code defined by a
table supplied by the programmer. The only optional hardware facility is the decimal multiplydivide feature. Floating-point arithmetic hardware is not available for the 1050 Processors.

An extensive system of interrupts and testable indicators provides program control.
Interrupts are divided into three priority levels and may occur upon detection of such conditions
as internal parity errors, operator request, decimal overflow, successful completion of an inputoutput operation, or malfunction of an iIWut-output device. All interrupts except the one for
internal parity errors can be program-inhibited. Recognition of an interrupt signal causes a
transfer of control to a specific location, depending on the type of interrupt, which can contain a
branch to a routine to handle the condition.
Testable indicators which enhance the interrupt system include indicators for decimal
overflow and for the status of the device on a particular input-output channel. The results of
comparisons and arithmetic functions are also indicated by testable indicators. Conditions that
can be tested include equal, not-equal, greater-than, and less-than after comparisons, and resultzero, result-negative, and decimal or binary overflow after arithmetic operations.

@

1969 AUERBACH Corporation and AUERBACH Info, Inc.

1/69

777:011.101

UNIVAC 1050

There are eight input-output channels available for the 1050 Central Processor, each
of which is assigned to a single peripheral subsystem (i. e., there are no facilities for programmed switching of the channel assignments). Three (any three) channels are included in
the basic price of the processor. The synchronizers (control units) for the card reader, card
punch, and printer are internal to the main frame of the processor, each using one input-output
channel. Two other channels are reserved for a magnetic tape subsystem, leaving three generalpurpose input-output channels.
The peripheral subsystems available for use in UNIVAC 1050 systems are summarized
below. Each peripheral subsystem requires the full use of one input-output channel, except that
a magnetic tape subsystem requires the use of two. Additional synchronizers will be required
(on special order) to add a second card reader, card punch, or printer to the system.
•

Card Reader: column reader; 800/900 or 600 cards per
minute; 80 or 90 columns.

•

Card Punch: row punch; 300 or 200 cards per minute;
80 or 90 columns.

•

Printer: 700/922 or 600/750 lines per minute; 128 (132
optional) print positions; 63 printable characters.

•

Uniservo m A Magnetic Tape Subsystem: 1 to 6 tape
handlers; up to 133,000 characters per second; readcompute or write-compute overlap; read-after-write
modulo-3 check; read-backward capability.

•

Uniservo IV C Magnetic Tape Subsystem: 1 to 6 tape
handlers; up to 90, 000 characters per second, readcompute or write-compute overlap; read-after-write
parity check; IBM 729- compatible.

•

Uniservo VI C Magnetic Tape Subsystem: 1 to 16 tape
handlers; up to 34,100 characters per second; read-writecompute overlap; read-after-write parity check; IBM 729compatible; relatively low-cost.

•

Fastrand IT Mass Storage Subsystem: 1 to 8 storage units;
over 132 million characters per storage unit; 93 milliseconds
average access time; up to 185,000 characters per second
data transfer rate; search feature; read-compute, writecompute, or position-compute overlap.

•

UNIVAC 1004 Subsystem: connects a UNIVAC 1004 Model
I, IT, or m (Computer System Report 770:) on-line with
the UNIVAC 1050.

•

Punched Paper Tape Subsystem: 1, 000 or 300 charactersper-second reader; 110 characters-per-second punch; 5-,
6-, 7-, or 8-level tape.

•

Standard Communications Subsystem: Communications
Multiplexer controls 1 to 32 Communication Line Terminals; console typewriter can monitor messages;
messages on all lines can be handled simultaneously;
will accept various telephone and teletypewriter communications services with data transfer rates of up to 800 characters per second.

In typical business data processing systems using the Uniservo VI C tape handlers,
simultaneity of input-output operations can be quite good; but in systems using the faster
Uniservo m A or IV C tape handlers or Fastrand IT, simultaneity can be severely restricted.
The complete picture of the UNIVAC 1050's capabilities for simultaneous operations is complex
and is presented in detail in Section 777:111 of this report.

Software
UNIVAC has developed a generously large array of software for the 1050 system. All
of the programming systems and routines summarized below are scheduled for delivery before
or during the first quarter of 1965.

1/69

•

PAL JR: Basic symbolic assembly system, usable on minimum
UNIVAC 1050 configurations with card input-output units and
4,096 character positions of core storage.

•

PAL CARD: Basic symbolic assembly system, usable on expanded
UNIVAC 1'050 card-oriented systems with at least 8,192 positions
of core storage. PAL CARD features more versatile input-output
control routines than PAL JR.

A.

AUERBACH

'"

(Contd.)

SUMMARY

777'011 102

•

PAL 1004: A basic assembly utilizing the UNIVAC 1004 for
input and output.

•

PAL TAPE: A more advanced assembly system requiring two
magnetic tape units, card punch (optional if tape-only output
is desired), card reader, printer, and at least 8,192 positions
of core storage on the translating computer. Macro, file control, and library facilities are provided.

•

PAL DRUM: A Fastrand-oriented version of PAL TAPE; a
Fastrand magnetic drum replaces the magnetic tape units.

•

PATCH Assembler: Provides the capability for patching
object programs without the need for full reassembly. Facilities are provided for addition, deletion, or alteration of sections
of coding. Input and output are on cards only.

•

COBOL: The COBOL compiler for the 1050 consists of a limited
subset of the COBOL-65. It includes almost none of the features
that distinguish COBOL-65 from COBOL-61 among these omissions
are SORT, REPORT-Writer, COMPUTE, any facilities for program segmentation or access to library routines. The result is
a simple fJompiler that was obviously introduced when UNIVACS
plans for offering a full COBOL-61 Extended compiler did not
come off. The minimum configuration for COBOL compilation
include:

•

•

•

Three magnetic tape units (must all be of the same type).

•

Card reader

•

Printer

the FORTRAN Operating System: Consists of a series of programs
and subroutines which translate other programs and subprograms
written in FORTRAN IV source language into object code capable of
execution on the UNIVAC 1050 system. Object programs produced
by the FORTRAN Operating System are relocatable, relative programs which will operate singly or concurrently under the control
of the OPR Executive Routine. The minimum configuration required
for compilation includes:
•

4 or more memory modules (16,384 positions of storage).

•

Multiply and divide hardware feature

•

Four tape units

•

Printer

•
Card reader
REGENT CARD: Provides facilities for generating report programs
for card-oriented systems. Sections of the user's own coding in PAL
symbolic language can be included.

•

REGENT TAPE: Generates report programs for tape-oriented systems. Sections of the user's own coding in PAL symbolic language
can be included.

•

UNIVAC 1050 SORT Routine: Performs tape sorting operations,
using decimal, binary, or alphanumeric keys, on a system with
3 to 6 magnetic tape units. The Sort routine can be used with as
little as 8,192 core storage positions, but additional storage will
increase its efficiency. Sections of the user's own coding can be
incorporated.

•

OPR: An executive routine for the UNIVAC 1050, which provides
the linkages between the "worker program" and routines from the
PAL Library. OPR can be modified to control data communications devices. One version of OPR provides for the loading and
execution of two programs in parallel.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

1/69

777:221.101
UNIVAC 1050
PRICE DATA

UNIVAC 1050
UNIVAC does not include monthly maintelWloe charges in their published prices; these charges have been
included in the Monthly Rental column in this Digest to permit convenient comparison with other systems.

IDENTITY OF UNIT
CLASS

PROCESSOR

Model
Number
*0500-02
·0500-01

Feature
Number

Name
Central Processor (including 8,192
characters of core storage)
Core Memory Expansion (4,096 character
modules; maximum total storage is 32,768
characters)

PRICES
Monthly
Monthly
Rental Purchase Maint.
$
$
$
1,930

68,600

215

625

23,400

40

1,245

47,000

70

1,250

46,400

90

790
780
835
225

36,500
28,360
29,050
7,400

175
70
110
40

425
750

13,000
19,400

100
265

370
818

12,400
26,000

60
168

Magnetic Tape

INPUTOUTPUT

·0551-00

*0850-00
*0851-00
*0851·04
*1353-00

Mag Tape Control (controls 1 to 16

Uniservo m A Tape Units)
Mag Tape Control (controls 1 to 12
Uniservo m C Tape Units)
Mag Tape Unit (Unlservo m A)
Mag Tape Unit (Unlservo m C)
Mag Tape Unit (Unlservo IV C)
Power Supply
Punched Card

·0701-31
·0650-21

Card Reader (1,000 cards/min)
Card Punch (300 cards/min)
Printer

·0751-03
·0751-04

Printer Buffer
Printer (700/922 cards/min)

NOTES
·No longer in production.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/6'

UNIVAC 1050

777:221.102

UNIVAC l050-A CARD PROCESSING SYSTEM
UNIVAC does not include monthly maintenance charges in their published prices; these charges have been
included in the Monthly Rental column in this Digest to permit convenient comparison with other systems.

PRICES

IDENTITY OF UNIT
CLASS

PROCESSOR

Model
Number

Feature
Number

*0500-05
*F0593-00
*F0726-00
*F0725-00
*F0701-00
*4002-01
*F0727-00
*0670-00

Name

Monthly
Monthly
Rental Purchase Maint.

$
Central Processor (includes 4,096 characters
of core storage)
Memory Expansion (4,096 - character module;
maximum total storage is 32,768 characters)
I/O Channel Expansion (1 I/O channel; 5
additional maximum)
Multiply and Divide
Inquiry Typewriter
Console (desk type)
Console (integrated)
Power Supply Expansion

$

$

1,230

44,000

130

335

12,800

15

45

1,600

5

155
100
75
45
160

6,000
3,200
2,800
1,600
5,600

5
20
5
5
20

415
240
715
425

12,985
7,000
21,560
13,570

115
65
230
130

570
195
870
615

21,640
7,000
25,970
17,940

30
20
275
225

Punched Card

INPUTOUTPUT

Card
Card
Card
Card

*0706-00
*0706-01
*0600-00
*0600-12

Reader (800/900 cards/min)
Reader (600 cards/ min)
Punch (300 cards/min)
Punch (200 cards/min)

Printer
*5003-00
*F0663-03
*0755-05
*0755-01

Printer
Printer
Printer
Printer

Control (for second printer)
Buffer
(700/922 lines/min)
(600/750 lines/min)

Notes:
*No longer in production

2/69

fA

AUERBACH

'"

777:221.103

'RICE DATA

UNIVAC 1050·111
UNIVAC does not include monthly maintenance charges in their published prices; these charges have been
included in the Monthly Rental column. in this Digest to permit convenient comparison with other systems.

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

PROCESSOR

Name

PRICES
Monthly
Monthly
Rental Purchase Maint.

$

$

$

Processing Unit
*0500-05
*F0593-00
*F0725-00
*F0726-00
*4002-01
*4002-03
*F0727-00
*F0710-00
*1353-00
*1353-01
*1353-02
*1353-03
*F0670-00
*F0635-00

1050-m Processor (includes 4,096 characters
of core storage and 3 I/O Channels)
Memory Expansion - 4K (4,096 character
module; maximum total storage is 32,768
characters)
Advanced Logic (multiply and divide)
I/O Channel Expansion (1 I/O channel;
maximum of 8 total)
Free Standing Console
Free Standing Console
Integrated Console
Inquiry Typewriter
Power Supply
Power Supply
Power Supply
Power Supply
Power Supply Expansion
1004 Control (used with 1004 or 1005 processor)

1,230

44,000

130

335

12,800

15

155
45

6,000
1,600

5
5

75
75
45
100
225
225
225
225
160
215

2,800
2,800
1,600
3,200
7,400
7,400
7,400
7,400
5,600
7,840

5
5
5
20
40
40
40
40
20
20

4,050

164,640

300

30
215
995

1,040
8,235
39,800

5
25
100

1,035

36,880

115

55

2,225

-

515

17,350

115

515

17,350

115

310
620

10,.t70
23,480

70
35

1,025

38,920

55

415
240
715
425
870
585

12,985
7,000
21,560
13,570
24,320
17,320

115
65
230
130
260
150

Drum Storage

MASS
STORAGE
6010-00
F0688-01
F0686-01
*5002-02
*5002-03
*F0710-00
INPUTOUTPUT

Fastrand II Unit (132, 120, 576 characters
per unit)
Write Lockout
Fastbands
Drum Control (controls up to 2 Fastband II
units or 8 modular Fastband units)
Fastrand Control (Controls up to 8 Fastrand
II Units)
Search All Words
Magnetic Tape

*OS53-16
*OS58-00
0858-01
*5307-00
*0551-01

Uniservo VIC Master (7-track; 200/556/800 bpi;
handles up to 3 Uniservo Vic Slaves)
Uniservo VIC Master (7-track; 200/556/800 bpi;
reads backward; handles up to 3 Uniservo
VIC Slaves)
Uniservo Slave (7-track; 200/556/S00 bpi)
Uniservo VIC Control (controls 1 to 4
Uniservo VIC Master units; maximum
total units per control is 16)
Mag Tape Control (controls l-e Uniservo III A
tape units)
Punched Cards

*0706-00
*0706-01
0600-00
*0600-12
*0600-02
*0600-14

Card
Card
Card
Card
Card
Card

Reader (80 column; 800/900 cards/min)
Reader (80 column; 600 cards/min)
Punch (80 column; 300 cards/min)
Punch (80 column; 200 cards/min)
Reader/Punch (80 column; 300 cards/min)
Reader/Punch (80 column; 200 cards/min)

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

UNIVAC 1050

777:221.104

IDENTITY OF UNIT
CLASS

INPUTOUTPlIT
(Contd. )

Model
Number

Feature
Number

Name

PRICES
Monthly
Monthly
Rental PurchasE Maint.

s;

$

870
615
570
195
210

25,970
17,940
21,640
7,000
7,680

275
225
30
20
20

Paper Tape Reader (1,000 char/sec)
Paper Tape Reader (300 char/sec)
Paper Tape Punch (110 char/sec)
Paper Tape Control (control and cabinet
for paper tape reader or punch)
Reader Spooler
Punch Take Up Reel

415
215
175
225

10,320
5,31\0
4,400
6,900

155
80
65
95

95
5

2,280
200

35

Multiplexer (8 simplex positions)

770

21,640

230

$
Printer

*0755-05
*0755-01
*5003-00
*F0663-03
*F0663-01

Printer
Printer
Printer
Printer
Printer

(700/922 lines/min)
(600/750 lines/min)
Control (control for second printer)
Buffer (128 ciaracters)
Buffer (132 characters)

Paper Tape
*0903-00
*0903-01
*F0606-02
*5005-00

COMMUNICATIONS

*F0636-00
*F0637-00
*F0900-03

NOTES:
*No longer in production.

2/69

fA..

AUERBACH

-

784:011. 100

A.

SI ..... I

UNIVAC 1107
SUMMARY

~EDP

AUERBAC~

'E"ln

•

SUMMARY: UNIVAC 1107
.1

BACKGROUND
The UNIVAC 1107 Thin-Film Memory Computer. first delivered in September 1962. is a
large-scale data processing system suitable for both scientific and commercial applications.
The 1107 was the solid-state successor to the vacuum-tube UNIVAC 1105 and 1103 scientific
processors; it provided greatly increased speed and flexibility in both internal processing
and input-output operations. Program compatibility with previous UNIVAC systems was
not maintained. Monthly rentals for most 1107 systems originally fell within the $40.000
to $70.000 range.
In July of 1964. the 1108. a largely compatible but faster and more powerful version of the
1107. was announced (see Computer System Report 785). Shortly after delivery of the first
1108 system in December 1965. UNIVAC placed the 1107 on a "discontinued" basis. although
1107 production facilities. to date. have not been dismantled. UNIVAC 1107 central processor prices were revised downward approximately 20%. while prices of peripheral units
common to both 1107 and 1108 systems were not reduced. The Price List included in this
Summary Report reflects guideline prices only; actual sales or rentals are individually
negotiated when machines become available. Many of the 40-plus existing 1107 systems
have been purchased outright. so the present availability of 1107 equipment is difficult to
assess. UNIVAC is continuing normal software maintenance for its 1107 customers. although no new software releases are expected.
The UNIVAC 1107 features both multiprocessing and limited multiprogramming capabilities.
an efficient central processor with a comprehensive instruction repertoire. and a powerful
input-output and communications capability.
Many of the features of third-generation computers. such as advanced interrupt facilities.
a separate. high-speed control memory. and a high degree of hardware modularity. were
provided in the UNIVAC 1107.
Chief deficiencies of the 1107 with respect to its successor. the 1108. include:
•

•2

Core memory capacity is limited to a maximum of 65.536 36-bit words.

•

Basic cycle time is 4.0 microseconds.

•

Double-precision arithmetic facilities are not provided.

•

Memory protection features are lacking.

•

A maximum multiprocessor system is limited to two 1107 central processors .

HARDWARE
A basic UNIVAC 1107 system consists of a central processor with console and 8. 12. or 16
input-output channels; 16.384. 32.768. 49.159. or 65.536 36-bit words of core storage; and
peripheral subsystems as required. Tables I and II summarize the subsystems and the
number of I/O channels that each subsystem requires.
Each core memory location can hold one instruction. one single-precision floating-point data
item. from one to six fixed-point data fields, four alphanumeric eight-bit bytes, or six
alphanumeric characters in 1107 internal code. No provision exists for recording or
checking parity of data words held in core storage. The memory can consist of dual asynchronous banks. Read access time is 1.8 microseconds and cycle time, including data
regeneration, js 4.0 microseconds for each 36-bit word. By storing instructions in one
bank and data in another, it is possible to overlap the operation of the two banks; this reduces the effective cycle time by a factor of two. The "alternate banks" storage allocation
technique decreases the execution time for most instructions by 4.0 microseconds; e. g. ,
each add. subtract, load, or store instruction takes 8.0 microseconds when the operand
and corresponding instruction are in the same bank and only 4.0 microseconds when they
are in alternate banks.
The 128-word Thin-Film Memory, used by the 1107 for control purposes, has a read access
time of 0.333 microsecond and a cycle time of 0.667 microsecond, allowing references to be
made at rates approaching 1. 5 million cycles per second. Specific functions are assigned to
63 of the 128 Film Memory locations; 15 locations serve as index registers and 16 as arithmetic registers (accumulators). This abundance of arithmetic and index registers contributes
heavily to the power and flexibility of 1107 programming. The remaining 65 Thin-Film Memory locations are available as general-purpose working storage, but there are certain programming restrictions on their use.
© 1967 AUERBACH Corporation and AUERBACH Info. Inc.

12/67

784:011. 200

UNIVAC 1107

TABLE I: UNIVAC 1107 MAGNETIC DRUMS
Characteristic

FH-880

Fastrand

17

92

60,000

25,625

6,291,456

176,160,768

8

8

lor 2

lor 2

Average access time, msec
Peak transfer rate, words/sec
Maximum storage per subsystem,
36-bit words
Maximum drum units per subsystem
Number of input-output channels
per subsystem

TABLE II: UNIVAC 1107 INPUT-OUTPUT SUBSYSTEMS

Subsystem

Uniservo IIA
Magnetic Tape

.2

Number of I/O
Channels per
Subsystem

Maximum Number
of Devices per
Subsystem

1

Peak Speed

12

25,000 char/sec.

Uniservo IlIA
Magnetic Tape

lor 2

16

100,000 chari sec.

Uniservo I1IC
Magnetic Tape

1

12

62,500 char/sec.

Punched Card

1

1 reader;
1 punch

Printer

1

Punched Paper
Tape

1

UNIVAC 1004

1

Communications
(multi -line)

1

4 multiplexors,
each serving up
to 32 half- or
full-duplex lines

Communications
(single-line)

1

1

4
1 reader;
1 punch
1

1

read 600 cpm;
punch 300 cpm.
9001pm.
read 400 char/sec;
punch 110 char/sec.
cards: read 615 cpm,
punch 200 cpm;
print 600 lpm;
paper tape: read 400
char/sec, punch 110
char/sec; magnetic
tape: 34,200 chari
sec.
4800 bits/sec per
line; 51,000 chari
sec total.
40, 800 bits/ sec.

HARDWARE (Contd.)
The UNIVAC 1107 Central Computer can perform fixed-point and floating-point arithmetic on
one-word binary operands. The 16 arithmetic registers, 16 index registers, a versatile
repertoire of 7-part instructions, recursive indirect addressing, and a partial-word transfer facility permit efficient processing of most scientific problems. Commercial processing
is somewhat less efficient because the 1107 lacks automatic facilities for editing, decimal
arithmetic, and radix conversions.
Although the 1107 uses a I-address instruction format, a limited 2-address capability is
provided since most instructions can specify the use of anyone of the 16 arithmetic registers.

12/67

A

(Contd. )

.,

AUERBACH

SUMMARY

.2

784:011. 201

HARDWARE (Contd.)
The partial-word load and store instructions can transfer any half, third, quarter, or sixth
of a word to or from the least significant bit positions of any arithmetic register. A wide
variety of logical, shift, search, and block transfer operations can be performed. All instructions can be indexed, and each index register can be automatically incremented or
decremented each time it is referenced. Indirect addresses can be "chained", and indexing
can be performed upon each address in the chain.
A program interrupt facility causes a transfer of control to one of 74 fixed core locations
upon completion of an input-output operation, upon detection of a processor or input-output
error, or upon count-down to zero of the real-time clock (its contents are decremented by
1 every millisecond). The interrupt facility permits full utilization of the Central Computer
and all peripheral devices under the control of an integrated operating system that handles
multiprogramming.
The 1107 has 16 input-output channels, and each channel is capable of transmitting data in
or1.e direction at a time. One channel is normally occupied by the Control Console, which
provides keyboard input and typed output at 10 characters per second. Each of the remaining
15 channels can accommodate 1 peripheral SUbsystem, and each subsystem can consist of
any of the following groups of devices and their associated control units.
•

1 to 8 Flying Head 880 Magnetic Drums. Each drum stores 786,432 words,
with an average access time of 17 milliseconds. Peak data transfer rate is
60,000 words per second. This rapid-access auxiliary storage plays an important role in the operation of several of the software systems.

•

1 to 8 Fastrand Mass Storage Units. Each unit has 2 drums served by 64
movable heads, and stores 12,976,128 words with an average access time
of 92 milliseconds. Peak data transfer rate is 25,000 words per second.

•

2 to 16 Uniservo lIlA Magnetic Tape Handlers. Read forward or backward at
a peak transfer rate of 100,000 rows per second. Nine tracks are recorded
on ~-inch-wide tape at a density of 1,000 rows per inch, with read-afterwrite row parity checking.

•

2 to 12 Uniservo IlA Magnetic Tape Handlers. Read forward or backward at a
peak transfer rate of 12,500 or 25,000 rows per second. Eight tracks are recorded on ~ -inch-wide tape at a density of 125 or 250 rows per inch; there is no
read-after-write checking.

•

2 to 12 Uniservo IIIC Magnetic Tape Handlers. Read forward only at a peak
transfer rate of 22,500 or 62,500 rows per second. Seven tracks are recorded
on ~-inch-wide tape at a density of 200 or 556 rows per inch, with read-afterwrite checking of longitudinal and row parity. The tape format is fully compatible with the IBM 727, 729, and 7330 Magnetic Tape Units.

•

1 Card Reader and 1 Card Punch. These units read standard 80-column cards at
600 cards per minute and punch them at 150 or 300 cards per minute. Reading
and punching can be performed in alphameric, row binary, or column binary mode.

•

1 High-Speed Printer. Two models are available: one uses a 51-character set
and prints up to 600 alphameric lines per minute; the other uses a 63-character
set and prints up to 700 alphameric lines per minute (or up to 922 lines per
minute when a restricted set of 40 characters is used).

•

1 Paper Tape Reader and 1 Paper Tape Punch. These units (housed in a single
cabinet) can read standard 5-, 6-, 7-, or 8-track punched tape at up to 400
characters per second and punch it at 110 characters per second.

As the above summary indicates, three different types of magnetic tape handlers are available for the 1107, and there is no format compatibility between any two of them. This situation resulted from the manufacturer's decisions to provide a tape handler compatible with
earlier UNIVAC systems (the Uniservo IIA), a tape handler compatible with IBM systems
(the Uniservo lIIC), and a high-performance tape handler for use where compatibility is not
a primary concern (the Uniservo lIlA) •
•3

SOFTWARE
Two different basic software packages have been developed for the UNIVAC 1107, and there
is little or no compatibility between them. UNIVAC states that maintenance for both packages will continue, but they will not be expanded to include new facilities. The "SLEUTH I
Package," also called the "A Package," was developed by UNIVAC's Scientific Computer
Department in St. Paul; it includes the following routines:
•

SLEUTH I - a symbolic assembly system with macro instruction facilities that
translates symbolic source programs into either relocatable or absolute machine language object programs.

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

12/67

784:011. 300

.3

UNIVAC 1107

SOFTWARE (Contd.)
•

EXEC I - an operating system designed to facilitate effective use of 1107 systems
by providing the means for automatically processing a scheduled set of jobs with
a minimum of operator intervention. '

•

CLAMP - a Relative Load Routine that loads either absolute or relocatable object
programs independently or under control of EXEC I.

•

Librarian - a library maintenance routine that creates a library tape and adds,
deletes, corrects, resequences, lists, and catalogs programs on existing library
tapes.

•

UON (!!ibrary of ,!nput-Qutput ~umerical Subroutines) - a set of subroutines,
called by SLEUTH macro instructions, that perform the following functions in
connection with EXEC I:
Opening and closing of files and reels;
Input and output on tape, drum, cards, or printer;
Conversions between decimal and binary radix;
Data transcriptions (cards to tape or drum, tape or
drum to cards, and tape or drum to printer).

•

MIDAS (Macro Instructions for Qumping Areas of .§.tore) - a set of subroutines
designed to aid debugging by providing printed listings of the contents of specified
areas of storage. A valuable option permits listing only the contents that have
been altered during execution of the program being tested.

•

Sort/Merge - a generalized, relocatable subroutine for sorting or merging files
into ascending or descending order. Control parameters are supplied on cards.
From 4 to 12 magnetic tape units can be used, and FH-880 Magnetic Drums
provide increased sorting speed when available.

The "SLEUTH IT Package," also called the liB Package," was developed by Computer Sciences
Corporation and includes the following routines:
•

SLEUTH II - a symbolic assembly system with macro instruction facilities, that
translates symbolic source programs into relocatable machine language object
programs. A magnetic drum is required, but magnetic tape is not. (There is
no compatibility between SLUETH I and SLEUTH II; even the mnemonic codes
for machine instructions are totally different.)

•

EXEC II - an operating system designed to monitor the compilation and execution
of programs, maximize utilization of the available hardware, and minimize operator intervention. The system utilizes an FH-880 Magnetic Drum as a high
capacity buffer store to keep the card readers, punches, and printers fully occupied and as a fast access allxiliary store for program segments. An integrated
set of diagnostic aids and library maintenance facilities is included.

•

COBOL - a compiler for COBOL-61 source programs that operates under control
of EXEC IT. Language facilities include nearly all of Required COBOL-61 (there
are a few minor deficiencies) ; several COBOL-61 electives (but not the extremely
useful COMPUTE verb); a MONITOR verb (which provides dynamic printouts of
the values of specified items); and a SORT facility (but not the one defined as part
of Extended COBOL-61). A magnetic drum is required for COBOL compilations,
but magnetic tape is not.

•

FORTRAN - a compiler for FORTRAN IV source programs that operates under
control of EXEC II. Language facilities are largely compatible with those of
FORTRAN IV as defined for the IBM 7090/7094. FORTRAN II source programs
can be converted to FORTRAN IV by means of the SIFT Translator, which has
been compiled and successfully run on the 1107. The FORTRAN compiler achieves
rapid compilation speeds through use of an FH-880 Magnetic Drum.

•

SORT II - a generalized sort/merge routine that will operate under control of
EXEC II.

The SLEUTH II Package is the more widely used of the two software packages because it includes the COBOL and FORTRAN compilers. Most UNIVAC 1107 users belong to USE, the
UNIVAC Scientific Exchange, which distributes user-developed programs. Furthermore,
the FORTRAN compiler and the SIFT Translator enable 1107 users to utilize the extensive
libraries of FORTRAN-coded routines that are now available.

12/67

fA

AUERllACH

co

784:221.101

su ......

EDP

UNIVAC 1107

•.,•• t,

PRICE DATA

UNIVAC 1107
UNIVAC does not include monthly maintenance charges in their published prices; these charges have been
included in the Monthly Rental column in this Digest to permit convenient comparison with other systems.

PRICES

IDENTITY OF UNIT

CLASS

Model
Number

Feature
Number

Monthly
Monthly
Rental Purchase Maint.

Name

$

$

$

20,590

888,750

2,365

4,925
7,215
9,330
13,985
18,380

213,750
315,000
405,000
607,500
798,750

325
480
540
775
955

65
125
50

2,700
5,400
2,250

5
10

335

13,000

205

7,840

20

1,550

58,800

190

2,845
3,430
2,150

135,000
160,000
85,165

115
290
190

1,615

77,500

150

2,100

124,800

230

3,935
1,025
5,180

182,400
35,460
240,000

195
100
300

1,040

48,000

60

790
780

36,550
28,360

175
70

475
260

20,000
10,000

110
30

835
105
570

29,050
3,600
25,300

110
5
45

675

25,350

125

Processing Unit

PROCESSOR

Central Processor

*7200

Main Storage

ATTACHMENTS,
ADAPTERS,
AND
CHANNELS

*7230
*7231
*7232
*7233
*7234

Core
Core
Core
Core
Core

*8048
*8049
*8158
*2502-01

Transfer Switch (Single)
Transfer Switch (dual)
Cabinet (houses 8048 or 8049)
Cabinet (same as 8049 but with integral
operator panel)
*Adapter (for connection to UNIVAC
1004/1005)

*F0597-03

Memory
Memory
Memory
Memory
Memory

(16
(16
(32
(49
(65

K words,
K words,
K words,
K words,
K words,

one
two
two
two
two

bank)
banks)
banks)
banks)
banks)

-

Drum Storage

MASS
STORAGE

Drum Control (controls 1 to 8 FH-880 Drums)
Drum Control (controls up to 8 Fastrand lor
II Drums)
Drum Fastrand I Mass Storage Unit
Drum FH-880 Magnetic Drum

*7427
*7433
*0900-00
*7432
INPUTOUTPUT

Magnetic Tape
*7214
*7273
*8003-08
*0952-00
*8003-99
*7424
*0850-02
*0851-03
*0854-02
*7274-02
*0851-04
*F0684-00
*8142

Mag Tape Control (controls 1 to 6 Uniservo
IIA Tape Units)
Mag Tape Control (controls 1 to 12 Uniservo
mc Tape Units)
Mag Tape Control (controls 1 to 16 Uniservo
IlIA Tape Units)
Mag Tape Control (for Uniservo mc Controls)
Mag Tape Control Dual
Tape Adapter Cabinet (for Uniservo mc
Control)
Mag Tape Unit (Uniservo IlIA; 1,000 bpi;
133.000 chari sec)
Mag Tape Unit (Uniservo mc; 200/556 bpi)
Mag Tape Unit (Uniservo IIA Magnetic Tape
Unit)
Translate Option
Mag Tape Unit (Uniservo IVC Tape Unit;
200/556/800 bpi)
800 PPI Option
Power Supply
Paper Tape

*7423-02

Paper Tape Subsystem

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2./69

UNIVAC 1107

78":221.102

IDENTITY OF UNIT

CLASS

INPUTOUTPUT
(Contd. )

Model
Number

Feature
Number

Name

PRICES
Monthly
Monthly
Rental IPurchase Maint.

$

$

2,335

83,250

265

2,335

83,250

265

1,320
535
365

57,500
25,000
17,500

340
205
85

80,000
1,825
2,855 ~25.000

185
245

90,000
36,000
40,500

245
275
275

$

Punched Card
*7240
*7277
*0650-01
*7224-00
*7906-01

Card Control (controls one 7906 Reader and
one 7224 Punch)
Card Control (controls one 7906 Reader and
one 0650 Punch)
Card Punch (300 cards/min)
Card Punch (150 cards/min)
Card Reader (600 cards/min)
Printers

*7299-00
*7299-01
*7319
*0751-01
*7400-01

Printer Control (controls one 0751-01 Printer)
Printer Control Dual (controls two 0751-01
Printer)
Printer Control (controls one 7400-01 Printer)
Printer (700/922 lines/min)
Printer (scientific printer) (102 - symbol
character set including upper and lower case
alphabetic s)

NOTES:
*No longer in production.

2/69

A

AUERBACH

'"

2,085
850
955

£

IA'-

AUERBAC~

785:000.001
su .....

UNIVAC 1108
REPORT UPDATE

EDP
UPOUS

L.-----~----.J

REPORT UPDATE
~

UNIVAC ANNOUNCES 1100 INPUT/OUTPUT SYSTEM FOR 1108 SYSTEMS
UNIVAC on December 9, 1968 announced the 1100 INPUT/OUTPUT System (lOS). This, declared by the manufacturer to be "a major new product enhancement", is a new processing unit
designed to assume input/output processing for the large-scale UNIVAC 1108 including servicing on-site peripherals and remote terminals. These processing functions could include card
handling, line printing, requests for retransmission, communication line termination, message
buffering, translation, data formatting and editing.
The hardware is compatible with the 1108. It has a 900 nanosecond main memory of 32K to 131K
36-bit words, an instruction subset of the 1108 and a typical add time of less than 2 microseconds. Other features include hardware tabling of communications interrupts, buffer chaining,
and up to 16 bi-directional channels with a total channel rate in excess of 2 million words per
second.
The instruction subset includes the instructions necessary for the lOS to perform computation
on its own, but does not include such features of the 1108 set as floating point and double precision which would be of doubtful value.
The various tasks to be performed by the 1100 lOS such as message switching, data reduction
and data manipulation will be programmed by the user. UNIVAC will provide an Assembler
for this purpose.
UNIVAC's marketing strategy is focused on the 1108 user that needs additional input/output
capability. The 16 bi-directional channels provide the high channel rate needed and reduced
usage of the 1108 for I/O. They, UNIVAC, conceive of this device as being a logical step upward for the users of their Input/Output Controllers (IOC), providing all of the functions of the
IOC plus the formatting, editing and translating that was handled by the 1108 processor.
UNIVAC considers that the 1100 lOS will be an effective counter to competitors who would
probably propose a small-to-medium scale general purpose system to serve the same purposes.
The only pricing information released to date is that the purchase prices are set to start at
approximately $185, 000. UNIVAC has scheduled initial deliveries for the fourth quarter of
1969.

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

12/68

785:000.002

1. "'"...
fD..'

AUERBAC~

UNIVAC 1108
REPORT UPDATE

EDP
.HORlS

~

REPORT UPDATE
..

UNIVAC INTRODUCES 1100 FAMILY
On March 28, 1969 UNIVAC announced a new addition to their product line - the 1106. The 1106
heralds the development of the 1108 into a full family of computers that will include the new 1106
and the 1108 together with more powerful models to be announced in the future. The UNIVAC
1106 is a medium to large scale computer with the capability to execute most instructions in one
memory cycle (1. 5 microseconds). Like the 1108, the 1106 is based around a binary single-address
processor. Memory will be available in capacities of 65,536, 98,034, 131,072, 196,608 or
262,144 36-bit words. The Control Memory consists of 128 integrated-circuit registers with a
cycle time of 166 nanoseconds. From both a hardware and a software point of view UNIVAC has
maintained compatibility with the more powerful 1108 with the exception of multi-processing
which is not available on the 1106. This will not only make it possible for the 1106 user to
graduate to the larger system, but may prove attractive to the 1108 user who needs an "1108and-a-half. "
In the competitive market, the 1106 will be confronting the IBM 360/50. UNIVAC would appear
to have a significant price/performance ratio advantage based on internal speeds. The 1106 has
four standard and up to twelve optional I/o channels each within a capacity of 333,000 words per
second or a total system capacity of 667.000 words per second. This is more than twice as
fast as the capacity of the 360/50. UNIVAC, however, does not have available a multiplexer
channel, which means that a low-speed subsystems, such as card reader/punch, will have to be
connected through a high-speed channel, thus negating some of this advantage. Another deficiency of the 1106 with respect to the Model 50 is a direct result of using the 1108 architecture;
it lacks hardware editing, decimal arithmetic, and radix conversion capabilities. This will
necessitate the use of either software routines or a "front-end" processor, especially in the
commercial market, which is the province of the IBM system.
All peripherals currently available with the 1108 will be available on the 1106.
The basic Control Console will include a Display Console (incorporating a CRT display and
keyboard); a Day Clock showing the time of day in hours, minutes and hundredths of minutes;
and the Operator Control and Indicator panel. Instead of the Display Console, a keyboard and
printer may be specified. Other features available are an auxiliary console to accommodate
Communications subsystems and the UNIVAC Pagewriter.
Because of the compatibility of the 1106 with the 1108, all software created for the 1108 is
directly executable on the 1106. This includes both EXEC 2 and EXEC 8 operating systems.
UNIV AC expects that systems dedicated to batch processing, as well as those smaller than
131K, will use EXEC 2.
EXEC 2 is part of a package that was, in its original form, offered to 1107 users as the "B"
software package. Since then it has undergone several enhancements. EXEC 2 is an operating
system that is designed to control the compilation and execution of programs, maximize utilization of the available hardware, and minimize operator intervention. The system uses a magnetic drum as a high-capacity buffer store to keep the card readers, card punches, and printers
fully occupied and as a fast-access auxiliary store for program segments. One notable characteristic of EXEC 2 is its lack of facilities for true multiprogramming; concurrent program
execution is limited to one main program plus multiple data transcriptions (Symbionts). The
EXEC II package also includes, an assembler, a FORTRAN V compiler, LIFT (a FORTRAN II
to FORTRAN V translator), a COBOL '65 compiler, a sort/me.rge program, a tape pre-mount
package, and a variety of applications packages. The minimum configuration required to
operate EXEC 2 on the 1106 is:
•

1 1106 CPU with 65K of memory;

•

1 Operator's console with:
Keyboard and Printer or
Display Console and UNIVAC Pagewriter;

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

3/69

REPORT UPDATE

785:000.003

•

1 FH-432/1782 Magnetic Drum Subsystem with:
3 FH-432 Drums or
1 FH-1782 Drum;

•

1 UNISERVO Magnetic Tape Subsystem with:
2 UNISER VO VI C Magnetic Tape Units or
2 UNISER VO VIII C Magnetic Tape Units;

•

1 Online UNIVAC 9300 System with:
8K storage
132 print position Printer
Multiplexer I/O channel
600 Card/min card reader with Multi-Strobe
Read-Feature
Inter-Computer Control Unit
200 Card/min card punch.

The EXEC 8 Executive System, the chief component of the EXEC 8 package, is a comprehensive
group of routines designed to control all activities of an 1106 computer system, including job
scheduling, hardware allocation, I/o control, and run supervision in a multiprogramming
environment. Other facilities provided by EXEC 8 include library facilities, I/O control, file
control, automatic writing of checkpoints and segmentation. The principal orientation of the
EXEC 8 system is toward maximizing the throughput of batch operations, while providing
facilities for handling useful amounts of real-time and demand processing. Program areas are
protected from the actions of another program (except for I/O operations) by hardware provisions, under control of the Executive. Protection of program areas from the input-output
operations of other programs is accomplished through software checks. In addition to the
Executive, other components of the EXEC 8 system include an assembler, FORTRAN V and
conversational FORTRAN compilers, a COBOL '65 compiler, LIFT (a FORTRAN II to FORTRAN
V translator), a JOVIAL compiler, a sort/merge program, and a large variety of applications
packages. BASIC is available from the UNIVAC Program Library Interchange, but is not supported by UNIVAC. ALGOL and PL/l compilers are not available. The minimum configuration
required to operate EXEC 8 on the 1106 is:
•

1 1106 CPU with 131K of memory;

•

1 Operator console with:
Keyboard and Printer or
Display Console and UNIVAC pagewriter;

•

1 FH432/1782 Magnetic Drum Subsystem with:
3 FH432 Drums or
1 FH1782 Drum;

•

1 FASTRAND Subsystem with:
1 F ASTRAND II or
1 FASTRAND III;

•

1 UNISERVO Magnetic Tape Subsystem with:
2 UNISERVO VIC Magnetic Tape Units or
2 UNISERVO

vIlle

Magnetic Tape Units.

UNIVAC plans to begin deliveries of the 1106 to customers in the first quarter of 1970.
3/69

fA

AUERBACH
@

REPORT UPDATE

785:000.004

UNIVAC 1106 PRICE DATA
Type/Feature
Number

Name

Monthly Rental
(I-Year), $ **

Purchase, $

Monthly
Maintenance, $

7,065

283,230

1,165

545

21,000

60

3011-20

1106 Processor: includes 128 word
control memory; Double Precision
Floating Point; four I/O Channels
with ESI; and Power Distribution
Center. Quarter word ESI logic
may be selected at time of installation. Display Console Type 4009-99
is required.

F0680-99

I/O Channel Expansion: four additional I/O Channels; maximum of
three Expansions may be added
per 1106 Processor

F1272-00

Conversion Kit: permits Processor
Type 3011-20 to be converted to
Type 3011-95, 1108-II, Processor
(field installable; storage units
must be exchanged for Types 700570, -71, -72, and -73)

9,660

346,170

1,160

4009-99

Display Console: includes Control
Console, Entry Keyboard, a CRT
Display capable of displaying 16
lines of 64 characters each and a
freestanding Pageprinter capable
of printing 80 character lines at
25 characters per second

1,000

32,625

250

Multi-Module Storage - 65K:
65,536 36-bit words (2 modules
of 32,768 words. 1. 5-microsecond
cycle time; includes cabinet and
power supply)

4,665

205,875

375

*7005-59

Multi-Module Storage - 98K:
98,304 36-bit words (3 modules of
32,768 words; 1. 5-microsecond
cycle time; includes two cabinets
and power supply. Not field upgradable to 131K; one cabinet and
module must be exchanged for Type
7005-09 Multi-Module Storage Unit)

7,000

308,815

565

*7005-58

Multi-Module Storage - 131K:
131,072 36-bit words (4 modules of
32,768 words; 1. 5 microsecond
cycle time. Includes two cabinets
and power supply)

9,325

411,750

745

*7005-57

Multi-Module Storage - 196K:
196,608 36-bit words (6 modules of
32,768 words; 1. 5-microsecond
cycle time; includes 3 cabinets and
power supply)

13,985

617,625

1,120

*7005-56

Multi-Module Storage - 262K:
262,144 36-bit words (8 modules of
32,768 words; 1.5 microsecond
cycle time; includes 4 cabinets and
power supply)

18,645

823,500

1,490

*7005-60

* 1. 5-microsecond core with overlap capability.
* * UNIVAC does not include maintenance prices in the rental charges in their published price data; for
convenience of comparison, maintenance charges have been included in the monthly rental column.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

3/69

-A.

pt,

-

AU£RB"~\

785:0 II. 100

"'....,
ED]?

UNIVAC 1108
INTRODUCTION

units

•

SUMMARY: UNIVAC 1108
.1

BACKGROUND
The UNIVAC 1108 computer system is a large-scale data processing system, oriented toward both scientific and business applications, that features multiprogramming and multiprocessing capabilities for increased hardware utilization. Monthly rentals for typical
single-processor 1108 configurations generally fall in the $45,000 to $60,000 range. The
cost of an 1108 system is typically lower than the original price of an equivalent UNIVAC
1107 system - the 1108's predecessor in the UNIVAC line - despite the 1108's five-fold
advantage in internal speeds. Typical multiple-processor configurations rent for upwards
of $100,000 per month.
The UNIVAC 1108 was originally announced in July 1964 as a faster, expanded version of
the UNIVAC 1107 (see Computer System Report 784). Shortly afterward, UNIVAC stated
that the development of multiple-processor hardware for the 1108 was under way. This
hardware was first documented and described to prospective users in August 1965, but its
official announcement was delayed until December 1965, when the multiple-processor version was introduced as the UNIVAC 1108-II. This, however, is only a marketing designation;
the single-processor and multiple-processor versions of the 1108 use the same hardware
components, and both versions are described in this report. The first single-processor
1108 system was delivered in December 1965, and the first multiple-processor configuration
was delivered in the third quarter of 1967.
The 1108 is currently enjoying widespread acceptance within the large-scale computer market. Certainly, some of this success is due to the setbacks suffered by other manufacturers
in implementing large and complex programming systems on their third-generation computers. Another factor Significantly enhancing acceptance of the 1108 is its time-tested
software and proven ability to function effectively in a variety of scientific and commercial
applications.
UNIVAC has had several years of experience in multiprogramming (i. e., the capability to
maintain several independent programs in core storage at the same time and to switch control among them to more fully utilize the hardware). This experience was gained with the
UNIVAC III, 490, and 1107 computer systems. Many new features in the 1l08, not found in
other UNIVAC systems, are oriented toward optimizing the use of the hardware through
multiprogramming.
The 1108 is largely program-compatible with the 1107, although existing 1107 programs will
need to be reassembled or recompiled to run under control of the 1108 EXEC 8 Operating
System. Most existing 1107 object programs can be run directly on an 1108 under control
of a modified version of EXEC II, one of the standard operating systems for the 1107.

Figure 1. An 1108 System showing main operator's console.
© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

12/67

UNIVAC 1108

785:011. 200

.1

BACKGROUND (Contd.)
Chief extensions of the 1108 with respect to its predecessor, the 1107, include:
•

Expanded core memory capacity - up to 262,144 36-bit words.

• Significantly faster core memory and internal processing speeds basic cycle time is 0.75 microsecond.
•

Double-precision fixed-point and floating-point arithmetic facilities.

•

Greatly improved memory protection and addressing techniques.

•

Provisions for I/O Controller units that can access memory independently of the Central Processor(s).

•

Capability for up to five Central Processors and I/O Controllers to
share a common core memory.

•

Provisions for up to eight independent core memory modules •

•2

HARDWARE

.21

System Configuration
A single-processor 1108 configuration consists of one 1108 Central Processor with console
and 8, 12, or 16 input-output channels; two, four, six, or eight 32, 768-word modules of
core memory (a maximum of 262,144 words) ; and peripheral subsystems as required. Tables I and II summarize the available subsystems and the number of input-output channels
that each subsystem requires. The console also requires one input-output channel. In addition, up to three I/O Controller units, each having 4, 8, 12, or 16 input-output channels,
can be included.
A multiprocessor (1108-II) configuration consists of up to five 1108 Central Processors and

I/o Controllers in any combination; four, six, or eight 32, 768-word modules of core mem-

ory (a maximum of 262,144 words) ; and peripheral subsystems as required. The inputoutput channels provided by the I/O Controllers are in addition to the Processor I/O channels.
Normally, in a multiprocessor configuration, critical peripheral subsystems such as magnetic drums are connected to both an I/O Controller channel and a Processor channel to
provide redundant data paths for increased reliability. Although up to five Central Processors can share the same core memory, the operating system being provided by UNIVAC
for the 1108 will contain provisions for a maximum of three Processors and two I/O Controllers •
. 22

Core Memory
Core. memory can consist of up to 262,144 word locations in increments of 65,536 words.
Each 36-bit word location can hold one instruction, one single-precision floating-point data
item, from one to six fixed-point data fields, four 8-bit bytes (quarter-words), or six
alphameric characters. Unlike the UNIVAC 1107, which has no checking provisions for
data stored in core memory, the 1108 records a parity bit with each half-word of memory.
Each 32, 768-word module can be accessed independently.
The basic cycle time of the 1108's memory is 0.75 microsecond - over five times as fast
as the 1107's 4-microsecond cycle. Like the 1107, the 1108 Central Processor can simultaneously access two different memory modules. If instructions and operands are stored
in separate modules, the effective execution time for most instructions can be reduced by
one memory cycle (0.75 microsecond). In all multiprocessor (1108-II) configurations, the
physical address locations are interleaved within each pair of memory modules; the even
locations are in one module and the odd locations in the other. Such interleaving reduces
the frequency of conflicts when two Central Processors are executing programs that are
physically located in the same memory modules .

• 23

Control Memory
The 1108's Control Memory, which corresponds to the 1107's Thin-Film Memory, consists
of 128 36-bit word locations. The Control Memory utilizes integrated circuits and has a
cycle time of 0.125 microsecond. In the 1108, 70 of the 128 locations are reserved for use
by supervisory routines; these reserved locations include a separate complete set of index
registers, arithmetic registers, and control registers, as well as the Input/Output Access
Control Registers. The 48 locations available to the user's program include 15 index registers, 16 arithmetic registers, aJ.l.d 4 control registers; the remaining 17 locations can be
used by the programmer for intermediate storage. In both the reserved and user's area
of Control Memory, four locations can be used as eitner index registers or arithmetic registers, permitting sOl1le unusual and powerful address modification operations •

• 24

Central Processor
The UNIVAC 1108 Central Processor can perform fixed-point and floating-point arithmetic
on one-word or two-word binary operands (although double-precision fixed-point arithmetic

12/67

A

(Contd.)

AUERBACH

co

SUMMARY

• 24

785:011. 240

Central Processor (Contd.)
is limited to addition and subtraction). The 16 arithmetic registers, 15 index registers, a
versatile repertoire of 7-part instructions, recursive indirect addressing, and a partialword transfer facility permit efficient processing of most scientific and commercial applications, although commercial processing will be somewhat less efficient because the 1108,
like the 1107, lacks automatic facilities for editing, decimal arithmetic, and radix conversions.
Although the 1108 uses a one-address instruction format, a limited two-address capability
is provided since most instructions can specify the use of anyone of the 16 arithmetic registers. The partial-word load and store instructions can transfer any half, third, quarter,
or sixth of a word to or from the least significant bit positions of any arithmetic register.
A wide variety of logical, shift, search, and block transfer operations can be performed.
All instructions can be indexed, and each index register can be automatically incremented or
decremented each time it is referenced. Indirect addresses can be "chained," and indexing
can be performed upon each address in the chain. Straightforward programming of the
UNIVAC 1108 is not unusually complex, but only skilled, highly-trained programmers will
be able to take full advantage of the powerful optional elements offered in most instructions.
A program interrupt facility causes a transfer of control to one of 42 dynamically reassignable core memory locations upon completion of an input-output operation, upon detection of
a Processor or input-output error, or upon count-down to zero of the real-time clock (whose
contents are decremented every 200 nanoseconds). A programmable day clock that can interrupt the executive system is also provided. The interrupt facility permits full utilization of
the Central Processor and all peripheral devices under the control of an integrated operating
system that handles multiprogrammed operations.
Like other recent computer systems designed for multiprogramming, the 1108 Central Processor can operate in one of several modes, which vary in the facilities they make available
to the program.
In the Guard mode, relative addresshg is in effect; i. e. , program addresses are modified
by the contents of Basing Registers prior to all operations that require access to core memory. The upper and lower program address limits are specified by the contents of the Storage
Limits Register. Any reference to the reserved set of registers in the Control Memory, any
attempt to read or write into areas outside the program limits, or any attempt to execute a
reserved instruction results in an interrupt. The reserved instructions include loading of
certain control registers and all input-output instructions. All users' programs are executed
in the Guard mode.
The Privileged mode is similar to the Guard mode, except protection is provided only against
writing, not against reading. Thus, programs stored in protected locations can be executed
but cannot be overwritten.
In the Open mode, the complete facilities of the processor are available; there are no restrictions on the core locations accessed or on the instructions used, and relative addressing
is not in effect. Only the supervisory routines are allowed to operate in this mode .

. 25

Peripheral Equipment
Three different magnetic drum units are available for use in 1108 systems. Two, the FH-432
and FH-1782, are rapid-access, word-addressable units designed to facilitate the rapid exchange of programs or routines between core storage and drum storage. One FH-432 drum
subsystem or equivalent with at least 786,000 words of storage is required for use of the
standard EXEC 8 Operating System. The third drum storage unit, Fastrand II, is also used
with several other UNIVAC computer systems. Fastrand II employs movable access mechTABLE I' UNIVAC 1108 MAGNETIC DRUMS
FH-432

FH-1782

Fastrand II

Average access time, msec

4.25

17

92

Peak transfer rate, words/sec

240,000

240,000

2,097,152

16,777,216

176,160,768

9*

8*

8

1 or 2

1 or 2

1 or 2

Characteristic

Maximum storage per subsystem,
36-bit words
Maximum drum units per subsystem
Number of input-output channels
per subsystem

25,625

* Up to 8 FH-432 and FH-1782 drum units, in any combination, can be connected to the
same controller.
© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

12/67

785:011.250

. 25

UNIVAC 1108

Peripheral Equipment (Contd. )
anisms to provide somewhat slower access to much larger quantities of data than the headper-track FH-432 and FH-1782 drums. A summary of the principal characteristics of each
of these drum units is contained in Table I. Notably absent from the current UNIVAC line
of random-access peripheral devices is a changeable-cartridge unit.
UNIVAC now offers only mM-compatible magnetic tape units in the standard line of peripheral
devices for the 1108. The characteristics of the Uniservo VIC and vmc Magnetic Tape
Handlers are summarized in Table II. Both tape handlers are available in 9-track versions
compatible with IBM's 2400 Series Magnetic Tape Units, as well as in the more common,
mM 729-compatible 7-track versions.
UNIVAC offers both multi-line and single-line controllers to serve as interfaces between an
1108 computer system and communications lines. The Communication Terminal Module
Controller is capable of controlling up to 32 half-duplex or full-duplex narrow-band or voiceband lines. Various Communication Terminal Modules permit communications with remote
terminals operating at up to 4,800 bits per s~cond, synchronously or asynchronously, and
utilizing transmission codes of up to eight levels. Up to four Communication Terminal
Module Controllers can be included in a Communications Subsystem, via a Scanner/Selector
Unit; each Communications Subsystem fully occupies one 1108 input-output channel. All
communications lines can be active simultaneously, subject to a maximum data rate of
51,000 characters per Communications Subsystem.
The Word Terminal Synchronous (WTS) and Communications Terminal Synchronous (CTS)
are single-line controllers capable of controlling data communications over a single voiceband line at 2,000 or 2,400 bits per second, or over a broad-band line at 40,800 bits per
second. Each Terminal fully occupies one 1108 input-output channel. The chief difference
between the twc:> 1'erminals is that the WTS transfers data in units of one 36-bit word between
the 1108 and the controller, whereas the CTS transfers data in units of one character. With
TABLE II: UNIVAC 1108 INPUT-OUTPUT SUBSYSTEMS
Subsystem

12/67

Maximum Number
of Devices per
Subsystem

Number of I/O
Channels per
Subsystem

Peak Speed

Uniservo VIC
Magnetic Tape

lor 2

16

34,200 char/sec.

Uniservo VIIIC
Magnetic Tape

lor 2

16

96,000 char/sec.

Punched Card

1

1 reader;
1 punch

Printer

1

4

Punched Paper
Tape

1

1 reader;
1 punch

UNIVAC 1004

1

1

Communication
Controller
(multi-line)

I

Communication
Controller
(single-line)

I

UNISCOPE 300
Visual Communication
Terminal

1

4 multiplexors,
each serving up
to 32 half- or
full-duplex lines
1

24 (I6-line)
48 (8-line)

A

read 900 cpm;
punch 300 cpm.
16001pm.
read 1000 char/sec;
punch 240 char/sec.
cards: read 615 cpm,
punch 200 cpm;
print 600 lpm;
paper tape: read 400
char/sec; punch 110
char/sec; magnetic
tape: 34,200 chari
sec.
4800 bits/sec per
line; 51, 000 char/
sec total.
40,800 bits/sec.

400 char/sec.

(Contd.)

AUERBACH

e

SUMMARY

.25

785:011. 260

Peripheral Equipment (Contd.)
appropriate adapters or features, any of these communications controllers can operate over
the public telephone network and can be equipped for unattended operation and automatic dialing.
The other peripheral devices available for use with an 1108 computer system are listed in
Table II.
UNIVAC 1107 users who are installing an 1108 system can carryover the same magnetic
drum and magnetic tape units they are currently using with the 1107. Specific provisions
are included in the standard 1108 software for handling FH-880 Magnetic Drums and Uniservo
lIA, lIlA, mc, and IVC Magnetic Tape Handlers, even thc..ugh these 1107 peripheral units
will not be actively marketed with the 1108 .

. 26

Simultaneous Operations
The 1108's capability for simultaneous operations is high. In addition to the overlapped operations of one or more Central Processors, all input-output channels in the Central Processors and I/o Controllers can be active simultaneously, subject to the peak data rate
limitations of each channel and the associated Processor or I/O Controller. Each channel
can handle a maximum of 250,000 data transfers per second; maximum capability of each
Processor or I/o Controller is L 33 million transfers per second. Most data transfers
consist of one 36-bit binary word (or six 6-bit data characters), in which case each Processor or I/O Controller can handle up to 8 million characters per second. Some peripheral
units, such as the Punched Paper Tape Subsystem, Communications Subsystem, and Communications Terminal Synchronous, transfer only one character per data transfer. In general, only one core storage cycle is required for each data transfer •

•3
• 31

SOFTWARE
EXEC II Software
Prior to delivery of the full 1108 software package, as outlined in Paragraph. 32, UNIVAC
is supplying 1108 users with a slightly modified version of the ttBtt Software Package for the
1107. The principal components of this package are:
•

EXEC II - an operating system designed to monitor the compilation
and execution of programs, maximize utilization of the available hardware, and minimize operator intervention. The system utilizes a
magnetic drum as a high-capacity buffer store to k.:ep the card readers,
punches, and printers fully occupied and as a fast-access auxiliary store
for program segments. An integrated set of diagnostic aids and library
maintenance facilities is included. One notable characteristic of EXEC II
is its lack of facilities for true multiprogramming; concurrent program execution is limited to one main program plus mUltiple data transcriptions.

•

SLEUTH II - a symbolic assembly system, with macro-instruction facilities, that translates symbolic source programs into relocatable machinelanguage object programs. SLEUTH II is synonymous with 1107 Assembler.

•

COBOL - a compiler for COBOL-61 source programs that operates under
control of EXEC II. Language facilities include nearly all of Required
COBOL-61 (there are a few minor deficiencies) ; several COBOL-61 electives (but not the very useful COMPUTE verb); a MONITOR verb (which
provides dynamic printouts of the values of specified items) ; and a SORT
facility (but not the one defined as part of Extended COBOL-61). A magnetic drum is required for COBOL compilations, but magnetic tape is
not.

•

FORTRAN - a compiler for FORTRAN V source programs that operates
under control of EXEC II. Language facilities are largely compatible with
those of FORTRAN IV as defined for the IBM 7090/7094. FORTRAN II
source programs can be converted to FORTRAN V by means of a LIFT
Translator. The FORTRAN compiler achieves high compilation speeds
through use of a magnetic drum.

•

SORT II - a generalized sort/merge routine that operates under control
of EXEC II.

•

Applications Packages - a comprehensive set of application programs
has been developed for the 1108 by both UNIVAC and users' groups; the
set includes: APT III, PERT, Linear Programming, Simscript, Simula,
Mathpack, Statpack, and GPSS II General Purpose Simulator. All can
run under control of EXEC II or EXEC 8.
With this software, the 1108 will, in effect, perform as an 1107 that is over five times as
fast internally as the original 1107; the peripheral performance is determined by the 1108's
configuration. UNIVAC states that the majority of 1107 object programs can be executed
directly by an 1108 under control of the modified EXEC II system. Significant exceptions
© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

12/67

785:011.310

.31

UNIVAC 1108

Interim Software (ConW.)
include some 1107 FORTRAN and COBOL programs, because the 1107 FORTRAN and
COBOL compilers used "illegal operation code" interrupts to provide entries into certain
standard functions. Such programs would not, in general, be executed properly by an 1108.
The "B" Software Package for the 1107 was developed by Computer Sciences Corporation
and is serving as a foundation for the 1108 software development program. UNIVAC states
that certain portions of the EXEC 8 software package described below, including 1108
COBOL, n08 FORTRAN V, and 1108 SORT/MERGE, are capable of operation on an 1108
under control of the modified EXEC II operating system .

. 32

EXEC 8 Software
The chief component of the standard software being prepared for the 1108 is the EXEC 8
Executive System, a comprehensive operating system designed to control all activities of
an 1108 installation. The software support package for the 1108 will include an assembler,
compilers for COBOL, FORTRAN, and ALGOL, and several useful application programs.
The scheduled delivery date for the full 1108 software package is the first quarter of 1968.
Most 1108 software, with the exception of the Executive System, will be similar to, or direct
extensions of, 1107 software. Programs written for an 1107 will, in general, need to be
recompiled or reassembled on an 1108 in order to run under supervision of the 1108 Executive System.
The EXEC 8 Executive System is a comprehensive group of routines designed to control all
activities of an 1108 computer system, including job scheduling, hardware allocation, I/O
control, and run supervision, in both a multiprogramming and a multiprocessing enviroment.
Other facilities provided by EXEC 8 include library facilities, I/O control, file control,
automatic writing of checkpoints, and segmentation.
EXEC 8 recognizes three types or levels of processing: real-time, demand, and batch.
Real-time processing is characterized by the need for a computer response to an external
event quick enough to achieve a desired goal. Real-time processing is normally, but not
exclusively, associated with data communications or process control applications where
delay in obtaining computer time could result in lost data or process malfunctions. Demand
processing is typified by the need for "conversation" between the computer and the user;
i. e. , the user will specify the execution of certain tasks dependent on the results of previously-initiated tasks. Batch processing is the normal execution of independent tasks
(programs) or groups of tasks that are not highly time-dependent. The order of priority for
scheduling and execution, in descending order, is real-time, demand, and batch.
The principal orientation of the EXEC 8 System is toward maximizing the throughput of batch
operations, while providing facilities for handling useful amounts of real-time and demand
processing. The type of processing is specified in the control statements initiating a run,
and sometimes within each task of a run (i. e., the type of processing can vary for each task
within a run).
Program areas are protected from the actions of another program (except for I/O operations)
by hardware provisions, under control of the Executive. Protection of program areas from
the input-output operations of other programs is accomplished through software checks.
The EXEC 8 Executive System can be utilized on any 1108 configuration incorporating at
least 786,000 words of FH-432 Magmetic Drum storage or equivalent. The Executive System contains provisions for handling any 1108 configuration that includes up to three Central
Processors and two I/O Controllers. The minimum resident core storage requirement is
20,000 words.
The other major items in the software support package for the 1108, which operate under
control of the EXEC 8, include:

12/67

•

1108 Assembler - a symbolic assembly system that is virtually
identical with SLEUTH II for the 1107 (see Paragraph. 31), with
additional mnemonics for the new 1108 instructions.

•

1108 COBOL - a compiler for programs written in COBOL-61.
Language facilities include those of Required COBOL-61, except
for a few minor deficiencies, and many COBOL-61 electives, including the COMPUTE verb and the extended version of the SORT
verb.

•

1108 FORTRAN - a compiler for programs written in a language
that UNIVAC calls "FORTRAN V." The language facilities represent
significant extensions of FORTRAN V as implemented for the 1107,
including proviSions to facilitate the writing and deletion of debug
statements, and to assign types implicitly, according to the first
letters of variable names. The 1108 FORTRAN V language includes,
as proper subsets, all the language facilities of 1107 FORTRAN IV,
IBM 7090/7094 FORTRAN IV, and the USASI FORTRAN language.

IA
AUERBACH

'"

(eontd. )

SUMMARY

• 32

785:011. 320

1108 Software (Contd.)

FORTRAN II source programs can be accommodated through use
of the LIFT translator; LIFT converts the source-language statements
into 1107 FORTRAN V statements, which can then be compiled by the
1108 FORTRAN V compiler. Two distinct versions of the FORTRAN V
compiler are offered for the 1108. One is a fast, efficient compiler
for batch programs. The second is an interactive, "conversational
mode" compiler for servicing users, who desire statement-by-statement program execution at remote terminals.
•

ALGOL - a compiler for programs written in ALGOL 60; its
language facilities conform to the ACM and GAMM committee
specifications.

•

1108 SORT/MERGE - a generalized subroutine used in conjunction
with a series of parameter lists to produce sor~rograms. The
complete program specifications can be entered via the control
stream or can be incorporated into a larger program. Fastrand II
magnetic drum storage can be utilized to speed sorting.

Application packages available for the 1108 include: Linear Programming, PERT/COST,
APT III (for computer-assisted programming of numerically-controlled machine tools),
BEEF (an extensive series of subroutines developed by Westinghouse Electric Corporation's
Baltimore Defense and Space Center to enhance FORTRAN's capabilities as a scientific
processing language), Mathpack routines, Statpack routines, several general-purpose
system simulators (GPSS II and SIMULA) , a Biomedical Support package (BIOMED), an
analog simulator (MIMIC), and a powerful matrix manipulation package (BEMAT).

© 1967 AUERBACH Corporation and AUERBACH Info. Inc.

12/67

785:221.101

A

AUlIlACH

snl""

EDP

UNIVAC 1108
PRICE DATA

"Pins

UNIVAC 1108
UNIVAC does not include monthly maintenance charges in their published prices; these charges have been
included in the Monthly Rental colwnn in this Digest to permit convenient comparison with other systems.

