Auerbach_Standard_EDP_Reports_196609_Volume_5_IBM Auerbach Standard EDP Reports 196609 Volume 5 IBM
User Manual: Pdf Auerbach_Standard_EDP_Reports_196609_Volume_5_IBM
Open the PDF directly: View PDF 
.
Page Count: 978
| Download | |
| Open PDF In Browser | View PDF |
STANDARD EDP "REPORTS
. An Information Service for the Electronic Data Processln, Field.·.. .
Prepared & Edited by
AUERBACH CORNlAnDN
"11III.1P1i. PH•.
Printed and Distributed by
INA IlCo..,.md
WASHINGTON, D. C.
5
I AUERBACH I $J
STANDARD
BDP
REPORTS
.. RINTED IN THE UNITED STATES 0,.. AMERICA
The information contained herein has been obtained from reliable sources
and has been evaluated by technical teams with extensive working experience
in computer design, selection and application. The information, however, is
not guaranteed.
Acknowledgement is made oHhe inspiration and guidance provided by the
Information Systems Branch of Office of Naval Research which has supported
data gathering activity by Auerbach Corporation in fields similar to some
covered in these reports. The data contained and formats used in STANDARD
EDP REPORTS were not prepared under any contract with the U. S. Government; and they are the exclusive property of the copyright holders.
,/
18.M 1620
Model· 1
International Business Machines Corp.
/
.'-.--.
AUERBACH INFO, INC.
PRINTED IN U. S. A.
IBM 1620
Model· 1
International Business Machines Corp.
AUERBACH INFO, INC.
PRINTED IN U. S. A.
412:001.001
STANDARD
REPORTS
IBM 1620 Modell
Contents
CONTENTS
1.
2.
3.
4.
5.
6.
7.
8.
10.
11.
12.
13.
14.
15.
Introduction
•
.
Data Structure.
.
System Configuration
Configuration IX; Desk Size Scientific
Configuration X; Punched Card Scientific
Internal Storage
Core Storage .
1623
Core Storage . • . .
2301
Core Storage Adapter
2302
Core Storage Adapter
Auxiliary Storage
1311 Model 3
Disk Storage Drive
1311 Model 2
Disk Storage Drive
Central Processor • •
••
Console..
•..••.
•
Input-Output; Punched Tape and Card
1621
Paper Tape Reader
5514
Paper Tape Reader Adapter
1624
Tape Punch •
•.
•
5514
Adapter.. . . .
.
1622
Card Read Punch (Reader)
1632
Card Read Punch Adapter
1622
Card Read Punch (Punch)
1632
Card Read Punch Adapter
Input-Output; Printers
I/O Console Typewriter •
Input-Output; Other
Calcomp 560-R Digital Recorder
Simultaneous Operations
Instruction List
Coding Specimens
SPS
FORTRAN I
FORTRAN II
GOTRAN
Data Codes
Internal Numeric
Internal Alphameric •
I/O Console Typewriter (numeric)
I/O Console Typewriter (alphameric)
Paper Tape Input-Output . •
Card Input-Output (numeric)
Card Input-Output (alphameric) .
Problem Oriented Facilities
IBM 650 Simulator Program
1710 Simulator/7090
1620 5-Channel Tape Translation Program .
Floating point function subroutines
©
1962 by Auerbach Corporation and BNA Incorporated
412:011
412:021
412:031
412:031
412:041
412:041
412:041. 4
412:041. 4
412:042
412:042
412:051
412:061
412:071
412:071. 4
412:072
412:072.4
412:073
412:073.4
412:074
412:074.4
412:081
412:101
412:111
412:121
412:131
412:132
412:133
412:134
412:141
412:142
412:143
412:144
412:145
412:146
412:147
412:151. 11
412:151.12
412:151. 15
412:151.171
12/62
412:001.002
IBM 1620 MODEL 1
CONTENTS-Contd.
15.
16.
17.
18.
19.
20.
21.
22.
12/62
Problem Oriented Facilities (Contd.)
Matrix Inversion (Tape)
Complex FORTRAN (Tape) ••
SPS to FORTRAN conversion.
Format Control Subroutines for
1620 Card FORTRAN
Interpretive systems
Other ••
AUTOMAP •
Process Oriented Languages
FORTRAN I
FORTRAN II •
GOTRAN
FORGO •
Machine Oriented Languages
SPS
SPS I-Pass ••
Program Translators
SPS • • • .
SPS I-Pass.
FORTRAN I
FORTRAN II
Operating Environment
General .
GOTRAN •
FORGO •
System Performance
Notes on System Performance
Worksheet Data • • • . . •
Matrix Inversion • • • • • •
Generalized Mathematical Processing.
Generalized Statistical Processing
Physi~l Characteristics
Price Data. • • . . • • • • . • • • • • • • • . .
412:151.172
412:151.173
412:151.174
412:151.175
412:151.176
412:151.177
412:151.3
412:161
412:162
412:163
412:164
412:17l
412:172
412:181
412:182
412:183
412:184
412:191
412:192
412:193
412:201. 001
412:201. 011
412:201.3
412:201.4
412:201. 5
412:211
412:221
412:011.100
STANDARD
EDP
IBM 1620 Modell
Introduction
INTRODUCTION
§ OIl.
The IBM 1620 Modell is a solid-state desk size computer oriented toward scientific
applications. The basic system consists of the 1620 Modell Central Processing Unit and
Console, and the Input/Output Console Typewriter, used for input with hard copy and for output. Typewriter output occurs at about ten characters per second. This minimum configuration, including core storage of 20,000 decimal digits, rents for $1,375 per month.
The processor performs the two-address instructions sequentially. Data processing
is performed serially by digit on variable length decimal fields; no input-output radix conversion is required. Alphameric data may be input and output; each alphameric character is
stored internally as a pair of decimal digits. Instructions are fixed in length at twelve digits.
A digit consists of four numeric bits, one check bit, and one flag bit used for storing the sign
of a numeric field and for delimiting a field. The core store has a 20 microsecond readrestore cycle. Fixed-point addition of two fields is performed at 80 microseconds per digit,
and field movement requires 40 microseconds per digit. Each digit in storage is individually
addressable. Core storage is expandable to a total size of 40,000 or 60,000 digits.
Punched tape and card equipment can be used with the 1620. Paper tape may be input
at 150 rows per second or output at 15 rows per second; no buffering is available. Two
independent buffered card channels are available, permitting reading at 250 cards per minute
and punching at 125 cards per minute. A few 1620 installations have installed Model 7330
Magnetic Tape Units and Model 1403 Line Printers on an RPQ basis. Other devices which
require an RPQ are the Model 1940 Printer which has a speed of 50 characters per second,
and Model 1402 Card Read Punch. Model 1402 can read cards at 800 cards per minute and
can punch cards at 250 cards per minute, and is the standard read-punch unit for the 1401
Data Processing System.
A new disc storage system has been announced for the IBM 1620, 1401, 1440, and
1710 systems. This system is the 1311 Disk Storage Drive, and features interchangeable
Disk Pack units as a replaceable storage medium. The peak _~'ansfer rate is 50,000 digits
per second when used in the 1620 system.
Each 1311 Disk Storage Drive holds one Disk Pack at a time, providing on-line
storage for 2,000,000 digits per drive in addressable sectors of 100 digits each. A maximum of four drives can be connected. Up to 20,000 digits can be read or recorded without
movement of the 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. Disk Storage seek operations (but not read -write operations) can be overlapped with other system functions. IBM has announced programs utilizing the disc file for
SPS and FORTRAN II, and also a Disk Utility Program to aid in the maintenance of programs
or data in disc storage.
Each Disk Pack consists of a stack of six discs with ten magnetic recording surfaces
and a cover that forms a sealed container when the Disk Pack is not mounted on a drive.
Diameter is 14 inches, height is 4 inches, weight is less than 10 pounds, and time to interchange two Disk Packs is about one minute.
The basic 1620 includes 33 instructions, many of which allow the use of literals (the
"Immediate" instructions, as Add Immediate). Fixed point multiply is provided, but divide
is done either by subroutine or by an extra-cost Special Feature. Floating point operations
al'e performed by subroutines, or by extra cost floating point instructions. Other processor
instructions are available at additional cost, including indirect addressing. The input-output
operations transfer groups of characters rather than a single character or word, and no radix
conversion is needed since data is stored in decimal form. Punched tape operations are term-
©
1962 by Auerbach Corporation and BNA Incorporated
IBM 1620 MODEL 1
412:011.101
§Oll
INTRODUCTION-Contd.
inated by a delimiter code, and card operations are stopped after 80 characters are transferred. Card operations are checked by comparing the buffer contents to information read
at a checking station.
The assembly language for the 1620 is the 1620/1710 Symbolic Programming System,
which includes macros for floating point subroutines for arithmetic and mathematical functions.
The translation may be performed on the 1620 or on the 709/7090 systems.
Problem oriented facilities are oriented towards industrial design applications,
mathematical applications, and linear programming. They include AUTOMAP, a program for
machine tool control.
FORTRAN I and II are available for mathematical program writing. The languages
have some restrictions relative to 709/7090 FORTRAN II; see sections :161 and :162. With
minor changes, 1620 FORTRAN source programs can be compiled and·executed on the
IBM 7070, 704, and 709/7090 systems.
IBM provides the GOTRAN interpretive system based on restricted FORTRAN
language statements. The number of arithmetic operations allowable in single statement is
one. The FORGO interpretive system, developed at the University of Wisconsin Engineering
Computing Laboratory, is more useful than GOTRAN.' See sections :163 and :164 and the
associated Operating Environment sections for descriptions of these two systems.
IBM has announced the 1620 Model 2 system, available by the end of 1963. This
system offers faster processing and has a modified IBM Selectric typewriter, thus providing
on-line printing at approximately 15 characters per second; see Computer System Report 413.
The 1620 Processor is used as the digital computer in the IBM 1710 Control System,
used in process control and data collection applications. 1620 Model 2 is not available for
the 1710 system.
12/62
412:021.100
.STANOARDEDP
•
REPORTS
IBM 1620
Data Structure
DATA STRUCTURE
§
021.
.1
•.2
Type of Information
STORAGE LOCATIONS
Name of Location
Size
Purpose or Use
Digit position:
4 bits plus flag
bit plus odd parity
bit
Row:
6 bits plus odd
parity bit
single column
code
20 digits plus
address
100' sectors
100 bands
10 bands
1 decimal digit;
basic addressable
location in core
storage.
cbaracter on
punched tape.
character on
punched cards.
smallest addressable
location on disc.
one band on disc.
one surface of disc.
the 10 bands addressa ble on Disk Storage
Drive with no head
movement required.
Column:
Sector:
Band:
Surface:
Cylinder:
INFORMATION FORMATS
©
Representation
Numeral: • . . . . .
1 digit (num,eric mode
storage).
2 digits (alphameric mode
storage).
1 row (tape or card).
Alphameric character: 2 digits (storage).
1 row (tape).
1 column (card).
Instruction:
12 digits.
Number:
group of numerals delimited by flag bit in storage.
Field: .
data between starting location and flag bit in
storage.
Record: . . . . . . . . data between starting location and record mark in
storage.
PuncheL tape record:
variable length group of
characters.
Card: • . . . . . . . .
80 characters on card.
80 digits in numeric mode
in storage.
160 digits in alphameric
mode in storage.
1962 by Auerbach Corporation and BNA Incorporated
12/62
./
412:031.100
•
STANDARD
EDP
•
IBM 1620 Modell
REPORTS
System Configuration
SYSTEM CONFI GURATI ON
§
031.
.1
DESK SIZE SCIENTIFIC SYSTEM; CONFIGURATION IX
Deviations from Standard Configuration: • • • • • .•
Rental: •
core storage larger by 8,000 decimal digits.
paper tape input faster by 140 rows/sec.
paper tape output faster by 5 rows/sec.
$2,455 per month.
Additional Core Storage, Model 1623-1
(20,000 digits).
Core Storage.
Processor', Console, and Input-Output
Typewriter.
Punched Tape Reader and Controller.
Tape Punch and Controller.
Optional Features Included:
Automatic Divide Feature.
Core Storage Adapter 2301.
Paper Tape Reader Adapter.
©
1962 by Auerbach Corporation and BNA Incorporated
12/62
412:031.200
§
IBM 1620 MODEL 1
031.
.2
PUNCHED TAPE/CARD SCIENTIFIC SYSTEM; CONFIGURATION X
Deviations from Standard Configuration:
no index register.
Rental:
$3,580 per month.
Additional Core Storage,
Model 1623-2 (40,000 digits).
Core Storage.
Processor, Console and Input-Output
Typewriter.
Card Read Punch and Controller.
Optional Features Included: . . . . • • • • • • . • . Indirect Addressing Feature.
Automatic Divide Feature.
Floating Point Feature.
Card Read Punch Adapter.
Core Storage Adapter 2301.
Core Storage Adapter 2302.
12/62
412:041.100
.STAliOAAO
EDP
•
IBM 1620 Modell
REPORTS
Internal Storage
Core Storage
INTERNAL STORAGE: CORE STORAGE
§
041.
.16
.1
GENERAL
. 11
Identity: . . . . . . . . basic core storage.
part of 1620 Central Processor, ModelL
Purpose
Basic Use: .
.13
Description
Card Load area:
working storage.
Basic core storage is part of the Central Processor
and has 20,000 storage locations. Each location
stores one decimal digit and is individually addressable by a five-decimal-digit address code. In the
computer numeric mode, one location stores one
decimal digit; and in the alphameric mode, two
digits are used to represent either an alphabetic
character, special character, or decimal digit. One
digit consists of four BCD numeric bits, one odd
parity check bit, and one flag bit. Storage cycle
time is 20 microseconds.
One access to storage retrieves two digits, but only
the addressed one is used. Core storage uses
"wrap-around" addressing: address 00000 follows
the highest-numbered address when incrementing
addresses; the highest-numbered address (19999 for
example) follows 00000 when decrementing addresses. The core store can be increased from 20, 000
to 40, 000 or 60, 000 positions by a separate storage
unit, Model 1623-1 or 2. A 1623 Modell provides
for the additional 20, 000 locations, and a 1623 Model 2 contains 40,000 additional storage locations.
Core storage is used for all input-output areas, instructions, and working storage. Instructions require 12 digits of storage. Basic storage includes a
300-digit reserved area for arithmetic tables. No
lock is provided to protect this area; it is used for
storing the loader routines when a program is being
loaded. Power may be turned on and off without loss
of information in storage when following normal operating procedures.
Operands (fields and records) may be any length desired. Numeric fields are delimited by a flag bit
in the most significant digit position, while complete records are delimited by a record mark code.
.14
Availability: . .
3 to 4 months.
.15
First Delivery:
basic storage with processor -- October, 1960.
1623 additional storage -June, 1961.
©
Number of Locks
locations
Add table:
100
Multiply table:
200
Product or dividend working area: 20**
additional core storage.
1623 Models 1 and 2.
• 12
Reserved Storage
80
no. *
no.*
no; may also be used
as working storage.
no; used with card
reader Load key;
may be used as
working storage.
* Used for loader routine before table is inserted.
** With Automatic Divide special feature.
.2
PHYSICAL FORM
. 21
Storage Medium: .
.22
Physical Dimensions
. magnetic core.
.221 Magnetic core storage
Array size:. . • .
100 by 100 by 12 bits.
.23
Storage Phenomenon:
direction of magnetization.
.24
Recording Permanence
.241 Data erasable by
instructions: . .
. 242 Data regenerated
constantly: .
. 243 Data volatile:
..
.244 Data permanent: .
. 245 Storage changeable:
.28
no .
no (with normal power
on/Off. procedures).
no .
no.
Access Techniques
.281 Recording method: .
.282 Reading method: . .
. 283 Type of access:
..
.29
yes •
coincident current.
same as recording .
uniform •
Potential Transfer Rates
.292 Peak data rates
Cycling rate: .
Unit of data: .
Conversion factor:
Data rate: . . . .
Compound data rate: .
1962 by Auerbach Carporation and BNA Incorporated
50,000 cycles/sec.
2 digits.
6 bits per digit.
100,000 digits/sec.
100,000 digits/sec.
12/62
IBM 1620 MODEL 1
412:041.300
§
041.
.• 5
.3
DATA CAPACITY
.31
Module and System Sizes
Minimum
Storage
Identity:
basic
storage
Characters:
Instructions:
Digits:
Modules:
. 32
10,000
1,666
20,000
1
Rules for Combining
Modules: . . . . . .
Maximum
Storage
basic
storage
+ 1623-1
20,000
3,333
40,000
2
basic
storage
+1623-2.
30,000.
5,000.
60,000.
3.
all combinations are listed
above.
ACCESS TIMING
.51
Arrangement of Heads:
· 52
Simultaneous
Operations:. . . . . . none.
.53
Access Time Parameters and Variations
.531 For uniform access.
Acce.,ss time: .
Cycle time:. .
For data unit of:
CHANGEABLE
STORAGE: .. . . . . none.
.7
PERFORMANCE
· 71
Data Transfer
,/
Pair of storage units pos.sibilities
With self:. . • • . . . yes.
Transfer Load Size
With self:
GONTROLLER
.41
Identity:
.42
.422 Off-line:
. 43
.73
none for basic storage.
2301 Core Storage Adapter
for 1623 Model 1.
.8
2301 and 2302 Core Storage
Adapters for 1623 Model
2.
With self:
Connection to Device
. 431 Devices per controller: 1.
.432 Restrictions: .
none.
12/62
25, 000 digits/ sec.
ERRORS, CHECKS AND ACTION
1 adapter of each type.
Model 2301 must be present in order to use Model
2302 Adapter and 1623
Model 2 Storage.
none.
.
up to size of core storage.
Effective Transfer Rate
Connection to System
. 421 On-line:
20 j.I sec.
20/.1 sec.
1 digit.
.6
· 72
.4
1 access device.
.
Error
Check or
Interlock
Invalid address:
parity and limit check
Invalid code:
Receipt of data:
parity only.
parity check
Recording of data:
Recovery of data:
records parity bit.
parity check
Dispatch of data:
Timing conflicts:
Reference to locked
area:
transmits parity bit.
no conflicts•
I AUERBACH / .@
none.
Action
indicator aiarm,
halt•
I
indicator, alarm;
system halt
optional.
indicator, alarm;
system halt
optional•
412:042.100
•
II
STANDARD
EDP
IBM 1620
R[PDRTS
Internal Storage Drive
Disk Storage Drive
INTERNAL STORAGE: 1311 DISK STORAGE DRIVE
§
042.
.1
GENERAL
. 11
Identity:
.12
Basic Use:
.13
Description
.13
. . . • • . . Disk Storage Drive.
1311 Models 3 and 2.
auxiliary storage.
The 1311 Disk Storage Drive is a new development
in low cost random access storage. It is available
for the IBM 1401 and the new IBM 1440 Data Processing Systems, as well as the 1620, and features
rapid interchangeability of the "disk pack" storage
cartridges. The system is suitable for either random or sequential processing methods.
Each disk pack consists of six discs on a common
vertical axis. Data can be recorded on ten disc surfaces; the top and bottom surfaces of the pack are
not used. Each recording surface is divided into
100 concentric bands, each band is divided into 20
sectors, and each sector holds a 5-digit address
and 100 numeric BCD-coded digits (six bits plus a
parity bit). Thus the data capacity is 2, 000 digits
per band, 200, 000 digits per surface, and
2, 000, 000 digits per pack. Up to four Disk Storage
Drives can be connected to a 1620 system, so that
the maximum on-line data capacity is 8, 000, 000
digits.
The disc is accessed by means of a comb-like
mechanism containing five arms that move horizontally between the discs. Each arm has one readwrite head mounted on the top and one on the bottom,
and each head serves one disc surface. The entire
access mechanism moves as one unit, so all ten
read-write heads are always positioned at corresponding bands on their respective surfaces. The
term "cylinder" is applied to the ten bands (one on
each disc surface) that can be read or recorded with
no movement of the access mechanism. There are
100 cylinders per Disk Storage Drive, and each cylinder can hold 20, 000 digits.
Time for access mechanism movement ranges from
zero (for successive references to a previously selected cylinder) to 400 milliseconds; average random
access time is 250 milliseconds. The access arms
cannot move directly from one cylinder to another;
instead, the arms retract all the way to the "home"
position (beyond track 00) and then move back to the
selected cylinder. The result is that movements
between adjacent cylinders require from 85 milliseconds (track 00 to 01) to 390 milliseconds (track
98 to 99). Maximum delay due to rotation is 40
milliseconds. There is also a head select delay of
2 milliseconds. Total reference cycle time to read
a randomly-placed lOa-digit record, update it, rewrite it, and execute a programmed write check is
©
Description (Cont'd)
354 milliseconds. If no access motion is required,
the total reference cycle time is reduced to 104
milliseconds.
Peak data transfer rate is 50, 000 digits per second,
and the effective bulk transfer rate when reading
a cylinder from the disc file is 30, 000 digits per
second. The bulk transfer rate, when recording
from storage and checking by re-read, is approximately 18, 000 digits per second in a full cylinder
operation, and approximately 5, 100 digits per second when recording one band of data. Moving data
from the disc file to core storage and to another
disc file location, with re-read checking, is done at
11,200 digits per second using a full cylinder of
data (20, 000 digits).
The standard processor instruction format is used
for referencing a fourteen-digit Disk Control Field
in core storage. The Disk Control Field contains
the sector address, count of sectors to be transferred, storage address, and a drive number digit.
A single read or write instruction can transfer from
1 to 200 consecutive sectors of information; i. e.,
from 100 digits to 20, 000 digits in multiples of 100
digits. The number of sectors is designated by the
programmer. An additional instruction automatically reads or records the 20 sectors in one band,
including the 20 addresses (Read or Write Disk
Track instruction set). Thus no additional data may
be stored using this instruction. An additional set
of instructions allows transfer operations to be performed with a check on record length by means of
the group mark delimiter.
In 1620 and 1710 systems, each sector of data may
be interlocked against recording. This is accomplished by a flag bit present in the sector address
preceding the sector to be protected. This protection system is practical insofar as a missing flag
bit should be detected by the parity checking circuits. There is no positive method of protecting a
complete disk pack file from being over-written.
Checks are made for parity errors and unequal address comparisons. The "write disc check" instruction causes a character-by-character comparison of data just written on the disc with the data in
core storage. It usually follows each write operation. All disc errors cause the setting of testable
indicators. These indicators are: Address Check,
Wrong Length Check, Cylinder Overflow, and Any
File. Lamps associated with these indicators are
on the 1620 Processor console. The usual Processor
error lamps and indicators are also used. A File
Check switch is on the 1620 console; when set to
STOP, disk file and input-output errors cause an
immediate halt of the 1620, as well as the normal
halt of the disc file operation.
1962 by Auerbach Corporation and BNA Incorporated
12/62
IBM 1620
412:042.130
§
042.
.13
Description (Cont'd)
Disk Storage Drive seek time can be fully overlapped with internal processing on all four drives.
A "Branch If Access Mechanism Busy" instruction
is not provided for the 1620, although it is available
on the 1710 Control System. No processing is possible during disc read or write operations.
The removable disk packs are 14 inches in diameter,
4 inches high, and weigh less than 10 pounds, including covers. A disk pack can be removed from a
Disk Storage Drive and replaced by another disk
pack in one minute. When a disk pack is not
mounted on a drive, the pack and its cover combine
to form a sealed container that can be conveniently
stored and transported. One disk pack is supplied
with each 1311 Disk Storage Drive. Additional disk
packs cost $490 each, f.o.b. SanJose.
..
.14
Availability:
.15
First Delivery:
?
.16
Reserved Storage: .
none. Note that each 100digit sector is preceded
by a S-digit address, but
these address digits are
not counted as storage.
.2
PHYSICAL FORM
• 21
Storage Medium: .
.22
Physical Dimensions
?
.281 Recording method:. • . by one of the magnetiC
heads on access arms
which move horizontally
in unison.
· 283 Type of access
Description of stage Possible starting stage
, Wait for selected
sector for reading
or recording: • . . if same band was
previously selected.
Move heads to home
position and then
to selected band: . mandatory to access new
band.
.29
· 291 Peak bit rates
Cycling rates:
Bit rate per track:
• 292 Peak data rates
Unit of data: .
Conversion factor: •
Gain factor:
Data rate: .
.3
DATA CAPACITY
.31
Module and System Sizes
Identity:
14 inches O. D.
thin.
• 23
Storage Phenomenon:
magnetization.
· 24
Recording Permanence
Discs:
Instructions:
Digits:
Modules:
6.
.32
. 243 Data volatile: . . .
· 244 Data permanent: . .
· 245 Storage changeable:
. 25
digit.
7 bits per digit (6 plus
parity).
1 track/band.
52,500 digits/sec, counting
address digits as above.
Maximum
Storage
0
0
0
0
1311 Model
3
1311 Model
2
6
167,000
2,000.000
1
6
167,000
2,000,000
1
Rules for Combining
Modules: • . . . . .
yes.
no.
1,500 rpm.
367,500 bits/ sec/track,
counting the additional
S-digit address normally
used with each sector •
Minimum
Storage
multiple magnetic disks .
. 222 Disc
Diameter:
Thickness:
Number on shaft: .
• 241 Data erasable by
instructions: •.
• 242 Data regenerated
constantly: . . .
Potential Transfer Rates
.4
CONTROLLER
.41
Identity:
.42
Connection to System
no.
no.
yes.
1-1311 Model
3 and 3-1311
Model 2 •
24.
668,000.
8,000,000.
4.
first module must be Model
3.
next three modules must
each be Model 2.
part of 1311 Model 3 .
Adapter required as
follows:
3339 for 1620 Model 1.
3340 for 1620 Model 2 .
Data Volume per Band of I Track
Characters: .
Digits: • . . .
Instructions: .
Sectors: .
Cylinder:. . ~
1,000.
2,000.
166.
20.
0.1.
,.421 On-line:
.422 Off-line:
• 43
1-1311 Model 3.
none.
Connection to Device
.26
Bands per Physical Unit: 100 per disc surface.
.431 Devices per controller: 4 modules counting Model
.27
Interleaving Levels: •
.28
Access Techniques
.432 Restrictions:. . . . . . 1 Model 3 and 1 to 3 Model
2; no other restrictions.
Operation is not buffered.
3.
12/62
1.
412:042.440
INTERNAL STORAGE: 1311 DISK STORAGE DRIVE
§
042.
.44
.53
Data Transfer Control
.441 Size of load
Variable length:
Fixed length: .
.442 Input-output area:
.443 Input-output area
access: .
.444 Input-output area
lockout: .
• 445 Synchronization: .
.447 Table control:
. 448 Testable conditions: .
.5
ACCESS TIMING
. 51
Arrangement of Heads
. 511 Number of stacks
Stacks per system:.
Stacks per module:.
Stacks per yoke: • •
Yokes per module: •
• 512 Stack movement: ••.
. 513 Stacks that can access
any particular
location: • • . . • .
.514; Accessible locations
By single stack
With no movement: .
With all movement: •
By all stacks
With no movement: .
.532 For variable access
Example
Wait for selected
sector for reading
or recording:
0 to 40 m. sec
Move heads to
home position and
then to selected
band:
85 to 390 m. sec
yes.
automatic.
none.
Seek Complete on 1710
Control System, as
optional feature.
.6
CHANGEABLE STORAGE
.61
Cartridges
20 m. sec.
250 m. sec.
2,000,000 digits (6 discs).
.611 Cartridge capacity:
.612 Cartridges per module: 1.
yes .
• 613 Interchangeable: . . •
.62
40 max.
10.
10.
1.
horizontal •
1.
20 sectors.
2,000 sectors.
Loading Convenience
.621 Possible loading
While computing
system in use:
While storage system
in use:
.....
. 622 Method of loading: •
•. 623 Approximate change
time: • • • .
• 624 Bulk loading: • • . • .
yes.
yes, if particular module
not addressed •
operator •
one minute.
yes; 1 cartridge.
200 sectors per module.
200 to 800 sectors per
system •
.7
AUXILIARY STORAGE PERFORMANCE
• 71
Data Transfer
Pair of storage units possibilities
seeking a specified sector.
reading.
recording.
= at most 1 per module.
a + b + c = at most N) for a disc storage
=0
Time
digit.
a +b+c
be
Stage
1 to 200 sectors of 100
digits per sector; number
of sectors set by programmer.
20 sectors of 100 digits per
sector (one band).
core storage; demand on
processor is 100%.
• 515 Relationship between
stacks and locations:. three most significant
digits of Sector Address
denote head and band
(cylinder) number.
.52 Simultaneous Operations
A:
B:
C:
Access Time Parameters and Variations
system of
1 to N modules, where N is
at most 4. (*)
With self: . . . . . .
With core storage: .
. 72
Transfer Load Size
With core storage: .
With core storage:. . .
(*) claimed by the manufacturer.
©
yes, by programming.
yes.
1962 by Auerbach Carporation and BNA Incorporated
1 to 200 sectors; number of
sectors selected by
programmer.
1 block of 20 sectors (one
band).
12/62
IBM 1620
412:042.730
§
042 •
• 73
Effective Transfer Rate: see table of effective rates.
EFFECTNE TRANSFER RATE, DIGITS PER SECOND
Including access movement
.8
OPERATION
1 band of
2, 000 digits
lor more
cylinders of
20, 000 digits
each
1 band of
2, 000 digits
Reading to core storage
6,440
30,000
33,333
47,700
Recording from core
storage without check
read
6,440
30,000
33,333
47,700
Recording from core
storage with check
read
5,130
18,000
14,300
23,300
Reading to core storage
and recording without
check read
3,220
15,000
16,667
23,850
Reading to core storage
and recording with
check read
2,850
11,200
10,000
15,000
ERRORS, CHECKS AND ACTION
Error
Invalid address:
Invalid code:
Receipt of data:
Recording of data:
Recovery of data:
Dispatch of data:
Timing conflicts:
Physical record
missing:
Reference to
locked area:
Attempt to read or
record a sector beyond end of cylinder:
12/62
No access movement
Check or
Interlock
Action
check on nonexistent
drive unit
check on matching
sector address
alarm, indicator,
optional halt.
alarm, indicator,
optional halt.
?
parity check at 1620
alarm, indicator,
optional halt.
alarm, indicator,
optional halt.
parity check in 1620
and optional use of instluction to read rec0rd back and compare
to core storage
alarm, indicator,
parity check in 1620
optional halt.
parity bit included.
wait.
interlock
optional check on
record length
ala~m t
check on optional lock
flag bit in sector
addre..
alarm, indicator,
optional halt.
check
alarm, indicator,
optional halt.
indicator.
optional halt.
1 cylinder of
20, 000 digits
412:051.100
.STAIIDIRD
EDP
•
IBM 1620 Modell
Centro I Processor
REPORTS
CENTRAL PROCESSOR
§
051.
.12
.1
GENERAL
• 11
Identity: . . . . . . . . Computer.
1620 Model 1.
.12
Description
The 1620 Computer is a two-address sequential
processor oriented toward scientific applications.
Operands are held as variable-length fields of decimal digits, delimited by a flag bit in the six- bit code
of the most significant digit of the field. Data may
also be stored in alphameric fields, two digits per
character. Instructions are twelve digits long and
are performed sequentially. The instruction format
is two digits for the operation code, five digits for
the P operand address, and five digits for. the Q operand address. Data operations are performed serially by digit upon the operands, which may be any
length. The Central Pl;'ocessor cabinet includes the
Console, desk work area with Console I/O Typewriter, 20,000 digits of core storage, and space for
adapters used with the input-output devices.
Internal records are defined by the presence of a
record mark code digit. Records, as well as fields,
may be moved within core storage. A numeric field
is addressed at the least significarit digit of the field,
and a record is addressed at the high order end of
the field. The record mark terminates a write operation on punched tape. Record marks are generated in storage by the following: end of record on
punched tape; Record Mark key on Console I/O
Typewriter; record mark as data on cards or
punched tape.
Although the core storage cycle is 20 microseconds,
instruction times are a good deal longer because of
serial operation. Add-Subtract-Compare operations
require 560 microseconds for five-digit operands,
while data movement is performed almost twice as
fast as this.
Add, Subtract, and Multiply are standard instructions in the Central Processor. Divide is performed
by subroutine. Optional features are provided at
extra cost and are listed below. A complete set of
conditional branch instructions is standard, and allows branching on presence or absence of the condition specified. Four sense switches are available
for interrogation. The Branch and Transmit instruction provides a jump to the P-address, and also
transmits the field specified by the Q-address to the
storage area immediately pre,ceding the P field.
This field may contain parameters needed by the
subroutine starting at P.
©
Description (Cont'd)
The arithmetic, comparison, and data movement instructions have counterparts for handling litera~s.
For example there is Add, and Add Immediate. The
Add instruction adds the operand addressed by the
Q-address to the operand addressed by the P-address, while the Add Immediate instruction treats
the Q-address field of the instruction as a number (a
literal) and adds this number to the operand addressed by the P-address. The Add Immediate instruction is helpful in address modification, since
there is no index address modification in the 1620.
The normal Add instruction could be used, however,
to add a constant to an address since the operating
times of the Immediate instructions are no faster
than the operating times of the normal instructions.
No editing instruction exists, but the typewriter can
be commanded to space, tabulate, or perform a carriage return. Normally, carriage return occurs
automatically at the end of each line.
Comparisons are performed digit- by-digit, whether
the field is numeric or alphameric.
Arithmetic (add, subtract, and multiply) is performed using data stored in the table area of storage
(locations 00100 through 00399). The operand dIgits
are used to address the results present in the tables.
The add-subtract table contains 100 locations.
Optional Features
Automatic Divide Special Feature: Replaces fixed
point division subroutines. The feature includes a
Load Dividend instruction for moving and pOSitioning
the dividend, and a Divide instruction for performing
the division operation.
Indirect Addressing Special Feature: Provides facility for a Q-address to be interpreted as an address
location rather than an operand location. It applies
to most instructions, and is recursive.
Additional Instructions Special Feature: Three instructions are provided as a package. They are
Move Flag: Moves the presence or absence of a
sign, field definition, or indirect address flag
from one storage location to another, and clears
the flag, if present, from the former location.
Transfer Numeric Strip: Compresses numeric
data field which is in two-digit alphameric coding form to single digit numeric form.
Transfer Numeric Fill: Expands single-digit
numeric data field to two-digit alphameric form.
1962 by Auerbach Corporation and BNA Incorporated
12/62
412:051.120
§
IBM 1620 MODEL 1
051.
· 12
· 232 Instruction layout:
Description (Contd. )
Automatic Floating Point Operations SpeciaJ.. Feature'
Provides a"full set of floating point arithmetic, shift,
and movement instructions. Operand sizes may be
fixed at eight digits or variable from 2 to 100 digits
for the fixed-point part. The exponent part has a
maximum value of ± 99.
• 13
. 14
.2
.21
.211
.212
.213
.214
· 215
· 216
· 217
. 218
· 22
.221
.222
· 223
· 23
.231
Part
Size (digits)
• 233 Instruction parts
Name
Operation:
Q-Address
5
Purpose
specifies operation to be
performed.
P-Address:
1. operand address in core
storage;
2. address of result of adAvailability: . .
presently available .
dition or subtraction;
3. jump address;
First Delivery:
October. 1960 .
4. address for transfer; or
5. input-output starting
PROCESSING FACILITIES
address.
Q-Address: . . . . . 1. operand address;
Operations and Operands
2. operand literal;
3. starting address for
Operation and Provision
Radix
Size
transfer;
Variation
4. address of testable
indicators;
Fixed point
automatic
decimal
2 to N digits,
5.· typewriter control funcAdd -subtract:
limited by core
tion; or
storage.
6. select input-output
Multip)y
device.
Shott:
none"
.234 Basic address structure: 2-address.
2 to N digits.
automatic
decimal
Long:
.235 Literals
Divide
Arithmetic: . . . .
any; 5 digits practically.
No remainder:
none.
Comparisons
and
decimal
Remainder:
subroutine
2 to 45 digits.
tests: . . . .
any; 5 digits practically.
optional feature decimal
2 to 100 digits.
Incrementing
Floating point
modifiers: •
no modifiers (no indexing);
decimal
Add -subtract:
2 to 45 digits.
subroutine
however literals in arith2 to 100 digits.
optional feature decimal
metic instructions are
Multiply:
decimal
2 to 45 digits.
subroutine
useful for incrementing
2 to 100 digits.
optional feature decimal
decimal
2 to 45 digits.
Divide:
subrout~ne
addresses.
, optional feature decimal
2 to lOO digits.
· 236 Directly addressed operands
decimal
Shift:
2 to 45 digils.
subroutine
.2361 Internal storage type: core storage.
optional feature decimal
2 to 100 digits.
Minimum size: . . .
2 digits.
none.
Boolean:
Maximum size: . . .
complete store.
Comparison
Volume accessible:.
complete store.
Numbers:
automatic
2 to N digits.
· 2362 Increased address
Absolute:
none.
capacity: • . .
none.
automatic
Letters:
1 to N char.
none; literals in arithmetic
.237 Address indexing: .
Mixed:
automatic
1 to N char.
instructions are useful
Collating sequence: special symbols, A to I, 0, J to Z, 0 to 9.
for incrementing
addresses.
Code translation:
none.
·
238
Indirect
addressing
Radix conversion:
none.
: 2381 Recursive: .
yes.
alter size by Set Flag
Edit format: .
..2382 Designation:
flag bit in least significant
command.
digit of address •
.
Table look-up:
none .
absolute address has no
· 2383 Control:. . .
flag bit.
Special Cases of Operands
· 2384 Indexing with indirect
addressing:
not possible (no indexing).
Negative numbers: •
flag bit over least signifinone.
cant digit of decimal field. i. 239 Stepping:. ~ . . .
Ze:r;o:.
both + and - zero. Signs
are ignored when compar- .24 Special Processor
. Stor~: . . . . . . . none.
ing 2 all-zero fields.
Operand size
.3
SEQUENCE CONTROL FEATURES
determination:
flag bit over most significant digit of decimal
.31 Instruction Sequencing
field.
.311 Number of sequence
Instruction Formats
control facilities: .
1.
.314 Special sub-sequence
Instruction structure:. 12 digits.
none.
counters: . . . . . •
12/62
'
CENTRAL PROCESSOR
§
412:051.315
051.
. 315 Sequence control step
size: • . . . .
• 316 Accessibility to
routines: . . .
1 instruction (12 digits).
by BT instruction; can store
next address in sequence;
address used by BB
instruction .
• 31'7 Permanent or optional
modifier: • •
none.
. 32
Look-Ahead:.
none.
.33
Interruption:.
none.
. 34
Multi -running:
none.
,35
Multi -sequencing:
none.
.4
PROCCESSOR SPEEDS
,41
Instruction Times in p, secs
.415
. 416
.417
.418
Counter control: •
Edit: •.
Convert:
Shift: . .
.42
Processor Performance in p, secs
.421 For random addresses
Fixed point
c = a + b: . . • •.
b = a + b: . . . . .
Sum N items; per
item:
c = ab: . . . . . .
c = alb
U sing macro for
subroutine: . , .
no
no
no
no
counters •
edit instruction.
convert instruction.
shift instruction.
320 + 1200.
160 + 800 .
160+ 800.
no + 800 +
168D2.
9,000 + 3.2(980 + 8200 +
52002 ).
Using Automatic Divide
special feature: • 720 + 8600 + 5200 2 •
. 411 Fixed point
Add-subtract:
160 + 80D.
560 + 40D + 168D2.
Multiply: . • .
Divide
By subroutine called
by DlV macro: . . 3.2(980 + 820D + 520D2).
By Automatic Divide
(special feature):. 160 + 740D + 520D2.
.412 Floating point
Using Automatic Floating Point
Operations special feature (Automatic Divide
special feature required)
Add-subtract (FADD,
FSUB): . . . . • . 400 + 100D.
Multiply (FMUL):. 1,120 + 80D + 168D2.
Divide ,(FDIV): •• 880 + 9400 + '5200 2 .
Using subroutines called
by macros
Add- subtract (FA, FS)
Fixed length mantissa
(8 digits): . .
9,800.
Variable-length
mantissa: • . . 6,854 + 4820 + 50 2 .
Multiply (FM)
Fixed length mantissa
(8 digits): . .
18,000.
Variable length
mantissa: . . . 7,400 + 2400 + 16802 .
Divide (FO)
Fixed length mantissa
(8 digits)
With Automatic
Divide: . . . 55,000.
Without Automatic Divide: 70,000.
Variable length mantissa
With Automatic
Divide: • . . . 7,890 + 1,5000 + 5200 2 .
Without Automatic Divide:. 1.9(7,890+ 1,5000 +
52002 ).
.413 Additional allowance for
80.
Indirect addressing:
800 (fixed or floating
Re-complementing:
point).
.414 Control
Compare: •
160+ 800.
Branch: •.
200.
© 1963
Floating point subroutines called by macros; 8
decimal digits
c = a + b: • • . . •
28,520.
b = a + b: • . . • •
18,800.
Sum N items; per
18,800.
item:
c = ab: . . . . . .
36,700.
c = alb
With Automatic Divide
special feature: . 73,700.
Subroutine only:. . 88, 700.
Floating point using Automatic Floating Point
Operations special feature; Automatic Oivide
special feature required; 8 decimal digits
c = a + b: • . • •.
1, 760.
b = a + b: . . . ••
1, 200.
Sum N items; per
1,200.
item: •
c = ab: . . . . •
13,100.
41,700.
c = alb: . . . . .
.422 For arrays of data
Fixed point
2, 520 + 120D.
ci = ai + b j : . . .
1,560+ 800.
b j = ai + b j : . . .
Sum N items; per
1,400+ 800.
item: . • .
2,280 + 800 + 168D2.
c = c + aibj:'
Floating point subroutines called by macros; 8
decimal digits
30,700.
ci = ai + b j : . . • •
20,400.
b j = ai + b j : . . . •
Sum N items; per
20,000.
item: •.•
47,400.
c = c + aibj:' . . •
Floating point using Automatic Floating Point
Operations special feature; Automatic Divide
special feature required; 8 decimal digits
ci = ai + b j : . . .•
3,960.
bj = ai + bj : . . . ,
2, 760.
Sum N items; per
2,440.
item: •.•
15,800.
c = c + aibj: .
by Auerbach Corporation and BNA Incorporated
3/63
412:051.423
§
051.
.423 Branch based on comparison
Numeric data: •
2, 060 + BOD; D = no. digits
in key.
Alphabetic data:
2, 060 + 16OC; C = no.
alpha char in key.
. 424 Switching
Unchecked:
960.
Checked: •
1,980.
List search
No indirect
addressing:
I, 960N + I, 120.
Using Indirect
Addressing feature: I, 960N + 840.
.425 Format control per character
Unpack
Scientific:. . • • . . 0; usable in input area.
Compose
Scientific
Fixed point: . . . . 60; for moving data.
Floating pOint: . . 70; for moving data.
. 426 Table look up per comparison
For a match: . . . •. 1,960.
For least or greatest: 2,680.
For interpolation
point: • . . . . •
I, 960.
.427 Bit indicators
Set bit in separate
location: . • . .
200.
Test bit in separate
location:
240 .
• 428 Moving: • . . . . . .
160 + 40D.
3/63
IBM 1620 MODEL 1
.5
ERRORS, CHECKS, AND ACTION
Error
Check or
Interlock
Overflow:
check
indicator, alarm;
halt or programmed
action•
Exponent underflow
and overflo w:
check
indicator. alarm; halt
or programmed
action.
Zero divisor:
Invalid data:
overflow check.
parity c hec k
Invalid operation:
Arithmetic error:
Invalid address:
check
none.
parity check
Receipt of data:
parity check
Dispatch of data:
parity check
Action
alarm. indicator;
halt or programmed
action.
halt. alarm. indicator.
alarm. indicator;
halt or programmed
action,
alarm, indicator;
halt or programmed
action•
alarm, indicator;
halt or programmed
action.
Incorrect operand
overflow check.
length:
Mispositioned divisor: overflow check.
412:061.100
IBM 1620
Console
CONSOLE
§
061.
.1
. 24
GENERAL
Name
· 11
Identity:........ Console; built into 1620
Processor.
· 12
Associated Units: . . . Console I/O Typewriter
stands upon console desk.
· 13
Description
Stop/Single Instruction
Execute (SIE):
Instant Stop/Single
Cycle Execute (SCE):
. 25
The Console is built into the 1620 Processor. The
Console consists of a work area, a two-panel control and lamp are.a, and a typewriter. The typewriter is described in Section :081, and is located
on the right side of the Console.
The Console panels contain: operating switches,
sense switches, and keys; error lamps; operation,
address, and storage register lamps; and a number
of lamps used primarily for diagnostic testing by
mM Customer Engineers. The diagnostic indicators
include a lamp to indicate that the last card has been
read by the card reader. Four sense switches are
provided, and also three toggle switches which are
used with the Parity, Input-Output, and arithmetic
error lamps to select whether to stop or continue
under program control in case of errors. All digitindicating lamps are in 6-bit binary-coded form.
Although the Console does not contain a Load switch,
the Card Reader unit of the Card Read Punch has a
Load key which causes the contents of one card to
be transferred to a specific area of core storage to
initiate program operation.
.2
CONTROLS
.21
Power
Name
Form
\.
Connections:...... none.
.23
Stops and Restarts
Form
Function
Start:
momentary
switch
momentary
switch
sets automatic mode and
starts processing.
terminates I/O operation
and sets manual mode.
momentary
switch
stops computer after execut,ion of present instruction.
momentary
switch
stops computer. at end of
present machine cycle.
Stop/Single Instruction
Execute (SIE):
Instant Stop/Single
Cycle Execute (SCE):
causes one instruction to be
executed at a time.
momentary
s witch
computer executes one
machine cycle •
Name
Form
Function
Reset:
momentary
switch
resets indicators, alarms,
registers.
.26
Loading
Name
Form
Function
Insert:
momentary
switch,
steady lamp
momentary
switch
sets automatic mode and
activates keyboard for data
entry.
terminates keyboard entry
and turns off Insert lamp.
momentary
switch
initiates transfer of 1 card
(80 char) to core storage,
and starts processing.
Release:
Load key (on Card
Read unit):
.27
Sense Switches
Name
Form
Function
Program Switches:
4 two-posi -
provides 4 branch contrel
indicators.
tion
switches
Special
Name
Form
Function
Save:
momentary
indicating
switch
lamp re mains on: next address in sequence is stored
in special register: used in
Console operating
procedures.
turns off all
power.
Name
Release:
momentary
switch
Resets
2 pos. switch.
• 22
Function
momentary resets indicators, alarms.
switch
Note: Depressing Reset and Check Reset keys together in model 2
also resets core storage ta zeros.
Function
pull switch
Form
Check Reset (model
2 only):
.28
Power On/Off:
Emergency
Off Switch:
\
Stepping
©
.3
DISPLAY
.31
Alarms
Name
Form
Function
Parity:
lamps
I/O:
lamps
OFLOW:
lamp
Check Stop:
Reader No Feed:
Punch N'o Feed:
Thermal:
lamp
lamp
lamp
lamp
Parity error in memory registers and memory address
registers.
parity error in input-output
transfer.
overflow due to +, -, +,
compare.
machine or I/O parity failure.
reader not ready.
punch not ready.
temperature of any component in system too ~l;\!,.
1962 by Auerbach Corporation and BNA Incorporated
12/62
IBM 1620
412:061.320
§
061.
.32
Conditions
Automatic and Manual
indicate computer mode.
lamps:
Save lamp: . . . . . . indicates an address is
being stored.
Comparison
indicator lamps:
indicate results of comparisons.
Power On lamp_ •
lit when power is on.
lit when system is ready
Power Ready lamp:.
for operation.
on when internal temperaThermal lamp: •.•
ture of system component
is too high.
.34
ENTRY OF DATA
.41
Into Control Registers:
.42
Into Storage: . . . . . . done by inserting, via typewriter, a read instruction
to storage area desired,
then either typing or
starting paper tape
reader.
.5
CONVENIENCES
. 51
Communication:
none .
.52
Clock: . . .
none.
indirectly by manually
entered instructions.
Comment
Name
. 33
.4
Control Registers
Name
Form
Function
. 53
Desk Space:
Operation Register:
Multiplier:
Sense and Branch:
Memory Address
Register:
Memory Address
Register Display
Selector:
lamps
lamps
lamps
operation code; 2 digits.
multiplier digit.
sense and branch order.
approx. 15 by 40 inches,
30 inches from floor.
.54
View: .
lamps
1 of a addresse., 5 digits.
designed for operation by
operator seated at console desk; unobstructed
view in all directions.
a-position
switch
selects Address Register for
display.
.6
INPUT-OUTPUT UNIT:
Form
Function.
lamps
2 digits",f storage.
lamp
storage digit addressed.
Console I/O typewriter operates at 10 chari sec for
output in model 1 and at
15 char/sec in model 2;
and manual speeds for input. It is described fully
in Section :081.
Storage
Name
Memory Buffer
Register:
Memory Data
Register:
412:071.100
IBM 1620
Input-Output
1621 Paper Tape Reader
INPUT-OUTPUT: PAPER TAPE READER
§ 071.
.1
GENERAL
.n
Identity:
.12
Description
.23
Multiple Copies: . . . . none.
· 24
Arrangement of Heads
Paper Tape Reader.
1621.
The Model 1621 Paper Tape Reader reads eighttrack chad paper tape in the forward direction at 150
rows per second, using a photoelectronic sensing
head. Each row includes an odd row parity check
bit which is checked while reading takes place. One
track is reserved for an end-of-line character punch
(EL code), which terminates the read operation.
The reader is capable of reading chadJ.ess \ape, but
chad tape is normally used. Tape threading is not
convenient.
A read command specifies data to be entered numerically or alphamerically. In the 'numeric mode, the
digits read are placed in successive core storage
locations. A row containing a flag pit retains it
when stored. In the alphameric mbde, each row
read is automatically inserted into two successive
storage locations in the normal internal two-digit
alphameric code. In either mode, sensing a record
mark code on tape causes a record mark code to be
placed in storage, and the tape contip.ues moving
until the EL code is sensed. This also causes a
recqrd mark code to be inserted into storage.
If ,a parity error is found, tape continues moving but
an indicator is set. Depending on the setting of a
console switch, the system either stops after ~e EL
code is sensed, or continues to the next instruction.
The next instruction can be an indicator testing command.
A numerical read command will accept alphameric
data from the tape, but incorrect (garbled) characters are sent to core storage. Each core location
will have correct parity, however.
Use of station: .
Stacks: . . . . .
Heads/ stack:. .
Method of use: .
.3
EXTERNAL STORAGE
.31
Form of Storage
.311 Medium: . • .
.312 Phenomenon: . . .
.32
.321 Serial by:
. 322 Parallel by:
· 324 Track use
Data: . . . . . .
Redundancy check:
Timing: . . . . . .
Control signals:
Unused: .
Total: •
. 325 Row use
Data: .
Redundancy check: .
Timing: . . . . .
Control signals:
Unused: .
1 (end-of-line).
O.
8 plus sprocket .
all except 1 per block.
O.
O.
1 (end-of-line).
O.
. 34
Format Compatibility:. all devices using standard
I-inch paper tape.
· 35
Physical Dimensions
.351 Overall width:
· 352 Length:. . • .
October, 1960 .
.2
PHYSICAL FORM
.4
CONTROLLER
· 21
Drive Mechanism
.41
Identity:
.42
Connection to System
...
sprocket drive.
.421 On-line:
©
O.
as in Data Code Table No.
5.
First Delivery:
. 222 Sensing system: . . . . photoelectronic.
6.
1.
Coding: .
. 14
Sensing and Recording Systems
opaque paper tape.
full or partially punched
holes.
· 33
3 to 4 months.
· 22
1 row at a time.
1 to N rows at 10 rows/inch •
8 tracks at standard
spacing.
Availability: .•
2.
swinging arm.
each approx. 16 inches .
?
?
8.
Positional Arrangement
.13
· 211 Drive past the head: .
· 212 Reservoirs
Number:
Form: . .
Capacity: .
. 213 Feed drive:
.214 Take-up drive: .
sensing.
1.
.422 Off-line:
1962 by Auerbach Corporation and BNA Incorporated
1 inch.
1,000 feet max.
Model 5514 Paper Tape
Reader Adapter, in 1620
model 1 Computer.
Model 5515 Paper Tape
Reader Adapter, in 1620
model 2 Computer.
1, used for paper tape
reader and punch .
none.
12/62
412:071.430
§
IBM 1620
071.
. 43
Connection to Device
.431 Devices per controller: 1 paper tape reader and 1
paper tape punch .
• 432 Restrictions: . . . . . . only 1 device (reader or
punch) may be operated at
a time.'
.44
Read alphamerically:
.442 Input- output areas:
.443 Input-output area
access: .
. 444 Input-output area
lockout: . .
.445 Table control: • .
• 446 Synchronization: .
· ..
· ..
.. 5
PERFORMANCE
.61
Conditions
II:
.62
Speeds
• 621 Nominal or peak speed: 150 rows/sec .
.622 Important parameters
Tape speed:
15 inches/sec.
1 row/block (end-of-line
.623 Overhead:
char).
. 624 Effective speeds:
150N/(N+l) row/sec .
N = no. char/block.
1 to N char; N limited by
core storage.
1 to N char; each char requiring 2 core storage
positions.
core storage.
.63
yes.
no .
automatic •
Demands on S:t:stem
. ..
. ...
EXTERNAL FACILITIES
· 71
Adjustments:...... none.
• 72
Other Controls
PROGRAM FACILITIES AVAILABLE
1 to N char; N limited by
core storage.
.512 Block demarcation
Input: . . . ·
...
end-of-line char (EL).
.52
Input-Output Operations
• 521
. 522
• 523
. 524
• 525
• 526
Input: . . .
Output: . .
Stepping: .
Skipping: .
Marking: .
Searching:
input 1 block forward .
none .
none .
none .
none .
none.
. 53
Code Translation:
automatic.
.54
Format Control
Control: • . . .
Format alternatives:
Rearrangement:
Insert spaces:
Section sizes: .
Select code:
Rewind:
Unload: . . .
12/62
program.
2 (numeric or alphameric).
no.
no.
no.
Control Operations
Disable: . . . • . .
Request interrupt:
Select format:
.56
Form
Comment
Blocks
.511 Size of block:
. 55
100 .
.7.
Function
.51
processor.
I, .II.
IOOON/150.
Component:
Condition:
m. sec per block:
or
Percentage: ...
each character.
.'
read numerically.
read alphamerically.
I :
....
Data Transfer Control
.441 Size of load
Read numerically:
.6
no.
no.
yes (numeric or alphameric).
no.
no.
no.
Testable Conditions:. . none.
· 73
Select Reel
or Strip:
toggle switch
Reel power:
2-pos. button
Loading and Unloading
.731 Volumes handled
Storage
Reel: . . . . .
Center roll feed:
Strip: . . . • . .
• 732 Replenishment time: .
.733 Adjustment time: .
.734 Optimum reloading
period: . • • . . .
.8
supply reel locked
out when reading
strips.
energizes reels after tape is loadec.i;
puts reader in
ready status •
Capacity
1,000 feet.
1,000 feet .
?
1 to 2 mins.
no adjustments.
13.3 mins.
ERRORS, CHECKS AND ACTION
Error
Check or
Interlock
Reading:
parity check
Input area overflow:
Invalid code:
none.
parity check as
above.
interlock
parity check.
interlock
wait.
interlock
alarm set.
Exhausted medium:
Imperfect medium:
Timing conflicts
Reader not in ready
status:
Action
set indicator and alarm.
Computer may stop, de·
pending on console
sWitch settings.
alarm set.
412:072.100
IBM 1620
Input-Output
1624 Tape Punch
INPUT-OUTPUT: PAPER TAPE PUNCH
§
072.
.3
EXTERNAL STORAGE
.31
Form of Storage
.1
GENERAL
. 11
Identity:
· 12
Description
. 32
The Model 1624 Tape Punch punches eight-track
chad paper tape in the forward direction at 15 rows
per second. Each row contains a six-bit character
and an odd parity check bit; the eighth track is reserved for the end-of-line (EL) code which signifies
the end of the variable-length tape block.
.321 ,Serial by:
.322 Parallel by:
Tape Punch •
1624.
Numeric or alphameric data may be punched, depending upon the output command. A numeric command causes the contents of sequential positions of
core storage to be punched. An alphameric command causes sequential pairs of core storage positions to be decoded into single alphameric characters. The operation proceeds until a record mark
code is reached in storage; an EL code is then
punched and the operation terminates.
..
.311 Medium: .
.312 Phenomenon: . . .
opaque paper tape •
fully-punched (chad) holes .
PositlOna! Arrangement
row at 10 rows/inch.
8 tracks at standard
spacing.
.324 Track use
Data: .
Redundancy check: .
Timing: . . . . .
Control signals:
Unused: .
Total: .
.325 Row use
Data: •
Redundancy check: .
Timing: . . . • .
Control signals:
Unused: •
......
6.
1.
O.
1 (end-of-line).
O.
8 plus sprocket.
all except 1 per block.
O.
O.
1 (end-of-line).
O.
The tape punch unit is housed below the paper tape
reader in the same cabinet. '
.33
Coding: .•
The sensing of a parity error in the machine code
being decoded stops the computer.
.34
Format Compatibili!,y: . all devices using standard
I-inch paper tape.
Phxsical Dimensions
as in Data Code Table No.
5.
· 13
Availability:..
3 to 4 months.
.35
· 14
First Delivery:
October, 1960.
.351 Overall width:
.352 Length: .•
·2
PHYSICAL FORM
• 21
Drive Mechanism
. 211
.212
• 213
. 214
Drive past the head: .
Reservoirs: •.
Feed drive: • .
Take-up drive: •
• 22
Sensing and Recording Systems
..
•4
CONTROLLER
.41
Identity:
.42
Connection to S},:stem
....
sprocket drive.
none.
none .
clutch.
.421 On-line:
. 221 Recording system: •
.222 Sensing system:
die punch.
none.
.422 Off-line:
• 23
Multiple Copies: .
none.
.43
.24
Arrangement of Heads
Use of station: •
Stacks: •
. Heads/ stack: .
Method of use: .
punching.
1.
8.
1 row at a time.
©
1 inch.
300 feet on take-up reel.
Model 5514 Paper Tape
Reader-Adapter, in 1620
model'l Computer •
Model 5515 Paper Tape
Reader Adapter, in 1620
model 2 Computer.
1; used for paper tape
reader and punch.
none.
Connection to Device
.431 Devices per controller: 1 paper tape reader and 1
paper tape punch.
.432 Restrictions:. . . . •. paper tape reader must be
in system; only 1 device
(reader or punch) may be
operated at a time.
1962 by Auerbach Corporation and BNA Incorporated
12/62
412:072.440
§
IBM 1620
072.
.44
Data Transfer Control
· 441 Size of load
Write numerically:
Write alphamerically: . . . . . .
• 442 Input-output areas:
• 443 Input- output area
access: . . . . .
· 444 Input- output area
lockout: . . . . .
. 445 Table control: ..
. 446 Synchronization:.
.6
PERFORMANCE
· 61
Conditions
II:
.62
1 to N char, each char requiring 2 core storage
positions.
core storage .
each character.
yes.
no.
automatic •
.5
PROGRAM FACILITIES AVAILABLE
. 51
Blocks
• 512 Block demarcation
Output: . . . . . .
1 to N char; N limited by
core storage .
record mark in storage.
Input-Output Operations
· 521
. 522
• 523
· 524
· 525
• 526
Input:. .
Output: .
Stepping:
Skipping:.
Marking: .
