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 .
Page Count: 1280
Download | |
Open PDF In Browser | View 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 Sk 15, 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~ mo 80 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 - Mo rovided 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 AIIU 1):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 : 1280EXIF Metadata provided by EXIF.tools