PRICES

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

PROCESSOR

Name

Monthly
Monthly
Rental Purchase Maint.
$
$
$ (1)

Processing Unit
3011-95
Fl053-99
F0680-01
*3011-99
*3011-97
*FI052-00
*FI053-00
*F0680-00

1108-IT Processor (requires Display Console;
includes 8 I/O channels)
Multiprocessor Capability(2)
I/O Channel Expansion (additional 4
channels; maximum of 16 channel total
per Processor)
1108 Central Processor
110SA Unit Processor
Multi-Processor Feature
Multi-Processor Feature
I/O Channel Expansion

16,725

629,400

2,325

200
545

8,700
21,000

-

16,890
17,225
4,700
4,700
545

633,000
646,800
204,500
204,500
21,000

2,400
2,425

1,000

32,625

250

165

6,600

10

165

6,600

10

11,190

457,500

745

22,230
33,290
44,255
11,040

915,000
1,372,500
1.830,000
457.500

1,345
1,965
2,490
600

11,060

457,500

620

10,965

457,500

525

10,975
21,800
11.190
22,230
33,295
44,260

448,000
897,000
457,500
915,000
1,372,500
1,830,000

735
1,325
745
1,345
1,965
2,490

60

-

60

Console
4009-99
F0774-00
F0774-01

Display Console (includes keyboard, page
printer, and CRT display)
Auxiliary Console (left- or right-hand
additi.on; accommodates 1 to 4 CTMC's)
Auxiliary Console (same as F0774-00
except located between consoles of two
Processors)
Main Storage

7005-7:\
7005-72
7005-71
7005-70
7005-6:1
7005-64
7005-65
*7005-94
*7005-9:1
*7005-90
*7005-89
*7005-92
*7005-91

Storage (65K words; for use by EXEC II
operating system only)
Storage (13K words)
Storage (196K words)
Storage (262K words)
Storage Expansion (65K words; for field
expansion from 65K to l3lK)
Storage Expansion (65K words; for field
expansion from l3lK to 196K)
Storage Expansion (65K words; for field
expansion from 196K to 262K)
Core Memory (65K words)
Core Memory (131K words)
Core Memory (65K words)
Core Memory (l31K words)
Core Memory (196K words)
Core Memory (262K words)

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

UNIVAC 1108

785:221.102

PRICES

IDENTITY OF UNIT
CLASS

Model
Number

ATTACIIl\IENTS,
ADAPTERS,

Feature
Number

Na.me

Monthly
Monthly
Rental Purchase Matnt.
$
$
$ (1)

For Multiprocessor Configurations
5013-00

AND

ell .·\NNELS

F0778-00
F0790-00
F0832-00
F0832-01
F0833-00
F083:l-01
25(16-00

F0874-00
0954-99
F0879-00

()~55-0.f

0:153-05
F0789-00
F0789-01
F0597-03
FlOG7-00
F0597-97

I/O Controller (includes 4 I/O channels and
256-word index storage; can address 131K of
storage via MMA units)
I/O Channel Expansion (additional 4 I/O
channels; maximum of 16 I/O channels
per I/O Controller)
Storage Expansion (additional 256-word
index storage)
Processor Interface (provides interface for
second Processor)
Processor Interface (provides interface for
third Processor)
MMA Interface (extends core addressing
capability to 196K)
MMA Interface (extends core addressing
capability from 196K to 262K)
Availability Control (ACU) (includes control and
console; displays system status and controls
system configuration; provides access to 6
SPI's, 4 MMA's, 3 Processors, and 2 I/O
Controllers. )
ACU Expansion (provides access to 6
additional SPI's; maximum capacity of
ACU is 24 SPI's total)
Multi-Module Access (MMA) (permits' access
of 2 Processors and 1 I/O Controller to
1 65K Storage Module)
MMA Expansion (increases capability for
access to storage Module to 3 Processors
and 2 I/O Controllers)
Shared Peripheral Interface (SP!) (36-bit interface that permits control of a peripheral
subsystem bv multiple Processors or I/o
controllers):
"PI (for control by 2 Processors or r/o
controllers; includes cabinet with space
for second SP!)
SPT (same characteristics as 0955-04 and
occupies the extra space)
SPI Expansion (adds third interface)
SPI Expansion (adds fourth interface)
1004 Control (36-bit interface for direct communicatlOn with on-site UN IV AC 1004)
Redundant Read Capability
1004 Control (:l6-bit interface for direct COIDmunication with on-site UNIVAC 1004;
includes Redundant Read Capability)

4,120

174,000

120

5ilO

22,620

60

705

32,450

120

5,220

120

5,220

GO

2,610

60

2,610

1,315

52,200

115

75

2,830

10

1,350

56,550

50

60

2,610

475

19,575

25

420

17,400

20

85
60
205

3,480
2,395
7,840

5
5
20

65
270

2,390
10,230

10
30

168

6,540

18

405

12,740

6ll

405

12,740

60

598

20,000

88

-

-

Other
F089:J-00

:2,~O8-00

~,)0i)-01

2')08-02

2/69

Transfer Switch (permits manual switching of
peripheral subsystems and processors
between I/O Channels; 36-bit interface;
switches can be cascaded if desired)
Switch Cabinet (contains power supply. integral
operators panel, and control unit for up to 5
Transfer Switches in any combination)
Switch Cabinet (same as 2508-00 but with remote
operator panel)
Switch Cabinet (same as 2508-00 but i.neludes
auxiliary power supply and can accommodate
only 4 Switches)

781:221.1 03

PRICE DATA

IDENTITY OF UNIT
CLASS

ATTACHMENTS,
ADAPTERS,
AND
CHANNELS
(Cont'd)

Model
Number

Feature
Number

Name

PRICES
Monthl~

Monthly
Rental IPurchase Maint.

_1

(1)

$

$

Other (Cont'd)
2508-03

Switch Cabinet (same as 2508-01 but includes
auxiliary power supply and can accommodate
only 4 Switches)

MASS
STORAGE

598

20,000

88

FH 880 Drum (786,432 36-bit words)
2,150
FH 880 Control (controls 1 to 8 FH 880 Drums)
1,550
FH 432 Drum (bit words)
1,070
FH 1782 Drum (bit words)
2,940
Dual Channel (for FH 432)
70
Dual Channel (for FH 1782)
70
FH 432/FH 1782 Control (controls 1 to 8
2,145
FH 432 or FH 1782 Drums in any combination)
Write Lockout
30
440
Shared Peripheral Interface (provides
access by multiple Processors or
I/O Controllers in a multiprocessor
system):
SPI (second interface)
440
SPI (third interface)
70
SPI (fourth interface)
55
Fastrand II (44 Million - 18-bit words)
4,050
Fastbands
215
Write Lockout
30
Fastrand II Control (unbuffered control for
1,280
1 to 8 Fastrand II Drums)
Fastrand II Control (buffered control for 1 to
1,335
8 Fastrand II Drums)
Fastrand II Dual Control (dual unbuffered con!l,050
trol for 2 to 8 drums includes dual access
for first 2 runs)
Fastrand II Dual Control (dual buffered con3,160
trol for 2 to 8 drums includes dual access
for first 2 drums)
Search all words (for buffered controls only)
55
Control B.lffer (for field conversion from un55
buffered to buffered control)
Dual Access Drum Adapter (required for attach50
ment all subsequent drums to a Dual Control,
except fifth)
Dual Access Drum Adapter (required for attach50
ment of fifth drum to a Dual Control)
Fastrand III Drum
4,965
Fastrand III Control
1,565
Fastrand III Dual Control (includes dual
3,650
access adapter for first drum only)
Fastrand TIl Control
2,085
(Field Upgrade from Single to Dual Access)
Control Buffer
55
Fastband
205
Write Lockout
30
Fastrand TIl Dual Access Drum Adapter
50
(for all subsequent drums except fifth)
Fastrand III Dual Access Drum Adapter
50
(for fifth drum)

85,165
58,800
42,435
117,210
2,255
2,255
82,515

190
190
100
260
15
15
260

1,040
17,805

5
25

17,905
2,675
2,060

25
5
5

8,235
1,040
51,060

25
5
115

53,410

115

122,330

240

127,030

240

Drum Storage
7304-01
7427-03
6016-00
6015-00
F0786-01
F0767-00
5012-00
F0929-00
F0930-00

F0930-00
F0930-01
F0930-02
6010-00
F0686-01
F0688-01
5009-00
5009-04
5009-99
5009-98

* F0710-00
F0763-00
F0959-99
F0959-98
6010-10
5009-89
5009-85
5009-81
F0763-01
F0686-0l
F0688-0l
F0959-97
F0959-96

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

-

2,255
2,355

-

1,830

10

1,830

10

200,800
62,220
147,010

350
135
270

84,790

135

2,350
8,235
1,040
1,830

-25

1,830

10

5
10

2/69

785:221.104

UNIVAC 1108

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

Name

Monthly
Monthly
Rental ~rchase Maint.
$ (1
$
$

Magnetic Tape

INPUT OUTPUT
5008-00
5008-98
F0627-04
F0627-99
F0706-00
0858-00
0858-08
0858-01
0858-10
0858-12
0858-14
FI021-00
FlO72-00
FI072-01
5008-12
5008-89

5008-83
5008-79

Uniservo VIC:
Uniservo VIC Control (controls up to 4
Master units)
Uniservo VIC Control (controls up to 4 Master
Units; simultaneous read-read/read-write;
requires I/O channels)
Translator (for 5008-00 or -08)
Translator (for 5008-98 or -91)
9 Track Capability (required in 9-track controis; (1 for non-simultaneous, and 2 for
simultaneous controls)
Uniservo VIC Master (7-track, 8,540/23,741/
34,160 char/sec; can handle up to three
0858-01 slave units)
Uniservo VIC Master (Same as 0858-00 but with
provisions for simultaneous read-read/
read-write)
Uniservo VIC Slave (7-track)
Uniservo VIC Master (9-track; 800 bpi; 34,160
bytes/sec; handles up to three 0858-14 Slave
units)
Uniservo VIC Master (same as 0858-10 but
with provisions for simultaneous read-read/
read-write)
Uniservo VIC Slave (9-track)
7 to 9 Track Conversion (converts 7-track tape
unit to 9-track unit)
Simultaneous Capability (converts 0858-00 to
0858-08)
Simultaneous Capability (converts 0858-10 to
0858-12)
Uniservo VIDC:
Uniservo VIDC Control (controls 1 to 16
Uniservo VIDC units)
Uniservo VIDC Control (controls 1 to 16 Uniservo VInC units; provides read-read,
read-write, write-read, and write-write
simultaneously; require 2 I/O channels)
Control Expansion (converts 5008-12 to 5008-89)
Control Expansion (converts 5008-20 to 5008-87)

F0704-00

0859-00
0859-04
0859-08
0859-10

2/69

PRICES

VIC Capability (allows 7-track VIC and VIIIC
units to be intermixed on a 5608-12 control)
F0706-00 9 Track Capability (1 required for nonSimultaneous, 9-track controls; 2 required
for Simultaneous, 9-track controls)
F0627-04 Translator (for 5008-12)
F0627-98 Translator (for 5008-fi!}1
Uniservo vmc (7-track; 24,000/66,720/96,000
char/sec; non-simultaneous)
Uniservo VIDC (9-track; 800 bpi; 96,000 bytes/
sec; non-simultaneous)
Uniservo vmc (7-track; same as 0859-00, but
for simultaneous controls)
Uniservo vmc (9-track; same as 0859-04, but
for simultaneous controls)
F0999-00 7 to 9 Track Conversion (converts 0859-00 to
0859-04)
F0999-04 7 to 9 Track Conversion (converts 0859-08 to
0859-10)

A

AUERBACH
co

745

31,070

35

1,490

62,140

70

110
220
50

4,410
8,820
1,960

5
10
5

515

17,350

115

570

19,800

115

310
515

10,470
17,350

70
115

570

19,800

115

310

10,470

55

2,450

55

2,450

-

1,550

62,430

120

3,100

124,860

240

1,550
1,550

62,430
62,430

120
120

85

3,385

5

50

1,960

5

110
220
860

4,410
8,820
32,735

5
10
110

875

33,390

110

890

34,045

110

905

34,700

110

15

655

15

655

-

-

70

-

785:221.105

RICE DATA

PRICES

IDENTITY OF UNIT
CLASS

INPlJTOllTPUT
(Cont'd)

Model
Number

Featw-e
Number

Name

Monthly
Monthly
Rental !Purchase Maint.

$

(1)

$

$

Punched Card
5010-00
0706-97
*0706-99
F1022-00
0600-00

Card Control (controls one 0600 Card Punch
and one 0706 Card Reader)
Card Reader (900 cards/min; includes
check read feature and alternate stacker)
Card Reader (900 cards/min)
Check Read (second read station for crecking)
Card Punch (300 cards/min)

920

28,620

265

480

15.385

125

415
65
715

12,985
2,400
21,560

115
10
230

885
630

30,015
23,055

200
100

440

17,905

25

70
55
810
75

2,675
2,060
27,400
2,575

5
5
180
20

810

27,440

180

1,250

43,500

305

870

25,970

275

705

24,700

135

260
65
85
100
20
40
40
80
95

9,020
2,255
2,895
3,630
635
1,315
1,315
3,050
3,480

50
15
15
15
5
10
10
10
15

125
60
70
80
80
95
60
40
20

4,570
2,175
2,610
3,050
3,050
3,480
2,175
1,525
650

20
10
10
10
10
15
10
5
5

80
40
40

3,050
1,525
1,525

10
5
5

Printer
--5011-00
F0751-00
F0933-00

F0933-01
F0933-02
*7299-03
F0965-00
*7299-04
0758-00
*0755-00

COMMUNICATIONS

F0900-06
F0906-06
F0901-04
F0902-02
F0903-02
F0905-00
F0904-00
F0904-01
F0988-00
F0988-01
F099I-00
F0989-00
F0989-01
F0989-02
FI027-00
F1048-00
F1018-02
FIOI8-03
F1018-05
FIOI9-0!
F1019-03
FI019-05

Printer Control (controls one printer)
Printer Control Expansion (provides control
for second printer)
Shared Peripheral Interchange (SPI):
SPI (provides interface for second processor
or I/O Controller in multi-processor
system)
SPI (provides third interface)
SPI (provides fourth interface)
Printer Control (controls one 0755 printer)
Printer Control Capability (permits 7299-03
Control to handle 0758-00 Printer)
Printer Control (used with 7299-04 to control
second printer)
Printer (1,600 lines/min with 43-character
set; 1,200 lines/min with 63-character set)
Printer (700 lines/min alphanumeric; 900 lines/
min numeric)
Communication Terminal Module Controller
(CTMC)
Spare CTM Controller
CTM-LS Low Speed
CTM-MS Medium Speed
C'fM-HS High Speed
Automatic Dialing
Parallel Output
Parallel Input
C TM VII High Speed
CTM VII High Speed (includes block parity
checking)
CTM VI High Speed
CTM VII Medium Speed
CTM VII Medium Speed
CTM vn Medium Speed
CTM IV Low Speed
CTM VI Low Speed
HS Interface Module (for 6 full-duplex lines)
HS Interface Module (for 4 full duplex lines)
ExpanSion Kit (capability for handling 2 additional lines)
HS Interface Module (for 8 full-duplex lines)
HS Interface Module (for 4 full-duplex lines)
Expansion Kit (capability for handling
4 additional lines)

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

785:221.106

UNIVAC 1108

IDENTITY OF UNIT
CLASS

Model
Number

Feature

PRICES
Monthly
Monthly
Rental Purchase Maint.

Name

Number

$

f---

Word Terminal

C( 11\11\JllNICA--

S~nchronous

(1)

$

$

(Wts):

TIllN~

:\552-01

({'nn! 'd\

F0614-00
F0614-01
F0771-00
F0772-00
F0772-01
F0772-02
F0772-03

Basic Cabinet (space for up to 2 WTS Modules)
Power Supply (for first Module)
Power Supply (for second Module)
Wts Module
Voice Band Interface
Unattended Answering
Automatic Dialing
Broad ik"lnd

270
110
110

5

8,380
3,335
3,335
14,210
225
225
1,670
225

270

8, :180

80

120
120
270
120
5
56

3,335
3,335
8,380
3,335
225
1,670

30
30
80
30

4H:;
OJ

:;
:is

80
:lO

30
160

15

-

Communication Terminal Synchronous ((:1'8)
8552-00

I

F0614 .. 00
F0614-01
F0615-00
FOGI6-00
F0617-00
FOGI8-00

Basic Cabinet (space for up to 6 CTS
Modules)
Power Supply (first 3 Modules)
Power Supply (second 3 Modules)
CTS Module
Broad Band
Unattended Answering
Automatic Dialing

-

NOTES:

* ;-\0 longer in production.
(1)

p~1V AC will e,,1;end rental agreements to a five-year term for systems in current
production at a monthly rental of 85 percent of the figure shown in this column.

(:2) fieqnll'cd on each Processor in multiprocessor system.

A minimum of 65K core storage
is required for each Processor and each I/O Controller in system; maximum core storage
per system is 262K.

-

2/69

A

AUERBACH

'"

15

A. ;;~p

AUERBAC~

790:011. 100
UNIVAC 418 SERIES
SUMMARY

REPORTS

SUMMARY: UNIVAC 418 SERIES
.1

INTRODUCTION
There are currently three systems which carry the UNIVAC 418 designation. All three
are medium-scale, solid-state systems oriented primarily toward real-time and data communications applications. The UNIVAC 418-1 and -II systems are virtually identical and
have been successfully operating in customer installation since mid 1963. UNIVAC no
longer offers the UNIVAC 418-1; however. the faster UNIVAC 418-11 is being marketed. The
UNIVAC 418-III is a new and significantly different system announced on 5 June 1968.
The following summary is oriented primarily to the UNIVAC 418-11 with some discussion
of the UNIVAC 418-III in order to better highlight the significant improvements contained
in this system. Section 792:011 provides a summary of the UNIVAC 418-III and highlights
the many Significant features of this system which make it an outstanding system in the realtime and communications fields. The UNIVAC 418-III also provides a competitive capability
for business and scientific applications.
The UNIVAC 418-1 and 418-11 evolved from the UNIVAC CUT (Control Unit Tester), a
special-purpose computer designed to test peripheral equipment for the larger UNIVAC
computer systems. A modified version of this 418 processor serves as the central processor in the Westinghouse PRODAC 510 and 580 process control systems. The UNIVAC
418-III is a greatly improved design utilizing design features and components successfully
employed in the UNIVAC 1108 system.
The UNIVAC 418-1 and -II processors are virtually identical, with the primary difference
being core storage cycle time. The cycle time for Model I is 4 microseconds, while that
of Model II is 2 microseconds. The UNIVAC 418-III is significantly different from the
other models. It includes a Command!Arithmetic unit and up to two I/O Modules. These
units are independently operating and collectively perform the functions commonly found
in a single processor unit. It has a core cycle time of 0.75 microseconds and uses the
peripheral devices of the faster UNIVAC 11 08. It is, however, compatible with the UNIVAC
418-11 at the source language level and can utilize UNIVAC 418-11 peripheral devices.
A wide range of magnetic drum units, magnetic tape units, and communications devices
permit the UNIVAC 418 systems to serve as versatile message switching and/or data collection units in real-time environments. In these applications it is essential that the system be able to store and retrieve message data rapidly in order to provide continuous and
timely service to terminals.
While the UNIVAC 418-1 and -II systems are well suited for real-time applications, their
use in scientific applications will generally not be economical due to the short word length
(18 bits) and the absence of instructions for double-precision and floating-point arithmetic
operations. The UNIVAC 418-111, on the other hand, being significantly faster and including
an optional floating point arithmetic feature will be a strong competitor in scientific applications. although the short word length can still be a slight handicap.

Figure 1. A large UNIVAC 418-II1 installation.

© 1968 AUERBACH Corporation and AUERBACH Info, Illc.

8/68

790:011 . 200

.1

SUMMARY

INTRODUCTION (Contd.)
The principal characteristics that make the UNIVAC 418 systems suitable for real-time or
communications applications are:
•

High-capacity random-access storage for master-file data or in-transit message
storage.

•

A variety of magnetic tape units which provide storage capabilities appropriate
for backup data sets such as reference storage and for record-keeping such as
ledger and Journal storage.

•

Program interrupt facilities which permit concurrent processing of several
levels of programs.

•

Flexible communications linkages to virtually any common carrier for transmission of data between the computer and remote terminals.

•

Two electronic chronometers, which make the system time-conscious.

In addition to the above, the UNIVAC 418-III system includes:

•

Up to three-way truly simultaneous I/O data transfers and processing.

•

An advanced multiprogramming Executive including job control, file control,
and file access facilities.

•

A Real-Time Communications Controller for user ease of handling of communications data.

•

.2

Multiple interrupt levels and communications interrupt tabling features for
rapid efficient handling of interrupts .
HARDWARE

. 21

Data Structure
The UNIVAC 418 systems are word-oriented, with each word consisting of 18 data bits and
a parity bit. Each 18-bit word is individually addressable and may contain one instruction,
two 8-bit alphameric characters, three 6-bit alphameric characters, or an 18-bit binary
data item .

. 22

System Configuration
The UNIVAC 418 Models I and IT are identical with respect to system configuration
possibilities and components. The basic system consists of 4,096 words of core storage,
an operator's console, a power supply, and eight I/O channels. Additionally, a magnetic
tape unit is required by the executive routine, and a UNIVAC 1004 card subsystem is
required if punched card input-output is desired. Core memory can be expanded in 4,096word increments to a maximum of 16,384 words of 4-microsecond core for the Model I and
65,536 words of 2-microsecond core for the Model II. In both systems, eight additional
I/O channels can be obtained in four-channel increments.
Any peripheral subsystem can be connected to any input-output channel, with the exception
of the Progr~mer's Console, which includes a keyboard-printer. If the Programmer's
Console is used, it must be connected to channel O. Two types of input-output transfers
are performed in the UNIVAC 418. One type uses one input-output channel to transfer
data in units of 18 bits; the second uses two input-output channels to transfer data in units
of 36 bits.
The following subsystems require two 418-I/II input-output channels and transfer dllta in
units of 36 bits:
• FH-880 Magnetic Drum Subsystem.
• Fastrand II Mass Storage Subsystems.
The following subsystems require one 418-1/11 input-output channel and transfer data in
units of 18 bits:
• FH-330 Magnetic Drum Subsystem.
• Uniservo VIC Magnetic Tape Subsystem.
• UNIVAC 1004 Central Processor.
• Programmer's Console Keyboard-Printer.
• Paper Tape Subsystem: 1 reader and 1 punch. This subsystem uses the same
channel as the Programmer's Console.
An unusual case is the Communications Terminal Module Controller (CTMC). This subsystem requires two channels, but only one channel is used to transfer data. The second
channel transmits Externally Specified Index (ESI) addresses.

8/68

A

(Contd. )

AUERBACH


SUMMARY

.22

790:0 II. 220

System Configuration (Contd.)
The UNIVAC 418-m. differs significantly from the UNIVAC 418-II in system configuration
possibilities. Memory is organized into up to four banks. each of which may consist of
16.384 or 32.768 words of O. 75-microsecond core. I/O operations are controlled by one
or two input-output modules. each of which may have 8. 12. or 16 I/o channels. The basic
UNIVAC 418-m central system consists of a Command/Arithmetic Section with operator's
console. one input-output module with 8 I/O channels. and two 16. 384-word banks of core
memory.
While the Model III :can accommodate the peripheral devices utilized with Models I and II.
its greatly increased 1:/0 data rates make the use of faster devices currently available on
the UNIVAC 1108 preferable. The rules for connecting peripheral devices to the 418-III
I/O channels follow those given above for the slower models. When two channels are required for the subsystem. the channels must be an even-numbered channel and the next
higher channel. The standard peripheral subsystems for the UNIVAC 418-III system and
their channel requirements are as follows:
•

FH-432 Magnetic Drum Subsystem; 2 channels.

•

FH-1782 Magnetic Drum Subsystem; 2 channels.

•

FH-880 Magnetic Drum Subsystem; 2 channels.

•
•

Fastrand II Magnetic Drum Subsystem; 2 channels.
Uniservo VI C Magnetic Tape Subsystem; 1 or 2 channels.

•

Uniservo VIII C Magnetic Tape Subsystem; 1 or 2 channels.

•

UNIVAC 9000 Series Computing System; 1 channel.

•
•
•

0758 High Speed Printer; 1 channel.
Punched Tape Subsystem; 1 channel.
Communications Terminal Module Controller (CTMC); 2 channels.

• Word Terminal Synchronous (WTS); 2 channels .
. 23 Internal Storage
. 231 Core memory
Working storage in the UNIVAC 418 systems is provided by ferrite-core memory. In the
Model I, 4,096 to 16.384 18-bit words are available in 4.096-word increments. The cycle
time for the Model I core memory is 4 microseconds. From 4,096 to 65,536 words of
2-microsecond core are available in 4.096-word increments in the Model II. Core storage
in the Model III is organized into up to four independently functioning banks. each of which
may contain 16. 384 or 32. 768 18 -bit words. The minimum system consists of two banks
of 16.384 words each. for a total of 32.768 words. while the maximum available is 131,072
words in four banks of 32.768 words each. The memory cycle time for the Model III core
memory is 0.75 microsecond.
In the UNIVAC 418-1 and -II, one data path to core memory is shared by the processor
and I/O transfers. By contrast, each bank of 418-m memory has three data paths, ODe
data path is used by the Command/Arithmetic section. and one is used by each of the two
possible I/O modules. While the three data paths must share the available memory cycles
if they are referencing the same bank. three peak-rate data transfers could occur
truly simultaneously and without mutual interference, if each was carried out in a
separate memory bank from the others •
. 232 Random-access storage
The random-access storage available on the UNIVAC 418 systems consists of the following
magnetic drum SUbsystems:
• FH-330 Magnetic Drum &1bsystem. Each subsystem has a capacity of from
786,432 to 3,932, 160 alphanumeric characters and an average access time of
8.5 milliseconds. The subsystem is connected via the FH-330 Drum Control
and Synchronizer Unit to one input-output channel.
•

FH-880 Magnetic Drum Subsystem. Each subsystem has a capacity of from
4. i million to 37. i million alphanumeric characters and an average access time
of 17 milliseconds. The subsystem is connected via the FH-880 Drum Control
and Synchronizer Unit to two input-output channels.

•

Fastrand II Mass Storage Subsystem. Each subsystem has a capacity of from
132 million to 1,056 million alphanumeric characters and an average access time
of 92 milliseconds. The subsystem is connected via the Fastrand II Control and
Synchronizer Unit to two input-output channels.

© 1968 AUERBACH Corporation and AUERBACH Info. Inc.

8/68

790:011. 232

UNIVAC 418 SERIES

.232 Random-access storage (Contd.)

• 24

•

FH-432 Magnetic Drum Subsystem (41S-m only). Each subsystem has a capacity
of from 1,572,S64 to 12,592,912 alphanumeric characters and an average access
time of 4.27 milliseconds. The subsystem is connected via the FH-432/17S2
Drum Control and Synchronizer Unit to two input-output channels.

•

FH-17S2 Magnetic Drum Subsystem (41S-m only). Each subsystem has a oapacity
of from 12, 5S2, 912 to 100,663,296 alphanumerio characters and an average access
time of 17.0 milliseconds. The subsystem is connected via the FH-432/17S2
Drum Control and Synchronizer Unit to two input-output channels •

Central Processors
From the standpoint of electronics and operating speed, the UNIVAC 41S-11I differs significantly from the 41S-I and II. The UNIVAC 41S-I and -II systems contain a single processing unit than contains all processor controls and functions as well as I/O controls.
The UNIVAC 41S-11I system includes two separate and independently functioning units that
combined perform the functions commonly included in the Central Processor. The Command/Arithmetic unit contains all processing facilities and controls while the independently
operating I/O Module contains all controls necessary for proper I/o data transfer .

• 241 Instruction formats
The instructions of the UNIVAC 41S systems are of three types, and these types are also
grouped in two classes - privileged and nonprivileged. The privileged instructions are
reserved for use by the Executive and basically consist of the instructions associated
with input-output and storage protection.
The Type I instructions are those commands which reference main storage and are
sensitive to the SR Active bit. The format of Type I instructions is:

I"

f

..

Ill

01

U

f = six-bit function code (02-27, 32, 33, 40-47)
u= 12-bit displacement
When a Type I instruction is executed and the SR Active bit is set to I, the high-order five,
bits required for a 17-bit main storage address are obtained from the Special Register.
If the SR Active bit is not set to I, the required five bits are obtained from the high-order
five bits of the Instruction Address Register.
The Type II instructions are those commands which reference main storage and are not
sensitive to the SR Active bit, and those which supply an immediate operand.
The format of Type II instructions is:
u

117

f

o

12111

f = six-bit function code (30, 31, 34-37, 51-76)
u= 12-bit displacement or immediate operand
When a Type II instruction is executed, the high-order five bits required for a 17-bit
main storage address are obtained from the high-order five bits of the Instruction Address
Register.
If the instruction is one which supplies an immediate operand, the u portion is handled
in a manner specified by the function code. An immediate operand is a constant
contained in the u portion of the instruction itself. The function code specifies the method
of creating an IS-bit operand from the 12-bit u portion. Either zero extension or sign
extension is used. In the case of zero extension, the high-order six bits_ of the operand
are arbitrarily set to O's. In the case of sign extension, the high-order bit of the u
portion is used to fUl the high-order six-bit portion of the operand.
Type m instructions are those that contain special parameters which must be supplied
to internal circuitry for control of certain funotions. The basic format of Type m
instructions is:
f

k

m

6

5

o

f = six-bit function code (always 50)
m = six-bit minor function code (00-77)
k = special parameters (e. g., shift count, channel number, eto.)

8/68

A

(Contd.)

AUERBACIot

"

SUMMARY

790:011. 241

.241 Instruction formats (Contd. )
Some Type III instructions require an additional one or two storage locations in order to
convey parameters. For example, the commands which are used to supply Buffer Control
Words to the I/O module use the two storage locations immediately following the Type III
instruction for these parameters. The optional Floating-Point and Binary/Decimal
instructions for the 418-III are Type III instructions.
· 242 Registers and Interrupts
The UNIVAC 418 systems utilize several special registers and interrupts to effect efficient
system operation. Six primary special registers are used as follows:
•

Three Arithmetic Registers, each of which is 18 bits long, are used in the
arithmetic section for immediate working storage. The A-upper (AU) and Alower (AL) registers are independently loadable (and storable) under program
control, but they also operate as a pair (A) for double-precision operations. The
third is the adder, which cannot be directly referenced by the program.

•

The Instruction Address Register (lAR) is a 16-bit register which is part of the
control action and is used to control program flow.

•

The Special Register (SR) is a 6-bit register used for address augmentation when
the 12-bit address of an instruction is insufficient to address the desired location.
The high-order bit of this register is used to indicate address augmentation
activity; the one (1) condition indicates active. When the SR is active, its loworde:t' 5 bits become the high-order 5 bits of a 17 -bit address whose low-order
12 bit~ are obtained from the instruction.

•

The Index Control Register is a 3-bit register which designates the index register
currently in use. (There are eight index registers in reserved core storage
locations, bu~ only one register at a time, as selected by a special instruction,
can be used for address modification.)

The occurrence of an interrupt in a UNIVAC 418-1 or -II system causes all other interrupts
to be disabled until they are re-enabled by a special instruction. Then control is transferred
to a common interrupt location where a routine (supplied by EXEC when in use) must determine the cause and effect the cure. The primary interrupts are divided into two principal
categories:
•

Internal interrupts - those interrupts which do not involve input-output, and
including illegal function code, Delta Clock underflow, and Day Clock interval
completion.

•

External interrupts - input-output-related interrupts, consisting primarily
of I/O termination and drum head-positioning completion.

The UNIVAC 418-III Command! Arithmetic section includes a powerful interrupt system.
There are 14 discrete internal interrupt levels and 96 additional levels for the I/O channels. The significance of this large number of interrupt levels is that upon encountering an
interrupt condition, control can be automatically transferred to the specific location in
main memory corresponding to that interrupt condition. This eliminates the time and space
consuming test routines required to determine the type and source of the interrupt. Additionally, the UNIVAC 418-III I/O control system can automatically accept and store interrupt information from any number of communications lines without interfering in any way
with the Executive which may be busy servicing a prior interrupt. This permits acceptanct:
of concurrent interrupts while permitting the Executive to perform its critical functions
uninterrupted.
· 243 Processing facilities
The UNIVAC 418 systems contain a full complement of instructions for fixed-point arithmetic,
logical, comparison,. and shifting operations on 18-bit binary operands. Double-precision
addition, subtraction, and shifting operations can also be performed, and the Model III can
include an optional floating-point feature. A full line of Boolean instructions operate on
18-bit quantities usually contained in the A-lower register.
The Model III processor includes a slightly expanded instruction set. Notable among the
added instructions are a block transfer. which enables movement of up to 64 words from
core to core in a block, and optional data conversion instructions for binary-to-decimal and
decimal-to-binary converSions. No character-handling or editing instructions are provided.
· 244 Console
The operator's console of the UNIVAC 418 systems consists of two primary components,
a Maintenance Panel and a Programmer's Console. The Maintenance Panel provides
direct operational control over the system and is used primarily for system startup and
shutdown. This panel can be operated off-line for use in hardware check-out and repair
while the remainder of the system continues to operate. The Programmer's Console
includes minimum facilities for direct system control; its primary purpose is to provide
effective operator-system communications while the system is in operation.
© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

8/68

790:011.250

· 25

UNIVAC 418 SERIES

Input-Output Equipment
UNIVAC offers a range of input-output equipment for the 418 System that is somewhat
Umited in scope. but well suited for the intended system applications.

· 251 Magnetic Tape Units
The Uniservo VIC Magnetic Tape Handlers are available with all 418 systems. The highspeed Uniservo VlIIC model is available only with the UNIVAC 418-m system. These
models use 2.400-foot reels of 1/2-inch coated Mylar tape and are capable of 7- or 9track data recording. and are IBM 729-compatible units. The data transfer rates available range from 8,500 characters per second for the Uniservo VIC at 200 bpi to 128, 000
characters per second for the Uniservo vmc in 9-track mode .
. 252 Printers
The 0758 Printer Subsystem is avanable only on the Model m processor. This printer is
a high-speed drum printer connected to I/o channels via the printer controller. It has
132 print positions per line and the character set is the standard 64 character set. Using
the full 64 character set. the 0758 printer prints 1200 lines; however. 1600-line-per-minute
printing speed is obtained for any 43 contiguous character set. With the alphibetics and
numerics being included in a set of 43 characters. the higher 1600 lpm can generally be
expected.
· 253 Paper tape subsystems
Paper tape subsystems capable of reading or punching 5-. 7-, or S-level tape are available
with all UNIVAC 41S models. Although these units are capable of stand-alone operation,
it is expected that when a paper tape subsystem is included in a UNIVAC 418-il configuration, it will effectively be used as part of the Programmer's Console. The Paper Tape
Subsystem available with the Model I and il processors is capable of reading 200 rows per
second and punching 110 rows per second. The Punched Tape Subsystem for the UNIVAC
41S-ill is capable of reading 1000 rows per second and punching 240 rows per second.
Each has its own control unit and is connected to one input-output channel.
· 254 Punched card subsystems
No direct on-line punched card facilities are provided with any UNIVAC 41S System.
Punched card facilities, when desired, are obtained through an on-line UNIVAC 1004 or
UNIVAC 9000 Series computing system.
· 255 Computing subsystems
The primary purpose of a computing subsystem in a 41S configuration is to provide
efficient access to slow-speed I/o devices. These systems provide card input-output
capabilities for the UNIVAC 41S which are not otherwise available. A printer is avanable
for use in the computing subsystem and it is expected that the printing facnities of the
computing subsystem can be used for on-line and off-line printing. This relieves the 41S
processor of the slow-speed data load and makes available the print edit facilities of the
computing system.
A UNIVAC 1004 Computing System can be used on-line with any UNIVAC 41S system. The
UNIVAC 1004 subsystem usually consists of a central processor with 961 characters of 6.5microsecond core memory. a 600-line-per-minute printer. a 615-card-per-minute card
reader. and a 200-card-per-minute card punch.
A UNIVAC 9000 Series Computing System can be connected on-line to a UNIVAC 41s-m
system. When used in this manner. the UNIVAC 9000 Series system will normally consist
of a 9200 or 9300 central processor with plated-wire memory. a high-speed printer. a
card reader. and a card punch with these characteristics:
Storage
Cycle
Card
Card
Time
Printer
Punch
Processor
Reader
CaEaci~
9200

S,192 bytes

1. 2 IJsec/

250lpm

400 cpm

75/200
cpm

600/1200
lpm

600 cpm

75/200
cpm

byte
9300
.26

8,192 bytes

0.6 IJsec/
byte

Data Communications Facilities
The UNIVAC 41S systems are primarily intended for real-time and communications applications. A wide range of flexible data communications equipment is available for use in
these systems. The Communications Terminal Module Controller (CTMC) is the primary
communications control unit provided for use in UNIVAC 418 systems. These controllers
interface with Communication Terminal Module (CTM) units. which in turn interface with
the appropriate adapters through communications lines to terminal devices. A wide variety
of commercially available terminal devices can be used.

8/68

A ..

AUERBACH

(Contd. )

SUMMARY

.26

790:011.260

Data. Communications Facilities (Conul.)
The Communications Terminal Module Controller (CTMC) and its associated transmission
adapters form a UNIVAC Communications Subsystem that can be used to connect a wide
variety of UNIVAC computer systems to multiple narrow-band and voice-band lines.
Different interfaces are provided for connection to a UNIVAC 418, 490 Series, 1050, or
1108 Computer.
Several types of transmission adapters, called Communication Terminal Modules (CTM) ,
are available for handling different ranges of transmission speeds: up to 300 bits per
second, up to 1,600 bits per second, and up to 50,000 bits per second. These adapters
accomodate 5-, 6-, 7-, or 8-level codes and transmit data. serially by bit. other adapters
transmit or receive data in parallel-by-bit fashion at up to 75 characters per second and
are compatible with the Bell System 402 Series Data Sets.
A CTMC has 32 input and 32 output line positions and can control up to 16 transmission
control modules. Each module can accomodate 4 simplex lines (2 input and 2 output), or
2 half-duplex or full-duplex lines.
A special adapter permits programmed automatic dialing in conjunction with a commoncarrier Automatic Calling Unit. One Dialing Adapter is required for each line utilizing
automatic dialing; four Dialing Adapters occupy one module space.
UNIVAC also offers two single-line controllers, each of which is capable of controlling communications between a UNIVAC 418 system and a remote terminal at 2000 bits per second
over the public telephone network, at 2400 bits per second over a leased voice-band line, or
at 40,800 bits per second over a leased broad-band facility. These two controllers. the
Word Terminal Synchronous (WTS) and the Communication Terminal Synchronous (CTS), are
essentially t"\'o versions of the same unit.
A special UNIVAC 418 feature is the Externally Specified Index (ESI), which allows a number of communications networks to operate concurrently on a single pair of input-output
channels by providing automatic sorting of incoming data and automatic collation of outgoing
data.
Virtually any commercially available terminal device can be used with UNIVAC 418 systems.
The preferred terminal devices are those manufactured by UNIVAC: the Uniscope 300
Visual Communications Terminal, the DCT 2000 Data Communications Terminal, the
UNIVAC 1004 and 9000 Series Computing Systems, and Teletype ASR/KSR terminals •

. 27

Simultaneous Operations
The simultaneity obtainable within UNIVAC 418 systems is a function of the system itself. The
UNIVAC 418-III permits up to three-way true simultaneity because of its independently operating memory banks and CiA and I/O Modules. In the UNIVAC 418-1 and -II, the number
of I/O trunks available on the system. the data transfer rate of each device in operation.
and the numbers of non-data.-transfer operations, determine the simultaneity obtainable.
Each I/o channel in a 418 system is capable of an input or output data transfer operation
Simultaneously with all other I/o channels and internal processing, provided that the
cumulative gross transfer rate does not exceed the gross data rate capability of core storage.
Additionally. any number of magnetic tape rewind operations may be occurring, and each
magnetic drum synchronizer may be controlling a drum positioning operation.
In UNIVAC 418-1 and -II systems, each data word transferred to or from core storage
in the 18-bit interface mode requires 4 core cycles, while each pair of words transferred
in the 36-bit interface mode requires 5 core cycles. Thus, the gross data transfer rate
obtainable between core storage and all simultaneously operating peripheral devices
cannot exceed the following values:
Model I Processor 18-bit interface: 62,500 words or 187,500 characters per second.
36-bit interface: 100.000 words or 300,000 characters per second.
Model II Processor 18-bit interface: 125,000 words or 375,000 characters per second.
36-bit interface: 200,000 words or 600,000 characters per second.
The increased capability for simultaneity provided by the UNIVAC 418-III is one of its
strongest competitive advantages. Each memory bank is capable of fully simultaneous
operation with every other memory bank, and each transfer of an 18-bit word requires
only one 0.75-microsecond memory cycle. Thus, each memory bank can support a gross
data transfer rate of 1. 33 million words or 4 million characters per second.
In the UNIVAC 418-111 system, the Command/Arithmetic section and the two possible I/O
Modules each have a separate data. channel to core memory. Each memory bank has
three independent data paths, enabling independent access by each of these data channels.
Thus, the Command/ Arithmetic section can be performing internal processing and each of
the I/o Modules can be performing I/O data transfers without interfering with each other,
so long as they are all referencing different memory banks.
© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

8/68

790:011. 300

. 27

UNIVAC 418 SERIES

Simultaneous Operations (Contd.)
Gross data transfer rates obtainable in the UNIVAC 418-III range from a minimum of 0.87
mUlion words per second with a demand on the processor of 67 per cent (if all operations
reference the same memory bank) to a maximum of 2.66 million words per second in addition to processing and with no demand on the processor (if the processor and the two II 0
Modules all reference different memory banks) •

.3

SOFTWARE

.31

UNIVAC 418-1 and 418-11 Software
The software packages available for use with the UNIVAC 418-1 and -IT systems can be
summarized as follows:
•

ART - An assembly system that translates symbolic source programs into machinelanguage object programs in relocatable or absolute form. ART is a two-pass
assembler unless a magnetic tape unit or drum is available for intermediate storage.

•

EXEC - An operating system capable of controlling four levels of programs: critical,
real-time, batch, and computational, all operating concurrently. EXEC is designed
to provide efficient utilization of the available system components and process scheduled
jobs with a minimum of operator intervention.

•

FORTRAN IV - A subset of IBM 7090/7094 FORTRAN IV that permits the use of most
of the facilities available in the 7090/7094 verSion, including integer, real, and a
form of double-precision constants and variables. Complex and logical constants and
variables, however, are not permitted.

•

Sort/Merge - A three-phase program that utilizes the polyphase method of merge
sorting with 3 to 12 Uniservo IIA or mc Tape Handlers or a Fastrand Mass Storage
Unit. A 12K 418 is required, and additional core storage can be utilized when available.

•

. 32

Utility Routines - include data transcription functions. dump and trace routines, tape
and drum maintenance, and inspect and change routines .
UNIVAC 418-m Software
The UNIVAC 418-11I software has been designed specificially for the unique characteristics
of the UNIVAC 418-11I hardware, and both the hardware and software have been designed
primarily for the unique characteristics of real-time applications. The software provided
for the UNIVAC 418-III includes:
• UNIVAC 418-111 Executive, an executive system which provides multiprogramming
capability with job control, file control and file access facilities together with I/O
Handlers;
•

Real-Time Communication Control designed to provide the user with a convenient
interface to real-time inPut/output operations;

•

Language processors which include COBOL. FORTRAN. and UNIVAC 418-111 Assembler; and

•

Systems Support Libraries which provide system support services of data file maintenance. program file maintenance. testing. and utility programs:
Sort/Merge
Program Monitor and Trace
Program Maintenance
General Tape/Drum Print
Common Procedures
Data Tape File Maintenance
Executive Independent Utilities

8/68

A

AUERBACH

'"

710:221.101

A
.--

Sf ••••••

UNIVAC 418 SERIES
PRICE DATA

BDP
'1"'11

UNIVAC 418·1/11
PRICES

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

PROCESSOR

Name

Monthly
Monthly
Rental Purchase Maint.
$
$
$

Processing Unit (Includes Core Storage)
*3010-11
*3010-12
3010-13
3010-14
F0631-00
F0604-00
F0724-00
F0633-00
F0632-00
F0632-01

ATTACHMENTS,
ADAPTERS,
AND
CHANNELS

• F0662-02
• F0632-03
F0597-01
• }<"O664-98
F0664-99
F0721-G"
2503-01

.-.

MASS
STORAGE

418-1 Processor (4,096 words of core storage;
8 I/O Channels; Control Console
418-1 Processor (4,096 words of core storage;
8 I/O Channels)
418-II Processor (4,096 words of core storage;
8 I/O channels; integrated console with
keyboard printer)
418-II Processor (4,096 words of core storage;
8 I/O channels)
Memory Expansion-4K (4,096 word module for
418-1 Processors; maximum of 16,384
words)
Memory Expansion-4K (4,096 word module for
418-II Processors; maximum of 65,536
words)
Console Alarm
Day Clock
I/O Channel Expansion (additional 4 channels
for 418-1 and II Processors)
I/O Channel Expansion (second set of 4 additional Channels for 418-1 and -II Processors)
I/O Channel Expansion
I/O Channel Expansion
1004 Control
Intercomputer Coupler
Intercomputer Coupler
Transfer Switch
Switch Cabinet

(for 418-1 Processors)
(for 418-1 Processors)
(with UNIVAC 1108)
(with another 418)

1,300

79,200

305

2,040

70,200

290

2,400

82,800

305

2,145

73,800

290

260

9,000

45

365

12,600

45

30
60
210

900
2,400
7,200

5
5
20

210

7,200

20

210
210
105
465
625
155
335

7,200
7,200
4,000
19,410
25,575
6,670
13,000

20
20
15
45
70
10

2,150
1,550
2,075

85,165
58,800
87,400

190
190
175

1,250

53,085

95

1,250

53,085
164,640
8,235
1,040
53,410
2,255
154,000

95
300
25
5
115
0
287
287
35

-

Drum Storage
7304-01
7427-02
6002~00

6002-01
F0625-00
6010-00
F0686-01
F0688-01
5009-04
F0710-00
*6010-02
*6010-04
*5009-xx

FH 880 Drum (1,572,864 18-bit words per unit)
FH 880 Control (controls 1 to 8 drums)
FH 330 Drum and Control (includes 1 drum and
control for up to a maximum of 5 drums;
262,144 words per drum)
FH 330 Drum and Cabinet (includes one drum
and space for second)
FH 330 Drum
Fastrand II (44 milli0n words)
Fastbands
Write Lockout
Fastband Control (controls 1 to 8 units)
Search All Words
Drum (29 million, 18-bit words)
Drum (14 million, 18-bit words)
Drum Control

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

4,050
215
30
1,335
55
3,637
2,192
1,035

-

36,000

2/69

UNIVAC 418 SERIES

790:221.102

IDENTITY OF UNIT
CLASS

Model
Number

Name

Feature

INPUTOUTPUT(2)

PRICES
Monthly Purchase Monthly
Maint.
Rental

Magnetic Tape
0858-00
0858-08
0858-01
5008-08
5008-91
*5008-09
*0952-00
*1345-01
*0952-01
F0627-04
F0627-99
*F0627-03

Uniservo VIC Master (7 track; handles up to
3 slaves)
Uniservo VIC Master (7-track; handles up to 3
slaves; simultaneous read-read!read-wrlte)
Uniservo VIC Slave (7-track)
Uniservo VIC Synchronizer (controls up to
4 0858-00 VIC Master units)
Uniservo VIC Synchronizer (controls up to
4 0858-08 VIC. Master units)
Uniservo VIC Auxiliary Synchronizer
Magnetic Tape Control
Power Supply
Magnetic Tape Control
Translate Option (for 5008-08)
Translate Option (for 5008-91)
Translate Option (for 5008-09)

515

17.350

115

570

19,800

115

310
745

10,470
31,0'10

70
35

1,490

62,140

70

74.5
1,025
235
1,130
110
220
110

31,000
35,460
8,600
39,OGO
4,410
8,H20
4,410

35
100
40
105
5
10
5

Paper Tape

COMl\lUNICATIONS

F0603-00

Paper Tape Subsystem

160

5,980

30

F0900-05

Communication Terminal Module
Controller (CTMC)
CTM-LS
CTM-MS
Communications Line Terminal
Communications Line Terminal
CTM-HS
Automatic Dialing
Parallel Output
Parallel Input
Spare CTMC
Word Terminal Synchronous (WTS):
Basic Cabinet (space for up to 2 WTS
Modules)
Power Supply (first Module
Power Supply (second Module)
WTS Module
Voice Band Interface
Unattended Answering
Automatic Dialing
Broad Band
Communication Terminal Synchronous (CTS):
Basic Cabinet (space for up to 6 CTS
Modules)
Power Supply (first 3 Modules)
Power Supply (second 3 Modules)
CTS Module
Broad Band
Unattended Answering
Automatic Dialing

705

24,700

135

65
85
50
50
100
20
40
40
260

2,255
2,895
1,360
1,360
3, G30
635
1,315
1,315
9,020

15
15
15
15
15
5
10
10
50

270

8,380

80

110
110
485
5
5
55
5

3.335
3,335
14,210
225
225
1,670
225

30
30
160

270

8,380

80

110
110
270
110
5
56

3,335
3,335
8,380
3,335
225
1, 670

30
30
80
30

F0901-04
F0902-02
*F0903-00
*F0903-01
F0903-02
F0905-00
F0904-00
F0904-01
F0906-05
8552-01
F0614-00
F0614-01
F0771-00
F0772-00
F0772-01
F0772-02
F0772-03
8552-00
F0614-00
F0614-01
F0615-00
F0616-00
F0617-00
F0618-00
~OTES:

*1\'0 longer in production.

(1)

(2)

2/69

l:!I.'1V AC will extend rental agreements to a five-year term for systems in current
productlOn at a monthly rental of 85 percent of the figure shown in this column.
Additional input-output units are normally implemented by attachment of a UNIVAC
1004.

fA

AUERBACH
~

---

15

--

--

15

I
I

790:221.103

RICE DATA

UNIVAC 418·111
UNIVAC does not include monthly maintenance charges in their published prices; these charges have been
included in the Monthly Rental column in this Digest to permit convenient comparison with other systems.
PRICES

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

PROCESSOR

Name

Monthly
Monthly
Rental IPurchase Maint.
$(1)
$
$

Processing Unit
3020-00
FlO83-00
F1082-00
F0676-09
4010-00
F1086-00
F0774-00
F0774-01

418-m Processor (includes 8 I/o channels)
Floating Point
Decimal Binary Conversion
Use Time Meter'
Operator's Console (required for 3020-00)
Day Clock
Auxiliary Console (accommodates two to four
CTMC's left- or right-hand addition to
Console)
Auxiliary Console (same as F0774-00 but
located between Consoles of two Processors)

280
135
165

92,875
5,220
2,395
525
11,310
5,655
6,600

275
5
5
5
20
5
10

165

6,600

10

Storage (-16K words; includes power for up to
65K words and cabinetry for up to 32K words)
Storage Expansion (16K words; permits
operation of memory in two 16K bands)
Storage Expansion (16K words; includes
cabinetry for 32K words; permits operation
of memory in one 32K and one 16K band)
Storage Expansion (16K words; permits
operation of memory in two 32K bands)

1,950

76,995

180

1,290

50,895

120

1,400

54,595

145

1,290

50,895

120

Second I/o Module (includes additional
8 I/o channels)
I/o Channel Expansion (includes 4 additional
channels)
I/o Channel Expansion (includes 4 additional
channels; requires F1084-01)
Shared Peripheral Interface (SPI):
SPI (permits 2 paired channels from one
processor or 1 paired channel from each
of two Processors to share the units of
a peripheral subsystem; includes, cabinet
with space for second SAl)
SPI (additional SPI; fits into cabinet included
with 0955-04)
SPI Expansion (ildds third Processor paired
channel capability to 0955-04 or -05)
SPI Expansion (adds fourth Processor
paired channel capability to 0955-04 or
-05 with F0789-00)

800

31,540

75

120

4,785

10

120

4,785

10

475

19,575

25

420

17,400

20

85

3,480

5

60

2,395

5

2,410
125
60

-

Main Storage
7009-00
F1087-00
7009-02
Fl087-01
FlO84-00

ATTACHMENTS,
ADAPTERS,
AND
CHANNELS

Fl084-01
F1084-02
0955-04

0955-05
F0789-00
F0789-01

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

UNIVAC 418 SERIES

790:221.104

IDENTITY OF UNIT
CLASS

ATTACHMENTS,
ADAPTERS,
AND
CIIANNFI$
(Contd. )

Model
Number

Feature
Number

5309-00

5309-02
FI095-00
F0893-00

F0893-02
2508-00
2508·01
2508-02
2508-03

l\L-\SS
STORAGE

Name

Sys tern Interconnections:
IntercoIDputer Coupler (provides 36-bit
interface for direct communication with a
second 41S-m Processor via 2 I/o
channels)
Intercomputer Coupler (permits communication with a second 41S-m Processor via
1 I/o channel)
41S/9000 1 CCU (provides IS-bit interface for
on-site communication with UNIVAC 9000
series computer){2)
Transfer Switch (permits manual switching of
peripheral subsystems and processors
between pairs of I/o channels; (36-bit
interface; switches can be cascaded if
desired).
Transfer Switch (Same as F0893-00, except
with 18-bit interf2ce for switching 1 I/o
channel)
Switch Cabinet (contains power supply, integral
operators panel, and control unit for up to
5 transfer switches in any combination)
Switch Cabinet (same as 2508-00 but with
remote operator panel)
Switch Cabinet (same as 2508-00 but includes
auxiliar~ower supply and can
accomm ate only 4 switches)
Switch Cabinet (same as 2508-01 but includes
auxiliary power supply and call.
accommodate only 4 switches)

PRICES
Monthly
Monthly
Rental IPurchase Maint.
$(1)
$
$

625

25,575

70

625

25,575

70

205

S,050

20

168

6,540

18

168

6,540

18

405

12,740

60

405

12,740

60

598

20,000

88

598

20,000

88

2,150
1,550
1,070
2,940
70
70
2,145

85,165
58,800
42,435
117,210
2,255
2,255
82,515

190
190
100
260
15
15
260

30
440
4,050
215
30
1,280

1,040
17,805
164,640
8,235
1,040
51,060

5
25
300
25
5
115

1,335

53,410

115

3,000

120,495

230

3,110

125,195

230

55
55

2,255
2,350

--

50

1,830

10

Drum Storage
7304-01
7427-03
6016.0Q
6015-00
F0786-01
F0767-00
5012-00

6010-00
5009-00

F0929-00
F0930-00
F0686-01
F0688-01

5009-04
5009-93
5009-92
F0710-00
F0763-02
F0959-1l9

FH880 Drum (1,572,864 18-bit words)
FH880 Control (controls 1 to 8 FH880 Drums)
FH ''32 Drum (524,288 18-bit words)
FH 17S2 Drum (4, 194,304 18-bit words)
Dual Channel (for FH 432)
Dual Channel (for FH 1782)
FH 432/FH 1782 Control (controls 1 to 8
FH 432 or FH 1782 Drums in any combination
Write Lockout
SPI (allows two Erocessors to access control
Fastrand II (44 Mil ion - 18-bit wordS)
Fastbands
Write Lockout
Fastrand n Control (unbuffered control for 1 to
8 Fastrand n Drums)
Fastrand II Control (buffered control for 1 to 8
Fastrand II Drums)
Fastrand II Dual Control (dual unbuffered
control for 1 to 8 drums; includes dual
access for first drum)
Fastrand n Dual Control (dual buffered control
for 1 to 8 drums; includes dual access for
for first drum)
Search all Words (for buffered controls only)
Control Buffer (for field conversion from
unbuffered to buffered control)
Dual Access Prum A.dapter (required for
attachment of all subsequent drums to a
Dual COntrol, except fifth)

I

1./69

A

AUERBACH
~

'780:221.105

PRICE DATA

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

MASS
STORAGE
(Contd. )

Name

PRICES
Monthly
Monthly
Rental Purchase Maint.
$
$
$

Drum Storage
F0959-98
6010-10
5009-89
5009-85
5009-81
F0763-01
F0686-01
F0688-01
F0959-97
F0959-96
FOnO-OI

Dual Access Drum Adapter (required for
attachment of fifth drum to a Dual Control)
Fastrand m Drum
Fastrand m Control
Fastrand m Dual Control (includes dual access
adapter for first drum only)
Fastrand m Control (field upgrade from
single to dual access)
Control Buffer
Fastband
Write Lockout
Fastrand m Dual Access Drum Adapter
(for all subsequent drums except fifth)
Fastrand m Dual Access Drum Adapter
(for fifth drum)
Search All Words

50

1,830

10

4,965
1,565
3,650

200,800
62,220
147,010

350
135
270

2,085

84,790

135

55
205
30
50

2,350
8,235
1,040
1,830

-

50

1,830

10

25

2,255

-

745

31,070

35

1,490

62,140

70

745

31,070

35

1,490

62,140

70

110
220

4,410
8,820

5
10

50
515

1,960
17,350

5
115

570

19,800

115

310
515

10,470
17,350

70
115

570

19,800

115

310

-

10,470

-

-

55

2,450

55

2,450

-

--

INPUTOUTPUT

25
5
10

Magnetic Tape
5008-00
5008-98
5008-08
5008-91
F0627-04
F0627-99
F0706-00
0858-00
0858-08
0858-01
0858-10
0858-12
0858-14
FI021-00
Fl072-00
F1072-01

Uniservo VIC:
Uniservo VIC Control (controls up to 4 Master
units; requires 2 I/o channels)
Uniservo VIC Control (controls up to 4 Master
units; simultaneous read-read7read-write;
requires 4 I/o channnels)
Uniservo VIC Control (same as 5008-00 but
requires only 1 I/o Channel)
Uniservo VIC Control (same as 5008-98 but
requires only 2 I/o Channels)
Translator (for 5008-00 or 08)
Translator (for 5008-98 or -91)
9 Track Capability (required in 9- Track
Controls, for non simultaneous and 2
for simultaneous controls)
Uniservo VIC Master (7-Track; 8,540/23, 741/
34,160 char/sec; can handle up to three
0858-01 Slave units)
Uniservo VIC Master (Same as 0858-00 but
with provisions for simultaneous read-read/
read-write)
Uniservo VIC Slave (7-Track)
Uniservo VIC Master (9-Traok; 800 bpi;
34,100 bytes/sec; handles up to three
0858-14 slave units)
Uniservo VIC Master (Same as 0858-10 but with
provisions for simultaneous read-read/
read-write)
Uniservo VIC Slave (9-Track)
7 to 9 Track Conversion (converts 7-Track
tape unit to 9-Track unit)
Simultaneous Capability (converts 0858-00
to 0858-08)
Simultaneous Capability (converts 0858-10
to 0858-12)

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

70

2/69

790:221.106

UNIVAC 418 SERIES

PRICES

IDENTITY OF UNIT

CLASS

Model
Number

Feature
Number

Name

Monthly
Monthly
Rental IPurchase Maint.
$(1)
$
$

Magnetic Tape (Cont'd)

INPUTOlTTPlTT
(C'ont'd)
5008-12
5008-20
5008-89

5008-87
5008-83
5008-79
F0704-00
F0706-00

F0627-04
F0627-98
0859-00
0859-04
0859-08
0859-10
F0999-00
F0999-04

Uniservo VIDC:
Uniservo VTIIC Control (controls 1 to 16
Uniservo VIDC units; requires 2 I/O
channels)
Uniservo VIDC Control (same as 5008-12
but requires 1 I/o channel)
Uniservo VIDC Control (controls 1 to 16 Uni
servo VillC units; provides read-read.
read-write, write-read. and write-write
simultaneously; requires 4 I/o channels)
Uniservo VillC ContrOlJsame as 5008-89
but requires only 2 0 channels)
Control Expansion (converts 5008-12 to
5008-89)
Control Expansion (converts 5008-20 to
5008-87) 7-track
VIC Capability (allows VIC and VIDC units
to be intermixed on 5608-12 or
5008-20 controls)
9 Track Capability (1 required for nonsimultaneous, 9-track controls; 2 required for simultaneous, 9-track controls)
Translator (for 5008-12 or -20)
Translator (for 5008-89 or -87)
Uniservo VIDe (7-track: 24,000/66,720/
96,000 char/sec; non-simultaneous)
Uniservo VIDC (9-track; 800 bpi; 96,000
bytes/ sec: non-simultaneous)
Uniservo VIDC (7-track: same as 0859-00,
but for simultaneous control)
Uniservo VIDC (9-track: same as 0859-04,
but for simultaneous controls)
7 to 9 Track Conversion (converts
OtJ59-00 to 0859-04)
7 to 9 Track Conversion (converts
0859-08 to 0859-10)

1550

62,430

120

1550

62,430

120

3100

124,860

240

3100

124,860

240

1550

62,430

120

1550

62,430

120

85

3,385

5

50

1,960

5

110
220
860

4.,410
8,820
32,735

5
10
110

875

33,390

110

890

34,045

110

905

34,700

110

15

655

-

15

655

-

500
400

18,490
14,790

75
60

1265

43,935

310

Reader and Control (1000 char/sec; requires
2 I/o channels)
Reader and Control (1000 char/sec; requires
1 I/o channel)
Punch (240 char/sec; requires Reader)
Punch with Verifier (240 char/sec; requires
Reader; includes post-punch photo-electric
read check)
Spoolers (for 0917-00 or -02 Readers)

600

19,575

150

600

19,575

150

250
300

8,700
9,350

50
85

150

5,000

35

Communication Terminal Module Controller
(CTMq

705

24,700

135

Printer
5011-12
F0751-99
0758-99

Printer Control (controls 0708-99 one printer)
Printer Control Expansion (for control of
second printer)
Printer (1600 lines/min. with 43-character
set; 1,200-lines/min. with !l3-character
set)
Paper Tape

0917-00
0917-02
0918-00
0918-01
F0962-00
CCnl:'.rumCATIm.rS

2/69

F0900-06

_.

A

AUERBACH

'"

PRICE DATA

790;221.107

"

PRICES

IDENTITY OF UNIT

CLASS

Model
Number

Feature
Number
F0906-06
F0901-04
F0902-02
F0903-02
F0905-00
F0904-00
F0904-01
F0988-00
F0988-01

COMMUNICATIONS

F0991-00
F0989-00
F0989-01
F0989-02
F1027-00
F1048-00
F1018-02
F1018-03
F1018-05
F1019-01
F1019-03
F1019-05
8552-01
F0614-00
F0614-01
F0771-00
F0772-00
F0772-01
F0772-02
F0772-03

Name

Monthly
Monthly
Rental IPurchasE Maint.

$(1)
Spare CTM Controller
CTM-LS
CTM-MS
CTM-HS
Automatic Dialing
Parallel Output
Parallel Input
CTM vn High Speed
CTM vn High Speed (includes black parity
checking)
CTM VI High Speed
CTM vn Medium Speed
CTM vn Medium Speed
CTM vn Medium Speed
CTM IV Low Speed
CTM VI Low Speed
HS Interface Module (for 6 full-duplex lines)
HS Interface Module (for 4 full duplex lines)
Expansion Kit (capability for handling 2
additional lines)
HS Interface Module (for 8 full-duplex lines)
HS Interface Module (for 4 full duplex lines)
Expansion Kit (capability for handling
4 additional lines)
Word Terminal Synchronous (WTS):
Basic Cabinet (space for up to 2 WTS Modules)
Power Supply (for first Module)
Power Supply (for second Module)
WTS Module
Voice Band Interface
Unattended Answering
Automatic Dialing
Broad Band

$

$

260
65
85
100
20
40
40
80
95

9,020
2,255
2,895
3,630
635
1,315
1,315
3,050
3,480

50
15
15
15
5
10
10
10
15

125
60
70
80
80
95
60
40
20

4,570
2,175
2,610
3,050
3,050
3,480
2,175
650

20
10
10
10
10
15
10
5
5

80
40
40

3,050
1,525
1,525

10
5
5

270

8,380

80

110
110
485
5
5
55
5

3,335
3,335
14,210
225
225
1,670
225

30
30
160

1~525

-15
-

NOTES:
(1) UNIVAC will extend rental agreements to a five-year term for systems in
current production nt n monthly rental of 85 per cent of the figure shown in
this column.
(2) UNIVAC 900 serieH computers Systems can be connected to the 418-ill
via the Intercomputer Control Unit (F1095-00) to provide additional
peripheral units •
.-

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

-.£.

791:011. 100
SI ......

~EDP

AUERIACH

UNIVAC 41&-1/11
SUMMARY

UPiITS

•

SUMMARY: UNIVAC 418-1/11
.1

INTRODUCTION

The UNIVAC 418 Models I and II were the first two models in the UNIVAC 418 Series.
A fully detailed Summary of this series is presented in section 790:011; this section briefly
summarizes those factors which relate specifically to the UNIVAC 418-1 and II systems. The
UNIVAC 418-11 is currently being marketed by UNIVAC; however, the UNIVAC 418-1 is no longer
available.
The UNIVAC 418-I/n is a medium-scale, solid-state computer that is oriented primarily toward real-time and message switching applications. Two central processors and
associated core memories are available for use in 418 systems: the Model I and the Model n.
Both models have the same set of instructions and, given the same peripheral equipment and
input-output channel assignments, are completely program-compatible. The core storage cycle
time Is 4 microseconds for Model I and 2 microseconds for Model n.
A wide range of magnetic drum units, magnetic tape units, and communication devices
permit UNIVAC 418-I/II systems to serve as versatile message switching centers and for realtime commercial applications. Use of the 418 in scientific applications will generally not be
econ: mical due to its short word length (18 bits) and the absence of instructions for doublepreciSion or floating-point arithmetic. UNIVAC 418-11 system rentals range from approximately
$7,000 to $18,000 'per month and average around $11,000.
Although UNIVAC 418-1 systems have been operating In customer installations since
June 1963, their general availability to commercia). users was not announced until August 1964.
This 418 is closely related to the UNIVAC 1218 Military Computer. Both the 418 and the 1218
evolved from the UNIVAC CUT (Control Unit Tester), a speciai-purpose computer designed to
test peripheral equipment for the larger UNIVAC computer systems. A modified version of the
418 serves as the central processor in the Westinghouse PRODAC 510 and 580 process control
systems.
The principal characteristics that make the UNIVAC 418-I/II systems suitable for realtime or message switching applications are:

.2

•

High-capacity random access storage for master file data or messages in transit.