Searching:
none.
output 1 block forward .
none.
none.
none.
none.
· 53
Code Translation:
automatic.
• 54
Format Control
.55
progra~.
2 (numeric or alphameric).
no.
no.
no.
Control Operations
• 56
12/62
Component: . . .
Condition: •••.
m. sec per block:
Percentage: . . .
Select code:
Rewind:
Unload: . . .
Testable Conditions: .
none.
processor.
I, II.
1000N/15.
100.
.7
EXTERNAL FACILITIES
· 71
Adjustments:...... none.
· 72
Other Controls
feed tape and punch tape
feed codes .
toggle switch.
Form:
· 73
Loading and Unloading
.731 Volumes handled
Storage:
Capacity: . . . .
. 732 Replenishment time: .
· 733 Adjustment time:
· 734 Optimum reloading
period: . . • . . .
.8
reel.
300 feet on take-up reel;
1, 000 feet on feed reel.
1 to 2 minutes .
none .
40 mins.
ERRORS, CHECKS AND ACTION
Error
Recording:
no.
no.
yes (numeric or alphameric).
no.
no.
no.
Disable: . . . . . .
Request interrupt:
Select format:
Demands on System
Function: .
. 52
Control: . . . .
Format alternatives:
Rearrangement:
Insert spaces: . . .
Section sizes: • . .
Speeds
· 621 Nominal or peak speed: 15 rows/sec.
· 622 Important parameters
Tape speed: ..
1. 5 inch! sec.
1 row/block (end-of-line
• 623 Overhead: . . . .
char).
.624 Effective speeds:
15N/(N+l) row/sec. N=
no. char/block, excluding
record mark •
.63
.511 Size of block:
write numerically.
write alphamerically.
I :
1 to N char; N limited by
core storage.
Parity at 1620:
Output block size:
Invalid code:
Exhausted medium:
Imperfect medium:
Timing conflicts:
Tight tape:
Check or
Interlock
parity check on
die positions
parity check on
internal code of
output char
variable size.
none.
interlock
none.
interlock
interlock
Action
computer halts; alarm
and ind icator set.
computer halts; alarm
and indicator set.
computer halts; alarm.
wait.
punch halts. loses
ready status •
412:073.100
_STANDARD
EDP
•
IBM 1620
REPORTS
Input-Output
Card Reader
INPUT-OUTPUT: CARD READ PUNCH (READER)
§
073.
.1
GENERAL
. 11
Identity:
. 12
Card Read Punch .
(Reader only).
1622.
. 13
AVailability:
. 3 to 4 months .
. 14
First Delivery:
. June, 1961 .
.2
PHYSICAL FORM
.21
Drive Mechanism
. 211 Drive past the head:
.212 Reservoirs:.
Description
The Model 1622 Card Read Punch provides punched
card input and output for the 1620 system. The
reader and punch feed units are separate and functionally independent and contain their own switches,
lights, checking circuits, and buffer storage. Nominal reading speed is 250 cards per minute, and almost all of card reading time is available for internal processing by the computer. Card hoppers have
a capacity of I, 200 cards.
The reader has an 80-character buffer which stores
the data from one card. A read command from the
processor transfers the buffer contents into core
storage in 3.4 milliseconds in the 1620 Modell
(1.7 milliseconds in Model 2), and initiates the refilling of the buffer from the next card. A second
sensing station reads each card and compares its
data to the buffer contents. If an error is detected,
card feeding stops. In addition, the 1620 checks
parity of data received from the buffer; an error
causes an indicator to be set which may be used by
the program to transfer to error-handling subroutines.
Record marks sensed by the reader are inserted into storage; reading is terminated only after a full
card of 80 characters is sensed. If a card requires
storage beyond the end of core storage, the remaining characters are stored starting at location 00000.
.22
Sensing and Recording Systems
. 221 Recording system: .
.222 Sensing system: .
.223 Common system:
. 23
Multiple Copies:
.24
Arrangement of Heads
Use of station:
Stacks:
Heads/stack:
Method of use:
Use of station:
Distance:
Stacks:
Heads/ stack:
Method of use:
.3
EXTERNAL STORAGE
.31
Form of Storage
.311 Medium: . .
. 312 Phenomenon:
.32
rollers .
none.
none in Reader .
brush.
no.
none.
sensing.
1.
80.
1 row at a time.
checking.
I card.
1.
80.
1 row at a time.
standard 80-column cards.
rectangular punched holes •
Positional Arrangement
.321 Serial by:
.322 Parallel by:
12 rows at standard spacing.
80 columns at standard
spacing.
all for data .
all for data .
Cards may be read in either the alphameric or numeric mode. The alphameric mode of reading
causes the 80 card characters to be inserted into 160
consecutive storage locations. When reading numerically, the 80 card columns are stored in 80 consecutive core storage locations. Several special symbol
characters can cause later record mark ambiguities
if data format is not known. In both modes, blank
columns are stored as zeros.
.33
Coding:
as in Data Code Tables Nos.
6 and 7.
.34
Format Compatibility:
all devices using standard
80-column cards.
The Model 1622 Card Read Punch may also be used
with the IDM 7040/7044 systems.
. 35
Ph},:sical Dimensions: .
standard 80 column cards.
©
. 324 Track use: .
. 325 Row use:
1962 by Auerbach Corporation and BNA Incorporated
12/62
412:073.400
§
IBM 1620
073.
.55
.4
CONTROLLER
.41
Identity
.42
1632 Card Read Punch
Adapter in 1620 Modell
Computer.
1633 Card Read Punch
Adapter in 1620 Model 2
Computer.
.43
.56
.432 Restrictions: .
.44
Data Transfer Control
.441 Size of load:
.442 Input-output areas:
.443 Input-output area
access:
.444 Input-output lockout:
•445 Table control:
.446 Synchronization:
80 char.
core storage.
each character.
yes.
no.
automatic.
.5
PROGRAM FACILITIES AVAILABLE
.51
Blocks
. 511 Size of block:
.512 Block demarcation
Input:
.6
PERFORMANCE
.61
Conditions:
.62
Speeds
Code Translation:
. 54
Format Control
Control: . . . .
Format alternatives:
Rearrangement:
Insert spaces:
Section sizes:
Component: . . .
. counter.
m. sec per card:
Percentage: . .
12/62
program.
2 (numeric or alphameric).
no.
no.
no.
...
none.
250 cards/minute.
3.4 m.sec, 1620 Modell.
1. 7 m. sec, 1620 Model 2.
2 point clutch .
250 cards/minute if processing time per card
does not exceed approximately 237 m. sec. t
Demands on System
. 80 char .
automatic.
yes.
• 623 Overhead: . . .
· 624 Effective speeds:
· 63
.521 Input: . . . . . . . . read 80 char from reader
buffer into core storage
and initiate reading of 1
card.
. 522 Output:
see Section :074 .
• 523 Stepping:
none .
. 524 Skipping:.
none.
.525 Marking:.
none.
· 526 Searching:
none.
no.
no.
no.
no.
yes (last card read).
no.
• 621 Nominal or peak
speed: . . . .
• 622 Important parameters
Buffer unload time:
. Processor.
1620 Modell
3.4 t
. 1.4
1620 Model 2
1.7+
0.7
t Normally processing time would include error
test and last card test instructions, thereby decreasing available processing time slightly.
Input-Output Operations
· 53
Testable Conditions
Disabled:
Busy device:
Nearly exhausted:
Busy controller:
Hopper empty:
Stacker full:
Read data transfer
error:
1; no restrictions .
none.
1 Model 1622 Card Read
Punch.
none.
no .
no.
no.
no.
yes (numeric or alphameric).
see format.
Select code:
Connection to Device
.431 Devices per controller:
.52
Disable: .
Request interrupt:
Offset card:
Select stacker: .
Select format:
Connection to System
•421 On-line: .
. 422 Off-line: .
Control Operations
.7
EXTERNAL FACILITIES
. 71
Adjustments: .
.72
Other Controls
Function
Restore ready status:
Remove ready status
and stop reader:
Read 1 card into storage and refill buffer:
none.
Form
Comment
momentary
switch
Start key, does not
actually start
reader •
momentary
switch
Stop key, computer
stops at next
read command.
momentary
switch
Load key.
INPUT·OUTPUT: CARD READ PUNCH (READER)
§
073.
.73
412:073.730
,8
ERRORS, CHECKS AND ACTION
Loading and Unloading
Error
.731 Volumes handled
Storage
Hopper:
Normal Stacker:
Error Stacker:
. 732 Replenishment
time:
.733 Adjustment time:
.734 Optimum reloading
period: . . . . .
Reading:
Capacity
1, 200 cards.
1, 000 cards.
1, 000 cards.
0.5 minute; device does not
need to be stopped.
none.
parity and data compari·
son with check station
Action
send card to
error stacker.
Stop, set alarm;
terminate
ready status.
Input area overflow:
Invalid code:
Exhausted medium:
none.
parity check only.
interloCk
Imperfect medium:
Timing conflicts:
Misfeed or jam:
see reading errors,
interlock
interlock
Dispatch of data:
parity check at 1620
4.8 mins.
©
Check or Interlock
1962 by Auerbach Corporation and BNA Incorporated
set alarm, in"
dicator •
wait.
ready status
removed.
set alarm, in"
dicator.
12/62
412:074.100
•
STANDARD
EDP
•
REPORTS
IBM 1620
Input-Output
Card Punch
INPUT-OUTPUT: CARD READ PUNCH (PUNCH)
§
074.
.1
GENERAL
.11
Identity:. .
.12
Card Read Punch.
(Punch only).
1622.
.13
Avai1abili~:
3 to 4 months.
. 14
First Delivery: .
June, 1961 .
.2
PHYSICAL FORM
.21
Drive Mechanism
.211 Drive past the
head:
. 212 Reservoirs:
Description:
The Model 1622 Card Read Punch provides punched
card input and output for the 1620 system. The
reader and punch feeds are separate and functionally independent and contain their own switches,
lights, checking circuits, and buffer storage. Nominal punching speed is 125 cards per minute, and almost all of card punching time is available for internal processing by the computer. Card hoppers
have a capacity of 1,200 cards.
.22
Record marks in storage are punched in the card;
punching is terminated after SO characters have been
punched. Columns may be left blank by storing the
"numeric blank" character code in the core storage
output area. If the end of the capacity of core storage is reached, the remaining characters for the
card are taken from storage starting at location 00000.
Cards may be punched in the alphameric or numeric
mode. The alphaml':!ric mode of punching causes SO
card characters to be punched from 160 consecutive
locations of storage. When punching numerically,
the SO card characters are taken from SO consecutive storage locations.
The Model 1622 Card Read Punch may also be used
with the mM 7040/7044 systems.
©
Sensing and Recording Systems
.221 Recording system:
.222 Sensing system:
.223 Common system:
die punches.
brush.
no.
.23
Multiele Coeies:
none.
.24
Arrangement of Heads
The punch has an SO-character buffer which stores
the data for one card. A punch command from the
processor transfers data for one card to the buffer
in 3 .4 milliseconds in the 1620 Model 1 Processor
(1. 7 milliseconds in Model 2). The processor
checks parity of the data sent to the punch buffer and
sets a testable indicator if an error occurs. Punching of the card is inhibited (Stop/N -Stop switch set
to Stop).
The data is punched, the punch buffer contents are
parity checked, and the card is read at a checking
station for agreement with the punch buffer contents.
Failure of the parity check or checking station comparison halts the punch (Stop/N-Stop switch set to
Stop) and selects the card to an error select stacker.
rollers .
none.
Use of station:
Stacks:
Heads/stack: .
Method of use:
recording.
Use of station:
Distance:
Stacks:
Head/stack:
Method of use:
checking.
.3
EXTERNAL STORAGE
.31
Form of Storage
. 311 Medium: .
.312 Phenomenon:
.32
1.
SO.
1 row at a time.
.,
1.
80.
1 row at a time.
standard SO-column cards.
rectangular punched holes.
Positional Arrangement
.321 Serial by:
.322 Parallel by:
.324 Track use: .
.325 Row use:
.33
Coding:
.34
Format Compatibility: . . . •.
1962 by Auerbach Corporation and BNA Incorporated
12 rows at standard
spacing.
SO columns at standard
spacing.
all for data.
all for data.
as in Data Code Tables Nos.
6and7.
all devices using standard
SO-column cards.
12/62
IBM 1620
412:074.350
§
074.
. 35
Physical Dimensions:
. . .
.4
CONTROLLER
.41
Identity:
.42
.432 Restrictions: .
.44
standard 80-column cards.
1632 Card Read Punch
Adapter in 1620 Modell
Computer.
1633 Card Read Punch
Adapter in 1620 Model 2
Computer.
Disable: .
Request interrupt:
Offset card:
Select stacker:
Select format:
.56
80 char .
core storage.
each character.
yes.
no .
automatic.
.51
Blocks
PERFORMANCE
.61
Conditions:
.62
Speeds
no.
no .
no.
no.
no.
no.
yes .
...
none .
.621 Nominal or peak
speed: . . . .
· 622 Important parameters
Buffer load time:
125 cards/minute.
3.4.m.sec, 1620 Modell.
1.7 m.sec, 1620 Model 2.
4 point clutch .
125 cards/minute if pro-·
cessing time per card
does not exceed approx
477m.sec. t
· 623 Overhead: . . .
· 624 Effective speeds:
80 char.
counter.
.63
Demands on S},:stem
Component: . . .
.523
.524
. 525
.526
Stepping: .
Skipping: •
Marking:.
Searching:
see Section :073 .
transfe;r 80 char from core
storage to punch buffer and
initiate punching of 1 card.
none.
.none .
none.
none.
.7
• 53
Code Translation:
automatic .
12/62
Testable Conditions
.6
Ineut-Output Operations
. 521 Input:
.522 Output:
no.
no.
no.
no.
yes (numeric or alphameric).
see format .
Disabled:
Busy device:
Nearly exhausted:
Busy controller: .
Hopper empty:
Stacker full:
Write data transfer
error:
1 Model 1622 Card Read
Pt!nch.
none.
PROGRAM FACILITIES AVAILABLE
. 511 Size of block: .
.512 Block demarcation
Output: .
Control Operations
Select code:
1; no restrictions .
none.
.5
.52
.55
Data Transfer Control
. 441 Size of load:
. 442 Input-output areas:
.443 Input-output area
access:
.444 Input-output area
lockout:
.445 Table control:
. 446 Synchroruzation:
program selection of mode .
2 (numeric or alphameric).
no.
columns may be left
blank.
no.
Section sizes:
Connection to Device
. 431 Devices per controller: .
Format Control
Control: .
Format alternatives:
Rearrangement:
Insert spaces:
Connection to S},:stem
.421 On-line: .
.422 Off-line: .
.43
.54
.71
. processor.
1620 Modell 1620 Model 2
3.4t
1.7.t
0.7
0.35.
m. sec per card: .
Percentage: . . .
Adjustments: . . • . • none.
t Usually, processing time would include an error
test instruction, thereby decreasing available
processing time slightly.
.
EXTERNAL FACILITIES
412:074.720
INPUT-OUTPUT: CARD READ PUNCH (PUNCH)
§
074.
. 72
Other Controls
Function
Restore ready status:
Stop punch unit on
1622 error:
Remove ready status
imd stop punch:
.73
.733 Adjustment time:
. 734 Optimum reloading
period: . . . . .
Form
Comment
.8
9.6 mins .
ERRORS, CHECKS AND ACTION
start key; does not
actually start
punch.
Error
Check or Interlock
Action
Data transmission:
parity check at 1620
2- position switch
Stop/N-Stop switch.
Recording:
parity check on buffer
momentary switch
stop key.
set indicator and
alarm.
halt before punching if Stop/NStop switch on
Stop.
send card to error
stacker. Stop,
set alarm if
Stop/N-Stop
switch on Stop.
momentary
switch
check station comparison with buffer
Loading and Unloading
.731 Volumes handled
Storage
Hopper: . . .
Normal stacker:
Error stacker:
. 732 Replenishment
time:
none .
Capacity
I, 200 cards.
I, 000 cards.
I, 000 cards.
0.5 minutes; unit does not
need to be stopped.
©
Output block size:
Invalid code:
Exhausted medium:
none.
parity check only.
interlock
Imperfect medium:
Timing conflicts:
Misfeed or jam:
see recording errors •
interlock
interlock
1962 by Auerbach Corporation and BNA Incorporated
ready status removed.
wait.
ready status removed.
12/62
412:081.100
II
•
STANDARD
EDP
IBM 1620 Modell
REPORTS
Input-Output
I/O Typewriter
INPUT-OUTPUT: CONSOLE I/O TYPEWRITER
§
.24
081.
.1
GENERAL
.11
Identity:.
. 12
Description
I/O Typewriter.
The Console I/O Typewriter is a modified singlecase IDM electric typewriter which stands upon the
console desk. It is usable only with the 1620; no
off-line use is possible. It types output data under
program control, and provides input under program
request or by operator initiation (console Insert
key). Up to 100 characters may be inserted into
storage by operator initiation" and any number of
characters by program request for typewriter data.
The record mark may be inserted in storage by a
typewriter key but is treated only as a data character. The decimal point (period) can be inserted
correctly into storage only with the use of a Read
Alphamerically instruction.
An output command transfers any number of characters to the typewriter from storage, termilUlted
by a record mark. Nominal output speed is 10 characters per second. Both input and output data transfers are parity-checked by the 1620. Incorrect output characters (parity or invalid codes) are marked
specially as they are typed.
Use of station:
Stacks:
Heads/stack: .
Method of use:
printing.
1.
l.
1 character at a time .
Use of station:
Stacks:
Heads/stack:
Method of use:
keyboard input.
1.
44 keys.
1 character at a time.
Range of Symbols
Numerals:
.10
Letters: .
26
Special:
14. *
Alternatives:
FORTRAN set:
Req. COBOL set:
Total:
50.
* including control codes not used as
.3
EXTERNAL STORAGE
.31
Form of Storage
.311 Medium:.
. 13
Availability:
3 to 4 months .
.312 Phenomenon:
. 14
First Delivery:
October, 1960 .
.32
.2
PHYSICAL FORM
. 21
Drive Mechanism
.211 Drive past the
head:
. 212 Reservoirs:
. 22
.223 Common system:
.23
none.
yes.
no.
data.
continuous fanfold stationery.
printing.
character at 10 per inch .
87 print positions .
all for data .
.33
Coding:
.34
Format Compatibility: . none.
engraved hammers.
typewri~er keyboard for
manual input.
no.
.35
PhJ':sical Dimensions
depends on stationery.
4
friction drive.
none .
o-9
A - Z.
Positional Arrangement
.321 Serial by:
. 324 Track use
Data:
. 325 Row use:
Sensing and Recording Systems
.221 Recording system:
.222 Sensing system:
"
.25
Arrangement of Heads
.351 Overall width:
.352 Length:
.353 Maximum margins:
as in Data Code Tables
No.3,4 .
8.875 inches.
no limit.
no limits.
Multiple Copies
.231 Maximum number
Interleaved carbon: .
.233 Types of master
Multilith:
Spirit:
.41
yes.
CONTROLLER
Identity: . . . .
yes.
©
1962 by Auerbach Corporation and BNA Incorporated
-------
no separate controller;
part of 1620 Console.
12/62
412:081.420
§
IBM 1620 MODEL 1
081.
.42
1.
not usable off-line.
. 55
Control Operations
Disable: . . . . .
Request interrupt:
no.
no .
. 56
Testable Conditions:
none.
.. 6
PERFORMANCE
.61
Conditions: ..
. 62
Speeds
Connection to Device
.431 DeviCes per controller: . . .
.432 Restrictions:
•44
Format Control: . . . fixed format; automatic
carriage return at end of
each line .
Connection to System
.421 On-line: .
.422 Off-line:.
.43
.54
1.
none.
Data Transfer Control
.441 Size of load
Input:
none .
/'
· 445
.446
no limit when requested by
program; 100 char when
manually initiated by Insert key.
Output: . . . . . . . any size up to limit of
storage; terminated by
record mark .
Input-butput areas:
core storage .
Input-output area
access:
each character.
Input-output area
lockout:
yes.
Table control: .
no.
Synchronization:
automatic.
.5
PROGRAM FACILITIES AVAILABLE
.7
EXTERNAL FACILITIES
.51
Blocks
.71
Adjustments:
.72
Other Control!!
. 442
. 443
.444
. 511 Size of block:
• 512 Bloc k demarcation
Input:
Output: .
.52
· 522 Output:
· 523 Stepping:
. 524 Skipping:
. 525 Marking:
.526 Searching:
. 624 Effective speeds: . . .
.63
typical typewriter adjustments .
end input and start computer.
momentary key.
R-S key.
Function:
Form:
Name:
input 1 block into core
storage.
output 1 block from core
storage, with automatic
carriage returns.
step 1 or 2 lines at end of
printed line; set byoperator.
none.
none .
none.
Loading and Unloading
.731 Volumes handled: .
.8
depends on feed facilities.
ERRORS, CHECKS AND ACTION
Code Translation: . . . automatic; data stored dependent on mode of operation (numeriC or alphameriC).
l
12/62
m. sec per char or Percentage
100
or 100.
Processor:
.73
10 char/sec for output;
manual typing speed for
input.
same as peak speeds, less
allowance for carriage
returns.
Demands on Sxstem
Component
console Release key or
100th char.
record mark in storage.
Input-Output Operations
. 521 Input: .
. 53
same as load size; see.
.441 above.
.621 Nominal or peal<
speed: . . . .
Error
Check or Interlock
Action
Parity:
check at 1620
Reading:
Input area overflow:
Output block size:
Invalid code:
parity check at 1620
count of 100 max. char
on manual Insert
operation.
any size possible.
check
indicator. alarm;
overprint a bar
over char acter.•
indicator. alarm•
Exhausted medium:
Imperfect medium:
Timing conflicts:
Dispatch of data:
none.
none.
interlock
attach parity bit.
Ir-A-U-ER-BA-C'!.--/-~
special char
printed •
wait.
-
412:101.100
•
II
STANDARD
EDP
IBM 1620
Calcamp Recorder
REPORTS
INPUT-OUTPUT: CALCOMP DIGITAL RECORDER
§
101.
.12
.1
GENERAL
. 11
Identity: . .
. 12
Description
paper (z-axis). The lateral movements of the pen
and paper are 0.1 inch long and result in the following lines being drawn:
Calcomp Digital Recorder.
Model 560-R.
a. + X, - X, + Y, - Y, each 0.1 inch long.
b. the four diagonals, each 0.141 inch long •
The Calcomp Digital Recorder is a point or line
plotter manufa~tured by California Computer Products, Inc., of Downey, California. It operates
via an adapter connected between the 1620 Computer and the 1624 Tape Punch. A switch on the
adapter selects either the plotter or punch. It accepts the numerics 0 to 9 output by a Write Numerically instruction.
The digits 1 to 8 cause independent lateral movements at right angles of the pen and paper, and
digits 0 and 9 cause pen movements to and from the
©
Description (Contd.)
Lines can be drawn at punch speed (15 per second);
the pen can be lifted or dropped at ten characters
per second. Plots are terminated by receipt of a
1620 record mark.
The plotter can be fitted with a roll-chart or sheets
of paper as desired, and has an operational plotting
width of approximately 10 inches. The length of
plot is limited only by the paper supply.
. 13
Availability:. . . . . 3 months (from Calcomp).
1962 by Auerbach Corporation and BNA Incorporated
12/62
412: 111.1 00
•
STANDARD
_EDP
."
IBM 1620
REPORTS
Simultaneous Operations
SIMULTANEOUS OPERATIONS
§
111.
·1
SPECIAL UNITS
• 11
Identity:
.2
CONFIGURATION CONDITIoNs: . . . . . . none.
.3
CLASSES OF OPERATIONS
buffers in 1622 Card Read
Punch unit.
A:
B:
· 12
Description
In a 1620 punched tape system, all operations are
performed sequentially. When a Model 1622 Card
Read Punch unit is added to the system, card reading and punching are overlapped with succeeding
operations. A card input or output operation transfers data from or to the respective card data buffer
and after the card operation is ~nitiated (read card to
buffer, punch card from buffer), program control is
transferred to the next instruction in sequence.
Disc data transfers are not buffered, but seeks may
proceed independently.
©
C:
D:
E:
F:
P:
.4
read or punch paper tape.
read or record on disc
storage.
read card.
punch card.
input or output on Console
I/O Typewriter.
seek in disc storage.
internal processing.
RULES
a + b + e + p = at most 1.
a + b + c + d + e + p = at most 3.
f = at most N, where N = number of disc drives in
system (4 max).
c = at most-l.
d =at most 1.
1962 by Auerbach Corporation and BNA Incorporated
12/62
412: 121.1 01
_SlANDA"
II
REPORTS
EDP
IBM 1620 Modell
Instruction list
INSTRUCTION LIST
§
121.
INSTRUCTION
OPERATION
OP
P
Q
Mnemonic Op.
Arithmetic: Fixed Point
21
11
22
12
23
13
28 t
18 t
29t
19t
P
P
P
P
P
P
P
P
P
P
A
AM
S
SM
M
MM
LD
LDM
D
DM
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
(P) + (Q)---» P.
(P) + Q ~ P.
(P) - (Q)---» P.
(P) - Q ---» P.
(P) X (Q)~ standard area.
(P) X Q ~ standard area.
(Q)---->-- standard area, positioned by P.
Q --...;. standard area, positioned by P.
(standard area) + (Q) positioned by P.
(standard area) + Q positioned by P.
Arithmetic: Floating Point
01*
02*
03*
09*
P
P
P
P
Q
Q
Q
Q
FADD
FSUB
FMUL
FOlV
(P) + (Q)--7P.
(P) - (Q)--7P.
(P) X (Q)---->-- P.
(P)
(Q)~P.
08*
P
Q
FSR
Move mantissa at Q to P, truncating previous low order
mantissa digits.
05*
P
Q
FSL
Move mantissa at Q to P, setting new low order mantissa
digits to zeros.
06t
P
.Q
TFL
07*
P
Q
BTFL
Store next instruction address. Move (Q) to P-1.
Branch to P.
Logic: Comparisons, Branching
24
P
Q
C
14
P
Q
CM
49
42
43
44
45
46
P
-
B
BB
BD
BNF
BNR
BI
-
-
P
P
P
P
Q
Q
Q
Q
Set HIP indicator if (P) > (Q).
Set E/Z indicator if (P) = (Q).
Set neither indicator if (P) < (Q).
Set HIP indicator if (P) > Q.
Set E/Z indicator if (P) = Q.
Set neither indicator if (P) < Q.
Branch to P to obtain next instruction.
Branch to address stored.
Branch to P if digit at Q not zero.
Branch to P if no flag bit at Q.
Branch to P if no record mark at Q.
Branch to P if indicator specified by Q is on.
Q Code
01.
02.
03.
04 .•
06.
07.
09.
11.
12.
13 .
* Special
Indicator
Program switch 1.
Program switch 2.
Program switch 3.
Program switch 4.
Read check
Write check.
Last card.
High/Positive
Equal/Zero.
HighlPositive or Equal/Zero.
Feature, models 1 and 2.
model 1.
t Special Feature,
©
1962 by Auerbach Corporation and BNA Incorporated
12/62
412:121.102
§
IBM 1620 MODEL 1
12f.
INSTRUCTION LIST -CoRld.
INSTRUCTION
OP
P
Q
OPERATION
Mnemonic Op.
Logic: Comparisons, Branching (Contd.)
Q Code
47
BNI
Indicator
14 .
Overflow.
•
Memory buffer register-even check.
16 .
17 .
. Memory buffer register-odd check.
19 .
Any data check.
Same as BI if the specified indicator is off.
Logic: Program Control
27
P
Q
BT
17
P
Q
BTM
32
33
P
P
P
Q
SF
CF
MF
nt
48
H
41
NOP
Store next instruction address. Move (Q) to P-l.
Branch to P.
Store next instruction address. Move Q to P-l.
Branch to P.
Set flag bit at P.
Clear flag bit at P.
Clear flag bit at Q, if any, and copy flag bit to P.
Halt in manual mode.
Advance to next instruction in sequence.
Data Transfers
25
P
P
72t
p
P
P
P
Q
TD
TDM
TF
TFM
TR
TNS
73t
P
Q
TNF
15
26
16
31
Q
Q
Q
Q
Q
Digit at Q~ P, including flag bit.
Digit in Q field of instruction ---? P, including flag bit.
(Q)---7 P, terminated by end of field.
Q ---7 P, terminated by end of field.
(Q)--;' P, terminated by end of record.
Transfer numeric digits of alphameric field starting
at P to numeric field starting at Q, terminated by
flag bit in numeric field. Transfer all flag bits.
Transfer numeric field starting at Q to digit location
of alphameric field starting at P, terminated by flag
bit in numeric field. Zone positions of alphameric
field filled with 7's. Old flag bits erased.
Input-Output
34
K
35
P
Q
DN
36
P
Q
RN
37
P
Q
RA
38
P
Q
WN
39
P
Q
WA
t
12/62
Q
Special Feature, model 1.
Execute typewriter control function specified by Q:
Space, Carriage Return, or Tab.
Output numerically on device selected by Q the contents
of storage starting at P, terminated by end of module
addressed, or end of card in process in standard
numeric manner.
Input numerically from device specified by Q to storage
starting at P, terminated by end of card, EL code on
tape, or release of typewriter.
Input alphamerically (2 storage locations per character)
from device specified by Q to storage starting at P,
terminated by end of .card, EL code on tape, or release
of typewriter.
Output numerically to device specified by Q from storage
starting from P, terminated by record mark for tape
or typewriter, or by 80 chars for a card.
Output alphamerically (2 storage locations per character)
to device specified by Q from storage starting at P,
terminated by record mark for tape or typewriter, or
by 80 chars for a card.
412:131.100
IBM 1620 Modell
Coding Specimen
1620/1710 SPS
CODING SPECIMEN: 1620/1710 SPS
§
131.
.1
CODING SPECIMEN
PRO GRAMI COMP,UTES THE AREA UNDER THE!. CURVE SQRT"3X**2a*ARCS I NEX
LIES BETWEEN 0 ANO 10 THE AREA IS COMPUTED BY SIMPSON5 RULE
IIfHER E X
NUME RICAL INTEGRATION. THE AREA IS E. VAL.UATI:,O US I NG THREE
OIFF EREN T VALUES FOR DEL-TAX. THEY ARE 0.100. 0.050. AND O.O~5.
• THIS
•
•
01732
01732 26 03394 -.J401
01744 26 03423 U3431
01756 26 03441 03448
01768 15 03089 O(')UOI
017Bu 15 02897 UQuUl
01792
•
FOR
STAR'T
15 02765 OOUOI
01804 3, 03449 03510
26 03580 U3458
0,8,6
01828
01840
01852
01864
01876
Ul8SH
0190U
01912
01924
01936
01946
01960
01972
01984
01996
02008
02020
020.32
02044
02056
02068
02080
02092
02104
02116
02128
02140
02152
02164
02176
02188
02200
02212
02224
02236
02248
02260
02272
02284
02296
02308
02320
02332
02344
02356
02368
02380
02392
0,?404
02416
02428
02440
02452
02464
02476
02488
0250U
O?'51Z
02524
02536
02548
02560
23
32
44
32
26
3,
2,
45
47
25
25
25
25
25
25
25
25
25
25
25
25
25
25
3_
39
26
22
3,
47
25
25
2S
25
25
25
25
34
39
26
26
49
2,
22
46
2,
3,
32
22
26
44
26
32
22
47
25
25
25
25
25
25
34
39
03580 03441
00084 UUiJO!,)
01876 llfJU99
00093 ouooo
0358.0 UUU93
ASINE
03449 lJ34~9
035HIJ O,J45U
01B28 03459
02,28 00100
03775 03571
03779 03572
03781 03~7J
03783 U3574
03785 U3575
03787 03576
03789 03577
0379l 03578
0.3793 U,J579
03795 0358U
03739 03435
03743 03436
03745 03437
03747 03438
oouoo
UUI02
03727
03588
03588
03588
02284
03841
03845
03847
03849
0.3851
03653
03855
00000
03799
03628
03654
02332
03646
03595
02320
03595
03581
03581
03646
03628
02332
03674
03589
03674
02572
03883
03887
03889
03891
00100
03448
03441
03603
UOIOO
0,3582
0,3583
03584
03585
03586
03587
03588
U0102
00100
U3640
0366J1
001,)00
03622
0.3654
01300
03654
03582
03893
03895
00000
03859
CONTA
CONTe
ROOT
00000
03628
03627
03623
03680
DQ'JOO
03594
00100
03669
03670
03671
03612
03673
03674
00,02
00100
OORG
TF
TF
TF
TOM
TOM
TOM
TR
TF
M
SF
BNF
SF
TF
TR
A
1732
DEL.TAX.)(.7.TRANSMIT VAL.UE OF INCREMENT
AREA.Z-3
XsuBN,UNIT
51113&1,1 •• SET SW3 OFF
SW2C,1 , 1 •• seT SW2 OFF
SWllrl •••• SET SWl OFF
ASUBN-9.CONST-9 •• TRANSMI T ASuB5 TO ASUSa
~NR
AS I NE. ASUBNC,.
0002
0004
0006
oooa
0010
0012
0014
0016
0018
0020
0022
0024
PSJX.ASUBN
PSIX.XSUi:SN
0026
002e
S4
00.30
*c..2*L,99
0032
0034
93
P'5I)(.93
ASUBN-9, ASUHNt.1
0036
0038
PSlx.AsuaN
0040
0042
SNel CONTA
0044
TO
TO
TO
TO
TO
TO
TO
TO
TO
TO
TO
TO
TO
TO
RCTY
WATY
TF
S
TR
BNCI
TO
TO
TO
TO
TO
TO
TO
RCTY
WATY
TF
TF
0046
0048
0050
POLVC,4B,PSI)C-9
POLYC,52, PS J X-8
POLYC,54,PSlx-7
POLYC,56,PSl)(-6
POLVC,60,PSIX-4
POL.yC,62,PSIX-J
POLVC,64,PSIX-2
POLyC,66. PS I X-I
0052
0054
0056
0058
0060
0062
POLyc..6B,PSIX
0064
POL yc, 12, XSUBN-6
POLVlII6. XSUBN-S
POLVC, J B. XSUSN-4
POLVC,20.XSUBN-3
0066
0068
0070
0072
0074
0016
0018
0080
0082
0084
00e6
008B
POLVC,SB,PSI X-5
POLY
RADCND.UNJT
RADCNO • xSUBN. ,RAD I CANO "
RAOCNO.ZNINES-13,
CONTS
ARGC,42. RAOCNO-6
ARGC,46. RAOCNO-S
ARGC,48. RAOCND-4
ARGC,SO.RAOCND-3
ARGC,S2. RAOCND-Z
ARGC,54,RAOCNO-l
ARGC,56.RADCND
.'-x
0090
0092
0094
0096
0098
0100
0102
0104
0106
0108
0110
0112
OJ 14
0116
0118
0120
0122
0124
0126
o,2e
01JO
0132
0134
01J6
0138
0140
0142
011&4.
0146
0146
0150
ARG
NINE.TWQ9
ODD I NT. ONE ONE
~
*C.2*L
A
ODD I NT-8. TWO
S
RADCNDC,1.0DDINT
eNN
ROOT
A
RADCNDfi.7.000INT
TR
RAOCND-7.RAOCND-6
SF
RADCND-7
S
aODINT-8,NINE
TF
NINE.NINE-l
8NF
RODTC,l *L. TWOC,l
TF
SORT.NINES
SF
RAOCNOC,I
S
SORT,RAOCNOC,6
8NC' CONTC
TO
GENRTC,24, SORT-5
TO
GENRTc,2a.SORT-4
GENRTC,30.SQRT-3
TO
TO
GENRTC,32.SQRT-.2
TO
GENRTC,34.SQRT-l
TO
GENRTL.36.SQRT
ReTV
WATV GENRT
PAGE
Figure 8.
Sample Progrnm Output Listing. Part 1
Copied from mM reference manual, mM 1620 / 1710 SYMBOLIC PROGRAMMING SYSTEM
©
1962 by Auerbach Corporation and BNA Incorporated
12/62
IBM 1620 MODEL 1
412: 131.1 01
§
131 •
•1
CODING SPECIMEN (CONTD.)
025'12 23 03674
02584 32 00065
02596 26 03690
0260H 47 02764
'02620 25 03909
02632 25 03913
02644 25 03915
02656 25 03917
03~BO
CONTe
00000
00094
UOIOO
03681
03662
03683
03684
02668 25 03919 03685
02680 25 03921 03686
02692 25 03923 03687
02704 25 03925 03688
02716 25
02728 25
02740 34
02752 39
02764 49
02776 23
02788 32
02800 26
02812 13
02824 32
02836 26
02848 26
02860 22
02672 15
03927 03689
SQRT.PSI
85
x
0152
0154
0156
0158
0160
0162
TEMPI.94
swl
FU'NCTf., 10. TEMPl-9
FUNCTb14.TEMPI-8
FUNCTb16.TEMPI-7
FUNCT& 18.TEMPP-6
FUNCT{,2Q. TEMP 1-5
0164
0166
0168
0110
0172
0174
0176
0178
0180
FUNCTc..22:, TEMP 1-4
FUNCTc..24.TEMPI-3
FUNCTC,26.TEMPl~2
·FU"!CT lr28, TEMP 1-1
FUNCTc..30. TEMP I
~CTY
r.~,ooo 00102
C3d99 00100
WATY
FuNCT
8
sw2
03441
M
XSUBN,XSUBN
0186
00087
03700
03700
00090
03588
03580
03580
02765
02884 49 02152
02896 49 03004
02906 21 03423
ouooo
SF
TF
87
QI88
TEMP2.96
TEMP2.3.10
0190
0192
0194
0196
0198
0200
0202
0204
0206
0208
0210
0'2920
02932
02944
02956
02968
'02980
02992
03004
03016
U3028
03040
03052
03064
03076
03088
03441
03063
02897
03711
03716
03718
01792
03711
03441
03441
01792
03711
00088
03423
03160
03394
03063
03711
03441
03089
01792
03423
00088
03690
0369P
03959
031J61
03963
03979
03983
000-2
03100
03112
03124
03136
03148
03160
03172
03184
03\1J6
26
16
15
26
26
21
49
21
21
24
47
13
32
21
49
16
26
26
15
49
23
32
26
23
03208 25
03220
03232
03244
03256
03266
03280
03292
03304
03316
03328
03340
03352
03364
03376
12/62
03929 03690
M
SF
TF
SNel
TO
TO
TO
TO
TO
TO
TO
TO
TO
TO
02896 00000
03441
SWI
UOU96
UOU-3
OOI,)UO
00096
03~69
03690
00009
00000
00000
03686
MM
SF
TF
TF
S
TOM
8
SW2
8
A
INIT IAL.I
*
000-4
00009
03434
03394
03718
00000
03690
03716
03680
011 00
-0000
00000
00095
00000
03434
0371B
00;)09
00000
03394
01)000
000!}7
03725
03389
0331J0
2"
U33YI
25
UOOB3
25
U0084
25
25 039a~ 00085
25 03987 OOOl::t6
25 03989 00087
25 03991 00088
34 00000 00102
39 03933 00100
II 01743 000-7
14 01743 -3422
47 01732 01200
48 00000 00000
TF
TFM
TOM
TF
TF
A
ODDVN
B
A
A
C
BNH
MUL.T
MM
SF
A
8
SW3
lNIT tAL.1
*
TFM
TF
TF
TOM
6
M
SF
TF
M
TO
TO
TO
TO
TO
TO
TO
TO
TO
0182
0184'
90
RADCNO.96
PSIX.CONSTc..50
PSIX,TEMPI
Sill 1 c.. I .9
ROOT-14*L.
ODDVN
AQEA, TEMPl-4 •• FOc..FN
ZATION FO~ Fsuaooo
'XSUBN.OE:L.TAX
0212
0214
0216
021b
0220
0222
0224
0226
0228
0230
0232
0234
0236
0238
0240
0242
0244
0246
0248
0250
MUL Tc..11 .4. 10
SW2&1.9
ACCUM,Z
TEMP3. DEL TA)(
TEMP3, TEMP3
AS I NE-3*L
ACCUM.TEMF'1
XSUBN,TEMF'3
XSUBN, NJ NES
ASINE-3*1.
ACCUM
88.
AREA.95
*&6*L
ZATtON FOR FSUI:::lEVEN
MULT&11.2,10
ACCUM.Z
X5UBN. TEMP3
SW3C,1.9
ASINE-3*L.
AREA,DEl. TAX
02~2
02~4
0256
0258
0260
0262
0264
0266
0268
0270
0272
0274
0270
0278
88
TEMPI.97
TEMPI.THREES
OUTPUTC,26. DEL. TAX-5
OUTPUT&2B, DEL TAX-4
OUTPUT(,30,OEL. TAX-3
OuTPuTc..46.83
OUTPUTc..SO.84
OUTPUTc..S2.85
OUTPUTC,54.86
OUTPUTc..S6.87
OuTPuTc..sa.86
OZ80
ReTY
WATY OUTPUT
AM
STARTCd 1 .7. 10
STARTld 1 .X(,21
CM
ENE:
START
0282
0284
0286
0288
0290
0292
0294
0296
0299
H
* AREA oEF
DEL TAX OS
DC
X
DC
DC
AREA
OS
INJTI0NS
03394
03401
03408
0)415
03423
00007
00007
00007
00007
0::100a
FigureS.
Sample Program Output Listing, Part 2
7
7.100000
7.50000
7.25000
8
030~
0305
PAGE
CODING SPECIMEN: 1620/1710 SPS
§
412:131.102
131 .
.1
CODING SPECIMEN (CONTD.)
03434
00011
03441
00007
00007
03448
034'58
03509
03519
03529
03539
03549
00010 00006
00001
00010
Z
XSUBN
UNIT
AsuBN
CONST
00010
00010
00010
00010
03559
03570
00012
00011
03580
03581
00010
00001
L
PSIX
I
7
7.10acaoo
0312
DSB
OS
10.6
I
0315
DC
10, -4337769
0319
DC
I:). 19349939
0.322
0325
0328
DC
10.-44958884
DC
DC
DC
OS
DS
OS
10.87876311
10.-214512362
1 I • I 570795207!1
_12
10
I
DS
7
00007
~AOCND
03601
00013
03616
00015
03622
00006
OS
ZNI NE5 DC
TWO
OS
NINE
OS
TW09
DC
ODOINT OS
ONEONE DC
~aRT
OS
NINES DC
TEMPI
OS
TEMP2 OS
ACCUM DS
TEMP3 OS
THREES DC
OAe
PPL.V
.RG
oAC
GENRT OAC
Fl!NCT OAC
OUTPUT OAC
DENO
03628
00006
00012
00014
000t4
03674
00006
03680
00006
00010
00010
00011
03690
03700
03711
03718
00007
03725
00007
03727
03799
00036
00030
03859
00020
o~e99
00017
00031
03933
017.32
0307
DC
03588
03640
03654
03668
Ita
os
DC
0310
0317
0331
0.3.34
0337
0339
0341
0343
034~
13
15 • 9999999~
0347
0350
035Z
0354
0357
6
6
12.200000090000
14
14. 10000000000001
0359
O~<'Z
6
6.999999
10
10
11
7
7.3333333
36.FOR x*o.OOO. PO~YNOMIA~.O.OOOOOOOOO.
3D.SOUARE ROOT ARGUMENTMO.OOOOOO"
20, SOU ARE ROOTMO.OOOOOI!!'
17. FXXDMO.OOODOOOOO(!l
31 .FOR OE~TAX"O.OOO_ AREA"O.OOOOO!!'
START
0364
0367
036<:.)
0371
0:173
0375
0378
0382
0386
O~e9
0392
0396
PAGE
Figure 8.
3
Sample Program Output Listing. Part 3
©
1962 by Auerbach Carporatian and BNA Incorporated
12/62
412:131.200
IBM 1620 MODEL 1
131.
§
.2
COnING SHEET
ID~lL
1620/1710 Symbolic Programming System
Coding Sheet
W
Program: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __
001.: _ _ _ _ _ _ __
Routine: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __
Programmer: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __
Line
3
Lobel
5 6
15 16
25
20
3.
35
••
50
'5
I
.~
•• 3 •
•••
I
I
tI 5 •
~
6JJl
I
.0.• 1 •
~
I
•
Q ,
•
1.0 •
I
I •
1 2 0
1 , 0
1 • 0
I ,
0
1 ,
0
..
I 7 0
I I
0
I ,
0
I
2 0 0
,
12/62
of - - - -
Operands & Remarks
pperation
11 12
Page No.
..
55
60
65
7.
75
412:132.100
.STAI
t-:I
t-:I
IB,.,
I< y
)
'!:.I
1. output only.
2. record mark symbol on input; stop code on output.
3.
symbol is output for invalid character with correct parity.
*
©
1962 by Auerbach Corporation and BNA Incorporated
12/62
412:145.100
.SIANDARD
IIE~~
IBM 1620
Data Code Table
Paper Tape
DATA CODE TABLE NO. 5
§
.23
145.
.1
USE OF CODE.
.2
STRUCTURE OF CODE
.21
Character Size:
.22
paper tape input-output .
6 data tracks + 1 odd
parity track + 1 track for
end-of-line symbol (EOL).
Character Structure
.221 More significant
pattern:
.222 Less significant
pattern:
2 zone bits; X = 32,
0=16.
4 numeric bits; 8, 4, 2, 1.
Character Codes
LESS
SIGNIFICANT
PATTERN
MORE SIGNIFICANT PATTERN
0
0
16
32
48
3/
0
- y
1
1
1
/
J,
2
2
S
K, 2
-
+ ?/
A
B
3
3
T
L, 3
C
4
4
U
M,4
-
D
5
5
V
N, 5
E
6
6
W
0, 6
F
7
7
X
P,
G
8
8
Y
Q,
9
9
Z
R,
7
8
9
H
I
:I: ~/
10
=
11
12
-
@
Y
,
$
(
*
)
13
14
15
§./
Notes: 1. 1 is minus one, etc.
2. :f: is record mark symbol; can be punched
only in Dump Numeric mode when tape
punched by computer. Becomes EOL
(End of Line) char. otherwise (single
punch in EOL track).
3. this code (blank or space) contains a
punch in parity track.
4. minus or hyphen in alphameric mode;
minus zero in numeric mode.
5. J-R and -1 to -9 interpreted as J-R in
alphameric mode. -1 to -9 stored as
-1 to -9, numerical mode.
6. Tape feed code.
7. minus zero in numeric mode.
8. code for numeric blank in numeric mode.
©
1962 by Auerbach Corporation and BNA Incorporated
12/62
412:146.100
.STANDIID
II
EDP
IBM 1620
REPORTS
Data Code Table
Card Code (Numeric)
OAT A CODE TABLE NO. 6
§
146.
.23
.1
USE OF CODE: .. card input-output (numeric mode).
.2
STRUCTURE· OF CODE
.21
Character Size: . . 1 column per character.
Character Codes
OVERPUNCH
UNDERPUNCH
None
None
+
12
11
0
1/
0
12
11
0
1
+
0
1
-
0
-...
1
-
2
2
3
3
.3
4
4
-4
5
5
6
6
7
7
8
8
9
9
8-2
2
-
5
-
6
7
-
8
-
9
of 3/
8-3
8-4
'!:/
8-5
8-6
8-7
Notes:
©
1. on input only.
2. read in as "numeric blank"; causes blank
column on output.
3. of is record mark symbol.
1962 by Auerbach Corporation and BNA Incorporated
12/62
412: 147.100
•
STANDARD
EDP
•
RfPORTS
IBM 1620
Data Code Table
Card Code (Alphameric)
DATA CODE TAI3LE NO.7
§
147.
.23
.1
USE OF CODE: .. card input-output (alphameric).
.2
STRUCTURE OF CODE
.21
Character Size: •. 1 column per character.
Character Codes
OVERPUNCH
UNDERPUNCH
None
None
Blank
12
11
+
-
0
12
11
+
0
0
+
0
4/
-
0
1
r
A
J. 1
2
2
B
K, 2
-
/
S
3
3
C
L, 3
4
4
D
M,4
U
5
E
N,
V
5
-
5
-
6
6
F
0, 6
7
7
G
P, 7
8
8
H
Q, 8
9
9
I
R, 9
-
-
8-2
T
W
X
Y
Z
of 1/
8-3
=
8-4
@
)
$
,
*
(
8-5
8-6
8-7
Notes:
1. of is record mark symbol.
2. i is minus one, etc.
+
3. 0 is positive zero.
4. input only.
©
1962 by Auerbach Corporation and BNA Incorporated
12/62
412: 151.100
.STAND.lRO
II
EDP
IBM 1620
REPORTS
P. O. Facilities
PROBLEM ORIENTED FACILITIES
§
151.
.15
.1
lITILITY ROlITINES
. 11
Simulators of Other Computers
1620 5-Channel Tape Translation Program
Reference: . .
File No. 1.6.014 .
Date available:
currently available.
Description:
Program reads 5-track punched tape, translates
data to legitimate 1620 characters, and punches
characters on tape.
IBM 650.
Reference:
File Nos. 2.0.004 (card)
and 2.0.005 (tape).
currently available.
Date available:
Description:
The simulator program 2.0.004 utilizes card input-output and simulates a 650 card system. The
storage required is as follows:
To Simulate:
. 12
Requires:
1,000 word basic 650
20,000 digit 1620 with
Automatic Divide
2,000 word basic 650
40,000 digit 1620 with
Automatic Divide
4,000 word basic 650
60,000 digit 1620 with
Automatic Divide
File Maintenance: . . . none.
. 17
Other Library Routines
.171 Floating point function subroutines (macros)
Note: data given is for 8-digit fixed length operands.
Routines are available for variable length
operands.
Macro: . . .
FSQR.
Function: • . . .
square root.
Time, /.L sec
1620 model 1:
120,000.
1620 model 2:
30,000.
Storage required, with
automatic divide: .
579 digits.
Macro: . . .
Function: . .
Time, /.L sec
1620 model 1:
1620 model 2:
Storage required, with
automatic divide: .
The tape simulator program is very similar to
the card simulator program. Either program allows execution of 650 programs on the 1620 without reprogramming.
Macro: . . .
Function: . .
Time, /.L sec
1620 model 1:
1620 model 2:
Storage required, with
automatic divide: .
Simulation by Other Computers
Reference:
File No. 171O-SI-002.
Date available:
currently available.
Description:
This program provides simulation of the IBM 1710
Control System, which uses the 1620 Processor,
on an IBM 7090.
. 14
.16
Simple modifications are given to simulate a 650
card system by means of the 1620 punched tape
and/or typewriter input-output. The Simulator
program occupies storage up to location 09021.
Execution time on the 1620 Model 1 is approximately 3 . 5 times as long as a very well optimized
program run in the 650.
IBM 7090.
.13
Data Transcription
Data Sorting and
Merging: . . .
none.
Report Writing:
none.
©
FSIN.
sine.
150,000.
37,500.
843.
FCOS.
cosine.
155,000.
39,000.
843.
Macro: . . .
Function: ..
Time, /.L sec
1620 model 1:
1620 model 2:
Storage required, with
automatic divide: .
FATN.
arctangent.
Macro: . . .
Function: ..
Time, /l sec
1620 model 1:
1620 model 2:
Storage required, with
automatic divide: . . .
FEX.
exponential, natural.
...
1962 by Auerbach Corporation and BNA Incorporated
260,000.
65,000.
989.
160,000 (positive power).
40,000 (positive power).
740 .
12/62
412:151.171
§
IBM 1620
151.
.171 Floating point function subroutines (macros) (cont'd)
Macro: . . . • .
Function: . . . .
Time, Il sec.
1620 model 1:
1620 model 2: . . .
Storage required, with
automatic divide: . . .
FEXT.
exponential, base 10.
Macro: . . . . .
Function: • . . .
Time, /J sec
1620 model 1:
1620 model 2:
Storage required, with
automatic divide: .
FLN.
logarithm, natural.
Macro: . . .
Function: ..
Time, /J sec
1620 model 1:
1620 model 2: . . .
Storage required, with
automatic divide: . . .
FLOG.
logarithm, base 10.
145,000 (positive power).
36,300 (positive power).
740.
290,000.
72,500.
842.
305,000.
76,000.
842.
.172 Matrix Inversion (Tape)
Reference: • . . . . . File No.5. 0 . 006.
Availability: . . • . . currently available.
Description:
This program will invert any non-singular square
matrix of size 22 x 22 or smaller. Provision is
made for re -inversion to check accuracy. Output
is on typewriter. The program is written in
FORTRAN language.
.173 Complex FORTRAN (Tape)
Reference: . . . . . . ·F. H. Maskiell
The Pennsylvania Transformer Division
McGraw-Edison Co.
Box 330
Canonsburg, Pa.
Availability:
currently available; IDM
File No. 6.0.008.
Description:
revision to FORTRAN
processor allows addition,
subtraction, multiplication, and division of complex variables .
. 174 SPS to FORTRAN
conversion:. • .
2 programs provide for
conversion of an SPS
object program into form
suitable for inclusion in
FORTRAN subroutine
library .
. 175 Format Control Subroutines for 1620 Card
FORTRAN
Reference: . . . . . . W. M. Fleischman
Worthington Corporation
401 Worthington Ave.
Harrison, New Jersey.
Availability: . . . . . currently available; IDM
File No. 1.6.017.
Description:
These subroutines permit the use of both fixed
length, variable point format, the standard FORTRAN print routine; and variable length, fixed
point format. Modes may be interchanged within
a single program.
12/62
.176 Interpretive systems
Interpretive Programming System (Tape, Card)
Reference: .
File No. 2.0.001,
No. 2.0.002.
Availability: . . . . currently available.
Description:
IPS is an interpretive programming system for
the 1620. The one-address language includes
the commands of the Intercom System, widely
used on the Bendix G-15.
An Interpretive System for Performing Operations
With Complex Numbers (Tape)
Reference: . . . . W. D. Glauz and J. O.
Hancock
School of Aeronautical and
Engineering Sciences
Purdue University
Lafayette, Indiana.
Availability:. . . . currently available; mM
File No.2. 0.003.
Description:
The program performs operations interpretively with complex numbers in floating point
form. Program uses SPS language.
"
.177 Other
Tracing, Dumping,
Debugging: . . . . . 10 programs.
Linear Programming,
Scheduling, etc.: .. 17 programs.
Industry applications: 27 programs.
Mathematical routines: 20 programs.
Graph plotting: . . . . 4 programs provide off -line
and on -line graph plotting
with annotations.
Fixed-point/iloatingpoint conversion:
1 program for each
conversion.
"Editing: . . . . . . . . 4 programs, which allow
automatic zero suppression and handle all alphameric characters. Floating dollar is not handled.
Also 2 programs handle
format when FORTRAN
typeouts occur.
.2
PROBLEM ORIENTED
LANGUAGES:
none.
.3
AUTOMAP
.31
Identity:
AUTOMAP.
Automatic Machining
Program.
.32
Origin and Maintainer:
IDM.
.33
Reference: . . . . . . . IDM Application Program
Bulletin H20-8097, dated
1962.
Machine Tool
Controllers: . . . . . 'l
.34
PRClBLL::M ORIENTED FACILITIES
§
412:151.350
151.
.35
.37
Description
AUTOMAP is a general-purpose program for the numerical control of machine tools. It consists of
three phases of computation:
a. Translation from English-type statements,
which describe: the outline of the part; holes
within the part; control of the cutter location;
control of the coolant; and tolerances, cutter
radius, etc.
·b. Computation on the output of the first phase,
which produces the path of the tool center during the actual cutting.
c. A post-processing phase which produces information for direct use by the particular machine
tool controller to be used.
The language can define circles, lines and angles,
and points. The cutter can be commanded to follow
pre-defined circular curves or to follow a path defined iIi the statement itself. Detailed information
must be obtained from ruM.
AUTOMAP Instruction List (Cont'd)
TLRAD:
ONKUL:
OFKUL:
FEDRAT:
GO (TO): .
DLTA: . . . . . .
axis.
INDR, VECTOR: .
GO LFT; GO RGT: .
PARLEL: . . • . . .
XSMALL, XLARGE: .
ATANGL:
LINE:
. 36
Availability: . . . . . . '!
POINT:
CIRCLE:
. 37
AUTOMAP Instruction List
FWD: ..
A complete list of instruction statements is not
available, but· a number of the representative ones
are listed below.
TANTO:
GO PAST:
FINI: . . .
FROM: .
TOLER:
specifies cutter radius.
coolant on.
coolant off.
specified feed rate during
cutting.
move cutter along a path
specified by additional information in the
statement.
cutter movement along Z
specifies a direction of
movement.
move to left or right in
specified path.
parallel relation involved.
pick condition involved with
smaller (or larger) value
of x.
at an angle of.
line passing through
specified points.
specified point .
specifies circle radius and
center .
move in same general
direction.
tangency relation.
pass by.
all parts are completed.
specifies starting point with
respect to an origin.
specifies tolerance on all
cuts, including non-linear
cuts.
©
1962 by Auerbach Corporatian and BNA Incorporated
12/62
412:161.100
•
II
STANDARD
EDP
IBM 1620
Process Oriented Language
FORTRAN I
REPORTS
PROCESS ORI ENTED LANGUAGE: FORTRAN I
.14
§161.
Description (Contd.)
Restrictions: (Contd.)
.1
GENERAL
• 11
Identity:
liM 1620 FORTRAN •
"FORTRAN I."
.12
Origin:
liM General Products
Division, San Jose,
California.
• 13
Reference:
liM Publication C26-5619-0 .
.14
Description
IBM: 1620 FORTRAN is basically the FORTRAN I
language. FORTRAN II is now available for the
1620 (see Sections :162 and :184), but the FORTRAN
II translator requires a 1622 Card Read Punch and at
least 40,000 positions of core storage. Therefore,
the earlier and considerably less powerful
"FORTRAN I" language, which can be compiled on
the minimum 1620 configuration using either punched
tape or card input-output, will continue to be widely
used.
The FORTRAN I system does not take full advantage
of the 1620's variable word length capability. All
integer items are represented internally by fourdigit fields, and all floating point items by ten-digit
fields (eight digits for the fixed pOint part and two
for the exponent).
The initial versions of IBM: 1620 FORTRAN, released
in December 1960 and April 1961, did not permit use
of the FORMAT statement; instead, a fixed format
was used for object program input and output data.
The FORMAT statement is available in the October
1961 versions, described here, though there is still
considerably less input-output flexibility than in
FORTRAN II.
Probably the most significant limitation of the
FORTRAN I system is its inability to compile subroutines and functions written in the FORTRAN
language and to combine separately-compiled subprograms at execution time. Other restrictions and
extensions of the liM 1620 FORTRAN I language relative to liM 709/7090 FORTRAN II are summarized
below.
Restrictions :
(1) The following statements are not permitted:
ASSIGN
Assigned GO TO
SENSE LIGHT i
IF (SENSE LIGHT i)
IF ACCUMULATOR OVERFLOW
IF DIVIDE CHECK
IF QUOTIENT OVERFLOW
READ INPUT TAPE
WRITE OUTPUT TAPE
READ TAPE
WRITE TAPE
READ DRUM
WRITE DRUM
END FILE
REWIND
BACKSPACE
(2) Arrays are limited to two dimensions.
(3) Double precision, complex, and logical operations are not permitted.
(4) Names are limited to five characters.
(5) Values of integer constants and variables cannot
exceed + 9999.
(6) Subscript expressions are limited
(VARIABLE ± CONSTANT).
~o
the form
(7) Statement length is limited to 66 characters;
continuations are not permitted.