•

A variety of magnetic tape units, which provide storage of slower access and
greater capacity for use as a reference store, a journal, and intercept and
overflow storage.

•

Program interrupt facilities which permit concurrent proceSSing of several
levels of programs.

•

Flexible communications linkages to any common carrier for two-way
transmission of data between the computer and remote pOints.

•

Two electronic chronometers, which make the systems "time-conscious. "

HARDWARE

UNIVAC 418 systems using the Model I Processor can have from 4,096 (included with
the basic processor) to 16,384 word locations of 4-microsecond core storage in 4, 096-word
modules. Systems using the Model II Processor can have from 4,096 (included with the basic
processor) to 65,536 word locations of 2-microsecond core storage in 4, 096-word modules.
Each core storage word contains 18 data bits and one parity bit. Each 18-bit word can hold one
instruction, three alphameric characters, or an 18-bit binary data item. Thirty-two locations
of core storage contain a permanently-wired "bootstrap" routine to facilitate program loading
and error recovery.
The Central Processor can perform a full complement of fixed-point arithmetic,
Boolean, comparison, and shifting operations on IS-bit binary operands. Double precision addition, subtraction, and shifting are also provided. The UNIVAC 418-I/II has no automatic facilities
for double-precision multiplication and diviSion, floating-point arithmetic, decimal arithmetic,
editing operations, multi-word internal transfers, or radix conversion. There are eight index
registers, with instructions to step and test, increment, store, and load them. Only one of the
eight index registers at a time, as selected by a separate instruction, can be used for address
modification. Indirect addressing is not provided except for ''.Jump Indirect" instructions.
Instructions which reference core storage contain a single 12-bit address. This restricts the number of locations directly accessible at any time to one module of 4, 096 words
unless Indexing is used. The particular module being addressed is indicated by a special
register, by the Program Address Counter, or by an index register.
ICJ 1968 AUERBACH Corporation and AUERBACH Info, Inc.

8/68

791:011.200

.2

UNIVAC 418-1/11

HARDWARE (Contd.)

Average execution time for UNIVAC 418 instructions is about 4 microseconds for the
Model II Central Processor and 8 microseconds for the Model I Processor. The longest instruction - Divide - requires 24 microseconds (48 microseconds for Model I), while a few instructions require as little as 2 microseconds (4 microseconds for Model I).
The program interrupt facility causes a transfer of control to one of 51 fixed core
locations upon completion of an input-output operation, upon detec:tion of an illegal operation code
or input-output error, upon underflow of the Delta Clock, or upon an interrupt from the Day Clock.
The,UNIVAC 418-I/II Processors can have 8, 12, or 16 input-output channels. If the
Console keyboard-printer is used, it must be assigned to channel 0; the remaining channels are
for general input-output use. Many of the standard peripheral subsystems require two channels,
as indicated below. In general, each subsystem can handle one data transfer operation at a time.
The maximum gross data transfer rate (or "saturation rate") for all simultaneously-operating
peripheral devices ranges from 135,000 to 200,000 words per second for the Model II Central
Processor, depending upon the channel requirements of the various SUbsystems. The gross data
rate for tl;le Model I Processor ranges from 62,500 to 100,000 words per second.
Two types of peripheral interfaces are used in the 418 system, depending upon the
channel requirements of the subsystem used. Subsystems which use one input-output channel
require 4 core memory cycles to transfer 18 bits (one word) of data. Subsystems using two
input-output channels require 5 memory cycles to transfer 36 bits (two words) of data.
Each UNIVAC 418-1/11 peripheral subsystem can consist of any of the following groups
of input-output devices and their associated control units:

•

1 to 5 Flying Head 330 Magnetic Drums. Each drum stores 262,144 words
with an average access time of 8.5 milliseconds. Peak data transfer rate
is 30,000 words per second with an interlace factor of 2. Interlace factors
of 4, 8, or 16 can alternatively be used to achieve lower effective transfelrates. One input-output channel is required.

•

1 to 8 Flying Head 880 Magnetic Drums. Each drum stores 1,572,864
words with an average access time of 17 milliseconds. Peak data transfer
rate is 60,000 words per second. Interlace factors of 2, 4. 8, or 16 can
be used for lower effective transfer rates. Two input-output channels are
required.

•

1 to 8 Fastrand II Mass Storage Units. Each unit has 2 drums, served by
64 movable read/write heads, and stores 44.040,192 words with an average
access time of 93 milliseconds. A single Fastrand I Subsystem can store
up to 528 million characters. Peak data transfer rate in UNIVAC 418 systems is 52,685 words per second with an interlace factor of 3. Interlace
factors of 5, 7, and 9 can be used for lower transfer rates. Two inputoutput channels are required.

•

2 to 16 Uniservo VIC Magnetic Tape Handlers. These units can read
forward only, at a peak transfer rate of 8,500, 24,000, or 34,00 rows
per second. The tape format is fully compatible with the IBM 729 and
7330 Magnetic Tape Units. One input-output channel is required; two
channels must be used if simultaneous read/read or read/write operations are desired.

•

1 Paper Tape Reader and 1 Paper Tape Punch. These units (housed in
a single cabinet on the console desk) can read punched tape of up to 8
levels at up to 200 characters per second and punch standard 5-, 6-, 7-,
or 8-level punched tape at 110 characters per second. One input-output
channel (the same channel as used by the console keyboard-printer) is
required.

•

1 UNIVAC 1004 Central Processor and associated peripherals. The
1004 provides the UNIVAC 418 with punched card input at 400 or 600
lines per minuie. An optional card punch is available which punches
cards at the rate of 200 pe r minute. The 1004 can also be equipped
with magnetic tape and paper tape units. See Computer System Report 770 for full details. The 1004 subsystem requires one 418 inputoutput channel.
Communications Terminal Module Controller, consisting of up to 32 input and
32 output Communication Line Terminals connected to a Communication
Multiplexer. These units enable the 418 to send and receive data via
any common carrier at transmission rates of up to 4, 800 bits per second.
Two input-output channels are required.

•

8/68

A.

AUERBACH

"

(Contd. )

SUMMARY

.3

791:011.300

SOFTWARE
Programs which are available for use with the UNIVAC 418-1/11 systems can be summarized as follows:
•

ART - An assembly system that translates symbolic source programs
into machine language object programs in relocatable or absolute form.
ART is a two-pass assembler unless a magnetic tape unit or drum is
available for intermediate storage. An 8K 418 is required, with a program input and program output device, a card reader or magnetic tape
unit for control input, and a 1004 printer. Assembly speed is approximately 400 lines per minute for a 12K 418-1 Processor with Uniservo
mc Tape Handlers.

•

EXEC - An operating system capable of controlling four levels of programs: critical, real-time, batch, and computational, all operating concurrently. EXEC is designed to provide efficient utilization of the available system components and process scheduled jobs with a minimum of
operator inte'rvention. EXEC requires at least 1.491 words of core
storage plus about 208 words per I/O handler routine. Times required
to perform the EXEC functions associated with each input-output operation range from O. 52 to 1. 19 milliseconds.
'

•

FORTRAN IV - A subset of IBM 7090/7094 FORTRAN IV that permits
the use of most of the facilities available in the 7090/7094 version,
including integer, real, and a form of double-precision constants and
vaJ,"iables. Complex and logical constants and variables, however, are
not permitted.

•

Sort/Merge - A three-phase program that utilizes the polyphase method
of merge sorting with 3 to 12 Uniservo IIA or mc Tape Handlers or a
Fastrand Mass Storage Unit. A 12K 418 is required, and additional core
storage can be utilized when available.

•

Utility Routines - include data transcription functions, dump and trace
routines, tape and drum maintenance, and inspect and change routines.

C 1968 AUERBACH Corporation and AUERBACH Info, Inc.

8/68

-.l.

792:011. 100
SlI .....

UNIVAC 418-111
SUMMARY

AEDP

-

AUERBAC~

REPORTS

•

SUMMARY: UNIVAC 418-111
•1

INTRODUCTION
The UNIVAC 41S-I11 Real-Time System was announced on 5 June 1965. In basic
processing characteristics, the UNIVAC 41S-1I1 is similar to the UNIVAC-II, but greatly
increased speed and I/O capability has been obtained by a system design that utilizes concepts and technology successfully employed in the UNIVAC 110S.
The UNIVAC 41S-I11 is a medium-scale computer system designed primarily for data communications and real-time applications but with sufficient power to permit concurrent scientific and business data processing. This system provides: up to 131,072 eighteen-bit
words of O. 75 microsecond core storage; truly simultaneous processing and input-output
data transfers; capability to handle up to 2.66 million word per second input-output data;
storage protection; Externally Specified Indexing (discrete communication line buffering);
and automatic tabling of communications interrupts.
The peripheral devices available with the UNIVAC 41S-III system include a wide range of
magnetic drum systems and magnetic tape units, as well as a flexible communications
capability that permits use of virtually any commercially available terminal device. With
respect to peripheral equipment, the UNIVAC 41S-ITI is compatible with all peripheral
equipment used in the UNIVAC 110S system, as well as that used in the UNIVAC 41S-11
system.
Key features and components of the UNIVAC 41S-I11 system include:
•

Independence of central components - central processor, I/O channels, and main
memory modules, leading to optimized performance through minimal cdmponent
interference.

•

Input/output and interrupt control mechanisms implemented in automatically functioning hardware.

•

Core memory access speed of 750 nanoseconds per IS-bit word.

•

Up to four independently accessible main memory banks of 16K or 32K IS-bit words
of storage per bank.

•

Up to 32 I/O channels controlled by two independently functioning I/O Modules
(Up to 16 channels per module).

•

A hardware/software memory protection scheme.

•

Extensive data communications capabilities, permitting control of more than 400
active duplex communication lines per UNIVAC 41S-IIT system.

•

High transfer rate, low access time magnetic drum subsystem.

•

A real-time operating system designed to exploit the functional independence of the
hardware components and to permit concurrent execution of multiple real-time
jobs in conjunction with scientific and commercial background processing.

•

A sophisticated Real-Time Operating System that consists of a set of control programs, programming aids, and utility services.

Figure 1. A Large UNIVAC 41S-III Installation
© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

8/68

792:011. 200

.2

UNIVAC 418-111

HARDWARE
The basic unit of data in the UNIVAC 41S-m is an IS-bit word which can contain one instruction, two S-bit alphameric characters, three 6-bit alphameric characters, or a fixed point,
floating point or absolute binary quantity.
Core memory is organized into four banks each of which can contain 16K or 32K words of
storage and each of which is controlled separately from the others. The vertical/
horizontal expansion capabilities of the UNIVAC 41S-m core memory allow the user to expand his memory capacity either by adding banks to provide more simultaneity or expanding
within a bank to provide more storage capacity or both.
Figure 2 illustrates the maximum storage unit and referencing unit configuration. The minimum configuration is the Command/Arithmetic (C/A) Unit, 10M#O and memory units mO of
Bank 0 and m1 of Bank 1. This provides two-way simultaneity with 16K core storage in each
bank. Three-way Simultaneity capability is achieved by adding memory unit m4 of Bank 2
and 10M#1. The addition of memory unit m5 of Bank 3 then provides greater flexibility of
storage allocation to ensure as much as possible the achievement of three-way simultaneity.

II

10M
#1

Command/Arithmetic
Unit

I

I

I

II

I I I

I I I

I

Bank 0

Bank 1

Bank 2

Bank 3

(~~

16K
(m1)
16K
(m3)

16K
(m4)
16K
(m6)

16K
(m5)
16K
(m7)

16K
(m2)

I

10M
#2

Figure 2. Storage Unit Assignment in a Maximum Storage Configuration
Increased core storage capability can also be obtained by the addition of memory unit m2 of
Bank 0 instead of or in addition to memory units m4 of bank 2 or m5 of Bank 3. Indeed,
each existing bank can be expanded to 32K by adding the second 16K memory unit without
first having subsequent banks added.
The main storage of the UNIVAC 41S-1I1 System is constructed so that the banks of storage
have a physical unit/logical unit relationship. This is accomplished by a series 'Jf address
switches on the maintenance panel of each storage cabinet. These switches are used to assign
the bank number to each of the memory modules.
Another Significant feature of the UNIVAC 418-m is the storage protection capability offered.
A Guard Limits Register is employed that enables storage protection of core memory in 256
word increments. The Executive determines required limits, sets the Guard Limits Register and activates Guard Mode before giving control to a particular program. When the Executive regains control, the Guard Mode is de-activated. Thus, in a multiprogramming
environment, the accidental alteration of one program area by another program, even one
being tested, is prevented.
The Command/Arithmetic Unit and I/O Module of the UNIVAC·41S-m System are independently operating units which collectively perform the functions normally found in one central
processor unit. The Command/Arithmetic Unit contains all necessary controls for program execution while the I/O Module contains the controls necessary for proper input-output
data transfers. This separation of controls into independently operating units enables the
establishment of independent data paths between the control units and the core storage banks
which allows the truly simultaneous operation of the UNIVAC 418-m not found in systems
with a single central processing unit.
The minimum UNIVAC 41S-I11 system consists of the Command/Arithmetic Unit, one I/O
Module with S I/O channels, and two banks of 16K words of core memory. A second I/O
Module with S. I/O channels can be added and increased simultaneity capability from twoway to three-way can then be obtained by adding an additional memory bank. An additional
four or eight I/O channels can be added to either or both of the I/O Modules providing a
maximum of 32 I/O channels in the fully expanded system. Memory can be increased from
the initial 32K words up to 131K words as previously indicated.
The UNIVAC 418-111 Command section includes a powerful interrupt system. There are 14
discrete internal interrupt levels and 96 additional levels for the I/O channels. The significance of this large number of interrupt levels is that upon encountering an interrupt condition, control can be automatically transferred to the speCific location in main memory

8/68

A

(Contd.)

AUERBACH

co

SUMMARY

.2

792:011.201

HARDWARE (Contel.)

corresponding to that interrupt condition. This eliminates the time- and space-consuming
test routines required to determine the type and source of the interrupt. Additionally, the
UNIVAC 418-m I/o control system can automatically accept and store interrupt information
from any number of commmlications lines without interfering in any way with the Executive
which may be busy servicing a prior interrupt. This permits acceptance of concurrent
interrupts while permitting the Executive to perform its critical flmctions Iminterrupted.
The following is a list of the perlpheral subsystems used in a UNIVAC 418-m system. A
peripheral subsystem can consist of any of the followlng groups of input-output devices and
their associated control Imits.
•

1 to 8 Flying Head 432 Magnetic Drums. Each drum stores 524,288 words wlth
an average access time of 4.25 milliseconds. Peak data transfer rate is
474,000 words per second. Interlace factors of 2, 4, 8, or 16 can be used to
obtain lower effective transfer rates. Two Input-output channels are required.

•

1 to 8 Flying Head 1782 Magnetic Drums. Each drum stores 4, 194,304 words
with an average access time of 17. 0 milliseconds. Peak data transfer rate is
454,000 words per second. Interlace factors of 4, 8, or 16 can alternatively be
used to achieve lower effective transfer rates. Two Input-output channels are
required.

•

1 to 8 Flying Head 880 Magnetic Drums. Each drum stores 1, 572, 864 words
with an average access time of 17 milliseconds. Peak data transfer rate is
60, 000 words per second. Interlace factors of 2, 4, 8, or 16 can be used for
lower effective transfer rates. Two input-output channels are required.

•

1 to 8 Fastrand n Mass Storage Units. Each unit has 2 drums, served by 64
movable read/write heads, and stores 44, 040, 192 words with an average access
time of 93 milliseconds. A single Fastrand n Subsystem can store up to 1,056
million characters. Peak data transfer rate in UNIVAC 418 systems is 52,685
words per second with an interlace factor of 3. Interlace factors of 5, 7, and
9 can be used for lower transfer rates. Two input-output channels are required.

•

2 to 16 Uniservo VIC Magnetic Tape Handlers. These units can read forward
only, at a peak transfer rate of 8, 500; 24,000; or 34, 000 characters per second.
The tape format is fully compatible with the 7-track IBM 729 and 7330 Magnetic
Tape Units, and an optional feature permits operation in the 9-track mode. One
input-output channel is required; two channels must be used if simultaneous read/
read or read/write operations are desired.

•

2 to 16 Uniservo vm C Magnetic Tape Handlers. These units can read forward
only, at a peak transfer rate of 24, 000; 67,000; or 96,000 characters per second.
The tape format is fully compatible with the 17-track IBM 729 and 7330 Magnetic
Tape Units, and an optional feature permits operation in.the 9-track mode. One
input-output channel is required; two channels must be used if simultaneous read/
read or read/write operations are desired.

•

1 0758 High-Speed Printer. This unit uses a 63-character set and prints up to
1600 alphameric lines per minute. One input-output channel is required

•

1 Paper Tape Reader and 1 Paper Tape Plmch. These units can read plmched
tape of up to 8 levels at up to 1000 characters per second and plmch standard 5-,
6-, 7-, or 8-level plmched tape at 270 characters per second. One input-output
channel (the same channel as used by the console keyboard-printer) is required.

•

1 UNIVAC 9000 Series Central Processor and associated peripheral's. The 9000
Series Computer provides the UNIVAC 418-ln with plmched card input at 400 or
600 lines per minute and printed output at up to 600 lines Per minute. An optional
card plmch is available which plmches cards at the rate of 200 per minute. The
9000 Series Computer can also be equipped with magnetic tape and paper tape units.
See Computer System Report 810 for full details. The 9000 Series Computer subsystem requires one 418 input-output channel.
•
1 Commlmications Terminal Module Controller (CTMC). Each CTMC can control
up t,o 16 Commlmications Terminal Modules (CTM) each of which will control a
varying number of commlmications lines depending upon the line speed. The
CTMC operates at a peak data transfer rate of 80,000 characters per second and
requires two I/o channels.
Figure 3 shows a UNIVAC 418-Ill System with each of preferred peripheral systems. The
minimum system is shown in heavy lines in the figure and consists of the Command/
Arithmetic Section, one I/O Module including 8 I/O channels, and two banks of 16K words
of core memory.

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

8/68

792:011. 300

UNIVAC 418-111

8AMK 0

lANK 1

r------,
I
I

r-----'

I

I

I

16K 1m2)

I

I

I

16K 1m3}

lANK 2

:
MAIN

STORAGE

--T--'

r--r----..----I

I

I

I EXPAN.I EXPANJ
I II I
I
I
I

1.2

INPUT
OUTPUT

INPUT
OUTPUT
MODULE .0

COMMAND/ARITHMETIC

MODULE II

SECTION

L2!:.''-L2~~..1~~.~~

CHAN

F~:~NG I :~~~:~

0-7

I
l
I"
I

EXPA.No EXPAN

..

I
I
I
I

PROCESSING

SECTION

~1!...L~-!.:J

I

ARITHMETId
____ I....CONYERS
___ ...JI

I DAY I
LI CLOCK
___ ...II

I
I

I

CD CD CO

I

EJ EJ
MAY BE MIXED

Figure 3. UNIVAC 418-111 System Components
.3

SOFTWARE
The UNIVAC 418-m software has been designed specifically for the unique characteristics
of the UNIVAC 418-II1 hardware and both the hardware and software have been designed
primarily for the unique characteristics of real-time applications. The software provided
for the UNIVAC 418-111 includes:
•
UNIVAC 418-II1 Executive, an executive system which provides multiprogramming
capability with job control, file control and file access facilities together with
I/o Handlers;
•
Real-Time Communication Control that has been designed to provide the user
with the same level of interface to real-time input/output operations that he has
to standard peripheral input/output;
•
Language processors which include COBOL, FORTRAN, and UNIVAC 418-II1
Assembler; and
•

Systems Support Libraries which provide system support services of data file
maintenance, program file maintenance, testing, and utility programs:
Sort/Merge
Program Monitor and Trace
Program Maintenance
General :rape/Drum Print
Common Procedures
Data Tape File Malntenanqe
Executive Independent Utilities

8/68

fA..
AUERBACH

(Contd. )

SUMMARY

.31

792;011. 310

UNIVAC 41S-m Real-Time Executive
The Executive provided with the UNIVAC 41S-m will be an advanced executive system providing multiprogramming capabilities and including job control, file control, and file access
facilities.
The UNIVAC 41S-m Real-Time Executive is expected to provide four levels of priority:
•

Critical,

•

Real-Time,

•

Batch, and

•

Computational;

and to allow the user to time-share each level with multiple programs. Further, it will
process all interrupts, including suspension and relocation of active jobs, as well as
assignment of control to the proper handling routine when an interrupt occurs. Scheduling
will be by priority and first-in-first-out within each priority level. Facilities allocation
has been extended to included dynamic acquisition and release of computer system facilities.
The Input/Output service provided by the UNIVAC 41S-Ill Real-Time Executive will include
the normal input/output control and device handler routines. Additionally, three methods
of file access will be provided, each conSisting of a comprehensive set of macro instructions. Details of these access methods are not yet available •
. 32

UNIVAC

41S~m

Real-Time Communications Control

The UNIVAC 41S-Ill Real-Time Communications Control (RTCC) provides an interface to
remote communication facilities which is free from the continuing influence of specific
devices. It provides control over internal routing of messages received from or destined
to remote communication terminals.
R TCC places an input message in a drum storage queue and delivers it to the user on a
GET basis. Similarly, RTCC takes an output message on a PUT basis and causes it to be
transmitted to its destination. Remote Communications terminals initially supported by
RTCC are:

• 33

•

UNIVAC OCT 2000 Data Communications Terminal

•

UNIVAC Uniscope 300 Cathode Ray Tube Display

•

UNIVAC 1004 System

•

UNIVAC 9200 Computer System

•

UNIVAC 2300 Computer System

•

Teletype ASR/KSR Units .

Programming Languages and Compilers
UNIVAC 41S-m users will have both FORTRAN and COBOL available as well as the
Assembly system. The UNIVAC 41S-m Assembler is a machine oriented language compatible with the ART Assembler of UNIVAC 41S-II. Both FORTRAN and COBOL are compatible with the UNIVAC 110S. FORTRAN and COBOL languages permit easy growth into
the larger UNIVAC 110S system should the UNIVAC 41S-Ill user encounter such a need •

• 331 FORTRAN
The UNIVAC 41S-III FORTRAN is expected to consist of the current IBM 7090/7094
FORTRAN language, and to include the USASI FORTRAN of 7 March 1966 as a subset.
No compatibility with FORTRAN II is expected. However. full compatibility with the
FORTRAN IV available on the UNIVAC 418-I/II and UNIVAC 1108 systems is expected •
• 332 COBOL
COBOL for the UNIVAC 41S-III is expected to be developed in accordance with the requirements of the proposed USASI standard for COBOL. Language characteristics provided in
the UNIVAC 41S-m COBOL are from the following modules:
•

Nucleus module

•

Sequential Access module

•

Table Handling module

•

Random Access module

•

Library module
C 1968 AUERBACH Corporation and AUERBACH Info. Inc.

8/68

792: 0 11. 333

UNIVAC 418 III

.333 UNIVAC 418-111 Assembler
ASSEMBLER is the machine oriented assembly system provided with the UNIVAC 418-III
system. ASSEMBLER is source code compatible with ART, the assembly system of the
UNIVAC 418-1/11 system, with expanded facilities to include instructions available on the
UNIVAC 418-111 only. For all intents and purposes, the language of ASSEMBLER is the
same as that of ~RT.
One significant difference between ASSEMBLER and ART is that ASSEMBLER allows sixteen
different location counters to be designated and used. This added feature of ASSEMBLER
enables effective programmer organization of program, data and buffer areas within the
memory banks. Proper use of this facility enables effective use of memory in order to take
full advantage of the higher data transfer rates and increased processing time available
when I/O data transfers and processor activity are occurring in different memory banks
from one another.
• 34

System Support Library
The UNIVAC 418-111 user will have available a System Support Library that provides services in the areas of data file maintenance, data file sorting, program maintenance, and
program testing. Included in this Library are also a number of common procedures as
well as independent utility programs •

• 341 Sort/Merge

The Sort/Merge package will provide a generalized Sort/Merge capability accepting parametrIc Information specifying input and output file descriptions and devices as well as key
size and position in the records. The Sort/Merge program can utilize UNISERVO VI C or
VIII C Magnetic tape handlers or FASTRAND II magnetic drums for input, intermediate,
and/or output data storage.
• 342 Program Maintenance Routing
The Program Maintenance routine provides the UNIVAC 418-ll1 Operating System user with
a method of performing maintenance services in the following major areas:
•

The creation of new source and/or object code libraries using both worker and
systems software programs as input.

•

The generation of updated libraries constructed by selecting desired portions of
one or more existing libraries.

•

The creation of control information and source input for the Assembler.

•

The ability to copy a program, one or more segments of a program, or any
number of programs.

•

The maintenance and correction of source programs.

• 343 Data Tape File Maintenance

The Data Tape File Maintenance program provides a utility routine which updates and makes
corrections to a data file. Several options are incorporated in the program to facilitate
mainten~ce of a file.
The program is designed to provide one compact program which allows for adjustments
that must be made to a data file in dally use. Corrections may be applied to segments or
an entire file. The file may be copied to another tape or printed; segments of the data file
may be compared to those on another tape, or the file may be poSitioned to a given segment.
The type of maintenance desired is specified by parameter cards.
The Data Tape File Maintenance program provides several options to the user. These
options are correct, copy, compare, print, and position.
•

Correct Option
The correct option allows the user to
- delete files or blocks
- replace files, blocks, or words
- insert corrections into files or blocks

8/68

•

Copy Option
The copy option provides the user with the ability to copy all or part of a data file
to another tape.

•

Compare Option
The compare option allows the user to compare all or part of two data files.

A

(Contd. )

AUERBACH

'"

SUMMARY

792:0 t t. 343

.343 Data Tape File Maintenance (Contd.)
•

Print Option
The' print option prints the first five words of the specified files or blocks, or
the number of words specified in a file or block.

•

Position Option
The position option allows the user to position a tape to the file or block specified.

· 344 General Print and Punch Routine
The General Print and Punch (GPp) routine provides the user with a means of printing and
punching from magnetic tape or drum storage without restriction by format conventions.
Additionally, the user has a choice of five conversion codes for display of the data on the
printer. The GPP routine operates in two modes:
•

the testing mode, where all the information to be output is edited including tape
block numbers or drum addresses;

•

the data mode, where only edited print or punch images are output.

The codes are:
USASCn (United States of America Standard Code for Information Interchange)
BAUDOT
FIE LDA TA
XS-3
OCTAL
The five conversion codes are available only in the testing mode.
· 345 Program Trace Routine
The Program Trace routine provides a testing aid designed to protect operational real time
programs that are running concurrently with program testing by monitoring each instruction
executed by the program placed under Program Monitor Control.
· 346 Common Procedures
These routines are a number of procedures and subroutines designed to facilitate programming for batch applications. The, Common Procedures are usually on the library tape in
symbolic format and must be merged with the worker program through a library run. '
The Common Procedures offered may be divided into five groups:
•

Two- and Four-Word Compare
Simulation of 36-bit or 72-bit registers for comparison purposes.

•

Conversion Routines
These provide single-word; double-word (36 bits) or four-word (72 bits) conversion
as follows:
Binary to X8-3
Binary to Fieldata
XS-3 to Binary
Fieldata to Binary

•

Edit Routines
These routines provide the user with aids to clear any area in storage, to left or
right shift, and to complement data in four simulated 36-bit registers.

•

Main Storage Search
This routine permits the operator to search any area of main storage for a specific bit pattern. This routine is controlled by specifying a mask containing the bit
pattern desired and the beginning and ending addresses of the storage area to be
searched. If an equal comparison between the mask and a word of storage within
the parameters is found, the address and contents are printed.

•

Storage Dump to Tape
This routine permits the operator to dump an area of main storage to a magnetic
tape. Any area of main storage may be dumped by specifying the beginning and
ending addresses of the storage area to be dumped. The tape is written in a
format having a predetermined block size and density which can be initially
loaded.

C 1968 AUERBACH Corporation and AUERBACH Info, Inc.

8/68

792:011. 346

UNIVAC 418-111

. 346 Common Procedures (Contd.)
•

Tape Copy
This routine permits the operator to copy any magnetic tape regardless of format,
density, and block size. This routine is controlled by specifying the format,
density, and block size of the tape to be copied.

•

Drum Storage Dump
This routine permits the operator to dump an area of the drum to the systems
printer. Any drum area may be dumped by specifying the beginning 'and ending
addresses of the area desired.

•

Logical Routines
These aid the user by providing simulated 36-bit logical sum, logical product, and
logical difference instructions.

•

Double Precision Arithmetic
Through simulation of auxiliary 36-bit and 72-bit registers, the following arithmetic
services are offered:
Two-word addition or subtraction
Four-word addition or subtraction
Multiplication

Division
.347 Executive Independent Utilities
The Executive Independent Utilities are a collection of routines which provide the user with
service routines for initial program testing and for initial software systems installation.
As the name implies, these routines do not run under the control of the Executive. The
routines included in the Executive Independent Utilities package are:
•

Main Storage Inspection and Change
This routine permits the operator to display any main storage address and change
the contents of that address if so desired.

8/68

•

Main Storage Fill
This routine permits the operator to place a specific data pattern in consecutive
addresses in any area of main storage. The parameters for this routine are the
data pattern to be entered in storage and the beginning and ending addresses of the
storage area to be filled.

•

Utilities Move
This routine permits the operator to move the entire Executive Independent
Utilities package from one storage area to another. The utility package resides
within one storage area and is designed to operate in any storage area.

•

Program Loader
This is a simple load routine which makes it possible to load the relocatable binary
coded output of the Assembler. Magnetic tape or punched cards may be used as the
load medium. Programs loaded by this routine are loaded absolutely and are not
relocated.

•

Main storage Dump
This routine permits the operator to dump any area of main storage to the systems
printer. This routine is controlled by specifying the beginning and ending addresses
of the storage area to be printed.

fA..
AUERBACH

-.6.

800:011. 100
m .....

~EDP

AUE~

UNIVAC 490 SERIES
SUMMARY

""'"

....-

SUMMARY
.1

SUMMARY
The UNIVAC 490 Series consists of three newer medium-to-large-scale computer systems
(the 491, 492, and 494) and one older system (the 490) which was iDitially delivered in
December of 1961. The three more recent systems, announced in JUDe of 1965, are also
mown as the "UNIVAC Modular 490 Real-Time Systems."
The original UNIVAC 490 system is no longer being actively marketed. The 491 and 492
systems are no longer in production but are available on an "as returned" basis. The 494
system, at this writing, is still in production and offered in UNIVAC's standard product
line.
The 490 Series is designed primarily for applications that require control based upon
continuously updated records. Examples of this type of real-time application, in which
it is essential or highly desirable to reduce the time lag between the occurrence of a
transaction and the corresponding updating of one or more master files, include airline
reservation systems, savings bank operations, production scheduling, inventory control, and order processing. Message switching is another important application. The
490 Series is also suitable for commercial applications of the more conventional batch
processing type, particularly when they are run as ·"background" programs to use the
processor time periods that would otherwise be idle between real-time transactions.
The principal characteristics that make the UNIVAC 490 Series suitable for real-time
applications are:
• A variety of fast, large-capacity random-access storage units for masterfile data and systems programs.
• Hardware and software facilities that permit concurrent processing (multiprogramming) of real-time and batch programs.
• Flexible data communications equipment that facilitates two-way communications between the computer and remote points.

Figure 1. Control Console of the UNIVAC 494 Real-Time System.
© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

9/68

UNIVAC 490 SERIES

800:011. 200

.1

SUMMARY (Contd. )
The original UNIVAC 490 system evolved as a commercial outgrowth of UNIVAC's Defense
Systems computer development work. Originally conceived as a special-purpose system
for airline reservatiOns, the 490 was later successfully applied to a wide range of other
commercial applications. A major factor in enhancing the saleability and effectiveness of
the 490 was the development of REX, an integrated, drum-oriented operating system capable of controlling the concurrent operation of one real-time program and one or more
batch-type programs. REX is used by the majority of 490 installations, and serves as the
standard operating system for the newer UNIVAC 491 and 492 as well.
The major change in the original UNIVAC 490 system during its four-year production cycle
was the introduction of an optional feature that improves its basic memory cycle time from
6 to 4. 8 microseconds, with proportional increases in internal processing speeds. About
60 UNIVAC 490 systems were delivered.

•2

The three more recent members of the 490 Series follow the industry trend by offering
significantly more performance per dollar than their predecessor. Using a typical 10-tape
system (our standard Configuration VIIA) as a basis for comparison, the original UNIVAC
490 system, with a 6-microsecond cycle time, rents for $31,270 per month. The newer
UNIVAC 491, with a 4.8-microsecond cycle time, rents for $23,715 per month - a 24%
reduction in rental. The UNIVAC 492 is identical to the 491 except that the 492 provides
six more I/o channels at a rental increase of $1,750 per month. The powerful UNIVAC
494, with twice the core storage capacity and with actual and effective cycle times of 0.75
and 0.375 microseconds, respectively, rents for $40,045 per month, or only 29% more than
the much slower 490. The 494 also provides an expanded instruction repertoire and improved multiprogramming capabilities. It is clear that UNIVAC's marketing strategy in
announcing the three new systems was to attract new customers through the lower price
tags on the 491 and 492, while retaining present customers by enabling them to trade up to
the more powerful but program-compatible 494 at modest increases in cost.
CENTRAL PROCESSORS AND CORE STORAGE
In all four of the UNIVAC 490 Series processors, each 30-bit word location in core storage can hold one instruction, one 30-bit or two 15-bit binary data items, or up to five
alphameric characters. Core storage capacity can range from a minimum of 16,384
words (in all models) to a maximum of 32,768 words in the 490, 65,536 words in the
491 or 492, and 131,072 words in the 494. Parity checks upon internal operations are
performed only in the 494. Cycle times and other features of the 490 Series processors
and core storage units are summarized in Table 1.
TABLE I: CHARACTERISTICS OF THE UNIVAC 490 SERIES PROCESSORS

9/68

Processor Model

UNIVAC 490

UNIVAC 491

UNIVAC 492

UNIVAC 494

Maximum No. of I/O Channels

14
(12 available)

8
(6 available)

14

24
(23 available)

Core Storage Cycle Time, ,",sec

6.0
(4.8 optional)

4.8

4.8

0.75
(0.375 effective)

Core Storage Capacity,
30-bit words

16,384 or
32,768

16,384 to
65,536

16,384 to
65,536

65,536 to
131,072

Core storage Protection

No

Yes;
1,024-word
increments

Yes;
l,024-word
increments

Yes;
64-word
increments

Core Storage Overlap

No

No

No

Yes

Core Storage Parity Checking

No

No

No

Yes

Floating-Point Arithmetic

No

No

No

Yes

Double-Precision Arithmetic

No

No

No

Yes

Decimal Arithmetic

No

No

No

Yes

Maximum I/O Data Rate,
characters/second

417,000

521,000

521,000

2,747,000

A

(12 available)

(Contd. )

AUERBACH

'"

800:011. 201

SUMMARY

.2

CENTRAL PROCESSORS AND CORE STORAGE (Contd. )
Facilities common to all of the 490 Series processors include a full complement of fixedpoint binary arithmetic, Boolean, comparison, and shifting operations. Facilities for
editing and radix conversion, however, are conspicuously absent. Anyone instruction
can be automatically repeated up to 32,767 times, permitting efficient table lookup and
accumulate operations. There are seven index registers, with a typical set of related
instructions for loading, testing, and storing them. (The 494 has two sets of seven index
registers to facilitate operating system control.)
Sixty-two basic single-address instructions are common to all of the 490 Series processors.
Each of these basic instructions consists of five distinct parts: a 6-bit operation code; a
3-bit field that can specify a variety of conditions under which a skip or jump shall occur;
a 3-bit field that specifies whether the operand shall be a full word, a half word, or a
literal; a 3-bit index register designator; and a 15-bit field that can specify an operand
address, a literal operand, or a shift count. This flexible instruction format permits
numerous variations of each of the 62 basic instructions.
The UNIVAC 494 has an expanded instruction repertoire that provides a full range of
double-precision arithmetic, floating-point arithmetic, decimal arithmetic, and enhanced
character-handling facilities. The 47 additional instructions which are unique to the 494
exceed the capacity of the 490 Series' 6-bit operation code field, so UNIVAC uses the
next 6 bits of the instruction word to specify the operation code for these additional instructions. As a result, the 47 instructions which are unique to the 494 cannot specify the use
of partial-word operands or transfers of control based upon the results.
Average execution time per instruction in a basic UNIVAC 490 Processor is about 10
microseconds. The longest instruction - Divide - requires 86.4 microseconds, while
a few instructions require as little as 6 microseconds. All instruction times for a 491,
492, or a 490 with the optional 4. 8-microsecond memory are exactly 20 percent shorter
than the times for the basic 490.
Average instruction execution time for a UNIVAC 494 system can approach the actual cycle
time of 0.75 microseconds when odd/even memory-bank overlapping is employed. This
means that the next instruction can be read from one memory bank while the processor is
executing an instruction that references an operand in the other memory bank.
The 490 Series processors have effective program interrupt facilities which cause a
transfer of control to one of 44 to 73 fixed core locations (depending upon the model)
upon completion of an I/o operation, upon detection of a processor or I/O error, or
upon overflow of either the real-time clock or the day clock. Interrupts from any or
all I/o channels can be enabled or disabled by means of special instructions.
Storage protection facilities, which prevent user programs from gaining unauthorized
access to specified areas of core storage, are an important factor to consider in evaluating computers with multiprogramming capabilities. The original UNIVAC 490 system
has no storage protection facility. The 491 and 492 contain hardware facilities that permit individual 1, o24-word blocks to be guarded against unauthorized access. The 494
provides effective protection through a combination of hardware facilities and the Omega
operating system. The "Guard Mode, " in which user programs will normally operate,
prohibits the use of input-output instructions and other instructions reserved for operating system use. Individual 64-word blocks of core storage can be protected against
writing only, or against both reading and writing. Attempted violations of storage
protection cause program interrupts.
The maximum number of input-output channels available for each of the 490 Series
processors is indicated in Table I. In every 490, 491, and 492 system, one channel is
reserved for the console and one for the real-time clock. In the 494, a single channel
serves both the console and the clock. Each of the remaining channels, in general, can
accommodate one peripheral subsystem and can handle one data transfer operation at a
time. The gross I/o data rates for all simultaneously-operating peripheral devices are
limited to the figures shown in Table I •

•3

PERIPHERAL EQUIPMENT
Probably the most noteworthy a&pect of the UNIVAC 490 Series peripheral equipment is
the numerous drum storage units and magnetic tape units that are available. Table II
summari:tes the characteristics of the three head-per-track "Flying Head" drums and
the two Fa strand units. The Flying Head drums provide rapid access to moderate
amounts of data, while the Fa strand units use movable access mechanisms and store
larger amounts of data, but with slower access times and data transfer rates. A smaller,
less expensive "Modular Fastrand" subsystem was announced along with the newer 490
Series processors, but it was withdrawn from the line later in 1965. UNIVAC's line
of mass storage devices for the 490 Series still lacks a unit with interchangeablecartridge capabilities.

© 1968 AUERBACH Corporation and AUERBACH Info. Inc.

q. 1i8

800;011,300

UNIVAC 490 SERIES

TABLE II: CHARACTERISTICS OF UNIVAC 490 SERIES DRUM STORAGE UNITS
FH-432
FH-880
FH-1782
Fa strand
Fa strand
Device
Drum
Drum
Drum
I
II
6
6
6
6
1.31 x 10
3.93 x 10
10.5 x 10
65.3 x 10
130.7 x 10 6
storage capacity, 6-bit
characters per unit
Storage capacity, 6-bit
11.8 x 10 6
characters per subsystem

31. 5 x 10 6

83.9 x 10 6

519 x 10 6

1,038 x 10

Average access time,
msec

4.25

17

17

92

92

Data transfer rate,
characters/second

240,000

60,000

240,000

25,150

25,150

Usable with 490

No

Yes

No

Yes

No

Usable with 491/492

No

Yes

No

Yes*

Yes

U sable with 494

Yes

Yes

Yes

Yes*

Yes

6

*Not actively marketed; available as a "compatibility option. "
.3

PERIPHERAL EQUIPMENT (Contd.)
Table m summarizes the characteristics of the five magnetic tape units available for 490
Series systems. UNIVAC now encourages use of the Uniservo VIC or VIIlC tape uIrlts,
which use "industry-compatible" (i. e., IBM 729-compatible) 7 -track tape. Optional dualchannel controllers permit read-write simultaneity within a single Uniservo VIC or VilIC'
subsystem. The other three tape units were available for the original UNIVAC 490 system,
and they are still offered as "compatibility options" to postpone or eliminate the need to
convert large existing tape inventories.
Other peripheral equipment available for the 490 Series systems includes the following:
•

Punched Card Subsystem: Consists of one Card Control and Synchronizer,
one Card Reader, and/or one Card Punch. In UNIVAC 491, 492, and 494
systems, cards are read at the rate of 800 cards per minute (or 900 cpm if
only the first 72 columns of each card are read) and punched at 300 cards
per minute. Reading and punching can be performed in Hollerith, row binary,
or column binary mode. (UNIVAC 490 systems use a 600-cpm card reader
and a 150-cpm punch. )
• High-Speed Printer Subsystem: Consists of a Control and Synchronizer Unit
and one Printer. Maximum speed is 700 alphanumeric/922 numeric or 1200
alphanumeric/1600 numeric 132-character lines per minute. There are 63
printable characters.
TABLE m· CHARACTERISTICS OF UNIVAC-490 SERIES MAGNETIC TAPE UNITS
Uniservo
Uniservo
Uniservo
Uniservo
Uniservo
Device
ITA
lIlA
mc
VIC
VIIIC
Tape Speed, inches/ second
100
100
112.5
42.7
120
Recording Density, rows/inch

125/250

1,000

200/556

200/556/800

200/556/800

Peak Data Transfer Rate,
RJIo-characters/second

12.5/25.0

125

22.5/62.5

8.5/23.7/34.1

24.0/66.7/
96.0

Tape Units per Controller

2 to 12

2 to 16

2 to 12

1 to 16

1 to 16

IBM 729-Compatible

No

No

Yes

Yes**

Yes**

Read Backward Capability

Yes

Yes

No

Yes

Yes

Read-After-Write Checking

No

Yes

Yes

Yes

Yes

U sable with 490

Yes

Yes

Yes

No

No

Usable with 491/492/494

Yes*

Yes*

Yes*

Yes

Yes

* Not actively marketed; available as a "compatibility option."
** Optional feature provides compatibility with the 9-track IBM 2400 Series Magnetic Tape Units
used with System/360.

9/68

A

(Contd. )

AUERBACH

'"

SUMMARY

.3

800;011 .400

PERIPHERAL EQUIPMENT (Contd.)
•

•4

UNIVAC 1004 Subsystem: The 1004 is a small, plugboard-programmed computer that oan be conneoted on-line to a 490 Series system and can perform
editing and input-output funotions. The 1004 oan read oards at 400 or 615 cards
per minute and can print at 400 or 600 lines per minute, depending upon the
model. other peripheral equipment that oan be connected to the 1004 inoludes
a 200-cpm card punch and one or two magnetic tape units.

• Data Communication Subsystem (For UNIVAC 491, 492, and 494 systems):
Consists of 1 Communioation Terminal Module Controller and 1 to 16 Communioation Terminal Modules, each of whioh can oontrol a maximum of 2
input lines and 2 output lines. Up to 64 communications lines oan thus be
multiplexed into a single I/o channel. This multiplexing equipment enables
the computer to send and receive data via most oommon-oarrier facilities
at transmission rates of up to 50,000 bits per second. The original UNIVAC
490 system uses similar communioations equipment, although its nomenclature is different and the cost for small configurations is higher .
SOFTWARE
The introduction of a series of new computer systems that are program-compatible with
an earlier system has obvious advantages for the manufacturer as well as for the user.
Software developed and perfected for the older system can be supplied with the newer
systems, thereby relieving many of the pressures usually associated with the software
development process.
UNIVAC 491 and 492 systems oan utilize all of the existing UNIVAC 490 software. When
operating in the special 490-compatible mode, UNIVAC 494 systems can use the existing
software, but this mode does not permit full utilization of the 494's expanded capabilities.
For this reason, recent software development work has been concentrated upon new
facilities for the 494. UNIVAC has made subset versions of most of the 494 software available for the 490, 491, and 492, as replacements for the software originally developed for
the 490. This approaoh to software development has two advalltage/::l for the UNIVAC 490
user who eleots to retain his present equipment: he is assured of continued maintenance
of the present software, and later he will be able to use a set of completely new, improved
software facilities •

. 41

UNIVAC 490 Software
Programs developed for the UNIVAC 490, all of which are curr~ntly available and usable
with UNIVAC 491, 492, and 494 systems as well, can be summarized as follows:
• REX - An operating system capable of controlling a single real-time program
and one or more batch programs, all operating concurrently. REX is designed
to provide for efficient utilization of the available system components and to
process a scheduled set of jobs with a minimum of operator intervention. REX
requires a magnetic drum, at least one magnetic tape unit, and an average
of about 4,000 core locations.
• SPURT - An assembly system that translates symbolic source programs into
machine-language object programs in relocatable or absolute form. At least
four magnetic tape units are required. Facilities for user-defined macroinstructions are available only for systems that include a Fastrand or Flying
Head Drum.
• COBOL - A compiler for COBOL-61 source programs that operates under
control of REX and produces a SPURT-coded symbolic program as output.
All of Required COBOL-61 and a number of useful electives and extensions
have been implemented. A magnetic drum and at least five magnetic tape
units are required for COBOL compilations.
• Sort/Merge - A generalized routine that sorts data on magnetic tape according to programmer-specified parameters. The cascade method is used for
the merge passes. From 3 to 12 Uniservo tape units on a single channel can
be used, and an FH-880 Drum can be used in the presort phase when available.
Sprting can be performed concurrently with a real-time program, under control of REX.
• Utility Routines - A series of generalized routines to perform such functions
as:
Transcribing data from one perlpheral medium to another;
Tracing and monitoring programs;
Maintaining program libraries on magnetic tape;
Transcribing programs from a library tape to a Master
Instruction Tape in a specified sequence.
© 1968 AUERBACH Corporation and AUERBACH Info. Inc.

9'68

800:011. 410

.41

UNIVAC 490 SERIES

UNIVAC 490 Software (Contd.)
•

Library Subroutines - Sixty subroutines designed to handle frequentlyencountered programming tasks such as:
Multi-precision arithmetic on binary or Fieldata-coded items;
Character insertion and extraction;
Radix conversions between Fieldata and binary formats;
Editing (zero suppression, floating dollar Sign, etc.);
File control;
Data movement, scaling, and rounding .

. 42

UNIVAC 494 Software
The software being developed especially for the UNIVAC 494 centers around a comprehensive
operating system called Omega. If the term "third-generation" can be applied to software,
as well as to hardware, then Omega is a true third-generation operating system. The lessons
learned in implementing and applying REX, the UNIVAC 490 operating system, were used
as foundations for the development of Omega.
Omega is designed to control the scheduling and execution of a mix of independent realtime and batch-type programs in a multiprogramming mode. Assigned prioritieS and
balanced utilization of the system's facilities are the governing factors. Conflicting user
programs are "rolled out" of core storage and restarted when the facilities required for
their continued operation again become available. Exclusive control and allocation of all
system facilities by Omega allows changes in configuration and/or operating procedures
without direct impact on user programs. Omega requires 20,000 words of core storage
for its resident routines, plus at least 262, 143 words of drum storage.
The collection and loading of the routines required for a particular task is facilitated
by having all source-language processors produce a common form of relocatable output.
An integrated test system facilitates debugging operations and permits testing of new
programs concurrently with the real-time operation of other programs.
The processing of batch-type programs is facilitated by Omega's facilities for automatic
job-to-job transitions, communication within and between jobs, and services such as
logging and accounting. An umsua! feature of the batch processing environment is
Omega's ability to provide multiprogramming within an individual activity. This "Fork
and Join" function allows, for example, the second pass of a sort to begin processing
the initial output of the first pass while the first pass is still transcribing data. The
ability for computer systems with random-access storage to perform this type of processing is not new, but including this ability as a general software option is quite novel.
The following source-language processors are being developed for use with Omega:

9/68

•

COBOL - The COBOL compiler for the UNIVAC 494 is based on the
language defined in the Department of Defense report, COBOL 1961
Extended. Source-language compatibility with the existing COBOL-61
compiler for the UNIVAC 490 is stressed. The 494 compiler, however,
generates a basically "straight-line" form of object coding, whereas the
490 compiler uses generalized subroutines. Compilation times, execution times, and object program memory requirements are said to be
reduced by the straight-line method. Additional time is saved by having
the compiler's output in the generalized relocatable-Ioader format, thereby eliminating the separate assembly phase that the 490 COBOL compiler
requires. The subset version of the 494 COBOL compiler, for use with
UNIVAC 490, 491, and 492 systems, was made available with the initial
release in the fourth quarter of 1966. A minimum of four magnetic tape
units and one drum are required for compilation.

•

FORTRAN IV - This one-pass compiler accepts a source language based
upon the A. S. A. working specifications for FORTRAN as published in the
Communications of the ACM, October 1964. No complex or logical operations are provided. Object-program execution speeds are much higher
on UNIVAC 494 systems than on the other 490 Series members because
of the 494's inherent speed advantage and its built-in facilities for floatingpoint arithmetic.

•

UNIVAC 494 ASM - The form of this new symbolic assembly system for
the UNIVAC 494 resembles that of the SLEUTH II assembly system for
the UNIVAC 1107, which features extensive macro-instruction facilities.
The new system facilitates utilization of the 494's expanded facilities.
The SPURT assembly system developed for the UNIVAC 490 is still used
in 494 installations where program compatibility with the smaller 490
Series processors is considered important.

A

AUERBACH

'"

800:221.1 01

-A

" ••••11

UNIVAC "90 SERIES
PRICE DATA

BDP
Iltim

UNIVAC 490
UNIVAC does not include monthly maintenance charges in their published prices;
these charges have been included in the Monthly Rental column in this Digest
to permit convenient comparison with other systems.

PRICES

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

Monthly
Monthly
Rental PurchaSE Maint.

Name

$

$

$

10,720

460,000

1,570

15,100

622,000

1,600

10,470

450,000

1,570

250

10,000

250

20,000

-

words,

1,545

60,000

15

words,

2,580

100,000

35

words,

3,920

152,000

50

words,

4,645

180,000

70

65
125
50
335
335

2,700
5,400
2,250
13,000
13,000

5
10

155
155
168
75

6,670
6,670
6,540
3,375

10
10
18
21

800
1,000
1,500
350
1,160
205

36,000
45,000
67,500
20,500
7,840

154
181
301
350
275
20

1,480

59,340

190

2,845

75,000

115

3,430
2,150

160,000
85,165

290
190

4,050

164,640

300

Processing Unit

PROCESSOR
*8186-01
*8187-01
*8186
* F0957-00
*F0957-01

Central Processor (16K words, 4. 8-/l sec
memory
Central Processor (32K words, 4. 8-/l sec
memory
Central Processor (16K words, 6. O-/l sec
memory
Conversion Set (for field conversion of
8186-00 to 8186-01)
Conversion Set (for field conversion of
8187-00 to 8187-01)
Main Storage
Core Memory Expansion (expands 32K
4. 8-/l sec memory to 40K words)
Core Memory Expansion (expands 32K
4.8-/l sec memory to 49K words)
Core Memory Expansion (expands 32K
4. 8-/l sec memory to 57K words)
Core Memory Expansion (expands 32K
4.8/l sec memory to 65K words)

ATTACHMENTS,
ADAPTERS,
AND
CHANNELS

*8048-00
*8049
*8158
*2502-00
*2502-01
*F0700-00
*F0700-00
* F0893-01
*8056
*8150
*8151
*8153
*2510-00
*0951-00
* F0597-02

MASS
STORAGE

Transfer Switch (Si~le)
Transfer Switch (dua )
Cabinet (for 8048 or 8049 transfer switches)
Cabinet (for F0700-00)
Cabinet (for F0700-00; contains integral
operator panel)
Transfer Switch (electronic)
Transfer Switch (electronic)
Transfer ~itch (electronic)
Special Peripheral Cabinet (central 8-foot
cabinet)
Scanner Selector (4 channel)
Scanner Selector (8 channel)
Scanner Selector (16 channel)
Line Isolator (wide band)
lntercomputer Coupler (490 to mM 7094)
Adapter (490 to UNIVAC 1004/1005)

-

-

--

Drum Storage
*8103-00
*8205-00
*0900-00
*8102
6010-00

Drum Control (buffered control for 1 to 8
FH880 Drums)
Drum Control (buffered control for Fastrand
I and II units)
Fastrand I Mass Storage Unit
FH-880 Drum (786,432 words)
Fastrand II Mass Storage Unit
(25,952,256 words)

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

UNIVAC 490 SERlE

800.221.1 02

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

Name

PRICES

Monthly
Monthl)
Rental !Purchase Maint.

$

$

$

Punched Card

INl'l'TOl 'T I' liT

1,675
365
535

80,000
17,500
25,000

265
85
205

675

25,350

125

1,510
1,820
2,855

72,500
27,440
54,880

160
180
245

530
850

25,000
36,000

150
275

(4 simplex positions)
(S simplex positions)
(16 simplex positions)
(32 simplex positions)
(64 simplex positions)
(32 external pOSitions)

720
777
850
l,On5
1,380
1,050

20,160
21,640
23,880
29,880
38,800
47,000

215
230
250
315
410
285

Multiplexer (64 external interrupt)

1,350

60,500

368

215
265
310
370
425
55

6,000
7,400
9,000
10,400
12,000
2,475

H5
80
95
110
125
15

65

2,250

15

55
30

2,475
1,350

15
8

30

760

10

21

560

7

33

H2O

10

30

760

10

38

1,040

12

30

760

10

50

1,360

15

50

1,3HO

15

40
40
20

1,315
1,315
645

10
10
5

Card Control
Card Header (600 cards/min)
Card Punch (150 cards/min)

'Sl04
'7!lO(;-01
*8124

Paper Tape
*813(;-00

Paper Tape Subsystem (read 400 char/sec;
punch 110 char/sec)
Printer
-.--

*810(;
*1-\120

Printer Control (controls one 8116 Printer)
Printer Control (controls one 0751-01 Printer)
Printer Control-Dual (controls two 0751-01
Printers)
Printer (600 lines/min)
Printer (700/922 lines/min)

*~l~O-Ol

'8llH
*0751-00

Multiplexers

CO;\DIUN1CATIONS
*F0900-04
*F0900-03
* F0900-0!'!
*F0900-0l
*F0900-00
*F0900-97
(XA"XX-02)
*F0900-98
(XA"XX-03)
*F0906-04
* F0906-03
*F0906-02
*F0906-01
* F0906-00
*F0878-00

Multiplexer
Multiplexer
Multiplexer
Multiplexer
Multiplexer
Multiplexer

Multiplexer (4 simplex positions)
Multiplexer (8 simplex pOSitions)
Multiplexer (16 simplex positions)
Multiplexer (32 simplex positions)
Multiplexer (64 simplex pOSitions
Comm Line Terminal (high speed output;
5-through 12-level; synchronous)
* F0878-01
Comm Line Terminal (high speed input;
5-through 12-1evel; synchronous)
* (X:>'"XX-04) C L T (high speed input; 5-through 12-1eve 1;
Asynchronous; external interrupt)
* (X).."XX-05) CLT (low speed input; 6-,7- and 8-level;
Asynchronous; external interrupt)
* F0901-00
Comm Line Terminal (low speed output;
5-1evel; asynchronous)
'" F0901-01
Comm Line Terminal trow speed input;
5-level; asynchronous)
*F0901-02
Comm Line Terminal (low speed output;
6-,7-, and 8-level; asynchronous)
*FODOI-03
Comm Line Terminal (low speed input;
6-,7-, and 8-1eve1; asynchronous)
*FOH02-00
Comm Line Terminal (medium speed output;
5-,6-,7-, and 8-1evel; asynchronous)
* F0902-01 Comm Line Terminal (medium speed input:
5-,6-,7-, and 8-1evel; asynchronous)
* FOD03-00 Comm Line Terminal (high speed output;
5-,6-,7-, and 8-level; asynchronous)
* F0903-01 Comm Line Terminal (high speed input:
5-,6-,7-, and 8-level; asynchronous)
* FO~)04-00 Comm Line Terminal (parallel output)
* F0904-01 Comm Line Terminal (parallel input)
Audodia ling
'" FOP05-00
'\OTES:
*:\0 longer in production

2/69

A

AUERBACH

800:221.103

JNIVAC 400 SERIES

UNIVAC 491/492
UNIVAC does not include monthly maintenance charges in their published prices; these charges have been
included in the Monthly Rental column in this Digest to permit convenient comparison with other systems.

PRICES

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

PROCESSOR

Monthly
Monthly
Rental Purchase Maint.

$

$

$

Processing Unit (includes core storage)
491 Processors (include 8 I/O channels):
4!'l Processor (16K words)
491 Processor (32K words)
491 Processor (40K words)
491 Processor (49K words)
491 Processor (57K words)
491 Processor (65K words)
492 Processors (include 14 I/O channels):
492 Processor (16K words)
492 Processor (32K words)
492 Processor (40K words)
492 Processor (49K words)
492 Processor (57K words)
492 Processor (65K words)

*8187-99
*8187-98
*8187-97
*8187-96
*8187-95
*8187-94
*8187-93
*8187-92
*8187-91
*8187-90
*8187-89
*8187-88
ATTACHMENTS,
ADAPTERS,
AND
CHANNELS

Name

*F0700-00
*2502-02
*2502-03
*2502-04
*2502-05

Transfer Switch (electronic)
Switch Cabinet (for up to 8 Switches; remote
opera to!' panel; cables enter through top)
Switch Cabinet (same as 2502-02 except
contains integral operator panel)
Switch Cabinet (for up to 8 Switches; remote
operator panel; cables enter through bottom)
Switch Cabinet (same as 2502-04 except
contains intergral operator panel)

7,370
11,500
13,030
14,180
15,500
16,130

280,000
440,000
500,000
540,000
592,000
620,000

1,570
1,600
1,630
1,680
1,700
1,730

9,190
13,310
14,860
15,890
17,230
17,950

350,000
510,000
570,000
610,000
662,000
690,000

155
335

6,670
13,000

335

13,000

335

13,000

335

13,000

2,150
1,550
4,050
215
30
1,335

85,165
58,800
164,640
8,235
1,040
53,410

190
190
300
25
5
115

2,670

106,820

230

60

2,355

-

55

2,225

-

515

17,350

115

570

19,800

115

310

10,470

70

745

31,070

35

745

31,070

35

1,690
1,

no

1,760
1,790
1,830
1,850

-

10

Drum Storage

MASS
STORAGE
*7304-01
*8103-01
6010-00
F0686-0l
F0688-0l
5009-08
*5009-09
F0763-00
FOnO-OO

INPUTOUTPUT

FH 880 Drum (786,432 words)
FH880 Control (controls 1 to 8 FH #880 Drums)
Fastrand II (25,952,256 words)
Fastbands
Write Lockout
Fastrand Control (buffered control for up
to 8 Fastrand II units)
Drum Control (buffered dual-channel control
for up to 8 Fastrand II units)
Control Buffer Kit (field conversion
unbuffered control to buffered)
Search All Words
Magnetic Tape

0858-00
0858-08
0858-01
5008-04
*5008-05

Uniservo VIC Master (7-track; 200/556/800
bpi; read backword; handles up to
3 VIC Slaves)
Uniservo VIC Master (same as 0858-00 but
for simultaneous read-read/read-write
operation)
Uniservo VIC Slave (7-track; 200/556/800
bpi)
Uniservo VIC Control (controls up to 4 VIC
Master units)
Uniservo VIC Control (controls up to 4 VIC
simultaneous Master units)

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

800:221.104

UNIVAC 490 SERIES

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

INPllTOl1TPllT

. Name

PRICES
Monthly
Monthly
Rental iPurchase Maint.

$

$

110
110
860
915

4,410
4,410
32,735
35,085

5
5
110
110

1,550

62,430

120

1,550

62,430

120

110
110
85

4,410
4,410
3,385

5
5
5

415
715
920

12,985
21,560
28,620

115
230
265

Printer (700 lines/min)
Printer Control (controls one 0751-00
Printer)

870
810

25,970
27,440

275
180

Communication Terminal Module Controller
(CTMC) Communications Terminal
Modules (CTM):
Communication Terminal Module (LowSpeed)
Communication Terminal Module (Medium
Speed)
Communication Terminal Module (High
Speed)
Automatic Dialing
Parallel Output
Parallel Input
Spare CTM Controller
Word Terminal Synchronous (WTS):
BaSic Cabinet
Power &tpply
Power &tpply
WTS Module
Voice Band Interface
Unattended Answering
Automatic Dialing
Broad Band
I/O Channel Expansion (for expansion of 491
Processor from 8 to 14 I/O channels)

705

24,700

135

65

2,255

15

85

2,895

15

100

3,630

15

20
40
40
260

635
1,315
1,315
9,020

5
10
10
50

270
110
110
485
5
5
55
5
1,830

A,380
3,335
3,335
14,210
225
225
1,670
225
70,000

80
30
30
160

$
Magnetic Tape (Contd.)

(Contct. I

F0627-04
*FO(i27-0:1
085!l-OO
0859-02
500:l-1(i
*5008-17
F0627-04
*F0627-0:1
F0704-00

Translate Option (for 5008-04)
Translate Option (for 5008-05)
Uniservo VIIIC (7-track; 700/556/800 bpi)
Uniservo VIllC (same as 0859-00 but permits
simultaneous, read-read/read-write
operations)
Uniservo VInC Control (controls 1 to 16
VIIIC Tape units)
Uniservo vrnc Control (required with
5003-16 for simultaneous operations)
Translate Option (for 5008-16)
Translate Option (for 5008-17)
VIC Control Option (allows intermixing of
VIC and vmc Tape units on a vnrc control)
Punched Card

*0706-9!)
0600-00
5010-01

Card Reader (900 cards/min)
Card Punch (300 cards/min)
Card Control (controls one Reader and
one Punch)
Printer

*0751-00
*8120-00
CO;\I:\IU\I-

F0900-05

C.\TIO~S

F0901-04
F0902-02
F0903-02
F0905-00
F0904-00
F0904-01
F0906-05
8552-01
F0614-00
F0614-01
F0771-01
F0772-00
F0772-01
F0772-02
F0772-03
*F0764-00

NOTES:

•

2/69

1\'0

longer in Production

fA..

AUERBACH

15
-

130

100:221.105

PRICE DATA

UNIVAC 494
UNIVAC does not include monthly maintenance charges in their published prices; tbese charges have been
included in the Monthly Rental column in this Digest to permit convenient comparison with other systems.

PRICES

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

PROCESSOR

Name

Monthly
Monthly
Rental Purchase Maint.

$

$

$

9,835

373,600

1,310

1,000
165

32,625
6,600

250
10

165

6,600

10

43,000

15

11,925
21,680
9,755

488,000
888,400
400,400

785
1,405
620

21,680
16,820
11,925
7.055

888,400
688,200
488,000
284.300

1,405
1,110
785
565

275

10,350

35

595

23,140

70

1,000

32,625

250

205
168
405

7,840
6,540
12,740

20
18
60

405

12,740

60

598

20,000

88

598

20,000

88

2,150
1,550
1,070
2,940
2,145

85,165
58,800
42,435
117,210
82,515

190
190
100
260
260

Processing Unit
3012-99
4009-97
F0774-00
F0774-01
*F1088-00

494 Processor (includes 12 110 channels;
Display Console required)
Display Console
Auxiliary Console (for mounting 2 to 4 communication control panels; left or right
addition to Display Console)
Auxiliary Console (same as F0774-00 but
located in center of Dual Console)
494 Console Indicator

-

Main Storage
Storage (65K words)
Storage (131K words)
storage Expansion (65K words; for field
expansion of 65K memory to 131K words)
Core Memory (131K words)
Core Memory (98K words)
Core Memory (65K words)
Core Memory (32K words)

7005-77
7005-76
7005-62
*7005-95
*7005-96
*7005-97
*7005-98
ATTACHMENTS,
ADAPTERS,
AND
CHANNELS

F0745-00
F0745-01
F0745-02

I

F0597-02
F0893-01
2508-00
2508-01
2508-02
2508-03

;\IASS
STORAGE

I/O Channel Expansion (4 additional 250KC I/O
channels; maximum of 24 channels total)
I/O Channel Expansion (4 additional 549KC
channels; maximum of 24 channels total)
I/O Channel Speedup (converts 4 250KC
channels to 549KC channels)
1004 Control
Transfer Switch (manual electronic)
Switch Cabinet (accommodates 5 Transfer
Switches; integral operator panel)
Switch Cabinet (accommodates 5 Transfer
Switches; remote operator panel)
Switch Cabinet (accommodates 4 Transfer
Switches; integral operator panel; auxiliary
power supply)
Switch Cabinet (accommodates 4 Transfer
Switches; remote operator panel; auxiliary
power supply)
Drum Storage

7304-01
8103-03
6016-00
6015-00
5012-02

FH
FH
FH
FH
FH

880 Drum (786K words)
880 Control (controls 1 to 8 drums)
432 Drum (262K words)
1782 Drum (2 million words)
432/PH 1782 Control (controls 1 to 8
FH 432 or FH 1782 Drums in any combination)

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

UNIVAC 490 SERIEl

800:221.106

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

l\IASS
STOHAt1E

F0786-01

(Contll. )

F0767-00
F0929-00
1i010-00
F0686-01
F068S-01
500!l-12
500111-13
5009-97
F076:1-00
F0959-99

F0959-98
(i010-10
5009-87
5009-83
5009-79
F0763-01
F068G-Ol
F0688-01
F0959-97
F095!l-!16

IMonthly
Monthly
Rental IPurchasE Maint.

$
Drum Storage (Contd.)
Dual Channel (provides dual channel access
to FH 432 Drum; required on each Drum)
Dual Channel (provides dual channel access
to FH 1782 Drum; required on each
Drum)
Write Lock Out
Fastrand II (25 million words)
Fastbands
Write Lockout
Fastrand Control (unbuffered control for 1 to S
Fastrand II units)
Fastrand Control (buffered control for 1 to 8
Fastrand units; required for dual-channel
operation)
Fastrand Dual Control (buffered dual-access
control; contains dual access adapter for
first two channels)
Control Buffer (for converting an unbuffered
control to a buffered control)
Dual Access Drum
Adapter (provides dual access to drums from
one of two controls; one required for all
subsequent drums except fifth)
Dual Access Drum
Adapter (same as F0959-90 but for fifth
drum)
Fastrand III Drum
Fastrand III Control
Fastrand III Dual Control
Fastrand III Control (for field upgrade from
single to dual access)
Control Buffer
Fastband
Write Lockout
Fastrand III Dual Access Drum Adapter (drum
2,3,4,6,7,8)
Fas.rand III Dual Access Drum Adapter (drum
5)

$

$

70

2,255

15

70

2,255

15

30
4,050
215
30
1,280

1,040
164,640
8,235
1,040
51,060

5
:100
25
5
115

2,560

102,120

2:10

3,050

122,3:10

240

55

2,350

50

1,8:10

10

50

1,830

10

4,965
1,565
3,650
2,085

200,800
62,220
147,010
84,790

350
135
270
1:15

55
205
30
50

2,350
8,235
1,0--10
1,830

50

1,830

10

515

17,350

115

570

19,800

115

:no
515

10,470
17,350

70
115

570

1!l,800

115

310
745

10,470
31,070

70
35

l,4!l0

62,140

70

-

25
5
10

Magnetic Tape

D:J>l"TOl'TPl"T

0858-00
0858-08
0858-01
0858-10
085R-12
0,-I5H-14
:iOOH-O--1
5008-%
FI02l-00
FI072-00

2/69

Name

PRICES

Uniservo VIC:
Uniservo VIC Master (7-track; 8,540/
23,741/34,160 char/sec; handles up to
3 slaves; nonsimultaneous)
Uniservo VIC Master (same as 0858-00 hut
allows read-read/read-write simultaneity)
l'niservo VIC Slave (7-track)
tTniservo VIC Master (9-track; 800 bpi.
:14,160 bytes/sec; handles up to 3 Slaves;
nonsimultaneous)
Uniservo VIC Master (same as 0858-10 but
allows read-read/read-write simultaneity)
Uniservo VIC Slave (9-tr ack)
Unlservo VIC Control (controls up to 4 nonsimultaneous 7-track VIC Master units)
Unlservo VIC Control (controls up to 4
Simultaneous, 7-track VIC Master units)
7 to 9 Track Conversion (converts 7-track
VIC Tape unit to 9-track)
Simultaneous Capability (converts nonsimultaneous 7-track VIC Master for
simultaneous operation)

A

AUERBACH

•

-

55

2,450

-

800:221.107

RICE DATA

IDENTITY OF UNIT
CLASS

Model
Number

INPUTOUTPUT
(Contd. )

Feature
Number
F1072-01
F0627-04
F0627-99

0859-00

!

0859-02
0859-04
0859-08

0859-10

5008-16
5008-88
5008-81
F0706-00

F0627-04
F0704-00
F0999-00
F0999-04

Name

PRICES
Monthly
Monthly
Rental Purchase Maint.

$
Magnetic Tape (Contd.)
Simultaneous Capability (converts nonsimultaneous, 9-track VIC Master for
simultaneous operation)
Translate Option (for 5008-04)
Translate Option (for 5008-96)
Uniservo VIIIC:
Uniservo VIIIC Tape Unit (7-track; 24,000/
66,720/96,000 char/sec); nonsimultaneous)
Uniservo VIIIC Tape Unit (same as 0859-00
but allows simultaneous read-read/readwrite operation)
Uniservo VIIIC Tape Unit (9-track; 860 bpi;
96,000 bytes/sec; non-Simultaneous)
Uniservo VIIIC Tape Unit (same as 0859-00
except allows dual access operation with
read-read/read-write/write-read/writewrite simultaneity)
Uniservo VIlIC Tape Unit (same as 0859-04
except allows dual access operation with
read-read/read-write/write-read/writewrite Simultaneity)
Uniservo VIIIC Control (controls 1 to 16
7-track 0859-00 units)
Uniservo VIIIC Control (controls 1 to 16
7-track 0859-02 units)
Control ExpanSion (expands 5008-16 nonsimultaneous control to simultaneous
5008-88 control)
Control Adapter (adapts 7-track controls
to 9-track format required on 9-track
controls)
Translate Option (for 5008-16)
VIC Control Option (permits intermixing
7-track VIC and VIIIC Tape Units on
5008-16 control)
7 to 9 Track Conversion (converts 7-track
0859-00 to 9-track 0859-04)
7 to 9 Track Conversion (converts 7-track
0859-08 to 9-track 0859-10)

$

$

-

55

2,450

110
220

4,410
8,820

5
10

860

32,735

110

915

35,085

110

875

33,390

110

890

34,045

110

905

34,700

110

1,550

62,430

120

3,100

124,860

240

1,550

62,430

120

50

1,960

5

110
85

4,410
3,385

5
5

15

655

15

655

480
415
65
715
920

15,385
12,985
2,400
21,560
28,620

125
115
10
230
265

1,250
870
885
810
75

43,500
25,970
30,015
27,440
2,575

305
275
200
180
20

810

27,440

180

705

24,700

135

65
85
100
300

2,255
2,895
3,630
635

15
15
15
5

-

Punched Card
0706-97
*0706-99
F1022-00
0600-00
5010-01

Card Reader
Card Reader (900 cards/min)
Check Read
Card Punch (300 cards/min)
Card Control (controls one 0600 punch and are
0706 header)
Printer
---

0758-00
*0755-00
5011-04
*8120-02
F0965-00
*8120-03
COMMUNICATIONS

F0900-06
F0901-04
F0902-02
F0903-02
F0905-00

Printer (1200/1600 lines min)
Printer (700/900 lines/min)
Printer Control (controls multiple 0758 Printers)
Printer Control (controls one 0755 Printer)
Printer Control Capability (permits control
of 0758 Printer)
Printer Control (controls second 0755 Printer)
Communications Terminal Module Controller
(CTMC)
CTM-LS
CTM-MS
CTM-HS
Automatic Dialing

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

2/69

UNIVAC 490 SERIES

800:221.108

IDENTITY OF UNIT
CLASS

COMMlINICATIONS
(Contd. )

Model
Number

Feature
Number
F0904-00
F0904-01
F0906-06
F1027-00
F0148-00
F0991-00
F0988-00
F0988-01
F0989-00
F0989-01
F0989-02
F1018-02
FI018-03
F1018-05
F1019-01
FlO19-03
F1019-05

Name
Printer (Contci.)
Parallel Output
Parallel Input
Spare CTM Controller
CTM IV Low Speed
CTM VI Low Speed
CTM VI High Speed
CTM VII High Speed
CTM VO High f1Jeed (includes block parity
checking)
CTM VII Medium Speed
CTM vn Medium Speed (includes character
parity checking)
CTM VII Medium Speed (includes character
and block parity checking)
HS Interface Module (for 6 full-duplex lines)
HS Interface Module (for 4 fun-duplex lines)
Expansion Kit (expands F1018-03 to 6 lines)
HS Interface Module (8 full-duplex lines)
HS Interface Module (4 full-duplex lines)
Expansion Kit (expands FI019-03 to 8 lines)

PRICES
Monthly
Monthly
Rental PurchasE Maint.
$
$
$
40
40
260
80
95
125
80
95

1,315
1,315
9,020
3,050
3,480
4,570
3,050
3,480

10
10
50
10
15
20
10
15

60
70

2,175
2,610

10
10

80

3,050

10

60
40
20
80
40
40

2,175
1,525
650
3,050
1,525
1,525

10
5
5
10
5
5

NOTES:
(1) UNIVAC will extend rental agreements to a five-year term for systems in current production at a
monthly rental of 85 per cent of the figure shown in this column.
*No longer in production.

2/69

A ..

AUERBACH

801:221. 101

£ "......
/A\\EDP

AUERBACH

UNIVAC 490 SERIES
490 COMPUTER SYSTEM
PRICE DATA

I[POITS

s

PRICE DATA: UNIVAC 490

UNIVAC 490 computer systems are no longer being actively marketed; the following prices
were those in effect while the system was in production. For price data on the UNIVAC
491, 492, and 494 systems, please turn to Page 800:221.101.
IDENTITY OF UNIT

Monthly
Maintenance
$

Purchase
$

With 16,384 Core Memory locations
& 6 I/O Channels

9,500

1,300*

427,500

With 16,384 Core Memory locations
& 14 I/O Channels

10,000

1,365*

450,000

With 32,768 Core Memory locations
& 6 I/O Channels

13,500

1,340*

580,000

With 32,768 Core Memory locations
& 14 I/O Channels

14,000

1,390*

602,000

4. 8-Microsecond Accelerator Package
For 16,384 Core Memory locations
For 32,768 Core Memory locations

250
500

CLASS
No.

INTERNAL
STORAGE

-

-

10,000
20,000

Core Memory
See Central Processor, above
8112
8122

FH-880 Ma!metic Drum Subsystem
FH-880 Drum
FH-880 Control & Synchronizer

2,000
1,420

165
165

92,000
71,000

8206
8205

Fastrand Mass Stor~e Subsystem
Fastrand Storage Unit
Fastrand Synchronizer

3,300
2,750

t

100

160,000
135,000

8143
8113
8142

Uniservo IlA Subsystem
Uniservo IrA Magnetic Tape Handler
Uniservo ITA Control & Synchronizer
Uniservo IrA Power Supply

450
1,530
550

95
130
40

20,000
76,500
25,300

750

155

36,500

3,700

170

177,600

4,800
550

260
40

230,400
25,300

8011
8003-1
8003-3,
8003-5
8142

t

Name
UNIVAC 490 Central Processor
(Including Control Console, Motor
Alternator, and Power Control
Cabinet)

CENTRAL
PROCESSOH

INPUTOUTPUT

PRICES
Monthly
Rental
$

Uniservo IlIA Subsystem
Uniservo IIIA Magnetic Tape Handler
Uniservo IlIA Control & Synchronizer
(Single Channel)
Uniservo IlIA Control & Synchronizer
(Dual Channel)
Uniservo IITA Power Supply

$250 for first unit; $120 for each additional unit.

* APplicable only when Central Processor is used for batch processing applications; maintenance
charges for real-time applications available upon request.
© 1968 AUERBACH Corporation and AUERBACH Info. Inc.

9/68

UNIVAC 490

-801:221. 102

PRICES

IDENTITY OF UNIT
CLASS
No.
INPUTOUTPUT
(Contd.)

Monthly
Maintenance

Purchase

$

$

$

8220
8208
8209
8142

800
1,000
2,250
550

62
50
173
40

38,400
48,000
108,000
25,300

8114
8124
8104

Punched Card Subs;y:stem
Card Reader (600 CPM)
Card Punch (150 CPM)
Card Control & Synchronizer

350
500
1,600

75
180
230

17,500
25,000
80,000

645

110

32,250

8121
8120-00

CLT-50L
CLT-51L
CLT-80L
CLT-81L
CLT-80M
CLT-81M
CLT-80H
CLT-81H
CLT-80P
CLT-81P
CLTDialing

P;mer T;me Subs;y:stem
(Paper tape reader, punch and
control unit)
High-SQeed Printer Subs;y:stem
Printer (700-922 LPM)
Printer Control & Synchronizer
(700-922 LPM)

800

240

36,000

1,750

160

80,000

Communication Line Terminals
Low speed output (5 level)
Low speed input (5 level)
Low speed output (6, 7, 8 level)
Low speed input (6, 7, 8 level)
Medium speed output
Medium speed input
High speed output
High speed input
Parallel output
Parallel input

25
20
30
25
35
25
45
45
35
35

#

1,125
900
1,350
1,125
1,575
1,125
2,025
2,025
1,575
1,575

Automatic Dialing Unit

20

900

675
725
800
1,000
1,300

#

C/M-64

Communication Multi2lexers
For up to 4 CLTs
For up to 8 CLTs
For up to 16 CLTs
For up to 32 CLTs
For up to 64 CLTs

30,375
32,625
36,000
45,000
58,500

8150
8151
8152
8153

Communication Scanner ISelectors
4 Channels
8 Channels
12 Channels
16 Channels

800
1,000
1,250
1,500

#

36,000
45,000
56,250
67,500

C/M-4
C/M-8·
C/M-16
C/M-32

9/68

Monthly
Rental

Uniservo mc Subs;y:stem
Uniservo mc Magnetic Tape Handler
Uniservo mc Tape Adapter Cabinet
Uniservo IIIC Control & Synchroni7.er
Uniservo IIIC Power Supply

8136

#

Name

Monthly maintenance charges will be quoted by the manufacturer upon request.

A

AUERBACH

•

-.£.

810:000.001
m .....

~BDP

AUlBaC~..

UNIVAC 9000 SERIES
REPORT UPDATE

I"un

REPORT UPDATE
On February 24, 1969 UNIVAC announced new members of the 9000 Series. The new models,
designated the 9200-II and the 9300-II, greatly extend the range of appUcations that can be
accommodated by the smaller members of the 9000 Series. The principal extensions are in
the size of main memory available and the type of peripheral devices that can be attached.
These extensions bring to the smaller 9000 Series computers hardware and techniques that
better take advantage of the high degree of processing power available. The lower 9000
Series computers now hold a far different position in the market than previously.
Hardware
The 9200 n will use essentially the same central processing unit as does the 9200; the principal differences are the availability of additional main storage and a change in the I/O structure.
The maximum 3torage size available with the 9200 n is 32,768 bytes. The multiplexor channel,
optional on the 9200, Is standard equipment on the 9200 II. Available as an option, is a selector
channel with a peak transfer rate of 350,000 bytes per second. This feature is intended primarily to handle the R411 Direct Access Subsystem.
Similarly, the 9300 II central processing unit is the same as the 9300, but with both the multiplexor channel ami the 350, OOO-byte-per-second selector channel standard.
UNIV AC has not announced any new peripherals for the 9200 II and the 9300 n; instead peripheral
devices for the 9400 wUl be made available for both models. In addition, peripherals available
on the 9300 will b" made available on the 9200 n.
The multiplexor and selector channels can accommodate up to eight device controllers each.
Essentially, the 9200 n and the 9300 n can accommodate any peripheral device announced for the
9400 except Uniservo 12/16 magnetic tape units and the 16-line data communications controller.
All options for the 9200 and 9300 are available for the 9200 II and 9300 n. The maximum number
of tape units support.e-t for the 9300 II is 16 and for the 9200 II is 8.
For further information on the peripheral devices released for the 9200 n and the 9300
pages in AUERBACH Standard EDP Reports may be consulted:

n.

the

followin~

•

•

•
•
•

Universo VIC: . . . . . . . • . . . . • . . . . . . . . . . . . . . ..
071l··0~

Card Reader (600 cards/min): . . . . . . . . . . . .

810:091.100
810:071. 100

0604 Row Card Punch (200 caros/min): . . . • . . . .

810.074.100

0768 Line Printers (~OO/1100 or 1200/1600
lines/min): . . . . . . . . . . . . . , . . . . . . . . • . . . . . . .

814:011. 315

R411 Disc File Subsystem • . . . . . . . . . . . . • • . • . . . .

814:011. 315

Software
All software for nOO/!l300 Cll.rd, Tape, Disc. and Communications Systems can be used for
the 9200 n and the 9300 II. In addition, UNIVAC has promised to develop several new packages
to make effective ulle of the high speed 0768 printers and the 8411 disc units. The 0768 Printer
will be considered strictly a data printer; any communications with the operator, such as diagnostic printouts from the language processors, will use the integral bar printer. Three levels
of file processing, using the 8411, will be supported:
•

Sequential loading and acceSSing of files;

•

Random access to sequentially loaded files (Indexed Sequential); and

•

Creation and access to randomly sequenced files.

Language processors to be available for 8411-oriented systems include:
•

Assembler/Linker,

•

Report Generator, and

•
COBOL.
Each of these processors operate with a minimum of 24K bytes of main storage.
In addition to the major packages, software for the 8411 will include a modular sort/merge.
library services, a disc-to-print symbiont, and miscellaneous testing aids.
© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

2/69

"

UNIVAC 9000 SERIES

110:000.002

Deliveries of tape-oriented 9200 n and 9300 D systems are soheduled to begin In six months.
Del1verles of diso systems are planned for Maroh 1970.

Name
9200 D Processor
Selector Channel
Storage (24K)
Storage (32K)
Storage Expansion
(16K to 24K)
Storage Expansion
(24K to 32K)
Storage Expansion
(16K to 32K)
9300 D Processor

Monthly Rental, $*
Purchase, $

Monthly Maintenance, $

355
80
1,020
1,350
340

14,270
4,320
47,630
62,860
15,890

70
10
90
120
30

380

330

15,230

30

780

670

31,120

60

910

795

32,890

165

I-year

5-year

400
100
1,185
1,565
400

*UNlVAC does not include maintenance prices in their published prioe lists; for ease
of comparison, maintenance prices ~ been included in this column.

2/69

A

AUERBACH

'"

-.£. "'. '"

810:000.001
UNIVAC 9000 SERIES
REPORT UPDATE

~EDP

-

AUER8AC~s

I£PIUS

REPORT UPDATE
~UNIVAC

INTRODUCES PAPER-TAPE SUBSYSTEM FOR 9000 SERIES

The UNIVAC Series 9000 acquired paper-tape input/output capabilities with the announcement
of the Type 0920-00 Paper Tape Subsystem on September 30, 1968. This will not only allow
users of UNIVAC's newest systems to enter new areas of application but it will also make transition to the 9000 easier for users of older, paper-tape oriented systems.
Consisting of the Paper Tape Control, Paper Tape Reader, Reader Spooler, Paper Tape Punch
and Punch Spooler, this subsystem is housed in a free standing cabinet connected via a multiplexor channel. The system is completely modular so that any combination, one of each, of the
components can be connected to the Paper Tape Control.
The 0920 reads at 300 cps and punches at 110 cps, handling any 5, 6, 7 or 8 level tape codes I
It can also read or punch binary tapes. The selection of codes is accomplished through the use
of a patch panel, called a program connection, mounted in the paper tape cabinet. UNIVAC
states that all conventional perforated tape with a light transmissivity of 40% or less can be
read by the tape reader.
IOCS routines will be provided for this subsystem.
First deliveries to customers are scheduled for the second quarter of 1969.
Pricing of the 0920-00 Paper Tape subsystems will be:
Purchase
Price

Monthly Rental
Short Term

Paper Tape Control
0920-00

$ 7,180

$190

Paper Tape Reader
F 1033-02

1,525

Reader Spooler
F 1034-00
Paper Tape Punch
F 1032-02
Punch Spooler
F 1035
Totals:

Monthly Rental
5 yr. lease

Monthly
Maintenance

$165

$25

50

45

15

1,525

40

35

5

5,220

140

120

20

655

20

18

5

$440

$383

$70

--$16,105

Note: The above rental prices include the monthly maintenance charges.

C 1968 AUERBACH Corporation and AUERBACH Info, Inc.

12/68

-1 . . . "
.
_-------.1

810:000.001

AEDP

AUERBAC~

UNIVAC 9000 SERIES
REPORT UPDATE

UPORTS

REPORT UPDATE
~ UNIVAC ANNOUNCES 16K COBOL FOR THE 9300
In January 1968, UNIVAC announced a reduction in the memory requirement for use of a COBOL
compiler with UNIVAC 9300 systems. The new 16K version of UNIVAC 9300 COBOL does not
replace the existing 32K version; both will be made available. The 16K COBOL differs from the
32K version in that it requires more compilation time and uses a different method of indicating
diagnostics. Instead of the message form of diagnostic output provided by the earlier compiler,
the new 16K program provides numeric keys that refer to a diagnostic catalog.
All COBOL language facilities provided in the 32K version are incorporated in the new and
smaller system, which therefore meets the USASI minimum requirements.
The minimum equipment configuration now required to compile a COBOL program includes a
9300 Proccssor with 16,384 bytes of plated-wire memory, a card reader, a printer, and four
tape units (9-track, or 7-track with the data conversion feature).
Customer availability of both the original 32K and the new 16K versions of 9300 COBOL is
scheduled lor September 1968.

~ UNIVAC ANNOUNCES DATA COMMUNICATIONS SUBSYSTEM-4
In February 1968, UNIVAC announced the addition of the DCS-4 to the existing communications
facilities of the 9000 Series. The DCS-4 provides all the capabilities of the DCS-l (see Report
Update of 11/67) plus multi-line communication between the 9000 Series central processor and up
to four remote systems. Each DCS-4 accommodates a maximum of four full-duplex or eight
simplex positions. A maximum of two DCS-4s can be incorporated into any 9200/9300 System.
The DCS-4 can control a wide range of speeds and services. Each Communication Subsystem
can service four input and/or four output lines, permitting communication via: (1) telegraph
lines at up to 150 bits per second; (2) public telephone networks at up to 2,400 bits per second;
or (3) leased broad-band lines at rates up to 230,400 bits per second (Telpac t C). Both synchronous and asynchronous line terminals are available, as well as communication interfaces
connecting line terminals with common carrier lines. Depending on DCS-4 line terminal choice,
a UNIVAC nOO/9300 can communicate with teletypes, remote terminals, CRT Displays, computer systems, and any device using switched network facilities.
A UNIVAC 9000 Series computer equipped with a DCS-4 can communicate with the following
UNiVAC systems: 418, 494, 1107, 1108, DCT 2000, 1004, 1005, and Uniscope 300.
A feature of the DCS-4 is that processor functions, including magnetic tape reading and writing,
can be overlapped with communication at voice-grade and telegraph rates.
Like the DCS-I, the DCS-4 can be connected to the optional I/O Multiplexer Channel of a 9200 or
9300. The DCS-4 is compatible with IBM's new Binary Synchronous Communications (BCS) data
transmission technique, as explained in the UNIVAC 9000 Series Report Update of 11/67. The
software commitment is essentially the same for the DCS-4 as with the DCS-1, with necessary
additions and modifications for greater flexibility. Line terminal character size can be from
4- to 10-level, depending on transmission speed.
Deliveries of the DCS-4 are scheduled to begin in the fall of 1968. The price of a DCS-4 configuration serving four telegraph lines is $522 per month or $19, 155 purchase; for a configuration
serving one wide-band line, one voice-grade line, and two telegraph lines, the rental is $566 per
month or $20,750 purchase.
tTrademark of A. T. & T. Co.

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

3/68

810:000.001
n ......

UNIVAC 9000 SERIES
REPORT UPDATE

EDP
11'lin

REPORT UPDATE
~ UNIVAC l:NVEILS THE 9olOO SYSTEM
January 15. 1968. marked the debut of the UNIVAC 9400. third member of the 9000 Series
family of compatible computers. Larger and more powerful than the 9200 and 9300 systems.
the 9400 features both real-time and multiprogramming capabilities. Continued use of
UNIVAC's fast plated-wire memory and monolithic integrated circuitry. blended with the 9400's
dual-byte memory access. 32 general-purpose registers. and strong communications capabilities. promises impressive processing power within the 9400's competitive price class.
Primary marketing efforts will be directed toward upgrading present users of smaller UNIVAC
9000 series and IBM System/360 Model 20 systems. In this respect. the UNIVAC 9400 offers
approxImately three to four times the internal processing speed of the IBM 360/30 at roughly
comparable system prices. Since the 9400 instruction set is a compatible subset of the IBM
360/30, present 360/20 programming staffs should experience little difficulty in converting to
a LI\IVAC 9olOO system.
In gaimng the 9400's cost/performance advantage. certain concessions had to be made:
•

The 9olOO implements only 68 of the 360/30's 142 instructions.

•

Floating-point. binary multiplication, and binary division hardware facilities
have not been implemented.

Plated-\llrememory, available incapacities of 24,576, 32.768.49.152,65,536.98,304. and
131, 07 2 b~·tl'S, has a cycle time of 600 nanoseconds per 2-byte access. Processor operating
efficlencu's are further enhanced by the use of two sets of general registers - 16 for user programs anrl16 that service the operating system. In addition, the UNIVAC 9400 provides separate
control rl'glsters for each Selector Channel and Multiplexor Channel, allowing direct access to every
channel wlll'n servicing I/O requests. An interval timer is a standard feature; storage protection is
available as an option.
The standa I'd MUltiplexor Channel has a gross data capacity of 85,000 bytes per second. One or
two optIOnal Sl'lector Channels, each having a 333KB/second capacity. can be added. Peripherals
offered InC IudI':
•

Thn'(' series of magnetic tape units. with transfer rates ranging from 34KB to

1!l2KB.
types of high-speed line printers - llOO or 1600 alphanumeric lines per
Illillute.

•

'1'110

•

Three communications subsystems that can accommodate up to 64 full-duplex
lines.

•

l'p to 58 million bytes of on-line disc storage. with a data transfer rate of
15GKB/second.

•

A GOO-cpm card reader.

•

A 250-cpm card punch.

A UNIVAC 1004 or 1005 Processor can be used either on- or off-line with the 9400 system, in
addition to, or in place of, the standard printer. reader. or punch.
Software will include multiprogramming provisions for five main programs. a Data Management
System, Job Control Stream operation, a Basic Assembler, a 32K COBOL-65 compiler. a
Report Program Generator, a 32K USASI FORTRAN compiler. a Systems Service Library. and
a nility Library.
Monthly rental for a basic 9400 system including a 24K memory. 4 magnetic tape units. 1100Ipm printer. 600-cpm card reader, and 250-cpm card punch is $5,880. Customer deliveries
are scheduled to begin during the second quarter of 1969. Basic system software will be delivered concurrently; advanced software will follow at a later date.

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

1/68

UNIVAC 9000 SERIES

810;000. 002

. . UNIVAC ANNOUNCES THE 8410 DISC STORAGE SYSTEM
The 8410 Disc Storage System, announced on December 6, 1967, represents a considerable
enhancement to the UNIVAC 9000 Series product line by providing random or sequential access
to a moderate amount of on-line storage. The recording medium is a nickel-cobalt-coated disc
in an interchangeable cartridge. Each single-disc cartridge can store up to 3.2 million bytes,
but only 1. 6 million bytes of each cartridge can be accessed "on-line". By physically removing
each disc cartridge, turning it over, and replacing it on the drive, the remaining 1. 6 million
bytes of data become accessible.
Two independently operating disc handlers are housed in a single cabinet, and each handler services a single disc cartridge. Up to four dual-disc units are allowed in a maximum 9000 Series
system configuration, providing a total on-line data capacity of 12.8 million bytes.
A seek function on one disc handler can be overlapped with seek, read, or write functions on
another. Once the desired sector address has been determined, the average time required to
position the head mechanism for randomly placed data is 110 milliseconds. The rotational
speed is 1,200 revolutions per minute, which corresponds to 50 milliseconds per revolution
(latency time). Thus, the average access time for randomly placed data, including 25millisecond average'latency time, is 135 milliseconds, an acceptable access time for use in
many random processing applications.
The fastband search technique, which may be employed at the user's discretion, provides the
address of the desired sector through use of a key (sector location data) contained in a special
50-sector track or "fastband" on each disc surface.
Although the 8410 Disc Storage System is capable of transferring data at the rate of 136,000
bytes per second, the effective speed, as limited by the data transfer capacity of the Multiplexor
channel, is 100,000 bytes per second. Buffering logic located in the disc drive cabinet provides
intermediate storage of up to 256 bytes. A maximum of 160 bytes can be transferred to or from
the disc unit with a single disc read/write command.
Accuracy control during read/write operations includes a longitudinal parity check by sector.
Control instructions include Write Check, Seek Track. and Sector Search commands.
Deliveries of the 8410 Disc Storage System will begin in approximately 12 months. The 8410
system's software support, scheduled for completion during the third quarter of 1968, will
include a disc-oriented Input-Output Control System (IOCS), a Disc Report Program Generator,
Library Services, and a Disc Sort Program. A completely tape-independent Disc Assembler
will also be provided.
UNIVAC states that disc-oriented FORTRAN and COBOL compilers are not currently planned.
PRICES
PRICES

IDE:-ITITY OF ['NIT
SYSTE~!

Monthly
Rental (1)

Lease

Lease

Monthly
Maintenance
$

295
295
180
195

255
255
ISS
170

125
125
80
35

12,835
12,835
7,830
8,485

1')5

170

35

8,485

lJ5

100

45

5.005

25

6.745
380

$

S'ame

No.

I-Year 5-Year
8410
DISC
STORAGE
SYSTDI

F 1016-00

Dual Disc File _ Master (2)
Dual Disc rile - Shlve
Single Disc File - Slave
Disc Dn\'c Control (3)
(for use \\ ithout 1001 Control)
Disc Drl"e Control(3)
(for use ,dth 1001 Control)
Disc D!1\'c

F 101;;"00
r 1102-00

Optional Features
Buffer and rastbond Search (2)
Cartridge

8410-00
8410-92
8410-02
F 1023-00

F 1023-01

155

-

(1) )'lonthlv rental does not Include maintenance cost.
(2) Hequlred for minimumconfiguration.
(3) Choice of one is required for mimmum configuration.

CONFIGURATION PRICES
~umber

of
Drives

On-Line
capacity.
millions of
bytes
3.2
4.8
6.4
8.0
9.6
11.2
12.8

PRICES'

1-Year 3-Ycar
$
S
9:10
1.0nO
1 250
1 310
1. 670
1,930
2,090

743
n80
1 12.1
1. 360
1 503
1.740
1. 885

PurcShase
28.065
35.8n3
40 nOD
48.730
53.733
61. 565
66.570

• Hental prices include maintenance charges. Fastband
~Nl.rch option. and requh'cd Disc File Control.

1/68

A

AUERBACH
."

135

-

-

Purchase
$

810:000.001

.A.. ..."...

~\\EDP

AUERBACH

UNIVAC 9000 SERIES
REPORT UPDATE

RlPOITS

~

REPORT UPDATE
• UNIVAC ANNOUNCES DATA COMMUNICATION SUBSYSTEM I
In October 1967, UNIVAC announced the addition of a modest data communications facility to its
9000 Series product line, thereby considerably broadening the applications areas for these lowpriced computer systems.
One or two of the new DCS I systems can be connected to the optional I/O Multiplexor Channel of
a 9200 or 9300 Processor, with each DCS I subsystem occupying one multiplexor subchannel.
Each DCS I has a single-line capability, can service one input and/or one output line, and permits communications via: (1) the public telephone networks at speeds up to 2,000 bps; (2) leased
voice-band lines at rates up to 2,400 bps; or (3) leased broad-band lines at rates up to 50,000 bps.
DCS I is physically housed within the existing computer cabinetry and is available with a wide
selection of options to provide compatibility with most UNIVAC equipment, including the 418, 494,
1004, 1107, and 1108 computer systems, the DCT 2000 Data Communications Terminal, and other
9000 Series computers. In addition, DCS I is compatible with IBM's new Binary Synchronous
Communications (BSC) data transmission technique, allowing communication with an IBM System/
360 computer that is equipped with a Model 2701 Data Adapter Unit or 2703 Transmission Control
Unit and the appropriate transmission adapter. Communication with a System/360, however, is
restricted to the non-transparent USASCII mode.
Each DCS I contains a communications interface that connects to a single line for synchronous
data communication at rates of up to 2000 or 2400 bps using voice-band lines, or up to 50,000
bps using broad-band lines. Message length can be varied from a single character up to the
available memory capacity.
A choice of 5-, 6-, 7 -, or 8-level (Plus parity) transmission codes is available, including userselected synchronizing, idling, start-of-message, and end-of-message characters. Odd or even
parity is generated for transmitted characters and checked for received characters.
Incoming calls on the public telephone network can be automatically answered under program
control.
CONFIGURATION
A DCS I Data Communications Subsystem includes a Line Terminal Controller which connects
to the I/O Multiplexor Channel in the processor. The Controller coordinates the performance
of its assoc iated Line Terminals, which may be an input terminal, an output terminal, or both.
Line Terminals, available in either Synchronous or Interprocessor models, perform parallelto-serial and seriaI-to-parallel data conversions and provide control-character insertion into
outgoing messages and corresponding control-character checks on incoming messages.
Line Terminals are connected, via a Communication Interface Unit, to an appropriate data
set as follows:
Dell System
Data Set
201A
20m
201B/303C

Facility
Public telephone
network
Leased voice-band
Leased broad-band

Speed, bps

UNIVAC
Communication
Interface

2,000

F 1002-04

2,400
50,000

F 1002-03
F 1002-05

SOFTWARE
UNIVAC is currently developing software support for the DCS I, including a Communication
Control Routine (CCR) supplement to the 9200/9300 Supervisor and several support packages
for specific communication with 1107, 1108, 494, or DCT 2000 systems. Software to support
comm,unication with systems other than those specified above must be supplied by the user
to handle interrupt proceSSing, issuance of I/o requests, and the performance of clocking
functions.
In addition to the CCR, UNIVAC is also extending its IOCS routines to include communicationsoriented subroutines.
Initial deliveries of the DCS I are expected in June 1968. Initial software support will be
delivered concurrently.
© 1967 AUERBACH Corporation and AUERBACH Info. Inc.

11/67

810:000.002

UNIVAC 9000 SERIES

PRICES

Type or
Feature No.

11/67

Product Name

Monthly Rental
(I-Year Rental
Agreement;
Excludes
Maintenance)

Purchase
Price

Monthly
Maint.

$4,350

$15

$100

565
565
870

2
2
5

13
13
20

F1000-00

Line Terminal Controller

F1002-03
F1002-04
F1002-05

Communications Interface Private Line
Data Phone
Broad-band

F1005-02
F1005-03

Line Terminals, Synchronous OUtput
Input

1,130
1,130

5
5

26
26

F1005-04
F1005-05
F1008-99

Line Terminals, Interprocessor OUtput
Input
mock Parity Feature

1,130
1,305
740

5
5
3

26
30
17

A•

AUERBACH

810;011. 100

£ "......

~EDP

-

AUER'AC~

UNIVAC 9000 SeRIES
SUMMARY

IU •• "

~

SUMMARY
.1

GENERAL
The 9200, announced on June 21. 1966, is a small-scale, card-oriented system that offers the
new user an easy transition from tabulating machines to computers. It offers the advantages of
familiarity of recording medium with the chance to use third-generation technology. all at a very
attractive price. A user having a Univac 1004 installation, can see in the 9200, a painless way
to upgrade his hardware.
The user of a small IBM 360 who is concerned with the complexities of the instruction repertoire
of that machine. would be attracted by the simplified set of 35 instructions of the 9200 (multiply,
divide, and edit are optional) that promise to do everything that he would need to do, with fewer
stumbling blocks and in a manner that would keep conversion pains to a minimum.
On the negative side of the ledger are the very limited and relatively slow i/o devices announced
with the 9200. A 250-lpm printer, a 400-cpm reader, and a 75-200 cpm punch do not seem to be
capable of using the 1. 2-microsecond cycle-time. plated-wire memory to its full capacity.

More recent expallB ions of the 9200 have added several other devices to increase 9200 capability. These include the 8410 Disc Storage Subsystem, the Paper Tape Subsystem, Data
Communications Subsystems capable of handling up to eight lines, an on-line 1001 Card
Controller and the ability to use an on-line UNIVAC 1004.
The larger 9300, also announced on June 21, 1966, is faster, larger and more versatile. The
UNISERVO VIC tape drive announced as available with this system made the 9300 attractive to
the users of small, tape-oriented, second-generation systems; e. g., the IBM 1401 and the RCA
301. As with the 9200, a strong effort was made to have the 9300 system appeal to the IBM 360
user by offering him simplified coding by using 35 instructions and data formats compatible with
the 360. The peripheral equipment originally announced, however, was relatively slow in view
of the potential shown by the 600-nanosecond per byte plated-wire memory.
Devices such as a 600-lpm printer, 600 cpm reader, 200 cpm carc\ punch, and the 34 KBS UNISERVO VIC tape unit indicate throughput rates closer to the second generation level than to the
third. In all fairness. most other manufacturers also do not provide high-speed magnetic tape
units for their small-scale business-oriented computers.
Significant expansion in the capabilities of the 9300 System have been achieved by the addition of several new devices and systems. These include Data Communications Subsystems
capable of handling up to eight lines, a Paper Tape Subsystem, the 8410 Disc Storage Subsystem, an on-11ne 1001 Card Controller, and the ability to use a UNIVAC 1004 on-line.
The prices that were announced for the 9200 and 9300 are among its chief competitive advantages.
A basic 9200 system, containing 8192 bytes of plated-wire memory, printer, card reader and
card punch can be rented for $1085 per month or purchased for approximately $40, ODD. The 9300
system rentals range from about $1740 to $9300 per month, and purchase prices range from
about $63,000 to $350.000. A tape-oriented system with sort/merge capability (three tape units)
can be rented for about $3000 per month.
The 9400 was announced on January 15, 1968. Larger and more powerful than its predecessors,
it brings to the 9000 series true medium-scale power together with realtime and multiprogramming capabilities. The use of the plated-wire memory and monolithic circuitry combined with
the 9400's dual-byte memory access, 32 general purpose registers, and strong communications
capabilities provide processing power that is quite impressive in comparison with other computers in its price class.
Throughput rates on the 9400 are more in keeping with internal processing power. The standard
Multiplexor Channel, optional on the 9200 and 9300, has a gross data capacity of 85,000 bytes
per second. One or two optional Selector channels, each having a 333-kilobyte per second capacity, can be added.
A Univac 1004 or 1005 Processor can be used, if available, either on line or off line with the
9400 system, in addition to, or in place of, the standard printer, reader, and punch. Monthly
rental for a basic 9400 system including a 24K memory, four magnetic tape units, 1l00-lpm
printer, 600-cpm reader, and 250-cpm card punch is $5880.
The latest announcement in the 9000 series is the 9200 II and 9300 II models. These additions are
basically the same processors as the 9200 and 9300, but offer substantial increases in capability
and versatility over the earlier models. This has been done by extending the line of available
peripherals to include some of the devices previously available only on the 9400.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

6/69

810:011. 101

.1

UNIVAC 9000 SERIES

GENERAL (Contd.)
The 9200 II, for example, can now be equipped with the UNISERVO VIC tape unit. Paper tape devices can also be attached, and both high-speed printers from the 9400 can be used. The Univac
8411 Direct Access Subsystem, which is IBM 2311 compatible, extends the capabU1ttes of the
lower end of the 9000 series to random-access processing. A 600-cpm card reader and 200-cpm
and 250-cpm card punches also aid in increasing the throughput rate.
Maximum memory capacity of the 9200 II is 32,768 bytes compared to 16,384 in the 9200. The
Multiplexor Channel, optional on the 9200, is standard eqUipment on the 9200 II. A 350-kilobyteper-second Selector Channel, capable of controlling up to eight subsystems, is available as an
option. All features currently available for the 9200 are available for the 9200 II.
Price increases over the 9200 have been modest. A basic 9200 II magnetic tape system with
24K of memory, 600-cpm card reader, 200-cpm card punch, 1100/900-lpm printer in addition
to the integrated 250-lpm printer, and three UNISERVO VIC magnetic tape units, will rent for
$4565 per month, including maintenance, on a one-year rental. A Selector Channel can be added
for an additional $100 per month. The purchase price of the system would be $159,935, with
monthly maintenance charges of $820; purchase o( the Selector Channel would increase the cost
by $4320 and the maintenance charges by $10 per month.
Changes in equipment for the 9300 II are not as extensive. The Multiplexor Channel, optional
on the 9300, is standard equipment on the 9300 II. The Selector Channel is standard equipment
on the 9300 II. This is not available on the 9300. Both the 1100/900-lpm and the 1600/1200-lpm
high-speed printers are available in addition to the 600-lpm integrated printer. As on the 9200
II, the 8411 Direct Access Subsystem can be used to extend the capabilities of the system to
random-access processing. All items available on the 9300 are available for the 9300 II.
A typical 9300 II magnetic tape system with 32K of memory, 600-cpm card reader, 200-cpm
card punch, 1100/900-lpm printer, and three UNISERVO VIC magnetic tape units, will rent for
$5775 per month, including maintenance, on a one-year rental. The purchase price would be
$208,205, with monthly maintenance charges of $1015.
All software of 9200/9300 Card, Tape, Disk and Communications Systems is available for the
9200 II and 9300 II. In addition, several new packages are being developed to provide effective
use of the newly added 0768 Printer and 8411 Disk. The 0768 will be supported as a data printer
only by an 10CS incorporated in the output of the RPG, the COBOL, and FORTRAN compilers, and
the Assembler. The Buffered Printing Module of the Concurrent Operating System will be modified so that report tapes can be optionally produced in the 0768 format. These tapes can then be
processed by an 0768 version of the Tape-to-Print Symbiont.
Supervisors of 8411 (Direct Access Storage Subsystem) oriented configurations will operate in
three environments. For machines with 16K and two 8411's, the Minimum Operating System or
the Non-current Operating System will be provided. A user with 24K or 32K or storage will be
able to run concurrent programs by using the Concurrent Operating System.
The 8411 IOCS will provide for three levels of file processing: sequential, indexed sequential,
and random or direct access. These will be incorporated in the new, 8411-oriented Assembler
Linker. RPG will be able to process Sequential or Indexed Sequential Files. COBOL will be
implemented in a disk version that will conform to the United States of America Standards Institute minimum requirements. All existing card and tape language processors will be upgraded
to optionally include Sequential and Indexed Sequential 8411 Processors. COBOL and FORTRAN
will also include the Random Processor.
Deliveries of the Univac 9200 II and 9300 II Tape Systems are scheduled to begin in August 1969.
Disk system deliveries are planned for March of 1970 •

.2

6/69

DATA STRUCTURE
The UNIVAC 9200 and 9300 provide facilities for convenient handling of variable-length- fields
composed of either 8-bit bytes or 4-bit digits. The minimum addressable unit in memory,
however, is one byte. Operand lengths can range from 1 to 256 bytes or digits, but in many
instructions are limited to 16 bytes in length.
Data represented in the 4-bit format can be either signed (with a 4-bit sign digit preceding
the least Significant numeric digit of the field) or unSigned. Data in the 8-bit format is always unsigned. and is treated (as in the IBM System/360) as a byte conSisting of eight data
bits and a parity bit. The eight data bits in a byte can represent one alphanumeric character. two packed decimal digits. or a portion of a binary field. Like the System/360, digits must be placed in packed decimal format prior to performing decimal arithmetic operations.
Bytes can be handled individually or grouped together into fields. A "halfword" is a group
of two consecutive bytes, or 16 bits. Binary numbers in the UNIVAC 9200 and 9300 are represented by signed halfwords (sign plus 15 data bits). No floating-point arithmetic operations
can be performed by hardware. Fixed point decimal operands can be up to 16 bytes (31 digits and Sign) in length. Machine instructions are either four or six bytes (38 or 42 bits) in
length.

A

(Contd. )

AUERBACH

'"

SUMMARY

810:011.200

DATA STRUCTURE (Contd.)
As in the System/360. data can be represented internally in either of two codes. EBCDIC or
ASCII. depending upon the setting of a program-controlled mode flip-flop In the processor.
Also, Translate instruction uses a table in main memory to accomplish efficient translations
between any 8-bit codes,

.3
.31

A hardware translator provides automatic translations between external codes and a
special UNIVAC internal compressed code which must be program-translated into EBCDIC
or ASCII before the data can be processed .
HARDWARE
?ystem Config\ll'ation
The basic UNIVAC 9200 system includes a 9200 Processor with a built-in 250-lpm printer;
8K-, 12K -, or 1GK -bytes of plated-wire memory; a 400-cpm card reader; and a column card
punch rated at 7:1 to 200 cpm. Only one of each of the basic I/O devices (1. e. , one card
reader, one printer, etc.) can be connected. Optional features available for the processor
and I/O devices are described in the appropriate sections of this Introduction.
The optional Multiplexor I/O channel permits connection of a large variety of devices to the
central processor. These include the following: 1001 Card Controller, 8410 Disc Storage
Subsystem, Paper Tape Subsystem, on-line UNIVAC 1004, and two Data Communications
Subsystems handling a maximum of eight lines. The Multiplexor can accommodate up to
eight control units. The 1001 is connected to one control-unit position via a 1001 Control.
The 9200 II differs from the 9200 principally in that the Multiplexor Channel above is standard equipment. UNISERVO VIC tape units are available for attachment to this channel.
Either of two card readers, the 0711-00 (400 cpm) or the 0711-02 (600 cpm) and two card
punches, the OG03-04 (75-200 cpm) or the 0604-00 (200 cpm) , can be attached directly to
the processor 11;; lng built-in control units. Also available are the 0768-00 and the 0768-99
high-speed printers. A 350 KB Selector channel, optionally available, permits the attachment of an 8411 Disc Storage Subsystem. Additionally, all peripheral equipment available
with the 9200 arc available with the 9200 II as well.
The basic UNIVAC 9300 system includes a 9300 Processor with a built-in 600-lpm printer;
8K-. 12K-, 16K-, or 32K-bytes of plated-wire memory; a 600-cpm card reader, and a
column card punch rated at 75 to 200 cpm. Only one of each of the basic I/O devices can
be connected.
The optional Multiplexor I/O Channel permits connection of a 100 Card Controller, a 200cpm Row Punch, an 8410 Disk Storage System, Uniservo VIC magnetic tape units, and a
data communications system. The Multiplexor Channel can accommodate up to eight control units. The 1001 Card Controller can be connected to one control-unit position of the
Multiplexor via a 1001 Control. Each Uniservo VIC Control unit occupies one control-unit
position and can control up to eight tape units. See Paragraph. 355 for Uniservo configuration details. The 0768-00 and 0768-99 high-speed printers are also available for attachment via the Multiplexor Channel.
The 9300 II includes the Multiplexor channel as well as the 35 KB Selector channel as standard equipment. These permit the attachment of all peripheral devices available to the
9200 II as well as those offered with the 9300.
The basic UNIVAC 9400 system includes a 9400 Processor with a built-in console and keyboard/printer; 24,576, 32,768, 49,152, 65,536, 98,304, or 131,072 bytes of plated-wire
memory; a standard 85KB Multiplexor Channel; a standard Interval Timer; and peripheral
devices as required.
One or two 33:JKB Selector Channels are optionally available for the 9400 system; each
channel can handle up to eight control ,units, on a one-device-at-a-time basis. Each Selector Channel can operate independently of normal processing operations, the other
Selector Channel, and/or the Multiplexor Channel.
The standard Multiplexor Channel has eight shared subchannels, seven of which can be
connected to peripheral device controllers. (One subchannel services the console Keyboard/Printer.) The Multiplexor Channel can transfer data between devices on all eight
subchannels simultaneously as long as the total data rate does not exceed 85,000 bytes
per second. 128 non-shared subchannels can be added with a communications adapter
option, which must be used in order to operate communication subsystems on the Multiplexor Channel. In general, the Multiplexor Channel is used to handle relatively lowspeed I/O devices, including printers. card readers, card punches, communications
terminals, display devices, and the low-speed Uniservo VIC Magnetic Tape Handlers.
The non-shared multiplexor subchannels permit data chaining operations.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

6/69

810:011.320
.32

UNIVAC

9000

Central Processors
The overall architecture of the UNIVAC 9200 and 9300 Processors is similar to that of the
IBM System/3BO processors. The UNIVAC 9200 and 9300 Processors are functionally identical to each other; they differ only in internal speeds (the 9300 is twice as fast) I in the
complement of I/O devices that can be connected. and in the fact that the Multiply, Divide,
and Edit facilities are extra-cost options in the 9200 and standard features in the 9300. Table I summarizes the basic characteristics and capabilities of the two systems.

TABLE I: CHARACTERISTICS OF THE UNIVAC 9200 AND 9300
SYSTEM

9200

9300

Memory cycle time, microseconds

1.2

0.6

Bytes accessed per cycle

1

1

Memory capacity, bytes

8K. 12K. or 16K

8K. 12K, 16K. or
32K

General registers
I/O Clontrol registers
Multiply, Divide, Edit Instructions
Processor speeds, microseconds
(signed 5-dlglt operands) c=a+b
b=a+b
c=axb
c = alb
Move a to b
Compare a to b

8

8

8
Optional

8

187.2
103.2
2,020.0
4,420.0
84.0
103.2

Multiplexor I/O Channel rate, bytes/
second

85,000

Card reading speed, cpm Basic reader
1001 Card Controller

400
1000/2000

Card punching speed, cpm

75-200

Alphanumeric printing speed, lpm

250

Magnetic tape speed, bytes/second

Not available

Standard

93.6
51. 6
1,010.0
2,210.0
42.0
51. 6
85,000
600
1000/2000
75-200 or 200
600
34,160

The architecture of the UNIVAC 9400 Processor is similar to that of the smaller UNIVAC
9200 and 9300 Processors. The effective basic memory cycle rate is double that of the
9300 and four times that of the 9200. The first 512 bytes of main storage are reserved for
use by the software and hardware for the handling of interrupts, indexing, I/O, and other
control functions. Two sets of 32-bit general-purpose registers are also implemented in
this low-order block memory. Features of the 9400 which are not found in the two smaller
UNIVAC 9000 Series models include:
•

Write storage protection (optional).

•

Two sets of 16 general registers.

•

Separate interrupt control registers for each I/O channel.

•

An interval timer.

•

Seven levels of processor interrupts.

•

Direct access to interrupt condition data.

•

Tabling of hardware status for communications interrupts.

Table n summari!/les the basic characteristics of the UNIVAC 9400 Processor, including
typical instruction execution times.
.

6/69

A

(Contd. )

AUERBACH

'"

SUMMARY

810:011.321
TABLE

n.

BASIC CHARACTERISTICS OF THE UNIVAC 9400

Memory cycle time, microsecond.
Bytes accessed per cycle
Memory capacity, bytes

0.6
2
24K. 32K. 49K. 65K. 98K. or 131K

General regleters, control
General regletere. user
I/O control registers

16
16
4

Processor specdH. microseconds
(signed &-dlglt operands):
c=a+b
b=a+b

46.8
23.4
125.3

c~axb

c=a/b
Move a to h
Compare a to b

280.8
23.4
23.4

Multiplexor I/O Chpnnei rate. bytes/second
Selector I/O Channel rate. bytes/second
Card reading speed, cpm
Card punching speeds, cpm
Alphanumeric printing speeds. ipm
Magnetic tape speeds, bytes/second
Disc storage peak transfer rate, bytes/second

85.000
333,000
600
250
1100 or 1600

8540 to 192. 000

156.000

.321 Instruction Formats
9200 - 9300
There are three types of instructions: I/O instructions, privileged or state control instructions. Both I/O and privileged instructions use the SI (Storage and Immediate Operand) instruction format, as found also in the IBM System/3S0. Processor instructions use either
the SI, SS (Storage to Storage), or RX (Register to Indexed Storage) formats. which are
identical to their System/3S0 counterparts. The RR (Register to Register) instruction format, however, is not implemented in either the 9200 or 9300 processors. Instruction types
RX and SI are four bytes in length, while type SS instructions are six bytes long.
The first two 4-bit digits of a processor instruction designate the operation code, which is
stated in hexadecimal notation. The remaining digits vary in their intent and purpose. depending upon the format type, as illustrated.
Operands in plated-wire memory can be addressed either directly or by means of the baseplus-displacement technique used in the System/3S0. If there is a 0 in the most significant
bit position of an instruction halfword containing a memory address, the remaining 15 bits
of the halfword a re interpreted as a direct address. If the most Significant bit is aI, the
"base address" contained in the general register specified by the next three bits is added to
the "displacement" contained In the last 12 bits of the halfword to form the required memory
address.

!!!QQ
The six types of UNIVAC 9400 instructions are two. four, or six bytes in length and are
structured identically to their System/3S0 cOWlterparts.
A two-byte instruction causes no reference to main storage, while a six-byte instruction
causes two storage references. The six basic instruction formats are:
Type RR - Register to Register (2 bytes)
Rl

Op

I I
R2

Type RX - Register to Indexed Storage (4 bytes)

I

Op

Rl I

I

x2 1

B2

I

D2

I

Type RS - Register to Storage (4 bytes)

Type SI - Storage and Immediate Operand (4 bytes)

I

Op

I

12

I I
B1

Dl

I

Type SS1 - Storage to Storage (6 bytes)

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

6/69

UNIVAC

810:011.322

.:J:'! 1

9000

SERIES

IlIsta'uction Formats (Contd)
Type SS2 - storage to Storage (6 bytes)

where

B
4-bit base register specification
D = 12-bit displacement
I
8-bit literal operand
L
8-bit operand length specification
Op
8- bit operation code
R
4-b:lt operand register specification
X
4-bit index register spec1f:icatlon

The first two 4-bit digits of a processor instruction desIgnate the operation code, which is
stated in hexadecimal notation. The remaining digits vary in their Intent and purpose,
depending upon the format type. as illustrated.
Operands in plated-wire memory are addressed by means of the base-plus-displacement
technique used in the System/360; 1. e. , the "base address" contained in the general
register specified by the instruction is added to the "displacement" contained in the last
12 bits of the instruction halfword to form the required memory address.
:J»

Processing Facilities
The UNIVAC 9200/9300 instruction repertoire emphasizes decimal arithmetic operations
upon variable-length fields. although it does provide facilities for adding and subtracting
fixed-length binary fields. Efficient data movement. comparison. and editing facilities are
provided. as are a number of logical commands to allow masking operations. Logical AND
and OR commands can be performed on both decimal and binary data.
Fixed point decimal multiplication (15 by 15 digits maximum) and division instructions are
standard with the 9300 and optional with the 9200 j in like manner. the Edit instruction is
standard with the 9300 and optional with the 9200. No floating-point hardware is available
with cither machine. A Translate instruction effects translations to or from any 8-bit code
through the use of a lookup table in the plated-wire memory. The Edit instruction unpacks
up to 31 digits plus sign from a field, either inserting "fill" characters where specified or
zero-suppressing the edited data. Zero suppression and character insertion operations can
be made in a single editing operation through the use of three different indicator bytes.
The UNIVAC 9200/9300 instruction repetoire is similar to, though not identical with. that of
IBM's small-scale System/3BO Model 20, and is far smaller and less comprehensive than the
instruction set used in the IBM 360/30. Nonetheless, the limited set of 35 instructions chosen
for implementation by UNIVAC should be entirely adequate for small scale business applications - and the smaller number of instructions should make programming easier and less
prone to error.
The philosophy adopted for the UNIVAC 9400 system has been: (1) to implement an instruction set that provides representation in each of the major classes, (2) to assure upward
compatibility from the 9200/9300 systems, and (3) to enable efficient programming - but
to achieve economies through exclusion of non-essential instructions. Many of the 142
instructions implemented in the IBM System/SSO Model 30, for example, were found to
be of limited value and rarely used; these instructions have been omitted from the repertoire of the 9400. Far from useless, but nevertheless also omitted, were all arithmetic
floating-point hardware features. UNIVAC feels, however, that since the primary market
for the 9400 is in the business data processing environment, the lack of floating-point
hardware is not 1I serious drawback.
A total of 68 instructions have been implemented in the 9400 Processor, including powerful
provisions for handling decimal arithmetic operations upon variable-length fields and for
adding and subtracting fixed-length binary fields. Binary multiplication and division hardware facilities, as in the 9200/9300, have not been implemented. Efficient data movement,
comparison, and editing facilities are provided, as are a number of logical commands to
allow masking operations. Logical AND and OR commands can be performed on both decimal and binary data.
A Translate instruction effects translations to or from any a-bit code through the use of a
lookup table in plated-wire memory. The Edit inst.ruction unpacks up to 511 digits plus
sign from a field, either inserting "fill" characters where specified or zero-suppressing
the edited data, Zero suppression and character insertion operations can be made in a
single editing instruction through the use of different indicator bytes.

fA

.,

AUERBACH

(Contd. )

SUMMARY

810:011. 323

,323 0pl'I'aUonal Stah's
The cl'ntl'al processor alway!! operates In one of two states: the Program State, in which user
progrnms m'l' l'Xt'cutcn and the I/o or Control State, during which control is given to the Bupel'visor progrum, Pl'ivilcgcd instructions (Load State, Store State, and Supervisor Request
Cull) control the pl'Ogram state by setting and clearing control registers and flip-flops. This
dual state processing capability is unique in small scale systems and is not prOVided by the
lBl'Il 360/:W, fOl' l'xamp!e,
An interrupt syRtelll e:lUSt'S the processor to enter the Control State and branch to the Supervisor whenever allY of the following conditions occurs: completion of an I/O operation, memory pat'ity e1T01', ilIt'lllory address erl'Or, execution of a Supervisor Request Call, and invalid
inst ruction.
Each of the two proc('ssing states, Pl'Ogram and 1/0, are assigned eight registers, which are
used solely within tht'! l' respective states; one group of eight is used for internal processing
functions, whil(' the other group is reserved for input-output control functions. The proceSSing
group is used Wht'Ilt'VCr the processor is operating in the normal program mode, called Processor Program Statl' Control mode (PPSC). Whenever an I/O interrupt occurs, the processor
switches automat ll'ally to the I/O Program State Control mode (I/OPSC) and uses the inputoutput group of 1'(·gI8ters. This system improves processing efficiency by eliminating the need
to store and then reload the contents of the general registers whenever an I/O interrupt occurs.
Conversely, programming flexibility will be somewhat restricted by the fact that only 8 of the
16 registers arp available for general use by the programmer.
Program interrupts occur upon completion of input-output operations and upon detection of
input-output or Pl'Cll'('SSOr errors, as described in the preceding paragraph. A status byte
is made available to the supervisor which examines this byte, determines the cause of the
interrupt, and init lat l'S the appropriate program action.
There are two functional processor states within the 9400: Supervisor and Standard. The
Supervisor state is used primarily by the software operating system. In this state, all
instructions (Supervisor and Standard) are valid and can be executed. The 16 privileged
general registers are selected and low-order storage can be addressed. Instructions
in the Supervisor set cannot be executed in a problem program.
The Standard (nonprlviJeged) state handles problem programs. In this state, the Supervisor
(privileged) instructions are invalid, the problem set of general registers is selected, and
low-order storage cannot be selected.
The dual-state general-register complement contributes substantially to the efficiency of
the operating systems, since it eliminates the requirement for storing the contents of the
user registers prior to performing any supervisor functions - as is the case in the IBM
360/30, for example .
. 324 Interrupt System
The interrupt hardware in the UNIVAC 9400 Processor allows the processing unit to respond
to a variety of service-demanding conditions that arise within the processor and/or the
peripheral units. A key element in the interrupt system is the "Program Status Word"
(PSW). a double (64-bit) word that indicates the operational status of the processor. When
an interrupt occurs, it causes the current PSW to be transferred automatically to old PSW
storage for that class of interrupt and the corresponding new PSW to be loaded from storage.
The detailed status of the processor is thus saved for subsequent examination, and the new
state of the processor is established. The 64-bit PSW holds enough information about a
running program to enable program interrupt and restart without risk of data or sequence
loss.
The 9400 interrupt hardware can handle seven possible interrupt conditions, which include
the following (listed on a decreasing priorIty basis):
Supervisor call
Program exception
Interval timer
Selector channel 1
Selector channel 2
Multiplexor channel (shared subchannels)
Multiplexor channel (non-shared subchannels)
Each interrupt level has its unique interrupt status register, enabling the operating system
to recognize the source of the interrupt immediately, without recourse to test routines.
Both selector channels and the multiplexor channel (in both shared and non-shared modes)
have separate interrupt status registers; each shared and non-shared multiplexor channel
has indirect access to its own status register. When an I/o interrupt occurs, its source
can be directly detected by the UNlVAC 9400 Supervisor,

(el

1969 AUERBACH Corporation and AUERBACH Info. Inc.

6/69

810:011.330

.33

UNIVAC 9000 SERIES

INTERNAL STORAGE

.331 Plated-Wire Memory
Probably the most significant technical innovation in the UNIVAC 9000 Series is the USe of
plated-wire memory for the main working storage. The plated-wire memory operates in a
nondestructive readout (NDRO) mode, eliminating the need for the regenerative cycle which
is required after every read operation in conventional magnetic core memories. Furthermore,
most of the plated-wire manufacturing and testing operations can be carried out in continuous,
automated processes. For these reasons, UNIVAC claims that its plated-wire memories can
bl' offered with higher speeds and at lower costs than the core memories which are used in
nearly all current computer systems.
The plated-wire memory is a magnetic storage device of the thin-film type. The substrate is
a beryllium copper wire, 0.005 inch in diameter. The manufal'turlllg process consists of electroplating an iron nickel alloy over an initial plating of copper. Plating is performed while
the wire is in the presence ot a circumferential magnetic field that is created by the passage
of current through the wire itself. The wire that provides a base for the thin-film material
also becomes an integral part of the read/write circuitry: it serves as the sense line during
read operations and carries write current during write operations.
The UNIVAC 9200 has a memory cycle time of 1. 2 microseconds per one-byte access, and
memory capacities of 8,192, 12,288, and 16,384 bytes are available. The UNIVAC 9300's
memory cycle time is 600 nanoseconds (0.6 microsecond) per one-byte access, and the available capacities are 8,192, 12,288, 16,384. and 32,768 bytes. Memory sizes can be increased
at any time by field-installing additional modules. Every byte read from memory is checked
for proper (odd) parity.
The 9400 main storage has a cycle time of 600 nanoseconds per half-word (two bytes), producing an effective memory cycle of 300 nanoseconds per byte accessed. Memory is available in capacities of 24,576, 32,768, 49,152, 65,536, 98,304, and 131,072 bytes, and can
be expanded in the field by adding modules. Every byte read from memory is checked for
proper (odd) parity.
The low-order 512 bytes of storage are reserved for use by the hardware and software for
handling interrupt control, indexing, I/O control, and certain internal control functions. The
two sets of general-purpose registers are also implemented within this low-order memory
block.
In order to provide a data communications capability, four additional bytes of low-order
main storage are required for each communications line in the subsystem. Since the 9400
can have 128 communications lines, a maximum configuration requires use of the first 5] 2
bytes of main storage. Unauthorized access to this area of memory causes a processor
interrupt •
. 34

Disc File Storage

.341 8410 Disc Storage Subsystem
The 8410 Disc Storage System, announced on December 6, 1967, represents a considerable
enhancement to the UNIVAC 9000 Series product line by providing random or sequential access
to a moderate amount of on-line storage. The recording medium is a nickel-cobalt-coated disc
in an interchangeable cartridge. Each single-disc cartridge can store up to 3.2 million bytes,
but only 1. 6 million bytes of each cartridge can be accessed "on-line". By physically removing
each disc cartridge, turning it over, and replacing it on the drive, the remaining 1. 6 million
bytes of data become accessible.
Two independently operating disc handlers are housed in a single cabinet, and each handler services a single disc cartridge. Up to four dual-disc units are allowed in a maximum 9000 Series
system configuration, providing a total on-line data capacity of 12.8 million bytes.
A seek function on one disc handler can be overlapped with seek, read, or write functions on
another. Once the desired sector address has been determined, the average time required to
position the head mechanism for randomly placed data is 110 milliseconds. The rotational
speed is 1,200 revolutions per minute, which corresponds to 50 milliseconds per revolution
(latency time) . Thus, the average access time for randomly placed data, including 25millisecond average'latency time, is 135 milliseconds, an acceptable access time for use in
many random processing applications.
The fastband search technique, which may be employed at the user's discretion, provides the
address of the desired sector through use of a key (sector location data) contained in a special
50-sector track or "fastband" on each disc surface.
Although the 8410 Oisc Storage System is capable of transferring data at the rate of 136,000
bytes per second, the effective speed, as limited by the data transfer capacity of the Multiplexor
channel, is 100,000 bytes per second. Buffering logic located in the disc drive cabinet provides
6/69

A

(Contd.)

AlJERBACH
~

SUMMARY

810:011.341

.341 8410 Disc File Subsystem (Contd.)
interm{'(hat!.' storage of up to ;. A gruup
of pseudo-instructions is provided to control the assembly process itself.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

6/69

UNIVAC 9000 SERIES

810:011.413

.11:1 Pl'obh'm Oriented Facilities

The card-oriented Report Program Generator (RPG) accepts problem-oriented specifications and generates programs in Machine language to produce the specified reports. Input
to the object programs is from punched cards; output can be in printed and punched-card
form. The UNIVAC RPG uses essentially the same coding forms and specifications as the
lBl\l System/360 RPG; its principal purpose is to ease the transition from punched card
tabulating with unit record equipment to stored-program computing.
Th" Gangpunch-Reproduce Program Generator permits functions normally associated with
tabulating l'quipment to be described in problem-oriented terms, and generates a program
to perform the following specific functions: straight reproducing; master card, interspersed
m;18(er c[ll'd, and offset gangpunching; sequence checking; and consecutive punching.
,\ group of standard subroutines is provided to perform functions such as simulation of the
hanlw

-15

--

--

(Contd.)