(8) Alphameric information can be handled only in
the form of Hollerith items (FORMAT specification wH), which cannot be named or manipulated in storage; the Aw specification is not
available.
(9) Implied 00 loops in input-output statements are
not permitted.
(10) Arithmetic statement functions are not permitted.
(11) Numeric displays in connection with PAUSE and
STOP statements are not permitted.
(12) Statement numbers may not be greater than
9999.
(13) Symbolic coding cannot be incorporated into the
FORTRAN source program.
SUBROUTINE
FUNcrION
CALL
RETURN
COMMON
EQUIVALENCE
FREQUENCY
Extensions:
(1) The ACCEPT TAPE and PUNCH TAPE statements
provide punched tape input and output.
©
1962 by Auerbach Corporation and BNA Incorporated
12/62
412:161.140
IBM 1620
§l61.
.14
: • 242 Designators
Procedures
Statement label:
Description (Contd.)
unsigned integer, 4 digits
max.
Function name:.
Extensions (Contd.)
same as variable being defined.
Data
Integer variables:.
Real variables: •.
Equipment
Card:
(2) The ACCEPT and TYPE (or PRINT) statements
provide typewriter input and output.
(3) Floating point data values from 10- 99 to 10+99
can be accommodated.
initial I,}, K, L, M, N.
any other initial letter.
implied by verbs READ,
PUNCH.
Paper tape:
. 15
Publication Dates
IBM 1620 FORTRAN for
Paper Tape: • . • . . . . December, 1960.
IBM 1620 FORTRAN for
Cards: • . . • . . •
IBM 1620 FORTRAN with
FORMAT for Paper Tape: .•
IDM 1620 FORTRAN with
FORMAT for Cards:
.
IBM 1620 FORTRAN
Reference Manual,
C26-5619-0: • • •
••
Typewriter
input: ••
Typewriter
output: .
April, 1961.
October, 1961.
PROGRAM STRUCTURE
.21
Divisions
Procedure statements: .
Data statements: •
May, 1962.
. algebraic formulae.
comparisons and
jumps.
input and output.
• FORMAT: describes
the .layout, size,
scaling and code of
input- output data.
DIMENSION: describes the elements
in each dimension of
an array or s!'!t or
arrays.
statements.
characters; blanks are
ignored.
Structure of Data Names
· 251 Qualified names: •
• 252 Subscripts
Number per item:.
Appllcable to:
Class may be
Special index
variable: •
Any variable:
Literal:
Expression:
Form may be
Integer only:
Signed: ••
Truncated
fraction:
Rounded
fraction:
• 253 Synonyms:.
• 26
Data Entities
Arrays:
Items:
Hollerith item:.
• 24
Names
.241 Simple name formation
Alphabet: •
•
Size:.
•
•
• .'
Avoid key words:
Formation rule: .
12/62
floating point or integer
variables.
floating point variables or
constants.
integer variables or constants.
Hollerith item.
alphameric item that can
only be used for oupput.
A to Z, 0 to 9.
1 to 5 char.
yes.
first char must be a letter.
. none .
.0 to 2.
· all variables.
.no.
· only integers.
· yes,
· at most N + C; where C is a
literal. · C and N.
. no.
.no.
. no.
. none.
Number of Names
.261 All entities: •
.23
implied by verbs PRINT,
TYPE.
C in column 1.
key wOilCd DIMENSION.
October, 1961.
Procedure Entities
Program:
Statement:
implied by verb ACCEPT.
Comments:
Translator control:
• 25
.2
'. 22
implied by verbs ACCEPT .
TAPE, PUNCH TAPE •
• 262 Procedures:
.263 Data
Files: .
Items: •
Data levels:
· 264 Equipment
Card reader:.
Card punch:
Paper tape reader:
Paper tape punch: .
Typewriter:.
•
• symbol table holds up to 238
entries with 20,000 stoJ"- •
age positiOns and 2,000
entries with 40,000 or
60,000 positions. An entry
is required for each name,
fixed or floating pOint number, and statement label
in the source program.
• see • 261, above.
. limited by hardware;
• see . 261, above.
.1.
: 1.
.1.
.1.
• 1.
.1.
PROCESS ORIENTED LANGUAGE: FORTRAN I
§161
. 27
• 346 Choice of external code:
.347 Possible external codes:
Region of Meaning
of Names:
• all names are universal with
respect to the program in
which they are defined, but
are not available to separately-compiled programs;
i. e., no COMMON statement.
.3
DATA DESCRIPTION FACILITIES
.31
Methods of Direct Data Description
. 311
.312
.313
.314
.315
Concise item picture: .
List by kind:
Qualify by adjective:
Qualify by phrase:
Qualify bv code:
• 316
. 317
• 318
.319
Hierarchy by list:
Level by indenting: .
Level by coding:
Others
Array size:
Four-digit integer:
Four-digit integers, 5:
Floating points items;
• 32
Files and Reels: •
• 33
Records and mocks
.331
. 332
• 333
.334
Variable record size:
Variable block size: .•
Record size range:
Block size range:
.335 Choice of record size: •
.336 Choice of block size:.
· 337 Sequence control: •
.338 In-out error control:
• 339 Blocking control:
• 34
412: 161.270
FORMAT statement.
no.
no.
no.
yes; first letter of
name •
no.
no.
no.
DIMENSION (4,7).
FORMAT (14).
FORMAT (514).
FORMAT (F8. 3, EW.
4) for +999.999 and
+.9999E+99.
own coding•
Logical: • • •
• 343 Choice of external radix:
. 344 Possible radices:.
.345 Justification:.
Range
Integer numeric:
Floating point numeric:
.349 Sign provision:
• 35
• 351 Constants
Possible sizes
Integer: •
Floating point:
Alphameric
(Hollerith); •
Subscriptable; •
Sign provision:
.352 Literals:.
.353 Figuratives:
• 354 Conditional variables:
yes.
no.
yes.
no.
no.
no; Hollerith constants
only.
.
no.
none.
decimal only•
integers automatic
right justified.
©
±9999.
10- 99 to 10+ 99 .
1 to 49 characters.
no.
optional.
same as constants.
own coding; e. g. ,
TEN = 10.0.
computed GO TO.
Special Description Facilities
.362 Re-definition: •
• 363 Table description
Subscription:
Multi-subscripts:
Level of item: .
.364 Other subscriptable
entities:
.4
OPERATION REPERTOIRE
.41
Formulae
by mUltiple reference
to single FORMAT
statement.
none.
yes.
2 max.
variables.
none.
.411 Operator list
+:
*.
I:
by name •
• fixed; 4 digits.
• fixed; 8 digit fixed
pOint part, 2 digit
exponent.
• optional.
Data Values
.361 Duplicate format: •
• preset variable.
• preset variable •
• no limit.
• maximum of 80 char
for cards, 87 char
for punched tape and
typewriter.
· input- output and
FORMAT statements.
• input-output and FORMAT statements.
· 1 logical record per
input or output statement.
• automatic.
• none; 1 or more blocks
per logical record.
fixed.
none.
.348 Internal item size
· 36
Data Items
• 341 Designation of class:
• 342 Possible classes
Integer:
•
Fixed point:
Floating point:
•
Double precision: •
Complex:
Alphameric: •.•
.348 Internal item size
Variable size:
Designation:
none •.
standard IBM 1620
punched tape and
card codes; octal or
binary codes not permitted.
** :
= ••
LOG ():
SIN ( ):
COS ():.
EXP ():
SQRT ():
ATAN ():
.412 Operands allowed
Classes:
Mixed scaling:
Mixed classes:
Mixed radices:
Literals: ••
1962 by Auerbach Corporation and BNA Incorporated
addition; also unary.
subtraction; also unary.
multiplication.
division.
exponentiation.
is replaced by.
natural log.
sine.
cosine.
exponential.
square root.
arctangent.
numeric only.
yes •
only in exponentiation
and functions.
no.
yes.
12/62
412:161.413
IBM 1620
§161
· 45
• 413 Statement structure
Parentheses
a - b - c means:
a + b x c means:
a .;. b .;. c means:
abC means:
Open: • . . . •
Close: . • • • •
Advance to next record: .
(a-b) - c.
a + (b x c).
(a';' b)';' c.
Step back a record:
Set restart point:
Restart:
Start new reel:
Start new block:
illegal; parentheses
must be used.
Size limit: . . . • .
Multi-results: • •
. 414 Rounding of results:
66 char; statements cannot
be continued from one
line to another.
no.
truncation of integers at
each step in expression.
floating
.415 Special cases - fixed
x = -x: . . . . • K =-K
X = -X.
.K=K+1
X=X+l.
x = x + 1:
x = 4.7y: ••• • K = 47*K/10
X = 4.7 * Y.
x = 5x107 + y2: · too large
X = 5. E7 + Y**2.
.416 Typical
examples:. X = (-BtSQRT(B*B-4. 0*A*C»/(2. O*A)
Search on key: .
Rewind:
Unload:
.46
.421 Matrix operations: •
. 422 Logical operations: •
.423 Scanning: . . • • •
.464 Accept option:
Object Program Errors
Error
Overflow:
none.
none.
none.
Discovery
automatic
In-out:
Invalid data:
. none.
.43
Other Computation:
.44
Data Movement and Format
.441
. 442
. 443
.444
.445
Data copy example:
Levels possible: .
Multiple results: .
Missing operands:
Size of operands
Exact match:
Alignment rule
Numbers:.
• 446
.
Alpha:
Filler rule
Numbers:
Alpha:
Truncating rule
Numbers:
Alpha:
Variable size
destination:
Editing possible
Change class:
Change radix:
Insert editing symbols
Actual point:
Suppress zeros: •
Insert: .
Float:
Special moves:
Code translation:
Character manipulation:
.
.447
. 448
.449
12/62
.
.
TYPE or PRINT uses
on-line typewriter.
same as log.
TYPE message and
PAUSE.
use SENSE SWITCH.
.462 Messages to operator:
· 463 Offer options:
Operations on Arrays
hardware
checks
automatic
stop or set indicator.
type error code and
continue.
PROCEDURE SEQUENCE CONTROL
.51
Jumps
.511 Destinations allowed: .
• 512 Unconditional jump:
• 513 Switch:
implied, except for
alpha or input- output.
.514 Setting a switch:
. .
statement .
Go to N•
Go to (35, 47, 18), 1.
(Note that Assigned
GO TO is not
available. )
1= 2; causes branch
to statement 47.
Go to (35, 47, 18), I.
.
.515 Switch on data:
right justified; normalized for floating
.52 Conditional Procedures
point.
left justified.
.521 Designators
Condition:
zeros.
Procedure:
blanks.
.522 Simple conditions:
IF.
.523 Conditional relations:
no .
yes.
no.
automatic.
automatic.
automatic point.
- sign only.
none.
automatic .
none.
Special Actions
type error code and
continue .
.5
Y= X.
items.
none.
not possible.
truncate at left.
truncate at right.
own coding•
own coding.
READ, WRITE,
PUNCH, PRINT,
ACCEPT.
not possible.
none.
none.
own coding.
implied in each inputoutput statement.
none,.
none.
none •
Operating Communication
.461 'Log of progress:
· 47
. 42
File Manipulation
.524 Variable conditions:
.525 Compound conditionals: •
.528 Typical example:.
.
implied.
expression or variable
against zero.
IF (A) nl, n2, n3: if
value of expression
A is less than, equal
to, or greater than
zero, go to statement nl, n2, or n3,
respectively.
always zero.
no •
IF (X**2.0 - 3.0) 29,
37,18; go to 29, 37 or
18 if x 2 - 3 is respectively less than,
equal to or greater
than zero.
PROCESS ORIENTED LANGUAGE: FORTRAN I
§161
.75
.53
Subroutines:
not permitted.
• 54
Function Definition
by Procedure:
not permitted.
.55
.56
Operand Definition
b}': Procedure:
. .
none.
EXTENSION OF THE
LANGUAGE: •
.7
LIBRARY FACILITIES
.71
Identit}'::
.72
.751 Insertion of new item:
• 752 Language of new item:
.753 Method of call:
.76
.562 Control by count:
• 563 Control by step
Parameter
Special index:
Any variable: .
Step:
Criteria:
Multiple parameters:
.564 Control by condition:
. 565 Control by list:
• 566 Nesting limit:
.567 Jump out allowed:
• 568 Control variable
exit status:
.
current place to
named end; e. g. ,
00 173 I = 1, N, 2.
none.
no.
integer only.
positive integer.
greater than.
no.
no.
no •
about 90.
yes.
.73
Storage Form:
• 74
,varieties of Contents:
separate run .
SPS o+", machine .
functions by name in
procedures; arithmetic and input-output routines are inserted automatically.
T}':pes of Routine
.761 Open routines exist:
. 762 Closed routines exist:
.763 Open-closed is variable:
no •
yes.
no.
.8
TRANSLATOR CONTROL
.81
Transfer to
Another Language:
no •
Optimizing Information
Statements:
none.
.83
Translator Environment:
no.
.84
Target Computer
Environment:
no •
.82
.85
Program Documentation
Control: •
-.
no •
available.
can write new function
in library (in SPS or
.9
machine language).
.91
FORTRAN subroutine
library.
Kinds of Libraries
.721 Fixed master: •
• 722 Expandable master:
Mechanism
none •
Loop Control
.561 Designation of loop
Single procedure:
First and last
procedures:
.6
412: 161.530
TARGET COMPUTER ALLOCATION CONTROL
Choice of
Storage Level:
none.
.92
Address Allocation:
none •
• 93
Arrangement of Items in
Word in Unpacked Form:
no.
yes.
paper tape or punched .94
cards.
subroutines (25 standard). functions (6
standard, up to 19
user- defined).
©
.95
1 item per field is
standard for
numerics •
Assi~ment
of InputOutput Devices:
InEut-OutEut Areas:
1962 by Auerbach Corporation and BNA Incorporated
implied by input-output verbs •
automatic; but overlapping operations
are not possible.
12/62
, I
412: 162.100
•
STANDARD
EDP
•
IBM 1620
REPORTS
Process Oriented Language
FORTRAN II
PROCESS ORI ENTED LANGUAGE: FORTRAN II
§
Restrictions:
162.
(1) The following statements are not permitted:
.1
GENERAL
.11
Identity:
mM 1620 FORTRAN II
.12
Origin: .
mM General Products
Division, San Jose,
California.
.13
Reference: .
mM Publication J26-5602-1.
.14
Description
ASSIGN
Assigned GO TO
SENSE LIGHT i
IF (SENSE LIGHT i)
IF ACCUMULATOR OVERFLOW
IF DIVIDE CHECK
IF QUOTIENT OVERFLOW
READ INPUT TAPE
WRITE OUTPUT TAPE
READ TAPE
WRITE TAPE
READ DRUM
WRITE DRUM
END FILE
REWIND
BACKSPACE
Compilation of source programs written in mM
1620 FORTRAN II requires a 1620 system with a
1622 Card Read Punch, at least 40,000 positions of
core storage, Automatic Divide and Indirect Addressing. The FORTRAN II system offers the following major advantages over the earlier mM 1620
FORTRAN system described in Sections :161 and
:183.
(2) Double preCision, complex, and logical operations are not permitted.
(1) Subroutines and functions can be written in the
FORTRAN language and compiled independently
of the main program whith which they will be
used.
(3) Statement length is limited to 330 characters
on up to five cards.
(4) Only seven standard library functions are provided (see Paragraph .411).
(2) The COMMON and EQUIVALENCE statements
can be used to reduce object program data
storage space requirements.
(5) Symbolic coding cannot be incorporated into
the FORTRAN source program.
(3) More flexibility is permitted in the FORMAT
statement, including the use of implied DO
loops.
Extensions:
(1) Internal item size is preset variable: 4 to 10
digits for all integers and 4 to 30 digits (including a 2-digit exponent) for all floating pOint
items within a program.
(4) The FORMAT specification Aw permits w
alphameric characters to be read into or
written from a named location and manipulated internally.
(2) The ACCEPT TAPE and PUNCH TAPE statements provide punched tape input and output.
(5) Arrays may have up to three dimensions.
(6) Names may be up to six characters long.
(3) The ACCEPT and TYPE (or PRINT) statements
provide typewriter input and output.
(7) Continuation cards can be used to handle source
statements of more than 66 characters.
(8) Internal item sizes can be preset to any value
from 4 to 10 digits for integers and from 4 to
30 digits (including a 2-digit exponent) for floating point items.
Restrictions and extensions of the mM 1620
FORTRAN II language relative to mM 709/7090
FORTRAN II are summarized below.
©
(4) Floating point data values from 10- 99 to 10+99
can be accomodated.
.15
Publication Date: . . . June, 1962.
.2
PROGRAM STRUCTURE
.21
Divisions
Procedure statements:
1962 by Auerbach Corporation and BNA Incorporated
algebraic formulae.
comparisons and jumps.
input and output.
12/62
412:162.210
§
IBM 1620
162.
Divisions (Contd.)
.21
Data statements:
.22
Procedure Entities
Program:
Subroutine:
FUnction:
Statement:
. 23
statements.
subroutines.
functions.
statements.
statements.
characters; blanks are
ignored.
Data Entities
Arrays:
Items: .
Hollerith item:
Alphameric:
.24
FORMAT: describes the
layout, size, scaling
and code of input-output
data.
EQUIVALENCE: used to
cause two variables to
have a common location
or to specify synonyms.
COMMON: used to cause
a name to be common to
more than one segment
rather than local to each.
DIMENSION: describes
the elements in each
dimension of an array or
set of arrays.
floating point or integer
variables.
floating point variables or
constants.
integer variables or
constants.
Hollerith item.
alphameri~ item.
alphameric item that can
only be used for output.
alphameric item that can
only be input during a run.
It can be used for output,
or as a format statement.
.242 Designators (Contd.)
Typewriter
input:
Typewriter
output:
Comments:
Translator control: .
.25
Form may be
Integer only:
Signed:
Truncated
fraction:
Rounded
fraction:
.253 Synonyms
Preset:
Dynamically set:
Function name:
Subroutine name:
Data
Integer variables:
Floating point
variables:
Equipment
Card:
Paper tape:
12/62
A to Z, 0 to 9.
1 to 6 char.
yes.
first char must be a letter .
unsigned integer, 5 digits
max.
same as variable being
defined.
no designator.
none.
o to 3.
all variables.
no.
only integers.
yes.
at most C * N + C'; where
C and C' are literals.
C, C', andN.
no.
no .
no.
EQUIVALENCE statement.
no.
Numer of Names
.261 All entities:
Names
.241 Simple name formation
Alphabet: .
Size: .
Avoid key words:
Formation rule:
. 242 Designators
Procedures
Statement label:
implied by verbs PRINT,
TYPE.
C in column 1.
key words DIMENSION,
EQUIVALENCE,
COMMON.
Structure of Data Names
.251 Qualified names:
.252 Subscripts
Number per
item:
Applicable to:
Class may be
Special index
variable:
Any variable:
Literal:
Expression:
.26
implied by verb ACCEPT.
.262 Procedures:
.263 Data
Files:
Items:
Data levels:
.264 Equipment
Card reader:
Card punch:
Paper tape reader:
Paper tape punch:
Typewriter:
symbol table holds up to
about 500 entries with
40, 000 storage positions
and about 900 entries
with 60, 000 storage
positions. An entry is
required for each name,
fixed or floating point
number, and statement
label in the source program.
see .261, above.
limited by hardware.
see .261, above.
1.
1.
1.
1.
1.
1.
initial I,J,K, L, M, N.
any other initial letter .
. 27
Region of Meaning
of Names:
implied by verbs READ,
PUNCH.
implied by verbs ACCEPT
TAPE, PUNCH TAPE.
•
Ir-A-U-ER-BA-CH-.L~~
all names are local to the
program, subroutine, or
function in which they are
defined unless specified
in a COMMON statement .
PROCESS ORIENTED LANGUAGE: FORTRAN II
§
162.
.346 Choice of external
code:
. 347 Possible external
codes:
.3
DATA DESCRIPTION FACILITIES
.31
Methods of Direct Data Description
.311 Concise item
picture:
,312 List by kind:
. 313 Qualify by adjective:
.314 Qualify by phrase:
. 315 Qualify by code:
.316 Hierarchy by list:
.317 Level by indenting:
.318 Level by coding:
.319 Others
Array size:
Four-digit integer:
Four-digit
integers,S: .
Floating point
items:
.32
Files and Reels:
. 33
Records and Blocks
.331 Variable record
size:
.332 Variable block
size:
. 333 Record size
range:
.334 Block size range: .
. 335 Choice of record
size:
. 336 Choice of block
size:
.337 Sequence control:
.338 In -out error
control:
. 339 Blocking control:
.34
FORMAT statement.
no .
no.
no .
yes; first letter of name.
no.
no.
no.
. 343 Choice of external
radix:
. 344 Possible radices:
. 345 Justification: .
.348 Internal item size
Variable size:
Designation:
Range
Fixed point
numeric:
Floating point
numeric:
DIMENSION (4, 7).
FORMAT (14).
Alphameric:
FORMAT (514).
FORMAT (F8.3, ElO.4)
for +999.999 and
+9999E+99.
own coding.
dynamic.
dynamic.
no limit.
Maximum of 80 char for
cards, 87 char for
punched tape and typewriter •
input-output and FORMAT
statements .
.349 Sign provision:
.35
.351 Constants
Possible sizes
Integer:
Floating point:
Alphameric
(Hollerith);
Subscriptable:
Sign provision:
.352 Literals:
.353 Figuratives:
.354 Conditional
variables:
.36
automatic.
none; 1 or more blocks
per logical record.
none .
standard IBM 1620 punched
tape and card codes;
octal or binary codes not
permitted.
preset variable.
by parameter card.
4 to 10 digits.
2 to 28 digits for fixed
point part, 2 for exponent.
1 to 14 characters; maximum of 5 char if name
begins with I,J,K,L,M,
or N.
optional.
Data Values
±lOk , where k is a preset
value between 4 and 10.
10-99 to 10+ 99.
1 to 87 characters (1 line).
no .
optional.
same as constants.
own coding; e. g., TEN =
10.0.
computed GO TO.
Special Description Facilities
.361 Duplicate format:
input-output and FORMAT
statements.
1 logical record per input
or output statement.
Data Items
.341 Designation of
class:
. 342 Possible classes
Integer:
Fixed point:
Floating point:
Double precision:
Complex:
Alphameric:
Logical:
412: 162.300
.362 Redefinition:
.363 Table description
Subscription:
Multi - subscripts:
Level of item:
. 364 Other subscriptable
entities:
by multiple references to
single FORMAT statement.
COMMON statement.
EQUIVALENT statement.
yes .
1 to 3.
variables .
none •
.4
OPERATION REPERTOIRE
.41
Formulae
by name .
yes.
no.
yes.
no.
no.
yes.
no.
.411 Operator list
+:
*.
I:
**.
none .
decimal only.
alpha automatic left
justified.
integers automatic right
justified.
©
ABSF( ):
LOGF ( ):
SINF ( ):
COSF ( ):
EXPF ( ):
SQRTF ( ):
ATANF ( ):
1962 by Auerbach Corporation and BNA Incorporated
addition; also unary.
subtraction; also unary.
multiplication.
division.
exponentiation.
is replaced by .
absolute' value.
natural log.
sine •
cosine.
exponential.
square root.
arctangent.
12/62
412: 162.412
§
IBM 1620
162.
.45
. 412 Operands allowed
Classes: . . .
Mixed scaling:
Mixed classes:
Mixed radices:
Literals: . . .
. 413 Statement structure
Parentheses
a - b - c means:
a + b x c means:
a ~ b ~ c means:
abC means:
Size limit: •
Multi-results:
.414 Rounding of results: .
.415 Special cases fixed
x = -x:..
K =- K
X = -X.
x=x+1:
K=K+1
X=X+l.
x=4.7y: •. K=47*L/10
X=4.7*Y.
x = 5x107 + y2. 50000000 + L**2 X = 5.E7 + Y**2.
. 416 Typical
examples:. X = (-BtSQRTF(B*B-4.0*A*C»/(2.0*A)
.42
Operations on Arrals
none .
none.
none.
.43
Other ComEutation:
none.
. 44
Data Movement and Format
.441
. 442
.443
. 444
.445
Data copy example:
Levels possible:
Multiple results: .
Missing operands:
Size of operands
Exact match: •
Alignment rule
Numbers:
.446
. 447
. 448
.449
Alpha:
Filler rule
Numbers:
Alpha:
Truncating rule
Numbers: . .
Alpha:
Variable size
destination:
Editing possible
Change clas s:
Change radix:
Insert editing symbols
Actual point: . .
Suppress zeros:
Insert:
Float: . . . .
Special moves:
Code translation:
Character
manipulation: .
12/62
Step back a record:
Set restart point:
Restart: . . .
Start new reel:
Start new block:
Search on key:
Rewind:
Unload: . . • .
.46
Y=X .
items .
none.
not possible .
implied, except for alpha
or input-output.
right justified; normalized
for floating point.
left justified.
same as log.
PAUSE and decimal display, or TYPE message
and PAUSE •
use SENSE SWITCH •
•464 Accept option:
Object Program Errors
Error
Overflow:
In-out:
Invalid data:
Discovery
automatic
hardware
checks
automatic
PROCEDURE SEQUENCE CONTROL
.51
Jumps
.511 Destinations
allowed:
.512 Unconditional jump:
. 513 Switch: •
.514 Setting a switch:
truncate at left.
truncate at right.
.52
none.
statement.
GOTON.
GO TO (35,47,18), 1.
(Note that Assigned GO
TO is not available. )
I = 2; causes branch to
statement 47 .
GO TO (35,47,18), 1.
Conditional Procedures
.521 Designators
Condition:
Procedure:
.522 Simple conditions:
.523 Conditional relations: .
automatic.
automatic.
automatic point.
- sign only.
none.
automatic •
Special Actions
type error code
and continue.
stop or set indicator.
type error code
and continue .
.5
.515 Switch on data:
yes.
no.
TYPE or PRINT uses online typewriter.
.462 Messages to
operator: .
.463 Offer options:
zeros.
blanks.
no.
READ, WRITE, PUNCH,
PRINT, ACCEPT.
not possible.
none .
none.
own coding.
implied in each inputoutput statement.
none.
none.
none.
Operating Communication
.461 Log of progress:
.47
. 421 Matrix operations:
.422 Logical operations:
.423 Scanning:
own coding .
own coding.
Open:
Close:
Advance to next
record: . . .
numeric only.
yes.
only in exponentiation and
functions.
no.
yes.
(a-b) - c.
a+(bxc).
(a .;. b) ~ c.
illegal; parentheses must
be used.
330 char, on up to 5 cards.
no.
truncation of integers at
each step in expression.
floating
File Manipulation
. 524 Variable conditions:.
.525 Compound
conditionals: . . .
IF.
implied.
expression or variable
against zero.
IF (A) nl, n2, n3: if value
of expression A is less
than, equal to, or greater than zero, go to statement n1, n2, or n3,
respectively .
always zero •
no.
PROCESS ORIENTED LANGUAGE: FORTRAN"
§
412:.162.528
162.
. 56
.528 Typical example:
. 53
Subroutines
.531 Designation
Single statement:
Set of statements
First:
Last:
.532 Possible
subroutines:
. 533 Use in-line in
program:
.534 Mechanism
Cue with
parameters:
Number of
parameters:
Cue without
parameter:
Formal return: . .
Alternative return:
. 535 Names
Parameter call by
value: . . . • .
Parameter call by
name:
Non-local names:
Local names:
Preserved own
variables:
. 536 Nesting limit: .
. 537 Automatic recursion
allowed:
.54
not possible.
SUBROUTINE.
END.
any number of statements.
no.
CALL XXX (X, Y, Z).
.561 Designation of loop
Single procedure:
First and last
procedures: . .
. 562 Control by count:
.563 Control by step
Parameter
Special index:
Any variable:
Step: . .
Criteria:
Multiple
parameters:
.564 Control by condition:
.565 Control by list:
.566 Nesting limit: . . .
.567 Jump out allowed:
.568 Control variable exit
status:
limited only by statement
length (330 char).
.6
CALL XXX.
RETURN at least once.
none .
.7
LIBRARY FACILITIES
.71
Identity:
.72
Kinds of Libraries
EXTENSION OF THE
LANGUAGE: . . . .
Set of statements
First:
Last:
.542 Level of procedure:
.543 Mechanism
Cue: . . . . .
Formal return:
. 544 Names
Parameter call by
value: . . . . .
Parameter call by
name: . • . . .
Non-local names:
Local names:
Preserved own
variables: . .
Operand Definition
by Procedure:
none.
current place to named
end; e.g., DO 173 I =
1, N, 2 •
none .
no.
integer only.
positive integer.
greater than.
no.
no .
no.
at least 80.
yes.
available.
can write new function
in library.
FORTRAN II subroutine
library.
none.
yes.
use COMMON.
all.
all.
no limit.
no.
.721 Fixed master: . .
.722 Expandable master:
no.
yes.
.73
Storage Form:
punched cards .
.74
Varieties of
Contents: .
Function Definition by Procedure
. 541 Designation
Single statement:
. 55
IF (X**2.0 - 3.0) 29,37,18;
~ to 29, 37 or 18 if
x 2 -3 is respectively
less than, equal to or
greater than zero.
Loop Control
"arithmetic statement
function" preceding first
executable statement;
e.g., FIRSTF (X) =
A*X+B.
.75
Mechanism
.751 Insertion of new
item:
.752 Language of new
item:
.753 Method of call:
FUNCTlON.
END.
any number of statements.
by name in expression.
RETURN .
none.
yes.
use COMMON.
all.
all.
none.
© 1962 by Auerbach
subroutines (27 standard).
functions (7 standard, up
to 43 user-defined) .
.76
separate run.
SPS or machine.
functions by name in procedures; arithmetic and
input-output routines are
inserted automatically.
Types of Routine
.761 Open routines exist:
.762 Closed routines
exist: . . . .
. 763 Open-closed is
variable:
no.
yes .
no.
.8
TRANSLATOR CONTROL
.81
Transfer to Another
Language:
Corporation and BNA Incorporated
no.
12/62
412: 162.820
IBM 1620
162.
.82 Optimizing Infonnation Statements
§
.821 Process usage
statements:
.822 Data usage
statements:
.83
.84
. 85
12/62
none.
COMMON.
EQUIVALENCE.
Translator
Environment:
no.
Target Computer
Environment:
no.
.9
TARGET COMPUTER ALLOCATION CONTROL
".91
Choice of Storage
Level:
no.
.92
Address Allocation:
none.
.93
Arrangement of Items
in Word in Unpacked
Form: • . . . . .-
.94
Program Documentation
. • . . . no.
Control:
.95
Assignment of InputOutput Devices:
Input-output Areas:
1 item per field is standard.
implied by input-output
verbs .
automatic.
PROCESS ORIENTED LANGUAGE: FORTRAN"
§
162.
. 56
.528 Typical example:
. 53
Cue without
parameter:
Formal return: .
Alternative return:
. 535 Names
Parameter call by
value:
Parameter call by
name:
Non -local names:
Local names:
Preserved own
variables:
. 536 Nesting limit:
.537 Automatic recursion
allowed:
.
not possible.
SUBROUTINE.
END.
any number of statements.
no.
CALL XXX (X, Y, Z).
limited only by statement
length (330 char).
CALL XXX.
RETURN at least once.
none .
Loop Control
.561 Designation of loop
Single procedure:
First and last
procedures: • .
. 562 Control by count:
.563 Control by step
Parameter
Special index:
Any variable:
Step: . .
Criteria:
Multiple
parameters:
.564 Control by condition:
.565 Control by list:
.566 Nesting limit: . • .
.567 Jump out allowed:
.568 Control variable exit
status:
.6
EXTENSION OF THE
LANGUAGE: . . . .
.7
LIBRARY FACILITIES
.71
Identity:
. 72
Kinds of Libraries
Set of statements
First:
Last:
.542 Level of procedure:
.543 Mechanism
Cue: . . . . .
Formal return:
.544 Names
Parameter call by
value: . . . . .
Parameter call by
name: . . . . .
Non-local names:
Local names:
Preserved own
variables: • .
Operand Definition
by Procedure:
none.
current place to named
end; e.g., DO 173 I =
1, N, 2 •
none .
no.
integer only.
positive integer.
greater than.
no.
no .
no.
at least 80 .
yes.
available.
can write new function
in library.
FORTRAN n subroutine
library.
none .
yes.
use COMMON.
all.
all.
no limit.
no.
.721 Fixed master:. .
.722 Expandable master:
no.
yes.
.73
Storage Form:
punched cards .
.74
Varieties of
Contents: .
Function Definition b,y Procedure
. 541 Designation
Single statement:
. 55
IF (X**2.0 - 3.0) 29,37,18;
~ to 29, 37 or 18 if
x 2 -3 is respectively
less than, equal to or
greater than zero.
Subroutines
.531 Designation
Single statement:
Set of statements
First:
Last:
.532 Possible
subroutines:
. 533 Use in-line in
program:
. 534 Mechanism
Cue with
parameters:
Number of
parameters:
.54
412:·162.528
"arithmetic statement
function" preceding first
executable statement;
e.g., FIRSTF (X) =
A*X+B.
FUNC'qON.
END.
.75
Mechanism
.751 Insertion of new
item:
.752 Language of new
item: . . .
.753 Method of call:
any number of statements.
by name in expression.
RETURN.
none.
yes.
use COMMON.
all.
all.
none.
©
subroutines (27 standard).
functions (7 standard, up
to 43 user-defined) •
.76
separate run.
SPS or machine.
functions by name in procedures; arithmetic and
input-output routines are
inserted automatically.
Types of Routine
.761 Open routines exist:
.762 Closed routines
exist: . . . .
. 763 Open-closed is
variable:
no.
yes .
no.
.8
TRANSLATOR CONTROL
.81
Transfer to Another
Language:
1962 by Auerbach Corporation and BNA Incorporated
no.
12/62
412:162.820
IBM 1620
§ 162.
.82 Optimizing Infonnation Statements
.821 Process usage
statements:
.822 Data usage
statements:
.83
.84
.85
12/62
none.
COMMON.
EQUIVALENCE.
Translator
Environment:
no.
Tar&!:t Computer
Environment:
no.
.9
TARGET COMPUTER ALLOCATION CONTROL
".91
Choice of Storage
Level:
no.
.92
Address Allocation:
none.
.93
Arrangement of Items
in Word in Unpacked_
Form:
.94
Program Documentation
Control:
no.
.95
Assignment of InEutOutEut Devices:
Input-output Areas:
1 item per field is standard.
implied by input-output
verbs.
automatic.
.-
412: 163.100
II
•
STANDARD
EDP
IBM 1620
REI'ORTS
Process Oriented Language
GOTRAN
PROCESS ORI ENTED LANGUAGE: GOT RAN
§
163.
• 14
.1
GENERAL
• 11
Identity:
GOTRAN .
.12
Origin:
IBM General Products
Division, San Jose,
California.
.13
Reference: •
.14
Description:
IBM Publication C26-5594-0.'
The GOTRAN language is a severely restricted version of FORTRAN that serves as the source language
for IBM 1620 GOT RAN, a TTload-and-go" algebraic
programming system. The GOTRAN program (Section :192) accepts source statements from punched
tape, cards, or keyboard; compresses and loads
them into core storage; and then executes them interpretively.
The GOTRAN language is so limited in scope and
flexibility that it has been found inadequate even for
use as an educational tool by a number of universities. These limitations are summarized in the following list of significant restrictions and extensions
of the GOT RAN language relative to IBM 709/7090
FORTRAN.
Description:
(Contd. )
Restrictions
(Contd.)
(3) Arrays are limited to one dimension, and integer
arrays are not permitted.
(4) Subscripts must be single integer variables; no
subscript arithmetic is permitted.
(5) The only integer variable names permitted are
I, J, K, L, M, and N, so no more than six different integer variables can be used in a program.
(6) Nested DO loops are limited to six levels.
(7) Floating point variable names are limited to four
characters.
(8) Values of integer constants and variables cannot
exceed 999.
(9) Integer multiplication, division, exponentiation
and functions are not permitted.
(10) The only functions available are LOG, EXP,
SQR, SIN, COS, and ATN.
(11) Statement numbers may not be greater than
999.
Restrictions
(1) The following statements are not permitted:
SUBROUTINE
CALL
RETURN
FUNCTION
COMMON
EQUIVALENCE
FREQUENCY
ASSIGN
Assigned GO TO
Computed GO TO
SENSE LIGHT i
IF (SENSE LIGHT i)
IF (SENSE SWITCH i)
IF ACCUMULATOR OVERFLOW
IF DMDE CHECK
IF QUOTIENT OVERFLOW
FORMAT
READ INPUT TAPE
WRITE OUTPUT TAPE
READ TAPE
WRITE TAPE
READ DRUM
WRITE DRUM
END FILE
REWIND
BACKSPACE
(2) Each arithmetic statement may contain only one
operation symbol; e. g., C = A + B is permitted,
but C = A + B - D is not.
©
(12) No continuation cards are permitted; statement
length cannot exceed 72 characters.
(13) Input and output formats are fixed: each READ
statement inputs one to five variables (one per
card or tape record); each PUNCH statement
outputs one to five variables (one per card or
tape record); and each PRINT statement types
one to five variables on a single line.
(14) Implied 00 loops in input-output statements are
not permitted.
(15) Hollerith constants and alphameric items are
not permitted.
(16) Double preCision, complex, and logical operations are not permitted.
(17) Symbolic coding cannot be incorporated into the
source program.
Extensions
(1) The PLOT statement can plot curves on the console typewriter, using any typewriter character.
Maximum width of the plotting field is 66 characters, and the variable(s) to be plotted must be
scaled between 0 and 66.
(2) Floating point quantities from 10- 50 to 10+4 9
can be represented.
1962 by Auerbach Corporation and BNA Incorporated
12/62
412: 164.100
•
•
STANDARD
EDP
IBM 1620
REPORTS
Process Oriented Language
FORGO
PROCESS ORI ENTED LANGUAGE: FORGO
Restrictions:
§l64
(1) The following statements are not permitted:
.1
GENERAL
.11
Identity:
FORGO
.12
Origin:
Charles W. McClure,
Kenneth R. Sanderson,
and Joel Davis; Engineering Computing Laboratory,
University of Wisconsin.
Source of information: . C. H. Davidson, Associate
Professor, University of
Wisconsin.
.13
Reference: • . . . . . 1620 General Program
Library No.2. O. 008 and
2. O. 009.
.14
Description
The FORGO language is a dialect of FORTRAN that
is used in writing source programs for the FORGO
and FOR-TO-GO systems. FORGO is a "load and
go" algebraic programming system developed primarily for educational purposes at the University
of Wisconsin. FOR-TO-GO is a closely related
system that uses virtually the same source language,
but in order to accommodate larger object programs,
the FOR-TO-GO translator is overwritten by the subroutine package before the object program is exe~
cuted. The FORGO and FOR-TO-GO programs and
their configuration requirements are described in
Section :193.
The FORGO source language can best be described as
a restricted version of IBM 709/7090 FORTRAN.
Though there are certain incompatibilities, the compatible subset of the two languages will be adequate
for defining many routines. In general, only minor
changes will be required to convert a working FORGO source program into 709/7090 FORTRAN or 1620
FORTRAN I or II.
The use of FORMAT statements for input and output
data is optional. When used, all the fleXibility of the
709/7090 FORMAT statement is available. When
FORMAT numbers are omitted from input statements,
a free style of input is permitted, and "almost any
number recognizable to a human being" will be correctly handled. A comma or one or more blanks are
used to separate numerical fields. Output statements
without FORMAT numbers result in a fixed output
format with up to ten items per line.
It is assumed that input and output will be on punched
cards and that the output cards will be listed on an
IBM 407. The only alternative input-output device is
the console typewriter; no paper tape version is available. The restrictions and extensions of the
FORGO language relative to mM 709/7090 FORTRAN
are summarized below.
©
FUNCTION
COMMON
EQUN.ALENCE
FREQUENCY
ASSIGN
Assigned GO TO
SENSE LIGHT i
IF (SENSE LIGHT i)
IF ACCUMULATOR OVERFLOW
IF DNIDE CHECK
IF QUOTIENT OVERFLOW
READ INPUT TAPE
WRITE OUTPUT TAPE
READ TAPE
WRITE TAPE
READ DRUM
WRITE DRUM
END FILE
REWIND
BACKSPACE
(2) Arrays are limited to two dimensions.
(3) Double precision, complex, and logical operations are not permitted.
(4) Names are limited to five characters.
(5) Values of integer constants and variables cannot exceed 99, 999.
(6) No continuation cards are permitted in FORGO;
one is permitted in FOR-TO-GO.
(7) Arithmetic statement functions are not permitted.
(8) Statement numbers may not be greater than 9999.
(9) Symbolic coding cannot be incorporated into the
source program.
(10) No subroutine arguments can be specified in the
SUBROUTINE or CALL statements; all names
are universal to the main program and all subroutines compiled with it.
(11) In raising a floating point quantity to a fixed
point power, the power term may not involve
fixed point operations; e. g., A**(I+J) is not
permitted.
(12) IF (SENSE SWITCH i) tests one of ten internal
pseudo switches that can be set or reset only
via the typewriter; the console Program Switches
are used to control the FORGO system.
(13) No more than eleven statement numbers may
appear in the list of a Computed GO TO
statement.
1962 by Auerbach Corporation and BNA Incorporated
12/62
412: 164.140
§
IBM 1620
.14
164 .
. 14
r:fescription (Contd.)
Description (Contd. )
Extensions (Contd.)
Restrictions:
and up to 10 items per line. Output format 112
or IPEI2.4 is automatically supplied, depending
on whether the variable has a fixed or floating
point name.
(Contd.)
(14) The only functions available are SIN, COS,
ATAN, EXP, LOG, SQRT, and ABS.
(15) It is not possible to ·read in all the elements of
an array by simply specifying the array name
(which has previously been listed in a
DIMENSION statement) in a READ statement.
Extensions:
(1) Use of FORMAT statements is optional; when not
used, input format is virtually unrestricted, and
output format is fixed at 12 positions per item
12/62
(2) Floating point quantities from 10- 51 to 10+ 49
can be represented.
(3) Use of the terminal F is optional in all function
names, and other variations in function naming
are permitted; e.g., SQR, SQRF, SQRT, and
~QRTF are all acceptable names for the square
root function.
(4) Subscripts may be of the form: 1* J ± K.
412:171.100
•
STANDARD
EDP
•
REf'{JRTS
IBM 1620
M. O. Language
SPS
MACHINE-ORIENTED LANGUAGE: SPS
§
171.
.14
.1
GENERAL
.11
Identity:
mM 1620/1710 Symbolic
Programming System.
. 12
Origin: .
mM General Products Division' Applied Programming.
San Jose, California.
. 13
Reference: . . . . . . . IBM Reference Manual C265600-0; IBM 1620/1710
Symbolic Programming
System.
.14
Description
1620/17l0 SPS is an extension of 1620 SPS, which has
been in use since late 1960. 1620/1710 SPS is available for punched tape or card oriented systems.
Card systems have a speed advantage because data
transfer is buffered. The name SPS in this section
refers to 1620/1710 SPS. SPS translates codes used
for the 1710 Control System, but these are not described here_ The 1710 Control System is designed
to control industrial processes and is described in
appropriate mM manuals. It uses the 1620 system
as a digital computer.
The number of labels permissible in a basic card
system (20, 000 digits store) is limited to 176 threecharacter symbols (fewer if larger symbols are
used). The larger storage configurations are useful
in increasing the symbol table size to 1,610 or
3, 020 three-character symbols, using additional
storage of 20, 000 and 40, 000 digits. 1710 SPS II
language permits an unlimited number of labels by
employing repeated-pass assembly; see 1710 Control System Bulletin, mM Form J26-5643-0 .
It may be noted that there is little similarity between SPS for the 1620 and SPS for the mM 1401.
.15
Publication Date:. . .• 1962.
.2
LANGUAGE FORMAT
.21
Diagram: •
BPS is a straightforward language allowing symbols
to name procedures, operands, and constants. It
has a number of pseudo operations for defining symboIs, areas, and tables, and two which provide control of object program segmentation. BPS has the
facility of providing identifying header characters
which modify labels of procedures following the
heading statement. In this manner groups of labels
containing some identical labels can be differentiated
from each other. Addl:-ess adjustment (i. e., relative addressing) may be performed using actual or
symbolic bases and adjustment values. Adjustment
operations allow several adjustments to one base;
each adjustment can be an add, subtract, or multiplication operation. A flag indicator operand allows
simple specification of locations where flags are to
be set in instructions.
©
refer to SPS Coding Sheet.
412: 131. 2.
• 22
Page No.:
Line: .
sheet identification.
line identification; every
tenth number used.
Label:
names the location of an
instruction or data item.
actual or mnemonic operaOperation:
tion code, or mnemonic
representation of a pseudo
operation or macro code.
Operands & Remarks: . contains actual or symbolic
operands, flag indicator
operand, and comments.
Number of operands varies depending on machine
instruction, pseudo or
macro code.
BPS provides macro codes which call subroutines for
16 floating point arithmetic and mathematical functions, and one fixed point divide function. Up to 12
additional macros may be specified for each installation, calling available subroutines or ones written at
the installation. The subroutine library is available
in five forms, to handle the following conditions:
Fixed or variable length mantissa floating point
numbers; 1620 with or without Automatic Divide
special feature; and variable length mantissa floating point operations using the Automatic Floating
Point special feature.
Description (Contd.)
• 23
Corrections: • . • . . . spare lines at bottom, of
coding sheet, and gaps in
line number sequence.
.24
Special Conventions
.241 Compound addresses:
.242 Multi -addresses:
. 243 Literals: . . . • •
1962 by Auerbach Corporation and BNA Incorporated
BASE + ADJUSTMENT.
BASE - ADJUSTMENT.
BASE * ADJUSTMENT
(multiplication).
Form is BASE (op) ADJI
(op) ADJ2 (op) ADJ3 ••.. ;
limited by Operands and
Remarks field length
(59 char).
none.
only occur in address
fields of specific instruction codes,
12/62
412:171.244
§
IBM 1620
171.
.244 Special coded addresses:
.245 Other
Actual core storage
address: . .
Flag indicator
operand: . .
1 to 5 decimal digits.
specifies digit locations for
flags in instructions.
Flag in literal to
delimit field: . •
automatically set at 5 decimal digits unless flag
indicator operand is used
for flag control.
$ sign used in label.
Heading indication: .
Record mark
insertion: • . . . .
Comment insertion:
Flag for indirect
addressing: .
.3
LABELS
. 31
General
refers to this address in
machine instructions.
01< refers to storage address
last allocated when used
in pseudo operation.
01<
@sign used.
in label, column 6.
01<
may use minus sign preceding operand.
. 311 Maximum number of labels
Labels are used for addresses of constants, data
items, procedures. Number depends on size of
core store and size of label according to table
shown below:
LABEL
SIZE
2
3
4
5
6
Char.
Char.
Char.
Char.
Char.
STORE SIZE (DECIMAL DIGITS)
20,000
40,000
60,000
Card Tape Card Tape Card Tape
205
176
154
137
137
304
261
228
203
203
1880
1610
1410
1250
1250
1970
1690
1480
1310
1310
3540
3020
2650
2360
2360
3630
3120
2730
2420
2420
· 322 Labels for library
routines: . . . . .
. 323 Labels for constants:
· 325 Labels for records: .
.326 Labels for variables:
new macro code mnemonics
must be written for added
subroutines; these must
be alphabetic, and 1 to 4
char long.
same as Procedures •
same as Procedures.
same as Procedures.
• 33
Local Labels:
none.
.4
DATA
.41
Constants
.411 Maximum size constants
Machine Form
External Form
Integer
Decimal:
50 decimal digits.
Octal: . . .
none.
Hexadecimal:
none.
Fixed numeric: .
none as such; decimal point
is implicit.
Floating numeric
Decimal:
by programming, using DC
pseudo op twice.
Octal: •.
none .
Hexadecimal:
none.
Alphabetic:
50 char max .
Alphameric:
50 char max.
.412 Maximum size literals
Machine Form
External Form
Integer
limited to 5 decimal digits
Decimal:
for practical use.
Octal: . .
none.
Hexadecimal:
none.
none as such; decimal point
Fixed numeric: .
is implicit.
none.
Floating numeric:
Alphabetic: .
none.
Alphameric:
none.
..
.42
Worki!!& Areas
Number of Labels
.421 Data layout:
.312 Common label formation
rule: . . . . . . .
. 313 Reserved labels: . .
. 314 Other restrictions:
• 315 Designators: • . . .
• 316 Synonyms permitted:
. 32
yes.
none .
none .
none .
yes; 2 or more labels may
refer to same storage
area address if assigned
by a defined symbol or an
actual value.
.422 Data type:
.423 Redefinition: •
.43
Input-Output Areas
.431 Data layout:
Universal Labels
• 321 Labels for procedures
Existence: . . . .
Formation rule
First character:
Others: • . .
Number of
characters:
12/62
specified in program by
pseudos.
tabulated in program by
pseudos.
more than 1 label may have
same address. Address
should be defined symbol
or actual value.
.432 Data type:
.433 Copy layout:
specified in program by
pseudos .
impliCit by layout.
no.
optional.
numeric, alphabetic, or
special; specials are equal
sign (=), slash (/), at
sign (@), and period (.).
same as first.
1 to 6; at least one of them
must be an alphabetic or
special char.
.5
PROCEDURES
.51
Direct 0eeration Codes
.511 Mnemonic
Existence:
Number:
Example: .
Comment:
optional.
93.
A= Add.
includes all instruction
variations.
MACHINE·ORIENTED LANGUAGE: SPS
§
412:171.512
171.
.64
. 512 Absolute
Existence:
Number:
Example: .
Comment:
. 52
area of library .
optional.
47.
21 = Add.
variations specified by
values in operand fields.
Macro-Codes
• 521 Number available
Input-output: . .
Arithmetic: . . .
Math functions: .
Error control: .
Restarts: . . . .
Data movement:
· 522 Examples: . .
.523 New macros:. . .
.53
Interludes: . . . . .
· 54
Translator Control
. 541 Method of control
Allocation counter: •
Label adjustment:
Annotation: . . . .
.542 Allocation counter
Set to absolute: .
Set to label: .•
Step forward:. .
Step backward: .
Reserve area:
· 543 Label adjustment
Set labels equal:
Set absolute value: •
Clear label table:
. 544 Annotation
Comment phrase:
O.
7.
.. 8.
O.
O.
2.
FA = Floating Add.
up to 12; into library after
separate assembly.
.65
psuedo operation.
pseudo operations.
see.544.
DORG pseudo.
DORG pseudo.
by address adjustment.
by address adjustment.
no.
only by assignment of
actual address to each.
by pseudos.
not possible.
* in label column,
.6
SPECIAL ROUTINES AVAILABLE
· 61
Special Arithmetic
Input-Output Control:
.66
Sorting: . . . . . . . . by insertion into expandable
area of library .
.67
Diagnostics: . . . . . . by insertion into expandable
area of library •
.7
LffiRARY FACILITIES
. 71
Identity: . . . . . .
. 72
Kinds of Libraries: . . fixed master of 17 subroutines and expandable
master of up to 12 subroutines. Expandable
master may be standard
in installation, or
private.
. 73
Storage Form: . . . .
. 74
Varieties of Contents: . floating point arithmetic
and mathematical functions; fixed point divide.
See Paragraph. 52.
,.75
. 612 Method of call: . .
· 62
Special Functions
· 621 Facilities: . . .
· 622 Method of call: .
. 63
.631 Facilities: . . .
.76
Types of Routines
.761
.762
.763
. 764
Open routines exist: . .
Closed routines exist:.
Open-closed is optional:
Closed routines appear
once: . . . . . • . . .
. 632 Method of call: •
©
no.
yes; all macros .
no.
yes.
.8
MACRO AND PSEUDO TABLES
.81
Macros
Description
FA (Floating Add): . .
FS (Floating Subtract):
FM (Floating
Multiply): . . .•• • .
FD (Floating Divide):
DIY (Fixed Point
Divide): . • . . . .
FSRS (Floating Shift
Right): . . . . . .
read and execute segment
of object language.
use TCD and TRA pseudos .
cards .
punched tape.
.751 Insertion of new item: . special assembly run, and
insertion into subroutine
deck or punched tape.
.752 Language of new item:. machine language assembled from SPS.
• 753 Method of call: •.
macro code .
8 floating point functions.
macro codes.
Overlay Control
subroutine group for
1620/1710 SPS.
Mechanism
Code
floating point operations,
fixed point divide, floating
shifts and mantissa length
adjustment.
macro codes .
by insertion into expandable
area of library.
none.
or following operands in operand and remarks field.
Title phrase: . . . • . none (use comments).
. 545 Other
Ignore a label'following Heading
statement: . • . . • $ sign prefixed to label
concerned.
.611 Facilities: . . . . •
by insertion into expandable
Data Editing:. . . . .
FSLS (Floating Shift
Left): . . . .
1962 by Auerbach Corporation and BNA Incorporated
(A) + (B) -;:,A.
(A) - (B) ~A.
(A) x (B) -':>A.
(A) 7 (B) -">-A.
(A) 7 (B)---:;" standard area.
Positioning control of operands included.
specified significant portion
of mantissa shifted right.
specified significant portion
of mantissa shifted left.
12/62
412:171.810
IBM. 1620
§ 171 •
• 81
.82
Macros (Contd.)
Code
TFLS (Transmit
Floating): . . .
BTFS (Branch and
Transmit Floating):
(B)~A.
address of next instructions
stored; (B) ~A-1;
branch to A.
FSQR (Floating
Square Root): .
.
FSIN (Floating Sine):.
FCOS (Floating
Cosine): . . . . • . .
FATN (Floating
Arctangent): . . . .
FEX (Floating Exponential (Natural»: •
FEXT (Floating Exponential (Base 10»: •
FLN (Floating Logarithm (Natural»: .
FLOG (Floating Logarithm (Base 10»:
.82
{(B)~A.
sin(B) ~A.
cos(B)-?>-A.
arctan(B)
e
~A.
exp(B)~A.
10 exp(B) ~ A.
naturallog(B) ~A.
common 10garithm(B) ~A.
Pseudos
Code
DS (Define Symbol
(Numerical»: • .
Description
assigns address (also considered a value) and field
length, at low-order end
of field. *
DSS (Define Special
Symbol (Numerical»: like DS, but assigns address to high-order end
of field.
DAS (Define Alphalike DS, but allows twue
meric Symbol): . .
field length specified because of alphameric field.
DC (Define Constant
assigns field length, con(Numerical»: . . .
stant, and address at loworder end of field. *
Field is flagged.
DSC (Define Special
Constant
like DC, but assigns ad(Numerical»: . . .
dress to high-order end
of field. Field is not
flagged.
DAC (Define Alphaassigns field length, alphameric Constant): .
meric constant, and address at left-hand end of
field plus I position.
Specified field length is
doubled. *
12/62
Pseudos (Cont'd)
Description
DSA (Define Symbolic
Address): • . • . • . assigns table of up to 10
actual addresses of symbolic or actual addresses
in pseudo. Address is
location of left-most address in table. Address
is assigned by processor.
DSB (Define Symbolic
Block): . . . . • . • assigns left-most address
of array of numbers, with
element size and number
of elements speCified. *
DNB (Define Numerical
Blank): . . • . • • • defines field of blanks,
specifying number of
blanks. Field is not
flagged. *
DORG (Define Origin): sets a value into location
assignment counter, defined by symbolic or
actual address.
DEND (Define End): . halts computer after object
program is loaded. If
symbolic or actual address is specified, computer jumps to address
(start) and starts execution, when start key is
hit.
SEND (Special End)
(Tape system only):
halts assembly until start
key is hit.
HEAD (Heading): . . . precede following labels of
5 char or fewer with the
specified symbol.
TCD (Transfer Control and Load): . .
causes loading of object.
program intO object computer,to halt, and a
branch to occur to start
·of program after arithmetic tables are loaded.
TRA (Transfer to
Return Address): •
causes a) stop of object
computer program execution, b) reading of loader
routine, and c) loading of
next segment of object
program.
* address may be actual or symbolic.
If address is
left blank, it is assigned by processor.
412:172.100
•
II
STANDARD
EDP
IBM 1620
M. O. Language
One-Pass Tape SPS
"l'ORTS
MACHINE ORIENTED LANGUAGE: ONE-PASS TAPE SPS
§
172.
.14
.1
GENERAL
.11
Identity:
IBM 1620 One-Pass Tape
Symbolic Programming
System (Punched Tape).
.12
Origin:
IBM.
. 13
Reference: .
IBM 1620/1710 Symbolic
Programming System
(1962); IBM Reference
Manual C26-5600-0.
Description (Contd. )
Numerical Constant, and Define Symbolic Address);
and using the control pseudos DORG, DEND, and
HEAD, for allocation counter control.
The restrictions on the pseudo operations prevent
easy program segmentation. Automatic floating
point operations and automatic divide operations are
not permitted. Indirect addressing (special feature)
may be used with the system.
Additional restrictions are listed below:
(1) Numeric operation codes are not allowed.
SPS One -Pass for Paper
Tape; Program 1620-SP007 (December, 1960).
.14
Description
(2) Address adjustment is limited to addition
and subtraction.
(3) An instruction operand may contain at the
most one previously undefined symbolic term.
The 1620 One-Pass SPS program has been withdrawn
by IBM, but is still used at a number of installations.
The One-Pass SPS language is designed for the punched tape 1620 system, and reduces tape handling requirements and assembly time from that of the full
two-pass SPS system (described in section: 171). It
operates on a restricted language format and a restricted set of pseudo operations, compared to the
1620/1710 two-pass SPS; otherwise, the two languages are essentially the same. No macro codes
and their corresponding subroutines may be used.
These restrictions limit the language to specifying
standard 1620 machine operations; using the pseudos
DS, DC, and DSA (Define Numerical Symbol, Define
©
(4) In area definition operands, all symbols must
have been previously defined.
(5) The flag indicator operand may specify only
one flag position in an instruction. However,
a negative operand will receive a flag in the
least significant position, as usual.
(6) The maximum number of labels permitted is
199.
(7) The first character of a label must be alphabetic; no special characters are allowed in
the label.
1962 by Auerbach Corporation and BNA Incorporated
12/62
412:181.100
.STANDAAD
II
EDP
IBM 1620
REPORTS
Program Translator
SPS
PROGRAM TRANSLATOR: SYMBOLIC PROGRAMMING SYSTEM
§
181.
.12
.1
GENERAL
. 11
Identity: • . . . . . . . 1620/1710 SPS, Card
System. 1620-SP-020.
There is no check on overflow of the storage allocation counter. The translator cannot allocate storage
greater in capacity than that of the translating computer.
1620/1710 SPS, Tape
System (punched tape).
1620- SP- 021.
. 12
Description (Cont'd)
The translator program can be modified easily to
run on a computer with increased storage capacity;
a translator program so modified may not be run on
a system with the smaller amount of storage .
Description
Up to 12 additional subroutines may be written for
insertion into the translator library. A macro code
and number must be assigned, and the subroutine
written in SPS language and assembled. The assembled program is added to the macro operation
library (card or punched tape).
This is a two-pass translator. A restricted version
of it, now withdrawn, is known as the 1620 One-Pass
Tape Processor and is described in section : 182.
The first pass of the source punched tape or cards
provides information for setting up a symbol table,
and for certain checking and error typeouts. Input
to the first pass can be from the typewriter, in
which case punched tape or cards are output. Errors in either pass may be corrected as they occur,
or after the pass, which is re-run. On the second
pass assembly, listing, and error checking is per
formed. Subroutines used in the program (macro
operations) are automatically read, relocated, and
punched out.
Assembled output includes a loader routine, arithmetic tables, and the relocated machine-language
subroutines used by the program. Also included is
the PICK subroutine, a supervisory subroutine which
coordinates the use of the other subroutines at run
time. PICK is used once per macro, requiring nine
milliseconds per use. Once the object program is
loaded, it may be executed immediately when the
start key is depressed.
In a card system, object language output can be in
uncondensed or condensed form, selected by the
setting of a sense switch. In condensed form, one
card contains five machine language instructions.
In uncondensed form, procedures are punched in the
form of two cards per symbolic instruction, including comments. Pseudos require two or three cards,
except for the TCD pseudo (11 cards). Each macro
requires several cards. In a punched tape system,
a machine language tape is produced. Full listing
on the typewriter is optional. An adjunct to the
translator program can provide a condensed card
deck from an uncondensed one in a separate pass.
The uncondensed deck can be listed off-line on a
407 printer.
The maximum size of the symbol table (card system) is 176 three-character labels; this may be increased to I, 610 and 3, 020 labels in machines with
40,000 and 60, 000 digits of storage, respectively.
Maximum label size is six characters. Note that
1710 SPS II translator permits an unlimited number
of labels by employing repeated-pass assembly; see
1710 Control System Bulletin, Form J26-5643-0.
Translating may be done alternatively on an IBM
700/7090 system at a speed of approximately 1,000
source statements per minute. The source program
must be transcribed off-line to magnetic tape.
. 13
Originator:. • . . . . . IBM General Products Division, Applied Programming.
San Jose, California.
. 14
Maintainer:
as above.
.15
Availability:
spring, 1962.
.2
INPUT
.21
Language
.211 Name:
.212 Exemptions:
.22
The translator keeps track of object computer addresses, and allocates the start of the subroutine
area following the program area. However, no
check is 'made that there will be sufficient room to
store the program and subroutines in the object
7omputer; this is the programmer's responsibility.
©
1620/1710 Symbolic Programming System (SPS).
none.
Porm
.221 Input media:
.222 Obligatory ordering: .
. 223 Obligatory grouping: .
1962 by Auerbach Corporation and BNA Incorporated
punched cards, punched
tape, typewriter, depending on processor.
none.
none.
12/62
IBM 1620
412:181.230
§
181.
· 23
Size Limitations
.231 Maximum number of
source statements:.
· 232 Maximum size source
statements: . • . . .
· 233 Maximum number of
data items: .
.234 Others
Maximum size of
symbol table: . .
.4
TRANSLATING PROCEDURE
.41
Phases and Passes
Pass 1: . . . . . . •
checks for valid mnemonic
operation codes.
forms symbol table.
assigns storage locations to
instructions, work areas,
constants.
performs checking for producing error messages.
punches tape or cards If
input was from the typewriter.
Pass 2: . . . . . . . . . punches loader.
forms numeric operation
codes.
forms numeric operands,
adjusts operands, sets
flags.
performs checking for producing error messages.
punches (and types) listing.
reads, assigns and punches
subroutines called for.
punches arithmetic tables.
no prat:ticallimit if overlays are used in object
program.
75 char.
depends on space available
in symbol table.
see table below.
STORE SIZE (DECIMAL DIGITS)
LABEL
SIZE
2
3
4
5
6
Char.
Char.
Char.
Char.
Char.
20,000
40,000
60,000
Card
Tape
Card
Tape
Card
Tape
205
304
261
228
1880
1610
1410
1970
1690
1480
203
203
1250 1310
1250 1310
3540
3020
2650
2360
2360·
3630
3120
2730
2420
2420
176
154
137
137
Number of Labels
.42
Optional Modes
.421 Translate: • . .
. 422 Translate and run: .
· 423 Check only: • . • •
.3
OUTPUT
· 31
Object Program
. 311 Language name:
. 312 Language style:
. 313 Output media:
· 32
mM 1620 machine code.
card deck or punched tape
is in numeric machine
code. Card deck is in
condensed or uncondensed
form. Usually contains
loader, program, followed by storage assignments, subroutine linkages, subroutines, and
arithmetic tables.
cards or punched tape .
Conventions
· 321 Standard inclusions:
loader routine and arithmetic tables. Also TCD,
TRA control instructions
if overlays are used.
PICK subroutine included
if macros are used.
yes .
no.
yes; translation with error
type outs but no punching
of object language .
• 424 Patching:. . . . . . . . no; correct individual
statements as error occurs, or reassemble after
a number of corrected
statements are collected .
.425 Updating:. . • . .
no.
.43
Special Features
.431 Alter to check 'only:
.432 Fast unoptimized
translate:. . . . .
· 433 Short translate on
restricted program:
no.
.44
Bulk Translating: . .
yes .
. 45
Program Diagnostics:
none in fixed master
library; may be incorporated as user wishes.
· 46
Translator Library
.461 Identity: . . . . .
. 462 User restriction:
.33
Documentation
Subject
Source program: •
Object program: .
Provision
optional typeout.
optional typeout and optional punching.
Storage map (symbol
table):. . . . • .
optional typeout.
Language errors: • . typeout.
12/62
. 463 Form
Storage medium: .
Organization: .
· 464 Contents
Routines: .
Functions:
Data descriptions:
see.423.
no.
subroutine group for 1620/
1710 SPS .
17 fixed master routines
and up to 12 additional
picked or written by user •
cards or punched tape.
fixed order, by call
numbers.
closed.
yes.
no.
PROGRAM TRANSLATOR: SYMBOLIC PROGRAMMING SYSTEM
§
181.
412:181.465
.612 Larger configuration
advantages:. . . . .
· 465 Librarianship
Insertion:. .
Amendment:
Call procedure:
assembly run.
assembly run.
macro codes used by subroutine processor program.
·5
TRANSLATOR PERFORMANCE
. 51
Object Program Space
.511 Fixed overhead
Name: . . . .
Space: . . . .
.512 Space required for each
input-output file: . . .
.513 Approximate expansion
of procedures: .
. 62
PICK subroutine supervisor.
approx. 900 decimal digits. .7
Error
1. 0 (exclusive of macros,
Missing entries:
Unsequenced entries:
Duplicate names:
Improper format:
Incomplete entries:
Target computer overflow:
Inconsistent program:
Symbol table full:
3 + O. 029S minutes; no
listing and uncondensed
deck (*). This is approximately 70% of card punch
speed; no macros used.
Translation with a condensed deck output would
proceed at reader speed,
250 cards per minute input for each pass (**).
7 + O. 029S minutes (*).
Punched tape system:
.522 Checking only
Card system: . . • . . 3 + O. 004S minutes (*).
Punched tape system: 7 + 0.0048 minutes (*).
.53
Optimizing Data: .
· 54
Object Program
Performance: .
ERRORS, CHECKS AND ACTION
as coded.
Translation Time
· 521 Normal translating
Card system: . . .
Target Computer
. 621 Minimum configuration: same as translating
computer .
. 622 Usable extra facilities: cards, larger core storage .
which vary).
· 52
none.
essentially unaffected when
few macros are used.
increased if many macros
are used; each macro
utilizes PICK subroutine,
which requires approximately 9 m. sec to
execute.
.6
COMPUTER CONFIGURATIONS
.61
Translating Computer
. 611 Minimum configuration: 1620 with punched tape input and output, and 20, 000
decimal digits of storage.
©
cards give faster assembly
because of buffered inputoutput.
larger storage holds larger
symbol table.
t
.8
Check or
Interlock
Action
not applicable.
none.
check
check
check
t
none.
none.
check
t
t
t
immediate halt and typeout of error type and reference to location of statement; or typeout with no
halt: selected by sense switch setting.
ALTERNATIVE TRANSLATORS
Computer:
Identity:
mM 709; mM 7090.
7090 Processor for Assembling 1620/1710
Programs.
Date: • . . . • • . . . . April, 1962.
Description
The 709 or 7090 requires a minimum of 32, 768
storage positions, two channels and ten magnetic
tapes. Processor runs are under the control of the
m SOS Monitor. Different diagnostic messages are
printed than on the 1620. Input can only be from
magnetic tape, prepared off-line. It is a two-pass
processor, and allows a symbol table of 3, 000
names. Output is on two magnetic tapes; one for
listing, and one for punching cards. If a punched
tape object program is required, these cards may
be converted to punched tape .
The 7090 Processor will assemble approximately
1, 000 SPS statements per minute (*).
1962 by Auerbach Corporation and BNA Incorporated
12/62
412:182.100
IBM 1620
Program Translator
One-Pass Processor (SPS)
PROGRAM TRANSLATOR: ONE·PASS TAPE SPS
§
.1
GENERAL
. 11
Identity:..
.12
Description
1620/1710 Symbolic Programming System.
all macros, some pseudos,
automatic floating point
instructions, automatic
divide, See description
in Section :172, One-Pass
Tape SPS.
.211 Name:
182.
.212 Exemptions:
SPS One-Pass Processor
for Paper Tape; Program
1620-SP-007 (December,
1960) .
.22
The Otie-Pass SPS program has been withdrawn by
IBM, but is still used at a number of installations.
The program translator is a one-pass system which
operates within the restricted language input of the
one-pass SPS language; otherwise it is similar to
the translator for the two-pass SPS language (section : 181). It assembles as the source statements
are read in; any statements which are correct but
which are not able to be completely assembled at
that time are placed in "suspense" until more complete information is available. The maximum number of labels allowed is 199.
There may be an undefined symbol in an instruction;
therefore, the instruction cannot be assembled completel)! as it is read. In such a .case, the instruction is assembled as completely as possible and
placed in temporary storage, and a note is made of
the undefined symbol. The typewriter listing contains the symbolic and incomplete machine coding of
the instruction, with a reference number assigned
to the symbol. When the symbol is defined, the instructions in temporary storage containing it are
completely assembled, and immediately punched out.
Although the object program instructions will not be
in correct sequence on the output tape, the loading
process will place the instructions into their proper
locations. The end of the program listing will contain each undefined symbol, its reference number,
and its defined address. When writing the source
program, care must be taken that an assembled instruction. if loaded out of order, will not modify any
locations improperly.
Form
.221 Input media:
.222 Obligatory ordering:
.223 Obligatory grouping:
.23
Size Limitations
.231 Maximum number of
source statements:
.232 Maximum size source
statements:
.233 Maximum number of
data items:
.3
OUTPUT
.31
Object Program
.311 Language name:
.312 Language style:
199 labels.
loader routine and arithmetic tables.
Documentation
Subject
The number of instructions that can be in temporary
storage at any time is equal to 454 - O. 77N, where
N is the number of instructions having undefined
symbols.
75 char.
Conventions
.321 Standard inc1usions:
.33
limited by size of object
computer, unless overlay control is coded.
IBM 1620 machine code.
loader. instructions with
storage address, arithmetic tables.
punched tape.
.313 Output media:
.32
punched tape or typewriter.
according to coding sheet
page and line numbers.
none.
Provision
Source program:
Object program:
Storage map (symbol table):. . . .
Language errors: .
optional typeout.
optional typeout and optional punching.
optional typeout.
typeout.
. 13
Originator:
IBM .
. 14
Maintainer:
none; withdrawn by IBM .
.4
TRANSLATING PROCEDURE
.15
Availabili!i'::
December, 1960.
. 41
Phases and Passes:
.2
INPUT
.42
Optional Modes
. 21
Language
.421 Translate:
©
1962 by Auerbach Corporation and BNA Incorporated
1 pass .
. . . . yes .
12/62
412:182.422
§
IBM 1620
182.
.611 Minimum configuration: .
. 422 Translate and run:
.423 Check only: .
.424 Patching:
.425 Updating:
no .
yes; translation with error
typeouts but no punching
of object language.
. 62 Target Computer
no; correct individual
statements as error occurs, or reassemble after .621 Minimum configuration: .
a number of corrected
statements are collected.
.622 Usable extra facilino.
ties: .
none .
same as translating computer.
1622 Card Read Punch.
Special Features
.43
.7
.431 Alter to check
only: .
.432 Fast unoptimized
translate:
.433 Short translate on
restricted program:
no .
.44
Bulk Translating:
yes.
.45
Program Diagnostics:
none.
. 46
Translator Library: .
none.
.5
TRANSLATOR PERFORMANCE
.51
Object Program Space
.511 Fixed overhead: .
. 512 Space required for
each input-output
file:
.513 Approximate expansion of procedures: .
.52
.612 Larger configuration
advantages: .
basic 1620 with punched
tape I/O; 20,000 decimal
digits of core storage.
see .423.
Missing entries:
Un sequenced entries:
Duplicate names:
Improper format:
Incomplete entries:
Target computer
overflow:
Inconsistent program:
Invalid character
present:
More than 1 symbol
in instruction has
not been defined:
Symbol in declarative statement not
previously defined:
Symbol table full:
Area for temporary
storage of instructions has been exceeded:
no.
none .
as coded.
1.0.
Translation Time
.521 Normal translating:
.522 Checking only:
.54
Object Program Performance:
'l
'l
Action
not applicable:
none.
check
check
t
t
none .
none .
none.
check
t
check
t
check
check
t
t
check
t
t immediate halt and typeout of error type and
reference to location of statement; or typeout
with no halt: selected by sense switch settings.
unaffected.
.8
.6
COMPUTER CONFIGURATIONS
.61
Translating Computer
12/62
ERRORS, CHECKS AND ACTION
Check or
Error
Interlock
ALTERNATIVE TRANSLATORS:
. IDM 709/7090. See section
:181.8, program translator for full SPS language.
412:183.100
.ST""RD
tl
EDP
REPORTS
IBM 1620
Program Translator
FORTRAN I
PROGRAM TRANSLATOR: FORTRAN I
§
.13
Originator:
IBM General Products
Division, San Jose,
California .
.14
Maintainer:.
as above .
.15
AVailability:
October, 1961 (Versions
without FORMAT were
released in November
1960 for paper tape and
April 1961 for cards).
Description
.2
INPUT
These translators convert IBM 1620 FORTRAN
source programs into machine language object programs in a single pass. The paper tape version can
be run on a minimwn IBM 1620 system; the card
version provides higher input and output speeds in
systems that include the 1622 Card Read Punch.
Bulk: translating is not possible because two segments of the translating program partially overlay
one another.
.21
Language
183.
.1
GENERAL
. 11
Identity:
FORTRAN with FORMAT
for Paper Tape.
Program 1620-FO-003.
FORTRAN with FORMAT
for Cards.
Program 1620-FO-004.
"FORTRAN 1."
.12
.211 Name:
.212 Exemptions:
.22
Form
.221 Input media:
The object tape or deck consists of a loading routine
that loads the compiled instructions in the proper
order to form the object program, the compiled instructions themselves, and the symbol table, which
is modified to form a data table. The subroutines
necessary for executing the object program (described in Paragraph .46) can either be punched into
the object tape or deck during compilation or loaded
separately before execution.
The "trace feature" facilitates debugging by inserting instructions that can cause the result of each
executed arithmetic statement to be typed. Insertion
and execution of the "tracers" is controlled by Program Switch settings. Each arithmetic statement
in a trace-compiled program requires 12 additional
digits of storage and l. 2 milliseconds additional
execution time when tracing is not elected.
The IBM 1620 FORTRAN Pre-Compiler is a useful
independent diagnostic routine that checks FORTRAN
source programs for 51 different format and logical
errors in a pre-compilation run. When an error is
detected, the error code and the statement containing the error are typed. The statement can then be
typed in correctly, and a corrected source program
tape or card deck can be produced. A swnmary
lists all undefined and unreferenced statement numbers and all relocatable subroutines called. The
Pre-Compiler is available in paper tape and punched
card versions; program nwnbers are 1620- FO-005
and 1620-FO-006, respectively. Use of the PreCompiler is optional.
©
.222 Obligatory
ordering:
· 223 Obligatory
grouping:
• 23
paper tape, punched cards,
or typewriter keyboard.
all statements in logical
sequence.
none.
Size Limitations
.231 Maximum nwnber of
source statements:
.232 Maximum size source
statements:
· 233 Maximum nwnber of
data items:
.3
OUTPUT
.31
Object Program
. 311 Language name:
.312 Language style:
. 313 Output media:
.32
IBM 1620 FORTRAN.
none, but note language
limitations in Section:
161.14.
limited by target computer
storage.
72 characters.
see 412:161.261.
IBM 1620 machine code .
condensed - 2 to 5 instructions. per card .
paper tape or punched card.
Conventions
.321 Standard inclusions:
1962 by Auerbach Corporation and BNA Incorporated
loading routine, arithmetic
and input-output subroutines (both may be
loaded separately at
execution time).
12/62
412: 183.330
§
IBM 1620
183.
. 33
.465 Librarianship
Insertion: .
Documentation
Subject
Source program:
Object program:
Symbol table:
Restart point list:
Language errors:
Provision
typed listing (optional).
none.
typed listing (optional).
none.
typed messages.
.4
TRANSLATING PROCEDURE
.41
Phases and Passes:
.42
0Etional Mode
.
.421 Translate:
.422 Translate and run:
.423 Check only:
.424 Patching:
.425 Updating:
.43
. 44
Amendment: . .
Call procedure:
one-pass compiler .
Note: The 1620 FORTRAN subroutine library contains 31 standard routines. Twenty-five are
used by the FORTRAN system only and are
not available to the programmer (arithmetic,
input, output, fix, float). Two are available
to either the FORTRAN system or the programmer (LOG, EXP). Four are available
only to the programmer as functions (SQRT,
SIN, COS, ATAN). Up to 19 additional
functions can be added by the user. The
functions are in relocatable form and are inserted only if used in the source program .
All subroutines can either be punched into
the object deck or tape at compile time or
loaded separately at execution time.
yes.
no.
yes, using IDM 1620
FORTRAN Pre-Compiler.
no.
no.
SEecial Features
.431 Alter to check
only:
.432 Fast unoptimized
translate:
. 433 Short translate on
restricted program:
Bulk Translating:
.45
no •
no .
.5
TRANSLATOR PERFORMANCE
not possible; the translating program consists of
two segments which
partially overlay one
another, so reloading is
necessary before each
source program.
.51
Object Program Space
instructions that will cause
optional printing of results of each executed
arithmetic statement can
be automatically inserted
at compile time. The·
trace facility is controll6i
by Program Switch
.settings.
.452 Snapshots:
.453 Dumps:
Translator
none .
none.
.511 Fixed overhead
Name
Space
Comment
Multip~y-add
tables:
Standard
subroutines:
Relocatable
subroutines:
300 digits.
7,200 to
8,000
digits.
varies
arithmetic, input,
output, fix, float.
called by function
names in source
program.
.512 Space required for each
input-output field:
single I/O area serves all
files.
.513 Approximate expansion
of procedures:
averages 5 to 6 machine
instructions per
FORTRAN statement.
(**) •
Libra~
.461 Identity:
.462 User restriction:
. 463 Form
Storage medium:
Organization:
.464 Contents
Routines:
Functions:
Data descriptions:
12/62
no.
Program Diagnostics
.451 Tracers:
• 46
assemble and manually
insert into subroutine
tape or deck.
manually insert amended
records.
use of a function name in
an arithmetic statement
causes insertion of the
appropriate routine;
arithmetic and inputoutput subroutines are
inserted automatically.
.52
FORTRAN subroutine tape
or deck.
none •
paper tape or punched
cards.
machine code (written in
mM 1620 SPS language) .
yes; 25 standard closed
subroutines.
yes; 6 standard, up to 19
user-defined.
no.
Translation Time (**)
.521 Normal translating
Card input-output:
PUnched tape
input-output:
.522 Checking only
(FORTRAN
Pre-Compiler):
2+ 0.025S to 3 + 0.125S
minutes, depending upon
statement complexity.
5 + O.OSS to 6 + 0.25S
minutes, depending upon
statement complexity.
maximum of about 35
statements per minute.
PROGRAM TRANSLATOR: FORTRAN I
§
412:183.530
.62
183.
.53
Optimizing Data:
.54
Object Program Performance (**)
Type
Elementary
algebra:
Complex
formulae:
Deep nesting:
Heavy
branching:
Complex
subscripts:
Data editing
(FORMAT):
Overlapping
operations:
none.
Space
Time
unaffected
unaffected.
increased
increased
increased.
increased.
unaffected
unaffected.
Target Computer
.621 Minimum
configuration:
.622 Usable extra
facilities: . .
.7
increased.
increased
unaffected.
COMPUTER CONFIGURATIONS
.61
Translating Computer
. 611 Minimum configuration
Paper tape version: .
Card version:
. 612 Larger configuration
advantages:
1620 Processing Unit,
1621 Paper Tape Reader,
1624 Tape Punch,
Typewriter.
1620 Processing Unit,
1622 Card Read Punch,
Typewriter.
1623 Core Storage unit
permits larger symbol
tables.
©
Check or
Interlock
Missing entries:
Unsequenced entries:
Duplicate statement
numbers:
Improper format:
Incomplete entries:
Target computer
overflow:
Inconsistent program:
Symbol table
overflow:
Mixed mode
expression:
Excessive record
length:
not possible.
.6
all.
(Special subroutine tape
or deck is available for
systems with floating
point hardware.)
ERRORS, CHECKS AND ACTION
Error
increased
any mM 1620 system.
Action
none.
none.
various checks
various checks
type error message.
type error message.
check
none •
type error message.
check
type error message.
check
type error message.
check
type error message.
Note: See also the FORTRAN Pre-Compiler
description in Paragraph 12 .
.8
ALTERNATIVE
TRANSLATORS:
1962 by Auerbach Corporation and BNA Incorporated
none.
12/62
412:184.100
_STANDARD
II
EDP
IBM 1620
REPORTS
Program Translator
FORTRAN II
PROGRAM TRANSLATOR: FORTRAN II
.§
184.
.1
GENERAL
. 11
Identity: .
.12
Description
.22
Form
.221 Input media:
IBM 1620 FORTRAN II.
Program 1620-FO-019.
.222 Obligatory ordering:
This translator converts IBM 1620 FORTRAN II
source programs into machine language object programs in two card passes. The first pass translates the source program into an abbreviated intermediate form, assigns storage locations for variable and constants, and checks for 50 types of
source program errors. The second pass completes the translation process and punches a selfloading object deck. The subroutines necessary for
executing the object program (described in Paragraph .46),can either be included in the object deck
during compilation or loaded separately before execution.
.223 Obligatory grouping:
.3
OUTPUT
Compilation of FORTRAN II source programs requires a 1620 system with at least 40,000 positions
of core storage, a 1622 Card Read Punch, and the
Automatic Divide and Indirect Addressing special
features. The same configuration requirements
apply to execution of FORTRAN II object programs,
except that only 20,000 storage positions may be required. A special subroutine deck is available to
take advantage of the Automatic Floating Point feature.
.31
Object Program
The "trace feature" facilitates debugging by inserting instructions that cause typing of the result of
each executed arithmetic statement and/or the expression calculated in each IF statement. Insertion
and execution of the "tracers" is controlled by Program Switch settings.
.23
.232 Maximum size source
statements: .
.233 Maximum number of
data items:
.311 Language name:
.312 Language style:
.313 Output media:
.33
Originator:
IBM General Products
Division, San Jose,
California.
. 14
Maintainer: .
as above.
.15
Availability:
. June, 1962.
.2
INPUT
.21
Language
.211 Name:
.212 Exemptions:
IBM 1620 FORTRAN II.
none, but note language
limitations in Section
:162.14.
©
limited by target computer
storage.
330 characters.
see 412:162.261.
IBM 1620 machine code.
relocatable.
punched cards.
Conventions
.321 Standard inclusions:
loading routine, arithmetic
and input-output subroutines (both may be loaded
separately at execution
time).
Documentation
Subject
Source program:
Object program:
Symbol table: .
Restart point list:
Language errors:
Compared to the "FORTRAN I" translator described
in Section :183, the FORTRAN II translator offers
expanded language facilities, lower compiling speeds,
and essentially unchanged object program execution
speeds and object program storage utilization.
.13
Size Limitations
.231 Maximum number of
source statements:
.32
punched cards or typewriter keyboard.
all statements in logical
sequence; arithmetic
statement functions must
precede the first executable statement.
none.
Provision
typed listing (optional).
none.
typed listing (optional).
none.
typed messages.
.4
TRANS LATING PROCEDURE
.41
Phases and Passes
Pass 1:
Pass 2:
1962 by Auerbach Corporation and BNA Incorporated
translates source program
to an abbreviated form,
assigns locations for variabIes and constants,
checks for certain errors,
and punches intermediate
output deck.
translates output from Pass
1 into a machine code object program and punches
a self-loading object deck.
12/62
412:184.420
§
IBM 1620
184.
.465 Librarianship (Contd. )
.42
Optional Mode
.421
.422
.423
.424
.425
Translate:. .
Translate and run:
Check only:
Patching:
Updating:
.43
Special Features
.431 Alter to check
only:
.432 Fast unoptimized
translate:
.433 Short translate on
restricted program:
.44
. 45
Bulk Translating::
.452 Snapshots: .
.453 Dumps:
.46
no.
.462 User restriction:
.463 Form
Storage medium:
Organization:
.464 Contents
Routines:
Functions:
Data descriptions: .
.465 Librarianship
Insertion:
Amendment:
Call procedure:
.5
TRANSLATOR PERFORMANCE
.51
Object Program Space
no.
yes, in both Pass 1 and
Pass 2. Symbol table
must be punched out after
Pass 1 when bulk translating; otherwise, this is
unnecessary.
instructions that will cause
optional printing of the
result of each executed
arithmetic statement can
be automatically inserted
at compile time; the
trace facility is controlled
by Program Switch settings.
none.
none.
Translator Library
.461 Identity: .
12/62
no.
Program Diagnostics
.451 Tracers:
Note: The 1620 FORTRAN II subroutine deck contains 27 standard subroutines which are
needed for proper execution of the object
program (arithmetic, input, output, fix, and
float) and seven relocatable subroutines
which are inserted in the object program
only if the functions they implement are used
in the source program (LOGF, EXPF, COSF,
SINF, ATANF, SQRTF, and ABSF). Up to
43 additional functions can be added by the
user. All subroutines can either be punched
into the object deck at compile time or loaded separately at execution time.
yes.
no.
yes, by halting after Pass 1.
no.
no.
FORTRAN II subroutine
library.
none.
.511 Fixed overhead
Name
Space
Comment
Multiply-add
tables:
300 digits.
Standard subroutines:
approx.
arithmetic, input,
12,000
output, fix, float .
digits
Relocatable
subroutines:
varies
called by function
names in source
program .
. 512 Space required for
each input-output
file: . . . . . .
single I/O area serves all
files.
.513 Approximate expansion of procedures:
averages 5 to 6 machine
instructions per FORTRAN
statement .
. 52
.521 Normal translating
Card input-output: .
punched cards.
machine code (written in
illM 1620 SPS language).
yes; 27 standard closed
subroutines.
yes; 7 standard, up to 43
user-defined.
no .
Translation Time (* * )
O. 05S to O. 25S minutes,
depending upon statement
complexity .
Punched tape inputoutput: . . . . .
O.lOS to O. 50S minutes, depending upon statement
complexity.
.522 Checking only: . . . • above translating times are
roughly halved when compilation is halted after
Pass 1.
assemble and manually insert into subroutine deck.
manually insert amended
cards.
use of a function name in an
arithmetic statement
causes insertion of the
appropriate routine; arithmetic and input-output
subroutines are inserted
automatically.
. 53
Note: FORTRAN II translation times for a single
program are about twice as long as FORTRAN
I times (412:184.521), but bulk translating is
not possible in FORTRAN I. Therefore, a
"break-even" point between the 2 systems
will be reached at about 3 programs. When
more than 3 programs are compiled at one
time, FORTRAN II will generally be faster.
Optimizing Data: . . . none.
PROGRAM TRANSLATOR: FORTRAN II
§
412:184.540
184.
.54
.62
. 621 Minimum configuration:
Object Program Performance (* *)
Type
Time
Space
Elementary
algebra:
unaffected unaffected.
Complex
formulae:
increased
increased.
Deep nesting:
increased
increased.
Heavy branching:
unaffected unaffected.
Complex subscripts:
increased
increased.
Data editing
(FORMAT): increased
unaffected.
Overlapping
operations: not possible.
.6
COMPUTER CONFIGURATIONS
. 61
Translating Computer
. 611 Minimum configuration:
..... .
. 612 Larger configuration
advantages:
60,000 core storage positions permit extension of
symbol and name tables.
1620 Processing Unit, 1622 •
Card Read Punch, Automatic Divide, Indirect Addressing.
.622 Usable extra facilities: . . . . . . .
.7
1621 Paper Tape Reader,
1624 Paper Tape Punch,
1623 Core Storage Unit,
Automatic Floating Point
Operations (with special
subroutine deck).
ERRORS, CHECKS AND ACTION
Check or
Error
Interlock
Missing entries:
Unsequenced enuies:
Duplicate statement
numbers:
Improper format:
Incomplete enuies:
Target computer overflow:
Inconsistent program:
Symbol table overflow:
Mixed mode expression:
Invalid literal:
1620 Processing Unit with
1623 Core Storage Unit
(total of 40, 000 core storage positions), 1622 Card
Read Punch, Automatic
Divide, Indirect Addressing.
©
Target Computer
Action
limited checks
none.
type error message.
check
various checks
various checks
type error message.
type error message.
type error message •
check (Pass 2)
limited checks
check
checks (Pass 2)
check
type
type
type
type
type
error
error
error
error
error
message •
message.
message •
message.
message.
Note: All the above checks except "target computer
overflow" and "mixed mode" are made in
Pass 1. In most cases, detection of an error
halts punching of the intermediate output;
but compilation continues to check for additional errors .
.8
ALTERNATIVE TRANSLATORS: . . . . . . none.
1962 by Auerbach Corporation and BNA Incorporated
12/62
412:191.100
•
IJ
STANDARD
ElDP
IBM 1620
Operating Environment
General
"PORTS
OPERATING ENVIRONMENT: GENERAL
§
191.
.1
GENERAL
.11
Identity:
.12
Description
.511 Tracing
All routines used in a program are called by standard or new macro codes and are incorporated into
the program at translation time.
.13
Availability:
library routines described
are available now.
, .14
Originator:
IDM Corporation and individual users.
PROGRAM LOADING
.21
Source of Programs
. 211 Programs from on-line
libraries: •
.212 Independent
programs:
.22
those areas, address of
each instruction and complete instruction is typed
out, as well as addressed
operands. No alternative
operation.
no integrated supervisor
available.
No comprehensive supervisor routine has been announced for the 1620 system. The facilities covered
in this section, therefore, must be provided by the
incorporation in each program of specific routines,
either library routines or individually written
routines.
.2
(Contd. )
Note: additional traCing routines exist in the IDM
Program Library, both more complex and
simpler.
.512 Snapshots: • . . . . . none as such; can use Selective Trace and run program less efficiently.
.52
Post Mortem:. . . . . Post Mortem Dump for Card
1620 (Card), File No.
1. 5. 004. It is loaded after
the running of the main
program. The specified
area of storage is punched
on cards at punch speed
and listed off-line .
.6
OPERATOR
CONTROL:
1.7
none.
punched tape, punched cards,
or typewriter keyboard.
.8
LOGGING: . . . . . . as incorporated in user's
program.
PERFORMANCE
Library
Subroutines:
punched tape or cards.
.81
.23
Loading Sequence:
manually controlled.
.811 Minimum
configuration:
.3
HARDWARE
ALLOCATION:
fixed.
.812 Usable extra
facilities: .
console alarms; typeouts as
incorporated in user's
program.
.813 Reserved equipment:
.4
.5
.51
RUNNING
SUPE RVISION:
PROGRAM
DIAGNOSTICS:
.82
S},:stem Requirements
Dynamic
. 822 Reloading
frequency:
. 511 Tracing
Flow Trace, program
1620-AT-013: . . . typeout of addresses of
:.83
branch instructions and the
addresses branched to.
No alternative operation.
[.84
Selective Trace, program
1620-AT-0l4:
operator specifies which
group or groups of instructions are to be traced. In
©
1620 with Console I/O Typewriter .
paper tape, card input-output .
none.
System Overhead
.821 Loading time:
called in at translation time
if incorporated.
as incorporated in user's
program .
Program Space
Available:
Program Loading
Time:
1962 by Auerbach Corporation and BNA Incorporated
cards: at reader speed.
punched tape: at reader
speed.
must be reloaded for each
new routine if routines are
long.
C-402 decimal digits.
at reader speed: 250 cards/
minute for cards; average
of 148 char/sec. for
punched tape.
12/62
412:192.100
•
STANDARD
EDP
•
REPORTS
IBM 1620
Operating Environment
GOTRAN
OPERATING ENVIRONMENT: GOTRAN
§
192.
.23
Loading Sequence: . . . manually controlled.
Interpreter Input
.1
GENERAL
. 24
.11
Identity: . . . . . . . . GOTRAN for Cards.
Program 1620-PR - 011.
. 241 Language
Name: .
Exemptions:
GOTRAN for Paper Tape.
Program 1620-PR-01O.
.12
Description
GOTRAN.
none (but note severe language restrictions in
Section :163).
.242 Form: . . . • . . . . . paper tape, punched cards,
or typewriter keyboard.
GOTRAN is a "load-and-go" algebraic programming
system that was developed to minimize the overall
.3
time required for program compilation and testing.
It uses a severely restricted subset of the FORTRAN .31
language, as described in Section :163. GOTRAN is
available in two versions for mM 1620 systems with
either paper tape or punched card input-output, and
.32
it can be used on a minimum 1620 system with
20, 000 core storage positions and no optional
features.
.4
The GOTRAN compiler/interpreter program and
subroutines remain in core storage continuously and
occupy 14,470 digit positions. Source programs are
read, compressed, and loaded into core storage. A
typed listing of the source statements is optional.
Source programs can also be entered manually via
the typewriter keyboard. Interpretive execution of
the program can begin as soon as the last source
statement has been processed.
Availability :
February, 1961.
.14
Originator:
mM General Products Division' San Jose, California.
. 15
Maintainer:
.2
PROGRAM LOADING
.21
Source of Programs
.211 Programs from on-line
libraries: .
.212 Independent programs:
. 213 Data: .
. 214 Master routines: .
I.22
Librarr Subroutines:
as above .
none; standard subroutines
for floating point arithmetic and functions are part
of the GOTRAN system.
© 1962
no choice of level; core
storage is always used.
Input-Output Units: . . specified in source program.
RUNNING SUPERVISION
Simultaneous Working:. none.
. 42
Multi-prog!:amming: .
not possible.
.43
Multi - sequencing: . .
not possible.
.44
Errors, Checks, and Action
Error
Loading input error:
In-out error:
Symbol table overflow:
Storage overflow:
Any indicator on:
Arithmetic overflow:
Underflow:
Invalid operation:
Improper format:
Mixed mode arithmetic:
45
none.
paper tape, punched cards,
or typewriter keyboard.
paper tape, punched cards,
or typewriter keyboard.
paper tape or punched
cards.
Storage: . . . . . .
.41
Error checking of the source program is limited,
but fairly thorough arithmetic checks are made during program execution. There are no built- in facilities for diagnostic aids such as tracing, snapshots,
or dumps.
.13
HARDWARE ALLOCATION
!.5
Restarts: • . . . . .
PROGRAM
DIAGNOSTICS: .
by Auerbach Corporation and BNA Incorparated
Check or
Interlock
parity check
parity check
check
nane.
check
check
check
check
none.
Action
halt or set indicator.
halt or set indicator.
type message & halt.
type
type
type
type
message
message
message
message
&halt.
& halt.
& halt.
& halt.
none.
can be restarted at any
time by the console operations STOP, RESET, INSERT, RELEASE,
START. Depending upon
Program Switch 3 setting,
a new source program
can be accepted or the
object program currently
in storage can be re-run
from the beginning.
none, except for the builtin error messages
described in Paragraph
.44.
12/62
412:192.600
§
IBM 1620
192.
none; i. e., optional
features and expanded
core storage cannot be
used without program
modification.
.813 Reserved eqUipment: . 14,470 core storage locations for GOTRAN system, including subroutine
package .
.812 Usable extra facilities:
·6
OPERATOR CONTROL
.61
Signals to Operator:.
. 62
Operator's Decisions:. program switches or console typewriter.
.82
Operator's Signals
.821 Loading time: . .
· 63
· 631 Inquiry: . . . . .
· 632 Change of normal
progress: . . . .
typed messages, provided
automatically for errors,
or incorporated into
source program.
none.
LOGGING: . . . . • . . as incorporated into source
program.
.8
PERFORMANCE
.81
S},:stem Reguirements
.811 Minimum configuration: basic mM 1620 system with
either paper tape or card
input- output.
I
12/62
.822 Reloading frequency:
approx. 2 minutes (for the
GOTRAN system).
can be maintained in core
storage continuously.
by restarting (Paragraph
.45) and/or manually interchanging source
programs.
.7
S},:stem Overhead
. 83
Program Space
Available: . .
.84
Program Loading Time: not available. (Note that
loading indudes translation to compressed form.)
.85
Program Performance in
m. secs: •. . . . . . overall execution times
average 2 to 3 times as
long as for routines compiled by mM 1620
FORTRAN. (***)
14S + 100L + C) .:5 4953,
for typical program of S
source statements, L
labels, and C constants.
412:193.100
II
•
STANDARD
EDE»
"'D.1S
IBM 1620
Operating Environment
FORGO
OPERATING ENVIRONMENT: FORGO
§
193.
.12
.1
GENERAL
• 11
Identity: .
each IF statement. Seven oilier modes of operation
are available for diagnostic and restart purposes .
FORGO.
1620 General Program
Library No.2. O. 008.
When the FORGO system is used in a 1620 with
40,000 storage positions, object program size limitations are rather sever.e: It is estimated that a
typical program with 50 source statements and a
total of 100 variables and constants will fill the
5, i40 available storage locations. Using either the
FOR-TO-GO system in a 40K 1620 or the FORGO
system in a 60K 1620, a program four to five times
as large can be accommodated.
FOR-TO-GO.
1620 General Program
Library No. 2.0.009.
.12
Description (Cont'd)
Description
The FORGO system was developed to serve as an
educational programming system and as a debugging
tool for testing FORTRAN-coded routines before
running them on larger computers. The primary
objects were to minimize the overall time required
for program compilation and testing, and to provide
comprehensive error checks and diagnostic facilities. The result is a "load-and-go" system. The
FORGO program remains in core storage continuously and occupies 34,860 digit positions. Source
programs, in the FORTRAN-like language described
in Section :164, are read from punched cards at the
rate of around 100 statements per minute and simultaneously translated into object program pseudo instructions which are loaded into core storage. As
soon as the last source statement has been translated, interpretive execution of the object program
begins. Execution speeds average about one-half to
two thirds as fast as those of object programs compiled by the ffiM 1620 FORTRAN system.
The FOR-TO-GO system uses virtually the same
language as FORGO, but in order to accommodate
larger object programs, the FOR-TO-GO translator
is ov!'!rwritten by the subroutine package before the
object program is executed. Operation of the FORTO-GO system is therefore less straightforward and
rapid, and batch compiling (a feature of FORGO) is
not possible in FOR-TO-GO.
A 1620 system with Card Read Punch, at least
40, 000 positions of Core Storage, Automatic Divide,
and Indirect Addressing is required to use the FORGO and FOR-TO-GO systems. No provision has
been made for paper tape input and output. Normal
input and output are on punched cards, and it is assumed that an ffiM 407 is available for listing the
output cards. The console typewriter can serve as
an alternate input device.
A wide variety of error checks is made during
source program translation and object program execution. When any source program error is detected, compilation is terminated but the remainder
of the program is checked for additional errors. In
the optional "full trace" mode, the number of every
statement executed is typed, together with the result
of each arithmetic statement, the index value for
each DO loop, and the value of the expression in
©
.13
Availability:
released for general distribution December, 1961.
. 14
Originator: •
Charles W. McClure, Keneth R. Sanderson, and
Joel Davis; Engineering
Computing Laboratory,
University of Wisconsin.
Source of information: . C. H. DaVidson, Associate
Professor, University of
Wisconsin.
.15
Maintainer:
Engineering Computing
Laboratory, University
of Wisconsin.
.16
First Use:
October, 1961.
.2
PROGRAM LOADING
.21
Source of Programs
.211 Programs from on-line
libraries: .
.212 Independent programs:
• 213 Data: •
. 214 Master routines:
.22
Library Subroutines:.
• 23
Loading Sequence
FORGO: . . . . .
1962 by Auerbach Corporation and BNA Incorporated
none.
punched cards.
punched cards (typewriter
keyboard only if control
digit set in memory) .
punched cards.
standard subroutine package is loaded along with
translator in FORGO and
overwrites the translator
in FOR-TO-GO; addition
of machine language
library functions is possible but inconvenient.
FORGO deck.
Program 1 source deck.
Program 1 data.
Program 2 source deck.
Program 2 data.
etc.
12/62
412:193.230
§
IBM 1620
193.
.45
.23
Restarts: (Contd.)
Loading Sequence (Contd.)
FOR-TO-GO A deck (translator).
Program 1 source deck.
FOR-TO-GO B deck (subroutines).
Program 1 data.
(batch operation not
possible).
FOR-TO-GO: . .
• 24
Interpreter Input
.241 Language
Name: .
Exemptions:
FORGO.
none (but note restrictions
relative to mM 709/7090
FORTRAN in Section
412:164).
• 242 Form: . . . . . • . . . punched cards; 1 statement
per card .
.3
HARDWARE ALLOCATION
.31
Storage: . . . . . .
no choice of level; core
storage is always used.
.
Card Read Punch or console typewriter; specified
in source program.
Object program can be
dumped on punched
cards at any time and
restarted from the point
at which it was halted.
·5
PROGRAM DIAGNOSTICS
.51
Dynamic
• 511 Tracing: . . . . . . . . several optional traCing
modes are controlled by
Program Switches and the
typewriter; unusually
comprehensive diagnostic
information can be
obtained.
none.
.512 Snapshots: . . . .
.52
Post Mortem: . •
.6
OPERA TOR CONTROL
.61
Signals to Operator: . . typed messages, provided
automatically by the system or incorporated into
sOllrce program.
.62
Operator's Decisions:. Program Switches or console typewriter.
Operator's Signals
.32
Input-Output Units:
.4
RUNNING SUPERVISION
.41
Simultaneous Working:
none.
. 42
Multi-programming: .
not possible.
.63
.43
Multi - seguencing:
not possible.
.44
Errors, Checks, and Action
.631 Inquiry: . . . . .
.632 Change of normal
progress: . . . .
.
Error
Check or
Interlock
Action
Loading input error:
Statement too complex:
Storage overflow:
Invalid statement:
Undefined variable:
Mixed mode arithmetic:
Improper format:
Arithmetic overflow:
Underflow:
In -out error:
sequence check
check
check
check
check
check
various checks
check
check
check
type message & halt. f
punch message.
punch message. •
punch message. •
punch message. •
punch message. •
punch messa ge. •
type message & halt.
set to zero.
type message 8. halt.
t May resume loading from check point when sequencing is
none .
by restarting (Paragraph
.45) and/or manually interchanging card decks.
·7
LOGGING:....... typed messages and
punched card records,
produced automatically
by the system.
.8
PERFORMANCE
.81
System Reguirements
•
• Detection of any of these errors terminates compilation, but remainder of source program is checked for other errors.
none, but see . 45 above .
· 811 Minimum configuration: 1622 Card Read Punch,
40, 000 core storage positions' Indirect Addressing, Automatic Divide.
corrected.
.45
Restarts:. . . . . . . FORGO and FOR-TO-GO
can be restarted at any
time by the console
operations STOP, RESET,
INSERT, RELEASE,
START. Depending upon
Program Switch 3 setting,
a new source deck can be
read or the object program currently in storage
can be re-run from the
beginning.
12/62
.812 Usable extra facilities:
.813 Reserved equipment
FORGO:
I AUERBACH I
FOR-TO-GO:
tl
60, 000 core storage pos i tions accommodate larger
object programs. FORGO
and FOR-TO-GO are selfadapting, and can use
80,000 or 100,000 pos itions
if they are available.
34,860 core storage
positions.
18, 280 core storage
positions.
OPERATING ENVIRONMENT: FORGO
§
412: 193.820
193.
.82
System Overhead
. 821 Loading time: . .
. 822 R,eloading frequency
FORGO: ••.
FOR-TO-GO: • . . .
. 83
2 minutes. (*) •
(Based on 470 program
cards at 250 cards/min.
Ready to go immediately
if first program deck
loaded behind FORGO
deck.
can be maintained in core
storage continuously.
must be reloaded for each
program: A deck before
source program and B
deck before data.
Program Space Available
FORGO: . • .
FOR-TO-GO:
80S + 100 :::: C - 34, 860.
80S + 100 :::: C - 18, 280.
Note: These are estimates for typical programs of
5 simple source statements and 0 data items,
where C is total number of core storage locations (40,000 or 60,000).
. 84
Program Loading Time: approx. 100 statements per
minute, including translation. (*).
.85
Program Performance
in m. secs(*): . . •
overall execution times average 1. 5 to 2 times as
long as for routines compiled by IBM 1620
FORTRAN. (*).
©
.851 Conditions:. . . . . • none.
.852 For random address (floating point data)
c =a + b: . .
14
b =a + b: . .
19.
Sum N items:
19 + (N-2)4.5.
c = ab: .
26.
c=a/b:.
55.
b = fa:.
530.
b=loga:
340.
b = ea : .
180.
b = sin a:
340 .
.853 For arrays of data
Ci = ai + br . . .
127.
c =C+aiOr . . . . . 122.
. 855 Moving, per item (X=Y): 18 .
.856 Data input, per item
Free format
Overhead:
300.
Time: . .
125.
F format
Overhead:
450.
Time: . .
200.
E format
Overhead:
450.
Time: : .
200.
. 857 Data output, per item
Free format
Overhead:
490.
Time: . .
O.
F format
Overhead:
490.
Time: • .
O.
E format
Overhead:
490.
Time: • .
O.
1962 by Auerbach Corporation and BNA Incorporated
12/62
412:201.001
IBM 1620 Model 1
System Performance
NOTES ON SYSTEM PERFORMANCE
§
201 .
.1
GENERALIZED FILE PROCESSING
Because the IDM 1620's output speed is low on punching and typing alphanumeric data,
it was considered unsuitable for this type of data pr·ocessing application at this.time.
(Where the master file is small enough to be held in internal storage, the 1620 can be
quite useful.)
.2
SORTING·
Magnetic tape is not generally used with the IDM 1620 system .
.3
MATRIX INVERSION
The standard problem estimate of the Users' Guide was used, which is based on the
time for floating point cumulative multiplication .
.4
GENERALIZED MATHEMATICAL PROCESSING
Both fixed point coding and floating point subroutines are timed for Configuration IX.
Floating point calculations are performed by the Automatic Floating Point Operations
special feature in Configuration X. Input is read by the 1621 Paper Tape Reader for
Configuration IX and by the 1622 Card Read Punch for Configuration X.
Results are printed on the on-line typewriter for Configuration IX, and punched on the
1622 Card Read Punch for Configuration X .
.5
GENERALIZED STATISTICAL PROCESSING
Fixed point machine coding is used. Input is read by the 1621 Paper Tape Reader for
Configuration IX and by the 1622 Card Read Punch for Configuration X.
©
1962 by Auerbach Corporation ond BNA Incorporated
12/62
412:201.011
IIS"'' "D
II
EDP
WIlR!S
IBM 1620 Modell
System Performance
IBM 1620 MODEL 1
SYSTEM PERFORMANCE
©
1962 by Auerbach Corporation and BNA Incorporated
12/62
IBM 1620 MODEL 1
412:201.012
IBM 1620 MODEL 1 SYSTEM PERFORMANCE
WORKSHEET DATA TABLE 2
Configuration
-,
Item
Worksheet
Reference
IX
5
Fixed/Floating point
Fixed Point
X
Floating Point-.
Floating Point ••
input
1621 P.T. Reader
1622, Card R'eader
Punch
output
I/O Typewriter
1622 Card Reader
Punch
input
100 digits
100 digits
input
100 digits
100 digits
Unit name
Size of record
Standard
Mathematical
Problem
A
input
Tl
667
307
output
T2
12,000
483
input
T3
667
307
output
T4
12,000
240
m. sec/record
T5
0
0
m.sec/5 loops
T6
m. sec/report
T7
m.sec/block
4:200.413
m.sec penalty
7
Unit name
1,867
0
1621 P.T. Reader
Size of block
Standard
Stati sti cal
Problem
A
678
3
1622 Card Reader
Punch
60 digits
60 digits
1
1
Records/block
B
m.sec/block
Tl
400
3.4
m.sec penalty
T3
400
3.4··
m. sec;block
T5
0
0
m.sec/record
T6
1.92
1.92
m.sedtable
T7
7.71
7.71
4:200.512
C.P•.
.By subroutines.
I
12/62
706
I AUERBAC@
.*By Automatic
Floating Point
Special Feature.
* •• N greater than 31.
412:201.300
•
II
STANDARD
EDP
IBM 1620
System Performance
REI'ORfS
SYSTEM PERFORMANCE
§
.312 Timing Basis:.
201.
.3
MATRIX INVERSION
•31
Standard Problem Estimates
using estimated procedure
outlined in User's Guide,
4:200.312 .
see graph below .
. 313 Graph:. . . . . . • .
.311 Basic Parameters:. . .
general, non-symmetric
matrices, using floating
point to at lea st 8
decimal digits.
100.00
9
7
If
4
I
A II
!.
If
2
II
10.00
I
7
I
4
6
~I
..., :::t
fJ~
~§
CIo t/j
.$ t5
2
:J"&
Time in Minutes for
Complete Inversion
.<:l;
1.00
Q)
§
r;
lr.,Qi
.~.....
fi·.!g
~Jj
lr.,~
urdi
ctI r:::
gos3
ctI
&)
7
'{I
4
J
J
2
I
,
.,
I ,
0.10
7
IJ
1/.8 td
;}ot:J.
-!.t5
;:j:f
~
II
J
4
I
If I
2
/ J"
0.01
1
2
4
7
10
2
4
7
100
2
4
7
1,000
Size of Matrix
© 1963
by Auerbach Corporation and BNA Incorporated
1/63
IBM 1620
412:201.400
§
201.
5 fifth -order polynomials •
5 divisions .
1 square root •
using estimating procedure
outlined in User's Guide,
4:200.413
see graph below .
•412 Computation:.
.4
GENERALIZED MATHEMATICAL PROCESSING
.41
Standard Mathematical Problem A Estimates
• 413 Timing hasis:.
.411 Record sizes:. . • • .
.414 Graph: • . • .
10 signed numbers, avg.
size 5 digits, max. size
8 digits.
Configuration IX; Single Length (8 digit precision); Fixed point.
R
=Number of Output
Records per Input Record
1,000,000
7
4
2
100, 000
7
l..iIlI
4
-- / ,
......
2
1.0
R
Time in Milliseconds 10, 000
per Input Record
~
~
,
~
~ 11""1'
~ I~
~
7
~
l.'
4
2
1,000
.JI' ~
R
=0.1
-
_
-
I- R - O. 01
"-
I
7
...
~
. ""
/'
~
I-"
1/
V
j.- -.100
""""
4
2
100
0.1
2
4
7
1.0
2
4
7
10.0
2
C, Number of Computations per Input Record
1/63
4
7
100.0
412:201.415
SYSTEM PERFORMANCE
§
201.
.415 Graph: . . . . . . . . see graph below.
Configuration IX; Single Length (8 digit precision); Floating point.
R
= Nwnber of Output Records
per Input Record
1,000,000
7
4
2
~
100,000
7
'I
~
4
2
R
Time in Milliseconds
per Input Record
~~
.----
1.0
-
JI
~
10, 000
7
.A
~
-'
"./
V~
V
..... ""
f~
--
/"~
4
/'"
.,,---
2
~
1,000
.-- ~~
R = O.O~jIIII
fI'
--
7
4
2
100
0.1
2
4
7
1.0
2
4
7
10.0
2
4
7
100.0
C, Number of Computations per Input Record
© 1963
by Auerbach Corporation and BNA Incorporated
1/63
IBM 1620
412:201.416
§
201.
.416 Graph: • . . . . . . . see graph below.
Configuration X; Single Length (8 digit precision); Floadng point.
R
=Number of Output Records
per Input Record
1,000,000
7
4
2
100,000
7
1;1
4
IJ
2
/
Time in Milliseconds
per Input Record
10,000
7
2
~
1,000
4
V
,
1.iII"
4
7
,
V
"
-
~",~,....
f\,0.0l,..
~
,~
~
f - ~'"
O· ""
~.-
2
100
0.1
2
4
7
1.0
2
4
7
10.0
2
C, Number of Computations per Input Record
1/63
4
7
100.0
412: 201.500
SYSTEM PERFORMANCE
§
20l.
augment T elements in
cross -tabulation tables .
using estimating procedure
.513 Timing basis:.
outlined in User's Guide,
4:200.513 •
. 514 Graph:. . . • . . . . see below .
.512 Computation:.
.5
GENERALIZED STATISTICAL PROCESSING
.51
Standard Statistical Problem A Estimates
. 511 Record size: . . . . .
thirty 2-digit integral
numbers.
10,000
7
I..!
Ii'!!
~
4
~~
2
~
1,000
~V
V
~
7
I",0Il'
---
IX
4
IIii""
I~
X
/
~
2
Time in Milliseconds
per Record
100
7
4
2
10
7
4
2
1
1
2
4
7
10
2
4
7
100
2
4
7
1,000
T, Number of Augmented Elements
Roman numerals denote Standard Configurations
© 1963
by Auerbach Corporation and BNA Incorporated
1/63
412:211.101
IBM 1620 Modell
Physical Characteristics
IBM 1620 MODEL 1
PHYSICAL CHARACTERISTICS
©
1962 by Auerbach Corporation and BNA Incorporated
12/62
412:211.102
IBM 1620 MODEL 1
IBM 1620 MODEL 1 PHYSICAL CHARACTERISTICS
Central
Processing Unit
Care Storage
Core Storage
Pa~er
Tape eader
Tape Punch-
Card
Read Punch
1620 model 1
1623-1
1623-2
1621
1624
1622
44x63x44*
44x 60x 27
44x 69)(h
44X :tilc 26
*
45x57x30
1,165
810 max.
915 max.
350
?
1,225
---
20
20
10
10
20
Temperature, of.
?
?
?
?
?
?
Humidity, '7.
?
?
?
?
?
?
Temperature, of.
60 to 90
60 to 90
60 to 90
60 to 90
60 to 90
60 to 90
Humidity, '7.
20 to 80
20 to 80
20 to 80
20 to 80
20 to 80
20 to 80
10,000
3,000
4,500
.(1)
(1 )
5,500
Air Flow, cfm.
840
420
700
(1 )
(1)
280
Internal Filters
?
?
?
?
?
?
208/230
,.
,.
,.
,.
*
± 10'1.
,.
,.
,.
,.
,.
60
,.
,.
,.
,.
,.
±0.5
*
*
,.
,.
*
Up, 3 wire
,.
,.
•
•
1.87
.85
1.25
(1)
?
Unit Name
IDENTITY
Model Number
Height x- Width
X
Depth,
I in.
Weight, Ibs.
PHYSICAL.
Maximum Cable Lengths,
Ft.
Storage
Ranges
Working
Ranges
ATMOS.
PHERE
Heat Dissipated, BTU/hr.
Nominal
Voltage
Tolerance
Nominal
ELECTRICAL
Cycles
Tolerance
Phases and Lines
Load KVA
*Inciudes work *From 1620.
shelf.
*From 1620.
NOTES
12/62
I AUERBACH I ~
,.
1.46
• Installed
• From ..1.1,20.
*From 1670
(1) Included in within 1621
1620.
Reader.
(1) Included in
1620.
PHYSICAL CHARACTERISTICS
412:211.103
IBM 1620 MODEL 1 PHYSICAL CHARACTERISTICS-Contd.
Disk
Storage Drive
Unit Name
IDENTITY
Disk
Storag e Dri ve
\
\
Model Number
1311- 3
1311- 2
38X43X24
38X30X24
700
350
\
Height X Width XDepth,
in.
Weight, lbs.
PHYSICAL
Maximum Cable Lengths,
Ft.
20 to 1620;
10 to power
receptacle
Temperature, of.
?t
?t
?t
?t
Storage
Ranges
Humidity,
or.
Temperature, of.
60 to 90
60 to 90
20 to SO
20 to SO
4,000
2,000
Working
Ranges
ATMOSPHERE
Humidity,
or.
Heat Dissipated, BTU/hr.
Air Flow, cfm.
?
?
Internal Filters
?
?
Nominal
20S/230
from 1311- 3
Voltage
± 10or.
Tolerance
Nominal
ELECTRI·
CAL
60
Cycles
±0.5
Tolerance
Phases and Linea
Load KVA
t
NOTES
from 1311- 3
HP, 3 wire
1cp, 3 wire
1.4
0.75
disk pack:
40-12a'F
10-S0or.
©
t
disk pack:
40-l20oF
10-S0or.
1962 by Auerbach Corporation and BNA Incorporated
12/62
412:221.101
•
STANDARD
EDP
•
IBM 1620 Modell
"'ORTS
Price Data
PRICE DATA
§ 221.
IDENTITY OF UNIT
PRICES
CLASS
Name
No.
CENTRAL
PROCESSOR
1620
1623
$
Purchase
$
76.75
64,000
2.75
1. 00
2,400
670
135
8.50
1. 50
1. 75
.50
.75
1.00
6.75
12,400
1,150
2,400
1,300
450
500
6,400
750
1,200
27.25
33.50
37,100
58,700
375
665
29.00
55.50
17,000
29,250
490
190
25
615
13.75
4.75
50.00
8,650
1,400
30,000
Core Storage
20,000 digits
40,000 digits
Disk Storage
Disk Storage Drive
Disk Storage Drive
Disk Pack
Modell
Model 2
1311
Model 2
Model 3
1621
1624
1622
Raper Tape Reader
Tape Punch
Card Reader Punch
Calcomp 560- R Digital Recorder
-
NOTE:
$
Optional Features
Automatic Divide
Additional Instructions
Automatic Floating· Point
Operations
Indirect Addressing
Core Storage Adapter
Core Storage Adapter
Paper Tape Reader Adapter
Card Read Pun'ch Adapter
Disk Storage Drive Adapter
4650
2301
2302
5514
1632
3339
INPUT-OUTPUT
Monthly
Maintenance
Central Processing Unit, including
basic storage, Console and Input/
Output Typewriter.
1285
1021
1288
STORAGE
Monthly
Rental
Y
176 hours per month usage.
*
See manufacturer (California Computer Products, Inc.).
Price is approximately $3,300 purchase.
©
1962 by Auerbach Corporation and BNA Incorporated
y
1,375
55
30·
225
25
50
25
10
10
*
*
12/62
IBM 1620
Model 2
International Business Machines
Corp~
'.-- ..
-,'
,/'-
(
~-
AUERBACH INFO, INC.
PRINTED IN U. S. A.
IBM 1620
Model 2
International Business Machines Corp.
/
AUERBACH INFO, INC.
PRINTED IN U. S. A.
413:001.001
IBM 1620 Model 2
Contents
CONTENTS
1.
2.
3.
4.
5.
6.
7.
8.
10.
11.
12.
13.
14.
15.
Introduction
.
•
Data Structure • •
System Configuration
Configuration IX; Desk Size Scientific
Configuration X; Punched Card Scientific
Internal Storage
1625
Core Storage.