810:221.105

PRICE DATA

IDENTITY OF UNIT
CLASS

Model
Number

Feature
Number

Name

PRICES
Monthly
Monthly
Rental PurchasE Maint.

$

(I)

$

$

Magnetic Tape

INPUTOUTPUT
0858-99
F0828-00
F0827-00
0858-14
0858-10
0858-98
0858-01
0858-00

Uniservo VIC Subsystem (includes 9-track
Master Handler and first Slave unit;
maximwn of 8 handlers per subsystem)
7 Track Feature
Data Conversion
Uniservo VIC Slave (9-track: 800 bpi: 34,000
bytes/sec; Maximwn of 3 Slaves per Master)
Uniservo VIC Master (includes 9-track tape
handler and control electronics)
Uniservo VIC Subsystem (includes 7-track
Master Handler and first slave unit;
maximum of 8 handlers per subsystem)
Uniservo VIC Slave (7-track; 200/556/800
bpi; 8540/23,741/, 34,160 char/sec;
maximum of 3 Slaves per Master)
Uniservo VIC Master (includes 7-track tape
handler and control electronics)

895

30,655

195

55
55
310

2,030
2,030
10,470

5
5
70

515

17,350

115

895

30,655

195

310

10,470

70

515

17,350

115

205
45
45
205
80
10
320
155
20

6,315
1,355
1,355
6,315
2,705
410
9,920
4,735
410

60
10
10
60
15

1,250

40,675

315

300

9,790

75

1,550

50,465

390

Punched Card
0711-02
F0872-00
F0872-01
0603-04
F0870-00
FOS71-00
0604-00
FOS75-00
F1054-01

Card Reader (600 cards/min)
Short Card-51 Column
Short Card-66 Column
Card Punch (75 to 200 cards/min)
Read/Punch Feature (pre-punch read station)
Selective Stacker
Card Punch (200 cards/min)
Read/Punch Feature (Pre-punch read station)
90 Column Read

-

90
45

-

Printer (9300 II)
0768-00
FI071-00
076S-99

COMMUNICATIONS

F1000-00
F100S-99

FI002-03
F1002-04
FI002-05
FI005-02
F1005-03
F1005-04
F1005-05

Printer and Control (1,100 lines/min
with 48 char. set; 900 lines/min with
63 char. set)
1,600/1,200 LPM Rate (converts
076S-00 to 076S-99)
Printer and Control (1,600 lines/min with
48 char. set; 1,200 lines/min with 63
char. set)
Line Terminal Controller
(controls one input and/or one output Line
Terminal)
Redundance Check (provides block parity
checking and generation)
Communications Interface (provides
interface between Line Terminal and
data set):
2400-bit/sec line
2000-bit/sec line
Broad-band line
Line Terminal Sync:
Output; 5-, 6-, 7-, or S-level characters;
up to 230,400 bits/sec
Input; 5-, 6-, 7-, or 8-level characters;
up to 230,400 bits/sec
Output; 10-level characters; for
communication with remote 9000 series
computers
Input; 10-level characters; for communication with remote 9000 series computer

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

.~

115

4,350

15

20

740

3

15
15
15

565
565
565

2
2
2

25

870

5

31

1,130

5

31

1,130

z 5

35

1,305

5

4/69

810:221.106

UNIVAC 9300

IDENTITY OF UNIT
CLASS

Model
Number

Feature

Name

PRICES
Monthly
Monthly
Purchase
Maint.
Rental
$
$
$

NOTES:

(l)UNIVAC will extend rentall\greements to a five year term for systems in current production at a monthly
rental of 85 per cent of the figure shown in this column.

4/69

fA..

AUERBACH

(Contd. )

810:221.107

PRICE DATA

UNIVAC 9400
I \' .\C t' is a small label printer that can be interfaced to the Elbit 100
computer. The Shepard SHIl h:JS an ASCII 64-character set and prints 80 characters per line at
600 lines per minute. The Shepard 880 sells for $9850 .

.3

SOFTWARE
The programming system fOl' the Elbit 100 is designed for ease of use. All programs are
written in Elbit Symbolic Language (ESL) for which there is an assembler. ESL instructions are written in mnemonic code and decimal form.
The Elbit Assembly Program (EAP) accepts source programs written in ESL and converts
them into the machine code ('l{uivalent. This conversion is accomplished in two passes.
The first pass is for display of the labels with assigned associated addresses while the
second pass provides the complete listing of the program with the error messages and the
self-loading object tape. The EAP operates with a minimum configuration of 2048 words
of core memory and an on-line Teletype ASR 33/620.
Elbug is an on-line debugging program. Eight powerful operations provide monitoring of
the executed program, tracing of a selected location, altering, printing, and punching of
parts of core memory. Communication with Elbug is via Teletype, and a self-protection
feature does not allow an operator to destroy the Elbug program. Elbug occupies 512
words of core storage.
The Elbit 100 is supported by a fairly wide range of software punched on paper' tape. Subroutines are provided to simulate logical AND, OR, and Exclusive OR. The Arithmetic

© 1969

AU~

RBACH Corporation and AUERBACH Info, Inc.

11/69

1490:011.103

ELBIT 100

Library includes subroutines for fixed point binary operations (both single and double
word) and decimal and floating point calculations. The utility routines provide for BCDto-binary and binary-to-BCD conversion. Also available are decimal, hexadecimal, bitby-bit, and alphanumeric print routines and a provision for table look-up.
The Elbit 100 is supplied with special-purpose programs for such applications as data
concentration, incremental digital plotting, display control, and multichannel integration.

11/69

A

AUERBACH
®

A

1490:221. 100
STANDARD

EDP

ELBIT 100
PRICE DATA

.(PORTS

AUERBACH
~

PRICE DATA

PRICES

IDENTITY OF UNIT
CLASS

CENTRAL
PROCESSOR

Model
Number

Feature
Number

Name

Monthly
Rental
$

Purchase
$

100

Central Processor (4096 words)

7600

100

Central Processor (2048 words)

7000

100

Central Processor (1024 words)

6400

Monthly
Maint.
$

Main Frame Options

INTERFACE
SYSTEMS*

Real-Time Adapter (recognizes 4
levels of Priority Interrupt signals)

500

Real-Time Clock

300

Dt'Hk-Top Cabinet

100

ASH :J:J Interface

825

BellHoll Plotter Interface
Dil;ltronics 2500 Interface

1100
770

Kt'lllledy 1600 Interface

1100

Magnetic Drum Interface

4400

l' E C Incremental Tape Interface

2200

SIll'PP:tI'd 880 Line Printer Interface

1425

Tally 1'120 Interface

-

990

.

*Elbit offers a family of stand a I'd interface cards and mounting hardware to facilitate user-designed
interface systems. These cardH I'ange in price from $110 to $550.

© 1969 AUERBACH CorporatIon and AUERBACH Info, Inc.

11/69

JAPAN

AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

AUE'RBACH
@
Printed in U,S.A.

-1.
-

1540:011.100

Sf"''''

~EDP

AUERBAC~

FUJITSU FACOM 270 SERIES
SUMMARY REPORT

.('ORTS

•

SUMMARY REPORT: FUJITSU FACOM 270 SERIES
.1

BACKGROUND
Fujitsu Limited of Tokyo, Japan, the only independent computer manufacturer in that country,
announced its FACOM 270 Series in 1963. The 270 Series consists of three models, namely,
270-10, 270-20 and 270-30. Models 270-20 and 270-30 have both upward and downward compatibility.
Memory capacities for this series of binary processors range from 1,024 words on the
smallest 270-10 to 65,536 words on the largest 270-30.
The purchase price for the FACOM-270-10 ranges from ¥5,400,000 (about $19,500), to
¥10,000,000, while the FACOM 270-20 sells for from ¥18,000,000 to ¥100,000,OOO. The
largest model in the series, FACOM 270-30 has a purchase price ranging from ¥30, 000, 000
to ¥240, 000, 000. All prices are as of June, 1969.
The 270-10 was first delivered in February, 1963, the 270-20 in July, 1966 and the 270-30
in March, 1968.
Fujitsu, a world-renowned manufacturer of communications and electronics equipment,
should not be dismissed lightly in their independent venture into the computer market. This
is especially true since they more recently announced the F ACOM 230 Series which shares
compatible peripherals with the 270 Series; the FACOM 230 Series is discussed in Summary
Report 1541 .

.2

HARDWARE

.21 System Configuration
The largest FACOM 270-10 system can contain a central processor, an IBM Input-Output
Typewriter, a magnetic drum, a paper tape reader, a paper tape punch, a real-time controller with display systems and other real-time equipment.
The largest FACOM 270-20 system can include a central processor with a built-in magnetic
drum unit, a direct channel to which can be connected a F ACOM WRITER, a paper tape
reader and punch, a line printer, aplotting device, a card reader and a magazine file. Also
included in the configuration are two data channels, 6 magnetic tape units and up to eight of
the other available peripherals. Data communication equipment and real-time 1-0 devices
can also be added.
The largest F ACOM 270-30 system can handle all the peripherals handled by the 270-20 and
additionally can have up to seven data channels, allowing a greater number of peripherals
and data communications devices to be connected .
. 22 Data Structure
The basic addressable unit of the 270 Series is the word, which consists of sixteen binary
bits and one parity bit. The sixteen binary bits can represent two characters or four
decimal digits. The 8-bit character corresponds to the 8-bit byte terminology currently in
use in the U. S. A. Instructions are of one address form and instruction lengths range from
one to two words.
F ACOM 270-20 and 270-30 represent a floating point number as a fraction having 24 or 56
bits and an exponent of 7 bits .
. 23 Central Processor
As presently implemented the FACOM 270 Series appears as fixed word length, binary processors with one address per instruction.
FACOM 270-10 provides no floating point facilities and allows fixed point multiplication and
division only through the use of subroutines.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

7/69

FUJITSU FACOM 270 SERIES

1540:011.230

.23 Central Processor (Contd.)
Parity checking is provided along with three index registers and one level of program
interrupt.
As an auxiliary storage, a 32,768 word magnetic drum unit can be connected to the system.
It has an average access time of 17 milli-seconds.
F ACOM 270-20 and FACOM 270-30 are designed specifically for scientific, engineering, and
process control applications. The two systems differ in the capacity of their central processors, and auxiliary storage, their internal speeds and the number of peripheral devices
that can be attached. The Series 270 central processors are compared in Table I. Two
instructions are provided on both systems for floating point arithmetic to handle operations
of two words and four words in length.
TABLE I.

FACOM 270 SERIES CENTRAL PROCESSOR CHARACTERISTICS

Model
No.

Minimum
Storage
Capacity
(words)

Maximum
Storage
Capacity
(words)

270-10

1,048

4,096

270-20

4,096

270-30

8,192

Cycle
time, J.Lsec

Program
Interrupt,
Levels

Numbers
of
Index
Registers

Checking

Storage
Protection

2/word

Parity

None

1

3

32,768

2.4/word

Parity

Write
only

12

3

65,536

.9/word

Parity

Write
only

12

3

The F ACOM 270 Series accomplishes editing through subroutines in all models. Model 10
has no indirect addressing ar.d none of the models have table look-Up capabilities. Otherwise, the instruction repertoire (34 in all models) of the series includes instructions to
load and store information, to perform boolean operations (Model 10 has no inclusive OR instruction), for shifting, both logical and arithmetic, indexing operations and the four basic
arithmetic operations as well as double-length addition and subtraction •
. 24 Auxiliary Storage
The FACOM 270 Series is provided with four models of magnetic drum, three models of magmetic disc, and one model of magnetic disc pack drive.
Table II lists the characteristics of the available auxiliary storage devices.
TABLE II.

CHARACTERISTICS OF FACOM 270 SERIES AUXILIARY STORAGE DEVICES

Type of emt

Number of
Units On Lme

~!aXlmum

\laXlmum

~umber

l)f

Words Per emt

~

F6~2D

F623A

Drum

Drum

per

channel
131,000
bytes

F624B
Drum

F627A
Drum

F631A
Disc

:. .; per
channel

8 per
channel

8 per
channel

~G2,OOO

2,096,000
bytes

524,000
bytes

0
17
34

0
8.4
17

bytes

F461K
Disc Pack
Drive

F631B
Disc

F631K
Disc

8 per
channel

8 per
channel

4 per
channel

8 per
channel

33.5M
bytes

67.1M
bytes

901l!
bytes

7. 25I\1

0
130
270

0
130
270

bytes

\\'altlng Time, mseL' ~!tmmum
Avera~c

(Handom)

~!axlmum

l-.ffl'etl\'c Transfer Hate,
Chal';&'c
Data

7/69

Checkm~

0

a

10
20

20
40

25.000
bytes
Panty

27,000
bytes
Panty

120,000
bytes

150,000
bytes

a
150
290
56,000
bytes

107,500
bytes

130,000
bytes

0
87.5
160
156,000
bytes

Check bytes Check bytes Check bytes Check bytes Check bytes Check bytes

A

(Contd.)

AUERBACH

c.)

SUMMARY REPORT

1540:011. 250

.25 Magnetic Tape Units
The FACOM ~70 Series includes eight models of magnetic tape units. Anyone model or any
mixture of models can be attached to any of the 270 Series central processors. All models
except one (F401A) have lateral and longitudinal parity checking on reading and read-afterwrite checking on writing; model F401A has track parity checking when reading and double
write when writing. See Table III for data transfer rates and other characteristics of all
available models.
TABLE III.

CHARACTERISTICS OF FACOM 270 SERIES MAGNETIC TAPE UNITS

Peak Speed,
bytes per

Interlock
Gap
Lengths,

second

inches

n~cLlrcting

Tape 'peed
lDches per

second

Model No.

DenSIty
bytes per
inch

Number
of
Tracks

F606A

4

333
556

7

15,000
25,000

F603B

75

200
556

7

15,000
41, 700

F603C

120

200
556

7

F603D

75

556
BOO

F603E

120

F603F

5

IBM

Transfer Rate

Compatibility

Kilo-Char/sec.
lOOO-char
blocks

100-char
blocks

Rewind
Time,
Minutes

729

2400

11.00
15.60

3.26
3.57

5.5

X
X

X
X

.75

12.90
28.90

5.77
7.67

2.0

X
X

X
X

24,000
66,700

.75

20.70
46.40

9.33
12.40

1.5

X
X

X
X

7

41,700
60,000

.75

28.90
36.70

7.67
8.12

2.0

X
X

X
X

556
800

7

66,700
96,000

.75

46.40
59.00

12.40
13.20

1.5

X
X

X
X

BOO

9

60,000

.58

39.50

9.70

2.0

X

.58

63.70

15.80

1.5

X

1.18

.32

1.5

F603G

120

800

9

96,000

F401A

30

333

4

1,670

.26 Peripheral Equipment
The FACOM 270 Series is provided with one card reader model and one card punch model
(Table IV). Eight card readers per channel and eight card punches per channel can be connected to any of the central processors and two additional on-line readers or punches can be
connected via direct channel. The card reader features dual-read data checking and code
translation while the card punch features code translation and read-after-punch data checking.
Six models of paper tape input-output eqUipment are available (Table IV). All models can be
connected to any central processor in the 270 Series eight per channel plus two by direct
channel. The paper tape readers have no code translation but feature dual read data checking
and read 6 or 8 level codes. The paper tape punches feature feed checking and punches 6
or 8 level codes.
Table V lists the five models of typewriters which are available with the 270 Series.
range of speeds and character set sizes have been provided .
.3

A wide

DATA COMMUNICATIONS
The FACOM 270 Series is designed to handle data communications. The system is used to gather
data sets from remote locations, to process them and to output the processed data or send the
appropriate command information back to the remote locations.
To accomplish this, additional equipment such as data transmitters and terminal equipment have
been developed. Data transmitters use 6 or 8 level paper tape at transmission speeds of 50, 200
or 1200 bits per second. A longitudinal redundancy check is performed on all data.
The FACOM 1510 Terminal handles data transceiving on 6 or 8 level paper tape, keyboards and
printers. Data Transmission speeds are 50 to 100 bits per second with a longitudinal-vertical
check on all data.
The FACOM 1530 Terminal is designed exclusively for banking and financial institutions. It consists of a terminal controller and terminal writers. The input-output mode is by keyboard and
printer at data speeds of 200 and 1200 bits per second. Up to three terminal writers can be controlled by each terminal controller. The printing speed is 15.5 characters per second .

.4

SOFTWARE
The software of the 270 Series is designed with emphasis on compatibility among all the models
of the series. The FACOM 270 COBOL compiler language and FACOM 270 FASP assembly
language are typical business oriented languages. COBOL is a full COBOL equivalent to COBOL
'61 Extended, incorporating the SORT verb. F ASP is a real time assembler incorporating pseudo
instructions and macro instructions based on symbolic instructions which have one to one correspondence with machine language.
© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

7/69

1540:011.400

FUJITSU FACOM 270 SERIES

TABLE IV.

PERIPHERAL EQUIPMENT FOR THE FACOM 270 SERIES

Type of Unit

Model Number

Card Reader

F664A

Card Punch

F683A

250

Paper Tape Reader

F749A

400 or 200 char/sec. 6 or 8-level codes

800 Cards per minute

Paper Tape Reader

F749E

1200 or 600 char/sec. 6 or 8-level codes

Paper Tape Reader

F750A

240 char/sec

Paper Tape Reader

F748A

1000 char/sec

Paper Tape Punch

F766A

200 char/sec

Paper Tape Punch

F767A

100 char/sec

PrlOter

F642A

1000 to 1500 lines/mm. 64 or 128 character set

Prmter

F643A

240 to 480 lines/min. 50 or 100 character set

Prmter

F643C

240 to 480 lines/min. 50 or 100 character set

---

Electronic Printer

.4

Rated Speed

10.000 lines/min

XY Plotter

F6201B

400 steps/ sec

Optical Character
Reader

(Document
Scanner)

440 sheets per minute

OptICal Character
Reader

(Page Scanner)

360 lines per minute

Optical Character Reader
Reader

(Journal
Scanner)

440 lines per minute

SOFTWARE (Contd.)
FACOM 270 ALGOL and FACOM 270 FORTRAN are typical scientific compilers. The 270
FORTRAN is equivalent to FORTRAN IV.
A PL/l compiler is also available. For the language specifications, an importance is placed on
International Standards and Japanese Industrial compatibility as much as possible with compiler
languages for other types of computers.
A variety of subroutines and utility programs are available in the form of sorts, merges, file
handling routines and input-output control programs.
Application programs for operation research and linear programming are available, as well as,
PERT/TIME and PERT/MANSCHEDULE.
Simulation Programs such as SOL (Simulation Oriented Language), DYNAMO (DYNAmic Models)
and KEMPF (Kaigin Econometric Method Program by Fujitsu) are available.
Approximately 100 library routines are available for scientific and engineering applications.
TABLE V.

7/69

FACOM 270 SERIES TYPEWRITER CHARACTERISTICS.
FACOM 788A

FACOM 789A

FACOM 790A

FACOM 791A

FACOM 801A

Rated speed, char/sec

10

15.5

9.8

15

15

Size of character set

88

88

86

128

88

1\umber of columns

100

156

130

120

130

Character spacing,
char/inch

10

12

10

10

10

fA
AUERBACH

"

•

1540:221.101

A.

STUIUD

EDP

AUEReA,CH

FUJITSU FACOM 270 SERIES
PRICE DATA

REPORTS

PRICE DATA

Prices

Identity
Class
Model
Number
CENTRAL
PROCESSOR

Monthly
Rental, ¥

Purchase,
¥

265,000

11,925,000

70,000

3,150,000

FACOM 270-30 System
Central Processing Unit
(Including a Drum)

490,000

22,050,000

Main Storage --- 8 k 16 bits-Words

190,000

8,550,000

Main Storage --- 16 k
16 bits-Words

350,000

15,750,000

20,000
65,000
10,000

900,000
2,925,000
450,000

30,000
100,000
25,000
10,000

1,350,000
4,500,000
1,125,000
450,000

Feature
Number

F7200A
F7220A

F7300A
F7320A
F7310A

,
F7210A
F7211A
F7233A

Name
FACOM 270-20 System
Central Processing Unit
(Including Drum)
Main Storage --- 4 k 16 bits-Words
(Up to 32 k Words)

Features for 270-20 Systems
Memory Protection
Floating Point Arithmetic
Line Printer Adapter
Features for 270-30 System

F7310A
F7311A
F7312A
F7333A

Memory Protection
Floating Point Arithmetic
TIMER
Line-Printer Adapter
Channels

INPUTOUTPUT

F7232A 1
F7232A
F7232B 2
1

For 270-20 System
Data Channel (Single)
Data Channel (Dual)
Data Channel (Single)
(A magnetic tape controller)

F7232B2

105,000
180,000
130,000

4,725,000
8,100,000
5,850,000

Data Channel (Dual)
(Dual magnetic tape controller)

220,000

9,900,000

F7232C

Data Channel (A data channel and a
magnetic tape controller)

190,000

8,550,000

F7331A
F7332A

For 270-30 System
Data Channel (Single)
Data Channel (Single)
(A magnetic tape controller)

60,000
110,000

2,700,000
4,950,000

45,000
180,000
260,000

2,025,000
8,100,000
11,700,000

F567K
F664K
F683K

Punched Card and Printer
Card Reader (100 cards/min)
Card Reader (800 cards/min)
Card Punch (250 cards/min)

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

FUJITSU FACOM 270 SERIES

1540:221.102

Identity

Class
Model
Number
INPUTOUTPUT
(Contd)

Prices

Feature
Number

Name

F687K

Card Punch (Require F 1620 A/B)

F643A

Monthly
Rental, ¥

Purchase,
¥

65,000

2,925,000

Line-printer (240 lines/min 80 position)

110,000

4,950,000

F643C

Line-printer (240 lines/min 136 position)

135,000

6,075,000

F642K

Line-printer (1,000 lines/min 136 position)

330,000

14,850,000

FI021A

Paper Tape Adapter
Input-output adapter (For up to 4 paper tape
units)

96,000

4,320,000

FI006

Paper tape adapter

60,000

2,700,000

F1406

Paper tape adapter

93,000

4,185,000

F1620A

Paper Tape Adapter
(Up to 4 paper tape units)

100,000

4,500,000

Paper Tape Adapter
(Up to 8 paper tape units)

160,000

7,200,000

112,000

5,000,000

F1620B

F1211A

Magnetic Tape
Magnetic Tape Controller
(Single channel) for F603 B/ C

F1212A

Magnetic Tape Controller
(Dual channel) for F603 B/C

198,000

8,910,000

F1205A

Magnetic Tape Controller (Single channel)
for F606A only

105,000

4,725,000

Disk and Drum Storage

7/69

F1213A

Magnetic DISK Controller (For a single
disk)

130,000

5,850,000

FI2I4A

Magnetic DISK Controller (For one or
two disks)

220,000

9,900,000

F631A

Magnetic DISK Unit (33.5 MB )

750,000

33,750,000

F631B
F631K

Magnetic DISK Unit (67.1 MB)

1,000,000

45,000,000

900,000
130,000

40, 500, 000

F1223A

Magnetic DISK Unit (90MB)
DISK PACK Controller
(Up to 4 Devices)

F461B

DISK PACK Unit (5. 12MBytes)

185.000

8,325,000

FI718A

File Control Unit

330,000

14,850,000

F462K

DISK PACK Unit --7.25 M Bytes (Requires FI718A)

180,000

8,100,000

A

5,850,000

(Contd. )

AUERBACH

'"

PRICE DATA

1540:221,103

Model
Number
INPUTOUTPUT
(Contd)

Prices

Identity

Class
Feature
Number

Name

Monthly
Rental, ¥

Purchase,
¥

F622D

Magnetic Drum Unit (131 KB )

195,000

8,775,000

F623A

Magnetic Drum Unit (262 KB )

300,000

13,500,000

F627K

Magnetic Drum Unit (622KB)

210,000

9,450,000

F624B

Magnetic Drum Unit (2,097 KB )

450,000

20,250,000

Paper Tape Unit
F749A

Paper Tape Reader (200/400 char/sec)

39,000

1,305,000

F750A

Paper Tape Reader (240 char/sec)

38,000

1,710,000

F749E

Paper Tape Reader (600/1200 char/sec)

45,000

2,025,000

F748A

Paper Tape Reader (1,000 char/sec)

85,000

3,825,000

F748E

Paper Tape Reader (600 char/sec)

65,000

2,835,000

F766A

Paper Tape Punch (200 char/sec)

60,000

2,700,000

F767A

Paper Tape Punch (100 char/sec)

43,000

1,935,000

40,000

1,800,000

On- Line Type Writer

F801D

FACOM-WRITER

F80lD-l

Subpaper Tape Reader

3,800

171,000

F80lD-3

Verify Option

7,600

342,000

F790A

Typewriter

70,000

3,150,000

F791A

Typewriter

55,000

2,475,000

F788A

Typewriter

48,000

2,160,000

F789

Typewriter

60,000

2,700,000

165,000

7,425,000

55,000

2,475,000

Display
F-6221A

Character Display Unit

F6201B

X-Y Plotter

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

7/69

A.

1541:011.100

STANDARD

EDP

FUJITSU FA COM 230 SERIES
SUMMARY REPORT

REPORTS

AUERBACH

SUMMARY REPORT: FUJITSU FACOM 230 SERIES
.1

BACKGROUND
The F ACOM 230 Series is manufactured by Fujitsu Limited of Tokyo, Japan. Originally the 230
Series consisted of five models that were classified into three categories, small-size Model 10,
small-to-medium-size Models 20 and 30, and large-size Models 50 and 60. Upward data and
program compatibility is maintained within each category of the Series. More recently Fujitsu
has announced three new models- Models 25, 35 and 45. Conceived as real-time systems, these
models take full advantage of Fujitsu's complement of real-time equipment.
FACOM 230-10 rents for ¥270, 000 (about $970) per month. The system was first delivered in
December, 1965. The 230-20 was announced in October, 1965 and has monthly rentals of
¥1, 200, 000 to ¥4, 300, 000. FACOM 2:lO-30 was announced in June, 1964 and rents for
¥2, 000, 000 to '1"8,000,000. First dclivered in March, 1966, the FACOM 230-50 rents for
'1"2,900, 000 to '1"15,000,000. FACOM 2:30-60 was announced in mid-1967 and has a monthly
rental ranging from ¥10, 000, 000 to ¥72, 000, 000. All prices quoted were as of June, 1969 .

.2

HARDWARE

. 21 System Configuration
A 230 Series system can span tremendous ranges in size, processing power and areas of applicability. The smallest and most baSIC system, FACOM 230-10, consists of central processing
unit, 4,096 byte main memory, 6ri, ;,:H; byte internal magnetic drum, typewriter, paper tape
reader, and paper tape punch.
At the other end of the scale, the FA('OM 230-60 can have a multiprocessor configuration consisting of a maximum of 2 centralpl'oct'ssors and extended core to a maximum of eight units of
262,144 words each, eight selector channel units, three multiplexor channel units, four magnetic drum units, two disc units, eif\ht lllagnetic tape units, 2048 communication lines, three
card readers, a card punch, two lint' pl'inters, a paper tape punch and reader and up to 512
typewriters.
Models 230-23, 35 and 45 represent I'U] itsu' s entry into the real-time market. The configurations are formed much the same as tlH'lr scientific and business counterparts, Models 20, 30
and 30. However, special consideration has been given to their usefulness in the real-time
environment. These models all COil\(' equipped with the terminal facilities that are available
with the 230 Series (shown in Par:tf\l'aph .3) .
. 22 Data Structure
The word length on the 230-10 is H binary bits plus a word mark and a parity bit. On the 230-20
and 230-30 a word IS 4 bits long WIth an associated word mark and parity bit. The 230-50 and
230-60 have 36 bits per word for data plus 4 flag bits and 2 parity bits.
On Models 230-10, 230-20 and 230-:lO characters are eight bits in length and on Models 230-50
and 230-60 characters are represented by six bits.
Model 230-10 has variable length operands, up to six bytes per instruction and a binary or decimal arithmetic representation.
Models 230-20 and 230-30 have variable operands, 4, 8 or 12 digits per instruction and a decimal or hexadecimal arithmetic radix. Models 230-50 and 230-60 have fixed length operands, one
word per instruction and perform binary arithmetic.
The three smaller models have va riable size fraction, up to 20 digits on Model 10 and up to 128
digits on Models 20 and 30. All three models have two digit exponents. Models 50 and 60 have
27 or 62 bits fraction and 9 digit exponents.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

7/69

FUJITSU FACOM 230 SERIES

1541:011.230

.23 Central Processor
The 230 Series range of applicability is reflected in the varying span of its instruction sets.
Model 10 has an instruction repertoire of 30 instructions which is considerably less than the
other models of the 230 Series; Models 20 and 30 have 70 instructions each and Model 50 has
257 instructions while Model 60 has 235 instructions. Models 25, 35 and 45 have 84, 84 and 94
instructions respectively.
The basic arithmetic add and subtract operations are included in all the FACOM 230 Series
models. The other arithmetic operations are not included in the Model 10 although they are
included in the other models; subroutines are provided on the Model 10 to simulate the missing
operations. Model 10 also differs from the other models of the Series in having no address
indexing capabilities.
The instruction repertoires of all models include flexible facilities for character manipulation,
program control, shifting, logical, and test and transfer operations. In addition, Models 50
and 60 have an extensive repertOire of table look-up and translate instructions.
FACOM 230-10 is the smallest model of the 230 Series. Model 10 appears as a binary-decimal
arithmetic processor with one to six bytes per instruction and having 0, 1, or 2 addresses per
instruction.
FACOM 230-20 was announced as a small-to-medium-sizc computer for business applications.
Models 20 and 30 have similar processor features (Table I). Both models have hardware facilities for floating point arithmetic as well as indirect addressing, editing and boolean operations.
Instruction lengths are 4, 8 or 12 digits, with variable operands and 0, 1 or 2 addresses. The
arithmetic radix can be either decimal or hexadecimal.

TABLE L

FACOM 230 CENTRAL PROCESSOR CHARACTERISTICS

~3U-ll)

230-20

230-30

2:W-50

230-60

230-25

230-33

230-45

Storage Capacity
\hnlmum

4,096

4,096

4,096

)laXln1Um

R,192

32,768

32,768

](;, :lK,1

32,768

8,192

32,7G8

65,536

,"):w

262,144

65,536

131,072

I, 048,000

65,

Cy cle TIme, usee

2,0

1,8

2.2

'

\lemor y ProtectIOn

:\0

Yes

Yes

2

70

:';umber of Index Register
\"umber of Instructions

30

0.92

1.5

. ::;

,J

y(,s

Yes

Yes

Yes

Yes

2

R

8

8

8

8

70

2!J7

235

84

84

94

"

FACOM 230-30 is a medium-size, general purpose, variable word length system with a main
memory cycle time of 2.2 microseconds per two digits. Main memory is available in sizes
ranging from 8,192 to 65,536 core locations. Up to four selector channels, one of which can be replaced by a multiplexor channel, can be connected to the system. As an option, memory protection is available in units of 128 bytes. Both decimal and binary arithmetic can be performed.
Instructions are 4, 8 or 12 digits in length. Two index registers are provided and indirect addressing and boolean operations are possible.
F ACOM 230-50 is a large scale, fixed word length, binary, general purpose system. Each word
consists of 42 bits: 36 data bits, 4 flag bits and 2 parity bits. Thirty-six bits of binary data
corresponds to approximately 10,5 decimal digits. Main memory capacities range from 16,384
words to 65,536 words with a cycle time of 2.2 microseconds per word. Both fixed and floatingpOint operations can be performed and double precision is available for floating-point operations.
Up to seven selector channels and/or multiplexor channels can be connected to the system.
Eight index registers are available and indirect addressing and boolean operations can be performed.
The F ACOM 230-60 general purpose system is the largest model among the FACOM Series 230.
Model 60 is downward compatible with Model 50 in both program and data.

7/69

A

AUERBACH
@

(Contd. )

SUMMARY REPORT

1541 :011. 240

.23 Central Processor (Contd.)
F ACOM 230-60 is a binary machine having an instruction length of one word and one address per
instruction. Five levels of interrupt are provided on the 230-60. These include overflow conditions, Input-Output control and operator intervention. Eight index registcrs are provided.
There are facilities for indirect addressing, editing of formats, table look-up and boolean opertions. Multiprocessing is achieved by connecting up to two central processors and up to three
channel controllers. Each channel controller can control up to six multiplexor channels and/
or selector channels.
Main memory is configured as a multi-bank system. The memory is expandible in units of
32,768 words from a minimum memory capacity of 32,768 words to a maximum capacity of
262,144 words. Each unit of 32,768 is further divided into two banks, so that the minimum
capacity of main memory has two banks, and the largest has 16 banks. Each bank operates independently with a cycle time of 0.92 microseconds.
Optionally, an even larger memory unit can be added to the system. An additional 786,432
words in units of 262,144 words can be added to the previous maximum capacity of 262,144
words. Each extended 262,144 word unit is divided into two banks, each operating independently
at cycle rates of 6 microseconds. The console has facilities that partition the system, i.e., by
setting manual switches individual units can be made unavailable for use by other units.
Main memory accessing is 2-way, 4-way or 8-way interleaved. Each central processor can
access three words of data at a timc and each channel control unit can access two words of data
at a time, all units operating simultaneously.
The FACOM 230-60 has six base registers that are used for dynamic relocation of programs
within core memory making it suitable for time-sharing operations, on-line real time processing and multiprogramming.
F ACOM 230-60 can easily function as a data communications processor since each selector
channel can handle up to 256 input-output devices and each multiplexor channel, having 1024 subchannels, can be connected to communications lines .
. 24 Auxiliary Storage
The FACOl\1 230 Series is provided with four models of magnetic drum, three models of magnetic disc, and one model of magneti c disc pack drive. Three units of additional core memory
consisting of 262,144 words each are available for Model 230-60. This additional core memory has an average access time of (; microseconds and an effective transfer rate of 6 million
bytes per second. Parity checking is provided.
Table II lists the characteristics of the remaining auxiliary storage devices.
TABLE II. CHARACTERISTICS OF FACOM 230 SERIES AUXILIARY STORAGE DEVICES

Type of l nIt

:\la.'Cimum

~umher

F6.!2D Drum

of Lmts on-lme

,. )'la.'Gmum :\umber of v.ords/ umt

.... /chanoel
131,000 bjtes

F623A Drum

8/channel
262.000 bytes

F624B Drum

F627A Drum

8/ channel

8/ channel

2.096.000

j24 ,ODO bvtes

F631A DISC

F631B DISC

F631K Drum

8/channel

8, channel

4/channel

33. 5M bytes

65.1M bytes

90M bytes

F461K DISC

Pack Dnve
81

channd

7. 23M bytes

bytes
\\ altmg time. mst'c 0

0

0

0.0

0

0

0

Average (Random)

10

20

17

8.4

150

150

130

87.5

:\laXlmum

20

40

34

17.0

290

290

270

160.0

.:'whmffium

Effective Tr,lnsfc>r Hate

Data l 'h{'cklng

Char sec

25,000 bytes

27 , 000 byte!'>

Parity

Parlt)

120,000 bytes

130,000 bytes

Check bytes

Check bytes

0.0

56,000 bytes

107.500 bytes

130.000 bytes

156, 000 bytes

Check b.ytes

Check bytes

Check bytes

Check bytes

.25 Magnetic Tape Units
The FACOM 230 Series is provided eight models of magnetic tape units. Anyone model or
any mixture of models can be attached to any of the 230 Series central processors. All models
except one (F401A) have lateral and longitudinal parity checking on reading and read-afterwrite parity checking on writing.
Model F401A has track parity checking when reading and double write when writing.
Table III for data transfer rates and other characteristics of all available models.
© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

See

7/69

1541:011.260

FUJITSU FACOM 230 SERIES

TABLE lli. CHARACTERISTICS OF FACOM 230 SERIES MAGNETIC TAPE UNITS
Hel'ording
Model No.

Tape Speed

DensIty

mches per

bytes per
inch

second

Number
of

Peak Speed.

bytes per
second

Tracks

F606A

4

333
556

7

15,000
25,000

F603B

75

200
556

7

15,000
41,700

F60JC

120

200
556

7

F603D

75

556
800

F603E

120

F603F

Interlock
Gap
Lengths.

inches

IBM
Compatibility

Transfer Rate
Kilo-Char/sec.
1000-char
blocks

100-char
blocks

Rewind
Time,
Minutes

729

2400

11.00
15.60

3.26
3.57

5.5

X
X

X
X

.75

12.90
28.90

5.77
7.67

2.0

X
X

X
X

24,000
66,700

.75

20.70
46.40

9.33
12.40

1.5

X
X

X
X

7

41,700
60,000

.75

28,90
36.70

7.67
8.12

2.0

X
X

X
X

556
800

7

66,700
96,000

.75

46.40
59.00

12.40
13.20

1.5

X
X

X
X

5

800

9

60,000

.58

39.50

9.70

2.0

X

F603G

120

800

9

96,000

.58

63.70

15.80

1.5

X

F40lA

30

333

4

1,670

1.18

.32

1.5

.26 Peripheral Equipment
The FACOM 230 Series is provided with one card reader model and one card punch model
(Table IV). Eight card readers per channel and eight card punches per channel can be connected to any of the central processors and two additional on-line readers or punches can be
connected via direct channel. The card reader features dual-read data checking and code
translation while the card punch features code translation and read-after-punch data checking.
Six models of paper tape input-output equipment are available (Table IV). All models can be
connected to any central processor in the 230 Series eight per channel plus two by direct
channel. The paper tape readers have no code translation but feature dual read data checking and read 6 or 8 level codes. The paper tape punches feature feed checking and punches
6 or 8 level codes but have no code translation.
TABLE IV. PERIPHERAL EQUIPMENT FOR TIlE FACOM 230 SERIES
Type of Unit
Card Reader

F664A

Peak Speed
SOO Cards per minute

Card Punch

F6S3A

250

Paper Tape Reader

F749A

400/200 char/sec, 6/S level codes

Paper Tape Reader

F749E

1200/600 char/sec. 6/S level codes

Paper Tape Reader

F750A

240 char/sec

Paper Tape Reader

F74SA

1000 char/sec

Paper Tape Punch

F766A

200 char/sec
100 char/sec

Paper Tape Punch

F767A

Printer

F642A

1 • 500/1 • 000 lines/ min. 64/128 character set

Printer

F643A

480/240 lines/min. 50/100 character set

Printer

F643C

480/240 lines/min. 50/100 character set

Electronic Printer
XY Plotter
Optical Character Reader

7/69

Model Number

-

10.000 lines/min

F620lB

400 Steps/ sec

(Document Scanner)

440 sheets per minute

Optical Character Reader

(Page Scanner)

360 lines per minute

Optical Character Reader

(Journal Scanner)

440 lines per minute

fA

AUERBACH
~

(Contd. )

SUMMARY REPORT

t 54 t :0 t t • 300

. 26 Peripheral Equipment (Contd.)
Table V lists the five models of typewriters which are available with the 230 Series. A wide
range of speeds and character set sizes have been provided •
• 3 DATA COMMUNICATIONS
The FACOM 230-20 On-Line System is designed specifically for data communications. The
system is used to gather data from remote locations, to process them and to output the
processed data or send the appropriate command information back to the remote locations.
To accomplish this, additional equipment such as data transmitters and terminal equipment
have been developed. Data transmitters use 6 or 8 level paper tape at transmission speeds
of 50, 200 or 1200 bauds. A longitudinal redundancy check is performed on all data.
The FACOM 1510 Terminal handles data transceiving on 6 or 8 level paper tape, keyboards
and printers. Data Transmission speeds are 50 and 100 bauds with a longitudinal-vertical
check on all data.
The FACOM 1530 Terminal is designed exclusively for banking and financial institutions. It
consists of terminal controller and terminal writers. The input-output mode is by keyboard
and printer at data speeds of 200 and 1200 bauds. Up to three terminal writers can be controlled by each terminal controller. The printing speed is 15.5 characters per second •
. 4 SOFTWARE
The software of the 230 Series is designed with emphasis on compatibility among all the
models of the series. The Master Control Program for the series can control concurrently
up to four jobs. The FACOM 230 COBOL compiler language and FACOM 230 FASP assembly
language are typical business oriented languages. COBOL is a full COBOL equivalent to
COBOL '61 Extended, incorporating the SORT Verb.
FACOM 230 ALGOL and FACOM 230 FORTRAN are typical scientific compilers. The 230
ALGOL is full ALGOL based on Revised Report ALGOL 60. The 230 FORTRAN is equivalent
to FORTRAN IV.
A PL/I compiler is also available. As to the language specifications, an importance is
placed on International Standards and Japanese Industrial Standards (JIS) and, at the same
time, consideration is given to obtaining as much compatibility as possible with compiler
languages for other types of computers.
A variety of subroutines and utility programs are available in the form of sorts, merges,
file handling routines and input-output control programs.
Application programs for operation research and linear programming are available, as well
as, PERT/TIl\1E and PERT/MANSCHEDULE.
Simulation Programs such as SOL (Simulation Oriented Language), DYNAMO (DYNAmic
:\Iodels) and KEMPF (Kaigin Econometric Method Program by Fujitsu) are available.
Approximately 100 library routines are available for scientific and engineering applications.
TABLE V.

FACOM 230 SERIES TYPEWRITER CHARACTERISTICS

FACOM 788A

FACOM 789A

FACOM 790A

FACOM 791A

FACOM 801A

Rated speed, char/sec

10

15.5

9.8

15

15

Size of character set

88

88.0

86.0

128

88

Number of columns

100

156.0

130.0

120

130

Character spacing,
char/inch

10

12.0

10.0

10

10

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

7/69

1541:221.101

A.

STUDARD

EDP

AUERBACH

FUJITSU FACOM 230 SERIES
PRICE DATA

REPORTS

PRICE DATA

Identity
Class

CENTRAL
PROCESSOR

Model
Number

Feature
Number

F2100A
F2120A

Prices
Monthly
Rental, ¥

Name

FACOM 230-10 System
Central Processing Unit
(Including 4 k Bytes and a Drum)
Main Storage --- 4 k Bytes
(Up to 8 k bytes)

Purchase,
¥

230,000

10,350,000

58,000

2,610, 000

260,000

11,700,000

FACOM 230-20 System
F2200A

Central Processing Unit

F2220A

Main Storage --- 4 k Bytes
(4 k to 8 k Bytes)

70,000

3, 150, 000

F2221A

Main Storage --- 8 k Bytes
(16 k to 32 k Bytes)

130,000

5,850,000

410,000
238,000
292,000
400,000
616,000

18,450,000
10,710,000
13,140,000
18,000,000
27,720,000

79,000

3,555,000

184,000
133,000

8,280,000
5,985,000

1,640,000

73,800,000

FACOM 230-30 System
F2311A
F2321A-I
F2321A-II
F2321A-III
F2321A-IV

Central Processing Unit
Main Storage --- 4 k Bytes
Main Stora::;e --- 8 k Bytes
Main Storage --- 16 k Bytes
Main Storage --- 32 k Bytes

F2341A-I

Console

F2331A-I
F2331A-II

Index and Floating Arithmetic
Index Option
F ACOM 230-50 System

F2500B

Central Processing Unit
(Including 8 k 36-bit words)

F2520B

Main Storage --- 8 k - 36-bit words

380,000

17,100,000

F2540C

Console (Including 100 card/min
card reader)

129,000

5,805,000

Console (Including 400 char/sec paper
tape reader)

129,000

5,805,000

Console

100,000

4,500,000

F2541C
F2545C

FACOM 230-60 System
F2600A

Central Processing Unit

2,450,000

110,000,000

F2620A

Main Storage --- 32 k - 36-bit words
(Up to 262 k - 36-bit words)

1,800,000

81, ODD, 000

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

7/69

1541 : 22 1. 102

FUJITSU FACOM 230 SERIES

Prices

Identity
Class

CENTRAL
PROCESSOR
(Contd)

Model
Number

Feature
Number

Name

Monthly
Rental, ¥

Purchase,
¥

160,000
70,000
30,000
40,000
150,000

7,200,000
3,150,000
1,350,000
1,800,000
6,750,000

F2632A
F2612A
F2640A
F2642A
F2610A

High Speed Transfer Unit
Main Storage Adapter
Operating Console
Configuration Console
Multi-Processor Adapter
Bulk Core Memory

F2621A

Bulk Core Memory --- 131 k 36-bit words
(Only FACOM 230-60)

2,000,000

90,000,000

F2622A

Bulk Core Memory --- 262 k - 36-bit words 3,200,000
(Only FACOM 230-60)

144,000,000

Processor Options
For 230-10 System
F2130A
F2130A 1
F2130A2
F2131A 1

Input-Output Controller
Input-Output Controller I
Input-Output Controller II
Line Printer Adapter

36,000
12,000
12,000
10,000

1,620,000
540,000
540,000
450,000

20,000

900,000

For 230-20 System
F2210A

Memory Protection

Channels
For 230 -20 System
F2230A
F2231A

Data channel (Single)
Data channel (Single)

50,000
70,000

2,250,000
3,150,000

162,000
269,000

7,290,000
12, 105,000

197,000

8,910,000

1,320,000

49,400,000

For 230-30 System
F2332B 1
F2332B 2

Data channel (Single)
Data channel (Dual)
For 230-50 System

F2531B

Data channel (Single)
For 230-60 System

F2639A

7/69

Data Channel Controller
(Up to 6 channels)

A

AUERBACH

'"

(Contd. )

PRICE DATA

1541:221. 103

Prices

Identity
Class
Model
Number
CENTRAL
PROCESSORS
(Contd)

Feature
Number

F2638A
F2613A
F2631A

Name

Channel Adapter
Channel Power Supply Units
Data Channel (Single)

Monthly
Rental, ¥

Purchase,
¥

100, 000
60,000
320,000

4,500,000
2,700,000
14,400,000

For 230-20 System
F2233A

Multiplexor Channel
(256 Sub-channels)

150,000

6,750,000

F2233B

Multiplexor Channel
(512 Sub-channels)

170,000

7,650,000

For 2:l0-30 System
F2335A

Multiplexor Channel
(256 Sub-channels)

150,000

6,750,000

F2335B

Multiplexor Channel
(512 Sub-channels)

170,000

7, 650,000

For 230-50 System
F2533B

Multiplcxor Channel
(512 Sub-channels)

230,000

10,350,000

F2533C

Multiplexor Channel
(1024 Sub-channels)

250,000

11,250,000

380,000

17,100,000

For 230-60 System
F2633A

Multiplexor Channel
(1024 Sub-channels)

INPUTOUT peT
Punched Card and Printer
F567K
F664K
F683K
F687K
F643A
F643C
F642K

Card Header (100 C/M)
Card Reader (800C/M)
Card Punch (250 C/M)
Card Punch (Require F 1620 A/B)
Line-printer (240 L/M --- 80 position)
Line-printer (240 L/M --- 136 position)
Line-printer (1,000 L/M --- 136 position)

45,000
180,000
260,000
65,000
110,000
135,000
330,000

2,025,000
8,100,000
11,700,000
2,925,000
4,950,000
6,075,000
14,850,000

96,000

4,320,000

Paper Tape Adapter
Fl021A

Input-output adapter
(Up to 4 paper tape units)

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

7/69

1541:221.104

FUJITSU FACOM 230 SERIES

Identity

Prices

Class
Model
Number

INPUTOUTPUT
(Contd)

Feature
Number

Name

Monthly
Rental, ¥

Purchase
¥

F1006
F1406
F1620A

Paper
Paper
Paper
(Up to

tape adapter
tape adapter
tape adapter
4 paper tape units)

60,000
93,000
100,000

2,700,000
4,185,000
4,500,000

F1620B

Paper Tape Adapter
(Up to 8 paper tape units)

160,000

7,200,000

Magnetic Tal2e

FACOM
2aO-20

F1211A

Magnetic Tape Controller
(Single channel) for F60a B/C

112,000

5,000,000

F1212A

Magnetic Tape Controllcr
(Dual channel) for F60a B/C

198,000

8,910,000

F1205A

Magnetic Tape Controller
(Single channel) for F606A only
Magnetic Tape Controller
(Dual channel) for F606A only
Magnetic Tape Controller
(Single channel) for F603 F /D

105,000

4,725,000

160,000

7,200,000

140,000

6,300,000

235,000

10,575,000

140,000

6,300,000

F1206A
F1216A
F1217A

FACOM
230-30

FACOM
230-50

7/69

F1216B

Magnetic Tape Controller
(Dual channel) for F603 F /D
Magnetic Tape Controller
(Single channel) for F603 G/E

F1217B

Magnetic Tape Controller
(Dual channel) for F603 G/E

235,000

10,575,000

F1002A 1

Magnetic Tape Controller
(Single channel) for F603 B/C/D/E

112,000

5,040,000

FI002A 2

Magnetic Tape Controller
(Dual channel) for F60a B/C/D/E

198,000

8,910,000

F1201A

Magnetic Tape Controller
(Single channel) for F603 B/C/D/E
Magnetic Tape Controller
(Dual channel) for F603 B/C/D/E

123,000

5,535,000

218,000

9,810,000

165,000

7,425,000

10,000

450,000

165,000

7,425,000

F1202A
Fl716A

Magnetic Tape Controller
(Single channel) for F603 D/F

Fl716A 11

Cross-Call Option
(for Fl716A)

Fl716B

Magnetic Tape Controller
(Single channel) for F603 E/G

fA

AUERBACH

'"

(Contd.)

1541:221.105

PRICE DATA

Prices

Identity
Class
Model
Number
INPUTOUTPUT
(Contd)

FACOM
230-60

Feature
Number

Name

FI716B 11
FI717A

Cross Call Option
(for F 1716B)
Magnetic Tape Controller
(Dual channel) for F603 D/F

FI717A11

Cross Call Option
(for FI717 A)

FI717B

Magnetic Tape Controller
(Dual channel) for F603 E/G

FI717B11

Cross Call Option
(for FI717B)

F603B
F603C
F603D
F603E
F603F
F603G

Magnetic
Magnetic
Magnetic
Magnetic
Magnetic
Magnetic

Tape Unit (41. 7~g)
Tape Unit (66 Kt ) )
Tape Unit (60 K C)
Tape Unit (96 KB
Tape Unit (60 KB
Tape Unit (96

l

Monthly
Rental, ¥

Purchase
¥

10,000

450,000

275,000

12,375,000

20,000

900,000

275,000

12,375,000

20,000

900,000

123,000
184,000
145,000
185.000
170,000
220,000

5,535,000
8,280,000
6,525,000
8,325,000
7,650,000
9,900,000

130,000
220,000

5,850,000
9,900,000

750,000
1,000,000
900,000
130,000

33,750,000
45,000,000
40,500,000
5,850,000

185,000
330,000
180,000

8,325,000
14,850,000
8,100,000

195,000
300,000
210,000
450,000

8, 775, 000
13,500,000
9,450,000
20,250,000

39,000
38,000
45,000
85,000
65,000
60,000
43,000

1,305,000
1,710,000
2,025,000
3,825,000
2,835,000
2,700,000
1,935,000

Disk and Drum Storage
F1213A
FI214A
F631A
F631B
F631K
F1223A
F461B
F1718A
F462K
F622D
F623A
F627K
F624B

Magnetic DISK Controller (Only one)
Magnetic DISK Controller (Up to two
Devices)
Magnetic DISK Unit (33.5 MB )
Magnetic DISK Unit (67.1MB )
Magnetic DISK Unit (90 MB)
DISK PACK Controller
(Up to 4 Devices)
DISK PACK Unit (5. UMBytes)
File Control Unit
DISK PACK Unit --- 7.25 M Bytes
(Requires FI718A)
Magnetic Drum Unit (131 KB )
Magnetic Drum Unit (262 KB )
Magnetic Drum Unit (622KBin
Magnetic Drum Unit (2,097
)
Paper Tape

F749A
F750A
F749E
F748A
F748E
F766A
F767A

Paper
Paper
Paper
Paper
Paper
Paper
Paper

Tape
Tape
Tape
Tape
Tape
Tape
Tape

Reader (200/400 char/sec)
Reader (240 char/sec)·
Reader (600/1200 char/sec)
Reader (1,000 char/sec)
Reader (600 char/sec)
Punch (200 char/sec)
Punch (100 char/sec)

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

7/69

1541:221.106

FUJITSU FACOM 230 SERIES

Prices

Identity
Class
Model
Number
INPUTOUTPUT
(Contd)

Feature

Name

Monthly
Rental, ¥

Purchase,
¥

On-Line Type Writer
FBOlD
FBOlD-l
FBOlD-3

FA COM-WRITE R
Subpaper Tape Reader
Verify Option

40,000

F790A
F791A

Typewriter
Typewriter
Typewriter
Typewriter

70,000
55,000

F7BBA
F7B9A

3,BOO

7,600

I, BOO, 000
171,000
342,000

60,000

3,150,000
2,475,000
2,160,000
2,700,000

165,000

7,425,000

55,000

2,475,000

4B,000

Display

7/69

F-6221A

Character Display Unit

F6201B

X-Y Plotter

A

AUERBACH

'"

fA

AUERBACH
COMPUTER
NOTEBOOK

1555:011.100

INTERNATIONAL

HITACHI HITAC 3010
SUMMARY REPORT

AUERBACH

'"

SUMMARY REPORT . HITACHI HIT AC 3010
.1

BACKGROUND
The IDTAC 3010 is a small-to-medium-scale data processing system manufactured by
Hitachi, Ltd., of·Tokyo, Japan. Manufactured under license from the Radio Corporation
of America, the solid-state IDTAC 3010 is similar in most respects to the secondgeneration RCA 301, a popular U. S. data processing system that has performed reliably
in several hundred installations.
The IDTAC 3010 was announced in 1961, and initial customer deliveries were made in
May 1962. Approximately 80 systems have been delivered to date. Hitachi is currently
marketing the HITAC 3010 system in Japan.
Monthly rentals for typical HITAC 3010 configurations range from about $5,000 for a
system with 10,000 characters of core memory, 4 magnetic tape units, a line printer,
and card equipment to about $22,000 for a system with a 40,000 -character memory, 12
high-speed magnetic tape units, a line printer, high-speed card equipment, and an interrogating typewriter. The 3010 system uses a data structure based upon six-bit alphanumeric characters, and its internal operations are more suitable for general business
data processing than for scientific applications. Memory cycle time is either 3.5 or 7
microseconds per 2-character access.
The IDTAC 3010 system features a high degree of compatibility with the RCA 301 and with
the larger HITAC 4010 Realcom system, Hitachi's version of the RCA 3301 Realcom system. In addition, the third generation HITAC 8000 Series computers can be equipped with
an Emulator that enables them to execute programs written for a HITAC 3010 system .

.2

DATA STRUCTURE
The basic unit of the HITAC 3010's data structure is the "character", which consists of
six information bits and a parity bit. The six information bits in a character can represent either an alphanumeric character or a portion of a binary field.
Characters can be handled individually or grouped together into variable-length fields.
Instructions, however, have a fixed length of 10 characters. The instructions are of a
two-address form and are performed sequentially .

.3

HARDWARE

. 31

Central Processors and Core Memories
A HITAC 3010 system can be built around any of four types of central processors: Small,
Standard, High-Speed, or Scicntific. The Small and Standard systemsuse central
processors which are functionally the same, but the Small system is physically more
compact. The Standard system is fully compatible with the central processors of the
Small and High-Speed systems. The High-Speed system performs internal operations at
twice the speed of the other systems. The Scientific Processor includes a high-speed
arithmetic unit that performs fixcd- and floating-point arithmetic and address indexing.
Table I summarizes the principal characteristics of all the HITAC 3010 processor models.
There are 41 instructions in the standard HITAC 3010 repertoire. They can be classified
into four general categories: input-output, data handling, arithmetic, and decision and
control.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

5/69

HITACHI HITAC 3010

1555:011.311

TABLE I. HITAC 3010 PROCESSOR CHARACTERISTICS

· 31

Cycle Time,
microseconds
per 2 characters

Model

Memory
Capacity,
characters

Small

H-303Y
H-304Y

10,000
20,000

7
7

41
41

Standard

H-303
H-304
H-305

10,000
20,000
40,000

7
7
7

41
41
41

High-Speed

H-304S
H-305S

20,000
40,000

3.5
3.5

41
41

Scientific

H-355

40,000

7

51

System

Number of
Instructions

Central Processors and Core Memories (Contd.)
The instruction repertoire is enhanced by inclusion of the following specialized
instructions:
•

A Translate by Table instruction that permits efficient internal processing of
various code structures.

•

A Local Symbol instruction that permits an internal search, within core memory, for any character designated.

•

Transfer instructions that permit transfer of data to the left or right within
core memory.

•

A Repeat instruction that permits the next repeatable instructions to be
repeated a specified number of times.

There is no integrated editing facility, but edit routines are rapid and straightforward,
and there is a good variety of instructions for handling variable-length alphanumeric items,
including a convenient code translation operation and Boolean operations. Indirect
addressing is a standard feature, but indexing is provided only in the H-355 Scientific
processor. No program interrupt facility is available.
The arithmetic unit has a suitable speed for general data processing but is slow for
mathematical work. Except in the Scientific system, multiplication, division, and
floating-point operations are available only as subroutines.
· 32

Punched Card Equipment
One card reader and one card punch can be included in a 3010 system. Table II summarizes the characteristics of the available punched card equipment. All four devices
employ automatic code translation, and all but the H-334 Punch can accommodate binary
data (160 6-bit characters per card).

· 33

Paper Tape Equipment
One paper tape reader/punch or one paper tape reader and one paper tape punch can be
included in a 3010 system. All devices employ automatic code translation.
The H-321 Reader/Punch operates at 100 characters per second. Seven-channel tape is
standard, and an optional feature is available to accommodate 5- and 6-channel tape.
The H-322 Reader reads up to 1,000 characters per second from 5-, 6-, or 7-channel
tape.
The H -331 Punch punches 100 5 - or 7 -channel characters per second.

• 34

Optical Character Reader
One H-5S20 Optical Character Reader can be included in a 3010 system. This unit reads,
on demand, at a rate of approximately 1,500 documents per minute. A mark reading
feature is available for the H-5S20.

5/69

A

(Contd. )

AUERBACH
®

SUMMARY REPORT

• 35

1555:011.350

Magnetic Tape Equipment
Six different models of Hitachi magnetic tape units are available for the 3010 system. In
addition, one or two IBM 729 II tape stations can be included. Table III summarizes the
characteristics of the Hitachi units. All Hitachi units provide forward and reverse
reading capability.
TABLE II: PUNCHED CARD EQUIPMENT
Device

Model

Peak Speed,
cards per minute

Feed Type

Reader

H-323

600

continuous, by row

Reader

H-329

1470

demand, by column

Punch

H-334

100

demand, by row

Punch

H-336

200

continuous, by row

TABLE III: MAGNETIC TAPE EQUIPMENT
Tape
Width,
inches

Reel
Length,
feet

Rewind Speed,
inches per
second

Maximum
Number
of Drives

7

1/2

1200

19

12

500

7

1/2

1200

120

12

333

16

3/4

2400

150

14

55,000

550

16

3/4

2400

150

14

H-582

66,667

667

16

3/4

2400

150

14

H-3485

83,000

200/556/800

7

1/2

2400

300

14

Peak Speed,
cbaracters
per second

Density,
cbaracters
per inch

10,000

333

H-382

30,000

H-581

33,333

H-197

Model
H-381

Number
of
Tracks

or
120,000*
*With High-Speed systems only.

The 16-track units provide a dual recording capability that helps to ensure accuracy.
The 16 tracks accommodate two sets of 6 data bits, each set having a parity bit and a
timing bit. All models except the H-381 and H-581 provide read-after-write checking
and automatic rollback on reads. The H-381 and H-382 units include 3, 4, or 6 tape
drives per unit •
• 36

Typewriter Equipment
The typewriter equipment includes the H-338 Monitor Printer and the H-328 Interrogating
Typewriter, either or both of which can be included in a 3010 system.
The Monitor Printer prints at a rate of up to 10 characters per second. The Interrogating
Typewriter allows inquiry or control messages to be entered into the system via a keyboard, and prints data from the computer at a rate of up to 10 characters per second •

• 37

Line Printers
A maximum of two line printers can be included in a 3010 system. Two different models
are available. The H-333 has 120 print positions and prints at a peak speed of 1,000
lines per minute. The H-335 has 160 print positions and prints at a peak speed of 1,070
lines per minute. Both printers offer a print font of 64 characters .

. 38

Auxiliary Storage
The H-366 Disc File is available in four models with data storage capacities of 22, 44, 66,
and 88 million characters. A maximum of two disc files can be used in a HITAC 3010
system. Average random access time is 105 milliseconds, and peal, data transfer rate is
32,000 characters per second.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

5/69

1555:011.390

. 39

HITACHI HITAC 3010

Simultaneous Operations
Without the optional Simultaneous Mode Control, the HlTAC 3010's capability to perform
simultaneous operations is limited to rewinding magnetic tapes, advanCing forms on the
printer, and seeking disc records. The central processor is fully occupied in controlling
all I/O data transfer except during short periods at the end of each transfer load.
When the Simultaneous Mode Control is added, anyone input or output operation can be
overlapped with internal processing. Alternatively, two peripheral units can interleave
transfers of data into or out of core memory while processing is suspended .

.4

SOFTWARE
Software for the HlTAC 3010 is available in three different versions, for installations in
which the program library is held on cards, paper tape, or magnetic tape. In addition to
standard subroutines, assembly routines, and diagnostic routines, there is an elementary
operating system appropriate for this class of computer, and an integrated program testing procedure.
Software available to facilitate programming, program testing, and operation of the
HITAC 3010 system includes the following systems and routines •

• 41

. 42

.43

.5

Language Translators
•

COBOL (based on the COBOL-61 language).

•

FORTRAN (based on the FORTRAN II language).

•

Assembly System with File Control Processor.

•

Report Program Generator .

Service Routines

•

Sort program.

•

Merge program.

•

File maintenance program.

•
•
•
•

Program transcriber.
Consolidata.
Debugging routines.
Peripheral routines.

Application Programs
•

PERT/TIME.

•

Linear programming system.

•

Libraries for scientific and technical computations •

PRICE DATA
Price information was unavailable at press time; price data will be included in
future issues.

5/69

fA

AUERBACH
®

A•

AUERBACH

AUIIII.ACH
COM'UTIlIl
NOTII.OOK
INTllilNATIONAL

111117:011.100
HITACHI HITAC 8000 SERIES
SUMMARY REPORT

SUMMARY REPORT: HITACHI HITAC 8000 SERIES
.1

BACKGROUND
The HITAC 8000 Series is the third-generation family of central processors, peripheral
devices, and supporting software manufactured by Hitachi, Ltd., of Tokyo, Japan. Manufactured under license from the Radio Corporation of America, the HITAC 8000 Series
hardware and software is similar in many respects to the RCA Spectra 70 line.
Noteworthy characteristics of the HITAC 8000 Series include:
•

The high degree of program compatibility, both upward and downward, among
three of the five processor models. Compatibility is also achieved with the IBM
System/360 and RCA Spectra 70 processors through similar hardware design
and compatible source languages.

•

The wide range of input-output and storage devices.

•

The numerous arithmetic modes and data forms, and the resulting complexity
of machine-language coding.

•

The emphasis upon software support through several levels of integrated operating systems.

•

Two disc-oriented operating systems for both small-scale and medium-tolarge-scale system use.

•

The use of true monolithic mtegrated circuits in the 8210, 8300, 8400, and
8500 processors.

•

The availability of optional features that enable certain HITAC 8000 Series processor models to emulate a number of second-generation IBM computers and the
Hitachi HITAC 3010 system.

The HITAC 8000 Series was announced in September 1965, and initial customer deliveries
were made in March 1966. Approximately 300 of the 8000 Series systems have been delivered to date. Hitachi is currently marketing the HITAC 8000 Series in Japan •
.2

DATA STRUCTl:RE
The HIT AC 8000 Series data structure is identical in all respects with that of the IBM
System/360 and the RCA Spectra 70. The basic unit of data storage is the "byte, " which
consists of eight data bits plus (in most system components) one parity bit. The eight
data bits in a byte can represent one alphanumeric character, two decimal digits, or a
portion of a binary field.
Bytes can be handled indiVidually or grouped together into fields. A ''halfword'' is defined
as a group of 2 consecutive bytes, or 16 bits. A "word" in the HITAC 8000 Series is a
group of 4 consecutive bytes, or 32 bits. A "double word" consists of 2 consecutive words,
or 64 bits. The location of any field or group of bytes is specified by the address of its
leftmost byte.
Every fixed-length field (halfword, word, or double word) must be located in main storage
on an "integral boundary;" i. e., the storage address of the field must be a multiple of the
length of the field in bytes.
This restriction is particularly important for efficient operation of the HITAC 8500 processor, which accesses up to four bytes in parallel, and the same restriction has been
applied to the smaller processors in order to maintain compatibility. Variable-length
(decimal) fields are processed serially by byte in all models, so they may start at any
byte location.
At the low end of the HITAC 8000 line of processors, the 8210 and 8200 processors can
perform arithmetic operations on two basic types of operands: fixed-point binary and

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

8/69

1557:011.200
.2

HITACHI HITAC 8000 SERIES

DATA STRUCTURE (Contd.)
variable-length decimal. The larger HITAC 8000 Series processor models can perform
arithmetic operations on four basic types of operands. In addition to fixed-point binary
and variable-length decimal, these models can also perform arithmetic operations on two
sizes of floating-point binary operands. The basic arithmetic operand size is the 32-bit
fixed-point binary word. Most fixed-point instructions can alternatively specify the use
of 16-bit halfword operands.
Floating-point numbers can be represented in either a "short" (32-bit) or "long" (64-bit)
format. The fractional part occupies 24 bits in the short format and 56 bits in the long
fOl'mat. The hexadecimal character o~'!Iupies 7 bits in both formats and permits representation of numbers ranging from 10- to 1075 •
Decimal arithmetk is performed upon 4-bit BCD digits packed two to a byte, with a sign
in the rightmost four bits of the low-order byte. Decimal operands can be up to 16 bytes
(31 digits and sign) in length.
The 8-bit byte structure has certain basic advantages over the 6-bit data format: decimal
digits can be packed more conveniently, the standard 7-bit ISO code and the Extended
BCD Interchange Code can be used, and today's familiar character sets can be convenIently expanded .