Auxiliary Storage
1311 Model 3
Disk Storage Drive
1311 Model 2
Disk Storage Drive
Central Processor.
•
Console.
•••
. •
Input-Output; Punched Tape and Card
1621
Paper Tape Reader
5515
Paper Tape Reader Adapter
1624
Tape Punch
5515
Adapter. •
1622
Card Read Punch (Reader)
1633
Card Read Punch Adapter
1622
Card Read Punch (punch)
1633
Card Read Punch Adapter
Input~Output; Printers
731
I/O Console Typewriter
Input-Output; Other
Calcomp 560-R Digital Recorder
Sim ultaneous Operations
Instruction List
.
•
Instruction List Addendum
Coding Specimens
SPS
•
FORTRAN I
FORTRAN II
GOTRAN
Data Codes
Internal Numeric
Internal Alphameric .
I/O Console Typewriter (numeric)
I/O Console Typewriter (alphameric)
Paper Tape Input-Output . .
Card Input-Output (numeric) ••
Card Input-Output (alphameric) .
Problem Oriented Facilities
IBM 650 Simulator Program
1710 Simulator /7090
1620 5-Channel Tape Translation Program.
Floating point function subroutines
Matrix Inversion (Tape)
•
. ••
©
1962 by Auerbach Corporation and BNA Incorporated
413:011
412:021
t
413:031
413:031
413:041
412:042
412:042
413:051
412:061
t
412:071
412:071.4
412:072
412:072.4
412:073
412:073.4
412:074
412:074.4
t
t
t
t
t
t
t
t
t
t
413:081
412:101
412:111
412:121
413:121
t
412:131
412:132
412:133
412:134
t
412:141
412:142
412:143
412:144
412:145
412:146
412:147
t
t
412:151. 11
412:151. 12
412:151. 15
412: 151. 171
412:151.172
t
t
t
t
t
t
t
t
t
t
t
t
t
t
t
12/62
413:001.002
IBM 1620 MODEL 2
CONTENTS-Contd.
15.
16.
17.
18.
19.
20.
21.
22.
Problem Oriented Facilities (Contd.)
Complex FORTRAN (Tape) .
SPS to FORTRAN conversion
Format Control Subroutines for
1620 Card FORTRAN
Interpretive systems • • .
Other ••
AUTOMAP .
Process Oriented Languages
FORTRAN I
FORTRAN II
GOTRAN . . •
FORGO •
Machine Oriented Languages
SPS
SPS I-Pass
Program Translators
SPS • . • .
SPS I-Pass
FORTRAN I
FORTRAN II
Operating Environment
General •
GOTRAN
FORGO •
System Performance
Notes on System Performance
Worksheet Data • • • • • •
Matrix Inversion • • • • . •
Generalized Mathematical Processing •
Generalized Statistical Processing
Physical Characteristics
Price Data • • • • • • • • . • • • • • • . • • .
412:151.173 t
412:151.174 t
412:151.175
412:151.176
412:151.177
412:151. 3
t
t
t
t
412:161
412:162
412:163
412:164
t
t
t
t
412:171
412:172
t
t
412:181
412:182
412:183
412:184
t
t
t
t
412:191
412:192
t
41~:193
413:201. 001
413:201. 011
413:201. 3
413:201.4
413:201. 5
413:211
413:221
t Refer to indicated section of computer System Report 412: IBM 1620 ModelL
12/62
t
t
413:011.100
IBM 1620 Model 2
Introduction
INTRODUCTION
§Oll.
The IBM 1620 Model 2 is a faster version of the IBM 1620. Processing Unit times
to do fixed and floating point operations and floating point subroutines have been reduced to
one-fourth of the time required in model 1. The higher speed is found in the Processing
Unit, Core Storage, and Console I/O Typewriter; other input-output equipment remains unchanged. Some of the optional Special Features available to the 1620 Modell are standard
inclusions on model 2. Model 2 can perform all programs written for model 1 without reprogramming. A list of specific changes from model 1 is given at the end of this Introduction.
The 1620 Model 2 is a solid-state desk size computer oriented toward scientific applications. The basic system consists of the 1620 Model 2 Central Processing Unit and
Console, and the Input-Output Console Typewriter, used for input with hard copy and for
output. Typewriter output occurs at about fifteen characters per second. This minimum
configuration, including core storage of 20,000 decimal digits, rents for $2,200 per month.
The processor performs the two-address instructions sequentially. Data processing
is performed serially by digit on variable length decimal fields; no input-output radix conversion is required. Alphameric data may be input and output; each alphameric character
is stored internally as a pair of decimal digits. Instructions are fixed in length at twelve
digits. A digit consists of four numeric bits, one check bit, and one flag bit used for storing the Sign of a numeric field and for delimiting a field. The core store has a 10 microsecond read-restore cycle. Fixed-point addition of two fields is performed at 15 microseconds per digit; field movement also requires 15 microseconds per digit. Each digit in
storage is individually addressable. Core storage is expandable to a total size of 40,000
or 60,000 digits.
Punched tape and card equipment can be used with the 1620. Paper tape may be input at 150 rows per second or output at 15 rows per second; no buffering is available. Two
independent buffered card channels are available, permitting reading at 250 cards per
minute and punching at 125 cards per minute. A few 1620 Modell installations have installed Model 7330 Magnetic Tape Units and Model 1403 Line Printers on an RPQ basis.
A new disc storage system has been announced for the IBM 1620, 1401, 1440, and
1710 systems. This system is the 1311 Disk Storage Drive, and features interchangeable
Disk Pack units as a replaceable storage medium. The peak transfer rate is 50,000 digits
per second when used in the 1620 system.
Each 1311 Disk Storage Drive holds one Disk Pack at a time, providing on-line
storage for 2,000,000 digits per drive in addressable sectors of 100 digits each. A maximum of four drives can be connected. Up to 20,000 digits can be read or recorded without movement of the 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. Disk Storage seek operations (but not read -write operations) can be overlapped with other system functions. IBM has announced programs utilizing the disc file for
SPS and FORTRAN II, and alsoadisk Utility Program to aid in the maintenance of programs
or data in disc storage.
Each Disk Pack consists of a stack of six discs with ten magnetiC recording surfaces
and a cover that forms a sealed container when the Disk Pack is not mounted on a drive.
Diameter is 14 inches, height is 4 inches, weight is less than 10 pounds, and time to interchange two Disk Packs is about one minute.
©
1962 by Auerbach Corporation and BNA Incorporated
12/62
413:011.101
IBM 1620 MODEL 2.
INTRODUCTION-Contd.
The basic 1620 Model 2 includes 42 instructions, many of which allow the use of
literals (the "Immediate" instructions, as Add Immediate). Fixed point multiply and divide
are provided. Floating point operations are performed by subroutines, or by extra cost
floating point instructions. The input-output operations transfer groups of characters
rather than a sing'le character or word, and no radix conversion is needed since data is
stored in decimal form. Punched tape operations are terminated by a delimiter code, and
card operations are stopped after 80 characters are transferred. Card operations are
checked by comparing the buffer contents to information read at a checking station.
The assembly language for the 1620 is the 1620/1710 Symbolic Programming System,
which includes macros for floating point subroutines for arithmetic and mathematical functions. The translation may be performed on the 1620 or on the 709/7090 system.
Problem oriented facilities are oriented towards industrial design applications,
mathematical applications, and linear programming. They include AUTOMAP, a program
for machine tool control.
FORTRAN I and II are available for mathematical program writing. The languages
have some restrictions relative to 709/7090 FORTRAN II; see sections :161 and :162. With
minor changes, 1620 FORTRAN source programs can be compiled and executed on the IBM
7070, 704, and 709/7090 systems. FORTRAN I compiling should be faster on Model 2 1620
than on Model '1, since compiling was generally processor bound on Modell.
IBM provides the GOTRAN interpretive system based on restricted FORTRAN language statements. The number of arithmetic operations allowable in single statement is one.
The FORGO interpretive system, developed at the University of Wisconsin Engineering Computing Laboratory, is more useful than GOT RAN . See Sections: 163 and: 164 and the associated Operating Environment sections for descriptions of these two systems.
The 1620 Modell Processor is used as the digital computer in the IBM 1710 Control
System, used in process control and data collection applications. 1620 Model 2 is not available for the 1710 system.
Changes from 1620 Modell
Core Storage read-restore cycle changed from 20 to 10 microseconds.
Add table replaced by decimal adder.
Digits transferred in fewer memory cycles.
I/O Console Typewriter replaced by Model 731 typewriter, a modified IBM Selectric.
Standard inclusions:
Indirect Addressing feature
Automatic Divide feature
Additional Instructions (Transfer Numerical Fill, Transfer Numerical Strip,
Move Flag).
Transmit Floating instruction.
New Instructions:
Transmit Record No Record Mark (record mark not transferred).
Branch and Select (makes Indirect Addressing feature active or inactive).
12/62
/'
413:031.100
IBM 1620 Model 2
System Configuration
SYSTEM CONFIGURATION
§
031.
.1
DESK SIZE SCIENTIFIC SYSTEM; CONFIGURATION IX
Deviations from Standard Configuration: . . . . . . .
Rental:
core storage larger by 8,000
decimal digits.
printer faster by 2.5 lines/minute.
paper tape input faster by
140 rows/sec.
paper tape output faster by
5 rows/sec .
. . . . . . . . . . . . . . . . . • . . . . . . $ 3,090 per month.
Core Storage, Model 1625- 2
(40,000 digits).
Processor, Console, and
Input- Output Typewriter.
Punched Tape Reader and Controller.
Tape Punch and Controller.
Optional Features Included: . . . . . . . . . . . . . . . . Paper Tape Reader Adapter.
©
1962 by Auerbach Corporation and BNA Incorporated
12/62
IBM 1620 MODEL 2
413:031.200
§
031 .
.2
PUNCHED TAPE/CARD SCIENTIFIC SYSTEM; CONFIGURATION X
Deviations from Standard Configuration:
no index register.
printer faster by 2.5 lines/minute.
Rental:
$ 4,275 per month.
Core Storage, Model 1625-3
(60,000 digits).
Processor, Console and InputOutput Typewriter.
Card Read Punch and Controller.
Optional Features Included: . . . . . . . . .• . . . . . . . Automatic Floating Point Feature.
Card Read Punch Adapter.
12/62
413:041.100
IBIA 1620 Model 2
Internal Storage
Core Storage
ItHERNAL STORAGE: CORE STORAGE
§
041.
.1
.16
Reserved Storage
GENERAL
Purpose
.11
Identity:
.12
Basic Use: .
.13
Description
Core Storage.
Model 1625-1, 1625- 2,
1625-3.
Number of
locations
Multiply table:
Product or Dividend
working area:
200
20**
working storage.
Card Load area:
Core Storage for the 1620 Model 2 system is very
similar to storage for Model 1. However, its readrestore cycle time has been halved, and the previous add table area is available for general storage.
All core storage is now externalto the 1620 Processor cabinet.
Core storage is located in a separate cabinet which
can contain any of the three models. Model 1625-1
contains 20, 000 decimal digits of storage; model
1625- 2 contains ·kO, 000, and model 1625-3 contains
60, 000. One of the three models must be specified
for any 1620 Model 2 system. Each location stores
one decimal digit and is individually addressable by
a five-decimal-digit address code. In the computer
numeric mode, one location stores one decimal
digit; and in the alphameric mode, two digit locations are used to represent either an alphabetic
character, special character, or decimal digit. One
digit consists of four BCD numeric bits, one odd
parity check bit, and one flag bit. Storage cycle
time is 10 microseconds. Core storage uses "wraparound" addressing: address 00000 follows the
highest-numbered address when incrementing addresses; the highest-numbered address (19999 for
example) follows 00000 when decrementing addresses.
Core storage is used for all input-output areas, instructions, and working storage. Instructions require 12 digits of storage. Basic storage includes a
200-digit reserved area for a multiplication table.
No lock is provided to protect this area; it is used
for storing the loader routines when a program is
being loaded. Power may be turned on and off without loss of information in storage when following
normal operating procedures.
Operands (fields and records) may be any length desired. Numeric fields are delimited by a flag bit in
the most significant digit position, while complete
records are delimited by a record mark code.
.14
AVailability: .
4th Qtr., 1963
. 15
First Delivery:
see above.
CD
80
Locks
no.*
no; may also be
used as working storage.
no; used with
card reader
Load key; may
be used as
working
storage.
* Used for loader routine before table is inserted.
** With Automatic Divide.
.2
PHYSICAL FORM
.21
Storage Medium: •
. 22
Physical Dimensions
. . magnetic core.
.221 MagnetiC core storage
Array size:. . . .
100 by 100 by 12 bits.
.23
Storage Phenomenon: .
direction of magnetization.
.24
Recording Permanence
.241 Data erasable by
instructions:
. 242 Data regenerated
constantly:
. 243 Data volatile:
.244 Data permanent:
. 245 Storage changeable: .
.28
coincident current.
same as recording .
uniform.
Potential Transfer Rates
.292 Peak data rates
Cycling rate: .
Unit of data:
Conversion factor:
Data rate:
Compound data rate: .
• .3
no .
no (with normal power
on/off procedures).
no .
no.
Access Technigues
.281 Recording method:
.282 Reading method:
. 283 Type of access:
.29
yes .
100; 000 cycles/sec.
2 digits.
6 bits per digit.
200, 000 digits/sec.
200, 000 digits/sec .
DATA CAPACITY
1962 by Auerbach Corporation and BNA Incorporated
12/62
413:041.310
§
IBM 1620 MODEL.2
041.
.53
Maximum
Storage
Mininum
Storage
Identity:
Characters:
Instructions:
Digits:
Modules:
.32
Model
1625-1
10,000
1,666
20,000
1
Rule for Combining
Modules: . . . .
Model
1625-2
20,000
3,333
40,000
2
Model
1625-3
30,000.
5,000.
60,000.
3.
Access Time Parameters and Variations
.531 For uniform access
Access time:
Cycle timc: . .
For data unit of:
.6
CHANGEABLE
STORAGE:
.7
PERFORMANCE
.71
Data Transfer
CONTROLLER
.41
Identity: . . .
")
•...l:"'-
Transfer Load Size
With self: . . . . .
no separate controller.
.73
With self: . . . . . .
.8
Error
Check or
Interlock
Invalid address:
parity and limit check
In valid code:
Receipt of data:
parity only .
~arity check
Recording of data:
Recovery of data:
records parity bit.
parity check
Dispatch of data:
Timing conflicts:
Reference to locked
transmits parity bit.
no conflicts.
Connection to Device
Data Transfer
Control: . . .
.5
ACCESS TIMING
.51
Arrangement of
Heads: . .
.52
by processor.
. . . 1 access device.
Simultaneous
Operations: . . . . . none.
area:
I
12/62
66, 667 digits/sec.
ERRORS, CHECKS AND ACTION
.431 Devices per controller:. 1.
.432 Restrictions: .
1 module type.
• 44
one -half of storage.
Effective Transfer Rate
Connection to System
.422 Off-line: . . . . . . . none.
. i3
. . . none.
Pair of storage units possibilities
With self: . . . . . . yes
all combinations are
listed above.
.72
.4
lOp sec.
10 I.' sec.
1 digit.
I. . .A-U-ER-BA-CH--L-.~
none.
Action
indicator, alarm,
halt.
indicator, alarm,
system halt
optional.
indicator. alarm,
system halt
optional.
413:051.100
•
III
STANDARD
ElDP
IBM 1620 Model 2
Central Processor
REPORTS
CENTRAL PROCESSOR
§
.12
051.
.1
GENERAL
. 11
Identity: .
.12
Description
Computer.
1620 Model 2
The 1620 Model 2 Computer is a two-address sequential processor oriented towards scientific applications, and is offered as an optional version of
the 1620. It is a faster processor than 1620 Model
1, and offers programming compatibility with model
1 programs. Functionally, model 2 is almost identical to model 1.
Computation speed has been improved by a factor of
four over model 1. A number of optional Special
Features for model 1 are standard on model 2; other
Special Features available for model 1 are also
available with model 2. A list of changes in model 2
is given at the end of this description.
Operands are held as variable-length strings of decimal digits, delimited by a flag bit in the six-bit code
of the most significant digit of the field. Data may
also be stored in alphameric fields, two digits per
character. Instructions are twelve digits long and
are performed sequentially. Data operations are
performed serially by digit upon the operands, which
may be any length. The Central Processor cabinet
includes the Console, desk work area with Console
I/O Typewriter, and space for adapters used with
input-output devices.
Internal records are defined by the presence of a
record mark code digit. Records, as well as fields,
may be moved within core storage. The record
mark terminates a write operation on punched tape.
Record marks are generated in storage by the following: end of record on punched tape; Record Mark
key on Console I/O Typewriter; record mark as data
on cards or punched tape.
Although the core storage cycle is 10 microseconds,
instruction times are a good deal longer because of
the serial operation. Add-Subtract-Compare-Move
operations all require 140 microseconds for fivedigit operands.
Fixed point Add, Subtract, Multiply, and Divide are
standard instructions in the Central Processor.
Automatic Floating Point Operations Special Feature,
available at extra cost, provides fast floating point
arithmetic. A complete set of conditional branch
instructions is standard, and allows branching on
.13
presence or absence of the condition specified. Four
sense switches are available for interrogation. The .14
©
Description (Cont'd)
Branch and'Transmit instruction provides a jump to
the P-address, and also transmits the field specified
by the Q-address to the storage area immediately
preceding the P-field. This field may contain parameters needed by the subroutine starting at P.
Note that the optional features listed in the description of the Central Processor of the 1620 Modell are
all standard inclusions in model 2, except the Autom'atic Floating Point Operations. These features,
now standard, are Automatic Divide, Indirect Addressing, Transfer Numerical Fill, Transfer Numerical Strip, and Move Flag. Transmit Floating
is also standard in model 2.
The set of arithmetic, comparison, and data movement instructions includes instructions for handling
literals. For example, there is Add, and Add Immediate. The Add Immediate instruction is helpful
in address modification, since there is no index address modification in the 1620. The normal Add instruction could be used, however, to add a constant
to an address since the operating time of Immediate
instructions is no faster than the operating time of the
normal instructions. No editing instruction exists,
but the typewriter can be commanded to space, tabulate, or perform a carriage return. Normally,
carriage return occurs automatically at the end of
each line. Comparisons are performed digit-bydigit, whether the field is numeric or alphameric.
Two new instructions are included in the repertoire,
Branch and Select, and Transmit Record No Record
Mark. Branch anti Select allows the Indirect Addressing feature to be made operative or inoperative; a normal power-on operation causes Indirect
AddreSSing to be operative. Use of this instruction
also causes an unconditional transfer to the P address of the instruction. The Transmit Record No
Record Mark instruction moves a record in storage,
but the terminal record mark delimiter digit is not
moved.
Changes from 1620 Modell
Add tables replaced by decimal adder.
,Core storage placed in separate cabinet.
Optional features made standard as described above.
Transmit Floating instruction is standard.
Data movement requires fewer memory cycle times
than previously.
AVailability: . .
4th quarter 1963.
First Delivery:
see above.
1962 by Auerbach Corporation and BNA Incorporated
12/62
IBM 1620 MODEL 2
413:051.200
§
051.
.2
PROCESSING FACILITIES
. 21
Operations and Operands
Operation and
Variation
.211 Fixed point
Add -subtract:
Multiply
Short:
Long:
Divide
No remainder:
Remainder:
Provision
Radix
Size
automatic
decimal
2 to N digits,
limited by
core storage.
none.
automatic
decimal
2 to N digits.
none.
automatic
decimal
2 to N digits.
subroutine
optional feature
subroutine
optional feature
subroutine
optional feature
subroutine
optional feature
decimal
decimal
decimal
decimal
decimal
decimal
decimal
decimal
2
2
2
2
2
2
2
2
.212 Floating point
Add-subtract:
Multiply:
Divide:
Shift:
.213 Boolean:
. 214 Comparison
Numbers:
Absolute:
Letters:
Mixed:
none.
2 to N digits.
automatic
Done.
1 to N char.
1 to N char.
automatic
automatic
Collating sequence: special symbols, A to I,
.215 Code translation:
. 216. Radix conversion:
· 217 Edit format: .
.218 Table look-up:
.22
to 45 digits.
to 100 digits.
to 45 digits.
to 100 digits.
to 45 digits.
to 100 digits.
to 45 digits.
to 100 digits.
0, J to z,
0 to 9.
none .
none.
alter size by set flag
command.
none.
Special Cases of Operands
.221 Negative numbers: . . . flag bit over least significant digit of decimal
field.
.222 Zero: . . . • . . . . . . both + and - zero. Signs
ignored when comparing
2 all-zero fields.
.223 Operand size
determination:
flag bit over most Significant digit of decimal
field.
.23
Instruction Formats
· 231 Instruction structure:
• 23 2 Instruction layout:
Part
Size (digits)
· 233 Instruction parts
Name
Operation:
.P-Address:
12/62
12 digits.
Q-Address
.233 Instruction parts (Contd.)
Name
Purpose
P-Address (Contd.) 4. address for transfer; ot"
5. input-output starting
address .
Q-Address: . . . . . 1. operand adc\ress;
.2. operand literal;
3. starting address for
transfer;
4. address of testable indicators;
5. typewriter control function; or
6. select input-output
device.
.
. 234 Basic address
structure: .
2-address.
.235 Literals
Arithmetic: .
any; 5 practically.
Comparisons and
tests: . . . .
any; 5 practically.
Incrementing
modifiers:
no modifiers (no indexing);
however, use of literals
in arithmetic useful for
incrementing addresses .
. 236 Directly addressed operands
.2361 Internal storage type: core storage .
Minimum. size: . ..
2 digits.
Maximum size: . . .
complete store.
Volume accessible: .
complete store.
.2362 Increased address
capacity:
none.
.237 Address indexing: .
none; use of literals in
arithmetic useful for incrementing addresses.
.238 Indirect addressing
.2381 Recursive: .
yes.
.2382 Designation:
flag bit in least significant
digit of address .
. 2383 Control: . . .
absolute address has no
flag bit .
. 2384 Indexing with indirect
addressing: . .
not possible (no indexing).
.239 Stepping: . . . . .
none.
.24
Special Processor
Storage:
. . . none.
.3
SEQUENCE CONTROL FEATURES
.31
Instruction Sequencing
.311 Number of sequence
control facilities:
.314 Special sub-sequence
counters: . . . .
. 315 Sequence control
step-size:
.316 Accessibility to
routines: . . .
5
Purpose
specifies operation to be
performed.
1. operand address in core
storage;
2. address of result of addition or subtraction;
3. jump address;
1.
none .
1 instruction (12 digits).
by BT instruction; can store
next address in sequence;
address used by BB
instruction.
.317 Permanent or optional
modifier: .
none.
.32
Look-Ahead:
none.
.33
Interruption:
none.
.34
Multi -running:
none.
413:051.350
CENTRAL PROCESSOR
§
.422 For arrays of data (Contd.)
051.
.35
Multi - sequencing:
.4
PROCESSOR SPEEDS
. 41
Instruction Times in IJ secs
.428
Floating point subroutines called by macros;
8 decimal digits
ci =ai + bf .
10,200 (*) .
6, SOO (*) .
bj = ai + bf .
Sum N items:
6,700 (*).
c=c+aibj: . . . . . 15,SOO(*) .
Floating point using Automatic Floating Point
Operations special feature; S decimal digits
ci = ai + bj : .
961.
756 .
bj = ai + bf .
Sum N items:
661.
c = c + aibf . . . . . 4,050.
Branch based on comparison
Numeric data: .
615 + 15D; D = no. digits
in key.
Alphabetic data:
615 + 30C; C = no. alpha
char in key.
Switching
Unchecked:
260.
Checked: .
590.
List search:
. 535N + 210.
Format control per character
Unpack
Scientific:. . . . . . 0; usable in input area.
Compose
Scientific
Fixed point: . . . . 23; for moVing data.
Floating point: . . 27; for moving data.
Table look up per comparison
For a match: . . . . . 535.
For least or greatest: 740.
For interpolation
point: . . . . . .
535.
Bit indicators
Set bit in separate
location:
so.
Test bit in separate
location:
70 .
65 + 15D.
Moving: . . . . . . .
.5
ERRORS, CHECKS, AND ACTION
. none.
.411 Fixed point
Add - subtract:
65 + 15D.
Multiply: . . .
160 + 10D + 40~.
Divide: . . . .
60 + 135D + 9SD
.412 Floating point
Using subroutines called
by macros; S digits
Add-subtract (FA,
FS): . . . . .
2,400 (*).
4,500 (*).
Multiply (FM):
Divide (FD): . .
11,000 (*).
Using Automatic Floating Point
Operations special feature
Add-subtract
(FADD, FSUB): 150 + 22D.
Multiply (FMUL): 2S0 + 30D + 40D2.
Divide (FOIV): . . 345 + 2700 + 9SD2.
.413 Additional allowance for
Indirect addressing:
30.
Re-complementing:
100 (fixed or floating
point) .
. 414 Control
Compare: .
65 + 15D.
Branch: ..
60 .
. 415 Counter control:
no counters .
.416 Edit: . .
no edit instruction .
. 417 Convert:
no convert instruction .
. 41S Shift: ..
no shift instruction.
.423
.424
.425
.426
.427
.42
Processor Performance in J.t secs
.421 For random addresses
Fixed point
c = a+ b: . .
130+ 30D.
b = a + b: . .
65 + 15D.
Sum N items:
65 + 15D.
c = ab:
225 + 25D + 40D2
c = alb: . . .
300 + lS0D + 9s02 .
Floating point subroutines called by macros;
8 decimal digits
c = a + b: . .
10,000 (*).
b=a+b: . .
6,300(*).
Sum N items:
6,300 (*).
c = ab:
12,300 (*).
c = alb: . . . . . . . 24,600 (*).
Floating point using Automatic Floating Point
Operations special feature; S decimal digits
c = a + b: . .
541.
b = a + b: . .
326.
Sum N items:
326.
c = ab:
3, 295.
c = alb: . . .
8,960.
. 422 For .arrays of data
Fixed point
750+ 30D.
ci = ai + bf .
590+ 15D.
bj = ai + bj: .
Sum N items:
400+ 15D.
730 + 180D + 9802
c = c+ aibf.
© 1963
Error
Check or
Interlock
Overflow:
check
indicator, alarm: halt
or programmed action.
Exponent underflow and
overflow (float-pt):
check
indicator, alarm: halt
or programmed action.
Zero divisor:
Invalid data:
overflow check.
parity check
Invalid operation:
Arithmetic error:
Invalid address:
none.
parity check
Receipt of data:
parity check
Dispatch of data:
parity check
Incorrect operand
lengrh:
Mispositioned divisor:
overflow check.
overflow check,
by Auerbach Corporation and BNA Incorporated
Action
alarm, indicator: halt
or programmed action.
alarm, indicator: halt
or programmed action •
alarm, indicator: halt
or programmed action.
alarm, indicator: halt
or programmed action.
3/63
413:081.100
.SIANDARD
II
EDP
' ' ' IS
IBM 1620 Model 2
Input-Output
I/O Typewriter
INPUT-OUTPUT: CONSOLE I/O TYPEWRITER
§
081.
• 24
•1
GENERAL
• 11
Identity:
.12
Use of station: .
Stacks: . • . .
Heads/stack:
Method of use: •
. . . • . • I/O Typewriter, Model 731.
Part of 1620 Console.
. 24
Description
The Console I/O Typewriter is a modified IDM Selectric typewriter which stands upon the console
desk. It is usable only with the 1620; no off-line use
is possible. It types output data under program control, and provides input under program request or
by operator initiation (console Insert key). Up to
100 characters may be inserted into storage by operator initiation, and any number of characters by program request for typewriter data. The record mark
may be inserted in storage by a typewriter key but is
treated only as a data character. The decimal pOint
(period) can be inserted correctly into storage only
with the use of a Read Alphamerically instruction.
A flag key is used to indicate a negative number, or
a flag bit, during numeric output.
An output command transfers any number of characters to the typewriter from storage, terminated by a
record mark. Both input and output data transfers
are parity-checked by the 1620. Incorrect characters (parity or invalid codes) are marked specially
as they are typed.
. 13
Availability:. .
• 14
First Delivery:
. 4th Qtr., 1963 •
printing.
1.
1.
1 character at a t_ime •
Arrangement of Heads (Cont'd)
Use of station: .
Stacks: . . • .
Heads/ stack:
Method of use: .
_25
keyboard input.
1.
44 keys.
1 character at a time.
Range of Symbols
Numerals:
Letters: . .
Special: •.
Alternatives: .
FORTRAN set:
Req. COBOL set: _
Total: • • . • .
*
PHYSICAL FORM
• 21
Drive Mechanism
. 211 Drive past the head: .
• 212 Reservoirs: • . • • •
• 22
Sensing and Recording Systems
.221 Recording system:
• 222 Sensing system:
. 223 Common system:
. 23
friction drive .
none.
. 231 Maximum number
Interleaved carbon:
.233 Types of master
Multilith:
Spirit:
.
Form of Storage
.311 Medium:
.
.
. 312 Phenomenon: .
.
continuous fanfold stationery .
typeface impression .
Positional Arrangement
.33
Coding:
,
.34
Format
Compatibili~:
.35
Phlsical Demensions
.352 Length: • • • • • •
.353 Maximum margins:.
CONTROLLER
.41
Identity: . • .
depends on stationery.
· 42
Connection to Slstem
yes.
yes.
· 421 On-line:
.' 422 Off-line:
© 1962
52
.31
.4
Multiple Copies
none.
yes.
no.
EXTERNAL STORAGE
• 351 Overall width: . . .
engraved rotating moving
typing element.
typewriter keyboard for
manual input .
no •
*
0-9.
A- Z.
including control codes not used as data.
.321 Serial by: .
. 324 Track use
Data:
. 325 Row use:
•2
10
26
16.
.3
. 32
see above.
Arrangement of Heads
by Auerbach Corporation and BNA Incorporated
character at 10 per inch .
85 print pOSitions •
all for data.
as in Data Code Table No.
5.
. none •
8. 5 inch line of print;
platen size may be 9.375
or 10. 875 inches.
no limit •
no limits.
no separate controller; part
of 1620 Console .
1.
not usable off-line •
12/62
IBM 1620 MODEL 2
413:081.430
§
.55
081.
· 43
Connection to Device
Control Operations
Disable:
Request interrupt:
no.
no .
none .
• 431 Devices per controller:. 1.
. 432 Restrictions: • . . • . none.
.56
Testable Conditions: .
• 44
.6
PERFORMANCE
.61
Conditions: .
.62
Speeds
Data Transfer Control
· 441 Size of load
Input: . • .
.442
• 443
• 444
• 445
• 446
no limit when requested by
program; 100 char when
manually initiated by Insert key.
Output: • • . . • • . any size up to limit of storage; terminated by record
mark.
Input-output areas: .
. core storage.
Input- output area
char.
access: . • • .
Input- output area
lockout: . • .
yes.
Table control: .
no.
Synchronization:
automatic.
.5
PROGRAM FACILITIES.AVAILABLE
;51
"mocks
.511 Size of block:
console Release and Stan
key, or 100th char.
record mark in storage.
Output:
• 52
Input-Output Operations
.521 Input:
• 522 Output: .
.523 Stepping: •
.
• 524 Skipping: .
.525 Marking: •
.526 Searching:
• 53
.54
12/62
Code Translation: .
.621 Nominal or peak speed:
.624 Effective speeds:
.63
.
15.5 char/sec for output;
manual typing speed for
input.
same as peak speeds, less
allowance for carriage
returns.
Demands on System
m.sec per
char
or Percentage
100
or
100 .
Component
Processor:
.7
EXTERNAL FACILITIES
.71
Adjustments: • • • . • typical typewriter adjustments.
.72
Other Controls
same as load size; see. 441
above.
• 512 Block demarcation
Input: •
none.
Function:
Form: • • •
Comment:
.73
end input and sgtrt computer•
momentary key.
R-S key.
Loading and Unloading
.731 Volumes handled: • . • depends on feed facilities .
input 1 block into core
storage •
output 1 block from core
storage, with automatic
carriage returns.
step 1 or 2 lines at end of
printed line; set by
operator.
none.
none.
none.
automatic; data stored dependent on mode of operation (numeric or
alphameric) .
Format Control: . . . . fixed format; automatic
carriage return at end of
each line. Carriage return may also be
programmed.
•8
ERRORS, CHECKS AND ACTION
Error
Check or
Interlock
Action
Recording:
parity check at 1620
Reading:
lnput area
overflow:
parity check at 1620
indicator. alarm; overprints a bar thru
character.
indicator. alarm •
Output block
size:
Invalid code:
Exhausted
medium:
Imperfect
medium:
Timing
conflicts:
Dispatch of
data:
count of 100 max. char
on manual Insert
operation.
any size possible.
check when printing
special char printed.
none.
none.
interlock
attach parity bit.
wait.
413: 121.1 00
IBM 1620 Model 2
Instruction List
INSTRUCTION LIST: ADDENDUM
§
121.
GENERAL
.1
The 1620 Model 2 contains all the instructions for the 1620 Modell (see Section
412:121) and has two new ones in addition. They are given below.
OPCODE
Numeric
Mnemonic
.2
BS
60
TRNM
30
Branch and Select.
Makes Indirect addressing
feature active or inactive,
under program control,
using this BS instruction.
Transmit Record No Record
Mark.
Record moved internally as
in Transmit Record except
that the delimiting Record
Mark is not transmitted .
A number of optional instructions available to the 1620 Modell are now standard inclusions. They are
D
DM
LD
LDM
MF
TFL
TNF
TNS
.3
Description
Divide
Divide Immediate
Load Dividend
Load Dividend Immediate
Move Flag
Transmit Floating
Transmit Numerical Fill
Transmit Numerical Strip
The optional instructions for the 1620 Modell which remain optional for model 2 are
listed here for convenience.
FADD Floating Add
FSUB Floating Subtract
FMUL Floating Multiply
FDIV
Floating Divide
FSL
Floating Shift Left
FSR
Floating Shift Right
BTFL Branch and Transmit Floating
©
1962 by Auerbach Corporation and BNA Incorporated
12/62
413:201.001
STANDARD
IBM 1620 Model 2
REPORTS
System Performance
NOTES ON SYSTEM PERFORMANCE
§
201.
•1
GENERALIZED FILE PROCESSING
Because the IBM 1620's output speed is low on punching and typing alphanumeric data,
it was considered unsuitable for this type of data processing application at this time.
(Where the master file is small enough to be held in internal storage, the 1620 can be
quite useful. )
.2
SORTING
Magnetic tape is not generally used with the IBM 1620 system •
.3
MATRIX INVERSION
The standard problem estimate of the Users' Guide was used, which is based on the
time for floating point cumulative multiplication .
•4
GENERALIZED MATHEMATICAL PROCESSING
Both fixed point coding and floating point subroutines are timed for Configuration IX.
Floating point calculations are performed by the Automatic Floating Point Operations
special feature in Configuration X. Input is read by the 1621 Paper Tape Reader for
Configuration IX and by the 1622 Card Read Punch for Configuration X.
Results are printed on the on-line typewriter for Configuration IX, and punched on the
1622 Card Read Punch for Configuration X.
.5
GENERALIZED STATISTICAL PROCESSING
Fixed point machine coding is used. Input is read by the 1621 Paper Tape Reader for
Configuration IX and by the 1622 Card Read Punch for Configuration X.
©
1962 by Auerbach Corporation and BNA Incorporated
12/62
413:201.011
IBM 1620 Model 2
System Perfonnance
IBM 1620 MODEL 2
SYSTEM PERFORMANCE
©
1962 by Auerbach Corporation and BNA Incorporated
12/62
413:201.012
IBM 1620 MODEL 2
IBM 1620 MODEL 2 SYSTEM PERFORMANCE
WORKSHEET DATA TABLE 2
Configuration
Item
Worksheet
Reference
X
IX
5
F.ixed/Floating point
Fixed Point
Floating Point **
Floating Point ***
input
1621 P.T.
Reader
1621P.T.
Reader
1622 Card Reader
P.unch
output
I/O Typewriter
I/O Typewriter
1622 Card Reader
Punch
input
100 digits
100 digits
100 digits
output
100 digits
100 digits
100 digits
Unit name
Size of record
Standard
Mathematical
Problem
A
input
T1
667
667
480
8,000
8,000
960
667
667
7
8,000
8,000
7
m.sec/block
4:200.413
output T2
input
T3
m. sec penalty
output T4
7
m.sec/record
TS
0
0
0
m. see/5 loops
T6
170
290
181
m.sec/report
T7
0
0
1
Unit name
1621 P.T.• Reader
60 digits
Size of block
Standard
Stati sti cal
Problem
A
60 digits
1
1
T1
400
3.4
T3
400
3.4*
m.sec!b1ock
T5
0
0
m. sec/record
T6
0.740
0.740
m. sec/table
T7
2.255
2.255
subroutines.
* N greater than 104.
*** By Automatic
Floating Point
Special Feature.
Records/block
B
m.sec/block
m.sec penalty
4:200.512
C.P.
** By
I
12/62
1622 Card Reader
Punch
IAUERBAC~
413:201.300
IISTAN"RD
II
EDP
IBM 1620 Model 2
System Performance
REPORTS
SYSTEM PERFORMANCE
.312 Timing Basis:. . . • .
201.
§
.3
MATRIX INVERSION
• 31
Standard Problem Estimates
.311 Basic Parameters:. . .
using estimated procedure
outlined in User's Guide,
4:200.312
see graph below .
.313 Graph:. . . • . . . .
general, non -symmetric
matrices, using floating
point to at least 8
decimal digits.
100.00
7
I
4
J
J IA
2
~
J
10.00
I....
.$
I
7
'(
18~
II~
4
~b
JJ
JJ
'tt~'
~
Q;
Ce.$
tI
:;:: I<:,
fQ j/:>
2
§Jj
,$;-C;i
Time in Minutes for
Complete Inversion
§
I
f<.,~
1.00
~c8--
II
7
III
I
I
I
4
I I
I
J
I
2
J
0.10
7
,
I
4
A
If
2
~
/ IIIf
0.01
2
1
7
2
4
7
10
2
100
4
7
1,000
Size of Matrix
© 1963
by Auerbach Corporation and BNA Incorporated
1/63
IBM 1620 MODEL 2
413:201.400
§
201.
.4
. 41
. 412 Computation:.
5 fifth -order polynomials .
5 divisions .
1 square root.
using estimating procedure
outlined in User's Guide,
4:200.413
see graph below •
. GENERAl.JZED MATHEMATICAL PROCESSING
.413 Timing basis: .
Standard Mathematical Problem A Estimates
.411 Record sizes:. . . . • 10 signed numbers, avg.
size 5 digits, max. size
8 digits.
.414 Graph: • • • .
Configuration IX; Single Length (8 digit precision); Fixed point.
R
=Number of Output Records per Input Record
.1,000,000
7
4
2
100,000
7
4
2
R
Time in Milliseconds 10,000
per Input Record
7
= 1.0
-----
~
~
~
""~~~
"
~.J
I/~
/. ~
4
--"
2
1,000
R
0.1
R
0.0
~
7
l....oo
iI'"
~
~
~~~
V/
V
~i""
4
2
100
2
0.1
4
2
7
1.0
4
7
2
10.0
C, Number of Computations per Input Record
1/63
4
7
100.0
SYSTEM PERFORMANCE
§
413:201.415
201.
.415 Graph: . . . . . . . . see graph below.
Configuration IX; Single Length (8 & 2 digit precision); Floating point.
R = Number of Output Records per Input Record
1,000,000
7
4
2
100,000
7
4
2
Time in Milliseconds
per Input Record
10,000
R
/
V
1.0
-,' ',;
~
If
1..1
'I
7
Ai
.AT
W
4
2
1,000
o.
-'o.t
.",.".
R -
~
R
~
,."
..,.
io"~
~
~~
"
7
4
2
100
2
0.1
4
2
7
1.0
4
7
2
10.0
4
7
100.0
C, Number of Computations per Input Record
© 1963
by Auerbach Corporation and BNA Incorporated
1/63
413:201.416
§
IBM 1620 MODEL 2
201.
.416 Graph: • . . . . • . . see graph below.
Configuration X; Single Length (8 & 2 digit precision); Floating pOint.
R
= Number of Output Records
per Input Record
1,000,000
7
4
100,000
7
4
2
"
10,000
Time in Milliseconds
per Input Record
7
I
/
4
~
/
2
IfIf
R=1.0
1,000
7
I
R =1 0• 1, 0.01
4
2
100
2
0.1
4
7
2
1.0
4
7
2
10.0
C, Number of Computations per Input Record
1/63
4
7
100.0
SYSTEM PERFORMANCE
§
413: 20 1.500
201.
.512 Computation: .
.5
GENERAliZED STATISTICAL PROCESSING
.51
Standard Statistical Problem A Estimates
. 511 Record size: . . • . .
augment T elements in
cross-tabulation tables.
using estimating procedure
outlined in User's Guide,
4:200.513
see below .
.513 Timing basis: .
.514 Graph: . • . . . . .
thirty 2-digit integral
numbers.
10,000
7
4
2
~~,
~
1,000
,,-
7
~
~
f
1..00
IX
~ ~'
.J
4
/
/'
X
2
Time in Milliseconds
per Record
100
7
4
2
10
7
4
II>
2
1
1
2
4
7
10
2
4
100
2
4
7
1,000
T, Number of Augmented Elements
Roman numerals denote Standard Configurations
© 1963
by Auerbach Corporation and BNA Incorporated
1/63
413:211.101
IBM 1620 Model 2
Physical Characteristics
IBM 1620 MODEL 2
PHYSICAL CHARACTERISTICS
©
1962 by Auerbach Corporation Qnd BNA Incorporoted
12/62
413:211.102
IBM 1620 MODEL.2
IBM 1620 MODEL 2 PHYSICAL CHARACTERISTICS
Unit Name
Central
Processing Unit
Core
Storage
Core
Storage
Core
Storage
Paper
Tape
Reader
Tape
Punch
Card
Read
Punch
Disk
Storage
Drive
Disk
Storage
Drive
Model Number
1620
Mode12
1625-1
1625-2
1625-3
1621
1624
1622
1311-3
1311-2
44x60X27
44X60x27
44X60X27
44X31x26
*
45x57x30
38x43X24
38X30x24
1,115
1,200 max.
350
?
1,125
700
?
?
10
10
20
20 to 1620;
10 to power
receptacle
IDENTITY
Height x Width X Depth, in. 44x63x44*
Weight, Ibs.
1,200
985
Maximum Cable Lengths
---
?
350
PHYSICAL
Temperature,oF.
?
?
?
?
?
?
?
?t
?t
Humidity, %
?
?
?
?
?
?
?
?t
?t
Temperature, of.
60 to 90
60 to 90
60 to 90
60 to 90
60 to 90
60 to 90
60 to 90
60 to 90
60 to 90
Humidity, %
20 to 80
20 to 80
20 to 80
20 to 80
20 to 80
20 to 80
20 to',80
20 to 80
20 to 80
10,000
4,200
5,200
6,200
..
t
5,500
4,000
2,000
Air Flow, efm.
840
?
?
?
*
t
280
?
?
Intemel Filters
?
?
?
?
?
?
?
?
?
208/230
*
*
*
t
*
..
208/230
From
1311-3
± 10'7.
*
*
..
t
*
..
± 10'7.
Storage
Ranges
Working
Ranges
ATMOSPHERE
HeatDissipated, BTU/hr.
Nominal
Voltage
Tolerance
Nominal
ELECTRICAL
Phases and Lines
Load KVA
12/62
..
*
*
t
*
*
60
±0.5
..
*
*
t
*
*
±0.5
1CP, 3 wire
..
*
*
t
*
*
1.5
0.6
0.8
1.0
*
?
1.46
* From
* From
* From
* Included
1620.
1620.
* Installed
within
1621
Reader.
t Included
in 1620.
From
1311-3
Cycles
Tolerance
NOTES
60
* Includes
work
shelf.
1620.
in 1620.
tFrom
1620.
* From
1620.
1CP, 3 wire 1CP, 3 wire
1.4
0.75
t diBkpack: tdiskpack:
40-120oF 40-120oF
10-80'7.
10-80'7.
413:221.101
.SI"DARD
EDP
_
IBM 1620 Model 2
Price Data
R£PORIS
§
221.
PRICE DATA
PRICES
IDENTITY OF UNIT
CLASS
Nc.
CENTRAL
PROCESSOR
1620
Model 2
Name
Central Processing unit" Console
and Input/Output Type\triter .
Monthly
Rental
$
Monthly
Maintenance
$
Purchase
$
1,200
53.75
52,600
250
25
10
135
9.00
1.50
1. 25
6.75
13, 750
1,250
500
6,400
1,000
1,650
2,200
25.00
32.00
38.50
46,500
79,000
106,500
665
375
55.50
29.00
29,250
17,000
490
190
25
615
13.75
4.75
50.00
8,650
1,400
30,000
Optional Features
1289
5515
1633
3340
STORAGE
1625
Modell
Model 2
Model 3
1311
Model 3
Model 2
1621
1624
1622
INPUTOUTPUT
-
NOTE:
Automatic Floating Poillt Operations
Paper Tape Reader Aclli.pter
Card Read Punch AQaPQer
Disk Storage Drive Adapter
Core Storage
20,000 digits
40,000 digits
60,000 digits
Disk Storage
Disk Storage Drive
Disk Storage Drive
Disk Pack
Paper Tape Reader
Tape Punch
Card Reader Punch
Calcomp 560-R Digitill\ecorder
.!I
176 hourEj Pe!r month usage.
*
See Manufacturer (California Com puter Products, Inc.).
Price is I'4PPIloximately $3,300 purchase.
©
1962 by Auerbach Corporation and BNA Incorporated
*
*
12/62
IBM 1440
International Business Machines Corp.
~
..
(
(
(
AUERBACH INFO, INC.
PRINTED IN U. S. A.
..
IBM 1440
International Business Machines Corp.
AUERBACH INFO, INC.
PRINTED IN U. S. A.
414:001.001
IBM 1440
Contents
CONTENTS
001.
§
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
14.
15.
16.
Introduction...................
Data Structure • . . • • . . . . • . . • • . • .
System Configuration • • . . . . . • . . • . . .
Table of System Configurations
I
Typical Card System • . . • • .
4-Tape Business System . . • .
IID
6-Tape Business System • . . •
IIID
6-Tape Auxiliary Storage System
VD
Internal Storage
1441
Core Storage
1311
Disk Storage
1301
Disk Storage
Central Processor
1441
Processing Unit
Console
1447
Console.. . • •
Console Printer
Input-Output; Punched Tape and Card
1442
Card Read Punch
1442
Card Reader
1444
Card Punch
Input-Output; Printers
1443
Printer..
Input-Output; Magnetic Tape
7335
Magnetic Tape Unit
Input-Output; other
7080
Serial Input-Output Adapter
1009
Data Transmission Unit
1011
Paper Tape Reader . . • . •
1012
Tape Punch • • • • • • • .
1412
Magnetic Character Reader
3271
Direct Data Transfer • . .
3845
Expanded Serial Input-Output Adapter
1062
Teller Terminal • • • . . .
1448
Transmission Control Unit
357
Data Collection System
1031
Data Collection System
1050
Data Communication System
1060
Data Communication System
1231
Optical Mark Page Reader .
7770
Audio Response Unit . . • •
Communication Control System
7740
Simultaneous Operations •.
Instruction List . • • . • .
Data Codes • . • • • • • • •
Problem Oriented Facilities
Sort 5
Basic Report Program Generator
Report Program Generator
Disk Utility Programs . • • • . .
Tape Utility Programs • . • • . •
Disk File Organization Programs
Auto-Test • • • . . • •
Application Programs
Process Oriented Languages
COBOL-61 "
FORTRAN IV •
© 1964 Auerbach Corporation and Info. Inc.
414:011
414:021
414:031
414:031. 011
414.031.1
414:031. 2
414:031. 3
414:031.4
414:041
414:042
414:043
414:051
414:061
414:061.13
414:071
414:071
414:072
414:081
414:091
414:101
414:101. 121
414:101.122
414:101. 123
414:101.125
414:101. 127
414:101. 128
414:102
414:103
414:103.121
. 414:103.122
414:103.123
414:103.124
414:104
414:105
414:106
414:111
414:121
414:141
414:151.13
414:151.14
414:151.14
414:151.15
414:151.15
414:151.16
414:151.17
414:151.17
414:161
414:162
4/64
414:001.002
IBM 1440
CONTENTS (Contd.)
§ 001.
17.
20.
21.
22.
Machine Oriented Languages
Basic Autocoder • . . . . .
Autocoder
•.....
Input-Output Control System
System Performance
Physical Characteristics
Price Data • • . . . . . • •
RIP = Report in process.
4/64
414:171
414:171
414:171
RIP
414:211
414:221
414:011.100
IBM 1440
Introduction
INTRODUCTION
!i 011.
The IBM 1440 is a small-scale, stored-program data processing system that is
adaptable to a wide range of business applications. It features the 1311 Disk Storage Drive
with interchangeable Disk Packs. The 1440 was announced in October, 1962, and the first
customer deliveries were made in April, 1963. Monthly system rentals range from about
$1,800 to $10,000, with most installations falling within the $2,000 to $5,000 range.
Compatibility
The 1440 is the lowest-priced member of the IBM 1400 series of business-oriented
data processing systems. It is program-compatible with (and 3. 5% faster than) the widelyused IBM 1401 with respect to internal processing, but the 1440 uses slower card readers,
card punches, and printers than the 1401 and different instructions to control them. Whereas
the 1401 has fixed input-output areas in core storage that must be used for all printer and card
input-out, the 1440 can initiate direct transfers of data between any area of core storage and
any peripheral device. Through the use of the new 7335 Magnetic Tape Unit, the 1440 can be
made tape-compatible with other IBM and competitive computers that use one-half inch tape at
a recording density of 556 characters per inch.
Hardware
An IBM 1440 system can have from 2,000 to 16,000 alphameric character positions
of core storage. Each core position contains six data bits, a parity bit, and a word mark bit
that is used to denote the end of a variable-length field. Core storage cycle time is 11.1
microseconds (compared to 11.5 microseconds in the IBM 1401, 6.0 in the 1460, and 4. 5 in
the basic 1410).
The 1441 Processing Unit uses add-to-storage logic and has no accumulator. All
operations are performed serially by character, and both data fields and instructions can be of
variable length. The basic instruction format consists of a 1-character operation code and
two 3-character addresses. Operand length is not specified in 1440 instructions; instead, most
operations are terminated when a word mark bit is sensed in the operand itself. Facilities for
editing, high-loW-equal comparisons, and full-record internal transfers are standard, but
multiplication, diviSion, indexing, and sense switches are extra-cost options. Instructions are
executed at the rate of about 4, 000 per second in typical 1440 routines.
Operation of the 1440 system is basically serial in nature (i. e., one operation at a
time). Little overlapping of input-output operations with one another or with internal processing is possible, except that printing is buffered if the optional Print Storage feature is installed. Available computing time per card read on the 1442 Card Read Punch is increased if
less than the full 80 columns are read. Disk Storage seek operations (but not read-write
operations) can be overlapped with one another and with other system functions.
Each 1311 Disk Storage Drive holds one Disk Pack at a time, providing on-line
storage for 2,000,000 alphameric characters in addressable sectors of 100 characters each.
A maximum of five drives can be connected. Up to 20, 000 characters per drive can be read
or recorded without movement of the 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. With the optional Track Record feature, a single 2, 980-character record can be
recorded on each track, increasing the capacity of a Disk Pack to 2,980, 000 characters.
Each Disk Pack consists of a stack of six discs with ten magnetic recording surfaces
and a cover that forms a sealed container when the Disk Pack is not mounted on a drive.
Diameter is 14 inches, height is 4 inches, weight is less than 10 pounds, and time to interchange two Disk Packs is about one minute.
@1964 Auerbach Corporation and Info, Inc.
4/64
IBM 1440
414:011.1 01
§
011.
INTRODUCTION (Contd.)
In addition to the 1311 Disk Storage Drives, up to five arrays of 1301 Disk Storage
can now be attached to a 1440. The 1301 discs are not interchangeable, but they use the same
record formats as the 1311: each track can hold twenty addressable 100-character sectors or
(with the optional Track Record Feature) a single 2, 543-character record. There are four
models of 1301 Disk Storage: Model 11 or 21 has one disc array and one access mechanism,
and stores up to 25.43 million characters. Model 12 or 22 has two disc arrays, each with an
independent access mechanism, and stores up to 50.86 million characters. Total waiting time
for access to a randomly placed record averages 177 milliseconds; up to 80,000 characters per
disc array can be read or recorded without movement of the comb-like access mechanism, with
an average waiting time of only 17 milliseconds. A 1440 system with the maximum complement
of five 1311 drives and five 1301 arrays would have a total of 142 million character positions of
random access storage.
The basic card input-output unit is the 1442 Card Read Punch, which has a photoelectric read station followed by a punch station on a single card feed path. Models 1 and 2 have
rated'speeds of 300 or 400 cards per minute for reading and 80 or 160 columns per second for
punching, respectively. Results can be punched into the same cards from which the input data
was read, but a single 1442 cannot handle separate input and output card files at the same time.
The 1442 Model 4 Card Reader provides read-only ability at 400 cards per minute, while the
1444 Card Punch has an output speed of 250 full 80-column cards per minute.
The 1443 Printer uses a horizontal typebar that permits rapid interchangeability of
character sets. Two models are available, which differ only in speed. Peak speeds are 150
or 240 lines per minute with the standard 52-character set and 430 or 600 lines per minute
with an optional 13-character set. The optional Print Storage feature provides virtually complete overlapping of printing with other system functions.
The 7335 Magnetic Tape Unit is a recent addition to the IBM line that expands the
range of applications for the 1440 by giving it a limited capability for magnetic tape operations.
Only one 7335, consisting of a tape control and either one or two tape transports, can be used
in a 1440 system. The 7335 is similar to the 7330 Magnetic Tape Units used in other IBM
systems but has only one recording density (556 characters per inch) and one speed (20,016
characters per second).
The 1447 Console houses the operating controls and displays and, optionally, a console typewriter rated at 14. 8 characters per second. Anyone of the following peripheral
devices can be connected to a 1440 through the Serial Input/Output Adapter: a 500 characterper-second paper tape reader, a 150 character-per-second tape punch, a 950 document-perminute magnetic character reader, a 75 to 300 character-per-second data transmission unit, an
optical mark page reader, or a 7740 Communication Control System. A 1448 Transmission
Control Unit can be attached to the 1447 Console, and the 1448 can control a large network of
remote data communication or data collection units. The new 7770 Audio Response Unit, which
provides human-voice replies to telephoned inquiries, can be connected through the 1311 Disk
Storage Drive, Model 2.
Software
The programs and programming systems supplied by IBM for the 1440 include:
4/64
•
1440 Basic Autocoder: the basic symbolic assembly system, usable on
card-only 1440 systems with 4, 000 core storage positions, 1442 Card
Read Punch, and 1443 Printer. Two special versions of the Basic Autocoder system are available; one uses the 1444 Card Punch for object pro~
gram output and the other uses a restricted form of the language and requires only 2,000 core positions for assembly.
•
1440 Autocoder: a more advanced symbolic assembly system that provides
macro instruction facilities and is similar to (but not program-compatible
with) the Autocoder languages available for the other IBM 1400 series
systems. A 1440 with at least 4, 000 core positions, one 1311 or 1301
Disk Storage Unit, a card read punch, and a printer is required. (Programs
written in 1440 Autocoder can alternatively be assembled on a 6-tape IBM
1401 system by means of a special version of the 1440 Autocoder translator. )
•
1440 Input-Output Control System (lOCS): macro instructions and corresponding generalized routines which can be used in Autocoder programs to
factilitate the coding of input-output operations. Individual versions of
1440 lOCS are available for systems using 1311 Disk Storage, the 1448
Transmission Control, and the Direct Data Channel, and for 1440 systems
connected on-line to an IBM 1410 or 7010.
414:011.1 02
INTRODUCTION
§
INTRODUCTION (Conld.)
011.
•
1440 COBOL: translates programs coded in a restricted version of COBOL61 into symbolic form for Autocoder assembly. IBM 1311 Disk Storage is.
utilized as the primary file storage medium instead of magnetic tape,
requiring several additions to the COBOL-61 language. Required are
4,000 core positions, one 1311 drive, a 1442 Card Read Punch, a 1443
Printer, and the Indexing and Store Address Register feature.
•
1440 FORTRAN IV: compiles engineering and scientific programs written
in FORTRAN IV. The only important FORTRAN IV language facilities not
implemented in the 1440 version are complex and double precision operations and the Assigned GO TO statement. Precision is variable; operand
lengths can be preset at up to 20 digits. Machine requirements are 8, 000
core positions, one 1311 Disk Storage Drive, a 1442 Card Read Punch, a
1443 Printer, and several optional features. Magnetic tape units can be
used by the object program but are not required.
•
1440 Basic Report Program Generator: a "load-and-go"generator for
card-only 1440 systems that facilitates the preparation of printed reports
from data in punched cards .
.
'
1440 Report Program Generator: uses 1311 Disk Storage to generate
programs to produce reports from input files in punched cards or Disk
Storage.
•
1440 Sort 5: generates routines that utilize 1311 Disk Storage to sort
blocked, fixed-length records.
•
Auto-Test: an intElgratEld set of utility programs designed to expedite the
testing and debugging of programs coded in 1440 Autocoder. At least one
1311 Disk Storage Drive and 8,000 core storage positions are required.
•
1440 Disk File Organization Programs: thirteen programs designed to
assist in establishing and maintainmg data files in 1311 Disk Storage, in
either random or sequential arrangements.
•
Disk and Tape Utility Programs: a variety of routines to perform frequentlyneeded functions such as data transcription and label handling.
•
Application Programs: documented programs to perform specific industry
applications, which can be modified to fit individual users' requirements.
Included are programs for Demand Deposit Accounting, Mortgage Loan
Accounting, File and Casualty Insurance, Weekly Premium Insurance,
Secondary School Systems, Hospitals, Motor Freight Accounting, On-Line
Savings Accounting, Retail Accounts Receivable, Retail Fashion Inventory
Control, and Chain and Wholesale Grocery Billing.
© 1964 Auerbach Carporation and Info, Inc.
4/64
414:021.1 00
IBM 1440
Data Structure
DATA STRUCTURE
§
021.
.1
STORAGE LOCATIONS
Name of Location
Size
Purpose or Use
Character position:
8 bits (6 data,
1 parity, 1 word
mark)
100 characters
20,000 characters
basic addressable location; holds
1 character.
Sector
Cylinder:
(1311)
Cylinder:
(1301)
Column:
.2
80,000 characters
12 hole positions
record location in Disk Storage.
volume accessible without
repositioning in 1311 Disk
Storage.
volume accessible without repositioning in 1301 Disk storage.
punched cards.
INFORMATION FORMATS
TyPe of InformatIon
Representation
Numeral: . • • . . . . . . . . • . . . • . . .
Letter: • . . • . . . . • . . • . . . . . . . . .
mstruction: . . . . . . . . . . . • . . . . . .
Number (field): • . . . . . • . . .- .••..
Block: • . • . . . . . • • . • . . . • . . . . .
1
1
1
1
1
character or 1 card column.
character or 1 card column.
to 8 characters, delimited by word mark bit.
to N characters, delimited by word mark bit.
to N characters, delimited by record or group
mark.
where N is limited by size of core storage.
@1964 Auerbach Corporation and Info, Inc.
4/64
414:031.011
IBM 1440
System Configuration
SYSTEM CONFIGURATION
§ 031.
Table of Permissible Configurations
Model
A2
Model
A3
Model
A4
Model
A5
Model
A6
2,000
1
4,000
1
8,000
1
12,000
1
16,000
1
2
2
2
1
2
2
2
1
2
2
2
1
2
2
2
1
2
2
2
1
1443 Printer, Model 1 t
1443 Printer, Model 2 t
1
1
1
1
1
1
1
1
1
1
1311 Disk Storage Drive @
1301 Disk Storage (arrays)@
0
0
5
5
5
5
5
5
5
5
1447
1051
1448
7335
7770
Console, Model 2 or 4
Control Unit
Transmission Control
Magnetic Tape Unit, Modell or 2
Audio Response Urtit
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1009
1011
1012
1231
1412
7740
3271
Data Transmission Unit, or
Paper Tape Reader,' or
Tape Punch, or
Optical Mark Page Reader, or
Magnetic Character Reader, or
Communication Control, or
Direct Data Channel
0
1
1
1
1
Bit Test
Expanded Print Edit
Indexing
Multiply-Divide
Sense Switches
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
Card Image
Punch Column Skip
Add'l Print Positions (24)
Print Storage
Selective Character Set
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
Direct Seek
Scan Disk
Seek Overlap
Track Record
Buffer Feature (1447-2/4)
no
no
no
no
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
1441 Processing Unit Model:
Core Storage Positions
1447 Console, Model 1
(required unit)
Maximum number of peripheral units:
1442
1442
1442
1444
Card Read Punch, Model 1*
Card Read Punch, Model 2*
Card Reader, Model 4*
Card Punch*
Availability of Special Features:
1
0
0
0
~
* A maximum of two card I/O unIts (1442 Modell, 2, or 4, and 1444) in any combination can
be attached, except that only one 1444 can be used.
t Only one printer can be attached to a system.
@ Both 1311 and 1301s can be attached to one system.
© 1964 Auerbach Corporation and Info, Inc.
4/64
414:031.1 00
IBM 1440
System Configuration
SYSTEM CONFIGURATION
Ii 031.
.1
TYPICAL. CARD SYSTEM: CONFIGURATION I
Deviations from Standard Configuration:
printer is 76% slower.
reader is 60% slower.
punch is 25% faster.
Equipment
1441 Processing Unit,
Model A4:
8,000 core positions
1447 Console, Model 1
1442 Card Reader,
Model 4:
400 cards/min.
80
200
1444 Card Punch:
250 cards/min.
375
1443 Printer, Model 2:
240 lines/min.
5567 Printer Control
450
Adapters:
Card Read Punch
Printer
Optional Features Included:
$ 1,295
235
20
25
Indexing and Store Address
Registers
90
Multiply-Divide
325
Sense Switches
15
Expanded Print Edit
20
Print Storage
~
TOTAL:
© 1964 Auerbach Corporation and Info, Inc.
$ 3,295
4/64
414;031.200
IBM 1440
§
031.
.2
4-TAPE BUSINESS SYSTEM; CONFIGURATION IID
Deviations from Standard Configuration;
printer is 52% slower.
reader is 20% slower.
punch is 150% faster.
3 Disk Storage Drives are used
in place of 4 tape units.
Equipment
1441 Processing Unit, Model A4:
8,000 core positions
1447 Console, Model 1
80
200
1444 Card Punch:
250 cards/min.
375
1443 Printer, Model 2:
240 lines/min.
5567 Printer Control
450
235
1,355
Adapters:
Card Read Punch
Printer
20
25
Sense Switches
15
TOTAL:
4/64
$ 1, 295
1442 Card Reader, Model 4:
400 cards/min.
1311 Disk Storage Drives (3) and
3321 Disk Storage Control
Optional Features Included:
Rental
$ 4, 050
SYSTEM CONFIGURATION
414:031.300
§ 031.
.3
6-TAPE BUSINESS SYSTEM: CONFIGURATION IDD
Deviations from Standard Configuration:
4 disc storage drives are used in place
of 6 tape units.
reader is 20% slower.
punch if' 150% faster.
printer is 52% slower.
1441 Processing Unit, Model A6:
16, 000 core positions
1447 Console, Model 2, and
2260 attachment
290
1442 Card Reader, Model 4:
400 cards/min.
200
1444 Card Punch:
250 cards/min.
375
1443 Printer, Model 2, and
5567 Printer Control:
240 lines/min.
685
1311 Disk Storage Drives (4)
and 3321 Disk Storage Control
Adapters:
Card Read Punch
Printer
Optional Features Included: •
$ 1, 945
Indexing and Store Address
Registers
Multiply-Divide
Sense Switches
Expanded Print Edit
Print Storage
Direct Seek on 1311
TOTAL:
© I 964 Auerbach Corporation and Info, Inc.
1,715
20
25
90
325
15
20
165
50
$ 5, 920
4/64
414:031.301
IBM 1440
§
031.
6-TAPE AUXILIARY STORAGE SYSTEM: CONFIGURATION VD
Deviations from Standard Configuration:
4 disc storage units are used in
place of 6 tape units
Punch is 150% faster.
Printer is 52% slower.
Reader is 20% slower:
Equipment
1301 Disk Storage, Model 11, and
3832 Expanded Disk Storage
Control: 20, 000, 000 characters
1441 Processing Unit, Model A6:
16, 000 core positions
1,945
290
1442 Card Reader, Model 4:
400 cards/min.
200
1444 Card Punch:
250· cards/min.
375
1443 Printer, Model 2, and
5567 Printer Control:
240 lines/min.
685
Adapters:
Card Read Punch
Printer
Indexing and Store Address
Registers
Multiply-Divide
Sense Switches
Expanded Print Edit
Print Storage
Direct Seek on 1311
TOTAL:
4/64
$ 2,050
1447 Console, Model 2, and
2260 attachment
1311 Disk Storage Drives (4)
and 3321 Disk Storage Control
Optional Features Included: • • • • • ••
Rental
1,715
20
25
90
325
15
20
165
50
$ 7,970
414:041.1 00
IBM 1440
Interna I Storage
Core Storage
INTERNAL STORAGE: CORE STORAGE
§
041.
.1
GENERAL
. 11
Identity:
. 12
Basic Use: ..
.13
Description
. Core Storage .
Contained in 1441 Processing Unit, Models A2,
A3, A4, AS, AB.
Core storage for the IBM 1440 system is housed
in the 1441 Processing Unit. Models A2, A3, A4,
A5, andABcontain2,000, 4,000, 8,000,12,000,
and 16,000 character positions of storage respectively. Cycle time is 11. 1 microseconds
for each access of one alphameric character.
Each storage position consists of eight bits: six
data bits, one odd parity bit, and one word mark
bit.
. 16
Reserved Storage
Purpose
Arith registers:
Logic registers:
I/O control:
I/O areas:
.31
9 characters
(optional)
O.
Storage Medium: ..
.23
Storage Phenomenon:
.24
Recording Permanence
. magnetic core.
direction of magnetization.
Access Technigues
.281 Recording method: ... coincident current.
.283 Type of access: . . . . . uniform.
· 29
Potential Transfer Rates
.291 Peak bit rates
Cyc"ling rates: . . . . . 90,000 cycles/second.
.292 Peak data rates
Unit of data: . . . . . . character.
Conversion factor: .. 8 bits/char.
Data rate: •.•...• 90,000 char/sec.
Compound data rate: • 90,000 char/sec.
Number of
Locations
Index registers:
.21
.28
The word mark bit defines the size of data fields
and instructions. Internal transfer operations can
be terminated by a word mark or by a record or group
mark, depending upon the instruction used. The
"Move Record" instruction, which is part of the
optional Advanced Programming feature for the
IBM 1401, is standard in the 1441. The effective
internal transfer rate is 45,000 characters per
second. Parity checks are made on all internal
transfers.
First Delivery: . . . . . April, 1963.
PHYSICAL FORM
.241 Data erasable by
program: . . . . . . . . yes .
.242 Data regenerated
constantly: . . . . . . . no .
. 243 Data volatile:. . . .
. no.
· 244 Data permanent:. .
. no .
. 245 Storage changeable: . no.
. Working storage.
.15
.2
none.
O.
O.
O.
.3
DATA CAPACITY
.31
Module and System Sizes
(See table below. )
· 32
Rules for Combining
Modules: .•.•••.• one module per system,
of any size listed above .
.4
CONTROLLER: •.•.• no separate controller.
.5
ACCESS TIMING
.51
Arrangement of
Heads: ••••.•••• 1 access circuit per
--.
system.
· 52
Simultaneous
Operations:
•••..• none
Module and §1stem Sizes
Model:
Words:
Characters:
Digits:
Instructions:
Modules:
Minimum
Maximum
stor~e
stor~e
A2
variable.
2,000
2,000
variable.
1
A3
A4
A5
A6.
4,000
4,000
8,000
8,000
12,000
12,000
16,000.
16,000.
1
1
1
1.
©1964 Auerbach Corporation and Info,lnc.
4/64
414:041.530
IBM 1440
.8
.53
Access Time Parameters and Variations
; 531 For uniform access
Access time: ••••• 6.3 f.1.sec.
Cycle time: •••••• 11.1 f.1.sec.
For data unit of: •••• 1 character.
•6
CHANGEABLE
STORAGE: •••••• no.
.7
PERFORMANCE
• 72
Transfer Load Size
With self:
.73
•••••••• 1 to N characters, limited
by storage capacity.
Error
Invalid
address:
Receipt of
data:
Check or Interlock
limit check
stop.
parity check
indicator
and
alarm •
Recovery of
data:
parity check
indicator
and
alarm.
Dispatch of
data:
send parity bit
indicator
and
alarm.
Invalid
character
validity check
stop and
alarm.
Conflicting
commands:
not possible.
Effective Transfer Rate
With self: ••••••.•• 45,000 char/sec.
4/64
ERRORS I CHECKS AND ACTION
414:042.100
IBM 1440
Internal Storage
Storage
1311 Disk
INTERNAL STORAGE: 1311 DISK STORAGE
§
042.
.13
.1
GENERAL
.11
Identity: .•••••..•• Disk Storage Drive.
1311 Models 1 and 2.
.12
Basic Use: . . . . . . • • auxiliary storage.
• 13
Description
The 1311 Disk Storage Drive is a recent development in low cost random access storage which
features rapid interchangeability of the "Disk
Pack" storage cartridges. The system is suitable
for either random or sequential processing
methods.
Each Disk Pack consists of six discs on a common
vertical axis. Data can be recorded on ten disc
surfaces; the top and bottom surfaces of the pack
are not used. Each recording surface is divided
into 100 concentric tracks, each track is divided into
20 sectors, and each sector holds a 6-character
address and up to 100 alphameric characters of
data. Therefore, the data capacity is 2,000 characters per track, 200,000 characters per surface,
and 2,000,000 characters per pack. Up to five
Disk Storage Drives can be connected to a 1440
system, so the maximum on-line data capacity is
10,000,000 characters. Disk Storage Drives cannot be connected to the 1441 Model A2 Processing
Unit, which has only 2,000 positions of core
storage.
Access is by means of a comb-like mechanism
containing five arms that move horizontally between the discs. Each arm has one read-write
head mounted on the top and one on the bottom,
and each head serves one disc surface. The entire
access mechanism moves as one unit, so all ten
read-write heads are always positioned at corresponding tracks on their respective surfaces. The
term "cylinder" is applied to each group of ten
tracks (one on each disc surface) that can be read
or recorded upon at a single setting of the access
mechanism. There are 100 cylinders per Disk
Storage Drive, and each cylinder can hold 20,000
data characters.
Time for access mechanism movement ranges
from zero (for successive references to a
previously-selected cylinder) to 400 milliseconds;
average random access time is 250 milliseconds.
Unless the optional Direct Seek feature is installed,
the access arms cannot move directly from one
cylinder to another. Instead, the arms retract all
the way to the "home" position (beyond track 00)
and then move back to the selected cylinder. The
result is that movements between adjacent cylinders
require from 85 milliseconds (track 00 to 01) to
390 milliseconds (track 98 to 99).
Description (Contd.)
Rotational speed of the disks is 1,500 rpm. Maximum rotational delay is 400 milliseconds, and the
average is 20 milliseconds. There is also a head
select delay of 2 milliseconds. Total reference
cycle time to read a randomly-placed 100character record, up-date it, re-write it, and
execute a programmed write check is 354 milliseconds. If no access motion is required the total
reference cycle time is reduced to 104 milliseconds. Peak data transfer rate is 77,000 characters per second, and the effective bulk transfer
rate is just under 50,000 characters per second.
A single read or write instruction can transfer
from 1 to 200 consecutive sectors of information;
i. e., from 100 characters to the capacity of core
storage in multiples of 100 characters. The programmer can elect to read and write sector addresses along with the data records. Handling of
variable-length disc records is facilitated by
"sector count overlays" in which the first three
characters of a record specify the number of
sectors (from 2 to 200) comprising that record.
All capacities and transfer rates quoted here are
based on operation in the "move" mode, in which
six data bits and one parity bit are recorded for
each character. In the alternative "load" mode,
the word mark bit is also recorded for each character, and sector capacity is reduced from 100 to
90 data characters. All capacities and transfer
rates for the load mode are therefore ten percent
lower than the figures quoted here. Use of the
load mode is essential for program storage and
for data storage when field lengths vary from
record to record.
Checks are made for parity errors, wrong length
records, and unequal address comparisons. The
"write disc check" instruction causes a characterby-character comparison of data just written on the
disc with the data in core storage. It usually follows each write operation. All disc errors cause
the setting of testable indicators. .
Disk Storage Drive seek time can be fully overlapped
with internal processing. A 'branch if access
mechanism busy" instruction is provided. No
overlapping is possible during disc read or write
operations. Only one seek operation may go on at
a time, regardless of the number of Disk Storage
Drives in a system, unless the Seek Overlap
feature, described below, is added.
The removable Disk Packs are 14 inches in diameter, 4 inches high, and weigh'less than 10 pounds,
including covers. A Disk Pack can be removed
from a Disk Storage Drive and replaced by another
Disk Pack in about one minute. When a Disk Pack
is not mounted on a drive, the pack and its cover
© 1964 Auerbach Corporation and Info, Inc.
4/64
414:042.130
§
IBM 1440
042.
· 13
Description (Contd.)
combine to form a sealed container that can be conveniently stored and transported. The 1316 Disk
Packs can be either rented at $15 per month or
purchased at $490 each.
Optional Features
Direct Seek: Permits the access mechanism to
move directly to the specified cylinder without retuming to the "home" position. Access motion
time ranges from zero to 250 milliseconds and
averages 150 milliseconds.
Track Record: Permits reading and writing a full
track as a single 2,980-character record, thereby
increasing the capacity of each Disk Pack from
2,000,000 to 2,980,000 characters. The increased
capacity is achieved by using the areas that normally contain sector addresses for data storage.
Complete track records can be intermixed with
sector organization of records on other tracks on
the same Disk Pack.
Scan Disk: Permits an automatic search of data
recorded in disk storage for a specific identifier or
condition.
Seek Overlap: Permits a disc seek operation to
overlap one disc read or write operation plus any
number of other seek operations. The feature must
be installed on every Disk Storage Drive in a
system.
. 15
First Delivery: . . . . . April, 1963.
.16
Reserved Storage: ... none. (Note that each 100digit sector is preceded by
a 6-digit address, but
these address digits are
not counted as storage.)
·2
PHYSICAL FORM
.21
Storage Medium: ..
.22
PhySical Dimensions
· 26
Bands per Physical
Unit: . . . . . . . . . . . 100 per disc surface.
· 27
Interleaving Levels: .. 1; i. e. , no interleaving.
· 28
Access Techniques
.281 Recording method: ... by one of the magnetic heads
on access arms which
move horizontally in
unison.
· 283 Type of access
Description of stage
Possible starting stage?
Move heads to home
position* and then
to selected band: .. mandatory when new band is
selected.
Wait for selected
sector for reading
or recording: . . . . . if same band was previously
selected.
*Not necessary with Direct Seek feature.
· 29
Potential Transfer Rates
.291 Peak bit rates
Cycling rates: . . . . .
Bit rate per track: ..
.292 Peak data rates
Unit of data: . . . . . .
Conversion factor: ..
1,500 rpm.
539,000 data bits/sec/track.
character.
7 bits per character (6 plus
parity).
Gain factor: . . . . . . 1 track/band.
Data rate: . . . . . . . . 77,000 characters/sec, not
counting address digits .
.3
DATA CAPACITY
.31
Module and System Sizes
Maximum
Storage
Identity:
1311
Modell
1311
Model 2
Discs:
Words:
Characters:
Instructions:
Modules
6
variable
2,000,000
variable
1
6
variable
2,000,000
variable
1
. multiple magnetic discs.
.222 Disc
Diameter: . . . . . . . . 14 inches o. d.
Thickness: . . . . . . . thin.
Number of shaft: ... 6.
11311
Modell
and 4
1311
Model
2's.
30.
variable.
10,000,000.
variable.
5.
· 23
Storage Phenomenon: . magnetization.
.4
CONTROLLER
.24
Recording Permanence
.41
Identity: . . . . . . . . . . 3321 Disk Storage Control.
· 42
Connection to System
· 241 Data erasable by
instructions: . . . . . .
· 242 Data regenerated
constantly: . . . . . . .
. 243 Data volatile: . . . . . . .
. 244 Data permanent: . . . . .
.245 Storage changeable: ..
.25
no.
no .
no.
yes.
Data Volume per Band of 1 Track
Words: . . . . . . . . . . .
Characters: . . . . . . . .
Digits: . . . . . . . . . . .
Instructions: . . . . . . .
Sectors: . . . . . . . . . .
4/64
yes.
variable.
2,000.
2,000.
variable.
20.
.421 On-line: . . . . . . . . . . l.
.422 Off-line: . . . . . . . . . . none .
· 43
Connection to Device
.431 Devices per controller: 5 modules .
. 432 Restrictions: . . . . . . . 1 Model 1 and 1 to 4 Model
2 drives; no other restrictions.
In addition, up to 5 1301
Disk Storage modules can
be used in the same 1440
system.
INTERNAL STORAGE: 1311 DISK STORAGE
§ 042.
.44
414:042.440
.53
Data Transfer Control
.532 Variation in access time
.441 Size of load.
Variable length: .•.. 1 to N sectors of 100 characters per sector; number of
sectors is set by programmer and limited by size of
core storage.
Fixed length: . . . . . . 20 sectors of 100 characters
per sector (one band) .
. 442 Input-output area: ••. core storage.
.443 Inout-output area
access: . • . . . . . . . each character.
.444 Input-output area
lockout: . . . . . . . . . yes.
• 445 Synchronization: .••. automatic.
. 447 Table control: . . . . • . none .
.448 Testable conditions: .. none.
.5
ACCESS TIMING
.51
Arrangement of Heads
.511 Number of stacks
Stacks per system: .. 50 max.
Stacks per module: •. 10.
Stacks per yoke:. . . . 10.
Yokes per module: .. 1.
.512 Stack movement: •... horizontal.
.513 Stacks that can access
any particular
location: . • . . . . . . . 1.
. 514 Accessible locations
By single stack
With no movement: • 20 sectors.
With all movement:. 2,000 sectors.
By all stacks
With no movement: . 200 sectors per module.
200 to 1,000 sectors per
system.
. 515 Relationship between
stacks and locations:. three most significant digits
of sector address denote
head and band (cylinder)
number.
. 52
Access Time Parameters and Variations
Simultaneous
Operations: . . . . . . . maximum of 11311 Disk
Storage operation (reading,
recording, or seeking) at
a time per 1440 system.
(Optional Seek Overlap
feature permits simultaneous seek operations on
each 1311 drive.)
Variation
°Without Direct Seek
Move head to home
pOSition and then
to selected band: 75 to 392 msec.
Wait for selected
sector for reading
or recording:
o to 40 msec.
75 to 432 msec.
Total:
°With Direct Seek
Move head to
selected band:
54 to 248 msec .
Wait for selected
sector for reading
or recording:
0 to 40 msec.
Total:
54 to 288 msec.
.6
CHANGEABLE STORAGE
.61
Cartridges
250 msec.
20 msec.
270 msec.
154 msec.
20 msec.
174 msec.
· 611 Cartridge capacity
Without Track Record
feature: . . . . . . . . 2, 000,000 characters
(6 discs).
With Track Record
feature: . • . . . . . . 2,980,000 characters
(6 discs).
.612 Cartridges per module: 1 Disk Pack on-line at a
time .
.613 Interchangeable: . . . . . yes.
• 62
Loading Convenience
· 621 Possible loading
While computing
system is in use: .. yes.
While storage system
is in use: . • . . . . . yes, if the particular drive
is not addressed.
.622 Method of loading: .•. operator.
.623 Approximate change
time: .•.•...••.. 1 minute •
• 624 Bulk loading: . . . . . . . no; 1 Disk Pack of 5 discs at
a time.
.7
AUXILIARY STORAGE PERFORMANCE
· 71
Data Transfer
Pairs of storage unit possibilities
With self: . • . . • . . . no.
With core storage: .• yes.
© 1964 Auerbach Corporation and Info, Inc.
4/64
414:042.720
§
042.
. 72
Transfer Load Size
With core storage: ... 1 to N sectors; number of
sectors is selected by program and limited by size
of core storage.
1 block of 20 sectors (one
band).
. 73
Effective Transfer Rate
The times shown are the average for either reading
from or recording on disc storage with no checking.
With core storage
With Direct Seek:. • . 38, 200 char/sec.
Without Direct Seek: . 33, 800 char/sec.
4/64
IBM 1440
.8
ERRORS, CHECKS AND ACTION
Check or
Error
Interlock
Invalid address: check on matching
sector address
Invalid code:
none.
Receipt of data: parity check
Recording of
programmed readdata:
back and compare
Recovery of
data:
parity check
Dispatch of
data:
parity bit included.
Timir.g conflict: interlock
Physical record
missing:
check on record length
Reference to
check on optional lock
locked area:
Action
indicator.
indicator.
indicator.
indicator.
wait.
indicator.
indicator.
414:043.100
IBM 1440
Infernal Storage
1301 Disk Storage
INTERNAL STORAGE: 1301 DISK STORAGE
§ 043.
.13
.1
GENERAL
. 11
Identity: .
Disk Storage Unit.
1301 Models 11, 12, 21, ar."
22.
.12
Basic Use: ..
auxiliary storage.
.13
Description
The 1301 Disk Storage Unit provides substantially
larger on-line random access storage capacities
than the 1311 Disk Storage Drive described in the
preceding report section, but lacks the interchangeable Disk Pack storage cartridge feature. Models
11, 12, 21, and 22 of the 1301 Disk Storage Unit, used
in IBM 1240, 1440, and 1460 systems, also lack
the ability to handle variable record lengths that is
a feature of the 1301 Models 1 and 2 used in larger
IBM systems. Models 11 and 21 have one disc
array and one· access mechanism. Models 12 and
22 contain two disc arrays, each with an independent access mechanism. Models 11 and 12 are
master units (the first unit on a system); Models 21
and 22 are additional units.
Each disc array consists of 24 discs, 20 of which
are usable for data storage, on a common vertical
axis. Data can be recorded on 40 disc surfaces.
Each recording surface is divided into 250 concentric tracks, each track is divided into 20 sectors,
and each sector holds a 6-character address and
up to 100 alphameric characters of data. Therefore, the data capacity is 2, 000 characters per
track, 500, 000 characters per disc surface, and
20,000,000 characters per disc array. Up to five
disc arrays can be connected to a 1440 system, so
that the maximum on-line 1301 data capacity is
100,000,000 characters. In addition, up to five
1311 Disk Storage Drives (see Section 414:042) can
be connected, providing a total on-line disc storage
capacity per 1440 system of 110,000,000 characters.
Access is by means of a comb-like mechanism containing 20 arms that move horizontally between the
discs. Each arm has one read-write head mounted
on the top and one on the bottom, and each head
serves one disc surface. The entire access mechanism moves as one unit, so all 40 read-write heads
are always positioned at corresponding tracks on
their respective surfaces. The term "cylinder" is
applied to each group of 40 tracks (one on each disc
surface) that can be read or recorded upon at a
single setting of the access mechanism. There are
250 cylinders per Disk Storage Drive, and each
cylinder can hold 80,000 data characters.
Time for access mechanism movement ranges
from zero (for successive references to a
previously-selected cylinder) to 180 milliseconds.
Description (Contd.)
The disc surface is divided into five 50-track areas,
and each area is further divided into six sections.
Movement between tracks in the same section of
an arl'la takes 50 milliseconds. Movement between
Stld()t·s not in the same section requires 120 milli A -
B is greater than A.
I)
B < A -
B is less than A.
@1964 Auerbach Corporation and Info, Inc,
4/64
IBM 1440
414:061.130
Ii! 061.
.13
.13
Description (Contd.)
•
Control Keys and Switches (Contd.)
•
(h)
Storage Scan - scans core until an invalid
character is found.
(i)
Address'Stop - machine stops when the
address selected by the manual switches is
reached.
start key light operations.
Description (Contd.)
Write Disk switch - prevents data from being
written on any Disk Storage Drive.
On the 1447 Model 4, which is designed to control a
1448 Transmission Control Unit, the manual address switches and the power-on switch have additional functions. A section of display lights on the
indicator panel, a 1448 Start-Reset key, and a 1448
Check-Reset key are added.
initiates or resumes machine
e Start reset key - resets the system (except for
data in the storage address registers, data
registers, error indicators, and core storage).
II
Stop key light system.
stops all processing in the
e
Check Reset key light error.
C\l
Program Load key routine.
e
Type key light - indicates the Console I/O
Printer is selected.
resets a processing
initiates a program load
The Console Printer provides communication between the operator and tlle Processing Unit or Disk
Storage Units. On the 1447 Model 4, the Console
Printer contains controls for the 1448 Transmission
Control. The Console Printer is a Single-case IBM
Selectric typewriter rated at 14.8 characters per
second. Either 10 or 12 characters to the inch can
be specified, and either 6 or 8 lines to the inch
vertical spacing. Double-spacing can be manually
selected, resulting in 3 or 4 lines to the inch.
Maximum writing line is 13 inches. Fully buffered
operation is possible on the Model 2 or 4 with the
Buffer Feature, an optional unit. The Console
Printer is useful for entering and printing exception
data, for logging, and as an inquiry station.
Optional Features
o Sense switches - one controls last card operation; six more are optional and permit external
control of program branching.
4/64
o
Bit switches storage.
o
Enter key storage.
o
Power On and Off switches - turn main power
supply on and off.
o
Emergency Off switch - disconnects power
from the entire system in an emergency.
o
I/O Check Stop switch - causes system to stop
at the completion of an I/o operation if an error
has occurred during the operation.
Cil
Diagnostic switch - determines whether erroneous disk data is to be entered into core with
correct parity or in erroneous form.
Sense Switches: Provides six console s:witches and
corresponding testable indicators that can be used
for manual control of the stored program.
used to alter characters in
causes character to enter core
Buffer Feature (Models 2 and 4 only): Provides a
210-character buffer between the system and the
1447 and its associated units.
For local or remote inquiry or data communication
terminal operations, a 1051 Control Unit can be
connected to the 1447 Model 2 or 4 and can control
a local 1050 Data Communication System, as described in Paragraph 414:103.123. With a Remote
Terminal Attachment (#6149) on the 1447 Model 2
or 4 and a Master Station (#5050) on the 1051, the
local 1051 can communicate via a single half-duplex
line with up to twenty-four remote 1050 systems
over common carrier leased private line telephone
service, common carrier leased sub-voice grade
service, or a privately-owned communication
network.
414:071.1 00
IBM 1440
Input-Output
1442 Card Read Punch
INPUT-OUTPUT: 1442 CARD READ PUNCH
Ii 071.
.12
.1
GENERAL
. 11
Identity: . . . . . . . . . . Card Read Punch.
1442 Models 1 and 2.
Card Reader.
1442 Model 4.
. 12
Description
The 1442 Card Read Punch, Modell or 2, is a
combination input-output unit for standard 80column punched cards. The 1442 Card Reader,
Model 4, is an input unit only. From a single
1, 200-card hopper, the cards are fed serially by
column past a single photoelectric reading station,
past a single punching station, and into a 1,300card radial stacker. A second stacker, which may
, be program-selected, is standard on Models 2 and
4, and optional on Model 1. Cards can be loaded
and removed without stopping the unit. All format
control is by the stored program; there is no plugboard.
A single 1442 can handle either an input or an output file. Alternatively, it can read data from and
punch results into the same cards, or the results
can be punched into trailer cards in the same file.
Two 1442 Card Read Punches can be connected to a
1440 system.
When cards are being read continuously and no
punching is being done, maximum speeds are 300
cards per minute for Model 1 and 400 cards per
minute for Model 2. Model 4 reads cards at a
rated speed of 400 cards per minute. Each column
is read twice and the results are compared. Checks
are made for invalid characters, improper registration, and malfunctions of the light-sensing
mechanism. Conversion from standard IBM
Hollerith card code to the internal BCD code is
automatic.
Reading begins at column 1 and can be terminated
after one to 80 columns have been read by a group
mark character with a word mark in the core storage input area. Decreasing the number of columns
read has no effect on the input speed, but it does increase the available computing time per card. At
the maximum input speed, available computing
time or "overlap time" for Modell varies linearly
from 55 milliseconds when all 80 columns are read
to 157 milliseconds when only one column is read.
For Models 2 and 4 the corresponding range of
overlap times is 40 to ll8 milliseconds.
Card feeding is asynchronous; there are no discrete
clutch points as in the IBM 1402 Card Read-Punch.
This means that if processing time slightly exceeds
overlap time, card input speed will be reduced proportionately in the 1442; with a I-point clutch, the .
Description (Contd.)
input speed under the same circumstances would
be cut in half. When the peak card reading speed
cannot be maintained, the card input cycle is increased from 200 to 210 milliseconds for Modell
and from 150 to 160 milliseconds for Models 2 and
4. The increase is caused by the clutch pickup
time when feeding on demand. With the longer
cycles, overlap time is increased by 20 milliseconds for Model 1 and 15 milliseconds for Models
2 and 4.
Card punching, like reading, is performed serially
by column, always begins at column 1, and can be
terminated after any number of columns from one
to 80. The cards to be punched can be either blank
or pre-punched. Skipping is not possible; each
blank column requires a full punch cycle. An echo
check is made to insure that the correct dies have
been actuated for each character.
Card punching speeds depend upon the number of
consecutive columns punched in each card. Model
1 punches at 80 columns per second and Model 2 at
160 columns per second. To the punching times
must be added the time to feed and position each
card: 210 milliseconds for Modell and 160 milliseconds for Model 2. The total positioning time
(but none of the punching time) is available for
overlapped computing. Resulting punching speeds
can be summarized as follows:
Number of Columns
Punched, P
1
10
20
40
80
Formula:
Cards/min. , Cards/min. ,
1442 Modell 1442 Model 2
270
179
130
84
50
360
270
210
146
88
60,000
12.5P+210
60,000
6. 25P+160
When reading and punching are done on the same
card, the reading occurs during the punch positioning time of 210 or 160 milliseconds, and the overall
read-punch operation proceeds at the same speed as
punching alone. Overlap time per card for the combined operation varies with the number of columns
read and is the same as for reading alone.
Two different punch instructions are provided:
"punch and feed" is used when punching only, and
"punch and stop" is used in read-punch operations
where the next card will be fed and positioned for
punching by the succeeding "read card" instruction.
"Punch and stop" also allows successive punching
operations to apply to the same card.
© 1964 Auerbach Corporation and Info, Inc.
4/64
414:071.120
IBM 1440
!l 071.
. 12
Description (Contd.)
Optional Features
Card Image: Provides direct input to and output
from core storage of any card code; the code translation and character validity check are bypassed.
The feature is useful for reading and punching
column binary cards or any cards that use nonstandard codes.
4/64
. 12
Description (Contd.)
Selective Stacker: Provides a second 1, 300-card
stacker that can be selected under program control.
The feature is standard on the 1442 Models 2
and 4.
Punch Column Skip: Allows punch to space over
card columns without interlocking the system.
g
414:0n.100
l~~
IBM 1440
Input-Output
1444 Card Punch
INPUT-OUTPUT: 1444 CARD PUNCH
§ 072.
.12
.1
GENERAL
• 11
Identity: . . . • . . • . • . Card Punch.
1444 ModelL
. 12
Description
The 1444 Card Punch is a punched card output unit
for a 1240 or 1440 system. This unit can be attached to a 1440 in combination with an IBM 1442
Model 4 Card Reader or an IBM 1442 Card Read
Punch, Model 1 or 2.
The IBM 1444 has a rated speed of 250 cards per
minute. The card-punch hopper capacity is 1200
cards. Cards feed 12-edge first, face down. The
feed path is from right to left. Cards pass a
blank station, a punch station, and a punch-read
station. The punch station consists of 80 punch
magnets for recording information. The punchread station has 80 brushes that read the data
punched in the card for a hole-count check. Because of hole-count checking, only blank cards
should be placed in the punch hopper.
The 1444 Card Punch is equipped with two radial
stackers, each having a card capacity of 1000
cards. Cards can be removed from the stackers
without stopping card feeding. If the stacker becomes full, card-punch and system operation
stops.
The card punch operates at a rated speed of 250
cycles per minute (240 milliseconds per cycle).
Actual card punching, at an optimum rate of 250
cards per minute, is controlled by punch instructions in the program. These are four points in the
cycle (occurring at 60-millisecond intervals) when
the punch feeding mechanism can start the punch
cycle.
Description (Contd.)
The punch cycle is divided into three separate
functions:
•
Punch start time is 37 milliseconds. After the
feed mechanism has been pulsed, the time required for the card to be fed and pOSitioned
constitutes punch start time. The Processing
Unit is interlocked during punch start time.
•
Card Punching time is 181 milliseconds. The
actual punching of the card takes place during
this part of the cycle. The Processing Unit is
always interlocked during card punching time.
(i)
Processing time is 22 milliseconds. This is
the remainder of the punch cycle that·is available for overlapped internal processing.
The next Punch instruction must be given during
this 22-millisecond period, or the punch operation
will end and at least 60 milliseconds will elapse before the punch can start again.
Card punching speeds and available processing
times are:
Speed,
cards/min.
C:ycle Time,
msec.
Processing Time,
msec.
250
240
22
200
300
82
166
360
142
143
420
202
125
480
262
© I 964 Auerbach Corporation and Info, Inc.
4/64
414:081. 100
IBM 1440
Input-Output
1443 Prin ter
INPUT-OUTPUT: 1443 PRINTER
§
081.
.12
.1
GENERAL
.11
Identity: .
. 12
Description
Printer.
1443 Models 1 and 2.
The 1443 Printer utilizes a horizontal typebar printing mechanism whose chief advantage is rapid interchangeability of character sets. Peak speed,
using the standard 52-character set, is 150 lines
per minute for Modell and 240 lines per minute for
Model 2. Using the optional character sets described below, the peak speeds range from 120 to
600 lines per minute. Skipping speed is approximately 15 inches per second, and skipping and
spacing are controlled by the stored program and a
12-channel paper tape loop. Only one printer can
be connected to a 1440 system.
The standard number of print positions is 120,
spaced ten per inch. Line spacing can be set at
six or eight lines per inch by an external switch.
Continuous fanfold, sprocket-punched forms from
4 to 16.75 inches in overall width arf:: fed by an
automatic carriage. Maximum form length is 22
inches at six lines per inch and 16.5 inches at eight
lines per inch.
All editing and format control is performed by the
stored program; there is no plugboard. A "sync"
check is made to insure that each print hammer was
actuated at the proper time. A print error causes
a halt or the setting of a testable indicator, depending upon the console I/O Check Stop Switch setting.
When the printer is operating at its peak speed, 24
milliseconds of computing time are available during
the print cycle for each line regardless of the size
of the character set. The printer timing relationships are summarized below.
Modell
Size of
Char. Ser
Peak Speed,
Lines/Min.
13
39
52
63
430
190
150
120
Cycle Time,
msec.
140
316
400
497
Overlap Time,
msec.
24
24
24
24
Description (Contd. )
Model 2
Size of
Char. set
13
39
52
63
Peak Speed,
Lines/Min.
600
300
240
200
Cycle Time,
msec.
100
200
250
300
Overlap Time,
msec.
24
24
24
24
Optional Features
Selective Character Set: Required for use of any
typebar other than the standard 52-character
model.
Character Sets: Can be interchanged with the
standard typebar in approximately 15 seconds. A
dial must be manually set to the proper number of
characters. The following sets are available:
13-character:
39 - character:
52-character:
63-character:
0-9 and specials . - *
0-9, A-Z, and specials. , $
0-9, A-Z, and 16 specials
0-9, A-Z, and 27 specials
Print Storage: Provides a special buffer register
that stores the line to be printed and permits the
stored program to continue 2.4 milliseconds after
initiation of a print operation. Available computing
time with the 52-character set is increased from
24 to 397.6 milliseconds per line. A "branch if
printer busy" instruction is provided, and interlocking will occur if a print instruction is given
before printing of the previous line has been
completed.
Additional Print Positions: Provides 24 additional
print positions, expanding the print field from 120
to 144 positions. Printer operation is unchanged.
@1964 Auerbach Carparatian and Info, Inc.
4/64
IBM 1440
414:081.120
§ 081.
EFFECTIVE SPEED
1443 PRINTERS WITH 52-CHARACTER SET
8,000
5,000
4,000
3,000
2,000
1,000
900
800
700
Printed
Lines per
Minute
600
500
400
300
.....:
200
'-
...............
---- ---- ---- --- ~ - --~
/Model1
100
90
80
70
60
50
40
30
20
o
112
1
2
3
Inter-Line Spacing, Inches
4/64
4
5
414:091.1 00
IBM 1440
Input-Output
7335 Magnetic Tape Unit
INPUT-OUTPUT: 7335 MAGNETIC TAPE UNIT
§
091.
.1
GENERAL
. 11
Identity: . . . . . . . . . . Magnetic Tape Unit.
7335 Models 1 and 2.
. 12
Description
The 7335 Magnetic Tape Unit reads or writes magnetic tape at a density of 556 characters per inch
and a speed of 20, 016 characters per second. Two
models are available; the 7335-1 has a tape control
unit and one tape drive housed in the same frame,
while the 7335-2 contains a control and two tape
drives. (The frame of the second tape drive is
mechanically secured to the frame of the first tape
drive). Only one 7335, Model 1 or 2, may be connected to a 1440 system. Thus, although some advantages of magnetic tape are available, applications such as tape sorting are still not practical in
a 1440 system.
.2
PHYSICAL FORM
.21
Drive Mechanism
.211 Drive past the head: ..
· 212 Reservoirs
Number: . . . . . . . . .
Form: . . . . . . . . . .
Capacity: • . . . . . . .
.213 Feed drive: . . . . . . . .
. 214 Take-up drive: . . . . . .
· 22
The basic speed of the 7335 should be adequate for
most 1440 system applications, but its throughput
can be limited by the fact that high speed rewind,
which requires 2.2 minutes per full reel, is always
terminated by unloading of the tape from the vacuum
columns and r~ad-record head. Rewinding without
unloading requires 13.3 minutes.
The 7335 Magnetic Tape unit is available with all
models of the 1441 ProceSSing Unit except model
A2. The #7802 Tape Adapter is required on the
1441.
.14
First Delivery: .•... October, 1964.
2.
vacuum.
about 1. 5 feet .
motor •
motor.
SenSing and Recording System s
.221 Recording system: ... magnetic head.
· 222 Sensing system: ..•.. magnetic head.
· 223 Common system: .. . two-gap head provides
read-after-write checking.
.23
Multiple Copies: . . . . . none.
.24
Arrangement of Heads
The 7335 is similar in its physical characteristics
to the IBM 7330 Magnetic Tape Unit, and is compatible with tapes prepared on an IBM 729 or 7330
at 556 characters per inch only. Tape speed is 36
inches per second.
A read-after-write parity check detects most recording errors at the time of occurrence. Both
lateral and longitudinal parity are checked during
read operations. Data can be recorded and read in
either the "load" or "move" mode. In the load
mode, each word-mark bit in core storage corresponds to a special word-separator character on
tape. In the move mode, only the six data bits in
each core position are transferred to tape; wordmark bits in core storage are neither written on
tape nor affected by a tape read operation. Block
lengths are fully variable in either mode.
pinch roller friction .
Use of station: . . . . . .
Stacks: . • . . . . . . . . .
Heads/ stack: . . . . . . .
Method of use: . . . . . •
recording.
1.
7.
1 row at a time.
Use of station: . . . . . .
Distance: . . . . . • . . .
Stacks: . . . . . . . . . . .
Heads/stack: . . . • . . .
Method of use: . . . . . .
sensing.
0.30 ± 0.02 inch.
1.
7.
1 row at a time.
.3
EXTERNAL STORAGE
.31
Form of Storage
.311 Medium: • . . . . . . . . . plastic tape with magnetizable coating.
.312 Phenomenon: . . . • . . . magnetization.
.32
Positional Arrangement
.321 Serial by: • . . . . • . . . 1 to N rows at 556 chari
inch; N limited by available core storage.
.322 Parallel by: ••....• 7 tracks.
.324 Track use
Data: . . . • . . . . . • • 6.
Redundancy check: .. 1.
Timing: . . • . . . . . . 0 (self-clocking).
Control signals: . . . . O.
Unused: . . . . . . . . . O.
Total: • . . . . . . . . . 7 .
. 325 Row use
Data: • . . . . • . . . . . 1 to N.
Redundancy check: . . 1.
Timing: . . . . . . • • . O.
Control signals:. . . . O.
Unused: . . • . . . . . . O.
Inter-block gap: •... 0.6875 to 0.8750 inch
(0.75 inch nominal).
©1964 Auerbach Corporation and Info, Inc.
4/64
414:091.330
IBM 1440
§ 091.
• 33
~:
.•.••••..• as in Data Code Tabl~,
Page 414:141.100.
.34
Format Compatibility
Other device or
system: . . . . . . . . . IBM 729 or 7330 Magnetic
Tape Units at 556 bits/
inch density only.
Code translation: . . . . none required.
· 35
Physical Dimensions
. 351 Overall width: . . . . . . 0.5 inch wide .
. 352 Length: . . . . . . . . . . 2,400 feet per reel.
.4
CONTROLLER
· 41
Identity: . . . . . . . . . . contained in tape unit.
· 42
Connection to System
. 421 On-line: . . . . . . .
. 1 max. (7335 Model 1 or 2) .
. 422 Off-line: . . . . . . . . . . none.
· 43
· 52
Input-Output Operations
· 521
· 522
.523
. 524
Input: . . . . . . . . . . . .
Output: . . . . . . . . . . .
Stepping: . . . . . . . . .
Skipping: . . . . . . . . .
.53
Code Translation: .•.. matched codes.
.54
Format Control: . . . . . none .
· 55
Control Operations
1 block forward.
1 block forward.
none •
1 block backward (backspace);
erase long gap forward (to
skip defective tape areas).
.525 Marking: . . . . . . . . . . inter-block gap; write tape
mark.
.526 Searching: . • . . . • . . . none.
Disable: . . . . . . . . . .
Request interrupt: ...
Select format: . . . . . .
Select code: . . . . • . .
Rewind: . . . . . . . . . .
Unload: • . . . . . . . . . .
.56
Connection to Device
· 44
Data Transfer Control
. 441 Size of load: . . . . . . . 1 to N characters; N limited
by available core storage .
. 442 Input-output areas: ... core storage.
· 443 Input-output area
access: . . . . . . . . . each character.
· 444 Input-output area
lockout: . . . . . . . . . yes.
. 445 Table control: . . . . . . none.
· 446 Synchronization: . . . . automatic.
.5
PROGRAM FACILITIES AVAILABLE
.51
Blocks
· 511 Size of block: . . . . . . . 1 to N characters; N limited
by available core storage.
. 512 Block demarcation
Input: . • . . . . . . . . . limiting mark in core storage or inter-block gap on
tape.
Output: . . . . . . . . . . limiting mark in core
storage.
4/64
Testable Conditions
Disabled: . • . . . . . . .
Busy device: . . . . . . .
Output lock: • . . . . . .
Nearly exhausted: . . . .
Busy controller: . . . .
End of medium
marks: . . . . . . . . . .
. 431 Devices per controller: 7335-1: 1 drive.
7335-2: 2 drives.
. 432 Restrictions: . . . . . . . none.
disabled after unloading .
no.
no.
no.
yes.
yes .
.6
PERFORMANCE
.62
Speeds
.621 Nominal or peak
speed: . . . . . . . . . •
.622 Important parameters
Tape speed: . • . . . . .
Density: . . . . . . . • •
Rewind (high speed): .
Rewind (low speed): .
Inter-block gap: . . . .
Start time
Read: . . . . . . . . . .
Write: • . . . • . . . .
Stop time
Read: . . . . . . . . . .
Write: • • . . • . . . .
.623 Overhead: • • . . . . • . .
.624 Effective speeds: •...
no.
no .
no.
yes •
no.
yes .
20,016 char/sec.
36.0 inches/sec .
556 char/inch.
2.2 minutes/reel.
13. 3 minutes/reel.
0.6875 to 0.8750 inch (0.75
inch nominal).
10.3 msec.
7.2 msec.
9.8 msec .
4.4 msec.
20.8 msec. /block.
20, 016N/ (N + 416), where
N = characters per block
(see graph).
INPUT-OUTPUT: 7335 MAGNETIC TAPE UNIT
§
414:091.630
091.
.63
Demands on System
Condition
Coml2onent
Read
Write
Processing Unit:
.72
. 73
Percentage of data
transfer time
msec l2er block
10.3 + o. 05N
7.3+ 0.05N
or
or
100.
100.
Other Controls
Function
Form
Comment
Start:
key
Low Speed Rewind:
High Speed Rewind:
key
key
Reset:
key
turns Ready light on and
places tape unit in ready
status.
rewinds at 36 inches/second.
rewinds at high speed (pulls
tape from vacuum columns).
resets tape to manual control
(can stop a rewind operation) •
Loading and Unloading
.731 Volumes handled
Storage: . . . . . . . . . 1 reel.
Capacity: . • . . . . . • 2,400 feet; for 1, OOO-char
blocks, capacity is
11,300,000 characters .
. 732 Replenishment time: .. 1. 0 to 1. 5 minutes; tape
unit needs to be stopped .
• 734 Optimum reloading
period: . . . . . . • . . • 13 minutes .
.8
ERRORS, CHECKS AND ACTION
Recording:
Reading:
Input area overflow:
Output block size:
Invalid code:
Exhausted medium:
Imperfect medium:
Timing conflicts:
Recording level:
Check or Interlock
Action
read after write; lateral
parity
lateral and longitudinal
parity
none.
none.
all codes acceptable.
reflective spot or tape
mark
none.
interlock
signal strength check
indicator & alarm.
indicator & alarm.
indicator.
wait.
indicator & alarm.
© 1964 Auerbach Corporation and Info, Inc.
4/64
414:091.900
~
IBM 1440
091.
EFFECTIVE SPEED:
IBM 7335 MAGNETIC TAPE UNIT
10,000,000
7
2
1,000,000
7
4
2
100,000
7
Characters
per Second
4
2
~
t-"'"""
10,000
7
i.oII"'"
L
4
-'
~
2
/
1,000
~
,
7
1
4
2
100
2
10
4
7
2
100
4
7
2
1,000
Characters per Block
4/64
4
7
10,000
414: 101.1 00
IBM 1440
Input-Output
Serial I/O Adapter
INPUT-OUTPUT: SERIAL I/O ADAPTER
.122 1011 Paper Tape Reader: Reads data from 5- to
8-track punched tape at a peak speed of 500 characters per second. (See also Section 401:073.)
f3 101.
.1
GENERAL
.11
Identity: . . . . . . . . . . Serial Input/OutputAdapter.
#7080.
Expanded Serial I/O
Adapter.
#3845.
Data Transmission Unit.
1009 Model 1.
Paper Tape Reader.
1011 Model 1.
Tape Punch.
1012 Model 1.
Optical Mark Page Reader.
1231 Model 1.
Magnetic Character
Reader.
1412 Modell.
Communication Control
System.
IBM 7740.
Direct Data Transfer.
#3271.
. 12
Description
The Serial Input/Output Adapter. when added to the
1441 Processing Unit. permits the on-line use of
anyone of the serial input-output devices described
in Paragraphs .121 through. 127 below. Only one
Adapter may be attached.
. 121 1009 Data Transmission Unit: Transmits and receives data over telephone or telegraph lines at
speeds of 75 to 300 characters per second. (See
also Section 401:101.)
.123 1012 Tape Punch: Punches data into paper or
Mylar tape at a peak speed of 150 characters per
second. Five. six. seven. or eight tracks can be
punched. Read-after write checking is performed.
Code translations must be programmed. (See also
Section 401 :074. )
.124 1231 Optical Mark Page Reader: Reads marked
data from 8-1/2 by 11 inch data sheets at varying
speeds. In "Continuous" mode. speed is constant
at 2,000 documents per hour. (See also Section
414:104.)
.125 1412 Magnetic Character Reader: Reads magnetically inscribed data from card and paper documents at a maximum rate of 950 documents per
minute. (See also Section 401:104.)
.126 7740 Communication Control System: A TeleProcessing system for control and supervision of
a communication network. (See also Section 414:106.)
. 127 3271 Direct Data Transfer: Permits direct
system-to-system data transfers between the 1440
and a 1401, 1460, or second 1440. The Indexing
and Store Address Registers feature is a prerequisite .
.128 3845 Expanded Serial I/o Adapter: For direct
data transfer between the 1440 and a 1410 or 7010,
the Expanded Serial I/O Adapter cable-connects a
1440 with a 1448 Transmission Control Unit to the
Processing Unit in a 1410 or 7010 system. When
not used for inter-system data transfers, this
feature can be used in lieu of a #7080 Serial I/O
Adapter. The Bit Test and Indexing features are
required on the 1441 Processing Unit.
©1964 Auerbach Corporation and Info, Inc.
4/64
414:102.100
IBM 1440
Input-Output
1062 Teller Terminal
INPUT-OUTPUT: 1062 TELLER TERMINAL
§
102.
.12
Description (Contd. )
.1
GENERAL
Additional Equipment Requirements (Contd.)
.11
Identity: • • . • • . . . Teller Terminal.
1062 Models 1 and 2.
.12
Description
control circuits for the Teller Terminals in both online and off-line operations. Each 1061 requires one
half-duplex communication line; i. e., a line capable
of transmitting data in only one direction at a time.
Up to 40 such lines can be connected to a 1440 system via one 1448 Transmission Control Unit. All
lines can operate simultaneously at 14. 8 characters
per second each; but when two Teller Terminals are
connected to a single 1061 Control Unit, only one of
the two can operate at a time. The 1061 can be connected to the 1448 either by direct two-wire cables
for in-house operation, or by data sets and leased
voice or subvoice grade communication lines for remote operation.
The 1062 Teller Terminal provides teller stations
with access through communication lines to customer
account information. The unit has a keyboard for
entering transaction data into the files and a printer
upon which to record the results of the transactions
after processing. The 1062 is primarily intended for
use as an on-line inquiry and posting unit for savings
accounting and other applications involving a passbook or receipt where access to and updating of records is critical.
The unit encompasses a:
o Keyboard, including nine columns of numerical
data keys for entry of account number, old balance, and transaction amounts; and three columns
of transaction code keys for transaction identification. Also provided are functional control keys
for initiation of action, control, and status lights;
and three locks for security over teller and auditor operations.
o Print unit which records 12 characters per inch
serially at a speed of 14. 8 characters per second
upon the 5 - inch wide terminal record tape (for
transmitted data) or on a document inserted in the
document feed (for reply data). The printing
mechanism is similar to that used in the mM Selectric typewriter and provides for printing 10 numeric, 26 alphabetic, and 8 punctuation characters.
o Document feed capable of accommodating documents up to 8 inches wide and passbooks up to 6
inches wide.
o Pair of accumulators for maintaining the net cash
positions for each of two tellers. Transaction
amounts are added or subtracted only upon successful completion of a given transaction. A third
accumulator provides arithmetic ability for the
1062 when operating off-line, and also serves as
a message storage unit during on-line transmission.
Additional Equipment Requirements
One or two Teller Terminals can be connected to a
1061 Control Unit. The 1061 provides the logical and
The 1448 Transmission Control Unit has a one-character assembly register and two single-character
buffers for each transmission line. Automatic interruption of the 1440 at the end of a processor instruction can be initiated by any of the following conditions:
1. Buffer filled (receive).
2. Buffer empty (transmit).
3. Status condition signal (e. g., end of message).
4. Request for polling address (i. e., terminal
identification ).
The 1440 program must store the contents of the program counter and initiate a scan of ail lines. Each
line is assigned an area in core storage and uses 10
core storage positions to keep track of the status of
the messages for each line. Interrupts must be
spaced at least 65 milliseconds apart (determined by
the transmission speed over the line; 14.8 characters
per second = 67 milliseconds per character). A scan
of all lines under peak traffic condition is less than 6
milliseconds, whereas an average scan time should
be about 4 milliseconds. It is possible to write a program for the 1440 which will initiate scan operations
on a fixed time cycle basis.
Optional Features
The 1062 Modell provides full off-line (independent
of communication lines and central processor) as
well as on-line operations, whereas the,Model 2 can
be used on-line only. The 1062 Modell can utilize
a punched program tape loop that provides off-line
control over printing format (tabbing, punctuation,
date, terminal identification, etc.) and arithmetic
operations. An adding machine feature permits use
of the 1062 Modell as an off-line adding machine. A
tape punch for punching l6-channel program tapes is
available.
© 1964 Auerbach Corporation and Info, Inc.
4/64
414:103.100
IBM 1440
Input-Output
1448 Transmission Control Unit
INPUT-OUTPUT: 1448 TRANSMISSION CONTROL UNIT
§
103.
.12
Description (Contd.)
The three basic purposes of transmission control
in a data processing system are:
A Transmission Control Unit Attachment, the Indexing and Store Address Register feature, and a
1447 Model 4 Console are prerequiSites on the 1440
or 1460 system for connection of a 1448. Through
the use of various special features, the 1448 can
accommodate up to 40 half-duplex 1050 or 1060
lines, up to 20 half-duplex 1030 or 1408 lines, up
to 40 half-duplex telegraph lines, or certain combinations of two or more different types of lines.
e To establish a connection between the central
processor and the terminals in the communication network.
Only one 1448 can be attached to a 1440 system.
First delivery is expected in the third quarter of
1964.
o To detect and compensate for distorted transmission or lost signals.
The data collection and communication systems
that can be connected to an IBM 1440, 1460, or
1240 Data Processing Systems by means of a 1448
Transmission Control Unit are briefly described
below.
.1
GENERAL
.11
Identity: . . . . . . . . . . Transmission Control Unit.
1448.
. 12
Description
•
To expand the flexibility of the data processing
equipment.
The 1448 Transmission Control Unit operates in
conjunction with an IBM 1240, 1440, or ·1460 Data
Processing System and data collection/communication systems such as the IBM 357, 1030, 1050,
and 1060 (described below). The 1448 communicates with the 1050 or 1060 at 14.8 characters per
second, with 1031's at 60 characters per second,
and with 357's (via 1408 Transmission Units) at
20 characters per second. Information from the
communication lines enters the 1448 in a serialby-bit, serial-by-character mode. A minimum of
two characters of information is accepted by 1448
core storage for each line. The bits are assembled into characters and transferred to the Processing Unit, where each communication line is
assigned its own message assembly area in core
storage. The reverse happens when the 1448 is
transmitting, in that characters are converted into
bit-serial form and transmitted serially. The exchange of information is· initiated by a "scan operation" associated with a priority interrupt (a temporary interruption of the processor's main
routine by an external signal).
Information is either written into or read out of a
preassigned area in core storage. The following
conditions in the 1448 can interrupt the processor's
main program: a buffer-full condition on a receiving line; a buffer-empty condition on a transmitting line; a status condition with an EOB (endof-block) bit; or a 1448 request for the next polling
address (different terminals on a multipoint
channel). The interrupt causes a program jump to
location 181 in the processor, where the first instruction of the interrupt routine is located. An
interlock prevents interruption while the system is
in the interrupt routine. Direct Data Channel
Interrupt is standard with the 1448. Various
checks are made in the 1448, such as buffer overflow, vertical redundancy (parity check), and
status-character validity check.
. 121 IBM 357 Data Collection System
Each 357 system transmits data from up to 20 remote plant locations to a central point. Input data
can be read from 80-column cards, from 22column identification badges, from portable data
cartridges, and/or from up to 12 manually set 11position slides. Transmission of data is automatic
upon insertion of a card or badge into the proper
reader slot. Data is transmitted at the rate of 20
characters per second.
Up to 20 input stations can be connected by multiwire cable to a 358 Input Control Unit. The 358
can be connected to a 24 Card Punch or a 26 Printing Card Punch for off-line punched card output.
For on-line operation with a 1440, the 358 is connected to a 1408 Transmission Unit, which relays
the input data via voice-grade private or common
carrier lines to the 1448 Transmission Control Unit
at the computer site.
. 122 IBM 1030 Data Collection System
The 1030 system permits two-way communication
between remote plant locations and a central processing area. Input can be from 80-column punched
cards, 10-column identification badges, preset
data cartridges, or up to 12 manually set slides.
Various models of the 1031 Input Station handle
different combinations of these input modes. Up to
24 input stations (in various combinations of "control" and "satellite" stations) can be connected to a
single 2-wire transmission line. Up to twenty 2wire transmission lines can be connected to a 1448
Transmission Control Unit for on-line operation
with a 1440 system. Input data is transmitted in
serial-by-bit form at 60 characters per second.
© 1964 Auerbach Corporation and Info, Inc.
4/64
IBM 1440
414:103.122
• 123 IBM 1050 Data Communication System (Contd.)
§ 103 .
. 122 IBM 1030 Data Collection System (Contd.)
Up to 24 Model 1033 Printers can also be connected
to each 2-wire line. The 1033 provides printed
copy at remote locations at a rated speed of 14.8
characters per second. Used in conjunction with
the 1031 Input station, the 1033 Printer provides
two-way on-line inquiry and reply capabilities.
When on-line operation is not required, up to 24
input stations can be connected via 2-wire lines to
a 1034 Card Punch for punched card output at 60
characters per second.
• 123 IBl\'[ 1050 Data Cownmllication System
The 1050 is designed to serve as a general-purpose
communication system to link the user's branches,
plants, and warehouses to his central data processing system. Input can be from punched tape,
punched cards, or keyboard. Output can be printed
or punched on tape or cards. Input, output, and
data transmission all occur at a rated speed of 14.8
characters per second. A 1050 system can communicate over leased or privately-owned 2-wire
telephone, telegraph, or TWX circuits with any of
the following IBM equipment:
o another 1050 system.
I)
a 1401 system (via a 1447 Model 3 Console and
a 1409 Console Auxiliary).
o a 1440, 1460, or 1240 system (via a 1448 Transmission Control Unit).
o a 1410 or 7010 system (via either a 1414 I/o
Synchronizer or an on-line 1440 system).
4/64
•
a 1401, 1410, 1440, 1460, or 7000 series system (except 7072) via a 7740 Communications
Control System.
•
a 1410 or 7000 series system (except 7072) via
a 7750 Programmed Transmission Control.
•
a 1060 Data Communication System (via a 1448).