.3

HARDWARE

. :11

System Configuration
The HITAC 8000 Series peripheral devices and their controllers are connected to the central processors through input-output channels of various types and capacities. A single
multiplexor channel is provided as standard equipment for the 8210, 8300, 8400 and 8500
systems. Selector channels are provided as standard equipment for the 8200 and 8210,
and as optional equipment for the larger systems. A selector channel provides direct
control of one high-speed input-output operation at a time.
Table 1 shows the various combinations and capacities of multiplexor and selector channels
possible for all HITAC 8000 systems, together with the maximum number of simultaneous
I/O operations per system .

. 32

Internal Storage

.321 !\lain core storage
The range of core storage sizes and speeds available with the various HIT AC 8000 Series
processors is shown in Table II.
The optional storage protection feature can protect the contents of specified 2, 048-byte
blocks of core storage from being altered as a result of program errors or misguided
input data. This feature prevents overwriting by unauthorized programs, but it does not
guarantee privacy since any program can still read the contents of any desired portion of
core storage •
. :122 Auxiliary Storage
Six different models of auxiliary storage devices are available in the form of magnetic
drums, discs, and cards. All except the H-8566 Drum Memory Unit use interchangeable
cartridges. The storage capacities of these devices range from 400,000 bytes for a smallcapacity disc drive to over 530 million bytes for the magnetic-card mass storage unit.
Similarly, average access times range from 8.6 milliseconds to 488 milliseconds.
Table III lists the various HITAC 8000 Series auxiliary storage devices with their principal functional characteristics. The H-8564, H-8568, H-8566, and H-8577 can be used
with the HITAC 8300, 8400, and 8500 processors; the single controller used for all these
storage devices allows an installation to tailor 'its complement of storage devices to satisfy
its specific capacity and access-time requirements, The H-8564-21, H-8564-12, and
H-8564-11 Disc Storage Units can be used only with the HlTAC 8210 processor. No auxiliary storage devices can be used with the HITAC 8200 processor .
. :13

Central Processors
Five processor models currently form the nucleus of the HITAC 8000 Series. Three of
the processor models - the 8300, 8400, and 8500 - are program-compatible and suitable
for a broad range of business and scientific applications. The 8200, with its restricted

,

8/69

£

/A\
AlIlRHAlH

(Contd.)

1557:011.330

SUMMARY REPORT

TABLE I: HITAC 8000 SERIES INPUT-OUTPUT CHANNEL COMBINATIONS
Processor Model

standard
Channel Complement

8200

8210

8300

8400

8500

0

0

0

-

-

-

1

1

1

1

Trunks per channel

6

2

Number of simultaneous
data transfer operations

2

1

-

0

1

1

Selector Channels -

Multiplexor Channels Trunks per channel

-

6

8

9

9

Number of simultaneous
data transfer operations

-

20

192

256

256

Processor Model

Fully Expanded
Channel Complement

8200

8210

8300

8400

8500

1

1

2

3

6

Trunks per channel

6

2

2

2

4

Number of simultaneous
data transfer operations

2

1

2

3

6

1

1

1

1

Selector Channels -

Multiplexor Channels -

0

I

Trunks per channel

-

6

8

9

15

I

Number of simultaneous
data transfer operations

,..

20

192

256

1024

I
I

Combined total of possible
simultaneous data transfer
operations

2

21

194

259

1030

TABLE II: HITAC 8000 SERIES MAIN CORE STORAGE CHARACTERISTICS
Core storage
Capacity, Bytes

I

.33

8210

8200

-

4,096
8,192
16,384
24,576
32,768
65,536
131,072
196,608
262,1444
393,216
524,288

8200-4
8200-8
8200-16

Cycle Time, Ilsec

2.0

1.4

Bytes Accessed
per Cycle

1

Effective Cycle
Time per Byte, IlSeC

2.0

----

8210-8
8210-16
8210-24
8210-32

----

8300

8400

8500

-

---

--

8300-16
--

8300-32
8300-65

-

-

8400-65
8400-131
8400-196
8400-262

8500-65
8500-131

1.44

1.44

0.84

1

2

2

4

1.4

0.72

0.72

0.21

--

---

--

-

8500-262
8500-393
8500-524

Central Processors (Contd.)
instruction repertoire, Is best suited for use as a satellite system for the larger HITAC
8000 Series processors. The 8210 also has a limited instruction repertoire, but its expanded throughput capability for disc-oriented applications makes it suitable for certain
single-processor installations.
Comparative arithmetic execution times for the various HITAC 8000 processors are illustrated in Table IV.
© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

8/69

1557:011.331

. 33

HITACHI HITAC 8000 SERIES

Central Processors (Contd. )
The remainder of this discussion of central processors concerns itself exclusively with the
HITAC 8300, 8400, and 8500 models. These processors are designed to be compatible with
the IBM System/360 computer. They offer the full System/360 instruction repertOire except for the "Test and Set" and "Privileged" instructions, which are normally reserved
for operating system use and are not used in users' programs. Thus, HITACHI expects
to achieve two-way System/:lGO program compatibility - to a limited extent at the machinelanguage level and to a much greater extent at the assembly, COBOL, and FORTRAN
language levels.
The HITAC 8000 Series processors contain facilities for addressing main storage, for
fetching and storing lnfonnation, for executing stored-program instructions in the desired
order, for arithmetic and logical processing of data, and for initiating all communication
between main storage and peripheral devices. Each program uses sixteen 32-bit general
registers and four 64-bit floating-point accumulators or as index registers. These registers are contained in a 128-word scratchpad memory in the HITAC 8400 and 8500 processors. In lieu of scratchpad memory, the HITAC 8300 processor provides 128 words of
additional core storage for use equivalent to that of scratch-pad memory. Different parts
of these memory units are used as the operational registers, depending upon which processor state is being used.
Instructions can be two, four, or six bytes in length. Main storage addresses are formed
by adding a 12-bit "displacement" contained in the instruction to a 24-bit ''base address"
contained in one of the 16 general registers. The addresses in many instructions (including most binary arithmetic and logical instructions) can be further modified by adding
a 24-bit "index" contained in another general register; this effectively provides a double
indexing capability. The base-register technique of address formation facilitates program
relocation and segmentation, at the expense of increased programming complexity.

TABLE HI: HITAC 8000 SERIES AUXILIARY STORAGE UNITS

Device

C apaci ty Range,
(millions of
bytes per
control unit)

Average
Access Time.
msec

Data
Transfer
Rate
bytes/sec

8.6

210,000

H-83GG Drum Memory
Unit

1. G to G.-!

H-8577 Mass Storage
l'nit

233.4

87.5

312,000

H-85G4 Disc Storage
Unit

7.25 to 58.0

87.5

156,000

H-83G8 Mass Storage
l'nit

53G.9 to 4,295

H-83G4-12 Disc Storage
l1nit*

2.5G to 5.12

72.5

156,000

H-8564-11 Disc Storage
l'nit*

5.12 to 10.24

87.5

156,000

H-8564-21 Disc Storage
l'nit*

5.12

87.5

156,000

488

70,000

'l'sable only with the HITAC 8210 processor.
The basic arithmetic mode of the HIT AC 8300, 8400, and 8500 processors is fixed-pOint
binary, using 32-bit operands and two's complement notation. Most instructions can alternati\'ely specify the llse of lG-bit "halfword" operands to improve storage utilization.
l\!ost products and all dividends are 64 bits long. Fixed-point arithmetic and comparison
instructions specify one operand in a general register and a second operand in either main
storage or a general register; these instructions are four bytes long when they specify an
operand address in main storage and two bytes long when both operands are in registers.
The System/360-compatible instruction set includes instructions which perform fixedpoint arithmetic, comparison, branching, moving, loading, storing, shifting, radix

8/69

fA.

Alit RHACH

(Contd. )

SUMMARY REPORT

. 33

1557:011.340

Central Processors (Contd. )
conversion. code translation, packing, unpacking, and Boolean operations. The radl.x
conversion operations perform automatic conversions between signed. packed decimal
fields up to 15 digits in length and 32-bit signed binary integers. The code translation
instruction uses a table to translate any 8-bit data code to any other 8-bit code. The
packing and unpacking instructions convert numeric BCD data between the one-characterper-byte format used in most input-output devices and the two-digits-per-byte format
used for decimal arithmetic.
The decimal arithmetic facility provides additional instructions for addition, subtraetion,
multiplication, division, comparison, and editing of decimal numbers. Decimal arithmetic
is performed upon 4-bit BCD digits packed two to a byte, with a sign in the rightmost four
bits of the low-order byte. Decimal operands may be up to 16 bytes (31 digits and sign)
in length. The length of each deeimal field is specified in the instruction referencing it.
Two-address (6-byte) instructions of the storage-to-storage type are used for all decimal
operations; the general and floating-point registers are not utilized.
The floating-point arithmetic facility provides additional instructions for addition, subtraction, multiplication, division, loading, storing, and comparison of both "short" (32bit) and "long" (64-bit) floating-point numbers. Floating-point instructions specify one
operand in a floating-point register and a second operand in either main storage or a
floating-point register.
Following are the five basic instruction formats:
•

Type RR - Register to Register (2 bytes)

I

Op

RI

R2

•

Type RX - Register to Indexed storage (4 bytes)

•

Type RS - Register to Storage (4 bytes)

•

Type SI - storage and Immediate Operand (4 bytes)

•

Type SS - Storage to Storage (6 bytes)
Op

Ll
B
D
I
L
Op
R
X

.34

I

I I
L2

Bl

4-bit base register specification
12-bit displacement
8-bit literal operand
8-bit operand length specification
8-bit operation code
4-bit operand register specification
4-bit index register specification

Input-Output Equipment
A fairly wide range of conventional input-output units is available for the HITAC 8000 Series
computer family. These units and their capabilities are summarized in Table V .

. 35

Data Communications Equipment
The Communications Controller Multichannel (CCM), usable with the 8300, 8400, and
8500 processors, terminates from 1 to 48 communications lines serving a wide variety of
remote terminals. Each of the 1 to 48 scan positions requires a communications buffer,
and in some cases a data set, to interface with the communications line. The CCM is
connected to a HITAC 8300, 8400, or 8500 computer by one trunk of the multiplexor channel. Each scan position of a CCM uses one multiplexor subchannel.
© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

8/69

1557 :011.360

HITACHI HITAC 8000 SERIES

TABLE IV; ARITHMETIC EXECUTION TIMES FOR THE HITAC 8000 SERIES PROCESSORS
Processor Model
Task

8200

8210

8300

8400

8500

Fixed Point Binary
c a+b
c -. ab
c ~ u/b

88

65.88

#
#

'*
'*

51. 36
140.64
232.44

25.20
81. 96
111. 21

4.48
9.41
14.18

Fixed Point Decimal
c - a +b
c - ab
c - u/b

88

63.12
340.15
731. 16

35.28
119.9
203.03

15.64
31. 32
45.93

'*
'*

'*
'*
'*

79.17
182.08
391.52

37.44
67.66
101.24

7.10
11.62
15.02

#
#
#

#

'#*

116.87
465.09
1218.21

52.65
211.51
305.23

9.69
24.49
40.85

Floating Point c a+b
c = ab
c'" alb

63.51
415.6
647.8

Short

Floating Point - Long
c -- a + b
c = ab
c = alb
NOTE:

#
#

#

All times are expressed in microseconds. The fixed-point decimal times
are based on 5-digit (3-i.Jyte) decimal operands. The floating-point times
are based on both the short-form (32 bits) and the long-form (64 bits)
binary operands. The 8200 and 8210 do not require programmer-initiated
operand movement to a fixed accumulator register. The symbol "#"
indicates that the facility is not available.
TABLE V: HITAC 8000 SERIES INPUT-OUTPUT UNITS
Peak Speed

Unit

I

Processor Models
to Which Unit Can
be Connected

400/91
400
91
750
1,470
100
250
500
110
200/100
200/100
1. 000/100
300

cpm
cpm
cpm
cpm
cpm
cpm
cpm
cps
cps
cps
cps
cps
lpm

8200
8200
8200
8200
8200
8200
8200
8210
8210
8210
8200
8200
8210

l! -8243 Printer

600

lpm

8200 through 8500

11-8246 Printer

1,250

lpm

8200 through 8500

H-8239-11 Card Reader/Punch
H-8239-21 Card Reuder
H -8239 -31 Card Punch
H-8233 Card Reader
H -8238 Card Reader
H-8234 Card Punch
H-8235 Curd Punch
H-8226-1 Paper Tape Reader
H-8227-1 Paper Tape Punch
H-8229-22 Paper Tape Reader/Punch
H-8221 Paper Tape Reader/Punch
H-S222 Paper Tape Reader/Punch
11-8244 Printer

9 -track :\Iagnetic Tape Units
(11-8432; 8442/ 8445)-800 bpi density
9 -track l\lagnetic Tape Units
(11-8431/8433)-1600 bpi density

.:Hi

30,000, 60,000
120,000 byte/sec.
60,000,
120,000 byte/sec.

through
through
through
through
through
through
through
only
only
only
through
through
only

Comments

8500
8500
8500
8500
8500
8500
8500

8500
8500

8200 through 8500
8200 through 8500

63 or 110 character
print drum
63 or 110 character
print drum
63 or 110 character
print drum
7-track adapters
available
800 bpi adapters
available

Simultaneous Operations
A HITAC 8000 Series central processor (except for models 8200 and 8210) can concurrently execute:
•
•
•

One machine instruction;
Vp to eight high-speed input-output operations (one per selector channel); and
Multiple slower input-output operations via a multiplexor channel.

fA

All! IIBACH

(Contd. )

SUMMARY REPORT

.36

1557:011.361

Simultaneous Operations (eontd. )
Table I shows the mix possibilities and simultaneous operations capabilities of the various
HIT AC 8000 Series input-output channels.
In general, the relationships between HITAC 8000 Series peripheral devices and inputoutput data channels are determined at installation time and cannot be altered under pro"ran1 control except by the inclusion of special optional features. Since it is not normally
possible to a.<;sign by program any free channel to any available peripheral device, the
number of input-output operations that can actually occur simultaneously may in many
eases be considerably fewer than the theoretical maximum. However, special features
are available to switch a limited number of devices to free data channels under program
control.

.·1

SOFTWAHE
----Seven levels of programming systems and operating systems are supplied by Hitachi for the
HI'TAC 8000 Series: 8200 Programming System. 8210 Disc-Oriented Programming System
(DOPS). Primary Operating System (POS). Tape Operating System (TOS). Tape/Disc Operating System (TDOS). Disc Operating System (DOS). and Extended Disc Operating System
(EDOS). The 8200 programming system. DOPS. TOS. TDOS. and DOS are already in use.
and EDOS is scheduled to be released during 1970. Since the original announcement of the
HITAC 8000 Series. Hitachi has greatly improved its standard software by adding "thirdgeneration" software facilities such as disc-oriented control systems. disc file language facilities. automatic on -line file management techniques. and comprehensive data communi('ations control routines.
POS, 1'08, TD08. DOS and EDOS are applicable to HITAC 8300. 8400 and 8500 systems.
POS requires a minimum of 32.768 bytes of core memory; and TOS. TDOS. DOS and EDOS
require at least 65.536 bytes of core memory. Multiprogramming control for up to six jobs
is provided for HITAC 8000 Series systems that have at least 65.536 bytes of core memory.
DOS-II is an enhanced version of DOS that provides improved throughput capabilities especially for large-scale. on-line real-time use .

. 41

Programming System
The 8200 programming system provides an assembler. RPG. COBOL. FORTRAN. tape
sort/merge and a set of service routines. This software system makes the 8200 suitable
for magnetic-tape-oriented data processing applications and for use as a satellite with
larger HITAC 8000 Series systems .

.p

Disc-Oriented Programming System (DOPS)
The 8210 Disc-Oriented Programming System provides an assembler. RPG. COBOL.
FORTRAN. POPS. tape sort/merge. disc sort/merge. control programs. and a set of service routines. Control programs consist of master control. program loader. physical IOCS
and logical IOCS. A minimum of 8192 bytes of core memory and one disc drive are required
for use of DOPS. Multiprogramming control for up to two jobs is provided for HITAC 8210
system and this requires at least 24.576 bytes of core memory.
POPS (Problem Oriented Programming System) is characterized by its capability for on-line
updating of multiple files based on simulations data input from several local typewriters. A
batch program and a real-time program are executed simultaneously by a roll-in/roll-out
controller included in the POPS control .

. 43

Primary Operating System (POS)
The Primary Operating System. for use with the HITAC 8300. 8400. and 8500 systems. is
a magnetic-tape-oriented software system that provides basic supervisory control for the
sequential execution of programs. interrupt control, and input-output control. as well as a
COBOL compiler. FORTRAN compHer. assembler. report program generator. and stanc1ard utility routines.
POS COBOL is a subset language of full COBOL 65. and POS FORTRAN is a subset of
FORTRAN IV. Both require 32.768 bytes of core storage and four magnetic tape units. No
routines are supplied for the automatic control of these devices. They can be programmed
in assembly language. The only forms of multiprogramming supported by POS and the
HITACHI-provided Peripheral Control Routine. which permits concurrent operation of up
to three data transcription routines.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

8/69

1557:011.440

. 44

HITACHI HITAC 8000 SERIES

Tape Operating System (TOS)
The second major level of HITAC 8000 Series software support for the 8300, 8400, and 8500
systems is designated the Tape Operating System (TOS). TOS is a magnetic-tape-orlented
integrated software package that provides supervisory control programs, language processors, and utility programs for installations that have a minimum hardware configuration of
65,536 bytes of core storage, five magnetic tape units, console typewriter, card reader,
and Une printer. The facility to control multiprogrammed operation of up to six programs
concurrently is the primary feature of TOS software.
The basic TOS executive program requires a minimum of 16KB of core storage. The
Monitor program that coordinates the operations of stacked-job processing requires an
additional 4KB, and the File Control Processor for input-output device and file control
requires another 4KB to 8KB of core storage.
Although the theoretical maximum number of problem and control programs that can be
processed concurrently is six, the actual limit will frequently be fewer than six, limited
by the amount of available core storage and the number of available peripheral devices.
As many as five magnetic tape units can be dedicated to system control and library functions when proceSSing in a stacked-job multiprogramming environment.
In addition to a comprehensive assembly system, TOS offers a COBOL language similar
to IBM's Operating System/360 COBOL F, as well as full-scale FORTRAN IV language
that includes all Operating System/360 FORTRAN IV facilities .

. 45

Tape/DiSC Operating System (TDOS)
Hitachi's Tape/DiSC Operating System (TDOS) is an improved and extended version of its
Tape Operating System (TOS). In addition to all TOS software facilities, TDOS offers
options that permit system control routines, problem programs, and library subroutines
to reside on either the H-8564 disc or H-8566 drum unit in order to improve the system
throughput capabilities. As a result, TOOS offers more efficient multiprogrammed operations than does the tape-oriented Tape Operating System. Also, with TDOS the number
of Job Control Language statements required to prepare and compile object programs is
reduced, increasing the efficiency of program preparation.
Another significant addition to the TOOS software package is a comprehensive set of
input-output routines for control of data communication devices. This communic/itions
package, called the Multichannel Communication Program (MCP) offers most of the same
facilities as the IBM System/360 Queued Telecommunications control. The MCP system
can accept remote messages either as they are entered or as polled, in contrast to thfil
polled-only acceptance technique of QTAM.
The minimum HITAC 8000 Series core storage requirements for use of TOOS remains at
65KB, of which 17KB is permanently reserved for the Executive. Both the Executive
Monitor and the optional data communications package require 4KB of storage. The
Monitor, however, is a transient routine and does not require permanent residence in core;
the data communications control routines are permanently resident .

. 46

Disc Operating System (DOS)
The Disc Operating System (OOS) is an improved and extended version of the Tape/DiSC
Operating System (TDOS). In addition to providing aU mos software facilities, DOS is a
disc-oriented operating system designed mainly to:
•

Improve throughput capabilities in on-line,

•

Reduce operator intervention in handling magnetic tape; and

•

Reduce the system use of magnetic tape units.

DOS includes several functional additions to TOOS: catalogued procedures, roll-in/rollout, disc seek prefetch function, device interchangeability between low-speed rio or
magnetic tape units and disc drives, and a disc sort/merge generator. A catalogued
procedure permits the cataloguing and retrieval of a job stream or a set of run-time
control parameters from a disc file. Roll-in/roll-out permits temporary storage of
lower-priority programs in a disc area •. For stacked-job processing controlled by the
Monitor, the device interchangeability function can be extended to higher-level languages
such as COBOL and FORTRAN. The minimum core storage requirement for use of OOS
remains at 65KB, of which 19KB are permanently reserved for the Executive.

8/69

A

AlIlRRACH

(Contd. )

1557:011.460

SUMMARY REPORT

. 46

Disc Operating System (DOS) (Contd.)
Throughput in on-line, real-time applications can be improved by using oos-n, which can
control up to 15 users' tasks. The multitask control function Is designed to enable a
specific user program to process a maximum quantity of input data per unit time interval.
OOS-U is an improved version of the OOS Executive and FOP, and It uses the same lanlanguage processors, utility routines, library maintenance routines, and initializers as
DOS. FCP is a re-enterable routine, permitting simultaneous use of one copy of the
program by a number of tasks. Although OOS-U is designed for use with as little as 65KB
of core storage, more storage is recommended for efficient use of the multitask control
feature .

. 47

Extended Disc Operating System (EDOS)
The Extended Disc Operating System is an enhanced and extended version of the Disc Operating System (DOS). In addition to all the DOS software facilities, EDOS offers the following:
•

Multi Job Stream
Three batch programs or two batch programs and one on-line program can be
operated under multiprogramming with all the batch programs executed under
job stream.

•

Input Job Stack
Multiprogramming can be carried out effectively as all jobs are executed after
being stacked in discs or drums.

•

Output Job Stack
Results of three multi job streams are temporarily stored in discs or drums and
then output to line printer or card equipment according to priority.

•

Job schedule
Jobs can be executed according to a sequential job schedule and also according to
a priority job schedule.

•

Job accounting
Logging facilities of job accounting, such as processor time of a job, input-output
device usage time, and the like are provided.

The basic control program of EDOS requires 10,240 bytes of memory. EDOS is compatible
with TOS, TDOS and DOS and fully compatible with the latter two on an object level .
. .)

COMPA TlBILITY

.31

Program Compatibility Within the HITAC 8000 Line
There is a high degree of program compatibility, in both the upward and downward directions, among the HITAC 8300, 8400, and 8500 models. Among these three models, any
valid program that runs on configuration A will run on configuration B and produce the
same results if:
•

Configuration B includes the required amount of main storage, the same or
compatible input-output devices, and all required special features; and

•

The program is independent of the relationships between instruction execution
times and input-output rates.

These limitations indicate that there will be a high degree of effective upward compatibility,
making it easy to expand an installation, but that the concept of downward compatibility will
be useful mainly in making pos'sible the common use of subroutines and software, rather
than in making it feasible to "shrink" an installation as its workload decreases or to back
up a large computer with a smaller one.
Although the HITAC 8200/8210 instruction repertoire is a subset of that for the iarger
processors (except for input-output instructions and address modification instructions),
the different software system designs restrict the compatibility of the 8200 and 8210 with
the larger 8000 Series systems. Some program conversion is required before an assemblylanguage source program written for an 8200 or 8210 can be run on an 8300. 8400. or 8500
System.
11;',

1969 AUERBACH Corporation and AUERBACH Info, Inc.

8/69

1557:011.520

.52

HITACHI HITAC 8000 SERIES

Program Compatibility with the IBM System/3GO
Hitachi provides, through its HITAC 8000 Series source languages, program compatibility
with the IBM System/3GO. The HITAC 8000 COBOL and FORTRAN languages are in many
cases identical with their System/3GO counterparts. Furthermore, since the instruction
repertoire of the large HITAC 8000 Series processors is virtually identical with that of the
corresponding IBM processors, Hitachi has been able to develop System/360 program
compatibility at the assembly-language level as well. The differences in the "privileged"
instructions, however, make it impossible to run machine-language System/3GO on a HITAC
8000 system, program recompilation or reassembly is always required. In many cases,
the System/3GO operational control cards can be retained and used directly in the HITAC
8000 program input stream.
When the HITAC 8000 software was first delivered during 1966, the lack of sourcelanguage-level support of random-access storage devices was a Significant limitation in
comparison to IBM System/360 software. Hitachi has since developed COBOL and RPG
software with capabilities for directly controlling random-access devices. A FORTRAN
compiler with similar capabilities is presently under development •

. 53

Program Compatibility with the HITAC 3010 System
The optional 3010 Emulator feature offers a high degree of compatibility between the
HITAC 8000 Series and Hitachi's second-generation HITAC 3010 system .

. 5-1

Program Compatibility with Second-Generation IBM Computers
HITACHI offers a series of Emulator features that enable certain models of the HITAC
8000 Series to run object programs written for certain second-generation IBM computer
systems. The IBM computers whose programs can be run by the various HITAC 8000
systems (when properly equipped) are as follows:
HITAC 8000 System

Systems Emulated

Release Dates

HITAC 8300

IBM 1401/1460

now in use

HITAC 8400

IBM 1401/1460

now in use

IBM 1410/7010

now in use

Emulation, in general, requires a HITAC 8000 system with an equivalent array of peripheral equipment, more proceSSing power, and more core storage than the secondgeneration system to be emulated. The functions of most of the common peripheral
devices (e. g •• card readers, card punches, printers, magnetic tape units, and console
typewriters) can be emulated, but the less common devices (e. g •• optical and magnetic
character readers, paper tape units, data communications devices, and random-access
storage devices) cannot be emulated. Time-dependent programs and programs not written
in accordance with IBM programming manuals, when emulated, may yield results which
differ from those obtained in the original system; the handling of many console operations
and error conditions will differ; and a variety of specific program restrictions and limitations apply to each Emulator feature. Nevertheless, most users of second-generation IBM
computers will be able to run most of their programs on a HlTAC 8000 system with little
or no need for immediate reprogramming.
The principal value of the Emulator features is that they enable users of second-generation
IBM computers to spread the task of reprogramming for the HITAC 8000 system over an
extended period of time. In nearly every case, the emulation mode involves additional
equipment costs and falls short of fully utilizing the performance capabilities of the HITAC
8000 system. Therefore, for maximum efficiency, most users will want to recode all of
their prinCipal applications for the HITAC 8000 system as soon as possible. The cost of
the additional core storage and of the optional features required for emulation must be
borne until all of the user's programs have been recorded •
.G

PRICE DATA
Pricing information was unavailable at press time; price data will be included in future
issues.

8/69

A

AUlRHALH

fA
~

AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

1575:011.100
NEAC SERIES 2200
SUMMARY REPORT

AUERBACH

'"
SUMMARY REPORT:
.1

NIPPON ELECTRIC NEAC-SERIES 2200

BACKGROUND
The Nippon Electric Company (NEC) was originally founded as a Japanese manufacturer of
telephone equipment, and has since extended its interests to cover a wide range of communications and other electronic equipment. Nippon first entered the computer field in 1957,
and currently claims to have the largest share of the Japanese domestic computer market
of any single manufacturer. Few Nippon computers are sold outside Japan.
Nippon has a licensing agreement with Honeywell EDP in the United States. Software development by the two companies tends to be a joint effort, while the hardware developments
of the two companies are complementary, with each company benefitting from the work of
the other.
The NEAC-Series 2200 is similar to the Honeywell Series 200. The NEAC-Series 2200/50
was developed by Nippon and is similar to the Honeywell 110, the smallest member of the
Honeywell Series 200 - since it has different peripherals from the other machines in the
family. it is discussed separately in Report 1576. The NEAC-Series 2200 models discussed
here are the Models 100, 200, 300, 400 and 500, of which the Models 100 through 400 correspond to the Honeywell 120 through 2200. The NEAC-Series 2200/500 has roughly the
same capabilities as the Honeywell 3200. Recently, Nippon announced the NEAC -Series
2200/700, which will be discussed in a future Report.
Nippon tends to act as a supplier of complete systems, with a heavy emphasis on communications. At their Tokyo data center. Nippon has NEAC-Series 2200/50, /100, /400 and
/500 - a NEAC-Series 2200/500 is being used for time-sharing at Osaka University.
Noteworthy characteristics of the NEAC-Series 2200 include:
•

The high degree of program compatibility, both upward and downward, among
all processor models. NEAC-Series 2200 processors are designed according
to the "Liberator" concept, which allows the users of various competitive
systems, e.g., the IBM i400 series, to use automated, one time translation of
programs for execution ot higher-performance NEAC-Series 2200 systems.

•

The wide range of input-output and storage devices.

•

The emphasis upon software support through several levels of integrated
operating systems.

•

The use of true monolithic equipment with a "storage protect" capability which
shields the contents of one or more designated memory areas against accidental
alteration by unrelated programs.

NEAC-Serics 2200/500 Installation.
© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

8/69

1575:011.101

.1

NEAC SERIES 2200

BACKGROUND (Contd.)
Information is stored in main memory locations either in pure binary form as 6-bit alphanumeric characters, or as signed decimal quantities. Parity checking in each character
position is automatic: Any number of consecutive locations can be grouped to form
fields; groups of consecutive fields can be delineated as items. Fields and items are
defined, respectively, by word marks which resemble those used in the IBM 1400 Series.
There are no reserved input/output areas in main memory - this allows both a high
dt'gree of programming flexibility and economical uRag-e of main memory •

.,

NEAC-Series 2200 processors have available a repertoire of instructions which, with
tremendous flexibility and power derived from the usc of variant characters, can handle
arithmetic, logical, control, editing, and input/output functions necessary for business
data processing. Also iMluded in all processors arc instructions for handling peripheral
and communication interrupts and for manipulating data in codes of up to 12 levels.
The NEAC-Series 2200 is based upon an improved version of the original NEAC-2200
svstem which was first delivered in 1964. The NEAC-Series 2200 was first announced
hi l\Iay 1965. The 400 processor was delivered in October 1966 and the 100 and 500 in
November 1966. Delivery of the first 300 system was in 1967.
NEAC-Series 2200 processors contain high speed memories, modular memory capacities,
powerful instructions, efficient addressing methods and suitable input/output facilities to
afford Simultaneity of operations, as well as high levC'l computing power. All processors
are equipped with:
•

Direct, indexed, and indirect addressing

•

2 to 4 character address interpretation

•

Program assignable read/write channels

•

Automatic program interrupt

•

Multi-level code handling facility

:-mAC-Series 2200 systems rentals range from approximately $800 to $55,500 per month.

Five central processors currentl~' form the nucleus of the NEAC-Series 2200. It is
considered that these processors Models 100, 200, 300, 400, and 500 span a range
equivalent to that spanned by the IBM System/360, Models 20 through 60. Table I summarizes the basic characteristics and capabilities of the NEAC-Series 2200. Comparative arithmetic times for the various Series 2200 processors are illustrated in Table II.
Instructions are variable in length. The basic instruction format consists of an operation
code which specifies the type of operation to be performed, two operand fields which
specif~' the binary addresses of fields to be used in the operation and a variant character.
The \'ariants are used to expand the meaning of the operation code or to specify literally
a piece of data to be used in the operation.
:\Iultiply and divide operations are standard in all processors except Model 100. Models
300 and 400 can be equipped with a floating point arithmetic facility for use in scientific
applications: floating point operations are standard on the Model 500 •
. 22

Working storage
;\Iain memory cycle times range from 2 microseconds to 188 nanoseconds per 6 bit
character and control memory cycle times from 500 nanoseconds to 188 nanoseconds.
The speeds of KEAC-Series 2200 memories are complemented by the wide range of
storage capacities available at each speed level. Memory size in the Model 100 processor ranges from 2,048 to 32,768 six bit cbaracters. At the other end of the scale, the
:\todel 500 processor is available with 65,536 to 524,288 character memories. The
modularity of the NEAC system is exemplified by the relatively small increments in
which main memory can be expanded.

3/69

fA

AlJlRAALH

(Contd. )

SUMMARY REPORT

1575:011.220

Working storage (Contd.)
All NEAC-Series 2200 main memory locations are directly addressable. Three additional features facilitate advanced programming and addressing of large memories indexed and indirect addressing and variable-length address interpretation.
Six main memory index registers are provided in configurations having 32,768 or fewer
storage locations; larger memories are equipped with 30 index registers. Indirect
addressing enables the user to reference stored information via one or more intermediaryaddresses. Variable length address interpretation allows instructions of either
two-character, three-character or four-character addresses.

NEAC-Series 2200 processors possess several features which enable them to provide
simultaneity which is powerful but easy to use; program assignable read/write channels,
multiple input/output trunks, and an Interrupt processing facility.
The use of program-assignable read/write channels enables NEAC-Series 2200 processors to compute while concurrently servicing from 3 (Model 100) up to 16 (Model 500)
input/output operations. In addition, Series 2200 processors provide facilities for
interfacing with a large number of peripheral controls, ranging from a possible 11, in
the case of the Model 100, up to 64 in the Model 500. The high internal speeds of these
processors insure that when simultaneity is fully exploited, the demands to service
peripheral devices will still be satisfied. That this capability does not depend upon complex software or expanded system configurations is more significant.
The basic Model 100 processor is equipped with integrated controls for a 500 line per
minute printer, a 400 card per minute card reader, a punch which processes cards at
rates from 100 to 400 per minute, a 300 frame per second paper tape reader, and a
paper tape punch which punches at the rate of 60 frames per second. The card devices
can be separate or combined as a reader-punch. Up to four non-simultaneous magnetic
tape units with a tape speed of 24 inches per second can be connected to the basic Model
100 processor. All other peripheral controls are connected to NEAC-8eries processors
"ia input/output trunks. The number of such trunks available in a processor ranges
from 11 in the Model 100 to 64 in the Model 500. A control unit which handles both
reading and writing (for example, a magnetic tape control) connects to a pair of trunks.
The number of peripheral controls possible in a system depends on the number of input/
output trunks available.
TABLE I: PRINCIPAL CHARACTERISTICS OF NEAC-SERIES 2200 PROCESSORS
Prlll'l'8~\lr

:\lalO

:\lo(It.-'1

~1('mof\

:\Icmnr\'
CapaC'tt\'

n housand.

Speed
JCH:\(, '1'1111('1

or

~umber of
Input !Oulpul
Trunka

Number of

I/O Operations
Simultaneous

Charactersl

Advanced

t'inanclal

Edit
Programming
InstructIons Instruchon

Multiply
and Divide
InstructIons

Scientific
Processing
Instructions

Memory

Protect
Faclhty

WIth

Computtng
IOQ

:: j.lSCl'!>. pl'r

or

2 to :12

is tu 11

2

:l

Optional

Optional

-\ til I;;)

H tn Hi

:i or-l

Optional

Optional

Standard

Sandard

Standard

Standard

Optional

OptIOnal

OptIOnal

Optional

l'harol'tt'r

21111

:!~,fWCfi.. pt'r

l'hara('tl'r
.!lIO

40t)

1. ,j .IS(>l'S. pcr
('hara,('\cr

Iii to 1:n

1 'j i'if'C, per

U; tn 21i2

li; to 32

,1 to M

Standard

standard

Standard

ti3 to 324

:t!

~

to Hi

Standard

Standard

Standard

16

{'har3('tl.'r
jlW

I,:J •. sees
~

per

('hara('tcrs

tf}

1).1

Standard

Optional

'I (' .....lfled m.m-

Pro"rammed

off_,...

ory ar. on error

Opt'lraUOnl

PUnob data from lpeel'led memory area

OIrd.
"'ck
cardao. . . luoy.

Punch data from lpeclfJed memory area, read
data and tranlf.r to
. _"led memory , _ .
rMd/puncb lame card

Da'" Tn"'er

4utomaUc tnnalation between HoUerith card and 6-blt ceatr.l proc...or code Ja standard. AddItiona. tranacrjpUon
modo
and ...ncb... capa.lty ,100 a .. llable.

Dall

lIIeplpu... b
.book

'_111

Modo"

ProtecUOIl

. ' Hoppor/OUIpu'
ltaokor capacity. OIrda

cycle oMcIk

lIIepljIUII.b
cbock
cycle oheck

:1000/21100

3000/1100

·TnoacrlpllcD mnda reodllll and

.42

_bIIII no'a .. llabJ.

Pu..,h cbock

Hole -('OWlt check

lIIopl punoh obock
on Reodll1ll ..... h
cback on puncblq

1100/1300

Ia MODEL 100 pro•• aaor·.

looo/noo
In~rllnd

1200/1300

.Ird .",Ipm.. ' con'rol •

Printers
As indicated in Table IV, printing speeds offered range from 420 to 950 single-spaced
lines per minute; 120 or 132 print positions per line are available.
Peripheral data transfer instructions are used to operate the printer, while error and
busy status is sensed through a peripheral control and branch instruction. A print roll
on which characters are embossed moves pass print hammers during the printing cycle.
A cyclic check is made for accuracy .

. -!3

Magnetic Tape Equipment
Two complete families of magnetic tape units are available for use in NEAC-Series 2200
systems:
•

Units which process 1/2 inch tape provide the standard means for storing 6bit data and IBM compatibility, including end-of-file mark recognition and the
ability to translate from card images in IBM even-parity tape code to NEAC
Series 2200 processor code. These units read forward, write forward, backspace one record, space forward one record, rewind, rewind and release and
erase; backward read Is also available.

•

Vnits which process 3/4 inch tape provide data compatibility with NEAC 2400/3400/2800/3800 systems and, in addition, feature a unique Orthotronic
control technique for data checking and regeneration. These units read forward, write forward, backspace one record, rewind and release, and
regenerate tape channel.

As indicated in Table \', data transfer speeds range from 7.200 to 96,000 characters
per second for units processing 1/2 inch tape, and from 32,000 to 88,000 characters
per second for units proceSSing 3/4 inch tape.
In the 1/2 inch tape units reading and writing can proceed simultaneously under the
direction of a single tape control unit at the same time that computing is in progress
(24 inch per second drives do not have this facility).
The vacuum technique has been incorporated in the design of all NEAC tape units. This
technique controls the mounting, driving and stopping of tape. A write enable ring and a
manual tape unit switch are used to guard against destructive unintentional write attempts.
All information written on 1/2 inch tape is immediately read and checked. During a
write operation, a parity bit is generated for each frame and also each data channel.
These bits are checked during reading. Failure of any of these checks causes a program
accessible indicator to be set. The 3/4 inch tape equipment has the further ability to
regenerate any tape channel on the basis of the parity established by the other channels
and the frame parity bits.
Table \' summarizes the characteristics of the available magnetic tape equipment.

8/69

A

AlJfRBAlH

(Contd.)

SUMMARY REPORT

1575:011.440

TABLE IV: SUMMARY OF PRINTER CHARACTERISTICS
MODEL

N222-4

N122A-l

N206A-l

950 (46 char. set);

500 (61 char. set);

500 (61 char. set);

750 (63 char. set)

420 (109 char. set)

420 (109 char. set)

Model 100: 19
Model 500: 3.6

. Model 100: 18.4
Model 500: 3.5

18.1

Print 5}leed, lines/min
Demand on Central
Processor, %
Print pos I tions per line

120 or 132
10 numeric. 26
alphabetic, and
27 special; or
48 Kana, and
5 special

Character Set

Skip speed inches/sec.

10 numeric
26 alphabetic,
48 Kana, and
25 special

35

10 numeric,
26 alphabetic,
48 Kana, and
25 speCial

20

Vertical specing

20

6 or 8 lines per inch

There are three models of paper tape readers, NI09A-l, N209A-l and N209A-2, which
process 5 through 8 level tapes at the rate of 300, 300 and 1000 frames per second,
rcspecti vely.
Four paper tape punches are available, N110A-l, N2l0A-l. N2l0A-3 and N110A-3.
These punch the 5 through 8 levels of tape at the rate of 50, 60, 110 and 110 frames per
second, respectively.
Frame parity is generated by programmed instruction during punching and checked in
reading. The reader can be equipped to check each frame for even or odd parity.
Table VI summarizes the characteristics of the available paper tape equipment .

. 51

Random Access Drum File and Control
The ;\Iodel N271A drum file and control offers high speed access to large quantities of
stored data. l'p to 8 drum storage units (with a total capacity of 2.6 million characters)
can be connected in a non-simultaneous manner to one drum control operating on-line
with a NEAC-Series 2200 System.
The drum, the power supply and a read/write rack are housed in a single cabinet. The
drum control is contained in one central processor logic drawer.
;\Iultiple subsystems, each composed of a drum control which requires two input/output
trunks and from one to eight drum units can be connected to a single computer.
TABLE V.

~Il)del ~umb('r

H.eturdlng Dl'nslt\.

CHARACTERISTICS OF NEAC-SERIES 2200 MAGNETIC TAPE UNITS
N204A-7

r-.;~u

:,.)li.2UU

5;'6.200

200.556.ijOO

:JOO.1'i51i, tJOO

:':20 . 1}1-.1

S20-lU-2

~:W-HI-I

.).)t-"

200

.. n~l"

N204H-fj

!\~O-lB-l

1'120411-11
1'120411-12

N204A-l

N204A-2

N204A-3

556.800

556.200

533

533

740

N204B-9

!,)t:. Inch

Pl'ah Transfer Rate,
hd()-('har Sl't

20.0.7.2

6-1.0 lh.O

66.7.24.0

7. 2.~O. O. 28, 8

16. O. H. 4, 64. 0

66.7.96.0

13.3.4.8

32.0

64.0

88.0

Tape Speoo,

36

KO

120

a6

0"

120

24

60

120

120

108

24U

360

108

240

360

72

180

360

360

Yes

Yes

Yes

Yes

Yes

Yea

Yes

No

No

No

~o

1'0

No

No

No

No

No

r.<204B-l 1<,
prc-rcquH.lte
for N:lO·Ui-2

N~U4

Inches set

Tape Re .... md Speed,
Inches's('c
IH~t 72~

Compatlhie

IBM 2-lUU Compatlhle
FC.tturcs

~nd

(-ommen!;.,

prc-rcqulslh'

N2048-11 Is
pre-requlalw

for K20-lB-4

[or 1'120411-12

B-:! is

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

8/69

NEAC SERIES 2200

1575:011.510

TABLE VI: SUMMAHY OF PAPEH TAPE EQUIPMENT CHAHACTERISTICS

-.,....,.....,._.,...rn...t_

-...-

I NlOIlA-l

300

......aI_

.
-

1,000

-

••
Japa: • level. IDternatioul
• level: " ..r ..\lonal ti, 7

Do ..

1• .,.1, 01' IAklr.. UonaI 5 level

Fol1lW.\

Prooee.r. '.t

.51

0.11

NlIOA-!

H210A-3

I

NUOA-3

liD

Dle-puach

100

•

Japan 8 lev", ...

Jal*n 6 level and Jlllpen 8
level

raterntlUonaI .,7,111 lnel,
"apan 6 Jeve!, and ....1'...Uonal 5,., 'I level; 01'
I)

naUoaal
~ Db

I

-

".pan

nala ProUtctioa

NIlOA, .. 1

NaOIA-a

PMto~rlllld

RMdIll.1IIaCI

...... _

N10lA-.

II

,

,,,yel ... II:llerij

llltem.tiona) 8 level. and
JapaD 6 level. Jalem.tlonal
6, " level and Japan 8
level; or lnlenaUoul 5
It'vel. aad Japan fi leveia

level

0 ••

O.I)!

Partey cbeck and lie rtlUdJ .. check

NfUU'

Random Access Drum File and Control (Contd.),
Data is stored at a density of approximately 406 bits per inch on the magnetic surface of
a 9-inch diameter cylindrical drum which has a constant speed of 3600 rpm. The drum
has 256 data tracks and read/write heads.
The standard data record contains 128 characters (one sector); one track can contain ten
such records. However, records can be variable In length. A drum unit has the capacity
for 2,560 standard records or 327,680 characters, making the total capacity of an eight
drum subsystem 2,621,440 characters.
The average access time for any record on the drum is 8.3 milliseconds. The average
transfer rate between the central processor and the drum is 103,000 characters per
second or 1. 67 milliseconds per standard record.
Recorded data is protected by a PERMIT/PROTECT switch on each drum which inhibits
writing when in the PROTECT position. A parity bit is generated for each character
during writing and automatic parity checking is performed when reading. Any check
condition generates a program-accessible error indication.
Data transfer interrupts the central processor one character cycle per character transfer.
The central processor is available for other operations during a search for a record area.
The instruction set for the drum is search and write, write, search and read, read, and
read address register.
The Random Access Drum file, Model N271 has the same features as Model N271A with
the following exceptions:
•

The data capacit~' is 20,480 records of 128 six bit characters or 2,621,440
characters;

•

The drum rotation speed is 1200 rpm;

•

The record access time is 27.5 ms (average); and

•

. 32

The transfer rate is 106,000 characters per second .
Disc Equipment
The :\261, :\262 Random Access Disc File and Control is a rapid memory device which
records and reads data from one to 36 (N261) and 72 (N262) magnetic discs. As many as
eight N261 or four 1'\262 disc files can be connected to a control unit. The storage capacity
of each disc file is 134 million (N261) or 268 million (N262) 6-bit characters, which means
that the storage capacity per control is expandable up to 10,000 million characters.
Discs rotate under 64 (N261\, 128 (N262) movable read/write heads at 1167 rpm. The
average access time is 78 milliseconds. The data transfer rate is 188,000 characters
per second. Only one memory cycle is required to transfer one character from main
memor.\'.
The Disc Pack Unit, Model N259, is a high-speed random access filing device. It combines the characteristics of both magnetic discs and tape units in that data records can
be variable in length and formatted under program control. It is connected to the N257
disc pack control which is connected to the input/output trunks of the central processor.
One disc pack unit has a storage capacity of 9.2 x 10 6 characters and has an average
access time of 110 milliseconds. Th~ data transfer rate is 208,000 characters per
second.

8/69

A

AIIU1):11 read-write channels and connects up to
eight input-output control units. AU Mo(kl r,o j)<'ripheral controls are connected to the
processor via input-output trunks. This lll(':II\;; that the Model 50 can be connected to as
many peripheral c()ntrols as the numbeJ' 01 :t,,:tilable input-output trunks.
Data is transferred between the main menlO)'\, and a trunk via a read-write channel which
is assigned by the instruction which initiat!'d th(' tnl11s[cr.

NEAC-Series 2200/50 Installation.
© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

8/69

1576:011. 210

, 21

NEAC SERIES 2200/50

Central Processor (Contd. )
The automatic program interrupt facility provides simple but efficient supervision of
processing involving combinations of input-output operations and computing. This allows
automatic branching between a main program and servicing routines for all input-output
devices.
The Model 50 interrupt processing facility consists of a hardware program interrupt
which signals a particular condition in an input-output control unit. An interrupt instruction subset is available for interrupt processin~.
A program interrupt occurs at the completion of any input/output operation. All interrupts can be inhibited by program control. A RlIlgle instruction is necessary to resume a
main program and to reset all indicators and registers .

. 22

Data Structure
The basic addressable unit of the Model 50 Is a 6-bit character. Any number of consecutive locations can be grouped to form fields; a group of consecutive fields form items.
Fields and items are defined by word marks and Item marks.
Instructions are variable in length. The basic instruction format consists of an operation
code, two operand fields which specify the binary addresses of fields to be used in the
operation, and a variant character. The variant characters expand the meaning of the
operation code or represent data to be used by the instruction. Both decimal and binary
arithmetic can be performed .

. 23

Card Equipment
The E214 Card Reader-Punch performs three operations: reading, punching and reading
a card and punching additional information lIlto the card on the same pass. The E214
has a reading speed of 400 cards per minute and a punching speed of from 100 to 400 cards
per minute. Cards are end fed and optically read. The appeal of the unit is enhanced by
the addition of a dual punching head for punching two columns simultaneously.
The E106-l Transcription Feature is optional. Table I gives the characteristics of the
E 214.
TABLE I: CHARACTERISTICS OF NEAC-SERIES 2200/50 CARD EQUIPMENT
Reading Speed
Punching Speed
Data Protection
Hopper/Stacker Capacity
Punch Feed Read
Direct Transcription
:\Iaximum :\umber of t'nits in System

.2-1

400 cards per minute
100 to 400 cards per minute
Validity check, cycle check on reading;
punch check on punching
1200/1300 cards
Standard
Optional
8

Printer
The Model 50 is equipped with a chain printer, the E206 High-Speed Printer. The
printer has a print speed of 333 lines per minute and has 60 characters available for
printing. The standard printer has 120 print positions but the EI030 Extended Print
Option permits the printing of a 132 character line.
The skipping rate is 100 lines per second and vertical spacing is 6 lines per inch. Control instructions for line and form spacing and editing instrucJions are available to the
programmer.
The NEAC-Series 2200 Model 50 can accommodate a maximum of eight E206 Printers •

. 25

Paper Tape Equipment
The E211 Paper Tape Controller connects both the E209 Paper Tape Reader and the E210
Paper Tape Punch to the central processor. Both,units pr.ocess 5, 6, 7 or 8 level tape.
The E209 Reader processes at 300 characters per second while the E210 Punch operates
at 110 characters per second. Both units come equipped with odd or even parity checks.
Table II lists the characteristics of both units.

8/69

fA

AUERBACH

(Contd.)

SUMMARY REPORT

1576;011.260

TABLE II: CHARACTERISTICS OF NEAC-SERIES 2200/50 PAPER TAPE EQUIPMENT
Characteristic

.2(,

E209 Paper Tape Reader

E210 Paper Tape Punch
110 char/sec
Die Punch
11 inches/sec

Speed
Read/Punch Mechanism
Tape Transport Speed

300 char/sec
Photoelectric
30 inches/sec

Data Format
Data Protection

5, 6, 7. or 8 Level Code
Program generated parity;
Dual read check

Magnetic Tape Units
The E204 Magnetic Tape Units process 1/2 inch, 7-track tape. The E203 Tape Controller
controls up to four tape units sharing a common input-output trunk. Write attempts can be
guarded by a write-enable ring and a manual tape unit switch. All information which is
written is read and checked. Parity is checked during reading and writing. Table III
shows the principal characteristics of the E204 tape unit.
TABLE !II: CHARACTERISTICS OF THE E204 MAGNETIC TAPE UNIT

. 27

Transfer Rate

8,900 char/sec

Tape Width

1/2 inch

Tape Length

450 feet per reel

Number of tracks

7

Read-Write Speed

Hi inches/sec

Recording Density

55G

Inter-record Gap

0.45 inches or 0.75 inches

Rewind Speed

48 inches/sec

Maximum Number of Units in System

:12

char/inch

Random Access Equipment
l\Yass storage for the NEAC-Series 2200 Model 50 system is provided by disc pack and
drum storage units. The E261 Random Acccss Disc Pack Storage consists of removable
disc pack storage cartridges. As many as four F.261 storage modules can be connected
to an E260 disc controller. Each module contains one disc pack which consists of five
discs of which eight surfaces can be used to storc data.
There are 64 tracks per surface which are further divided into 16 sectors. Each sector
has a capacity of 100 characters giving a total disc capacity of 819,200 characters.
The head positioning time ranges from 126 to GOO milliseconds. Data is transferred at a
rate of 83,333 characters per second. The 2200/GO system can contain up to 32 Disc Pack
Storage units.
The E271 Random Access Drum rotates at 3600 rpm which provides an access time of
8,3 milliseconds. The E270 Drum Controller can handle as many as four such units.
A read-write head is assigned to each of the 64 data tracks. A band consists of four
tracks giving 16 bands per drum. Bands are further divided into 40 sections each of
which can contain 128 characters giving one tenth the storage capacity of the disc pack of
81,920 characters. Data is transferred at 103,000 characters per second. The maximum number of drums which the 2200/50 can contain is 32 .

. 28

Input-Output Typewriter
The E220 Input-Output typewriter, or NEAC WRITER, consists of a typing mechanism, a
keyboard and a paper tape input-output secUon. tJp to four E220's can be connected to an
EI025 Adapter which may be attached to the E211 Paper Tape Controller. Six- or eightlevel paper tape can be processed, Table IV lists the peak speeds of the E220.
© 1969 AUERBACH Corporat.on and AUERBACH Info. Inc.

8/69

NEAC SERIES 2200/50

1576:011.300

TADLE IV: CHARACTEHISTICS OF NEAC-SERIES 2200/50
Printing Speed
Paper Tape Reading Speed
Paper Tape Punching Speed
Maximum Number of Units in System
.;1

560 char/min
560 char/min
900 char/min
32

SOFTWARE
The software system of the NEAC Series 2200 Model 50 incilldes the following:
•

Three E system

•

Programming systems

•

Application systems

The Model 50, being one model in the NEAC Series 2200 line, allows programs written
for all other models of the Series to be run.
• :11

Three E System
The Three E system is composed of EASY BILL, EASY PRO, and EASY COBOL. EASY
BILL is an automated program package which includes functions of billing processes and
can produce billing programs as directed by parameter cards. The minimum machine
configuration to use EASY BILL is a 4,096 character central processor, 1 paper tape
reader, 1 typewriter and an 819,200 character disc pack storage unit or a drum storage
unit of 81,920 characters.
EASY PRO contains functions for sorting. collating. merging and reproducing. The minimum machine reqUirements are a central processor with 4,096 characters of memory, 1
paper tape reader, 1-4 magnetic tape units and 1 printer.
EASY COBOL is a Japanese language version of basic COBOL. The minimum configuration required is a 4.096 character central processor, 1 paper tape reader, 1 typewriter,
and -1 magnetic tapes or 1 disc storage unit with a capacity of 819,200 characters.

· :12

Programming Systems
The programming systems for the Model 50 consist of the following:
•

Basic Programming System

•

Operating

S~'stem

:'IIOD I

The Basic Programming System is mainly for paper tape based users while the operating
system is card based. The Basic Programming System includes an Easycoder assembler,
a COBOL processor, a program for editing and updating files, control programs, data
manipulation routines and scientific subroutines. The minimum machine configuration
contains a central processor with a -1,096 character memory. 1 paper tape reader. 1
printer and one to four magnetic tape units.
Besides the contents of the Basic Programming System, the Operating System MOD 1
contains a FORTRAX compiler. two CODOL compilers (an English and Japanese version),
and an interrupt control package.
The minimum machine configuration required for the Operating System MOD 1 is a main
memory of -1,096 characters and an additional channel trunk, 1 card reader punch, 1
printer and -1 magnetic tapes.
· :13

Applications Systems
The XEAC Series 2200 :\!odel 50 is supplied with three basic application tools, the Forecasting for Inventory Control System (FIeS), Program Evaluation and Review Technique
(PERT), and Linear Programming Packages (LP).
FICS is used to forecast sales, processes files of sales demands and updated parameters
used in stock control and supply managem,ent operations. FICS requires a machine
configuration of 8,192 characters.

8/69

A

All! RRACH

(Contd. )

1576:011.330

SUMMARY REPORT

.33

Applications Systems (Contd.)
PERT is used to forecast the time required for completion of projects within specified
time periods and to evaluate progress and results. PERT Is based on the premise that
minor activities incorporated within larger projects exert a considerable influence upon
the overall project schedule. PERT identifies those portions of the overall project which
!'equirc greater control and supervision and produces a wide variety of rcports. The
minimum machine configuration required to run PERT is a memory Hize of 16,384
characters.
The LP system is used for solving problems which utilize linear programming techniques.
The system is divided into several segments which perform given functions and overlays
the next segment. A Simplex method is used in the system in which statistical tabular
data is maintained on magnetic tape Whlt'h makes it possible to handle numerous variables and limiting equations.
The minimum machine configuration required to use the LP system is a memory size of
16,384 characters.

C'

-1969 AUERBACH Corporiltlon ilnd AUERBACH Info, Inc

8/69

A•

AUERBACH

AU.IIMeH
COMI'UT.II
NOT••OOK
INT.IINATIONAL

1575:221. 101
NEAC SERIES 2200
PRICE DATA

PRICE DATA
CLASS

Model Feature
Number Number
CENTRAL
PROCESSORS

PRICES

IDENTITY OF UNIT

NlOO
NlOOM-l
NIOOM-2
NlOOM-3
N200
N200M-l
N200M-2
N300
N300M-1
N300M-2
N400
N400M-1
N400l\1-2
N500
NSOOl\l-l
N500l\1-2

Name

Processor (2,048 characters)
Additional Memory (2,048 characters)
Additional Memory (4,096 to 16,384 characters)
Additional Memory (4,096 to 32,768 characters)
Processor (4.096 characters)
Additional Memory ('1,096 to 32,768 characters)
Additional Memory (R, 192 to 65.536 characters)
Processor (16,384 characters)
Additional Memory (\(;,384 to 65,535 characters)
Additional Memory (16,384 to 131,072 characters)
Processor (16,384 characters)
Additional Memorv (\6, :l84 to 131,072 characters)
Additional Memory (:l2, 76A to 262,144 characters)
Processor (65,536 characters)
Additional Memory (Ii:i, :i3G to 262,144 characters)
Additional Memorv (1:11,072 to 524,288 characters)

Monthly
Rental
in ¥

Purchase
Price in
Thousand ¥-

13,500
300,000
40,000
1,800
80,000
3,600
70,000
3,150
450,000
20,250
80,000
3,600
110,000
4,950
930,000
41,850
260,000
11,700
150,000
6,750
1,600,000
70,000
200,000
9,000
300,000
13.500
2,700,000 121,500
800,000
36,000
54,000
1,200,000

Features for 2200/100 PrOCCS!lOrS
N10ll
N10l3
N10H
NlO15
NI016

Advanced Programming
Editing Instructions
8-bit Code Handling
Auxiliary Input/Output Trunk
Auxiliary Read/Write Chllnnel

27,000
18,000
9,000
54,000
108,000

1,215
810
405
2,430
4,860

36,000
32,000
54,000
18,000

1,620
1,440
2,430
810

108,000
18,000
18,000

4,860
810
810

36,000
18,000
108,000
18,000
40,000

1,620
810
4,860
810
1,800

100,000
100,000

4,500
4,500

Features for 2200/200 Processors
Advanced Programming
Editing Instructions
Auxiliary Input/Output Trunk
Auxiliary Read/Wrlt(' Channel

NOlO
:-;013
N015
NOl6

Features for 2200/300 PJ"(l('cssors
Scientific Vnit
Storage Protect
Optional Instruction Package

NllOO
Nll14
N0191

Features for 2200/400 Processors
N1115
N1l17
NllOO
NOl91
N1l21

Auxiliary Read/Write Channel
storage Protect
Scientific Unit
Optional Instruction Package
Extended Multi-programming Option
Features for 2200/500 Processors

Nl118A
Nll16A

Storage Protect
Auxiliary Read/Write Channel

© 1969 AUERBACH CorporatIon and AUERBACH Info, Inc.

8/69

1575:221.102

CLASS

PRICE DATA

IDENTIFY OF UNIT
Model
Number

HANDOM
ACCESS
STORAGE

PRICES

Feature
Number

N270
N271
N075
N270A
N271A
N075A
N257
N257-1
N259
N260
N261
N262
N077
N4005

Monthly
Rental
in ¥-

Name

Drum Storage Control
(Controls up to 8 drum units)
Drum Storage
Track Protect
Drum Storage Control (controls 8 drum units)
Drum Storage
Track Protect
Disc Pack Drive Control
Disc Pack Drive Control
Disc Pack Drive
Disc Control Unit
Disc File
Disc File
8-bit Transfer Feature
Disc Pack

103,000
:

Purchase
Price in
Thousand ¥
4,635

230,000
10,350
9,000
405
80,000
3,600
180,000
8,100
5,000
225
200,000
9,000
230,000
10,350
185,000
8,325
230,000 10,350
1,450,000 65,250
2,460,000 110,700
20,000
900
5,400
243

Punched Card
I~Pl'T-

Ol"TPl'T

N207
N20S
N227
N017
N040
N060
N061
N062
N223
N044
N043
N20SA-l
N22-lA-l
N061A-l
N060A-l
~20SA-2

N22-1A-2
N060A-2
NOlilA-2
:\207A-2
N062A-2
N064
1\20S-1
1\06-1-1
:\21-1-1
:\208-2
1\214-2
N066
l\12:l

Card Reader Control
Card Punch Control
Card Read/Punch
Stacker Select
Direct Transcription
Direct Transcription
Hole Count Checking
Punch-Feed Read
Card Reader
Direct Transcription
51 Column Adapter
Card Punch Control
Card Punch
Hole Count Checking
Direct Transcription
Card Punch Control
Card Punch
Direct Transcription
Hole Count Checking
Card Read Control
Punch Feed Read
Direct Transcription
Card Punch Control
Direct Transcription
Card Punch
Card Read/Punch Control
Card Read/Punch
High Speed Skip
Card Reader

86,000
76,000
258,000
9,000
18,000
18,000
36,000
38,000
140,000
9,000
15,000
76,000
78,200
36,000
18,000
76,000
140,000
18,000
36,000
86,000
38,000
20,000
54,000
36,000
108,000
81,000
126,000
9,000
72,000

3,870
3,420
11,610
405
S10
S10
1,620
1,710
6,300
405
675
3,420
3,519
1,620
810
3,420
6,300
810
1,620
3,870
1,710
900
2,430
1,620
4,860
3,645
5,670
405
3,240

99,000
35,000
105,000
12,000
81,000
45,000
81,000
33,000
10,000
10,000
10,000
10,000
10,000

4,455
1,575
4,725
540
3,645
2,025
3,645
1,485
450
450
450
450
450

Paper Tape
1\209
N209A-l
N209A-2
NI09A-l
1\210
N210A-1
N210A-2
Nll0A-l

8/69

Paper Tape Input
Paper Tape Input
Paper Tape Input
Paper Tape Input
Paper Tape Output
Paper Tape Output
Paper Tape Output
Paper Tape Output
N029A-l
Special Code Detection
N029A-2
Special Code Detection
NI029A-l
SPecial Code Detection
N028A-l
ISO Code Processing
N028A-2
ISO Code Processing

A

AUERBACH

(Contd.)

NETHERLANDS

AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

AUERBACH
®
Printed in U.S.A.

fA

AUERBACH

AUERBACH

1620:011.100

COMPUTER
NOTEBOOK
INTERNATIONAL

PHILIPS Pl000 SERIES
SUMMARY REPORT

'"

SUMMARY REPORT:
.1

PHILIPS Pl000 SERIES

BACKGROUND
Philips Gloeilampenfabrieken, who prior to 1968 was active in core memories, peripherals and computers, has made a full-scale entry into the computer market with the
announcement of its P1000 series, a line of three program-compatible computers which
parallel the IBM System/360 Models 30, 40, and 50.
The P1000 Series has many advanced features, including a very broad size range for core
memory from 16,384 to 517,288 octads of 1 J.!sec on the smallest computer to an extended
core of over 14,680,064 octads of 2.5 J.! sec on the two larger models.
Other features of the P1000 Series are the compactness of the hardware, the extensive
instruction set, and the special provisions for multiprogramming and dual system
processing. The appeal of the PIOOO Series is further enhanced by a balanced set of
operating and applications software. An extensive range of equipment is available, which
enables an installation to be extended if, and when, required. First deliveries are
scheduled for 1969 .

.2

DATA STRUCTURE
The basic unit of data in core storage is the octad, which consists of eight information
bits and an associated parity bit. A word is four octads or 32 bits in length.
In the P1000 Series, data can appear in two forms: fixed length and variable length.
Data of fixed length may have the following formats: the octad (eight bits), the half word
(sixteen bits), the word (thirty-two bits), and the double word (sixty-four hits). Fixed
length data can be grouped in the following way:
•

Alphanumeric characters (1 character pcr octad) in the 1':1\('1)1(, Wxtended
Binary Coded Decimal Interchange Code) or the lISASC II (l:nited States or
America Standard Code for Information Interchange)

•

Binary coded integers have half word and word formats

•

Floating point numbers have word and double word formats

•

Logical data may take any form up to the double word format or a combination
of four double words.
.

Decimal numbers occupy 1 octad at minimum, 1 sign tetrad and 1 digit tetrad and at
maximum 8 octads, 1 sign tetrad and 15 digit tetrads .
.3

HARDWARE

.31

Central Processors
There are three basic sizes of internal storage for each of the central processor models,
as shown in Table 1.
The processing units consist of a control section, registers, and provisions for arithmetic
and logical operations. The control section contains the necessary equipment for converting logic addresses into store addresses, for storage protection, for address modification and indirect addressing, and for transferring data.
Programs may be written in relation to a relative address zero. The ultimate location
in storage is determined by the control program monitor. A program base register
(PBR) is filled by the control program with the starting address of the program from
which, during processing, all addresses can be derived. This technique is called
address conversion. Usage of the PBR makes dynamic relocation of a program possible
even if that program is under execution. The length of a program is maintained in a program length register (PLR) and together with the PBR, this register prevents the program
limits from being exceeded.
© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

5/69

1620:011.310

• 31

PHILIPS PIOOO SERIES

Central Processors (Contdo )
Use can be made of the segmenting feature in the P1200 and P1400 to enable segment c, of
programs to reside in different locations in storage or, temporarily, on another st0rage
medium.
In the P1000 Series a distinction is made between direct, modified, and indirect addressing. In direct addressing the address portion of an instruction represents an explicit
address in store. A maximum of 65,536 octads are accessible through direct addressing.
With modified addressing, use is made of the contents of one of the 14 index registers
that are available. Two levels of indexing are permitted. With indirect addressing, the
address portion of an instruction does not contain the relevant operand address, but the
address of a storage location that does contain the operand address. Modification of
indirect addressing permits addressing of up to 224 locations.
Available on the P1000 Series are 2 arithmetic registers, (A and B), 14 index registers,
1 condition register and 1 point-location register.
Arithmetic operations include binary, decimal or floating point.
are available:

Five number formats

Half-word integers that are integral binary numbers with a length of 16 bits,
including the sign bit. Negative numbers are represented in two's complement
notation.
TABLE I: CENTRAL PROCESSOR CAPACITIES
Central Processor

Core Storage

Additional Core Modules

Model

Model

Octads

PllOO

PllOO-OOl
PllOO-002
PllOO-003

16,384
32,768
65,536

P1200

P1200-001
P1200-002
P1200-003
P1400-001
P1400-002
P1400-003

P1400

Model

Octads

65,536
131,072
262,144

P1200-008

2,097,152

131,072
262,144
524,288

P1400-008

2,097,152

TABLE II: CHARACTERISTICS OF MAGNETIC TAPE UNITS
Features
Model

Recording
Method

Number
of
Tracks

Tape Speed,
inches/second

PI061

001
002
003

NRZ
NRZ
NRZ

9
9
9

37.5
75
112.5

800
800
800

30,000
60,000
90,000

P1061 (with
P1061-13
installed)

001

NRZ

7

37.5

002

NRZ

7

75

7,500
20,000
30,000
15,000
40,000

003

NRZ

7

112.5

200
556
800
200
556
800
200
556
800

001
002
003

PM
PM
PM

9
9
9

37.5
75
112.5

1600
1600
1600

60,000
120,000
180,000

Type

PI064

5/69

A

Character
Density,
Char/inch

Peak Transfer
Speed,
char/second

GO,OOO

22,500
60,000
90,000

(Contd. )

AUERBACH
®

SUMMARY REPORT

1620:011.311

TABLE III: PUNCHED CARD EQUIPMENT
Feature

Card Reader

Card Punch

P1010

P10ll

P1012

P10l5

P1016

Reading Speed for SO-column cards,
cards/min

400

SOO

1500

-

-

Reading Speed for 51-column cards,
cards/min

500

1000

2000

-

-

Punching Speed for SO-column cards,
cards/min

-

-

-

100

300

Punching Speed for 51-column cards,
cards/min

-

-

-

-

400

Capacity of Hopper

1400

2500

2500

1000

2000

Capacity of Stacker

1000

2000

2000

1000

2000

100

2000

-

-

-

Capacity of Selection Stacker

TABLE IV: LINE PRINTERS
Model
Feature

Pl030-00l

P1030-002

Pl030-003

Printing Speed, lines/min

360

600

1000

Skipping Speed, inches/sec

20

20

35

Maximum number of copies

5

!i

!i

10

10

]0

6

G

G

Horizontal Spacing, characters/inch
Vertical Spacing, lines/inch
. 31

Central Processors (Contd. )
•

Word integers that are integral binary numbers with a length of 32 bits including
the sign bit. Negative numbers are represented in two's complement notation.

•

Decimals that are binary coded decimal numbers, consisting of an even number
of tetrads (4 bits) with an algebraic sign and 15 decimal digits maximum.

•

Word floating point numbers, consisting of a sign bit, a 7-bit exponent and a
24-bit fraction. Numbers in this notation range between 5.4 x 10- 79 and
7.2x10 75 .

•

Double-word floating point numbers, which are similar to the word format but
with a fraction of 56 bits. Numbers in this notation also range between 5.4 x
10- 79 and 7.2 x 10 75 •

Logical data can be of a fixed or variable length. When a logical item is used as an
operand of fixed length it is processed in a register. The length of variable data can be
up to 256 octads. Alphanumeric characters are expressed in EBCDI Code but it is possible
to process data in the United States of America Standard Code for Information Interchange
(USASCII).
The PlOOO Series has instruction lengths of one and two words. Only the field instructions
are of the double word format. Although all word instructions have the same format the
meaning of the different portions can vary according to the type of the instruction.
Both types of instruction are always located at word addresses in storage; i. e., a multiple
of 4 octads.
There are two types of channels in the P1000 Series: Character Allocated Transfer
Channel (CATCH) and Block Allocated Transfer Channel (BATCH).

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

5/69

1620:011.312

.31

PHILIPS PI 000 SERIES

Central Processors (Contd. )
CATCH executes the complete input-output procedure for the transfer of one character
only. The advantage of this is that characters can be transferred to and from different
input-output devices in succession.
CATCH has a lower data transfer rate than that of BATCH, and hence is used for lowspeed devices. BATCH is concerned with one input-output device at a time and is blocked
against any other transfer of data. BATCH is used on high-speed devices such as magnetic
tape units and disc units .

. 32

Magnetic Tape Equipment
The magnetic tape units are available in two nine-track, single-density types: the Pl06l
and the Pl064. As an option, the Pl06l can be provided with facilities for recording on
seven tracks in one of three densities. Each type of tape unit is available in three models
of different speeds. Details of the tape units are shown in Table II.
Data is recorded on the Pl06l Model by means of a non-return-to-zero (NRZ) method and
on Model Pl064 by means of the phase modulation (PM) method. On the three-density
version of Model Pl06l, the density is changed by means of a switch on the units. All
models allow reading backwards.
The 8-bit EBCDIC Code is used for data representation along with a parity bit. The
standard magnetic tape reel holds approximately 2400 feet of tape.
The Pl060 and Pl062 Tape Control Units may be used in conjunction with up to eight Pl06l
or Pl064 Magnetic Tape Units, respectively. Both control units allow as many as eight
tape rewinds simultaneously or one unit to be reading or writing and seven units rewinding.
The seven-track, three-density feature on Model Pl06l is provided through the installation of feature Pl060-l3.
The tape units controlled by the Pl062 may be of type Pl06l if the Pl062-0l2 NRZ Feature
is installed .

. 33

Disc Equipment
The Pl04l Disc Unit is provided with an interchangeable pack of six 14-inch diameter
discs. Data is recorded on 10 of the 12 sides available by means of an access mechanism
that has 10 read/write heads. Each side of a disc has 200 tracks for storing data; the
storage capacity of a disc pack is 7,250,000 octads. The average access time to a block
of data is approximately 90 milliseconds and the peak transfer speed is 156,000 octads
per second.
The self-contained Pl040 Disc Control Unit is used in conjunction with up to eight disc
units and allows positioning of the access mechanisms to proceed simultaneously on each
of up to eight disc units; data can only be written or read from one disc unit at a time.
The Pl040-012 Dual Entry Switch is designed for dual system operation and allows connected disc units to be accessed via two channels, one from each of the combined
processors .

. 34

Data Communication Equipment
The PIOOO data communication equipment includes the Pl086 Teleprinter, the 1085 Data
Collection Device together with the necessary control units, the PI080 Data Communication Control Unit, and the Pl080-0l0 and PI080-012 Line Control Units. The PI080 Data
Communication Control Unit has a maximum of 16 line connections. Four of the 16 can
be used by Pl080-012 Line Control Units and the remaining 12 are used by PI080-0l0
Line Control Units.
A maximum of twenty P1085 Data Collection Devices can be connected to the P1080-012
Control Unit. Data collection devices permit data communication in one direction. Data
can be input from the P1085 keyboard or from punched cards or from external sources by
normal telephone lines.
The P1086 Teleprinter can be connected to the Pl080-0l0 via a single line and permits
data communication in two directions. Data can be input by manual operation of the Pl086
keyboard or with punched tape. Output of data is accomplished by printing and/or punching tape. Here also use can be made of normal telephone lines.

5/69

fA

AUERBACH
®

(Contd.)

1620:011.400

SUMMARY REPORT

.4

INPUT-OUTPUT DEVICES

.41

Punched Card Equipment
Three card reader models are available: P10lO, P10n and P1012. As shown in Table
III they differ mainly in their speed of operation. Reading is carried out photoelectrically,
column by column, and the hopper and stacker can be loaded and emptied during operation. The control units for the readers are built into the equipment. All cards being
processed are checked in transport and any faults stop the reader.
Two card punch models are available: P1015 and P1016. These differ mainly in the
speed of operation and the size of the cards which can be punched (Table III). Both card
punch models check for faulty punches, and Model P1015 has an optional selection feature .

. 42

Printing Equipment
The Model P1030 Line Printer is a family of three models: P1030-001, P1030-002, and
P1030-003. These models differ mainly in their speed of operation (Table IV).
Paper sizes range from 4. 5 to 19 inches wide. All three models of line printers have 64
character sets arranged on a printing drum, and characters may be printed in up to 132
positions per line depending on the paper size .

. 43

Punched Tape Equipment
The P1020 Punched Tape Reader is very compact and reads paper, mylar or nylon tape at
a speed of 1000 characters per second. Tapes punched in five, six, seven or eight channels are able to be read. Checking is provided by a second read head.
The punched tape reader is attached to a control unit which controls one tape reader.
The P1025 Tape Punch has a maximum speed of 150 characters per second. About 10
characters can be punched per inch of tape in either five, six, seven or eight channels.
A tape punch control unit can control one tape punch .

.5

SOFTWARE

.51

Operating Systems
There are four operating systems provided for the P1000 Series: Support Package,
Basic System, Extended System and Multiprogramming System.
The Support Package is intended for smaller P1000 installations dealing principally with
punched card processing. It differs considerably from the other Operating Systems:
•

maximum core storage used will not exceed 64K octads

•

its operating programs are coded in Autocode S and are on punched cards which
require translation before use and the user must add them to the source program

•

each program must be run separately.

The system support routines for Support Package are Update S, Tape Sort S, Disc Sort S,
Rug S, Autocoder S, and system utility routines.
The Basic System is available for users with an installation having a central processor
with a minimum of 32, 767 octads of main memory, at least one card reader, an operator's
typewriter and one magnetic tape unit or one disc unit.
The following programs are available in the Basic System:
•

ALGOL B, FORTRAN B, COBOL B;

•

Autocoder B assembly program;

•

RUG B - Report and Update Generator;

•

Utilities, Linkage Editor B, Update B;

•

Disc Sort B, Tape Sort B;

•

Macro Processor;
© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

5/69

1620:011.510

.51

PHILIPS Pl000 SERIES

Operating Systems (Contd.)
•
•

Data Management B; and
~

Basic Monitor and Job Control B.

The Extended System requires a minimum central processor capacity of 65,536 octads,
one card reader, an operator's typewriter, one line printer and one disc unit or five
magnetic tape units. The greatest advantage of the Extended System is that the slow
peripheral equipment does not retard the execution of the user programs.
The Extended System consists of the same programs as the Basic System except for the
Extended Monitor and ,Job Control E.
The Multiprogramming System enables a maximum of 15 programs to be executed
Simultaneously. A priority system is used to determine the order of operations.
A Multiprogramming System requires the following minimum configurations: 131,072
octads of core storage, one card reader, one line printer, one operator's typewriter,
three disc units and a segmenting feature.
The hardware Dual System Feature is supported by the Dual System Routines B, E and M,
which are optional features of the Basic, Extended and Multiprogramming Monitors .
. 52

Data Communications Routines
Data Communications modules B, E and M are developed to handle messages entering
via data communications control units from remote terminals. When used in one of the
operating systems, these routines always have the highest priority.
For the Basic Operating system, extended by the data communication module, the arrival
of a message stops the execution of the current program which will be temporarily rolled
out and the data communication routine B are loaded and executed.
Two priorities are recognized by data communication modules E and M. A high priority
message is handled by stopping and rolling out the current Program.
The lower priority message is held in an external storage medium and the relevant data
communication program is executed as soon as possible after the end of the current
program. Within the Multiprogramming Operating System a message is handled by
giving control to the relevant data communication program as soon as the message is
completed.

5/69

fA

AUERBACH
®

A
~

AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

1620:221.101
PHILIPS PIOOO SERIES
PRICE DATA

AUERBACH

'"

PRICE DATA
Pricing information for the Philips PIOOO Series was unavailable at press time.
data will be included in future issues.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

Price

5/69

UNITED KINGDOM

AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

AUERBACH
®
Printed in U.S.A.

AUERBACH
COMPUTER
NOTEBOOK

A

1850:011.100
ICL SYSTEM 4
SUMMARY REPORT

INTERNATIONAL

AUERBACH

'"

SUMMARY REPORT: ICL SYSTEM 4
.1

BACKGROUND
International Computers Limited (ICL) of London, England was formed in 1968 and represents the culmination of a series of mergers among English computer manufacturers
that effectively consolidates almost all of England's computer manufacturing into single
corporate structure. ICL now manufactures three lines of computers of which the System
4, Models 30, 40, 50, 70 and 75, are compatible with IBM's System 360/25 through System
360/75.
The System 4 Series was designed for use in both the commercial and scientific data processing community and possesses the following general attributes:
•

A broad range of real-time communications equipment;

•

Compatibility within the series; and

•

System expansion and product line expansion.

ICL now markets in about 70 countries throughout the world and maintains marketing support in the United States. ICL can offer leasing agreements of 3, 5 or 7 years. Maintenance contracts are usually for seven years. Table I highlights some of the features of the
five processors. The Models 4-70 and 4-75 are, in fact, the same computer with the
addition of a paging unit in the 4-75. This unit, based on an associative memory, is
designed for use in time -sharing operations. The paging unit slows down the instruction
execution rate of the machine when it is in use, since special action is needed on each
access to main memory. It is possible to run a Model 4-75 without using the paging unit;
in this mode, it has the same instruction execution rate as the Model 4-70.
Typical rentals range from £1, 700 per month for a 4-:30 to£12, 000 per month for a 4-70.
The first 4-50 was delivered in June 1967, followed by the first 4-30 in October 1967.
Deliveries of 4-70 commenced in March 1968 and of 4-75 in September 1968.
The principal characteristics of System 4 include:
•

All processors use monolithic micro-integrated circuits.

•

A high degree of compatibility among the larger models and with IBM
System/360 and RCA Spectra 70.

System 4 has the same non-privileged instruction set as the full IBM System/360 and
RCA Spectra 70 Series instruction set, except for the 4-30, which uses the commercial
subset of instructions. The data codes and formats are also the same for System 4 and
System/360. Software is provided for various levels of operating system according to the
type of configuration available and the user's needs.
ICL maintains a technical information exchange agreement with RCA, and System 4 and
Spectra 70 share similar design concepts. Except for the 4-50, which closely corresponds
to the Spectra 70/45, there is not a close correspondence between individual System 4
models and Spectra models .
. 21

Data Structure
The data structure used in System 4 is identical with that of the IBM System/360 and the
RCA Spectra 70 Series. The basic unit is the 'byte', consisting of eight data bits and one
parity bit. A byte can represent one alphanumeric character, two decimal digits or part
of a binary field. Binary fields are either two consecutive bytes ('halfword'), four consecutive bytes ('word'), or eight consecutive bytes ('double word'). Halfwords, words and
doublewords must be located in storage on an integral boundary, that is the address of the
first byte must be a multiple of the number of bytes in the field. Decimal and alphanumeric
fields may be of variable length in all processors.
Instructions are two, four or six bytes in length. Decimal operands, in all processors are
from one to sixteen bytes long. They are made up of 4-bit BCD digits packed two to a byte,
© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

5/69

lel SYSTEM 4

1850:011.210

. 21

Data Structure (Contd. )
with a sign occupying the least significant half-byte. Fixed point arithmetic is performed
on halfwords and words in the 4-40 upwards. The 4-30 performs decimal arithmetic and
fixed point binary add and subtract only, using variable length operands from one to sixteen bytes, on words and halfwords. Floating point arithmetic can be simulated on the
4-30, but on the other processors floating-point instructions are implemented in hardware,
using either 'long' or 'short' operands. 'Long' operands (using a doubleword for one
number) consist of a 56-bit fraction and a 7-bit hexadecimal exponent; 'short' operands
(single word) have a 24-bit fraction and 7 -bit exponent.

.22

Central Processor
The 4-30 processor has instructions for performing decimal arithmetic, binary addition
and subtraction, comparison, branching, moving, loading, storing, packing, unpacking,
code translation, editing and Boolean operations. The 4-40, and higher numbered processors, also have facilities for radix conversion, binary multiply and divide, and floating
point arithmetic. The full instruction set of the 4-40 upwards contains 144 instructions,
while the 4-30 has 41 instructions. Instructions are two, four or six bytes in length. A
2 -byte instruction has no reference to main storage; a 4-byte instruction refers once to
main storage, and a 6-byte instruction refers twice to main storage.
Main storage addresses are formed by adding a 12-bit 'displacement' from the instruction
to a 24-bit 'base address' kept in one of the 16 general registers. This allows program
segmentation and relocation as well as one level of address modification. A second level
of modification is provided in many instructions where a 24-bit 'index', held in a general
register, is added to the base -plus -displacement address in instructions referring to one
storage address and a register. System 4 has no indirect addressing.
On the 4-30, there are five types of interrupt. On the larger processors, there are up to
32 types of interrupt. Interrupts are caused by input-output, illegal arithmetic operands,
overflow, illegal operation code, improper addressing, etc. The larger processors have
four processor states to give fast interrupt servicing and each state has its own set of
registers in the 'scratchpad store'. User programs run in the normal state until interrupted while the interrupt is analyzed and serviced in other states. The 4-30 has two
processor states; interrupt servicing takes place in the normal state.
Storage protection and the interrupt system permit efficient multiprogramming of a maximum of 14 programs on the 4-40 upwards. Multiprogramming allowed on the 4-30 is of
up to three media conversion routines with a main program.
In the 4-70 and 4-75, banks of store are interleaved, reducing the cycle time to 0.65J.lsec.
Store protection in all processors is against writing only, allowing a program to access
another area but not to change it.
In the 4-40, 4-50, 4-70 and 4-75, scratch pad storage is provided to hold registers, etc.
The 4-40 and 4-50 have 128 32-bit words, and the 4-70 and 4-75 have 72 32-bit words with
a separate scratch pad for input-output operations in the Multi-channel Control Unit.
The cycle times for the scratchpad memories for the different processors are included in
Table I. Table II gives some representative arithmetic times .

. 23

Simultaneity
Peripheral devices and their controllers are attached to System 4 processors via various
types of input-output channels. A selector channel controls one high-speed input-output
operation at a time. A multiplexor channel can control many low speed devices operating
concurrently.
The 4-30 has up to 8 concurrent selector channels, and a multiplexor channel to which up
to 115 devices can be connected. The 4-40 can have 3 concurrent selectors and a multiplexor prOViding for 256 devices. The 4-50 has the same capabilities as the 4-40 with an
additional feature for direct control of up to 5 other processors. The 4-70 and 4-75 can
operate up to 16 selectors concurrently, can have up to 506 devices connected to the
multiplexor and can have direct control of 5 other processors.
Each channel has one or more trunks. Each trunk is connected to one device control unit
which controls one or more devices depending on the type of device. All types of processor can be linked to exchange data. Note that, for the purposes of applying the
restrictions described in this Section, a communications network which is multiplexed
through a control unit counts as a single' concurrent device.

5/69

A

(Contd. )

AUERBACH
®

SUMMARY REPORT

1850:011.240

TABLE I: ICL SYSTEM 4 CENTRAL PROCESSOR CHARACTERISTICS

Storage SiL:e,
Bytes

4-40

4-30

Characteristics

16,384 to 65,536

Cycle Time,
11 sec

4-50

4-70/75

65,536 to 131,072 65,536 to 262,144 65,536 to 1,048,576

1.5

1.5

1.4

0.65

2

2

2

4

Bytes/ Access
Reserved
Storage, Bytes

None

2048

2048

512

Scratchpad

None

Yes

Yes

Yes

Scratchpad
cycle time,
Il sec •

-

0.5

0.3

0.25

Number of
Instructions

41

144

144

144

*

14

14

14

360

465

520

4,000

2

4

4

4

Levels of
Mult~programming

Data Transfer
Rates
(KB/Sec)
Interrupt Levels

*

One user program plus up to three media conversion routines.

TABLE II: ARITHMETIC EXECUTION TIMES ON SYSTEM 4 PROCESSORS
Operation

4-30

Processor Execution Time, 11 sec
4-70
4-50
4-40

4-75

Fixed 120int binar,l::
:l2-bit operands
c~a+b

c
c

~

~

ab
a/b

Fixed Point Decimal *
c-a+b
c c ab
c ... alb
Floating point-short
:12-blt operands
c~a+b

c

~

c·~

ab
a/b

53.70
673
691

33.84
119.40
162.09

25.20
81.96
111. 21

4.82
9.17
14.10

6.12
10.52
15.45

50.1
672.8
694.3

63.22
194.29
319.93

42.0
129.86
213.99

14.72
67.45
93.50

16.52
69.49
95.54

-

40.32
70.54
104.12

37.20
67.66
101.24

6.82
9.90
13.68

8.17
11. 25
15.03

-

55.53
214.39
308.11

52.60
211. 50
305.24

8.68
16.36
23.64

10.03
17.71
24.99

Floating: point-long
(iI-bit operands
c=a~b

c = ab
c = alb

*
• 24

-

a and b are five-digit operands for all the decimal arithmetic operations except division, for
which a has ten digits and b has five digits .
Mass Storage
Mass storage for System 4 is provided by a range of random access devices which allow
up to 700 million bytes per unit as shown in Table III. All the units have IBM-compatible
track formats •

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

5/69

1850: 0 11 .241

ICL SYSTEM 4

. 24

Mass Storage (Contd. )
The Replaceable Disc Unit uses interchangeable stacks of six discs. The 4-30 processor
can be outfitted with the fixed length field version of this unit but only Models 4-40, 4-50,
4-70 and 4-75 can use variable length fields. Different control units are required for the
two versions.
Information is stored on the ten inner surfaces of the six discs in a stack with a readwrite head for each surface which contains 200 operational tracks plus 3 alternate tracks.
Ten tracks can then be covered at anyone time and constitute a "cylinder". All the discs
rotate at 2,400 rpm with head movement between adjacent tracks in 23 milliseconds and
across all tracks in 145 milliseconds. The average head positioning time is 90 milliseconds and the average latency is 12.5 milliseconds. Up to eight replaceable discs can
be connected to a single control unit.
The Discfile contains four disc stacks operated as two independent pairs each with a
capacity of 350 million bytes. The discs rotate at 1,500 rpm with head movement between
adjacent tracks in 27 milliseconds and across all tracks in 140 milliseconds, giving an
average head movement time of 80 milliseconds. The average latency is 20 milliseconds.
With Single-track recording the peak data transfer rate is 265,000 bytes per second two-track, or parallel, recording gives a peak data transfer rate of 530,000 bytes per
second. Up to eight 350 million byte discfiles can be connected to a single control unit.
A Magnetic Drum unit is also available for System 40 This has a capacity of two million
bytes and is controlled by a drum control unit to which up to eight drum units can be connected. The access time for the drum unit is 10 milliseconds, and four tracks are read
in parallel, giving a peak data transfer rate of 875,000 bytes per secondo This high
transfer rate limits the use of this device to System 4 Models 70 and 75.
Table III lists the capacities and transfer rates of all the System 4 Mass Storage units.

TABLE
Type of
Storage

Model
Number

m.

SYSTEM 4 MASS STORAGE UNITS.

Capacity,
bytes

Peak Transfer Rate,
bytes per second

Average access Time,
milliseconds.

Replaceable
disc storage

4425

7,250,000

156,000

10205

Discfile

4440

350,000,000

265,000

100

or

Magnetic
Drum

.25

4443

700,000,000

530,000

100

4430

2,200,000

875,000

10

Magnetic Tape Equipment
Two types of magnetic tape recording are available for System 4 - the Non-Return-to-Zero
(NRZ) mode at 800 bits per inch and Phase Encoding at 1600 bits per inch. For each of
these recording modes, there are three different nine-track tape units with different
transfer rates. There is also a seven-track magnetic tape unit, which uses the NonReturn-to-Zero method of recording and can be used at three different recording densities.
The units provided for Phase Encoded recording can optionally record in the NRZ mode.
A device control unit links up to 8 magnetic tape devices to the processor, and can optionally be extended to handle 16 magnetic tape devices. Dual channel control can be provided
on each magnetic tape unit.
Seven track tapes hav€ two optional features The pack/unpack facility converts three
eight-bit bytes to four six-bit characters when writing and vice versa when reading. The
translation facility provides for the conversion from eight-bit bytes in E BCD I C code
to six-bit binary coded decimal characters. Pack/unpack and Translate can all be used
for both odd or even parity and any available character density. Table V contains the
various physical characteristics as well as the rated speeds of the various tape unit
models. Reverse reading is provided on all models.
0

5/69

A

(Contd. )

AUERBACH
®

SUMMARY REPORT

1850:011.242

TABLE IV: SYSTEM 4 INPUT-OUTPUT EQUIPMENT
Features and Comments

Rated Speed

Unit

includes controller.
no controller; up to 3 4581 readers
can be connected to a 4580 reader.

4580 Paper Tape Reader
4581 Paper Tape Reader

1500 char/sec
1500 char/sec

4585 Paper Tape Punch

150 char/sec

4513
4515
4520
4521
4522

Medium Speed Card Reader
High Speed Card Reader
standard Speed Card Punch
High Speed Card Punch
High Speed Card Punch

800 cards/min
1435 cards/min
100 cards/min
300 cards/min
300 cards/min

reads 51- or 80-column cards.
reads 51- or 80-column cards.
80-column cards only.
80-column cards only.
80-column cards only; includes
binary image punching feature.

4554
4555
4560
4561

High Speed Line Printer
High Speed Line Printer
Medium Speed Line Printer
Medium Speed Line Printer

1350 lines/min
1350 lines/min
750 lines/min
750 lines/min

160
132
160
132

print
print
print
print

positions.
positions.
positions.
positions.

TABLE V: SYSTEM 4 MAGNETIC TAPE UNITS

Model

No. of
Tracks

Recording
Density, rows
per inch

4450

7

800

Rewind Speed
inches per second

Peak Transfer Rate,
bytes per second

75

150

60,000

556

75

150

41,700

200

75

150

15,000

Speed inches
per second

4452

9

800

75

150

60,000

4453

9

800

150

300

120,000

4454

9

800

37.5

100

30,000

4458

7

800

75

150

60,000

556

75

150

41,700

200

75

150

15,000

4460

9

1600

37.5

150

60,000

4461

9

1600

75

150

120,000

4462

9

1600

125

250

200,000

TABLE VI: SYSTEM 4 DIGITAL PLOTTERS
Model

Paper Width, inches

Step Length

Step Rate, steps
per second

4710

31

.01 inches

200

4711

31

.005 inches

300

4712

12

.01 inches

300

4713

12

.005 inches

300

4714

31

.1

mm

300

4715

12

.1

mm

300

\

?

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

5/69

1850:011.260

.26

ICL SYSTEM 4

Input-Output Devices
The characteristics of the principal input-output devices for System 4 are summarized in
Table IV. Besides the devices listed, there is a range of digital plotters whose characteristics are given in Table VI. There is also a MICR (magnetic ink character) sorter/
reader and a mark-sensing document reader, both of which can be used on-line to a System
4 computer. The data communications equipment available for System 4 is discussed in
Paragraph .27.
.
Five, seven and eight track Paper Tape can be handled having width variations of 11/16
inch, 7/8 inch and 1 inch which can be adjusted under manual control. Spoolers are
available as an optional feature but tape dispensers are supplied. The ISO code is the
normal code used for paper tape equipment on the System 4, however program conversion
is possible for other codes. The ISO code is automatically converted to the internal
E BCD I C code.
Two models of Card Reader provide for the reading of 51 and 80 column cards at 800 and 1435
cards per minute. Automatic conversion between the System 4 extended Hollerith card
code and the internal E BCD I C code is provided on both models. In addition, the
readers can have the option to read in binary card images. The readers have two stackers
with capacities of 2,000 cards each. Both stackers are program selectable. The hopper
capacity is also 2,000 cards and loading and unloading is permitted during operation.
All Card Punches for System 4 have a read-after-punch feature for data checking. Two
stackers with capacities of 850 cards each are provided on high speed punches and an
input hopper with a capacity of 1,000 cards is available. The input hopper of the standardspeed punch has a capacity of 850 cards and the output stacker has a capacity of 850 cards.
All punches are row-oriented and Model 4521 and Model 4522 differ only in the addition of
a punch binary feature on Model 4522.
The printers for the System 4 are barrel-type and print a 64 character subset of E BCD I C
in the E C M A type B font. Paper for these printers varies in width between 4 and 20
inches.
The maximum print line is 160 characters in length. A printer acts independently of the
central processor via its own control unit once a buffer has been loaded.
Printing can be spaced at six or eight lines to the inch under operator control. Normal
multiple line feeding is 33 inches per second; however, when the high-speed paper motion
option is employed, feeding at 75 inches per second is possible. A special printing barrel
is available which gives increased printing speeds, for purely numeric data, of 1500 lines
per minute on the medium-speed printers and 2700 lines per minute on the high-speed
printers.
The characteristics of the various Digital Plotters available for System 4 are given in
Table VI. Both continuous lines and individual points can be plotted on all models. The
plotting paper can be translucent and can be plain or marked with a number of linear and
non-linear scales. Annotation is possible on all models.
The System 4 MICR Reader/Sorter, Model 4601 processes documents marked with a
single line of up to 84 magnetic ink characters at a rate of 1,500 documents per minute.
When the reader/sorter is operated off-line, documents are sorted into the 13 available
pockets according to the value of a specified digit. When the operation is on -line, characters are sent to core storage and the documents are sorted under program control. Any
document between 2.5 and 4.25 inches high and 5.75 and 9.5 inches wide is acceptable.
A Model 4652 Lector 2 Mark Sensing Document Reader can be used on-line to a System 4
computer. Document can have up to 16 data columns and one control column, w.ith up to 79
lines on the form. The line spacing can be up to five lines per inch. Marks can be entered
by hand, by computer printer, or by using an address plate. The range of acceptable document sizes is from 4 by 5 inches to 10 by 16 inches. The documents are moved through the
device at 30 inches per second, giving throughput rates ranging from 6,500 documents per
hour for 6" documents to 3,000 documents per hour for 16" documents .

. 27

Data Communications Equipment
A comprehensive range of data communications equipment is available for System 4. The
devices available can be broadly classified into two types - limited distance control units
and unlimited distance control units.

5/69

A

(Contd.)

AUERBACH

'"

SUMMARY REPORT

.27

1850:011.270

Data Communications Equipment (Contd. )
The limited distance control units are those used where the data communications requirements are localized. Communications between two System 4 computers within 200 feet of
each other is provided by the Data Exchange Controller, which allows two processors to
be connected together via selector or multiplexor channels, the peak data transfer rate
being governed by the speed of the slowest input-output channel involved in the transfer.
Communication between a System 4 computer and up to 16 local terminals is provided by
the Multi-Purpose Device Control Unit. The peripheral devices concerned can be up to
4000 feet from the controller, and the data transfer rate can be up to 800 bytes per second
for each device on the controller. The devices available for use with the Multi-Purpose
Device Control Unit include an output writer operating at 10 to 15 characters per second,
an alphanumeric keyboard and a slow card reader, which reads standard, 8u-column cards
at an effective rate of 100 cards per minute.
The unlimited distance control units are designed for communications with remote devices
over telegraph lines, or public, private or leased telephone lines. These devices include
the Single Channel Communications Controller, which is used for communication in a conversational mode between two System 4 computers. Various models are available to suit
different line speeds and using seven- or eight-bit character codes. Facilities for manual
calling or automatic dialing are available with this device.
Communication between a System 4 computer and a number of remote terminal devices is
provided by the Multi-channel Communications Control Unit. This is a special-purpose
computer which is used for the control of up to 112 remote devices connected to the multiplexor channel of a System 4 computer. The total overall data transfer rate on all the
lines handled by a single multi-channel communications control unit can not exceed 57,600
bits per second. The functions of the Multi-channel Communications Control Unit include
the transmission of characters between main storage and remote devices and the detection
of certain control characters and error conditions. The remote devices connected to a
System 4 computer via a Multi-Channel Communications Control Unit are typically terminal devices, but they can include other System 4 computers connected either via a Single
Channel Communications Control Unit or another Multi-channel Communications Control
Unit.
Remote terminal devices available from ICL for use with System 4 computers include
teletypewriters, video data terminals and various types of banking terminals. There is a
range of Remote Data Terminals, each of which handles up to four local devices and communicates with the computer via a data link operating at up to 4800 bits per second.
Devices available for use with the Remote Data Terminals include teletypewriters, paper
tape reader, a card reader and a line printer. Besides these devices, a range of off-line
paper tape transmission devices is available from ICL .

.4

SOFTWARE
The organizational structure of the software for the System 4 is similar to that of IBM for
System/360, and RCA for Spectra 70. Software for a particular installation is selected
on the basis of processor size and type and the available backing storage. All compilers,
operating systems, utilities, assemblers and application packages are classified by
regime. Each regime contains an automatic program trials system in which modules of
user programs can be amended, compiled in relocatable form, linked together and run
without operator intervention.
Multiprogramming of up to 14 programs is available in regimes which use a minimum
configuration of 65,536 bytes of storage on Model 4-40 and above. A remote multi-access
time sharing system is being developed for the 4-75.
Table VII shows the minimum configuration for each regime, and Table VIII shows the
main features of the software for each regime. All the software described is currently
available with the exception of the Multijob operating system, which is scheduled for
release in the first quarter of 1970.
The tape operating systems, regimes 3D, 4E and 5E, include COBOL, FORTRAN and a
Report Program Generator. Regimes 4E and 5E also include a compiler for CLEO - a
commercial autocode originally developed for ICL ' s LEO computers. Multiprogramming
of up to six programs is allowed in 4E and 5E, where the resident supervisor routines
occupy 16,384 bytes of -storage. In regime 3D the supervisor occupies 12,288 bytes of
storage.
The disc-based operating system for large machines, regimes 4J, 5J and 7J, includes an
ALGOL compiler and various application packages as well as the facilities of the tapebased systems. J regimes have a more sophisticated job control system than E regimes.
© 1969 AUERBACH CorporatIOn and AUERBACH Info, Inc.

5/69

ICL SYSTEM 4

1850:011.400

TABLE VII: MINIMUM EQUIPMENT REQUIREMENTS FOR SYSTEM
4 OPERATING SYSTEMS
Computer
Regime
Minimum Main
Memory, bytes
Input Device (Paper
Tape Reader or Card
Reader)
Line Printer
Magnetic Tape Units
Replaceable Disc Units

4-30
3D
32,768

3H
32,768

4-40, 4-50, 4-70 or 4-75
E-Level
J-Level
Multi-job
131,072
65,536
65,536

1

1

1

1

1

1
5
0

1
0
1

1
6(1)
0

1
0
2

0(2)
3 (2)

1

(1) Five of the units must be nine-track - the other can be either seven- or nine-track
(2) Two replaceable disc units and two magnetic tape units can be used instead of three
replaceable disc units.
TABLE VIII: FACILITIES AVAILABLE IN SYSTEM 4 OPERATING SYSTEMS
4 -40, 4-50, 4-70 or 4-75

4-30

Computer
3D

3H

E level

Type of
system

Tape based

Disc based

Tape based

COBOL
With Random
Access

X

X
X

X
X

X
X

X
X

FORTRAN

X

X

X

X

X

X

X

X

X

X

X

X

X

ALGOL
CLEO

.4

J level

Multi job

Regime

Disc based

Remote
job entry

Report Program
Generator

X

X

PERT

X

X

X

X

Linear
Programming Simplex
Transportation

X

X

X

X

X

X

X

X

Matrix Scheme

X

X

Statistics
Package

X

X

APT IV

X

X

CSL (Simulation
Language)

X

X

SOFTWARE (Contd.)
Within each of the J level operating systems, there is multiprogramming of up to 14 user
programs. This multiprogramming is based on a stream concept. The Supervisor allows
up to six streams (A to F). Each stream consists of a fixed amount of core storage and a
fixed number of each type of peripheral device; discs are excepted, since all programs in
all streams can access the disc concurrently. Since stream sizes are tailored for each
installation, it is presumed that programs would be developed to fit these streams.
Multiprogramming control is on a priority basis where each program is assigned a running
priority, 1 to 14 (where 14 is the highest priority), when it is loaded. The best use of the

5/69

A

(Contd,)

AUERBACH
®

SUMMARY REPORT

.4

1850:011.401

SOFTWARE (Contd. )
machine is achieved by peripheral-dominated programs being given the highest priorities
and calculation-dominated programs the lowest.
Whenever a program is waiting for a peripheral transfer, Supervisor passes control to
the highest priority program able to continue. Hardware checks are performed to inhibit
the destruction of anyone program by another.
Extensive software support is provided for communications equipment, particularly in the
J regimes. The Communications Control Package, a software package available with the
J regimes which is tailored to the needs of each installation, provides software support
for the use of a variety of terminals and other remote devices. A number of different
types of use are supported, and there are facilities to allow a number of separate, multiprogrammed user programs to use communications devices at the same time. There are
facilities for multiprogramming engineering test programs with other communications
programs, so that some of the terminals on a system can be tested while other parts of
the system are in use.
The Multijob operating system, scheduled for release in the first quarter of 1970, will
provide facilities for remote job entry and remote program testing. Batch processing
facilities will be available concurrently, and remote job entry will be performed on a
roll-in/roll-out basis using a replaceable disc unit .

.5

PRICE DATA
ICL does not issue a general price list. Price data for specific configurations will be
published in a future issue.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

5/69

\

AUERBACH
COMPUTER
NOTEBOOK

A

AUERBACH

1855:011.100

INTERNATIONAL

ICL 1900 SERIES
SUMMARY REPORT

'"

SUMMARY REPORT
.1

°

leL 1900 SERIES

BACKGROUND
The 1CL 1900 Series, manufactured by International Computers Limited of London, England,
presently consists of 14 central processors and a wide range of peripheral equipment and
supporting software. The 1900 Series systems are suitable for both business and scientific
applications, and most models provide facilities for multiprogramming and real-time operations. Most of the peripheral units and all of the programming languages are fully compatible within the series. The 1900 Series was introduced by ICT (International Computers
and Tabulators Limited) in September 1964. At that time the series consisted of the 1902,
1903, 1904, 1905, 1906, 1907 and 1909 processors. The small-scale 1901 processor was
added in September 1965. The 1900 Series design was largely based upon the earlier FP
6000, a computer developed by~Ferranti-Packard Limited in Canada to the specifications
of the Ferranti Computer Department in England, which merged with ICT in September
1963.
In 1967 ICT announced the 1904E, 1905E, 1906E, and 1907E processors, which use 1. 8microsecond core memories, and the 1904F, 1905F, 1906F, and 1907F processors, which
use O. 75-microsecond core memories. In October 1967, ICT introduced the large-scale
1906A processor, which uses integrated circuits. In January 1968 four more integratedcircuit processors were added to the line: the 1901A, 1902A, 1903A, and 1904A.
The A, E, and F series processors have effectively replaced the eight original 1900 Series
processors. Table I summarizes the principal characteristics of the 14 processors that are
currently being actively marketed.
In July 1968, ICT joined forces with English Electric Computers to form a new computer
company - by far the largest in the United Kingdom - called International Computers
Limited (ICL). The formation of ICL represented the culmination of a long series of
moves aimed at the consolidation of the various United Kingdom computer manufacturers.
ICL currently plans to continue marketing both the 1900 Series and the System 4, which
was English Electric's third-generation computer line and is discussed in Summary Report 1850.

The 1900 Series has enjoyed far greater market acceptance than any previous Britishbuilt computer line. To date, more than 1350 orders have been received and more than
900 systems have been installed. The availabi.li.ty of the faster "A" series processors and
more advanced software has considerably enhanced the marketing prospects for this wellconceived product line.
ICL is the largest organization not controlled from the U. S. in the commercial and scientific computer business. It has marketing activities in over 70 countries around the world
and it is estimated that it has some 45 to 50 per cent of the U. K. market. ICL' s new
super computer, the 1908A, projected for delivery in 1972, is said to be "one of the most
powerful computers in the world. "
.2

HARDWARE

.21

Central Processors
Any program wOritten for a 1900 Series central processor can be run on any other central
processor in the series; hardware features not included in a.processor, such as floatingpoint, are simulated by software routines. A part of this general philosophy of compatibility is the ICL standard interface for attaching peripheral devices to the central processor.
This interface insures that any peripheral can be connected to any processor in the range,
subject only to the data transfer capacity of the channel concerned, and also that older
peripheral devices supplied with the earlier 1900 Series computers can be used on the more
recently announced members of the Series.
Certain processors can be paired with another central processor of the same type to form
a dual-processor configuration.. Programs in either processor have access to peripherals
connected to either processor. The processors share a common core store. One executive and one operating system is shared by the two constituent processors, but program
© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

5/69

ICL 1900 SERIES

1855:011.210

.21

Central Processors (Contd. )
instructions are able to run in each of the processors simultaneously. Through store interleaving, up to eight words can be accessed at one time.
The first eight words of each program are used as accumulators for arithmetic, copying,
testing, and logical functions. Three of the accumulators can also be used as modifying
(index) registers. Additional hardware, to implement the accumulators as separate hardware registers, is available optionally.
There are two branch modes in the 1900 Series instruction set, normal branch mode (one
word instruction) and extended branch mode (two-word instruction). The normal branch
mode limits the address part of a branch instruction to 15 bits or 32,768 locations. The
extended branch mode allows 22 bits for the address of the branch instruction, which
theoretically allows the addressing of up to 4,194,304 locations.
The general word format of a program instruction is represented by four fields: an accumulator field, an operation field, a modifier field and an operand field.
The priority interrupt feature consists of a number of peripheral channels for attaching
non-ICL peripherals to certain 1900 Series Central Processors. Peripherals connected
by means of this feature will be serviced in Priority mode which has a higher priority
than Executive mode.
The principal characteristics of the ICL 1900 Series processors currently being marketed
are shown in Table I. These processors are divided into three principal groups: A series,
E series, and F series. The A series are the latest to be added to the 1900 line and incorporate integrated circuits. The principal difference between corresponding models of
the E and F series is the speed of the associated core storage. The four models within
the E and F series are fundamentally the same machine; differences in the models are
whether or not the basic machine is equipped with floating point arithmetic and whether it
is a single or dual processor. The newly announced 1908A is the only model in the A series
that can be equipped for dual processing. Multiprogramming is supported on all models
except the 1901A. Special-purpose paging hardware, which divides the available memory
into 1024-word pages, is available for the 1906A and 1908A.

TABLE I.

5/69

PRINCIPAL CHARACTERISTICS OF THE ICL 1900 SERIES PROCESSORS

Input-Output
Channels

Integrated
Circuits

Dual
Processors

FloatingPoint
Hardware

16,384

4 to 7

Yes

No

Optional

8,192

32,768

4 to 8

Yes

No

Optional

1.5

16,384

65,536

4 to 12

Yes

No

Optional

1904A

0.75

65,536

262,144

10 to 31

Yes

No

Optional

1906A

0.75

65,536

524,288

14 to 49

Yes

No

Optional

Yes

Optional

Standard

Minimum
Core Storage,
24-bit Words

Processor
Model

Cycle Time,
Microseconds

1901A

4.0

4,096

1902A

3.0

1903A

Maximum
Core Storage
24-bit Words

1908A

0.33

131,072

2,097,152

64 to 80

1904E

1.8

32,768

262,144

6 to 30

No

No

1905E

1.8

32,768

262,144

6 to 30

No

No

Standard

1906E

1.8

65,536

262,144

12 to 60

No

Yes

No

1907E

1.8

65,536

262,144

12 to 60

No

Yes

Standard

1904F

0.75

32,768

262,144

6 to 30

No

No

No

1905F

0.75

32,768

262,144

6 to 30

No

No

Standard

1906F

0.75

65,536

262,144

12 to 60

No

Yes

No

1907F

0.75

65,536

262,144

12 to 60

No

Yes

Standard

fA

AUERBACH
®

No

(Contd. )

SUMMARY REPORT

. 22

1855:011.230

Data Structure
The basic unit of storage on the 1900 Series Computers is called a word and consists of
24 consecutive bits plus parity.
A character of data is stored as six consecutive bits. Data in pure binary form is always
interpreted as having a numerical value. Apart from counter modifier words, words holding pure binary data has the most significant bit reserved as a sign bit. Negative numbers
are expressed as a complement.
Signed numbers may be fixed-point integers or fractions, mixed, or floating-point, and
may be held in a single word or a double word.
A counter modifier (index) word can be used in two forms: as a word-counter modifier or
as a character-counter modifier. The word-counter modifier uses 9 bits to hold a count
of the number of times an operation is to be performed, up to a maximum of 511; 15 bits
are used to hold the modifier. In the character-counter modifier word, the character modifier occupies two bits, the counter occupies 7 bits and can contain a maximum value of 127,
and the modifier occupies the remaining 15 bits of the word.
For a single length integer the binary point is assumed to be right justified. Single length
integers lie in the range _223 to +223 inclusive. Negative integers are stored as their
complements with respect to 224.
Fixed-point arithmetic operations are provided for single- or double-word operands for
either integer or fraction formats.
For a mixed number the binary point can be assumed to lie between any two bits of the
one or two words holding the number. However, the general 1900 Series convention is
to use two words to store a mixed number; one word for the integer part and one word
for the fraction part. In this case the number is referred to as a 'mid-point' number
since the binary point is assumed to be between the two words.
Floating-point numbers can be held in single, double, or quadruple words; the double
word form is standard and has a 37 -bit fraction plus sign and an 8-bit exponent plus sign .

. 23

Card Readers
The 1900 Series includes five different card readers. Models 2104, 2105 and 2106 read
600, 300 and 600 80-column cards per minute, respectively.
Also available are Model 2101 that reads 1,600 80-column cards per minute and Model
2103 that reads 600 40-column cards per minute.
Models 2106 and 2101 can be equipped with a binary image feature, which interprets each
card column as two 6-bit characters. Any of five card codes, including the IBM 1401
code, can be switch-selected on the 2102. Several other card readers have been discontinued, but could be connected to a new processor because of the common interface .

. 231 Card Punches
The 1920 card punch is a row punch, capable of punching 100 80-column cards per minute.
The 1922 is a column punch capable of punching 33 cards per minute, while the 2151 is a
row punch capable of punching 300 cards per minute .
. 232 Line Printers
The 1900 Series includes five different line printers. All are drum-type printers using
continuous-form paper. A 64-character print font, 10-character-per-inch horizontal
spacing, and vertical and horizontal tabs are provided on all models. All can print six
lines to the inch; in addition, Model 1933 can print eight lines per inch. Table II lists
the various printers models available on the 1900 series processors along with their
peak speeds and available print positions .
. 233 Paper Tape Equipment
All ICL paper tape I/O devices normally employ the ICL 8-track (7 data-bit) code based
upon the 1. S. O. 7 data-bit code. The characteristics of the readers and punches are
shown in Table III.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

5/69

1855:011.231

1900 SERIES

TABLE II. LINE PRINTERS
Model

Number of Print Positions

Speed, Lines/min.

120
96
96
96
96

1100 or 1350*
300
600
300
600

1933
2401
2402
2404**
2405**

or
or
or
or
or

160
120
120
120
120

*

The lower speed is with the full 64-character set. The higher
speed is using only a restricted 48-character portion of the
full set.

**

The 2404 or 2405 is integrated into the 1901A processor
cabinet.

TABLE III. PAPER TAPE EQUIPMENT
Peak Speed, char/sec

Name

Model

300
1000
250/110
1000/110
110

Reader
Reader
Reader/Punch
Reader/Punch
Punch

1915
1916
2601
2602
1925

.234 Disc Storage
Both fixed- and removable-disc storage is available for the 1900 Series. Fixed-disc
storage units have a capacity range of from 100.66 million to 741 million 6-bit characters.
Up to 14 fixed-disc storage units may be included in a system, permitting a maximum of
10.4 billion characters of on-line storage. Exchangeable Disc Storage units have a
capacity range of from 1. 64 million to 8.19 million 6-bit characters. Up to 8 Exchangeable Disc Storage units can be connected to one controller.
The characteristics of the ICL disc units are summarized in Table IV.

TABLE IV: ICL 1900 SERIES DISC STORAGE UNITS.
Model

Type of unit

Number
of discs

Storage
capacity,
char

Average
access time,
msec.

A verage data
transfer rate*,
char/sec

2801
2802
2805/1
2805/2
2805/3

Exchangeable disc store
Exchangeable disc store
Fixed disc store
Fixed disc store
Fixed disc store

6
6
7
14
26

4.1 million
8.19 million
100.66 million
218.10 million
419.43 million

97.5
97.5
150
150
150

208,000
208,000
135,000
135,000
135,000

2806/2
2806/3
2806/4
2820
2821

Fixed disc store
Fixed disc store
Fixed disc store
Twin exchangeable disc store
Twin exchangeable disc store

7
14
28
2
2

218 million
421 million
741 million
1. 64 million
3.28 million

150
150
150
162.5
162.5

135,000
135,000
135,000
208,000
208,000

*The average data transfer rate in the table represents the average peak data transfer rate
for the unit concerned. In some cases a unit has different data transfer rates on different
tracks of the disc.

5/69

fA

AUERBACH
®

(Contd.)

SUf1MARY REPORT

1855:011.232

.235 Magnetic Drums
Drum storage is available for the 1900 Series with average access times ranging from
6.5 to 21. 5 milliseconds and storage capacities ranging from 32,768 to 4.19 million
24-bit words (131,056 to 16.8 million characters). The larger figures- represent the
capacity of a four-drum system.
The characteristics of the ICL drum units are summarized in Table V.
TABLE V. ICL 1900 SERIES DRUM STORAGE UNITS
Words
Per Track

Magnetic Drum
32,768 words

11.5

256

50,000

1963

Magnetic Drum
131,072 words

11. 5

512

100,000

1964

Magnetic Drum
524,288 words

21.5

1,024

100,000

2851

Magnetic Drum
524,288 words

6.3

1,024

800,000

Description

1962

.24

Average Data
Transfer Rate,
Characters/Second

Average Access
Time, Milliseconds

Model

1004 Link
One ICL 1004/0 or 1004/2 Data Processor can be linked, via a standard interface, to any
ICL 1900 Series central processor provided it is not operating in a multi-programming
environment. The link enables those already using the plugboard-programmed, UNIVACdesigned 1004 to utilize the additional processing facilities and memory of a storedprogram computer. The 1004 is effectively both a card reader and a line printer contained in one unit. It is no longer manufactured by ICL, but second-hand 1004 units are
available.
For use as a 1900 Series peripheral, the 1004 must have a suitably plugged control panel.
ICL has designed a 1004 input-output program to perform the following operations:
•

Read an 80-column card.

•

Transfer 80 characters from the 1004 to a 1900 Series Central Processor.

•

Transfer 120 or 132 characters and one paper feed control character from
a 1900 Series Central Processor to a 1004 for printing.

•

Print a line of 120 or 132 characters.

•

Transfer 8 characters from a 1900 series central processor to a 1004
for punching.

•

Punch an SO-column card.

•

Translate from 1004 internal code to 1900 internal code and vice versa.

Within the limits of the 1004 program capacity available after the input-output program
is implemented, it is possible to include other routines. However, the successful incorporation of such routines is a user responsibility.
When not being used as a peripheral device for a 1900 series central processor, the 1004
can be used independently •
• 25

Magnetic Tape Systems
The 1900 Series is available with eight different magnetic tape systems. All the systems
are packaged in "clusters" containing from two to six tape drives. The 2501 cassette
tape system discussed in this Report is no longer being sold by ICL but is still supported.
All the tape systems except the 2501 record either 7 or 9 tracks on half-inch-wide magnetic tape in various sizes of reels containing up to 2400 feet of tape. The 2501, which

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

5/69

1855:011.240

ICL 1900 SERIES

.25

Magnetic Tape Systems (Contd.)
is a cassette tape system, records 8 tracks on a loop of one-inch-wide magnetic tape with
an average length of 240 feet. Except in the case of the 2501, one track is used for parity
bits. With the 2501, a cyclic check code is used. Table VI summarizes the basic characteristics of the tape systems available for the ICL 1900 Series.
The 2404, 2405, 2406, and 2407 tape systems write in the forward direction only and
can read in either the forward or reverse directions. All the other tape systems write
and read inlthe forward direction only.
TABLE VI: ICL 1900 SERIES MAGNETIC TAPE SYSTEMS

Recording
Tape
Drives Number
Tape Speed,
per
Density,
System
of
inches per sec
bits per inch
No.
Cluster Tracks

Peak Speed,
6-bit
char/sec

Interblock Character Full Rewind
Time,
Parity,
Gaps,
minutes
odd or even
inches

1971

2/4/6

7

37.5

200/556

20,800

0.56/0.75

Either

4

1972

2/4/6

7

75

200/556

41,700

0.56/0.75

Either

2.5

1973

2/4/6

7

75

200/556/
800

60,000

0.56/0.75

Either

2.5

2501

4

8

150

531

10,000

1.2

None

0.32*

2504

2/3/4

9

37.5

1600

80,000

0.6

Either

3

2505

2/3/4

9

75

1600

160,000

0.6

Either

3

2506

2/3/4

9

37.5

800

40,000

0.6

Either

3

2507

2/3/4

9

75

800

80,000

0.6

Either

3

'The 2501 does not actually rewind but passes tape in the forward direction only. The time shown is the
maximum for initializing the tape and includes the time to move the tape to the beginning point.

The 9-track tape units record data in eight tracks, with the ninth bit for parity. The data
may be in binary form (i. e., each 8-bit row represents a third of a word), or alphanumeric characters may be recorded as 6-bit characters. In the latter case, a row represents either one complete 6-bit character and two bits of the next, or two 4-bit subsets
of two consecutive 6-bit characters.
Except for the 2501 Cassette system, tape block lengths in the 1900 Series are limited
by standard software convention to a maximum of either 2,048 or 4,096 six-bit characters, depending upon the central processor model. The 2501 block length is similarly
limited to 512 24-bit words •
. 26

Optical and Magnetic Readers
The document sorter/reader reads hand-made or printed marks from forms of a wide
variety of sizes. The positions of the marks to be read are detected by means of a
sequence of pre-printed markings along one edge of each form which are used to control
the reading action of the device. Under program control, the sorter/reader can also
route documents to any of three output stackers. An additional feature is the reading of
80-column cards for which code translation is required. The document sorter/reader
is plugboard-controUed and operates at two speeds, 150 and 300 documents pel; minute.
The Universal Document Transport can be equipped to read characters printed in the
OCR'B' font and hand-made or printed marks from forms of a wide variety of sizes. The
reading function is controlled by marks along the edge of each form. Under program
control, the 8101 can also route documents to anyone of three stackers; additional
stackers may be included in modules of three. The 8101 can also be equipped to read
standard 80-column punched cards, though it lacks automatic code translation facilities.
The unit is plugboard-controlled and has a peak speed of '600 documents per minute.
The 8500 MICR Sorter/Reader reads characters encoded in the OCR'A' font at a peak
speed of 1,200 documents per minute. It has 18 stackers and can be used for off-line
sorting .

• 27

Digital Plotter
\

The 1934 Digital Incremental Plotter, available with the entire 1900 Series, plots one
variable against another to produce, under program control, annotated graphs complete
5/69

A

(Contd. )

.,

AUERBACH

,

SUMMARY REPORT

. 27

1855:011.270

Digital Plotter (Contd. )
with scale markings, calibrations, legends, and curve identification symbols. This
device is available in six models offering various incremental steps and plotting widths.
Step sizes range from. 1 millimeter to .01 inch. Plotting widths are 11 and 29.5 inches.
For all models, X- and Y-axis plotting speeds are 300 steps per second, and the 2-axis
speed is 10 operations per second. Chart length is 120 feet, and the data transfer rate,
through the standard interface, is 300 characters per second .

.3

COMMUNICATIONS EQUIPMENT
Communications equipment for the 1900 Series comprises remote or local terminals and
their associated peripherals, able to connect with 1900 Series processors by telephone,
telegraph, or local lines through computer and control equipment. Control equipment
attached to a 1900 Series processor can operate through either single channel or multichannel (multiplexor) connections .

. 31

Teletypewriters
Four models of the 7071 Typewriter can be connected to a 1900 Series processor, locally
or remotely over leased telegraph lines either through a single channel (7070) or a multichannel connection. Each teletypewriter is a character by character printer with a fourrow keyboard. Friction-feed and sprocket-feed platens can be fitted to the typewriters.
Each teletypewriter operates at speeds of 6-2/3 or Hi characters a second .

. 32

Alpha-numeric Display Units
Many display units are available for use with 1900 Series processors either locally or
remotely. The ICL 7153 may act as an out-station, communicating with a central processor over leased telephone lines. These devices can display 13 lines of characters,
with either 40 or 80 characters to a line, chosen from a 64-character font. Four function
keys are provided for editing •

. 33

Message Buffering Multiplexors.
The 7900 Series communications equipment includes three basic modules:

. 34

wh~ch

•

Line termination units (7922-7923)

allow connections to telegraph lines;

•

Line Scanners (7920/1, 7920/2) which sample each line, assemble received
bits into characters, transmit characters as bits, and perform code translation; and

•

Linking unit (7921) which connect line scanners to line control and messagebuffering units (7901, 7902) .

Character Buffering Multiplexors
Two character buffering multiplexors, Model 7007/2 and a 7900 module, are available
for 1900 Series processors.
The 7007/2 allows up to 62 half-duplex transmission lines to be connected to a 1900 Series
processor. Transmission lines are linked to the 7007/2 via data terminals, chosen to
suit the transmission characteristics of the line and data .

.4

SOFTWARE
ICL 1900 Series software includes Executive, which oversees the system operations; two
operating systems, GEORGE and Automatic Operator; the PLAN assembler; compilers
for NICOL, COBOL, ALGOL and FORTHAN; and a group of library routines and application packages .

. 41

Executive
Executive is supplied with each 1900 Series central processor and resides in a protected
area of core memory. Its general purpose is to take over, from both programmer and
operator, the execution of routine error-prone tasks and to organize the running of user
programs in the most efficient manner, including programs originally written for a less
powerful configuration. To achieve this purpose, Executive has the following four main
functions:

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

5/69

lel 1900 SERIES

1855:011.410

.41

. 42

Executive (Contd. )
•

To sanction, initiate, and check all peripheral transfers.

•

To execute all "extracode" functions. (These are routines contained in
Executive that generally simulate hardware functions, such as floating-point
arithmetic, which are not included in a particular 1900 Series configuration.)

•

To control, execute, or initiate all communications with the operator via
the console typewriter.

•

To organize and control multiprogramming and dual-programming .

Operating Systems
In addition to the various versions of the Executive, operating systems of varying degrees
of complexity are provided for the 1900 Series, to insure increasingly automatic operation
for the larger members of the 1900 Series. For the smaller processors of the range, an
Automatic Operator is provided, which increases the operating efficiency of the Executive
by providing for the input of control messages to the system on cards or paper tape.
Operating systems for the larger members of the Series are given the acronym GEORGE
(GEneral ORGanizational Environment). Each of the versions of GEORGE provides for
the processing of a stream of jobs interspersed with control messages. GEORGE 1 provides complete control of the running of a job stream, providing the operator with a log
indicating peripheral devices which require operator attention. Background jobs can be
multiprogrammed with the main job stream. GEORGE 2 provides additional facilities for
the maintenance of input and output buffers on magnetic tape or discs.
GEORGE 3 provides full control of multiple job streams, while GEORGE 1 and 2 control
single job streams with or without concurrent multi-programming. Integral with
GEORGE 3 is a multiple on-line programming module (MOP), which provides multiaccess facilities to a virtually unlimited number of users, up to 60 of which can be serviced at one time. Facilities are included for the interactive line-by-line input of programs. Multi-access facilities on smaller 1900 Series installations are provided by
Mini-MOP, a scaled-down version of MOP, which allows servicing of up to 16 enquiry
typewriters with the concurrent running of background jobs. The initial version of
Mini-MOP offered FORTRAN, ALGOL and JEAN, ICLl s conversational language developed
for the solution of technical calculations. The initial version catered for up to nine users
operating simultaneously.
The GEORGE 4 operating system will provide an operating system for 1900 Series machines with hardware paging units .

. 43

Language Processors
The primary language processors, available for the entire 1900 Series subject to their
main memory requirements, are the PLAN Assembler and the NICOL, COBOL, FORTRAN
and ALGOL compilers. These processors have exactly the same form regardless of the
operating system under which they are run. Output from the PLAN assembler and from
the FORTRAN, COBOL and ALGOL compilers is compatible in the sense that subprograms
compiled from any of these different source languages can be linked into a single object
program at load time .

. 44

Assembler
The basic 1900 Series programming language is PLAN, which is available in several
sizes, based on central processor memory capacity. The PLAN language enables the
programmer to describe the work to be done by a program in an unambiguous and logical
manner, using symbolic names rather than numbers. PLAN includes facilities for
"pseudo-operations", which are single line statements that generate calls to ready-made
subroutines at assembly time .

. 45

NICOL Compiler
NICOL is a commercial programming language for the smaller 1900 Series computers.
It is designed specifically for 4096-word computers without auxiliary storage, but it can

also be used in larger installations. The features of this language closely resemble the
facilities of a tabulator. NICOL is, therefore, particularly suitable for data processing
installations changing over from tabulating equipment to a computer.

5/69

fA

AUERBACH
®

(Contd. )

SUMMARY REPORT

.46

1855:011.460

Library Routines and Applications Packages
Several hundred programs and subroutines are available in the ICL program library for
the 1900 Series computers. These include routines of general commercial, scientific,
or mathematical interest and routines that facilitate peripheral device handling. Most
commercial routines are oriented to the Sterling monetary system. ICL provides complete programs for many particular commercial and scientific needs. Among these are
ICL PERT, PROMPT (for production control), SCAN (for inventory management), FIND
(for information retrieval), NIC (for indexing and cataloging), PROP (for investment
evaluation), and COMPAY (for company payrolls). Other packages cover such fields as
civil engineering, traffic, engineering, linear programming, matrix calculations, and
statistical analysis.

047

FORTRAN Compilers
The first version of FORTRAN implemented on 1900 Series computers was developed
before the publication of the American Standards Association (A. S. A.) draft. It is equivalent in general scope to A. S. A. Basic FORTRAN although ithas some FORTRAN IV features.
Presently available on the 1900 Series computers are FORTRAN IV or 1900 FORTRAN
and the original version developed for the series. However, 1900 FORTRAN does include
some extensions, and also some slight restrictions to obtain what is basically a one-pass
compiler.
1900 FORTRAN has a feature for multiple assi~nment statements which is not available
in standard FORTRAN. A multiple assignment statement gives the same value to more
than one variable or array element. It has the form
V 1 , V 2 , . . . . Vn = e
where each V is a variable or array element and e is an expression.
1900 FORTRAN allows TYPE INTEGER expressions for m1, m2, m3 in the statement
DO Ai = m1, m2, m3 provided they have prcviously been defined. Some input and output
subroutines are provided in 1900 FORTRAN to facilitate multiprogramming operations in
dynamically controlling which peripherals arc currently available to a program. Also
special forms of the I, D, E, F and G descriptions are provided to allow input records to
have a free format .

. 48

COBOL Compilers
There are two versions of 1900 COBOL, namely Compact COBOL and Full COBOL, for
which there are seven compilers available. The Compact COBOL compilers require
machine configurations having at least 8,192 storage locations while the Full COBOL
compilers require at least 16,384 core storage locations. Common to all configurations
on which these compilers are run is the following equipment: one central processor, one
paper tape reader or one card reader, one line printer, one card or paper tape punch if
the object program is required on cards or paper tape, and one console typewriter.
other peripherals such as discs and magnetic tapes are needed with their respective
versions.

049

other Packages
1900 Algol is almost a full implementation of Algol 60 and includes as subsets ECMA
ALGOL and SUBSET ALGOL 60 (IFIP).
Compilers are currently available on paper tape or magnetic tape using card or paper
tape input.
1900 EMA (Extended Mercury Autocode) is a considerably more powerful language than
Mercury Autocode, which it includes as a subset.
Simulation is a technique for obtainin~ information about the performance of a system
without ever putting that system into operation. A model is constructed so that the results obtained by operating or modifyin~ the model indicate the results to be expected
when the corresponding real system is operated or modified. SIMON and CSL, simulation languages, have been developed to simplify such processes. SIMON uses ALGOL
procedures and 1900 CSL uses FOHTRAN procedures.
The IBM 1401 and ICL 1500 Simulators operate under the SPAN system. Both simulators
operate on the 1903 or larger machines. The 1401 Simulator accepts even parity 1401
tapes (200, 556 or 800 bpi). A 1500/1900 file conversion program is available.
© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

5/69

1855:011.490

. 49

ICL 1900 SERIES

Other Packages (Contd. )
Students taking City and Guilds computer courses at technical colleges throughout Britain,
write programs in the City and Guilds language to be run on ICL' s 1900 Series computers
by using an ICL developed compiler.
ICL IS Rapidwrite for the 1900 Series is a commercial autocode based on COBOL. COBOL
being a rather wordy language sometimes seems cumbersome to the user not requiring
the self-documenting feature of COBOL. Rapidwrite is designed to relieve such a user of
that burden by enabling the user to write his source programs in an abbreviated shorthand form while retaining the full self-documenting features of COBOL via the Rapidwrite
system .

.5

PRICE DATA
ICL does not issue a general price list.
published in a future issue.

5/69

Price data for specific configurations will be

A

AUERBACH
@

WEST GERMANY

AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

AUERBACH
®
Printed in U.S.A.