Every 1050 system includes a 1051 Control Unit,
which provides the basic connection between the
input-output units and the modulator-demodulator
equipment for the transmission line. A 1050 system can also include one 1052 Printer-Keyboard
and one or two of each of the following units: 1053
Printer, 1054 Paper Tape Reader, 1055 Paper Tape
Punch, 1056 Card Reader, 1057 Card Punch, and
1058 Printing Card Punch. The only configuration
restrictions are that neither the total number of
readers (1054's and 1056's) nor'the total number of
punches (1055'S, 1057's, and 1058's) may exceed
two.
. 124 1060 Data Communication System
The 1060 system is designed for on-line banking
and savings and loan applications. It consists of
1062 Teller Terminals connected to 1061 Control
Units. The 1062 Teller Terminal is fully described
in Section 414:102. The 1060 system can be connected to a 1448 Data Transmission Control for online use with an IBM 1440, 1460, or 1240 Data
Processing System. Alternatively, the 1060 system
can be connected to a 7740 Communication Control
for on-line use with nearly any IBM 1400 or 7000
series system.
414:104.100
IBM 1440
Input-Output
1231 Optical Mark Page Reader
INPUT-OUTPUT: 1231 OPTICAL MARK PAGE READER
§
104.
. 12
Description (Contd.)
.1
GENERAL
• 11
Identity: .
. 12
Description
Data sheets are fed from a pneumatically-controlled hopper (600-sheet capacity) through the reading
area and directed to one of two stackers. The main
stacker holds 600 sheets. Sheets with detected
errors are directed to a separate stacker (50-sheet
capacity). Documents are stacked in reverse sequence in both stackers .
The Optical Mark Page Reader reads orginary pencil marks (not printed characters) from 8-1/2" x
11" data sheets directly into a 1401, 1440, or 1460
Data Processing System. A #7080 Serial Input/
Output Adapter is required on the 1401 or 1441
Processing Unit.
All marks read from a data sheet are stored as
bits in a sonic delay line storage unit until they
are transferred to the computer by execution of a
Read instruction. Less than 10 milliseconds are
required to transfer data from the sonic delay line
to the computer storage.
Optical Mark Page Reader.
1231.
The 1231 (or its off-line counterpart, the 1232) will
be useful in organizations that use standardized
forms for such functions as surveys, orders, applications, medical records, payroll time records,
inventory listings, and sales analyses. The 1231 's
chief advantage is the elimination of much of the
key punching and verifying normally associated with
the preparation of input for ,automatic data processing. In a single pass of the pencil-marked data
sheets through the 1231, the marks are read and
the data is transferred to the computer. (The 1232,
working off-line, converts the marked information
into punched cards.)
Documents are read at varying rates of speed, depending upon the mode switch settings. When set
to "continuous", feeding is at a constant speed of
2,000 documents per hour. When set to "demand",
feeding is controlled by the computer program with
speeds varying up to 1600 documents per hour.
The feeding mode selected depends upon the computer program control method used.
The user may have up to 1,000 mark positions on
one side of a sheet (2,000 on both sides). Mark
positions are pre-printed in rows of 20 positions.
Each row is divided into two 10-position groups.
Each 10-position group is called a "word" for the
purpose of defining a marking area. Each word
can be divided into two 5-position segments. Data
words and segments can be grouped in various
combinations to form fields.
First deliveries are scheduled for the third quarter
of 1964.
Optional Feature
Master Mark: A master data sheet, containing up
to 10 words of marked data, can be read and stored
in the delay line for transfer to the computer. The
master sheet is identified by a special preprinted
mark and contains data that is to be associated with
all subsequent data sheets until a new master sheet
is read. ThUS, data common to a series of data
sheets need be recorded and read only once.
© 1964 Auerbach Carporation and Info, Inc.
4/64
414: 1OS.l 00
IBM 1440
Input-Output
7770 Audio Response Unit
INPUT-OUTPUT: 7770 AUDIO RESPONSE UNIT
.12
Ii 105.
. 11
GENERAL
Identity: . .
.12
Description
.1
Audio Response Unit.
7770 Model 1.
The IBM 7770 Audio Response Unit is a buffered
communications unit that accepts telephoned requests and relays them to a central processing
unit which processes the data and returns a coded
reply to the 7770. The 7770 interprets the reply,
selects the proper words from its stored vocabulary, and transmits these words as a voice response back to the inquirer.
The Audio Response Unit is composed of three
sections: inquiry, control, and audio output.
The inquiry section accepts digital inquiries from
the connected inquiry terminals through a common
carrier receiver, in message lengths up to 40
digits. Transmission of the inquiry between the
receiver and the 7770 is accomplished in a 3-outof-14 or 2-out-of-8 inquiry code. The 7770 receives the inquiry from the receiver terminal in
serial-by-character, parallel-by-bit form. The
inquiry is then translated into BCD form, stored
in a buffer, and transferred tO'the computer when
the total inquiry has been received. An inquiry
is assumed to be complete when the 7770 receives
no data for five consecutive seconds.
The digital control section controls the data flow
between the CPU (central processing unit) and the
7770 through the 1311 Disk Storage control unit.
A Read File command from the CPU causes the
inquiry to be transferred from the 7770 to the
CPU. When the CPU completes its interpretation
of the inquiry, it composes a coded response
message. The CPU then issues a Write File command, and the response message is transferred to
the buffer of the 7770. The response message is
composed of vocabulary word locations (located on
the 7770's magnetic drum), and is sent one word
at a time. The maximum length of this response
message is 38 words. Positions 1 through 38 may
contain the drum addresses of the appropriate
audio response words. A group mark in any position from 2 through 39 will signify the end of the
message. The last position of the 40-character
area is always a blank.
The audio output section provides the actual
audio response to the inquirer. These words,
recorded on the magnetic drum in analog form as
an audio signal, are amplified and transmitted to
the terminal originating the inquiry.
Description (Contd. )
The vocabulary is recorded on the magnetic slepve
of a drum 4 inches in diameter and 10 inches long .
~'.ere are 128 tracks on the drum, addressable by
a seven-bit BCD address field associated with each
track. Two tracks are prerecorded; one having an
address of zeros (blank) and the other with ones
(group mark). The "blank" track indicates 500
milliseconds of silence (500 milliseconds per revolution), and the "group mark" track is used for endof-message control.
The prerecorded drum vocabulary is flexible in
three ways:
(1) The number of drum words can be 32, 48, or
64, and can be increased to 80, 96, 112, or
128 by a special feature.
(2) A master vocabulary provided by IBM is a
list of frequently used industry words, numbers from 0-9, and letters of the alphabet.
(3) The vocabulary can be changed, by exchanging
the analog recorded cylinder, at any time by
the customer.
A 7770 Modell can be used with a 1401, 1440, or
1460, and a 7770 Model 2 with a 1410 or 7010
System. Only one 7770 can be attached to a system, in lieu of a 1311 Disk Storage Drive Model 2,
thereby reducing by one the maximum number of
1311 Model 2's. For a 1401, a 7149 Adapter on the
1311 Model 4 is needed; for a 1440/1460, a 7150
Adapter on the 1311 Modell is needed.
The 7770 operates in a half-duplex mode over toll,
leased, or privately-owned voice-grade lines.
The basic 7770 handles up to four lines. The line
capacity can be expanded, in 4-line increments, to
a maximum of 64 lines. Random inquiries on any
or all input-output lines can be serviced simultaneously.
The 7770 operates with the following inquiry terminal devices, provided the proper arrangements
have been made by the customer with the common
carrier: 1001 Data Transmission Terminal,
rotary dial telephone with associated pushbutton
manual dialing device, rotary dial telephone with
pushbutton card dialer device, pushbutton manual
dialing telephone, and pushbutton type card dialer
telephone.
First 7770 deliveries are scheduled for the first
quarter of 1965.
@1964 Auerbach Corporation and Info, Inc.
4/64
414:106.100
IBM 1440
Input-Output
n40 Communication
Control System
INPUT-OUTPUT: n40 COMMUNICATION CONTROL SYSTEM
.12
§ 106 .
.1
GENERAL
. 11
Identity: . . . . . . . . . . 7740 Communication Control
System.
. 12
Description
The mM 7740 Communication Control System is a
stored-program computer designed to provide
message switching and data transfer using commoncarrier lines and switching facilities. The telecommunications terminals can be standard Teletype Corporation equipment, other standard
terminal equipment, or IBM transmission terminals
Description (Contd. )
Transmission terminals and systems that can be
connected to a 7740 include:
•
Telegraph terminals, operating at speeds up
to 180 bits per second (approximately 24
,Baudot characters per second).
•
mM 65/66 Card Data Transceivers.
•
mM 1009 Data Transmission Units, which
permit communication with other mM 1440,
1401, and 1410 Data Processing Systems.
Ci)
The 7740 system can stand alone as an independent
message switching system, but is more typically
used as a switching facility for an mM data processing system. As an independent system, the
7740 can connect to up to five mM 1311 Disk
Storage Drives, each capable of storing over two
million characters. (Magnetic tapes are not
available to the 7740.) As a system serving a
"host" computer, the 7740 transfers messages to
the host computer for disc or magnetic tape storage.
The 7740 can be connected directly to any of the
following mM computers: 1401, 1410, 1440, 1460,
7010, 7040, 7044, 7070, 7074, 7080, 7090, 7094.
The 7740 system communicates with a 1440 via
the #7080 Serial Input-Output Adapter (Section
414:101).
From 1 to 84 low-speed half-duplex lines (45 to 200
bits per second) and 1 to 4 high-speed half-duplex
lines (1200 to 2400 bits per second) can be connected to.the 7740 system. The actual number of
each type depends on the combination of lines
desired, and on whether or not a disc file is connected. Pairs of half-duplex connections can be
used to form full duplex connections.
The 7741 Processing Unit contains the arithmetic
and logical facility, the core storage, and the
terminal connections. The 7741 consists of one
large unit of six racks or bays.
The console for the 7740 system consists of the
Printer-Keyboard, Card Reader, and Control Unit
of an mM 1050 Data Communication System, connected to its own input-output connection on the
7741.
Communication lines can be leased (private) lines
or toll lines to exchanges. Telegraph or voicegrade lines or exchanges can be used by the system. The 7740 has facilities for automatic dialing
and for automatic answering of dialed input
messages.
The system accepts and transmits messages of any
length, and can perform code conversion under
program control.
Ci)
mM 1013 Card Transmission Terminals.
mM 1050 Data Communication Systems
(printer-keyboard, punched tape, and cards).
o
mM 1060 Data Communication Systems
(teller terminals).
o
mM 7701 and 7702 Magnetic Tape Transmission Terminals.
o
mM 7710 Data Communication Systems (connected to mM 1401 Data Processing Systems).
o
other mM 7740 systems.
o
IBM 7750 Programmed Transmission Controls
(connected to other mM Data Processing
Systems).
Functional Description
The 7740 is an electronic computer available with
three sizes of core storage: 4,096, 8,192, and
16,384 words of four 8-bit characters per word.
The programming package available for the 7740
requires the maximum amount of core storage.
Operation of the system is completely under program control. The 7740 supervises the communications network, and controls traffic by polling
terminals which may have input data and by addressing terminals prior to sending output data.
The program determines the order of polling, and
also maintains output queue control. Terminals
which are polled or addressed must give a response,
which is checked for by the program.
The 7740 can initiate a request for data transfer
between it and the host computer by setting an
indicator in the host. When the host recognizes the
request, programming in the 7740 controls the
transfer. When the host computer initiates or
services a request for data transfer between it and
the 7740, the 7740 must be in a suitable mode of
operation to perform the transfer.
© 1964 Auerbach Corporatian and Info, Inc.
4/64
IBM 1440
414:106.120
§ 106 .
• 12
Description (Contd.)
Functional Description (Contd. )
Character validity checking is done in the system.
It performs a 4-out-of-8 check on internal charac-
ters and on 8-bit data codes. Teletype and 6- and
7-hivel codes entering the system are changed to
an 8-bit form.
.12
Description (Contd.)
Program Processing (Contd. )
Message protection is accomplished by the program.
Procedure's used are message sequence numbering, checking on device addresses associated with
line numbers and line types, and checking of terminal response to polling and addressing. If the
latter check fails, indicating an unavailable or
inoperative device or a line malfunction, output
messages are put on the disc file Greater Than
? (Plus Zero)
A
8
C
D
E
F
8CD
CODE
C
12-(1-8
12-4-8
12-5-8
12-6-8
12-7-8
12
11-3-8
11-4-8
11-5-8
11-6-8
11-7-8
11
0-1
0-3-8
0-4-8
0-5-8
0-6-8
0-7-8
2-8
3-8
4-8
5-8
6-8
7-8
C
C
8 A 8
8 A 8 4
8 A 8 4
8 A 8 4
8 A 8 4
8 A
8
8
C
C
8
8
8
C
C
C
C
C
52
A
X
x
x x
X
X
2 I
X
X
1
2 1
X
H
X
)
+
X
X
X
X
X
X
X
X
(
I
A 8 4 2
A 8 4 2 I
, .
2 I
8 4
8 4
C
@
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
I
8 4 2
8 4 2 I
8 A 8
2
12-3
12-4
C
8 A
8 A
8 A
12-5
C
12-<>
C
8 A
8 A
8 A
X
I
2
2 I
4
4
4 2
I
52
63
DEFINED CHARACTER
CODE
8CD
X
G
12-7
8 A
X
X
X
X
X
X
H
I
I (- zero)
12-8
8 A
8
12-9
C 8 A
8
C 8
8
8
X
X
X
X
X
X
X
X
X
X
X
X
X
8
CARD
52
&
%
A 8 4
A 8 4
A
12-1
12-2
X
):(
X
X
X
A
A 8
C
C
12~
39
I
2
2 1
8 4
8 4
I
8 4 2
8 4 2 I
8
8
C
2 I
13
J
K
L
M
N
0
P
Q
R
4= Racad Merlo:
S
T
11-0
11-1
11-2
11-3
11-4
11-5
11-6
11-7
11-8
11-9
X
1
X
X
X
X
2
2 I
4
I
4 2
4 2 I
8
8
8
8
I
X
2
2 I
X
X
I
X
X
2
C
A
A
U
0-4
C
A
A
4
V
0-5
0-6
A
4
C
A
A
4 2
4 2 I
C
A
8
A
C
8
8
W
X
Z
X
X
X
X
X
X
X
o (Zero)
0
I
2
3
4
I
2
3
4
5
6
5
C
6
C
Y
7
7
8
9
8
9
X
X
X
I
2
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
4 2
4 2 1
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
I
X
X
X
X
X
X
1
2
2 I
4
4
I
8
8
Note 1
52
H
X
X
X
X
X
X
X
X
X
X
X
X
C
C
X
X
X
X
X
X
X
X
4
0-2
0-3
0-7
0-8
0-9
X
4 2 I
2
8
8
C 8
C 8
X
X
X
X
X
A
13
1
C 8
8
C 8
0-2-8
39
CODE
X
X
63
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Note 1
Note 1: 1443 Printer character sets; X indicates the defined character
is included.
Note 2: Codes are arranged in internal collating sequence.
Reproduced from IBM 1440 System Component Description,
IBM Publication A26-5666-0, page 28.
@1964 Auerbach Corporation and Info, Inc.
4/64
414:151.100
IBM 1440
Problem Oriented Facilities
PROBLEM ORIENTED FACILITIES
§
.14
151.
Report Writing (Contd.)
.1
UTILITY
Use of the Report Program Generator involves four
distinct phases:
.11
Simulators of Other
Computers: ••..... none.
.12
Simulation by Other
Computers: . . . • . . . none.
1. Writing the report specifications in a problem
oriented language on four different types of
specification sheets: Input, Data, Calculation,
and Format.
.13
Data Sorting and Merging
2. Translating the report specifications into a symbolic (1440 Autocoder) report program using the
RPG Processor.
IBM 1440 Sort 5:
3. Translating the symbolic program into 1440
machine language form using the 1440 Autocoder Processor.
Using specifications contained in control cards,
Sort 5 generates sort routines for sequencing
blocked, fixed-length records in 1311 Disk storage.
The files to be sorted may be in punched cards,
magnetic tape, or Disk Storage. Tags containing
the control fields and Disk Storage locations of the
records are sorted first. Then the records themselves can be directly rewritten in Disk Storage in
either ascending or descending sequence. Alternatively, by reference to the sorted tags, the
records can be punched or printed in the proper
sequence without internal rearrangement of the data
records themselves. Facilities are included for
adding, deleting, and selecting records. Minimum
configuration requirements are 4,000 core storage
pOSitions, one 1442 Card Read Punch, one 1443
Printer, and one 1311 Disk Storage Drive.
.14
4. Executing the machine language object program
to produce the desired report.
If the user does not require a symbolic deck and
listing of the object program, the second and third
phases can be combined into a single pass. Operation of the RPG Processor requires a 1441 with at
least 4,000 core storage positions, one 1442 Card
Read Punch, one 1443 Printer, and one 1311 Disk
Storage Drive. A 1444 Card Punch and a 1442
Model 4 Card Reader can be used in place of the
1442 Card Read Punch.
Execution of RPG object programs requires a minimum of 4,000 core storage positions (more for
complex reports) and one 1442 Card Read Punch.
Most of the 1440 peripheral devices and optional
features can be utilized when available.
Report Writing
IBM 1440 Basic Report Program Generator (BRPG):
This is a "load-and-go" program that processes
data in punched cards to produce a report in printed
and/or punched card form. The report specifications for the desired report are written in a problem oriented language, punched on cards, and read
into the 1440 system ahead of the data file. The
desired report is produced in a single pass. Minimum configuration is a 1441 with 4,000 core storage positions, one 1442 Card Read Punch, and one
1443 Printer. Disk storage cannot be utilized. A
1444 Card Punch and a 1442 Model 4 Card Reader
can be used in place of the 1442 Card Read Punch.
IBM 1440 Report Program Generator (RPG):
This generator facilitates the preparation of report
programs that will process an input file in punched
cards or 1311 Disk Storage to produce a report in
any combination of three forms: punched cards,
disc records, and/or printed reports. Disc records can be blocked, and master records can have
a variable number of trailer records. Block length
can be 100 to 1, 000 characters in multiples of 100
characters. Disc records can be processed either
sequentially or randomly. Processing sequence for
the random mode is specified by "processing-order
records" in punched cards or Disk storage.
.15
Data Transcription
Disk Utility Programs for IBM 1440/1311 or 14401
1301:
This is a set of nine related programs to facilitate
operation of 1440 systems with 1311 or 1301 Disk
Storage. The following minimum configuration is
required: 1441 ProceSSing Unit with 4,000 core
storage positions (8,000 if the Track Record feature
is used), one 1442 Card Read Punch, one 1443
Printer, and one 1311 Disk Storage Drive or 1301
Disk Storage array (two for the Copy Disk Program).
A 1444 Card Punch and a 1442 Model 4 Card Reader
can be used in place of the 1442 Card Read Punch.
One 7335 Magnetic Tape Unit is required for the
Disk-to-Tape and Tape-to-Disk programs. The
routines are supplied in the form of self-loading,
condensed card decks and are modified for each
processing run by specifications in control cards.
Each routine can operate in either the load or move
mode; i. e. , with or without word mark bits. The
individual routines are described below.
o Clear Disk Storage: Fills blanks or any specified character into all or any specified portions
of a Disk Pack or array. Minimum time to
process a full 1311 Disk Pack is 3.1 minutes.
©1964 Auerbach Corporation and Info, Inc.
4/64
IBM 1440
414:151.150
!! 151.
.15
.16
Data Transcription (Contd.)
IBM 1440 Disk File Organization Programs:
•
Disk-to-Card: Punches the contents of specified
areas of Disk Storage into cards - a maximum
of 70 characters per card in the move mode and
50 in the load mode. Minimum time to transcribe the contents of a full 1311 Disk Pack
would be 600 minutes.
CI
Card-to-Disk: Reloads cards produced by the
Disk-to-Card routine into the Disk Storage locations from which they were punched. Minimum
time to load a full 1311 Disk Pack is 107 minutes.
This series of routines facilitates the establishment and maintenance of data files in 1311 Disk
Storage. Minimum configuration requirements
are 4,000 core storage positions, one 1442 Card
Read Punch, one 1443 Printer, and one 1311 Disk
Storage Drive. A 1444 Card Punch and a 1442
Model 4 Card Reader can be used in place of the
1442 Card Read Punch. The series includes eight
different routines for random files and five routines for sequential files.
.17
o Copy Disk: Copies all or specified portions of
the data from one Disk Pack onto another Disk
Pack on a second Disk Storage Drive. The addresses written are the same as those on the
pack being read. Copying a full 1311 Disk Pack
takes from 4.3 to 7.0 minutes, depending upon
the relative positions of the index points on the
two drives.
o Print Disk: Prints the contents of specified
areas of Disk Storage without editing. Output
format is fixed at 100 data characters (one
sector) per line. Minimum time to dump the
contents of a full 1311 Disk Pack is 142 minutes.
o Disk-Record- Load: Loads data from punched
cards into any specified area of Disk Storage.
Each card contains the required control information in the first 20 columns and the data in the
remaining 60. In the load mode, only one field
can be entered per card. The routine can be
used to alter Disk Storage addresses as well as
data records.
o Disk-to-Tape: Writes the contents of specified
Disk Storage areas onto magnetic tape.
o Tape-to-Disk: Reads data from magnetic tape
into specified areas of Disk Storage.
o Disk-Label Program: Performs all necessary
maintenance operations on the label track of a
Disk Pack. It is used to set up the initial
header-label track, remove the entire label
track by clearing it and restoring the addresses
to the original range, enter a new label, delete
an existing label, make changes to labels, print
labels, and punch and print labels.
IBM 1440 Tape Utility Programs:
These programs are designed to transfer data files
from card to tape, tape to card, and tape to printer.
Output format flexibility is provided through fieldselection and exception record procedures. IBM
standard 120-chara.cter tape labels can be processed. Non-standard labels can be bypassed by the
Tape-to-Printer and Tape-to-Card programs.
Minimum machine requirements are 4,000 positions
of core storage, a 1442 Card Read Punch, a 1443
Printer, and a 7335 Magnetic Tape Unit. A 1444
Card Punch and a 1442 Model 4 Card Reader can be
used in place of the 1442 Card Read Punch.
4/64
File Maintenance
Others
Auto-Test:
Auto-Test is an integrated set of utility programs
designed to provide documentation for program
evaluation during a program testing run. It can be
used to test programs assembled by the 1440
Autocoder program for any IBM 1440 system. Because test runs can be planned to proceed with a
minimum of operator intervention, Auto-Test is a
useful tool for remote testing. The Auto-Test
system resides in a Disk Storage during testing.
It controls program s tests in which the following
functions can be selectively specified:
o Clear selected areas of Disk Storage.
o Load program data from cards to Disk Storage.
o Print selected areas of Disk Storage.
o Load the program to be tested from cards.
o Print data from selected areas of core storage
at specific times during program execution.
o Print data read from or written on Disk Storage
during program execution.
o Print core storage contents when program
execution terminates.
Several programs may be stacked for testing in a
. single run. A convenient card format for patching
program decks being tested is provided to facilitate
retesting of a program. Minimum machine require·ments are the following: 8K 1440 system (AutoTest 8K) or 16K 1440 system (Auto-Test 16K),
1442 Card Read Punch, 1443 Printer, one 1311
Disk Storage Drive, and the Indexing and Store
Address Register special feature.
IBM 1440 Application Programs:
This is a series of documented programs for specific industry applications which can be modified by
the user to fit his individual requirements. The
following programs are being provided:
o Demand Deposit Accounting
o Mortgage Loan Accounting
o On-Line Savings Program
414: 151.170
PROBLEM ORIENTED FACILITIES
§
151
. 17
.17
Others (Contd.)
o
Fire and Casualty Insurance:
Automobile Rating
Homeowners Rating
General Distribution
o Weekly Premium Insurance
• Secondary School Programs:
Student Scheduling
Grade Reporting
Payroll and Personnel
G Hospitals
Patient Accounting
Accounts Receivable
Accounts Payable
Payroll
Stores Inventory
General Ledger
Others (Contd.)
o Motor Freight Accounting:
Edit (revenue transaction data)
Transaction Register
Statement Writing Run
Cash Application to Accounts Receivable
Cash Application to Pro Control File
• Retail Accounts Receivable
Ci)
Retail Fashion Inventory Control
Q
Chain and Wholesale Billing (Grocery).
@1964 Auerbach Corporation and Info, Inc.
4/64
414:161.100
IBM 1440
Process Oriented Language
COBOL-61
PROCESS ORIENTED LANGUAGE: COBOL-61
§
161.
.14
.1
GENERAL
. 11
Identity: . . . . .
IBM 1440 COBOL.
. 12
Origin: . . . . . . . .
IBM Corporation.
. 13
Reference: . . . . .
IBM Publication C24-3112-0.
.14
Description
The 1440 COBOL language is a subset of COBOL61, the most widely implemented pseudo-English
common language for business applications. Although a number of electives have been incorporated,
1440 COBOL has several significant deficiencies
with respect to Required COBOL-61. These deficiencies result from the manufacturer's decision
to "defer" several important but not easily implemented language facilities. The deficiencies of
1440 COBOL relative to Required COBOL-61, as
well as the facilities of Elective COBOL-61 that
have and have not been implemented, are tabulated
at the end of this description.
The 1440 COBOL Processor converts programs
written in IBM 1440 COBOL language into 1440
Autocoder. The Autocoder processor then
assembles the machine-language object program.
The minimum confuguration required for the translating computer is 4,000 positions of core storage,
the Indexing and Store Address Register feature,
one 1311 Disk Drive, a 1443 Printer, and a 1442
Card Read-Punch.
For execution of machine language object programs
the object 1440 must have:
'
o
Sufficient core· storage to contain either the
entire object program or the largest single
overlay.
o
The Indexing and Store Address Register
feature.
o
All input-output units required by the source
program.
o
The Expanded Print Edit feature (if high order
CR, DB, -, or +; floating +, -, or $; check
protection; or decimal suppression for blank
or zero fields are used).
Instead of magnetic tape, which is the primary file
storage medium in most implementations of
COBOL-61, 1440 COBOL utilizes 1311 Disk Storage.
There are two types of file-description entries in
1440 COBOL. Unit record files such as those for
the card read-punch and the printer are described
by FD entries. Unblocked fixed-length disc
records can also be described by FD entries.
Description (Contd. )
Records for the card read-punch must not exceed
80 characters, while those for the printer must
not exceed the number of print positions. The
maximum size of a disc record is 999 characters.
Blocked files read from, stored on, or written on
1311 Disk Packs are described in MD (massstorage file) entries which contain the standard
FD clause and several additional ones. The user
can choose one of three access modes, depending
upon the disc file organization and processing
technique to be used:
o
Random Access - each record has a unique
disc address. Records can be arranged in
l'andom order, and facilities are provided
for specifying and computing the disc addresses
of the records to be processed.
o
Sequential Access - records are arranged in
sequence by control field and stored in consecutive disc locations.
o
Control-Sequential Access - records are
stored with a blank (sequence-link) field
appended to each record. Records to be inserted into the file can be stored in a separate
area and referenced by the link address.
All disc records are recorded and read without
word marks. Data can be recorded in disc
storage in one of four modes:
o
Sector - data records are written or read by
100-character sector.
o
Track-Sector-Address - a full track consisting of 2, 000 data characters (20 sectors)
and 100 address digits is written or read.
o
Track-Record - a full track consisting of
2,980 characters is written or read; the
Track Record feature is required on the
object 1440 system.
o
Track-Record-Address - a full track consisting of 2, 980 data characters and one 6digit address is written or read; the Track
Record feature is required.
IBM 1440 COBOL object programs can process disc
records in fixed-length unblocked, fixed-length
blocked, or variable-length blocked form. (The
latter form can be used only with the sequential
access mode.) Maximum record size is 999 characters in all cases. A block can contain a maximum
of 10 records in random files, 30 records in sequential files, and 100 records in control-sequential
files.
Among the el~ctives that have been incorporated
into 1440 COBOL, the COMPUTE verb is probably
@1964 Auerbach Corporation and Infa,lnc.
4/64
IBM 1440
414: 161.140
• 142 Deficiencies with Respect to· Required
COBOL-61 (Contd.)
I:J 161.
• 14
Description (Contd.)
the most valuable. COMPUTE permits arithmetic
operations to be expressed in a concise formula
notation similar to that of FORTRAN. For example,
the COBOL operations:
SUBTRACT B FROM A GIVING T
DIVIDE C INTO T GIVING X
can alternatively be expressed as:
COMPUTE X = (A - B)/C.
The library facilities of COBOL-61 are not available
in the 1440 COBOL system. The COBOL library
is a collection of prewritten entries for the Environment, Data, and Procedure Divisions that can
reduce the amount of writing involved in coding
programs and encourage standardization of coding
techniques. The other deviations from Required
COBOL-61 are relatively minor when COBOL
source programs are written specifically for the
1440, but are likely to cause serious problems
when COBOL programs written for other systems
must be compiled and run on a 1440.
Although the 1440 is a variable word length computer, the COBOL provisions for item lengths
which vary from run to run have not been implemented. Arithmetic operand sizes can be preset to any
value up to 18 digits.
The data description clauses USAGE, SIGNED, and
SYNCHRONIZED have no Significance in the 1440
system because of its variable word length capability and its use of the same representation for both
numeric and alphameric data. These three clauses
should not be used in 1440 COBOL source programs.
•
The RENAMING clause in the FILE CONTROL
paragraph, which enables more than one file
to utilize the same File Description without
the need to rewrite the description, is deferred.
•
All options that relate specifically to magnetic
tape files are inapplicable.
•
No ALTERNATE AREA may be specified in
the FILE-CONTROL paragraph.
•
No RERUN option is provided.
Data Division
Q)
The COpy options that enable File and Record
descriptions to be taken from the library are
deferred.
COl
Neither editing of a single-digit field nor
single-position zero suppression can be specified by editing clauses or picture clauses.
"
No item may exceed 999 characters.
•
The PICTURE clause characters "0" and "B"
are implemented as replacement characters
rather than as insertion characters, so nonstandard results may be obtained in editing
operations.
Procedure Division:
•
The REEL option of the CLOSE verb is not
applicable, since there are no provisions for
the use of magnetic tape •
•
A single-character field cannot be moved to
an editing field.
• 141 Availability
Language: ••••••••• 1963.
Compiler: • • • • • • . • • ?
• 142 Deficiencies with Respect to Required COBOL-61
Environment Division
4/64
o
The OPTIONAL option in the FILE-CONTROL
paragraph, which provides for files that will
not necessarily be present each time the object
program is run, is deferred.
•
The COPY options that enable SOURCE-COMPUTER, OBJECT-COMPUTER, and SPECIALNAMES paragraphs to be taken from the library
are deferred.
. 143 Extensions to COBOL-61
mM 1440 COBOL has added facilities for handling
files in 1311 Disk Storage. These consist of a
special MD entry in the file description section,
an INPUT-OUTPUT file option with the OPEN verb,
a branch on INVALID KEY option with the READ
and WRITE verbs, and a SEEK verb that permits
overlapping of disc seek operations with internal
processing.
The SORT verb and Report Writer facility of
COBOL-61 Extended are not provided.
PROCESS ORIENTED LANGUAGE: COBOL-61
§
414:161.144
161.
.144 COBOL-61 Electives Implemented (See 4:161.3)
Key No.
Elective
1
6
Characters and Words
Formula characters
Figurative constants
8
File DescriI!tion
BLOCK size
9
FILE size
11
SEQUENCED ON
19
Record DescriI!tion
SIZE clause option
20
Conditional range
22
24
Verbs
-COMPUTE
ENTER
26
Verb Options
USE
27
30
LOCK
ADVANCING
32
33
35
Formulas
Operand size
Tests
37
38
Compound conditionals
Complex conditionals
39
Conditional statements
41
Environment Division
OBJECT COMPUTER
42
SPE CIAL NAMES
Comments
+, -,
*, /, **, =.
HIGH-VALUE (S), LOW-VALUE(S).
allows an upper limit to be
specified, but not in the
standard way.
approximate size of disc files
can be shown by FILE-LIMITS
clause.
keys for disc files can be
specified by special ACTUAL
KEY and SYMBOLIC KEY
clauses.
can be used only to specify
size of variable length record.
allows a conditional value to be
a range.
permits algebraic formulas.
permits use of 1440 Autocoder
language within a COBOL program.
can be used only to specify a
KEY CONVERSION procedure
for disc files.
locks rewound tapes.
specifies paper advance of 1, 2,
or 3 lines, or to any channel
on carriage tape.
algebraic formulas can be used.
up to 18 digits.
IF. .• IS NOT ZERO form is
provided.
ANDs and ORs can be intermixed.
permits conditional statements
within conditional statements.
A T END, ON SIZE ERROR, and
INVALID KEY may follow imperative statements.
specifies differences from the
"standard" configuration.
specifies hardware for ACCEPT,
DISPLAY, and WRITE verbs.
© 1964 Auerbach Corporation and Info, Inc.
4/64
414: 161.145
IBM 1440
§
161.
.145 COBOL-6! Electives Not Implemented (See 4:161.3)
Key No.
Elective
2
3
4
Characters and Words
Relationship characters
Semicolon
Long literals
5
Figurative constants
7
Computer-name
10
12
13
File Description
Label formats
HASHED
Record Description
Table length
14
Item length
15
BITS option
16
RANGE IS
17
RENAMES
18
21
SIGN IS
Label handling
23
25
Verbs
--nEFINE
INCLUDE
28
Verb Options
MOVE CORRESPONDING
29
34
OPEN REVERSED
Relationships
36
Conditionals
40
4/64
,I
Environment Division
SOURCE-COMPUTER
43
44
FILE-CONTROL
PRIORITY IS
45
46
I/O-CONTROL
I/O-CONTROL
47
Identification Division
DATE-COMPILED
48
49
Special Features
Library
Segmentation
Comments
> and < not available.
always ignored by translator.
literals may not exceed 120
characters.
HIGH BOUND(S), LOW BOUND(S)
not available.
no alternative computer names.
Labels must be standard or
omitted.
hash totals cannot be created.
lengths of tables and arrays may
not vary.
variable item lengths cannot be
specified in a PICTURE.
items cannot be specified in
binary.
value ranges of items cannot be
shown.
alternative groupings of elementary items cannot be specified.
no separate signs allowed.
only standard labels (or none)
may be used.
new verbs cannot be defined.
no library routines can be called.
each item must be individual).y
moved.
tapes cannot be read backward.
IS UNEQUAL TO, EQUALS, and
EXCEEDS are not provided.
no implied objects with implied
subjects.
cannot specify differences from
the "standard" configuration.
cannot be taken from library.
no priorities can be specified for
multi-running purposes.
cannot be taken from library.
rerun methods and multi-file
tapes cannot be specified.
current date will not be inserted
automatically.
library routines cannot be called.
no provision for segmentation
of object programs.
/
414:162.100
IBM 1440
Process Oriented Language
FORTRAN IV
PROCESS ORIENTED LANGUAGE: FORTRAN IV
§
. 14
162.
.1
GENERAL
. 11
Identity: . . . . . • .
IBM 1401/1440/1460
FORTRAN IV.
. 12
Origin: • . . . . .
IBM Corporation.
. 13
Reference: .
IBM Publication C24-3155-1.
.14
Description
FORTRAN is an automatic coding system designed
primarily for scientific and engineering computations. The language closely resembles the language
of mathematics, and includes various types of
arithmetic, control, input-output and specification statements. The 1440 version also provides
for reading and writing in 1311 Disk Storage. The
source program, written in FORTRAN language
and punched into cards, is processed on a 1401,
1440, or 1460 system under control of the FORTRAN Compiler. The output is an object program
in machine language. Diagnostic messages are
provided automatically. The user may specify a
punched object deck; the system also provides for
automatic execution of object programs.
FORTRAN IV for the IBM 1401, 1440, and 1460
contains many facilities not found in 1401 FORTRAN. Four of the more significant are: the
CALL statement, Function and Subroutine subprograms, the DEFINE FILE statement, and the FIND
statement. Others are: the COMMON statement,
the DATA statement, the RETURN statement,
variable dimensions in subprograms, "Type" statements which allow explicit type specifications for
variables, logical expressions on the right side of
an equal sign, the Logical IF statement, new forms
for I/o statements, and the ability to read in
FORMAT statements at object time.
One of the most significant features of IBM 1440
FORTRAN IV is its ability to take advantage of the
variable word length capabilities of the Processing
Unit. The programmer can specify any desired
degree of precision up to 20 decimal digits for the
internal representation of numeric data. The precision (f) to be used for all real (floating point)
values within a single program is preset by a control card. If no specification is made, f is set at
eight digits. The number of core storage positions
required for each real variable of f digits precision
is f + 2, because two additional digits are required
to specify the exponent. Integer (fixed point) precision (k), which is also specified by a control card,
is set at five digits if no other specification is
made and applies to all integer values within a program. Object program execution times and storage
requirements will naturally increase when increased
precision is demanded.
Description (Contd. )
Minimum machine requirements for FORTRAN
compilation are 8, 000 core storage positions, one
1311 Disk Storage Drive, a 1442 Model 1 or 2
Card Reader Punch (or a 1442 Model 4 and a 1444
Card Punch), a 1443 Printer, Sense Switches,
Multiply-Divide, and the Indexing and Store Address Register feature. Magnetic tape units can
be used by the object program, but they are not
used by the FORTRAN Compiler.
The restrictions and extensions of the IBM 1440
FORTRAN IV language relative to IBM 7090/7094
FORTRAN IV are listed below. See Section
408:162 for a complete analysis of 7090/7094
FORTRAN IV.
.141 Availability
Language: • .
. .. 1963.
Compiler:. . . . . . . . . ?
.142 Restrictions (Relative to IBM 7090/7094
FORTRAN IV)
(1)
The following statements are not permitted:
ASSIGN
BLOCK DATA
COMPLEX
DOUBLE PRECISION
Assigned GO TO
PRINT
PUNCH.
(2) Double precision and complex operations are
not permitted (but note that precision can be
preset at up to 20 digits).
(3)
Symbolic coding cannot be incorporated into
the FORTRAN IV source program.
(4)
Input-output in octal form (O-type conversion)
is not permitted.
(5)
Complex and double precision functions are
not included; also, the TANH (real hyperbolic
tangent) function is not provided.
(6) In the PAUSE n statement, n can be an unsigned octal integer of only 1 to 3 digits.
(7)
The following library subprograms are not provided: OVERFL (test for floating point overflow condition), DVCHK (test divide check
indicator), EXIT (terminate execution of program), DUMP (dump core and terminate
execution), and PDUMP (dump core and
continue).
(8) An array name with adjustable dimensions
cannot appear in a COMMON statement.
©1964 Auerbach Corparation and Info,lnc.
4/64
IBM 1440
414: 162.142
§
162.
.142 Restrictions (relative to IBM 7090/7094
FORTRAN !Y) (Contd.)
(9) Block names may not be used in COMMON
statements.
(10) All variables forced to occupy the same locations as a result of an EQUIVALENCE statement must be of the same type.
(11) Dummy arguments in FUNCTION and SUBROUTINE statements follow different rules
from those of 7090/7094 FORTRAN IV.
.143 Extensions (Relative to IBM 7090/7094
FORTRI\.N IV)
(1) Data items can be represented internally with
any desired degree of precision between 2 and
20 digits for floating point items and between 1
and 20 digits for fixed point items, as preset
by control cards.
(2) A wider range of numeric magnitudes can be
represented in 1440 FORTRAN IV than in the
7090/7094 system, as follows:
1440
FORTRAN IV:
7090/7094
FORTRAN IV:
10-100 to
1 to 10k_1
(1-10- f ) x 10 99
10-38 to 1038
1 to 131,071
where f and k are real and integer precisions,
respectively, in decimal digits.
4/64
.143 Extensions (Relative to IBM 7090/7094
FORTRAN IV) (Contd.)
(3) Two new statements are available:
• DEFINE FILE divides the Disk Storage
units into any number of symbolic I/O
units for use in the READ, WRITE, and
FIND statements. For each symbolic
I/O unit, the DEFINE FILE statement
specifies: an integer constant that serves
as the file (or symbolic unit) name; the
number of records in the file; the length
of each record; whether or not the file is
edited (i. e., read with formatted READ/
WRITE statements); and the name of an
integer variable whose value is set to the
number of the next record at the conclusion of each READ/WRITE statement
referencing this file.
• FIND locates a particular record in Disk
Storage. The FIND statement indicates
positioning of the access mechanism to
the location of the Nth record of the
specified file while permitting computation to proceed concurrently. Use of the
FIND statement is not mandatory, but the
overlapping it makes possible can significantly improve object program execution
speeds.
(4) The FORMAT statement can specify a logical
(L-type) conversion.
414:171.100
IBM 1440
Machine Oriented Language
Autocoder
MACHINE ORIENTED LANGUAGE: AUTOCODER
§
. 141 1400 Basic Autocoder 2K (Contd.)
171.
.1
GENERAL
. 11
Identity: .
IBM 1440 Autocoder.
IBM 1440 Input/Output
Control System (IOCS).
(Specific versions of these
two basic systems are
identified and described
below.)
. 12
Origin: • . . • . . . .
IBM Corporation•
• 13
Reference: . . . . .
numerous IBM publications,
as listed in IBM 1440
Bibliography.
.14
Description
Autocoder is the basic machine oriented language
for the IBM 1440. The language format, and
facilities are very similar to those of the IBM 1401
Autocoder system, which is described in detail in
Section 401:172. Autocoder for the 1440 is currently available in six different versions designed
for different system configurations. The language
facilities and configuration requirements for each
version are summarized in Paragraphs . 141 through
.146.
The 1440 Input/Output Control System (lOCS) is
designed to reduce the amount of detailed coding
associated with input-output operations. The use
of lacs macro-instructions in Autocoder source
programs causes the Autocoder translator to
insert and generate linkages to the appropriate
generalized routines. (It is important to note that
laCS can be used only with the full Autocoder system, and not with Basic Autocoder.) The laCS
routines handle blocking and unblocking, error
correction, end-of-file procedures, and labeling.
Like Autocoder, rocs is available in several
versions designed for different system configurations. The facilities offered by these versions are
described in Paragraphs. 151 through. 155.
program after assembly, is part of the system.
Minimum machine requirements: 2K 1440 system
with 1442 Card Reader Model 4, 1444 Card Punch,
and 1443 Printer. (A 1442 model 1 or 2 may be used
as the reader in lieu of the 1442 model 4. )
.142 1440 Basic Autocoder
1440 Basic Autocoder is a symbolic programming
system designed for 1440 card systems with at
least 4, 000 core storage positions. Macro-instruction facilities are not available, and lacs cannot
be used. Source programs written in 1440 Basic
Autocoder language are punched into cards and
processed by the Basic Autocoder processor to
obtain a machine language object program. The
object program is punched, one instruction per
card, into the original source cards. A printed
listing is also prepared. The resultant deck can
then be loaded into core storage by a load routine
supplied with the system. Minimum machine requirements: 4K 1440 system with 1442 Card Read
Punch and 1443 Printer.
• 143 1440 Autocoder
The 1440 Autocoder system is an advanced symbolic
programming system that represents a significant
extension of 1440 Basic Autocoder. The language
provides continuity (but not direct compatibility)
with the 1401 and 1410 Autocoder systems. Some
of the important advantages of Autocoder over
Basic Autocoder are: macro-instruction facilities
(including laCS), more freedom with literals,
automatic assembly process through use of 1311
Disk Storage, and freedom from the need to reproduce source programs before reassembly.
Minimum machine requirements: 4K 1440 system
with a 1442 Card Read Punch, 1443 Printer, and
one 1311 Disk Storage Drive. A 1444 Card Punch
for punching and a 1442 Model 4 for reading may
be used in lieu of the 1442 Card Read-Punch, Model
lor 2.
.144 1440/1301 Autocoder
.141 1440 Basic Autocoder 2K
Basic Autocoder 2K is a symbolic programming
system designed to simplify the preparation of
programs for 1440 card systems with 2000 core
storage positions. The 1440 Basic Autocoder 2K
language is a subset of the 1440 Basic Autocoder
language, and offers all the facilities of the latter
language except literals and the use of symbolic
names as operands in Origin and Equate statements.
Source programs written in the Basic Autocoder
2K language and punched into 1440 Autocoder format
are processed by the Basic Autocoder 2K processor
to produce a macine language object program. A
listing routine to check for coding accuracy and consistency before assembling, and to list the object
This system is similar to 1440 Autocoder, described above, except that 1301 Disk Storage is
used instead of 1311 Disk Storage for automatic
assembly of programs. Minimum machine requirements: 4K 1440 system with a 1442 Card Read
Punch, 1443 Printer, and 1301 Disk Storage (one
array). A 1444 Card Punch for punching and a
1442 Modell, 2, or 4 for reading may be used in
lieu of the 1442 Card Read Punch.
• 145 1440 Autocoder - 1401 Processor
The 1440 Autocoder - 1401 Processor Programming
System is a 1401 assembly program which will
assemble programs written in 1440 Autocoder
@1964 Auerbach Corporation and Info,lnc.
4/64
IBM 1440
414:171.145
!I 171.
.145 Autocoder - 1401 Processor (Contd.)
language. By using a 1401 Data Processing System,
1440 users can write and assemble their programs
prior to the availability of their 1440 system. The
output of the assembly program is the 1440 machine
language object program punched into cards. A
1440 loader is supplied with the program. IOCS
macro-instructions are not processed by the 1401
Processor, and other minor language restrictions
must be observed when using this system. Minimum
machine requirements: 1401 system with 16,000
core storage positions, six magnetic tape units,
High-Low-Equal Compare, and Advanced Programming.
• 146 1440/1444 Basic Autocoder
1440/1444 Basic Autocoder is a special version of
Basic Autocoder for 1440 card systems that utilize
the 1444 Card Punch. Source programs written in
1440 Basic Autocoder language and punched into
1440 Autocoder format are processed by the Basic
Autocoder processor to obtain a machine language
object program. A printed listing is also prepared.
The resultant deck can then be loaded into core
storage by a load routine supplied with the system.
Minimum machine requirements: 4K 1440 system
with a 1442 Card Reader Model 4 (a 1442 Modell
or 2 may be used as the reader in lieu of the 1442
Model 4), 1443 Printer, and 1444 Card Punch.
.151 1440/1311 Input/Output Control System
1440/1311 IOCS is designed to facilitate the programming of input-output operations associated
with card reading and punching, printing, 1311
Disk Storage, and magnetic tape. No additional
features beyond those required by 1440 Autocoder
are required when assembling a program which
includes IOCS macro-instructions. The generated
input-output routines will, however, take advantage
of the Indexing and Store Address Register feature
if it is available on the object machine.
.152 1440/1301 Input/Output Control System
1440/1301 IOCS is similar to 1440/1311 IOCS,
except that routines for 1301 Disk Storage inputoutput are included, as well as those for 1311
Disk Storage, card read-punch, and printer. No
additional features beyond those required for 1440/
1301 Autocoder are required when assembling a
program which includes IOCS macro-instructions.
The generated input-output routines will, however,
take advantage of the Indexing and Store Address
Resigter Feature if it is available on the object
machine.
4/64
· 153 1440/1448 Input/Output Control System
IOCS for the 1440/1448 is designed to reduce the
programming effort required to control data transmission to and from remote terminals. It provides
pretested routines to handle such programming
functions as priority request, end-of-block detection, error detection, status alteration, control
word address initialization, output scheduling, and
coordination of the 1440/1448 system with other
input-output devices. It also schedules real-time
routines and defines what programming functions
can be performed within these routines. All functions are provided on an option basis so that the
user need not pay a storage penalty for routines
that are not required. The 1448 IOCS routines can
be added to the 1440 Autocoder Library in the same
way that 1440 IOCS is added. Minimum machine
requirements: no additional features beyond those
required when assembling a 1440 program using
1448 macro-instructions. The generated 1448
routines require the Indexing and Store Address
Register Feature and a minimum of 8, 000 positions
of core storage, in addition to the 1448.
• 154 1401/1440/1460 Direct Data Channel IOCS
IOCS for Direct Data Channel connection of two
1401, 1440, or 1460 systems in any combination
provides pre-tested routines to handle these
functions: program detection of a read request or
write request by either system, priority interrupt
request (included with the Direct Data Channel
Feature in 1440 and 1460 systems when the 1448 is
attached), error detection, output scheduling, system-to-system read-write, coordination with other
IOCS programs in either system, and scheduling of
the user's Direct Data Channel routine for each
system. No additional features beyond those required for 1440 Autocoder are required when assembling a 1440 program for transfer of data via
the Direct Data Channel. The generated routines
require the Indexing and Store Address Register,
Serial Input/Output Adapter, and Direct Data
Channel Features.
· 155 1440
Inte~system
IOCS
This version of IOCS provides all the functions of
1440/1448 IOCS plus the following functions for
servicing communication needs between interconnected 1440 and 1410/7010 systems: reading and
writing of data messages, reading and interpreting
of control messages, writing of sense data, writing'
of service messages, priority interrupt requests,
error detection, output scheduling, coordination
with other loeS programs, and scheduling of the
user's routine for processing records received
from the 1410/7010. No additional features beyond
those required for 1440 Autocoder are required
when assembling a 1440 program utilizing 1440
Intersystem IOCS. The generated routines require
the Indexing and Store Address Register, Bit Test,
and Expanded Serial Input/Output Adapter Features,
as well as 8, 000 positions of core storage.
414:211.101
IBM 1440
Physical Characteristics
PHYSICAL CHARACTERISTICS
§ 211.
Width,
inches
Depth,
inches
Height,
inches
Weight,
pounds
1441 Processing Unit
1447 Console, Modell
1447 Console, Model 2
45
48
55
31
29
29
60
44
44
1,500
175
400
2.5
0.1
0.3
1,700
150
550
7335 Magnetic Tape Unit, Modell
7335 Magnetic Tape Unit, Model 2
29
58
31
31
60
60
640
1,210
1.4
2.4
4,370
7,370
1301 Disk Storage
1311 Disk Storage Drive
86
30
33
24
69
38
3,625
430
7.5
0.8
16,700
2,000
1442 Card-Read Punch, Models 1, 2
1442 Card Reader, Model 4
1444 Card Punch
43
43
43
24
24
24
49
49
48
524
525
510
0.7
0.7
0.8
1,700
1,700
2,000
32
-32
112
56
44
37
37
42
43
24
60
60
61
46
45
529
570
2,745
725
620
1.8
1.8
2.7
0.8
1.15
4,100
4,100
6,300
2,000
4,000
29
31
40
500
0.3
1,000
26
25
45
38
15
16
31
31
27
25
60
70
195
90
1,500
500
0.375
0.1
2.0
3.1
1,280
275
5,100
3,000
Unit
1011
1012
1412
1443
1231
Paper Tape Reader
Paper Tape Punch
Magnetic Character Reader
Printer
Optical Mark Page Reader
1009 Data Transmission Unit
1051 Data Communications Control
Unit
1408 Transmission Unit
1448 Transmission Control Unit
7770 Audio Response Unit
Power,
KVA
BTU
per hr.
General Requirements
Temperature: . . • • •
Relative humidity:
60 to 90°F for all units except the following:
• 1412; 65 to 80°F.
• 1231, 1051, and 1408; 50 to 110°F.
10 to 80% for all units except the following:_
• 1412; 20 to 65%.
• 7335; 20 to 80%.
Power:
208!230V, 60-cycle, 3-phase, 4-wire for
all units except the following:
• 1231 and 1408; 115V, 60-cycle, I-phase,
3-wire.
• 1009 and 1051; 115/208/230V, 60-cycle,
I-phase, 3-wire.
• 1011, 1012-, 1412, and 7770 operate from
I-phase lines.
© 1964 Auerbach Corporation and Info, Inc.
4/64
414:221.101
IBM 1440
Price Data
PRICE DATA
§ 221.
IDENTITY OF UNIT
PRICES
CLASS
Name
No.
Central
Processor
1441
Model
Model
Model
Model
Model
A2
A3
A4
A5
A6
1470
4631
5275
7600
3835
8023
1447
Modell
Model 2
Model 4
1390
1490
6149
7600
Internal
Storage
Processing Unit
2,000 core storage
4,000 core storage
8,000 core storage
12,000 core storage
16,000 core storage
positions
positions
positions
positions
positions
Optional Features
Bit Test
Indexing and Store Address
Registers
Multiply-Divide
Sense Switches
Expanded Print Edit
Translate Feature
Console
Control section with power and
operator controls (required with
every system)
With inquiry printer and keyboard
(includes #2260 Console Attachment); used with Model 1
With inquiry printer and keyboard
plus controls for 1448 Transmission Control Unit (includes #2260
Console Attachment); used with
Modell
Optional Features
Attachment, 1051 Modell (for
Models 2 and 4)
Buffer Feature
Remote Terminal Attachment (for
Models 2 and 4; requires #1390)
Sense Switches, group of six (for
Modell only)
Monthly
Rental
Monthly
Maintenance
Purchase
$
$
$
770
970
1,295
1,620
1,945
37.50
38.50
39.50
40.50
41.25
53,100
56,350
73,850
91,350
108,850
20
0.50
800
90
325
15
20
60
1.00
9.00
0.50
0.50
2.50
4,950
11,700
550
750
3,000
80
1. 25
4,400
210
15.50
9,800
270
16.25
12,650
25
1.00
1,250
120
25
3.25
1. 00
6,000
1,250
15
.50
550
2,000
3,475
129.00
223.00
105,500
175,500
2,000
3,475
129.00
223.00
105,500
175,500
Core Storage: included in 1441
Processing Unit, above
1301
Model 11
Model 12
Model 21
Model 22
Disk Storage Unit
25.43 million characters max.
50.86 million characters max.
(First units on system; #3321 Disk
Storage Control and #3832 Expand
ed Disk Storage Control required)
25.43 million characters max.
50.86 million characters max.
(Additional units on system)
© 1964 Auerbach Corporation and Info, Inc.
4/64
414:221.102
IBM 1440
§ 221.
PRICE DATA (Contd.)
IDENTITY OF UNIT
CLASS
No.
Name
PRICES
Monthly
Rental
$
Internal
Storage
(Contd. )
1311
Modell
Model 2
1316
3281
6396
6400
8011
InputOutput
1442
Modell
Model 2
1442
Model 4
1531
6406
5880
4/64
28.00
17,610
360
27.00
16,510
Disk Pack (removable, inierchangeable disk storage for 1311)
15
Time/
Material
50
3.25
2,400
35
0.50
1,680
40
1. 75
1,950
40
0.50
1,920
Optional Features
Direct Seek (on 1441; one #3283
(no charge) required for Modell;
one #3282 (no charge) required
for Model 2)
Scan Disk (on 1441; one required
for both 1301's and 1311's)
Seek Overlap, on each 1311 (Seek
Overlap Adapter #6399 required
on 1441 - no charge)
Track Record (on 1441, for both
1301's and 1311's)
490
Card Read Punch (2 max.)
Reads 300 cpm, punches 80
col/sec.
Reads 400 cpm, punches 160
col/sec.
(1632 Adapter is required on
first 1442)
280
19.25
18,700
395
24.75
19,850
Card Reader
Reads 400 cpm (1632 Adapter is
required on first 1442)
200
23.00
11,500
30
20
0.50
0.50
1,650
1,100
20
0.50
1,000
375
43.50
21,600
325
450
34.25
45.50
19,750
19,900
45
165
25
2.25
4.00
2.00
2,475
9,800
1,375
-
-
Optional Features
Card Image
Selective Stacker (standard on
Model 2 and 4)
Punch Column Skip (Models 1 and
2; requires 5881 Punch Column
Skip Control)
1443
Printer (5567 Printer Control and
5561 Printer Attachment required)
Prints 150 lines/min.
Prints 240 lines/min.
Optional Features
Additional Print Positions (24)
Print Storage
Selective Character Set
Character Sets:
13 characters
39 characters
52 characters
63 characters
1890
1891
1892
1893
$
385
Card Punch
Punches 80-Column Cards at 250
cards per minute (Printer
attachment #5561 required)
5559
5585
6401
Purchase
$
Disk StQrage Drive
First drive on system (#3321 Disk
Storage Control required)
Additional drives (4 max)
1444
Model 1
Modell
Model 2
Monthly
Maintenance
400
450
475
500
PRICE DATA
§
414:221.1 03
PRICE DATA (Contd.)
221.
PRICES
IDENTITY OF UNIT
CLASS
No.
InputOutput
(Contd. )
7335
1009
1011
1012
1231
1412
7770
Modell
4667
4668
4669
8720
8721
Name
Magnetic Tape Unit (requires 7802
Tape Adapter)
Modell - one tape drive
Model 2 - two tape drives
Monthly
Rental
Monthly
Maintenance
Purchase
$
$
$
700
1,100
45.00
78.50
37,100.
57,100
Data Transmission Unit
Buffer for 1009
Paper Tape Reader
Tape Punch
Optical Mark Page Reader
Magnetic Character Reader
(The above five units require
the 7080 Serial I/O Adapter)
500
150
500
465
430
2,000
10.75
12.00
39.75
33.25
34.50
171.00
26,400
8,650
22,400
20,850
23,100
91,400
Audio Response Unit
(Requires 7770 Adapter #7150 on
1311 Modell)
1,200
38.50
57,600
175
200
75
100
100
12.50
2.50
2.00
2.00
2.50
8,400
9,600
3,600
4,800
4,800
100
3.50
5,790
1,150
29.00
56,700
150
250
50
20
10
25
235
115
100
2.00
8.00
4.50
0.50
0.50
0.50
5.50
1. 75
1.50
6,730
12,000
2,500
1,100
500
1,375
13,200
7,100
3,750
325
2.00
12,200
Optional Features
I/O Line Expander
I/O Line Frame
I/O Line Panel
Vocabulary Line Expansion
Vocabulary Words, Additional
'~
Controllers
1051
Modell
1448
8025
3321
3832
1632
5881
5561
5567
7802
7080
3845
Control Unit (Requires 1447 Model
2 or 4)
Transmission Control Unit
Transmission Control
Unit Attachment
Disk Storage Control
Expanded Disk Storage Control
Card Read Punch Adapter
Punch Column Skip Control
Printer Attachment
Printer Control
Tape Adapter
Serial I/O Adapter
(Required for 1009, 1011, 1012,
1231 or 1412)
Expanded Serial I/O Adapter
'"
© 1964 Auerbacn Corporatian and Info, Inc.
4/64
\
' - ____ J
IBM 1460
International Business Machines Corp.
c
AUERBACH INFO, INC.
PRINTED IN U. S. A.
IBM 1460
International Business Machines Corp.
/
AUERBACH INFO, INC.
PRINTED IN U. S. A.
415:001.001
IBM 1460
Contents
CONTENTS
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
l2.
14.
15.
415:011
415:021
Introduction....
Data Structure . • •
System Configuration
Notes on System Configuration
6-Tape Business System • . .
III
V
6- Tape Auxiliary Storage System
Internal Storage
1441 Core Storage . . .
1311 Disk Storage Drive
Disk Pack
Central Processor
1441 Processing Unit
1420 Bank Transit System
Console
1447 Console • . • • • .
Input/Output Printer
Input-Output; Punched Card
1402 Card Read-Punch (Reader)
1402 Card Read- Punch (Punch)
Input-Output; Printers
1403 Printer, Models 2 and 3
Input-Output; Magnetic Tape
729 Magnetic Tape Unit
7330 Magnetic Tape Unit
Input-Output; Other
7080 Serial Input-Output Adapter
1009 Data Transmission Unit
1011 Paper Tape Reader • • • .
1012 Tape Punch. • • . • • • •
1412 Magnetic Character Reader
1419 Magnetic Character Reader
1418 Optical Character Reader
1428 Alphameric Optical Reader
3271 Direct Data Channel •
Simultaneous Operations .. • • . • •
1461 Input/Output Control
Processing Overlap feature
Print Storage feature
Read Punch Release feature
Seek Overlap feature
Serial Input/Output Adapter
Instruction list . • • . .