A

1950:011. 100

AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

SIEMENS SYSTEM 4004
SUMMARY REPORT

AUERBACH

s

SUMMARY REPORT: SIEMENS SYSTEM 4004
.1

BACKGROUND
Siemens System 4004 is the third-generation family of central processors. peripheral
devices. and supporting software marketed by Siemens AG. Munich. West Germany.
Some noteworthy characteristics of the Siemens System 4004 are:

•

The high degree of program compatibility, both upward and downward, among
four of the six Siemens 4004 processor models. Compatibility also exists with
certain IBM System/360 and RCA Spectra 70 processors through similar hardware design and compatible source languages.

•
•

The numerous arithmetic modes and data forms.
The wide range of input-output and storage devices.

•
•

The availability of emulators on the 4004/35 and 4004/45.

•

The extensive software support as evidenced by several levels of integrated
operating systems.

The use of monolithic integrated circuits.

The 4004/45 and 4004/55 processors were first delivered in the second half of 1966 and
the 4004/35 was introduced in 1967. First deliveries of the 4004/16 and 40004/26 processors are scheduled for October 1969. Recently. Siemens announced the 4004/46 processor. which is a modified 4004/45 including a virtual memory for time-sharing use.

Figure 1. Siemens 4004/45 Computer System

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

10/69

SIEMENS SYSTEM 4004

1950:011.200

.2

DATA STRUCTURE
The Siemens System 4004 data structure is based on the Extended Binary Coded Decimal
Interchange Code (EBCDIC). The system is also capable of receiving, processing and
sending information in the United states of America standard Code for Information Interchange (USASCI!).
Alphanumeric data is represented by 8-bit bytes; each byte can represent one alphanumeric character, two decimal digits, or eight binary digits (bits).
Fixed length data of 16, 32, 64 bits, or variable length data of up to 256 bytes can be
processed. For purposes of data description, the terms "character" (one byte), ''halfword" (two bytes), "word" (four bytes), and "double word" (eight bytes) are used. In the
Siemens System 4004 a two-part system of memory addressing is used. According to
this system, a machine address consists of a base address and a displacement address.
The displacement address, specified in the instruction, consists of twelve bits and enables the programmer to address a region of up to 4096 bytes. The base address, stored
in one of the General Purpose registers (32 bits), is used for extended addressing. The
base address, in effect, subdivides memory into sections of 4096 bytes, and the displacement address specifies the individual byte within the 4096 byte section.
The Siemens System 4004 uses several instruction formats including two-, four- and sixbyte forms. Each instruction contains an operation code (one byte) and a maximum of
three operand addresses. Each main memory reference address can be indexed by one
of the General Purpose registers in combination with the Base Address register. The
Model 4004/16 and 26 processors contain instruction complements which are functional
subsets of the larger 4004/35, 45 and 55 instruction complements. Floating point operations are standard on the 4004/35, 45, and 55 processors. All models are capable of
decimal and binary arithmetic operations. Decimal operations are performed on variable- or fixed-length operands in "packed" format - two digits packed into one byte.
The maximum field size permitted is 31 digits plus sign; the operations are performed
from storage to storage.
Binary, fixed-point arithmetic operations use either storage or registers for computation
depending on the processor model. In the Model 4004/16 and 26 Processors, a maximum
field length of 127 bits plus sign bit is permissible. In the Model 4004/35, 45, and 55
Processors, field length is limited to 31 bits (Plus sign), but arithmetic operations are
performed in the registers and several can be coupled to increase precision.
Where floating-point arithmetic operations are provided, four additional double-length
registers are supplied for computation and storage of results. Both short and long precision (four and eight bytes respectively), are provided. The absolute value of floatingpoint numbers can range from 2.4 x 10-78 to approximately 7.2 x 1075 •

.3

HARDWARE

.31

Central Processors and Main Memory
Presently, six processors are offered within the Siemens System 4004: 4004/16, 4004/
26, 4004/35, 4004/45, 4004/46, and 4004/55. Four of the models are program-compatible and cover a broad range of business and scientific applications, 4004/35, 4004/45,
4004/46 and 4004/55, the 4004/46 is a 4004/45 modified for time-sharing. The 4004/
16, with a restricted instruction repertoire, is probably best suited as a satellite of
larger 4004 processors. The 4004/26 has a limited instruction repertoire. Memory
cycle times range from 1.44 microseconds for the 4004/35 to 840 nanoseconds on the
4004/55. One to four bytes are fetched per memory cycle, depending on the processor
model. Memory storage capacities range from 8,192 bytes to 524,288 bytes. The characteristics of the main memories of the processor models in the System 4004 are given
in Table I.
Simultaneous operation of peripheral devices is possible through the use of input-output
channels operating independently of each other. Depending on the processor model, one
multiplexor channel and several selector channels are available. The multiplexor operates in a time-multiplex mode and thus provides for time-shared input-output operation •

. 32

Scratchpad Memory
The Scratchpad Memory, a fast micro-magnetic storage device, contains the GeneralPurpose, Floating-Point and various other registers required for program and inputoutput control, and has a capacity of 128 four-byte words. The maximum cycle time is
300 nanoseconds for one word. Locations in scratchpad memory are uniquely addressed
by the Load Scratchpad and Store Scratchpad instructions.

10/69

A

(Contd.)

AUERBACH

'"

1950:011.330

SUMMARY REPORT

TABLE I: SIEMENS SYSTEM 4004 MAIN CORE STORAGE
CHARACTERISTICS
Core Storage
Capacity, Bytes

•
•

8,192
16,384
32,768
65,536
131,072
262,144
524,288

26

•
•
•

Cycle Time, jJ.Sec
Bytes Accessed
per Cycle
Effective Cycle
Time per Byte, ,."sec

.33

16

35

45

•
•

•

46

•

••

•

55

•
•
•
•

0.88

0.88

1.44

1.44

1.44

0.84

1

1

2

2

2

4

0.88

0.88

0.72

0.72

0.72

0.21

Console and Typewriters
The Model 4004/97 Console and Typewriter, available with Processors 35, 45, 46 and 55,
is a free standing self-contained unit which provides for operator control and supervision
of processing.
A set of pushbutton switches allows programs or data to be loaded into memory, allows
program start and interrupt and also displays the current processor state. The typewriter also allows the operator to communicate with the operating system and vice versa.
A parity bit is generated for each character sent from the typewriter to the processor
and a bit by bit echo check is performed on all characters coming from the processor to
the typewriter.
The Model 4004/4217 interrogating typewriter is a peripheral device intended for use
with processors 16 and 26. Data interchange with the processor is at a maximum of 20
characters per second. A line can contain 72 characters and one carbon can be prepared.
A parity bit is generated for each character sent to the processor and a bit by bit echo
check is performed on all characters sent to the typewriter •

• 34

Auxiliary Storage
Up to 12 random access devices can be connected to one controller. The controller has
a set of commands that are translated into specific commands for each of the different
devices. Seeking and selecting specific data within a file is accomplished independently
of the processor.
The Model 4004/564 Disc Storage Unit has a capacity of 7.25 million bytes of information
held on an interchangeable disc pack. Up to 8 units can be attached to one 4004/551
Controller. The unit provides 200 data tracks plus three alternative tracks per read/
write head and has ten such heads. The total disc pack capacity is 2,000 tracks, each
having packing densities of 433 bits per centimeter (1100 bits per inch) for a total of
3,660 bytes. The peak data transfer rate between the processor and the disc storage unit
is 156,000 bytes per second. The minimum seek time is 30 milliseconds; average seek
time is 75 milliseconds and maximum seek time is 135 milliseconds. The disc rotates at
2400 revolutions per minute giving an average latency of 12. 5 milliseconds.
The Models 4004/567-8 and -16 Drum Memory Units are used to provide virtual memory
for the time-sharing 4004/46 processor. They have an effective transfer rate of 277,000
bytes per second and a rotational time of 20 milliseconds. The Model 567-8 has a capacity
of 4, 128, 800 bytes per unit, while the 567-16 has a capacity of 8,257,600 bytes.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

10/69

1950:011.340

.34

SIEMENS SYSTEM 4004

Auxiliary storage (Contd.)
The Model 4004/568-11 Mass Storage Unit comprises from one to eight removable magazines. All card magazines are served by a common read/write station. The Model
4004/551 Controller can accommodate up to four Mass storage Units. The basic storage
element of the unit is a 16 by 4 1/2 inch magnetic card, which records data on one side
only. Each card has 128 separately addressable tracks containing 2,139 bytes each and
each card magazine houses 256 cards.
The Models 4004/4570-1 and 4004-4570-2 Direct Access Storage Units are operated by
their own control electronics. The Model 4570-1 provides random-access storage for
233.4 million bytes on 8 (plus one stand-by) interchangeable disc packs and the Model
4570-2 for 145.4 million bytes on 5 interchangeable disc packs. One disc pack provides
200 data tracks plus three alternative tracks; thc total disc pack capacity is 29. 17 million bytes, the track capacity 7,294 bytes. Data is transferred between the processor
and the Direct Access Storage Unit at the peak transfer rate of 312,000 bytes per second.
The average seek time is 60 ms. Since the disc rotates at 2,400 rpm there is an average latency of 12.5 ms. Read-read, read-write or write-write simultaneity is possible
on any two disc packs serviced by the Model 4004/4572 Auxiliary Storage Control.

. 35

Sequential Input-Output Units
The peripheral devices for the Siemens System 4004 are listed in Table II, together with
their rated speeds.
The Model 4004/243 High Speed Printers are drum-type printers capable of printing up to
1,250 lines per minute. A line contains up to 160 print positions. A maximum of 64 different characters can be printed. The printers are buffered, which means that once they are
loaded they act independently of the central processor.
Paper advance for the first line is 12 milliseconds; for lines 2 to 8 it is 6.67 milliseconds; for line 9 onwards it is 2.22 milliseconds. Fan-fold paper is used and its
width can vary from 4 to 8 3/4 inches.
Data is checked during transmission from processor to buffer and buffer to printer. A
low paper condition or an invalid character in the buffer causes an indication to be sent to
the processor.
The l\Iodel 4004/4247 high speed printer is a drum type printer which operates at a maximum of I, 500 lpm for numeric print. Each line is 132 characters in length. There are
64 different characters in the print set.
Paper advance time for the first line is 16.5 milliseconds; each subsequent line requires
6.35 milliseconds. Fan-fold paper is used varying in lengths from 2 to 213/4 inches including margins. An optional feature is a second paper transport which enables the
printer to print on two forms simultaneously and independently. The width of both forms
cannot exceed 22 3/4 inches.
The Model 4004/4241 and /4242 Printers are buffered train printers capable of printing up
to 910 lines per minute. A single line contains 80-, 136- or 160-print positions depending
on the model.
Paper advance time for the first line is 18.85 ms, each subsequent line requiring 6.35 ms;
high-speed paper advance time from line 6 onwards is 2.17 ms per line. The Model 4004/
4242 has a second paper transport for printing on two forms simultaneously at a maximum
of 143 print pOSitions per line.
Thc Model 4004/4245 Printer is a buffered drum type printer which operates at a maximum
of 285 lines per minute. Each line is 120 or 136 characters in length, the character set consisting of 48 or 64 different characters.
Paper advance time is 24 ms for the first line and 12 ms for each subsequent line.
density is 6 or 8 lines per inch.

10/69

A

AUERBACH
~

The line

(eontd. )

SUMMARY REPORT

.35

1950:011.350

Sequential Input-Output Units (Contd.)
The Models 4004/4250 to 4004/4250-5 are Videos can Document Reader-Sorters capable of
reading numerical characters in OCR-A, IBM-407 -1, or OCR-B font at a maximum speed
of 1600 documents per minute. The device has 16 sort-pockets.
The Model 4004/4251 Videoscan Document Reader reads OCR-A numerical characters at a
rated speed of 1600 documents per minute in 3 sort-pockets.
The Model 4004/4252 Videos can Document Reader reads OCR-A numerical characters at a
rated speed of 750 documents per minute in 12 sort-pockets.

TABLE II: SIEMENS SYSTEM 4004 INPUT-OUTPUT UNITS
Device
Card Punch 4004/234-10, -11
Card Punch 4004/236-10, -11
Card Reader 4004/237-10, -21
Card Reader 4004/4235
Card Punch 4004/4238
Card Reader 4004/4239-10, -20
Paper Tape Reader 4004/4223
Paper Tape Punch 4004/4225, -s
Paper Tape Reader 4004/4226
Paper Tape Reader 4004/4227, -s
Paper Tape Reader 4004/4227-0
Paper Tape Punch 4004/4228, -s
Edge-Punch Card Reader 4004/4280,
/4281
Edge- Punch Card Punch 4004/4282
Printer 4004/243-30, -40
Printer 4004/4241-18, /4242-1B
Printer 4004/4241-1D, /4242-1D
Printer 4004/4241-2B, /4242-2B
Printer 4004/4241-2D, /4242-2D
Printer 4004/4245-1
Printer 4004/4245-2
Printer 4004/4247-1-B1, -B2, -B3, -B4
Printer 4004/4247 -1-F1
Printer 4004/4247-2-Bl, -B2, -B3, -B4
Printer 4004/4247 -2-F1
Videoscan Document Reader-Sorter 4004/
4250, -1, -2,-3,-4,-5
Videos can Document Reader-Sorter 4004/
4251
Videoscan Document Reader-Sorter 4004/
4252
Console and Typewriter 4004/97
Console and Typewriter 4004/4217

Rated Speed
100
300
1,435
600
420
1,000
1,200
100
400
1, 000
850
150
120

cpm
cpm
cpm
cpm
cpm
cpm
char/sec
char/sec
char/sec
char/sec
char/sec
char/sec
char/sec

28 char/sec
1,2501pm
910lpm
765lpm
465lpm
390lpm
285lpm
200lpm
620lpm
750lpm
620lpm
750lpm
1600 documents/min
1600 documents/min
750 documents/min
10 cps
20 cps

Siemens System 4004 magnetic tape devices are capable of processing 7 or 9 track tape
and are available in several versions. All magnetic tape devices are capable of reading
in the forward and reverse directions. The Siemens System 4004 magnetic tape devices
are fully compatible with the IBM Series 2400 magnetic tape devices and tapes can be interchanged between the two systems as required. Table III summarizes the characteristics of the available devices, while Table IV summarizes the available tape controllers.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

10/69

SIEMENS SYSTEM 4004

1950:011.351

TABLE III: CHARACTERISTICS OF SIEMENS SYSTEM 4004
MAGNETIC TAPE TRANSPORTS

10/69

Peak Speed,
Char Per
Second

Recording Density,
Characters
Per Inch

432 (7 track)

37.5

200
556
800

7,500
20,850
30,000

0.75

254

432 (9 track)

37.5

800

30,000

0.6

254

441 (7 track)

50.0

333
500

16,600
25,000

0.6

96

442 (7 track)

75

200
556
800

15,000
42,000
60,000

0.75

195

442 (9 track)

75

800

60,000

0.6

195

4432 (7 track)

37.5

200
556
800

7,500
20,850
30,000

0.75

130

4432 (9 track)

37.5

800

30,000

0.6

130

4442 (7 track)

75

200
556
800

15,000
42,000
60,000

0.75

130

4442 (9 track)

75

800

60,000

0.6

130

4443 (7 track)

75

200
556
800

15,000
42.000
60,000

0.75

60

4443 (9 track)

75

800

60,000

0.6

60

4446 (7 track)

150

200
556
800

30,000
83,400
120,000

0.75

60

4446 (9 track)

150

800

120,000

0.6

60

4451 (9 track)

37.5

800
1600

30,000
60,000

0.6

130

4433 (9 track)

75

800
1600

60,000
120,000

0.6

130

A

Interblock
Gap Size,
filches

Full Rewind
Time,
Seconds

Tape Speed,
Inches Per
Second

Model
No.

(Contd.)

AUERBACH

'"

SUMMARY REPORT

1950:011.360

TABLE IV: SIEMENS SYSTEM 4004 MAGNETIC TAPE CONTROLLERS
Tape Devices Controlled

Tape Controller

Version

Model 4004/463

432
441
442
4432
4442
4443
4446

9 or 7-Track

Model 4004/472

432
442
4432
4442
4443
4446

9-Track

Model 4004/473

432
442
4432
4442
4443
4446

9 or 7-Track

Model 4004/4472

432
442
4432
4442
4443
4446

9-Track

:\lodel 4004/4475

432
442
4432
4442
4443
4446
4451
4453

9-Track

Model 4004/4476

4451
4453

9-Track

Each of the six tape controllers is available in four versions:
- 108
-

116

- 208
- 216
.36

single-channel; max. 8 devices
single-channel; max. 16 devices
dual-channel; max. 8 devices
dual-channel; max. 16 devices

Simultaneous Operations
The control of the transmission of data between the processor and an associated peripheral
device is accomplished through channels and the Siemens Standard Interface. A channel
may be considered as an independent unit controlling data flow to and from the processor
and releasing control to the input or output device. This release allows the processor to
function simultaneously with the input-output operation. Each channel utilizes its own set
of commands to perform input-output operations. These commands, referred to as channel commands, control the device once a start command has been issued by the processor. Chaining of channel commands provides a means by which several operations, such
as multiple reads, may be completed independent of processor operation.
The Siemens Standard Interface functions as a connector between the channel and a device
control. The interface establishes an identical relationship with each input-output device
in that any device may be connected to the interface regardless of type, size or speed.
The number of channels connected varies with the processor model. System 4004 offers
two types of channel: Selector Channels and Multiplexor Channels.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

10/69

1950:011.361

.36

SIEMENS SYSTEM 4004

Simultaneous Operations (Contd.)
The Selector Channels control the transfer of data to and from peripheral devices. A
Selector Channel has from one to four trunks, depending upon the processor model. The
type of controller cOlmected to a trunk determines the number of devices that can be connected to that trunk. For example, a tape controller controls as many as 16 tape devices.
Selector Channels can operate concurrently; however, each Selector Channel can operate
only one device at a time.
The Multiplexor Channel can control up to nine input-output trunks, through which up to
256 devices can be addressed (on the 4004/35 and higher-numbered processors, one
trunk is required to connect the operator's console). The Multiplexor has two modes of
operation - Multiplex Mode and Burst Mode. The Multiplex Mode is used for the simultaneous operation of a number of low-speed devices on the Multiplexor. while the Burst
Mode (which is not available for the Models 16 and 26) allows the operation of high-speed
devices on the Multiplexor. Only one device at a time can be used on the Multiplexor
when it is operating in Burst Mode •

• 37

Memory Protect
The protection of memory segments from destruction by overlaying is provided in Model
,1004/35, 45, 46, and 55 Processors through the optional Memory-Protect feature. This
feature can protect up to 15 memory segments; the basic segment is 2,048 bytes and can
be increased in multiples of 2.048 bytes.

The memory-protect feature consists of a set of registers that are constantly scanned
during instruction execution to ascertain that the data field operated upon by an instruction
resides within preset limits. An interrupt condition is set when a program attempts to
modify data located in an area of memory not allocated to this program or program
section.
This feature greatly enhances the multiprogramming capabilities of the Siemens System
4004 •
• 38

Elapsed-Time Clock
An elapsed-time clock for real-time processing applications is available in the Siemens
System 4004. This feature is standard on the Model 4004/26 Processor and optional in
the I\Iodel 4004/35, 45, 46, and 55 Processors. The clock is maintained on a power line
frequency (50 cycles per second) rate providing an interrupt capability every 20 milliseconds. The clock can be reset at any time by programmed instruction .

. 39

Data Communications Equipment:
The Data Exchange Control enables two, close-by Siemens System 4004 Processors to
communicate with each other. Data can be transferred at high speed in either direction,
but in only one direction at a time. The transmission rates depend upon the type of
channel used and the number of simultaneously active devices that are attached to the
transmitting and receiving processors.
The Data Exchange Control can be a powerful tool for multiprocessor applications. In a
real-time system, for example, status information and queue tables can be continuously
transferred from the operating processor and to the backup processor.
This permits the backup processor to assume immediate control and continue processing
operations in case of a failure in the operating processor.
The Model 4004/668 Communications Controller (multichannel) operates on the processors
/35, /45, /46, and /55 and terminates from 1 to 48 communications lines. The maximum
data rate per line is 2,400 bits per second. The maximum throughput that one 4004/668
can handle is 6,000 character per second. The Model 4004/4666 Communications Controller
(multichannel) operates on the /35, /45, and /55 processors and serves from 1 to 30
communications lines. The maximum data rate per communications line is 4,800 bits per
second; the maximum total communications data rate of one 4004/668 is 8,000 characters
per second.
Each of the scan positions requires a communications buffer, and in some cases a data set,
to interface with the communications line. Each scan position uses one Multiplexor subchannel.

10/69

fA

AUERBACH

(Contd. )

1950:011.390

SUMMARY REPORT

.39

Data Communications Equipment (Contd.)
The Communications Control (single-channel) Model 4004/656 permits half-duplex commuications between one System 4004 computer system and another System 4004 computer
system that is equipped with the appropriate communications equipment. Different
attachments permit communication over common-carrier leased voice-band or broadband line .

.4

SOFTWARE
Siemen's software systems for the System 4004 series, in general, parallel the structure
and contents of the software supplied by RCA for its Spectra 70 series. Siemens provides "third-generation" software through such facilities as disc-oriented control systems, disc file language facilities, automatic on-line file management techniques, and
comprehensive data communications control routines. Multiprogramming control for up
to six jobs is provided for Siemens 4004 systems that have a minimum of 32K bytes of
core storage.
The principal levels of Siemens 4004 software are deSignated Primary Operating System,
Tape Operating System, Tape-Disc Operating System, Disc Operating System, and Time
Sharing Operating System in order of increasing complexity and capability. Software for
the small-scale Siemens 4004/16 system, however, is a specially-designed, card-oriented
set of routines that provides assembly language, Report Program Generator, I/O control,
and service routine facilities at the 4K-byte corc storage levels and additionally includes
a Sort/Merge Generator at the SK-byte core storage level. The system can also be supplied in a magnetic tape-oriented version.
The Siemens 4004/26 operating system offers baSically the same supervised facilities as
the Primary Operating System (POS) for the larger Siemens 4004 systems and functions
with a minimum hardware configuration of 16K bytes of core storage. The principal
limitation of 4004/26 POS facilities in comparison with the ros facilities for 4004/35,
4004/45 and 4004/55 systems is the omission of a COBOL language processor.
A limited Disc Operating System (DOS) is available on the 4004/16 and 4004/26 systems
only and is available in card and paper tape versions •

. 41

Primary Operating System (POS)
The Primary Operating System for use with the Siemens 4004/35, 4004/45 and 4004/55
systems is a magnetic tape-oriented software system that provides basic supervisory
control for sequential execution of programs, interrupt control, input-output control, as
well as a COBOL compiler, assembler, report program generator, and standard utility
routines. POS COBOL is a subset of full COBOL 65 and requires a minimum of 32K
bytes of core storage for compilations. The POS Assembler also requires use of 32K
bytes of core storage.
No FORTRAN processor is provided under POS, nor are any routines supplied for the
automatic control of random-access devices, although the operation of these devices is
programmable at the assembly-language level. Multiprogramming capabilities are provided through the Peripheral Control Routine, which permits concurrent operation of up
to three data transcription routines •

. 42

Tape Operating System (TOS)
The second major level of Siemens 4004 software support designed for use with the
4004/35, 4004/45, and 4004/55 systems is designated the Tape Operating System (TOS).
TOS is a magnetic tape-oriented integrated software package that provides supervisory
control programs, language processors, and utility programs for installations that have
a minimum hardware configuration of 65K bytes of core storage. The facility to control
multiprogrammed operation of up to six programs concurrently is the primary feature of
TOS software. The basic TOS Executive program requires a minimum of 16K bytes of
storage. The Monitor program that coordinates the operations of stacked-job processing
requires an additional4K bytes, and the File Control Processor for input-output devices
and file control requires another 4K to SK bytes of core storage.
In addition to a comprehensive assembly system, TOS offers a COBOL language and a
full-scale FORTRAN IV language.

© 1969 AUERBACH Corporation and AUERBACH Into. Inc.

10/69

1950:011.430

.43

SIEMENS SYSTEM 4004

Tape/Disc Operating System (TOOS)
Siemens' Tape Disc Operating System (TOOS) is an improved and extended version of its
Tape Operating System (TOS). In addition to all TOS software facUities, TOOS offers
options that permit system control routines, problem programs, and library subroutines
to reside on discs or drums in order to improve the Siemens 4004's throughput capabilities.
As a result, more efficient multiprogrammed operations are possible than with the tapeoriented TOS. The efficiency of program preparation is increased due to a reduction in
the number of Job Control Language statements required to prepare and compile object
programs. The TOOS software package also includes a set of input-output routines for
the control of data communications devices.
TOOS requires a minimum hardware configuration of 65,536 bytes of core storage, of
which 16,384 bytes are permanently reserved for the Executive •

• 44

Disc Operating System (OOS)
The Siemens 4004 Disc Operating System (DOS) provides a comprehensive disc-oriented
multiprogramming control and operating environment for the 4004/35 and highernumbered processors. It requires a minimum memory capacity of 32,768 bytes, and
provides multiprogramming of up to six programs. Major features include a relocatableloader, a non-resident monitor, the ability to control input jobstreams concurrently, and
the sharing of random access file space by several programs. DOS is functionally composed of three component groupings: a system for controlling the entire system environment; language translators for compiling and/or assembling symbolic source programs
into machine-coded programs ready to be placed in executable form by the linkage editor;
and a utility system providing a variety of system service functions such as diagnostics,
library maintenance, Sort/Merge, etc •

. -15

Time Sharing Operating System (TSOS)
The newly announced Time Sharing Operating System (TSOS) is a specialized software system designed for control and support of the new Siemens 4004/46 time-sharing processor.
The system is scheduled to provide advanced time-sharing capabilities, as well as improved
facilities for handling batch processing in a multiprogramming mode.
Provided within the TSOS Executive program will be routines for handling task scheduling
(capable of using a time-slicing algorithm), memory management, device allocation,
physical-level input-output, and a combination command and job control language. Also
provided as a TSOS system program will be a File Control Processor (FCP) with extensive automatic data management capabilities. In addition to all of the language processors
and utility routines used with TOS and TDOS, TSOS will also provide a full conversational
FORTRAN compiler, plus conversational text editor and desk calculator programs. Conversational syntax checking will also be provided for the FORTRAN, COBOL, and Assembler
languages.
The minimum equipment required to use thc Time Sharing Operating System includes a
Siemens 4004/46 processor with at least 262K bytes of core storage, a 4004/567 Drum
Memory Unit, two 4004/564 Disc Units, two 9-track magnetic-tape units, one card reader,
and one printer. To support conversational users, the system must also include a 4004/
668 Communications Controller - Multi-channel and from 1 to 48 remote terminal units .

.5

COMPATIBILITY

.51

Program Compatibility Within the Siemens 4004 Line
Siemens emphasizes the high degree of program compatibility, both upward and downward, among the following models of Siemens 4004: 4004/35, 4004/45 and 4004/55.
Among these three models, any valid program that runs on configuration A will run on
configuration B and produce the same results if:

10/69

•

Configuration B includes the required amount of main storage, the same or
compatible input-output devices, IU'ld all required special features; and

•

The program is independent of the relationships between instruction execution
times IIlld input-output rates.

A

(eontd.)

AUERBACH
~

1950:011.520

SUMMARY REPORT

.52

Program Compatibility with the IBM System/S60 and RCA Spectra 70 Series
Siemens provides program compatibility with both RCA Spectra 70 Series and IBM
System/S60 through the Siemens 4004 source language. The Siemens 4004 COBOL and
FORTRAN languages are in many cases identical to their Spectra 70 and System/S60
counterparts. Since the instruction repertoire of the large Siemens 4004 processors is
virtually identical with those of the similar-sized IBM and RCA processors. Siemens is
able to realize program compatibility with t4em at the assembly-language level as well.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

10/69

A
..

AUERBACH

AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

1950:221. 101
SIEMENS SYSTEMS 4004
PRICE DATA

PRICE DATA

PRICES

IDENTITY OF UNIT
CLASS

Model
Number

Feature

Name

Monthly
Rental
DM

Monthly
Maint.
DM

Purchase
DM

3,395
5,870

135
230

165,000
286,000

6,575
10,475
14,980

260
415
590

320,000
510,000
730,000

13,163
18,428
556
224
878
1,755
1,433
702
1,209

518
730
24
10
38
72
58
29
48

641,250
897,750
26,980
10,735
42,750
85,500
69,825
34,200
58,900

18,428
23,693
30,713
47,385
556
224
878
2,204
2,305
2,204
2,305
965
1,658
2,418
2,515

730
936
1,210
1,867
24
10
38
86
95
86
95
38
67
96
100

897,750
1,154,250
1,496,250
2,308,500
26,980
10,735
42,750
107,350
112,100
107,350
112,100
47,025
80,750
117,800
122,600

58,385
540
215
855
525
3,920
5,775
7,450

2,300
25
10
35
25
160
230
295

2,844,30(
26,125
10,450
41,515
25,270
191,000
281,200
362,900

36,641
43,661
60,333
98,943
663

1,445
1,723
2,376
3,898
29

1,785,050
2,127,050
2,939,300
4,820,300
32,110

Processing Unit

PHOCESSOH

Model 4004/16:
Central Processor (8,192 Bytes)
Central Processor (16,384 Bytes)

4004/16-B
4004/16-C

Model 4004/26:
Central Processor (16,384 Bytes)
Central Processor (32,768 Bytes)
Central Processor (65,536 Bytes)

4004/26-C
4004/26-D
4004/26-E

Model 4004/35:
4004/35-D
4004/35-E
5001-35
5002-35
5003-35
5005-35
5006-35
5030
5031

Central Processor (32,768 Bytes)
Central Processor (65,536 Bytes)
Memory Protect
Elapsed Time Clock
Direct Control
301 Emulator
1401 Emulator
Selector Channel
Selector Channel
Model 4004/45:

4004/45-D
4004;45-E
4004/45-F
-l004/45-G
5001-45
;;002-45
5003-45
5005-45
5005-45
5006-45
5006-45
5015
5016
5026-45
5026-45

Ccntral Processor (32,768 Bytes)
Central Processor (65,536 Bytes)
Central Processor (131,012 Bytes)
Central Processor (262,144 Bytes)
Memory Protect
Elapsed Time Clock
Direct Control
Flrst 301 Emulator
Second 301 Emulator
Flrst 1401 Emulator
Second 1401 Emulator
Selector Channcl
Selector Channel
First 1410 Emulator
Second 1410 Emulator
Model 4004/46:

4004/46-G
5001-46
5002-46
5003-46
5019-46
5040
5041
5042

Central Processor (262,144 Bytes)
Memory Protect
Elapsed Time Clock
Direct Control
Elapsed Time Clock
Selector Channel
Selector Channel
Selector Channel
Model 4004/55:

4004/55-E
4004/55-F
4004/55-G
4004/55-H
5001-55

Central Processor
Central Processor
Central Processor
Central Processor
Memory Protect

(65,536 Bytes)
(131,072 Bytes)
(262,144 Bytcs)
(524,288 Bytes)

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

10/69

1950:221.102

SIEMENS SYSTEMS 4004

PRICES

IDENTITY OF UNIT
CLASS

Model
Number

Feature

PROCESSOH
(Contd. )

Name

Monthly
Rental
DM

Monthly
Maint.
DM

Purchase
DM

224
1,102
1,970
3,510
5,051

10
43
82
139
202

10,735
53,675
95,950
171,000
246,050

2,400
24,275
16,575

218
2,460
1,500

105,634
1,080,622
737,843

11,915
23,830

1,645
3,290

580,450
1,160,900

12,734

1,003

620,350

2,311

91

112,433

Processing Unit (Contd.)
5002-55
5003-55
5020
5022
5024

Elapsed Time Clock
Direct Control
Selector Channel
Selector Channel
Selector Channel
Disc
--

MASS
STORAGE
4004/564
4004/4570-1
4004/4570-2

Disc Storage (I. 25 Million Bytes)
Direct Access storage (233.4 Million Bytes)
Direct Access Storage (145.88 Million Bytes)
Drum
-

4004/567-8
4004/567-16

Drum Memory Unit (4, 128,800 Bytes)
Drum Memory Unit (8,257,600 Bytes)
Magnetic Card

4004/568-11

Mass Storage Unit (536,870,912 Bytes)
Cohtroller

4004/551
5501-18
5502-14
5508
5511
5512
5513
4004/4572
45572
4004/4573-1
4004/4573-2
INPUTOUTPUT

NC

NC

NC

775

34

37,620

430
161
63
488
11,061

20
10
5
19
166

20,710
7,695
3,000
23,750
510,906

NC
1,615
2,675

NC
70
105

NC
78,500
130,000

1,970
2,486
44
3,296
3,803
44
2,847
3,627
44
132
NC
10
930
68

274
346
10
456
528
10
394
504
10
10
NC
5
130
10

95,950
120,650
2,138
160,550
185,250
2,138
138,700
176,700
2,138
6,413
NC
247
45,000
3,325

pqnched Card
4004/234-10
4004/234-11
5213
4004/236-10
4004/236-11
5215
4004/236-10
4004/237 -21
5~Qa

5a04

oUI
4(l04/4~35

oliGO
45201

1()/69

Rendom Access Controller
Input/Output Attachment Feature to use
Model 4004/564
Input/Output Attachment Feature to use
Model 4004/568-11
Input/Output Attachment Feature to use
Models 4004/567-8, -16
Off- Line Scan Feature
Record Overflow Feature
Multichannel Switch
Auxiliary Storage Control forModels4570-1,-2
Inpqt/OIItpIIt Attachment Feature for Model
4004/4572
Attachment to use 4570-1,-2
Attachment to use 4570-1,-2

Card Punch (100 cpm)
Card Punch (100 cpm)
Scored Cllrd Feature for 234-10, -11
Card PlUlch (300 cpm)
Card Punch (300 cpm)
Scored Card Feature for 236-10, -11
Card Reador (1,435 opm)
Card HeRder (J.,435 cpm)

........... 0 .... · _..
Column Btnal')' Feature
t.:nd-of- FUe Feature
CCllumn III lnlUbU
Card ReMer (1100 opm)
li:nd·ot~J'Ue

J'ClIlture

A•

AUERBACH

1

for
237-10, -21

(Coutd.)

PRICE DATA

1950: 221. 103

PRICES

IDENTITY OF UNIT
CLASS

Model
Number

INPUTOUTPUT
(Contd.)

Monthly
Rental
DM

Monthly
Maint.
DM

Purchase
DM

90-Column Card Feature
Column Binary Feature
Card Punch (420 cpm)
Column Binary Feature

98
49
1,600
15

14
10
220
5

4,750
2,375
78,000
600

Card Reader (1,000 cpm)

1,435

205

70,000

Card Reader (1,000 cpm)
Column Binary Feature
Scored Card Feature
Mark Sensing Feature
90-Column Card Feature
End-of-File Feature

1,560
34
24
200
30
15

220
5
5
30
5
5

76,000
1,520
1,140
9,800
1,370
665

Paper Tape Reader (1,200 char/sec)
Paper Tape Punch (100 char/sec)

722

101

34,960

Paper Tape Punch (100 char/sec)
Paper Tape Reader (400 char/sec)
Paper Tape Reader (1,000 char/sec)

722
605
956

101
82
134

34,960
29,735
46,550

Paper Tape Reader (1,000 char/sec)

956

134

46,550

Paper Tape Reader (850 char/sec)
Paper Tape Punch (150 char/sec)

956
330

134
90

46,550
10,850

Paper Tape Punch (150 char/sec)
Edge-Punch Card Reader (120 char/sec)
Edge-Punch Card Reader (120 char/sec)
Edge-Punch Card Punch (28 char/sec)
Paper Tape Controller to use Models 4004/
4223, /4228, /4280, /4281, /4282
Paper Tape Controllers to use Models 4004/
4228, /4282
Paper Tape Controllers to use Models 4004/
4223, /4280, /4281
Paper Tape Controller to use Models 4004/
4225, /4226, /4227
Controller Attachment for second device

330
105
105
260

90
30
30
70

10,850
3,500
3,500
8,655

1,030

45

50,000

760

35

35,000

760

35

35,000

707
171

96
29

34,295
8,360

Printer (1,250 lpm)

4,840

670

236,OO(

Printer (1,250 lpm)

6,360

880

310,000

Printer (910 lpm)

4,950

690

241,000

Printer (765 lpm)

4,950

690

241,OO(

Printer (465 lpm)

4,000

555

195,OO(

Printer (390 lpm)

4,000

555

195,OO(

Feature

Name
Punched Card (Contd.)

45202
45206
4004/4238
45212
4004/
4239-10
4004/
4239-20
45207
45208
45209
45210
45211

Punched Tape
4004/4223
4004/4225
4004/
4225-8
4004/4226
4004/4227
4004/
4227-8
4004/
4227-0
4004/4228
4004/
4228-8
4004/4280
4004/4281
4004/4282
4004/4220
4004/4221
4004/4222
4004/4224
45224

Printer
4004/
243-30
4004/
243-40
4004/
4241-1B
4004/
4241-lD
4004/
4241-2B
4004/
4241-2D

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

10/69

1950:221.104

SIEMENS SYSTEMS 4004

PRICES

IDENTITY OF UNIT
CLASS

Model
Number

Feature

Monthly
Maint.
DM

Purchase
DM

~(Contd.)

INPUTOUTPUT
(Contd.)

Monthly
Hental
DM

Name

4004/
4242-1B
4004/
4242-lD
4004/
4242-2B
4004/
4242-2D
45256
45264
4004/
4245-1
4004/
4245-2
4004/
4247-1B1
4004/
4247-1F1
4004/
4247-1B2
4004/
4247-1B3
4004/
4247-1B4
4004/
4247-2B1
4004/
4247-2F1
4004/
4247-2B2
4004/
4247-2B3
4004/
4247-2B4

Printer (910 lpm)

5,460

755

266,000

Printer (765 lpm)

5,460

755

266,000

Printer (465 lpm)

4,525

625

220,000

4,525
110
290

625
15
45

220,000
5,200
14,000

Printer (285 Ipm)

2,985

415

145,000

Printer (200 Ipm)

2,985

415

145,000

Printer (620 Ipm)

3,803

528

185,250

Printer (750 Ipm)

3,803

528

185,250

Printer (620 lpm)

3,803

528

185,250

Printer (620 Ipm)

3,803

528

185,250

Printer (620 Ipm)

3,803

528

185,250

Printer (620 Ipm)

4,388

610

213,750

Printer (750 Ipm)

4,388

610

213,750

Printer (620 Ipm)

4,388

610

213,750

Printer (620 Ipm)

4,388

610

213,750

Printer (620 Ipm)

4,388

610

213,750

Videoscan Document Reader-Sorter

10,695

1,500

497,200

Vldeoscan Document Reader-Sorter

12,600

1,795

588,000

V,deosoan Pocumont Reader-Sorter

15,735

2,205

730,000

Videoscan Document Reader-Sorter

15,735

2,205

730,000

Printer (390 Ipm)
Hi-Speed Paper Advance
Train Cartridge

l

for
4004/4241, /4242

OCR - Devices
4004/4250
4004/
4250-1
4004/
4250-2
4004/
4250-3

10/69

A

.,

AUERBACH

(Contd. )

PRICE DATA

1950:221.105

PRICES

IDENTITY OF UNIT
CLASS
INPUTOUTPUT
(Contd. )

Model
Number

Monthly
Rental
DM

Monthly
Maint.
DM

Purchase
DM

Videoscan Document Reader - Sorter

16,970

2,380

789,100

Videoscan Document Reader - Sorter
Videoscan Document Reader - Sorter
Videoscan Document Reader - Sorter

19,230
6,715
5,465

2,695
940
760

894,200
312,200
255,000

1,458
455
360
65

62
70
15
5

70,918
22,000
17,500
3,000

Tape Controller (single channel)

3,725

149

181,450

Tape Controller (single channel)
382 Tape Mode Feature
for 463-108, -ll6
Pack/Unpack Feature

6,143
68
224

245
5
10

299,250
3,230
10,735

Tape Controller (dual channel)

4,934

197

240,350

Tape Controller (dual channel)
382 Tape Mode Feature
} for 463-208, -216
Pack/Unpack Feature

8,015
132
380

317
10
14

390,450
6,413
18,430

Tape Controller (single channel)

3,081

125

150,100

Tape Controller (single channel)

5,928

235

288,800

Tape Controller (dual channel)

4,290

173

209,000

Tape Controller (dual channel)

7,683

302

374,300

Tape Controller (single channel)

3,305

134

160,835

Tape Controller (single channel)
Pack-Unpack Feature

6,152
224

245
10

299,535
10,735

Tape Controller (dual channel)

4,626

187

225,245

Tape Controller (dual channel)
Pack-Unpack Feature

8,019
380

317
14

390,545
18,430

Tape Controller (single channel)

3,080

125

150,100

Tape Controller (single channel)

5,930

235

288,800

Tape Controller (dual challnel)

4,290

175

209,000

Tape Controller (dual chanllel)

7,685

305

374,300

Tape Controller (single channel)

4,485

180

219,000

Feature

Name
OCR - Devices (Contd.)

4004/
4250-4
4004/
4250-5
4004/4251
4004/4252

~
4004/97
4004/4217
4004/4215
45215

Console and Typewriter
Console and Typewriter
Control
Control ExpanSion
Magnetic Tape

4004/
463-108
4004/
463-116
5414-1
5415-1
4004/
463-208
4004/
463-216
5414-2
5415-2
4004/
472-108
4004/
472-116
4004/
472-208
4004/
472-216
4004/
473-108
4004/
473-116
5402-1
4004/
473-208
4004/
473-216
5402-2
4004/
4472-108
4004/
4472-116
4004/
4472-208
4004/
4472-216
4004/
4475-108

}

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

10/69

SIEMENS SYSTEMS 4004

1150:221. 106

PRICES

IDENTITY OF UNIT
CLASS
INPUTOUTPUT
(Contd. )

Model
Number

Feature

Name

Monthly
Rental
DM

Monthly
Maint.
DM

Purchase
DM

Magnetic Tape (Contd.)

4004/
4475-116
4004/
4475-208
4004/
4475-216
4004/
4476-108
4004/
4476-116
4004/
4476-208
4004/
4476-216
4004/
432-1
4004/
432-2
4004/
441-1
4004/
441-2
4004/
442-1
4004/
442-2
4004/
443.l-7
4004/
4432-9
4004/
4442-7
4004/
4442-9
4004/
4443-1
4004/
4443-2
4004/
4446-1
4004/
4446-2
4004/
4451
4004/
4453

Tape Controller (single channel)

6,770

2'70

330,000

Tape Controller (dual channel)

6,065

240

295,000

Tape Controller (dual channel)

10,300

410

502,000

Tape Controller (single channel)

3,630

145

177,000

Tape Controller (single cbannel)

6,240

250

304,000

Tape Controller (dual channel)

4,720

190

230,000

Tape Controller (dual channel)

8,210

330

400,000

Magnetic Tape Unit

2,637

365

120,650

Magnetic Tape Unit

2,637

365

120,650

Magnetic Tape Unit

3,101

432

151,050

Magnetic Tape Unit

3,101

432

151,050

Magnetic Tape Unit

3,949

547

180,500

Magnetic Tape Unit

3,949

547

180,500

Magnetic Tape Unit

1,320

170

60,325

Magnetic Tape Unit

1,320

170

60,325

Magnetic Tape Unit

1,975

275

90,250

Magnetic Tape Unit

1,975

275

90,250

Magnetic Tape Unit

1,974

274

90,250

Magnetic Tape Unit

1,974

274

['0,250

Magnetic Tape Unit

2,210

310

101,000

Magnetic Tape Unit

2,:nO

310

101,000

Magnetic Tape Unit

1,555

220

71,000

Magnetic Tape Unit

2,210

310

101,000

Data Exchange Control
Communication Control (fling1\'! channel)

1,970
1,820

82
75

95,950
88,500

Communication Control (multi-channel)

3,081

125

150,100

Communications

4004/
627-10
4004/656
4004/
668-11

10/69

fA

AUERBACH

'"

(eontd.)

PRICE DATA

1950:221. 107

PRICES

IDENTITY OF UNIT
CLASS
INPUTOUTPUT
(Cootd. )

SWITCHING
DEVICES

Model
Number

Feature

Monthly
Rental
DM

Name

Monthly
Maint.
DM

Purchase
DM

Communications (Contd.)
4004/
668-21
4004/
668-31
4004/710
4004/
720-21
4004/
720-22
4004/
721
4004/780
4004/4666
4004/
4713-1
4004/
4713-2
4004/752
4004/
310-21
4004/
310-22
4004/
310-23
4004/
310-24
4004/
310-25
4004/
310-26
4004/
310-27
4004/
310-28
4004/
K 310
4004/
350-2
,*004/
350-3
4004/
350-4

Communication Control (multi-channel)

3,959

158

192,850

Communication Control (multi-channel)
Telegraph Buffer

4,836
122

192
5

235,600
5,795

185

10

8,978

185

10

8,978

185

10

8,978

1,365

58

66,500

141

10

6,840

141
741

10
29

6,840
36,100

Standard Interface Switch

390

19

19,000

Standard Interface Switch

780

34

38,000

Standard Interface Switch

1,170

48

57,000

Standard Interface Switch

1,560

62

76,000

Standard Interface Switch

1,950

77

95,000

Standard Interface Switch

2,340

96

114,000

Standard Interface Switch

2,730

110

133,000

Standard Interface Switch

3,120

125

152,0.00

180

10

8,598

Switoh Controller

2,087

86

101,650

Switch Controller

2,535

101

123,500

Switch Controller

2,965

120

144,400

ADS Buffer
for 668-11, -21, -31

ADS Buffer

SDS Buffer
Time Generator/Buffer
Communication Control (multi-channel)
Telegraph Buffer
Telegraph Buffer
Video Data Terminal

I

for 4666

Switch Expansion Kit

© 1969 AUERBACH CorporatIon and AUERBACH Info. Inc.

10/69

fA
AUERBACH

AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

1955:011. 100
SIEMENS SYSTEM 300
SUMMARY REPORT

•

SUMMARY REPORT: SIEMENS SYSTEM 300
•1

BACKGROUND
Siemens AG of Munich, West Germany announced the System 300 in the third quarter of
1965. The System 300 originally consisted of four models, namely, 302, 303, 304 and 305.
More recently two additional models, 301 and 306, have been introduced to expand the range
of the System 300's processing capability. Memory capacities for this series of binary
processors range from 4, 096 24-bit words on the smallest 301 to 65, 536 words on the largest 306.
Models 301 and 306 represent Siemens' attempt to compete with the imports in those two
ranges (small and large scale systems) from Western Europe and the U. S. A. The Siemens
System 300, as originally introduced, was a second-generation range of computers; these
two new processors extend the range to compete with third-generation equipment.
The purchase price for the System 300 ranges from 60, 000 DM (about US $15, 000) to
165,000 DM for the Model 301 and from 410, 000 DM (about $103, 000) to 1,090, 000 DM for
the Model 306.
Siemens, a manufacturer of communications and electronics equipment, should not be dismissed lightly in its venture into the computer market. This is especially true since it
more recently announced the 4004 Series, which is discussed in Summary Report 1950 .

.2

HARDWARE

. 21

System Configuration
A Siemens System 300 includes a central processor that contains core memory, an arithmetic unit and a control unit. The peripherals communicate with the central processor
through channels. The System 300 is equipped with two types of channels; standard speed
channels for low-speed devices such as teletypes, and high-speed channels for fast devices
such as discs and magnetic core storage units.
A wide range of peripheral units is optionally provided for connection to the central processors. The available peripheral units are compatible on all models of the System 300 .

.........1IIIIIIII
11l1li0111111

Figure 1. Siemens System 300 Computer
© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

10/69

SIEMENS SYSTEM 300

1955:011.220

. 22

Data Structure
The basic addressable unit of the System 300 is the 6-bit binary character. Four such
characters constitute the 24-bit word. The 24 binary bits can represent four characters
or four decimal digits. Models 305 and 306 represent a floating point number as a fraction
having 24 or 34 bits and an exponent of 10 bits.
Instructions are one 24-bit word in length, with a 14-bit address, a I-bit accumulator designator (right or left), a flag bit to interrogate the console interrupt flag, a bit to test for address modification, an interrupt bit and a 6-bit operation code. The interrupt bit can be set
by the following conditions: an illegal operation has occurred in the instruction register, an
interrupt has been issued by a peripheral unit or the operator has depressed the interrupt
button on the console.
TABLE I: SIEMENS SYSTEM 300 CENTRAL PROCESSOR CHARACTERISTICS
Model No.

302

301

303

304

305

306

Minimum Storage
Capacity (words)

4,096

8,192

4,096

8,192

8,192

16, 384

Maximum Storage
Capacity (words)

16,384

16,384

16,384

16,384

16,384

65,536

1.5

33*

1.5

1.5

0.6

None

None

Parity
Write
only

Main Memory
Cycle Time, J,!sec

1.6

Checking

None

None

None

Storage Protection

None

None

Write
only

Write
only

Write
only

Program Interrupt
(levels)

2

Multiple

Multiple

Multiple

Multiple

Multiple

None

None

None

None

None

16

Number of Index
Registers

* The System 300 Model 303, an early member of the range, was substantially slower than
the other models. It was developed from the earlier Siemens 3003 .
. 23

Central Processor
As presently implemented, the System 300 processors appear as fixed word length, binary
processors with one address per instruction. Models 301 and 302 have only one accumulator while the remaining processors each have two. There is only a limited downward
compatibility of index registers because of the different instruction sets. Model 306 comes
equipped with 16 index registers. Index registers are not simulated through software on the
models without index registers. Parity checking is provided only on Model 306.
The basic instruction set is common to all models of the System 300. In successive models the basic instruction list is made more powerful through the use of address modification subsets to take advantage of the more powerful system. The instruction repertoire
of all models of System 300 include a full complement of shift instructions, logical and
transfer instructions, branching and arithmetic instructions and peripheral control instructions. Model 301 has 22 instructions, Model 302 has 23, Model 303 has 31, Model 304 has
40, Model 305 has 46 instructions and Jl1:odel 306 has a full complement of 55 instructions.
Models 301 through 304 provide no floating point facilities. Model 303 allows only fixed
point addition, subtraction and multiplication and Models 301 and 302 permit fixed point addition and subtraction only.
Multilevel indirect addressing is available on all models of the System 300. Siemens refers
to this as Address Substitution. Each instruction has an interrupt bit, bit 18, that is set by
the programmer to enable the interrupt system. Interrupts can also be enabled from the console and from peripheral units that request servicing.

10/69

A

(Contd. )

AUERBACH
®

SUMMARY REPORT

.23

1955:011.230

Central Processor (Contd. )
All models of the Siemens System 300 are capable of processing up to 23 independent programs on a multiprogramming basis. The individual programs are processed on 23 priority levels (program numbers) to ensure that the program with the highest priority is
processed first, if necessary by interruption of a less important program.
The contents of essential registers in the processors are displayed through a number of optical indicators, and toggle and pushbutton switches that allow access to any location in main
memory for manual input-output transfers.
The input-output channels are provided with standard interfaces for the connection of peripheral units. A distinction is made between standard interfaces which are provided in all
processors of the System 3000 and high-speed channel interfaces which are optional and
available on Models 304, 305, and 306 only.
The Siemens System 300 is fully upward compatible for programs written on the machine
language level; this means that all programs written for a specific processor model of the
system can be run without modifications on the larger models of the system.
A means for providing downward compatibility is offered on the assembly language level.
On this level, all instructions not included in the instruction complements of the individual
central processors are coded as macro instructions so that, when non-implemented instructions are found at assembly time, they are replaced by a subroutine which simulates the
illegal instructions .

. 24

Auxiliary storage
Auxiliary storage for the System 300 is provided by the Model 2027 Magnetic Core Storage
Unit, and by the drum and disc units whose charaeteristics are given in Table II. A storage
controller is available for the connection of drum units - up to five drum controllers can
be connected to Models 302 and 303, and up to tcn can be connected to Models 304 and 305.
From one to four drum units can be connected to a drum controller, but different drums
connected to the same controller cannot be usce! simultaneously.
The Disc Storage Unit contains an exchangeable disc stack with six magnetic discs. Disc
storage units are connected to the processor thnlUgh a disc controller to the high-speed
channel.
The Model 2027 Magnetic Core Storage Unit is connected to the processor via a high-speed
channel and provides secondary storage of 16, :HH words.
TABLE II: CHARACTERISTICS OF SIEMENS SYSTEM 300
AUXILIARY STORAta: DEVICES
201G Drum

2027 Magnetic
Core Storage
Unit

2051 Disc

4/Trunk

4/Trunk

l/Trunk

l/Trunk

65,536

131, 072

262, 144

16,384

1,792,000

0
32
64

0
32
64

0
32
64

-

0
87.5
16.0

Effective Transfer
Rate, char/sec

72,000

72, 000

72,000

2,668,000

208,000

Data Checking

Cyclic

Cyclic

Cyclic

Parity

Cyclic

Type of Unit

2013 Drum

2014 Drum

Maximum Number of
Units On-Line

4/Trunk

Maximum Number of
Words/Unit

~

~

Waiting Time (J.Isec)
Minimum
Average (Random)
Maximum

© 1969 AUERBACH Corporation and AUERBACH Info. Inc .

•

10/69

1955:011.250

.25

SIEMENS SYSTEM 300

Peripheral Equipment
The peripheral units for System 300 are listed in Table III.
The Siemens System 300 provides three different types of peripheral units capable of processing punched cards.
Model 2009 is a column by column reader/punch; the cards can be printed under program
control during punching.
Models 2010 and 2021 are a reader and a punch, respectively and feature automatic code
translation and a column binary feature on reading and punching.
Presently two printers are available as shown in Table III. Model 2023 is available at option with 120 and 104 print positions per line; the Model 2022 has 120 print positions per
line. Forty-eight different characters can be printed.
The Model 2022 Printer can print up to 1500 lines per minute if only numeric data is
printed; the speed for Model 2023 is 1600 lines per minute for numeric print.
Five models of paper tape input-output equipment are available. Model 0001, a console paper tape reader, is used only with processor Model 303, while the Model 0016 Console Paper
Tape Reader is available for Processor Models 302, 304, 305, 306.
The three remaining paper tape input-output units (Models 2006, 2007 and 2008) can be connected to any central processor in the System 300.
All models feature automatic code translation; Models 2006, 2007 and 2008 have parity
checking while Modcls 0001 and 001 (j contain a second read station for data checking.
TABLE lIT: PERIPHEHAL EQUIPMENT FOR SIEMENS SYSTEM 300

.26

Type of Unit

Model Numher

Card Reader/Punch

2009

53 cards/min

Card Reader

2010

660 cards/min

Validity

Card Punch

2021

420 cards/min

Echo

Paper Tape Reader

0001

30 char/sec

Second Read Station

Paper Tape Reader

001 (j

200 char/sec

Second Read Station

Paper Tape Reader

200(,

400 char/sec

Parity

Paper Tape Reader

2008

400 char/sec

Parity

Paper Tape Punch

2007

150 char/sec

Parity

Printer

2022

1500 lines/min

Echo, Timing

Printer

2023

1600 lines/min

Timing

Console Typewriter

2017

10 char/sec

Rated Speed

Data Checking

-

-

Process Signal Input-Output Units
The System 300 provides a range of analog/digital and control devices. A special feature of
the Siemens System 300 is the wide variety of devices for the automatic interchange of data
between the processor and the process to be controlled. With the System 300, two types of
process signal input-output units are provided, both of which connect to the processor via
a standard interface.
The Model P1K Process Signal Input-Output Unit handles a maximum of 12 process signal
input-output devices that are located in the immediate vicinity of the processor.

10/69

A

(Contd. )

AUERBACH
Co)

SUMMARY REPORT

.26

1955:011.260

Process Signal Input-Output Units (Contd.)
The Model P2K Process Signal Input-Output Unit handles process signal input-output devices as remote as 12. 5 miles from the central processor and the maximum information
transfer rate is 2400 bauds.
These process signal input-output units handle analog variables. binary signals, digital inputs, timing pulses, and alarm signals.
The Models P3 and P4E Process Signal Input-Output Units handle the input-output of digital
data at a speed of 220,000 words per second for P3 and 30,000 words per second for the
P4E in the immediate vicinity of the processor and remotely up to 3/4 mile .

.3

DAT A COMMUNICATIONS
The System 300 is designed to handle data communications. To accomplish this, additional
equipment such as data transmitters and terminal equipment has been developed.
The Model 2011 Telegraph Unit is one of the data communications units provided with the
Siemens System 300. This unit provides for connection of teleprinters and paper tape equipment to any Siemens System 300 processor through private or public lines. It can also be
used to connect System 300 processors with each other via telegraph lines. The maximum
data transmission rate is 200 bits per second. Over short distances it is possible to transmit characters in parallel at a maximum rate of 500 characters per second.
Siemens has under development, equipment for higher-speed data transmission at a maximum rate of 1200 bits per second. This equipment will make it possible to connect Siemens
System 300 processors and Siemens System 4004 processors via telephone channels.
The Data Exchange Control 2041 enables two adjacent Siemens System 300 processors to
communicate with each other at a rated speed of :;4. noo characters per second. The Data
Exchange Control 2071 transfers data between one Siemens System 4004 processor and
another Siemens System 300 processor; the transmission rate depends upon the type of
channel used and the connected processors .

.4

SOFTWARE
Control Systems consisting of a supervisor (the Control System) and other system programs
are used to extend the hardware capabilities of the Siemens System 300 processors. These
systems control the overlapped operations of hardware elements and of programs. The required Control System functions are initiated by calls in the user programs which consist of
an instruction sequence that is written as a macro instruction in the 300 assembler language.
The Control System of the System 300 provides for multiprogramming of up to 23 independent programs and one scheduling routine. The Control System allocates a software program counter to each of these programs which usc the hardware program counter in alternation. The program with the highest priority is always jJrocessed first, it necessary, after
the interruption of a lower priority program. When programs require input-output operations, they are held up until higher priority operations are completed.
Program priorities depend on the program numbers; program No. 1 has the highest priority.
It may be interrupted only if it must wait for completion of input-output operations or if Con-

trol System operations are required. The Control System also controls the shared use of
core memory areas (in a time-sharing environment) and of subprograms by concurrently
operating programs and automatically assigns, and prepares a directory of, storage areas
on external storage devices.
To prevent the blocking of program processing, the Control System maintains queue lists
for the peripheral units by storing the necessary information from the program requesting
the input-output operation and continuing the program.
Programs for the Siemens System 300 are generally written in the PROSA 300 Assembly
Language. ALGOL 300 is a multiple-pass compiler and corresponds to the IFIP Subset
of ALGOL 60.
FORTRAN IV-300 is a multiple-pass compiler and corresponds to the IBM 360 FORTRAN
E level.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

10/69

A

1955;221.100

AUERBACH
COMPUTER

NOTEBOOK

SIEMENS SYSTEM 300
PRICE DATA

INTERNATIONAL

AUERBACH
8

PRICE DATA

IDENTITY OF UNIT
CLASS

CENTHAL

Model
Number

Feature

Name
Central Processor (4,096 words)
301 Memory Expansion Options

301-4

PHOCESSOH
4-8
8-12
12-16

from 4,096 to 8, 192 words
from 8, 192 to 12,288 words
from 12, 288 to 16, 834 words

Monthly

Purchase
~~tal
DM
1,300
155

Monthly

~:t.
60,000

760
760
760

90
90
90

35,000
35,000
35,000

302-8
302-16

Central Processor (8,192 words)
Central Processor (16,384 words)

3,045
4,425

355
518

140,000
204,250

303-4
303-8
303-12
303-16

Central
Central
Central
Central

2,575
3,650
4,575
5,750

302
427
538
672

118,750
768,340
211,090
265,430

304-8
304-8
SK21
304-8
SK25
304-16
304-16
SK21
304-16
SK25

Central Processor (8,192 words)
Central Processor (8, 192 words) with one
high speed channel for 1 external unit
Central Processor (8, 192 words) with one
high speed channel for 5 external units
Central Processor (16, 384 words)
Central Processor (16, 384 words) with one
high speed channel for 1 external unit
Central Processor (16, 384 words) with one
high speed channel for 5 external units

4,650
4,835

542
566

214,700
223,250

5,555

648

256,500

6,030
6,215

706
725

278,350
286,900

6,935

811

320, 150

305-8
305-8
SK21
305-8
SK2-5
305-16
305-16
SK2l
305-16
SK25

Central Processor (8,192 words)
Central Processor (8,192 words) with one
high speed channel for 1 external unit
Central Processor (8,192 words) with one
high speed channel for 5 external units
Central Processor (16, 384 words)
Central Processor (16, 384 words) with one
high speed channel for 1 external unit
Central Processor (16, 384 words) with one
high speed channel for 5 external units

5,370
5,555

629
648

247,950
256,500

6,275

734

289,750

6,750
6,935

787
811

311,600
320,150

7,655

893

353,400

306

Central Processor (16,384 words)
306 Memor;y: EXEansion Options
from 16, 384 to 32, 768 words
from 32, 768 to 49, 152 words
from 49,152 to 65,536 words
Features
High speed channel for one external unit
High speed channel for five external units

8,860

1,040

410,000

4,750
5,185
4,750

560
610
560

220,000
240,000
220,000

185
905

19
106

8,550
41.800

Magnetic Core Storage Unit (16, 384 words)
Drum
Drum Storage (65, 536 words)
Drum Storage (131, 072 words)
Drum Storage (262, 144 words)
Controller
Drum Storage Controller with one Drum
Storage 0011
Drum Storage Controller with one Drum
Storage 0012

3,395

398

156, 750

1,654
2,265
2,880

192
264
336

76,000
104,500
133,000

3,085

360

142,500

432

171,000

16-32
32-49
49-65
SK21
SK25
MASS
STOHAGE
UNITS

PRICES

2027
0011
0012
0013
2013-1/-2
2014-1/-2

Processor
Processor
Processor
Processor

(4, 096 words)
(8,192 words)
(12, 288 words)
(16, 384 words)

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

3,705

.

10/69

SIEMENS SYSTEM 300

1955:221.101

PRICES

IDENTITY OF UNIT
CLASS

MASS
STOHA.{;E
II Nl'rS

Model
Number

Feature

Name

Monthly
Monthly
Purchase
Maint.
Rental
DM
DM

DM

Controller
2015-1/-2

(Contd. )

0010
0042

Drum Storage Controller with one Drum
Storage 0013
Drum Cabinet (for Drum Storage Expansion)
Drum Cabinet (for Drum Storage Expansion)

4,320

504

199,500

825
825

96
96

38,000
38,000

2,400
55

218

-

105,634
1,800

2,210

238

91,586

65

10

2,850

620

72

28,500

110
310
610
725
725
750
750
1,045

14
38
72
86
86
91
91
125

4,845
14,250
28,025
33,440
33,440
34,440
34,440
48,165

1,045

125

48,165

1,070

130

49, 1 G5

1,070

130

4 ~J, 165

Card Header/Punch (13 ••• 53 cpm)
Card Header (600 cpm)
Card Punch (104 •.. 420 cpm)
Column Binary (read)
Column Binary (punch)
Printer

885
1,360
1,545
55
15

120
158
182
10
5

40,660
G2,700
71,250
2,375
GOO

Printer (1500 lpm)
Printer (1600 lpm)
Printer (1600 lpm)
Printer (1600 lpm)
Expansion from 104 to 120 print positions
Expansion from 120 to 136 print positions
Data Exchange Controller

3,715
3,520
3,335
2,340
105
105

437
412
388
278
14

14

171,475
162,450
153,900
107,825
4,750
4,750

Data Exchange Controller
Data Exchange Controller

980
1,245

115
160

45,000
57,000

Disc
2051
0017

Disc Storage (1. 792 Million words)
Disc Pack
Controller

INPUT OUTPUT
DEVICES

2050-1/-2

Disc Storage Controller

0002

Channel Selector (Expansion only for
Model 303)
Interface Attachment
Punched Tape

2092

Paper Tape Reader Controller (30 char/sec)
Paper Tape Reader (200 char/sec)
Punched Tape Controller (400 char/sec)
Paper Tape Punch (100 char/sec)
Paper Tape Punch (100 char/sec)
Paper Tape Punch (150 char/sec)
Paper Tape Punch (150 char/sec)
Paper Tape Reader/Punch (400/100
char/sec)
Paper Tape Header/Punch (400/100
char/sec)
Paper Tape Header/Punch (400/150
char/sec)
Paper Tape Header/Punch (400/150
char/sec)
Punched Card

0001
0016
2006
2007 -1
2007 -2
2007-3
2007 -4
2008 -1
2008-2
200S -3
2illJ~-"

20(lf)
~OlO

2021

0026
0027

---

202Z-A
202:)-1
2023-2
2023-3
0018
0019
20H
2071

10/69

A

AUERBACH
®



Source Exif Data:
File Type                       : PDF
File Type Extension             : pdf
MIME Type                       : application/pdf
PDF Version                     : 1.3
Linearized                      : No
XMP Toolkit                     : Adobe XMP Core 4.2.1-c043 52.372728, 2009/01/18-15:56:37
Create Date                     : 2013:10:23 21:06:13-08:00
Modify Date                     : 2013:10:24 05:53:44-07:00
Metadata Date                   : 2013:10:24 05:53:44-07:00
Producer                        : Adobe Acrobat 9.55 Paper Capture Plug-in
Format                          : application/pdf
Document ID                     : uuid:47d95963-932b-4897-a9d1-b744b5b4ff79
Instance ID                     : uuid:3eda9760-d0ae-4b90-9c95-e861736a2018
Page Layout                     : SinglePage
Page Mode                       : UseNone
Page Count                      : 1280
EXIF Metadata provided by EXIF.tools

Navigation menu