Data Codes . . • • • . • • • . •
Problem Oriented Facilities . . • •
Data Sorting and Merging
1401 SORT 1 • . . . . .
1402 SORT 2 & MERGE 2
© 1963
by Auerbach Corporation and BNA Incorporated
415:031.001
415:031.1
415:031. 2
415:041
415:042
41,5:042.12
415:051
415:052
415:061
415:061.13
415:071
415:072
415:081
415:091
415:092
415:101
415:101.121
415:101.122
415:101.123
415: 101.124
415: 101.125
415:101.126
415: 101.127
415:101.128
415:111
415:111.12
415:111.12
415:111.12
415:111.12
415:111.12
415:111.12
415: 121
415:141
401:151
(IBM 1401)
401:151.13
401:151.13
401:151. 13
5/63
415:001.002
IBM 1460
CONTENTS (Contd.)
15.
16.
17.
18.
19.
20.
21.
22.
Problem Oriented Facilities (Contd.)
SORT 6 . • • • • • • • • . •
Report writing • . • • . . . •
1401 Card Report Program Generator •
1401 Tape Report Program Generator •
FARGO . • • • • • • • • • • . • •
Data Transcription . . • . • . • • .
Multiple Utility Program for IBM 1401 Tape System •
Programs for Card Systems.
• ••••.•
Disc Storage Utility Programs
1009 Utility Programs
7710 Utility Programs •
File Maintenance . .
Other • • • • • . .
Process Oriented Languages
COBOL . • . • . • •
FORTRAN . . . . . .
Machine Oriented Languages
Symbolic Programming System.
Autocoder . . . . . • . . •
Program Translators
Symbolic Programming System •
Autocoder . .
COBOL 4-8K .
COBOL 12-16K
FORTRAN •.
Operating Environment
System Performance
Notes on System Performance
Generalized File Processing
Physical Characteristics •
Price Data • . . . . . • . • •
RIP = Report in Process
5/63
401:151. 13
401:151.14
401:151. 14
401:151.14
401:151. 14
401:151. 15
401:151.15
401:151. 15
401:151.15
401:151.15
401:151.15
401:151.16
401:151.17
401:161 (IBM 1401)
401:162 (IBM 1401)
401:171 (IBM 1401)
401:172 (IBM 1401)
401:181 (IBM 1401)
401:182 (IBM 1401)
401:183 (IBM 1401)
401:184 (IBM 1401)
401:185 (IBM 1401)
401:191 (IBM 1401)
415:201
415:201. 001
415:201. 1
415:211 (RIP)
415:221
415:011.100
IBM 1460
Introduction
I NTRODUCTI ON
§ OIl.
The IBM 1460, announced in February 1963, is a new member of IBM's 1400 series
of small to medium scale business-oriented data processing systems. System rentals range
from approximately $4,000 to $20,000 per month, and most installations will probably fall
within the $S,OOO to $12,000 range. Flrst customer deliveries are scheduled for the fourth
quarter of 1963.
Throughput and Compatibility
The 1460 ranks between the IBM 1401 and 1410 systems in price and throughput, and
is dlrectly program-compatible with the 1401, Many 1401/1460 programs can also be run
without alteration on a 1410 by means of the 1410's built-in 1401 compatibility clrcuits. The
IBM 1440, the smallest member of the 1400 series, is program-compatible with the 1401 and
1460 with respect to internal processing, but uses different input-output units and different
instructions to control them.
The 1460 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 input-output units and optional features. The 140S (RAMAC)
Disk Storage Unit, 1404 Printer, and 1407 Console Inquiry Station are not currently available
for use in 1460 systems, so 1401 programs that utilize these devices must be reprogrammed.
To take full advantage of the 1460's increased internal speeds, it will be desirable to reevaluate the input-output timing considerations in existing 1401 programs and make changes
in timing loops and placement of input-output instructions. (Timing loops for 1412 or 1419
Magnetic Character Reader input must be changed.)
The principal advantages of the IBM 1460 over the IBM 1401 (described in Computer
System Report 401) can be summarized as follows:
o Core storage cycle time is 6 microseconds per character, compared to 11. S
microseconds in the 1401, so internal processing speeds are nearly twice as fast.
o The new 1403 Model 3 Printer can be used for alphameric output at a peak speed of
1,100 single-spaced lines per minute. (Models 1 and 2 of the 1403, used in 1401
systems, have peak speeds of 600 lines per minute.)
III
The 729 VI Magnetic Tape Unit, with a peak speed of 90,000 characters per sec0nd at a recording density of 800 characters per inch, can be used in 1460 systems
but not in 1401 systems.
o The free-standing, desk-style 1447 Console, with optional console input-output
typewriter, makes operation of the system more convenient.
o The Move Record and High-Low-Equal Compare processor facilities, which are
optional features in the 1401, are standard in the 1460.
The overall reduction in program run time that can be gained by replacing an IBM
1401 with a 1460 ranges from zero (for a run limited by the speed of a particular peripheral
unit or combination of units) to 48 per cent (for pure internal processing with no inputoutput). Test runs by the manufacturer of specific 1401 programs on a 1460 system have
shown the following reductions in over-all program run times:
o
Sorting of IS, 000 records: .
30% reduction.
•
Merging of 27,000 records:
38% reduction.
© 1963
by Auerbach Corporation and BNA Incorporated
4/63
IBM 1460
415:011.101
INTRODUCTION (Contd.)
§
OIl.
• Production control run: •
38% reduction.
•
FORTRAN compilation: •
26% reduction.
•
FORTRAN object program execution:
40% reduction •
Hardware
A 1460 system can have 8,000, 12,000 or 16,000 character positions of core storage. IBM 1401 systems can have as few as 1,400 positions of core storage, but because of
the severe programming limitations imposed by such a small store, the manufacturer has
wisely decided to require the inclusion of at least 8,000 positions in every 1460 system.
Each core position contains six data bits, a parity bit, and a word mark bit used to denote
the end of a variable-length field. Core storage cycle. time is 6.0 microseconds.
Up to five 1311 Disk Storage Drives can be used in a 1460 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 wlthout 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.
Tne 1441B Processing Unit is a solid-state, alphameric processor with add-tostore logic. A11 operations are performed serially by character, and both data fields and
instructions are variable in length. The basic instruction format consists of a onecharacter operation code, two 3-character operand addresses, and a one-character modifier; instruction length can vary from one to eight characters and averages about six characters. There is noaccumulator. Facilities for editing, three-way comparisons, and fullrecord internal transfers are standard, but multiplication, division, indexing, bit testing,
and sense switches are extra-cost options. Built- in floating point arithmetic is not available. Instructione are executed at the rate of about 7,000 to 8,000 per second in typical
routines.
Input-output control circuits for 1460 syste::ms are housed in a separate unit, the
1461 Input/Output Control, whereas they are cont.'lined in the ProceSSing Unit cabinet in
1401 systems . As in the 1401, system operation ~s basically serial in nature (Le., one
operation at a time). Little overlapping of input-output operations with 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 (described in Section
415:111) increases the system's capability for simultaneous operations, but also increases
programming complexity and input-output area storage requirements.
The 1402 Card Read-Punch provides a peak reading capability of 800 cards per minute. The 1403 Printers have 132 print positions and a 48-character print set. Peak speeds
are 600 and 1,100 single-spaced lines per minute for 1403 Models 2 and 3, respectively.
The 1403 Model 3 printer can nearly double the system throughput in printer-limited applications such as tape-to-printer data transcriptions. Only one Card Read-Punch and one
Printer can be used in a 1460 system.
Up to six 729 and/or 7330 Magnetic Tape Units can be connected. Peak data transfer rates range from 7,200 to 90,000 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 O;erlap,
internal processing can be overlapped with tape start-stop times and (at transfer rates of
41,667 characters per second or below) with character transfers to or from a tape unit.
4/63
415:011.102
INTRODUCTION
INTRODUCTION (Conld.)
§011.
The Serial Input/Output Adapter permits connection of anyone of the following devices: 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 IBM 1401, 1440, or another
1460. The 1448 Transmission Control Unit permits connection of 1062 Teller Terminals
for on-line processing of banking or savings and loan transactions.
Software
Software availability for the mM 1460 can be summarized by noting that all programs
and programming systems for the mM 1401 except those that require 1405 Disk Storage will
be directly usable on a Similarly equipped 1460. The extensive repertoire of 1401 programs
supplied by the manufacturer includes:
III
II)
SPS-l and SPS-2: basic symbolic assembly systems, usable on a card-only 1460.
1401 Autocoder: more advanced assembly system, providing macro facilities;
requires four magnetic tape units on the translating 1460.
o 1401-1311 Autocoder: utilizes 1311 Disk Storage instead of magnetic tape; otherwise similar to 1401 Autocoder.
o Input-Output Control Systems: provide macro instructions and corresponding
generalized routines to facilitate coding of input-output operations; four versions
are available for different system configurations.
\) Report Program Generator: facilitates preparation of programs to produce printed
reports from punched cards, magnetic tape, or 1311 Disk Storage.
o FARGO: a "load-and-go" report generator that produces mM 407-type printed
reports •
• Sorting and Merging: four generalized routines to handle sort/merge operations
using either magnetic tape units or 1311 Disk Storage.
o Auto-Test: expedites testing and debugging of programs coded in Autocoder, SPS,
and FARGO.
o Disk File Organization Programs: nine programs to assist in establishing and
maintaining data files in 1311 Disk Storage, in either random or sequential
arrangements •
., Utility Programs: a variety of routines to perform frequently needed functions
such as data transcription, multiplication, and program loading in card, tape,
and Disk Storage systems.
o COBOL: compiles programs coded in COBOL; 2 versions are available, for 1460
systems with 8,000 and 12,000 or more core storage positions; both versions
require 4 magnetic tape units.
o FORTRAN: compiles programs coded in a severly restricted subset of the full
FORTRAN language; magnetic tape is not required, and "load-and-go" operation
is possible.
For detailed descriptions of these programs and programming systems, refer to the
mM 1401 report, Sections 401:151 through 401:191.
mM 1420 Bank Transit System
The 1420 Bank Transit System is a variation of the 1460 that is specially engineered
for bank transit applications. The 1421 Bank Transit Processing Unit combines most of the
© 1963
by Auerbach C~rporation and BNA Incorporated
4/63
IBM 1460
415:011.103
INTRODUCTION (Contd.)
§Oll.
functions of the 1441B Processing Unit and the 1419 Magnetic Character Reader in a single
physical unit. Magnetically inscribed card or paper documents can be processed at speeds
of up to 1,600 documents per minute. From 4, 000 to 16,000 character positions of core
storage can be used, and cycle time is 6 microseconds per character as in the 1460 system.
The only input-output devices that can be connected to a 1421 are the 1442 Card Read-Punch
(which reads up to 400 cards per minute or punches up to 160 columns per second) and the
1403 Modell or 2 Printer (which prints up to 600 alphameric lines per minute). Programming of the 1420 system differs from programming of a 1401/1419 combination only in that
timing relationships are altered and the 1442 replaces the 1402 for punched card inputoutput. See Section 415:052 for a more complete description of the 1420.
4/63
415:021.100
.STAltDARD
EDP
_
REPORTS
IBM 1460
Data Structure
DATA STRUCTURE
§
021.
.1
.2
STORAGE LOCATIONS
Name of
Location
Character
position:
Sector:
Cylinder:
Colwnn:
Size
INFORMATION FORMATS
Type of Information
Representation
Nwneral: .
1 character or 1 card
Letter: . .
1 character or 1 card
Purpose or Use
8 bits (6 data, 1
parity, 1 word
mark)
100 characters
basic addressable
location; holds 1
character.
record location in
1311 Disk Storage.
20,000 characters volwne accessible
without repositioning in 1311 Disk
Storage.
12 hole positions punched cards.
© 1963
column.
Instruction:
Nwnber (field):
Block: . . • . •
by Auerbach Corporation and BNA Incorporated
colwnn.
1 to 8 characters, delimited
by word mark bit.
1 to N characters, delim ited
by word mark bit.
1 to Ncharacters, delimited
.by record or group mark.
where N is limited by
size of core
storage
4/63
415:031.001
IBM 1460
System Configuration
SYSTEM CONFIGURATION
§
031.
The basic rules for combining components to form an IBM 1460 Data Processing
System are listed below. For prices, refer to Section 415:221. For optional features
applicable to any unit, refer to the report section covering that unit.
CONFIGURATION RULES
Required in Every System:
•
one 1441 Processing Unit; choice of:
Model 84; 8, 000 core storage positions.
Model 85; 12,000 core storage positions.
Model 86; 16,000 core storage positions.
• one 1447 Console; choice of:
Modell; no console typewriter.
Model 2; includes console typewriter.
Model 4; includes console typewriter and controls for 1448 Transmission
Control Unit.
• one 1461 Input/Output Control; choice of:
Modell; for Card Read-Punch and Printer.
Model 2; for Card Read-Punch, Printer, and 729 Tape Units.
Model 3; for Card Read-Punch, Printer, and 7330 Tape Units.
For Punched Card Input/Output, Add:
• one 1402 Model 3 Card Read-Pundi.
For Printed Output, Add:
• one 1403 Printer; choice of:
Model 2; 600 alphameric lines/minute.
Model 3; 1,100 alphameric lines/minute.
For Magnetic Tape Input/Output, Add:
• one to six MagnetiC Tape Units; choice of:
729 II }
reqUire 1461 Model 2 Control; can be intermixed.
729 IV
729 V
729 VI
7330; require 1461 Model 3 Control; can be connected to 1461 Model 2
along with 729s if Tape Intermix Feature is added.
For Auxiliary Storage, Add:
• one to five 1311 Disk Storage Drives; first drive on a system must be
Modell, remainder Model 2; Disk Storage Control is required.
© 1963
by Auerbach Corporation and BNA Incorporated
4/63
415:031.002
IBM 1460
SYSTEM CONFIGURATION (Contd.)
§
031.
For Additional Input/Output, Add:
• one Serial Input/Output Adapter, connected to any one of the
following units:
1009 Data Transmission Unit
1011 Paper Tape Reader
1012 Tape Punch
1412 Magnetic Character Reader
1419 Magnetic Character Reader
1418 Optical Character Reader
1428 Alphameric Optical Reader
3271 Direct Data ehannel
For Data Transmission Control, Add:
• one 1448 Transmission Control Unit, which can be connected to a
1060 Data Oommunications System for on-line banking or savings
and loan applications.
4/63
·51. ' '
415:031.100
EDP
•
REPORTS
IBM 1460
System Configuration
§
SYSTEM CONFIGURATION
031.
.1
6-TAPE BUSINESS SYSTEM: CONFIGURATION III
Deviations from Standard Configuration: . . . . . . ,
Magnetic tape units are 11,667 char/sec faster.
Card reader is 300 cards/minute faster.
Card punch is 150 cards/minute faster.
Console typewriter provides input as well as output.
Rental
Equipment
Core Storage:
16,000 characters
Processing Unit:
1441 Model B6
)
Input/Output Control:
1461 Model 2
1,980
1447 Model 2 Console
and Attachment
290
1402 Card Read-Punch:
Reads 800 cards/minute
Punches 250 cards/minute
1403 Model 2 Printer: 600 lines/minute
729 II Magnetic Tape Units (6):
15,000 or 41,667 char/sec.
Optional Features Included: . . . . . . . . . . . . . Indexing & Store Address Registers
Multiply- Divide
Processing Overlap
Sense SWitches
Expanded Print Edit
Total Monthly Rental : .
© 1963
$ 3,230
by Auerbach Corporation and BNA Incorporated
560
775
4,200
90
325
250
15
20
$11,735
4/63
415:031.200
IBM 1460
§ 031.
.2
6-TAPE AUXILIARY STORAGE SYSTEM: CONFIGURATION V
Deviations from Standard Configuration: •
. Magnetic tape units are
11.667 char/sec. faster
Card reader is 300/cards/
minute faster.
Card punch is 150 cards/
minute faster.
Console typewriter provides
input as well as output.
Disk Storage is smaller by
5,100,000 characters.
Equipment
Rental
1311 Disk Storage Drives (5):
14,900,000 positions total
$2,150
Core Storage: 16,000 characters
Processing Unit: 1441 Model B6
}
Input/Output Control: 1461 Model 2
~-o
1,980
1447 Model 2 Console and Attachment
290
1402 Card Read-Punch:
Reads 800 cards/minute
Punches 250 cards/minute
560
1403 Model 2 Printer: 600 lines/minute
775
729 II Magnetic Tape
Units (6): 15,000 or 41,667 char/sec.
4,200
Optional Features Included: • . . . . . . . . . . . . . Indexing & Store Address Registers
Multiply- Divide
Processing Overlap
Sense Switches
Expanded Print Edit
Track Record
Direct Seek
Total Monthly Rental: .
4/63
3,230
90
325
250
15
20
40
50
$13,975
415:041.100
•
STANDARD
EDP
•
IBM 1460
REPORTS
Internal Storage
Core Storage
INTERNAL STORAGE: CORE STORAGE
§
041.
.1
• 13
GENERAL
• . . • . • • Core Storage.
contained in 1441 Processing Unit, Models B4, B5,
B6.
• 11
Identity:
. 12
Basic Use: .
.13
Description
working storage •
Core storage for the IBM 1460 system is housed in
the 1441 Processing Unit. Models B4, B5, and B6
contain 8, ODD, 12, ODD, and 16,000 character positions of storage, respectively. Cycle time is
© 1963
Description (Contd. )
6 microseconds for each access of one alphameric
character .. Each storage position consists of eight
bits: six data bits, one odd parity bit, and one word
mark bit.
The word mark bit defines the size of data fields and
instructions. Internal transfer operations can be terminated by a word mark or by a record or group
mark, depending upon the instruction used. The
"move record" instruction, which is part of the optional Advanced Programming feature for the IBM
1401, is standard in the 1441. The effective internal
transfer rate is 83, 300 characters per second.
Parity checks are made on all internal transfers.
by Auerbach Corporation and BNA Incorporated
4/63
.. ___ I
I
415:042. 100
IBM 1460
Internal Storage
Disk Storage
INTERNAL STORAGE: DISK STORAGE DRIVE
§
. 13
042.
.1
GENERAL
.ll
Identity: .
Disk Storage Drive.
13ll Models 1 and 2.
.12
Basic Use: .
auxiliary storage.
.13
Description
The 1311 Disk Storage Drive is a new development in
low cost random access storage. It is available for
the IBM 1401, 1410, 1440, and 1620 Data Processing
Systems as well as the 1460, and features rapid interchangeability o~ the "Disk Pack" storage cartridges. The system is suitable for either random
or sequential processing methods.
Each Disk Pack consists of six discs on a common
vertical axis. Data can be recorded on 10 disc surfaces; the top and bottom surfaces of the pack are not
used. Each recording surface is divided into 100
concentric tracks, each track is divided into 20 sectors, and each sector holds a 5-character address
and up to 100 alphameric characters of data. Therefore, the data capacity is 2, 000 characters per
track, 200, 000 characters per surface, and
2, 000, 000 characters per pack. Up to five Disk
Storage Drives can be connected to a 1440 system,
so the maximum on-line data capacity is 10, 000, 000
characters. (The optional Track Record feature,
described below, increases the capacity of each pack
from 2, 000, 000 to 2,980,000 characters.)
Access is by means of a comb-like mechanism containing five arms that move horizontally between the
discs. Each arm has one read-write head mounted
on the top and one on the bottom, and each head
'serves one disc surface. The entire access mechanism moves as 1 unit, so all 10 read-write heads are
always positioned at corresponding tracks on their
respective surfaces. The term "cylinder" is applied
to each group of 10 tracks (Ion each disc surface)
that can be read or recorded upon at a single setting
of the access mechanism. There are 100 cylinders
per Disk Storage Drive, and each cylinder can hold
20, 000 data characters.
Time for access mechanism movement ranges from
zero (for successive references to a previouslyselected cylinder) to 400 milliseconds; average random access time is 250 milliseconds. Unless the
optional Direct Seek feature is installed, the access
arms cannot move directly from one cylinder to
another. Instead, the arms retract all the way to the
''home'' position (beyond track 00) and then move
back to the selected cylinder. The result is that
movements between adjacent cylinders require from
85 milliseconds (track 00 to 01) to 390 milliseconds
(track 98 to 99).
© 1963
Description (Contd. )
Rotational speed of the discs is 1,500 rpm. Maximum rotational delay is 40 milliseconds, and the average is 20 milliseconds. There is also a head select delay of 2 milliseconds. Total reference cycle
time to read a randomly placed lOa-character record,
update it, rewrite it, and execute a programmed
write check is 354 milliseconds. If no access motion
is required the total reference cycle time is reduced
to 104 milliseconds. Peak data transfer rate is
77, 000 characters per second, and the effective bulk
transfer rate is just under 50, 000 characters per
second.
A single read or write instruction can transfer from
1 to 200 consecutive sectors of information; 1. e. ,
from 100 characters to the capacity of core storage
in mUltiples of 100 characters. The programmer can
elect to read and write sector addresses along with the
data records. Handling of variable-length disc records is facilitated by "sector counter overlays" in
which the first 3 characters of a record specify the
numbe.r of sectors (from 2 to 200) compriSing that
record.
All capacities and transfer rates quoted here are
based on operation in the "move" mode, in which six
data bits and one parity bit are recorded for each character. In the 'alternative "load" mode, the word mark
bit is also recorded for each character, and sector
capacity is reduced from 100 to 90 data characters.
All capacities and transfer rates for the load mode
are therefore 10 per cent lower than the figures quoted
here. Use of the load mode is essential for program
storage and for data storage when field lengths vary
from record to record.
Checks are made for parity errors, wrong length records, and unequal address comparisons. The "write
disc check" instruction causes a character-by-character comparison of data just written on the disc with
the data in core storage. It usually follows each
write operation. All disc errors cause the setting of
testable indicators.
Disk Storage Drive seek time can be fully overlapped
with internal processing. A "branch if access mechanism busy" instruction is provided. No overlapping
is possible during disc read or write operations. Only
one seek operation may go on at a time, regardless of
the number of Disk Storage Drives in a system, unless
the Seek Overlap feature, described below, is added.
The removable Disk Packs are 14 inches in diameter,
4 inches high, and weigh less than 10 pounds, including covers. A Disk Pack can be removed from a Disk
Storage Drive and replaced by another Disk Pack in 1
minute. When a Disk Pack is not mounted on a drive,
the pack and its cover combine to form a sealed container that can be conveniently stotl;d and transported.
by Auerbach Corporation and BNA Incorporated
4/63
415:042.130
IBM 1460
Description (Contd.)
• 13
Description (Contd. )
One Disk Pack is supplied with each 1311 Disk Storage Drive. Additional Disk Packs cost $490 each,
f. o. b. San Jose.
Opti~nal
Features
Direct Seek: Pennits the access mechanism to move
directly to the specified cylinder without returning to
the "home" position. Access motion time ranges
from zero to 250 milliseconds and averages 150
milliseconds.
Track Record: Permits reading and writing a full
track as a single 2. 980-character record, thereby
4/63
increasing the capacity of each Disk Pack from
2,000,000 to 2,980,000 characters. The increased
capacity is achieved by using the areas that nonnally
contain sector addresses for data storage. Complete
track records can be intennixed with sector organh
zation of records on other tracks on the same Disk
Pack.
Scan Disk: Pennits an automatic search of data recorded in disk storage for a specific identifier or
condition.
Seek Overlap: Penn its a disc seek operation to overlap one disc read or write operation plus any number
.of othe:.; seek operations. The feature must be installed on every Disk Storage Drive in a system.
415:051.100
IBM 1460
Central Processor
CENTRAL PROCESSOR
§
OS1.
.12
.• 1
GENERAL
.11
Identity: .
.12
Description
Processing Unit.
1441 Models B4, BS, B6.
The 1441 is a two-address, add-to-storage,
character-oriented processor. Its internal logic
and instruction structure are identical to those of the
IBM 1401 Processing Unit, described in Section
401:051, and programs written for a 1401 system
can, in general, be run without modification on a
similarly-equipped 1460 system. Models B4, B5,
and B6 have identical processing capabilities; they
differ only in having 8,000, 12,000, and 16,000
character positions of core storage, respectively,
with a cycle time of 6 microseconds. (A 1441
Processing Unit is the central component of the IBM
1440 system as well as the 1460. A 1440 system,
however, requires the 1441 Model A3, A4, AS, or
A6, with a core storage cycle time of 11.1 microseconds.)
Processor operations are performed serially by
character and terminated when a word mark bit or
a record or group mark character is sensed, so
operand sizes are variable from one character to the
limit of core storage capacity. Instruction length is
variable from one to eight characters. Most arithmetic and data transfer instructioITS. consist of a
one-character operation code and two operand addresses of three characters each. In other instructions, one or both addresses are eliminated and/or
a one-character modifier is added. Through careful
placement of data, instructions can sometimes be
"chained" so that a one-character instruction does
the w.ork of a seven-character one, saving both storage space and execution time.
The 1441 Processing Unit is well suited to general
data manipulation and has powerful editing capabilities. The ability to move full, multi-field records within core storage by a single instruction,
which is part of the optional Advanced Programming
feature for the 1401, is standard in the 1441. The
High-Low-Equal Compare feature is also standard.
Hardware facilities for multiplication, division·, indexing, and bit testing are all extra-cost options, as
described below. Built-in facilities for floating point
arithmetic are not available.
Execution of an instruction in the 1441 requires the
same number Of cycles as in the 1401, and the'
1441B's core storage cycle time of 6 microseconds
is 5.5 microseconds faster than that of the 1401.
Therefore, the 1441B. is nearly twice as fast internally as the 1401, and 1441B execution times can be
derived by multiplying the 1401 execution times in
Paragraph 401:051. 4 by the factor 0.522 (1. e. ,
Description (Contd.)
6.0/11.5). The following performance times are
based on fixed point operations upon fields D digits
in length and are expressed in microseconds.
For Typical Tasks
c =a +b: .••
108 + 24D.
b =a +b: ••.
60 + 120.
Sum N items: .
60 + 120.
c =ab, using
243 + 18000 + 73D2.
subroutine: .
c = ab, with MultiplyDivide feature:
108 + 66D + 3002•
c = alb, using
4530 + 15400 + 90D2.
subroutine: • . .
c =a/b, Wi~
Multiply-Dlvide
feature: ••.•
102 + 60D + 420 2•
For Arrays of Data
ci '=' ai + bj' with
Indexing. feature:
ci = ai + bj' without
Indexing feature:
c = c + aibj, using
subroutine: • . •
c = c + aibj. with
Multiply-Divide
feature: • . . .
Moving, per item: • .
+ 24D.
612 + 24D.
400 + 1800D + 73D2.
480 + 78D
48
+ 300 2•
+ 120.
Whereas the input-output control facilities for an
IBM 1401 or 1440 system are housed in its Processing Unit, the 1460 system requires a .separate control
unit: the 1461 Input-Output Control, Modell, 2, or
3. The 1461 and the optional features that can be
added to it are described in Section 415:111,
Simultaneous Operations.
Optional Features
Bit Test: Permits testing for the presence of any
speeified bit in any core storage location.
Expanded Print Edit: Adds check protection, floating
dollar sign, decimal control, and sign control left to
the standard editing capabilities.
Indexing and Store Address Registers: Makes three
3-character fields in core storage available as index
registers that can index any instruction address, and
provides instructions that store the contents of the A
and B registers to facilitate subroutine linking. (The
convenience of the index registers is somewhat limited by the fact that no special instructions are provided for setting, incrementing, or testing them. )
Multiply-Divide: Permits direct multiplication and
division; when this feature is not installed, subroutines must be used.
First Delivery:
© 1963
504
by Auerbach Corporation and BNA Incorporated
. • 4th quarter 1963.
4/63
415:052.100
_STANDARD
EDP
•
REPORTS
IBM 1460
Central Processor
IBM 1420 Bank Transit System
CENTRAL PROCESSOR: IBM 1420 BANK TRANSIT SYSTEM
§
.12
052.
.1
GENERAL
.11
Identity:
. . . • . . . IBM 1420 Bank Transit
System.
Bank Transit Processing
Unit.
1421 Models A3. A4. AS.
A6.
Card Read-Punch.
1442 Models 1 and 2.
Printer
1403 Models 1 and 2.
. 12
Description (Contd.)
Punch replaces the 1402 Card Read-Punch for
punched card input and output. The Indexing and
Store Address Register feature and the Bit Test
feature. which are extra-cost options for 1460 systems, are standard in the 1421.
Internal processing is not overlapped with reading of
data from magnetically inscribed documents unless
the optional Processing Overlap feature is added.
The optional Endorser feature enables the 1421 to
print a full bank endorsement on the back of ea-ch
check at no reduction in processing speed.
The only input-output units that can be connected to
the 1421 are the 1442 Card Read-Punch and the 1403
Printer. The 1442 is a combination input-output unit
Description
The IBM 1420 Bank Transit System is a specialpurpose variation of the 1460 that is designed for
Federal Reserve banks and commercial banks having
large transit volumes. The 1421 Bank Transit
Processing Unit combines most of the functions of
the 1441B Processing Unit and the 1419 Magnetic
Character Reader into a single physical unit. (See
Sections 4151051 and 401:103 for detailed descriptions of the 1441B and the 1419. respectively.) Card
or paper documents that have beeq magnetically inscribed with E13B type font can be processed at
speeds of up to 1,600 documents per minute. Document feeding and selection of one of the 13 pockets
are always controlled by the stored program. Documents ranging from 2. 75 to 3. 67 inches in width and
from 6. 00 to 8. 75 inches in length can be processed.
Card and paper documents of varying sizes can be
intermixed.
The 1421 can contain from 4.000 to 16.000 character
positions of core storage with a cycle time of 6 microseconds per character. Programming of the 1420
system is essentially the same as programming of a
1401/1419 combination, except that timing relationships are altered and the slower 1442 Card Read-
© 1963
for standard 80-column punched cards. From a
single I, 200-card hopper. cards are fed serially by
column past a photoelectric reading station, past a
die punching station. and into a 1,300-card radial
stacker. A second stacker, which can be programselected for reject or hold items, is standard on
Model 2 and optional on Modell. The 1442 Modell
has a peak reading speed of 300 cards per minute and
a peak punching speed of 80 columns per second (or
50 full cards per minute). Model 2 can read up to
400 cards per minute and punch up to 160 columns
per second (or 88 full cards per minute). A maximum of two 1442s, in any combination of models, can
be used in a 1420 Bank Transit System. Because each
1442 has only one card feed, two 1442s must be installed if the card reading and punching functions are
to be separated. A more detailed description of the
1442 Card Read-Punch is presented in Section 414:071.
One 1403 Modell or 2 Printer can be connected for
printed output at a peak speed of 600 alphameric lines
per minute. With the optional Numerical Print Feature,
up to 1,285 lines per minute of all-numeric output can
be printed. Models 1 and 2 differ only in having 100
and 132 print positions. respectively. The 1403
Printer is fully described in Section 401:081.
by Auerbach Corporation and BNA Incorporated
4/63
IBM 1460
415:052.120
§
052.
• 12
Description (Contd.)
First Delivery: 2nd quarter 1964.
Prices
Unit
Monthly
Rental
1421 Bank Transit Processing Unit
Model A3: 4K core storage
Model A4: 8K core storage
Model AS: 12K core storage
Model A6: 16K core storage
Monthly
Maintenance
Purchase
$4,305
4,755
5,205
5,655
$370.00
378.00
385.00
390.00
$210,800
235,100
259,400
283,700
1442 Card Read-Punch
Modell
Mode12
Adapter (req'd. on first 1442)
280
395
20
20.50
26.50
0.50
18,700
19,850
1,100
1403 Printer
Modell
Mode12
725
775
130.00
140.00
32,900
34,000
375
250
20
75
225
190
32.00
13.75
0.50
18,350
15,000
1,100
3,125
9,050
8,100
Optional Features
3791
5730
6406
4740
5381
6411
Endorser (on 1421)
Processing Overlap (on 1421)
Selective Stacker (1442 Modell)
Interchangeable Chain (on 1403)
Numerical Print (on 1403)
Selective Tape Listing (on 1403)
8.00
10. 75
Physical Characteristics
Unit
1421
1421 with
Endorser
1442
1403
Max. Weight
(lbs)
Height x Width x Depth
(inches)
Load
KVA
BTU/hr
3,900
61 x 157 x 41
5.4
14,900
3,900
525
750
61 x 157 x 41
49 x 43 x 24
53 x 48 x 29
5.9
O. 7
15,800
1,700
3,000
Power required: 208 or 230 volts AC, 3-phase, 4-wire, 60 cycles.
4/63
•
415:061.100
STANDARD
EDP
•
REPORTS
IBM 1460
Console
CONSOLE
§
061.
• 13
.1
GENERAL
• 11
Identity: .
• 12
Associated Units: • . . Input-Output Printer is
included in Models 2 and 4.
. 13
Console .
1447 Models 1. 2. and.4.
Description
The 1447 Console is a free-standing desk that houses
the manual controls for an IBM 1460 system. Models
2 and 4 include the Input-Output Printer (a Console
Typewriter). but Modell does not. Model 4 also includes the lights and switches required to control a
1448 Transmission Control Unit~ All controls and
displays are on a vertical panel mounted above and
to the rear of the desk top. just below eye level for
a seated operator. The controls and displays are
nearly identical to those on the IBM 1401 Console
(Section 401:061). tho,u&h their physical arrangement
© 1963
Description (Contd.)
is different. The 1447 Console will provide considerably more comfort and convenience for the operator than the 1401 Console. which is built into the
front panel of the Processing Unit cabinet and designed for operation from a standing position •
The Input-Output Printer is a single-case IBM
Selectric typewriter rated at 14. 8 characters per
second. Pitch is 10 characters per inch and line
spacing is 3 or 6 lines per inch. The maximum
writing line width is 8.5 inches. Input-Output Printer
operations cannot be overlapped with internal processing. The unit is useful for entering and printing
exception data. for logging. and as an inquiry station.
(The 1407 Console Inquiry Station, an optional unit
that performs console input-output functions in 1401
systems. is not available for the 1460.)
Optional Feature
Sense Switches: Provides six console switches a)ld
corresponding testable indicators that can be used
for manual control of the stored program.
by Auerbach Corporotion and BNA Incorporated
4/63
415:071.100
STANDARD
•
EDP
•
IBM 1460
Input-Output
1402 Card Reader
REPORTS
INPUT-OUTPUT: CARD READ-PUNCH (READER)
§
.12
071.
.1
GENERAL
. 11
Identity: .
• 12
Description
Card Read-Punch (Reader).
1402 Model 3.
Description (Contd.)
capacity and 3 stackers with 1, ODD-card capacities
(1 shared with the punch unit) can be loaded and unloaded without stopping the reader.
For a more detailed description of the 1402's card
reading capabilities, see Section 401:071.
The 1402 Card Read-Punch consists of a card reader
and card punch which are housed in the same cabinet
but are functionally independent of one another. The
punching capabilities are described in the next report
section, 415:072. The 1402 Model 3 differs from the
1402 Model 1 used in 1401 systems only in that the
Early Card Reader feature, which provides three
clutch points per cycle instead of one, is standard.
The card reader reads standard 8o-column cards, 1
row at a time, at a peak speed of 800 cards per minute. Conversion from the Hollerith card code to internal alphameric code is automatic. A hole-count
check is made on each column at a second reading
station, and the bit configuration of each character is
checked for validity as it is transferred into core
storage. As in the 1401, data read from a card is
always placed in a fixed input area, core storage positions 001 through 080. A hopper with a 3, DOD-card
© 1963
Optional Features
Interchangeable Read Feed: Permits reading either
80- or 51-column cards by interchanging hardware.
Binary Transfer: Makes it possible to process data
recorded in column binary form on cards and magnetic tape, providing format compatibility with the
IBM 700/7000 scientific systems. This feature
serves the same purpose as the Column Binary
feature for the IBM 1401.
Read Punch Release: Permits computation during the
start time of each card read or punch cycle; see
Section 415:111.
Processing Overlap: Permits computation during
reader and punch start times and between data transfer 'cycles; see Section 415:111.
by Auerbach Corporation and BNA Incorporated
4/63
415:072.100
_STANDARD
EDP
•
IBM 1460
Input-Output
1402 Card Punch
"PORTS
INPUT-OUTPUT: CARD READ-PUNCH (PUNCH)
§
072.
.12
.1
GENERAL
.11
Identity:
.12
Description (Contd.)
For a more detailed description of the 140.2's card
punching capabilities, see Section 401:072.
Card Read-Punch (punch).
1402 Model 3.
Description
The 1402 Card Read-Punch consists of a card reader
and card punch which are housed in the same cabinet
but are functionally independent of one another. The
reading capabilities are described in Section 415:071.
Standard SO-column cards can be punched, 1 row at a
time, at a peak speed of 250 cards per minute. Conversion from internal alphameric representation to
the Hollerith card code is automatic. As in the 1401,
data to be punched must be placed in the fixed corestorage punch area, positions 101 through 180. A
reading station makes a hole-count check on each
column and sets a testable indicator when an error is
detected. The single 1, 200-card feed hopper and
three 1, ODD-card stackers (one shared with the
reader) can be loaded and unloaded without stopping
the punch.
© 1963
Optional Features
Punch Feed Read: Adds a reading station ahead of
the punching station so that results can be punched
into the same card from which data was read_ Holecount and chara'cter validity checks are made on the
read operation.
Binary Transfer: Makes it possible to process data
recorded in column binary form on cards and magnetic tape, providing format compatibility with the
IBM 700/7000 scientific systems. This feature
serves the same purpose as the Column Binary
feature for the IBM 1401.
Read Punch Release: Permits computation during the
start time of each card read or punch cycle; see
Section 415:111.
Processing Overlap: Permits computation during
reader, and punch start times and between data transfer cycles; see Section 415:111.
by Auerbach Corporation and BNA Incorporated
4/63
415:081.100
·STAND"'
lO P
_ED
."
IBM 1460
Input-Output
1403 Pri nter
REPORTS
INPUT-OUTPUT: PRINTER
§
081.
.12
.1
GENERAL
.11
Identity-:
• 12
Description
Printer.
1403 Models 2 and 3.
Description (Contd.)
For a more detailed description of the 1403 Model 2,
see Section 401:081.
Optional Features
The 1403 Model 2 is the printer used in most IBM
1401 systems, and its operation is exactly the same
in 1460 systems. The 1403 Model 2 uses a horizontal chain printing mechanism, in which a solenoid
driven hammer presses the form and ribbon
against the moving chain while the desired character
slug is passing each printing position. Peak speeds
are 600 lines per minute at single spacing and 480
lines per minute at an average line spacing of 1 inch.
The 1403 Model 3 Prin,ter, announced in February,
1963, is program-compatible with Model 2 but has
peak speeds of 1,100 single-spaced lines per minute
and 750 lines per minute at an average spacing of 1
inch. The printing technique is similar to that of the
1403 Model 2, but instead of being attached to a steel
band to form a continuous chain, the type slugs in
Model 3 are separate three-character units that ride
in a precision-machined horizontal channel, forming
a "train". The linear speed of the moving train in
the Model 3 is 206 inches per second, versus 90
inches per second for the chain in Model 2.
Both models have 132 print positions and a 48character print set. Spacing is 10 characters per
horizontal inch and 6 or 8 lil!es per vertical
inch. Forms from 3.50 to 18.75 inches in width can
be handled. The dual-speed carriage used on both
models has a skipping speed of 33 inches per second
for skips of 8 lines or less and 75 inches per second
for longer skips. Skipping is controlled by a 12channel punched tape loop. Data to be printed must
be placed in the fixed core-storage print area (positions 201 through 332), and all editing and format
control must be performed by the stored program;
the unit has no plugboard.
© 1963
Interchangeable Chain Cartridge Adapter (Model 2
only): Permits the operator to insert a chain cartridge with a different type font or character set
quickly and without special tools.
Numerical Print (Model 2 only): Permits changing
from the standard 48-character chain to a chain
with only 16 different characters. Speeds up to
1, 285 lines per minute can then be obtained on allnumeric output. The Cartridge Adapter described
above is a prerequisite.
Auxiliary Ribbon Feed (Model 2 only): Recommended
for satisfactory utilization of polyester film ribbons,
which provide improved print quality for applications
such as optical character recognition; can also be
used for feeding conventional fabric ribbons.
Selective Tape Listing (Model 2 or 3): Permits printing up to 13 characters on each of 8 separate 1. 5-inch
wide tapes or up to 29 characters on each of four
3. I-inch wide tapes. Combinations of 1. 5- and 3. linch wide tapes can be used. Each tape can be individually spaced under program control, but no skipping is possible while the feature is in use.
Print Storage (required for Model 3; optional for
Model 2): Provides a core buffer that holds a full
line of data to be printed and provides overlapping of
all except 1 millisecond of each print cycle with any
other system function. Without Print Storage, the
Processing Unit is interlocked for 84 milliseconds
during each 1403 Model 2 print cycle. Testable indicators permit branching if the printer or forms
carriage is busy.
by Auerbach Corporation and BNA Incorporated
4/63
415:091.100
•
STANDARD
EDP
•
IBM 1460
Input-Output
729 Magnetic Tape Unit
REPORTS
INPUT-OUTPUT: 729 MAGNETIC TAPE UNIT
§
091.
. 12
.1
GENERAL
. 11
Identity: .
.12
Description
Magnetic Tape Unit.
729 II, IV, V, and VI.
The 729 Magnetic Tape Units are used in IBM's more
powe::rful 1410, 7040, 70.70, 7080, and 7090 series
Description (Contd. )
systems as well as in the 1401 and 1460. They are
compatible with the 7330 and 727 tape units in tape
width, recording density, and format. The only significant differences among the four 729 models are
in densities and tape speeds. These characteristics
are summarized in the following table. The effective
transfer rates shown are based upon 1, OOO-character
blocks.
Model
Tape speed,
inches/sec
Density,
char/inch
Peak
transfer rate,
char/sec
Effective
transfer rate,
char/sec
729 II
75.0
200
556
15,000
41,667
12,600
27,400
729 IV
112.5
200
556
22,500
62,500
18,800
40,300
729 V
75.0
200
556
800
15,000
41,667
60,000
12,600
27,400
34,200
729 VI
112.5
200
556
800 +
22,500
62,500
90,000 +
18,800
40,300
50,200 +
+ Usable in 1460 systems but not in 1401 systems.
Optional Features
From one to six 729 tape units, in any combination of
models, can be connected to a 1460 system via the
1461 Model 2 Input/Output Control. Only one tape
read or write operation can occur at a time, and no
computing can be done during a tape read or write
operation unless the Processing Overlap feature is
added.
A read-after-write parity check detects most recording errors at the time of occurrence. Both lateral
and longitudinal parity are checked during read operations. As in the 1401, data can be recorded and
read in either the "load" or "move" mode. In the
load mode, each word mark bit in core storage corresponds to a special word-separator character on
tape. In the move mode, only the six data bits in
each core position are transferred to tape; word
mark bits in core storage are neither written on tape
nor affected by a tape read operation. Block lengths
are fully variable. Backward reading is not possible.
For more details on the 729 Magnetic Tape Units,
see Section 401:091.
© 1963
Compressed Tape: Permits reading and correct expansion of tape records written with high-order zero
elimination by an IBM 7070 series system. The Indexing and Store Address Register feature is a prerequisite.
Binary Transfer: Makes it possible to process data
recorded in column binary form on magnetic tape ano.
cards, providing format compatibility with the IBM
700/7000 scientific systems. This feature serves
the same purpose as the Column Binary feature for
the IBM 1401.
Processing Overlap: Permits computation during
tape start-stop time. When 200 characters per inch
density recording is used in all models, and at 556
characters per inch in the 729 II only, computation
cycles can also be interspersed with character trans-'
fers to or from the tape unit. The Processing Unit
is delayed for one 6-microsecond cycle each time a
character is transferred. At tape transfer rates
above 41,667 characters per second, no interspersed
computation is possible. (See also Section 415:111.)
by Auerbach Corporation and BNA Incorporated
4/63
415:092.100
_STANDARD
EDP
•
REPORTS
IBM 1460
Input-Output
7330 Magnetic Tape Unit
INPUT-OUTPUT: 7330 MAGNETIC TAPE UNIT
§
092.
. 12
.1
GENERAL
.11
Identity: .
.12
Description
Magnetic Tape Unit.
7330 Modell.
The 7330 Magnetic Tape Unit is slower, simple1,
and less expensive than the 729 tape units described
in the previous section, but the two units are completely program-compatible. Tape speed is 36
inches per second. Peak data transfer rates are
7,200 and 20,016 characters per second at recording densities of 200 and 556 characters per inch, respectively. When reading or writing 1, OOO-character blocks, the effective data transfer rates are
6,260 and 14,100 characters per second. The principal disadvantage of the 7330 is the fact that rewinding a full reel without unloading tl;J.e tape from the
vacuum columns and read-write head requires 13.3
minutes. A full high-speed rewind takes only 2.2
minutes, but is always terminated by unloading,
which makes the rewound tape unavailable for further
processing (as required in sorting operations) without operator intervention.
From one to six 7330 tape units can be connected to
a 1460 system via the 1461 Model 3 Input/Output
Control. Alternatively, both 7330 and 729 tape units,
up to a total of six, can be connected to the 1461
Model 2 Input/Output Control through use of the Tape
Intermix feature. Only one tape read or write
operation can occur at a time, and no computing can
be done during a tape read or write operation unless
the Processing Overlap feature is added.
A read-after-write parity check detects most recording errors at the time of occurrence. Both lateral and longitudinal parity are checked during read
© 1963
Description (Contd. )
operations. As in the 1401, data can be recorded
and read in either the "load" or "move" mode. In
the load mode, each word mark bit in core storage
corresponds to a special word-separator character
on tape. In the move mode, only the six data bits in
each core position are transfer,red to tape; word
mark bits in core storage are neither written on tape
nor affected by a tape read operation. Block lengths
are fully variable. Backward reading is not possible.
For more details on the 7330 Magnetic Tape Unit,
see Section 401:092.
Optional Features
Compressed Tape: Permits reading and correct expansion of tape records written with high-order zero
elimination by an IBM 7070 series system. The Indexing and Store Address Registers feature is a prerequisite.
Binary Transfer: Makes it possible to process data
recorded in column binary form on magnetic tape and
cards, providing format compatibility with the
IBM 700/7000 scientific systems. This feature serves
the same purpose as the Column Binary feature for the
mM 1401.
Processing Overlap: Permits computation during tape
start-stop time and between character transfers to or
from the 7330. The Proc.essing Unit is delayed for
one 6-microsecond cycle for each character transferred. (See also Section 415: 111. )
Tape Intermix: Permits 7330 tape units, along with
729s, to be connected to a 1461 Model 2 Input/Output
Control. The total number of tape units cannot
exceed six.
by Auerbach Corporatian and BNA Incorporated
4/63
415:101.100
.STAI and < not available.
always ignored by translator.
literals may not exceed 120 characters.
UPPER-BOUND(S), LOWER-BOUND(S)
not available.
no alternate computer names.
File Description
8
BLOCK size
9
12
13
14
15
16
17
FILE CONTAINS
SEQUENCED ON
HASHED
Table length
Item length
BITS option
RANGE IS
RENAMES
18
19
20
SIGN IS
SIZ E clause option
Conditional range
11
no range in block size can be specified;
range in RECORD size is permitted.
approximate file size cannot be shown.
no list of keys can be given.
hash totals cannot be created.
length of tables and arrays may not vary.
variable item lengths cannot be specified.
items cannot be specified in binary.
value ranges of items cannot be shown.
alternative groupmgs of elementary
items cannot be specified.
no separate signs allowed.
no range in size may be given.
a conditional value cannot be a range.
Verbs
23
25
26
DEFINE
INCLUDE
USE
new verbs cannot be defined.
no library routines can be called.
no non-standard I/O error and label
handling routines.
Verb Options
29
30
34
OPEN REVERSED
ADVANCING
Relationships
36
39
Conditionals
Conditional statements
tapes cannot be read backward.
cannot specify paper advance.
IS UNEQUAL TO and EXCEEDS are not
provided.
no implied objects with implied subjects.
only AT END or ON SIZE ERROR may
follow imperative statements.
Environment Division
40
SOURCE-COMPUTER
43
44
FILE-CONTROL
PRIORITY IS
45
I-O-CONTROL
no differences from the "standard"
configuration may be specified.
cannot be taken from library.
no priorities can be specified for multirunning purposes.
cannot be taken from library.
Special Features
48
49
12/63
Library
Segmentation
library routines cannot be called.
no provision for segmentation of object
programs.
417:162.100
IBM 7080
Process Oriented Language
FORTRAN II
PROCESS ORIENTED LANGUAGE:
~
162.
.14
.1
GENERAL
• 11
Identity: . . . . . . . . . • 7080 Processor: FORTRAN .
• 12
Origin: •..
.. IBM Corp.
.13
Reference:
. . IBM Publication
J28-6247-1.
.14
Description
FORTRAN II
Description (Contd.)
WRITE TAPE instructions are not used for binary
data; they result in the transmission of one BCD
record whose size is determined by the associatcd
list and not limited to 132 characters. The PRINT
and PUNCH commands produce a magnetic tape
Tecord for off-line printing or punching .
There are four kinds of 7080 FORTRAN messages
which can occur during compilation:
The IBM 7080 FORTRAN language is a restricted
but useful version of FORTRAN II, the most widely
accepted process oriented language for scientific
applications. For a general description of the
FORTRAN II language, see Section 408:161. The
principal restrictions on IBM 7080 FORTRAN are
the inability to handle subroutines; the lack of
Boolean, complex, and double precision arithmetic;
and the absence of the optimizing statements
EQillVALENCE and COMMON. Other restrictions
on the 7080 version relative to IBM 709/7090
FORTRAN II are listed in Paragraph. 142 below.
An adjustable DIMENSION statement is provided
which permits the size of arrays to vary during the
execution of a program. Two other useful extensions are a TYPE instruction and an expanded print
line and tape record length of 132 characters.
Other extensions to the FORTRAN II language which
are implemented in 7080 FORTRAN are described
in Paragraph .143 below.
Compilation of FORTRAN programs is performed
by the 7080 Processor module of the 7080 Compiling System, which requires at least 80, 000 core
storage positions and 10 magnetic tape units. See
Section 417:181 for a description of the 7080 Processor. The FORTRAN sections of the 7080 Processor produce Autocoder "one-for-one" instructions
and macro instructions, which are expanded and
assembled by subsequent sections to provide object
programs which run on an IBM 7080 in the 7080
mode only. FORTRAN source statements may be
mixed with Autocoder statements or with other
higher-level languages such as Report/File or
DA TGEN (Paragraph 417: 151. 2). When languages
are mixed, all input-output operations initiated by
other languages must be completed before any
FORTRAN input-output operations are executed.
• Advisory Messages - direct the programmer's
attention to possible errors.
•
Minor Error Messages - occur when the compiling program makes an assumption, substitutes
for invalid or improper input, notices unnecessary duplication, finds a definite error wl;lich
can be corrected by moderate patching, or skips
over improper input.
•
Serious Error Messages - occur when the compiling program drops a statement because a
major assembly error, which would at best require extensive patching, has been made.
o
Cannot-Proceed Messages - occur when the
source program exceeds system capacity.
Messages may also be produced during the running
of a FORTRAN object program. Object program
error messages provide options which can be exercised via console interrupt keys. Errors associated with "trigger tests" (divide check, accumulator overflow, or quotient overflow) will cause
messages to be typed, a value substituted, and
computation continued Without an option.
Each fixed or floating point constant or variable
occupies 10 positions of core storage (2 positions
for the exponent and 8 for the mantissa in the
floating point mode). Library subroutines will
usually occupy a total of between 5,000 and 8,000
core positjons at object time. The compiler optimizes all subscripting under control of a DO statement, but there is no provision for optimizing
arithmetic statements by evaluating common subexpressions only once. For example, the statement
x
The 7080 FORTRAN version differs in several ways
from 709/7090 FORTRAN II. Arrays are stored
row-wise unless otherwise specified by an. inputoutput list. The operation of the control character
"A" in FORMAT statements, which permits alphameric information to be transmitted to and from
core storage, is based on words of 10 characters
rather than 6 characters. The READ TAPE and
© 1963
=
A*B - C + SQRT(A*B)
will cause the product A *B to be computed twice by
the object program. It should therefore be replaced by the two statements
T = A*B
x
=
Auerbach Corporotion and Info, Inc.
T - C + SQRT(T).
12/63
417:162.141
§
IBM 7080
162.
. 142 Restrictions (Contd. )
. 141 Availability: •••...• translator released in
January, 1963.
(6)
The CHAIN feature, which permits programs
too large to fit into core storage to be executed
as a series of independent "links, " has not been
implemented .
(7)
Since PRINT and PUNCH instructions cause a
magnetic tape record to be written, use of both
these instructions in a single program will result in mixing of PRINT and PUNCH output.
• 142 Restrictions
(1)
The following statements are not permitted:
CALL Name (aI, a2, ..•..• , an>
COMMON A, B,
END (Il, 12, 13, 14, 15)
.143 Extensions
EQUIVALENCE (a, b, c, .•...• ),
(d, e, f, ...... ), ......
FREQUENCY n (i, j, ...... ), m (k, 1,
...... ), ......
FUNCTION Name (aI' 3.2 ' ••.... ,
(1)
Printer lines and input-output tape records can
have a maximum length of 132 characters.
(2)
The PUNCH and PRINT statements result in
the writing of magnetic tape records for offline punching or printing. Thus, programs
written for computers with an on-line printer
or punch can be used without modification.
~)
READ DRUM i, j, List
RETURN
(3) A TYPE statement allows the writing of a list
on the console typewriter.
SUBROUTINE Name (aI, a2' ...... , an>
WRITE DRUM i, j, List.
(2)
Boolean, complex, and double precision operations are not permitted.
(3)
The following functions are not permitted:
ATAN2F (
): .•• arctangent (double length)
(4)
General READ and WRITE statem ents are provided which can be substituted for all executable input-output statements. These are useful
as shorter command words for tape inputoutput statements.
(5)
The DIMENSION statement is adjustable, which
permits the size of arrays to vary during the
execution of the object program.
TANHF ( ): ••... hyperbolic tangent.
(4)
(5)
12/63
Indexing is not allowed in the lists following
READ TAPE and WRITE TAPE statements.
These commands are used to transmit arrays
without FORMAT control.
The FORMAT specification "Ow" (used to
specify an octal integer of w digits) is not
allowed.
(6) A wider range of numeric magnitudes can be
represented in 7080 FORTRAN than in 709/7090
FORTRAN II, as shown in the following table:
Floating Point Integers
7080 FORTRAN:
10- 99 to 10 99 1 to 10 10 _1.
709/7090 FORTRAN: 10- 38 to 10 38 lto 131,071.
417:171.100
IBM 7080
M. O. Language
Autocoder
MACHINE ORIENTED LANGUAGE:
!l 171.
.1
GENERAL
.11
Identity:
.22
AUTOCODER
Legend
Pglin:
IBM 7080 Processor: Autocoder.
Tag: . . . . .
. 12
Origin: ...
IBM Corp .
Operation:
.13
Reference:
IBM Publications
C28-6263, C28-6130-1,
J28-6265, J28-6231.
Numeric:
.14
Description
Autocoder is basically a symbolic machine oriented
language, expanded through the addition of powerful macro generators and the Input/Output Control
System (IOCS80). The language permits utilization
of all hardware facilities of a 7080. Autocoder
programs can contain sections written in the 7080
FORTRAN, COBOL, Report/File, Decision, Arithmetic, and Table-Creating languages described
elsewhere in this report.
Each macro instruction in the source routine is
converted by the appropriate macro generator into
a series of symbolic instructions, which are then
converted to machine instructions. Parameters in
the operand of the macro instruction control the
coding that will be produced by the generator.
There is a SPEED macro which instructs the
translator to optimize either the execution
time or size of an object program. Other
standard macros perform such functions as loop
control, comparisons, data movement, table
search and maintenance, arithmetic operations, indirect transfers, and typing of a message. Additional macro generators can be coded by the user
and added to the systems tape.
The IBM 7080 Input/Output Control System (IOCS80)
is a supplement to Autocoder. It provides additional control and macro operations that handle reading,
writing, tape blocking and unblocking, file labeling,
checkpoints, and error checking. Information pertaining to the system configuration, file characteristics, record layouts, file labels, and checkpoints
must be indicated in the source routine in the operands of IOCS descriptive macros. From 35 to 51
operands must be used to describe each file. The
IOCS80 macros and their functions are listed in
Paragraph 417:171.81.
. 15
Publication Date:
.2
LANGUAGE FORMAT
. 21
Diagram: •.
Operand: . . . . . . .
Comments:
Flag:
...•.....
.23
Corrections: . . . .
.24
Special Conventions
.241 Compound addresses:
. 242 Multi-addresses:
.243 Literals: . . . . . . .
... December, 1962 .
. 244 Special coded
addresses:
see 7080 Processor Coding
Form, Page 417:171. 90U.
© 1963
Auerbach Corporation and Info, Inc.
sequences coding sheets
and lines on each coding
sheet.
names an area or instruction.
defines operation to be performed in mnemonic code.
indicates size of data field;
specifies an address
ending, size of character
code, bit code, console
switch code, number of
ASU, sign, or allocation
counter.
actual or symbolic address
of data to be operated
upon, with specification
of relative and/or iI).direct
addressing, address constants, literals, constants,
or format symbols.
additional information about
statement (for listing.
only).
communication to the Processor (see Paragraph
.82).
entered on back of coding
sheet; suggested gaps in
sequence numbers allow
for insertions.
BASE ± ADJUSTMENT,
where BASE is any basic
operand and ADJUSTMENT is a decimal integer.
,
BASE * or / ADJUSTMENT,
where BASE is any tag
and ADJUSTMENT is a
decimal integer.
PREFIX, BASE, where
PREFIX is a letter and
indicates operand modification or indirect address.
macros only .
preceded and followed by
#=; any number of characters except where restricted by instruction•
*
refers to low-order
position of instruction in
which it appears.
12/63
IBM 7080
417:171.245
§
171.
.245 Other
Actual:
Address constants:
.3
LABELS
.31
General
.311 Maximum number
of labels: .•.
.312 Common label
formation rule:
.313 Reserved labels:
. 314 Other restrictions:
.315 Designators:
.316 Synonyms permitted:
. 32
denotes an actual storage
location, a setting for accumulator or an ASU, or
size of a block; preceded
by @ if five or more numeric characters. No
notation needed if less than
five and used with certain
instructions.
ADCON statement creates
a NO-OP instruction with
address constants.
ACON4, ACON5, and
ACON6 statements give
address constants in desired size with specified
size for any operand.
literal-operand modifier
followed by @ precedes tag
or literal.
no practical limit; i. e. ,
limited by length of tape.
yes.
none.
if label is numeric, it must
be 5 characters or more.
none.
yes; TRANS pseudo (limited
to 50).
.321 Labels for procedures
Existence:
mandatory if referenced by
other instructions.
Formation rule
First character:
alphabetic, numeric, or
blank.
Others: ..
alphabetic and/or numeric
and blanks.
Number of
characters:
maximum of 10.
. 322 Labels for library
routines: . . . • . .
same as Procedures.
.323 Labels for constants: same as Procedures .
. 324 Labels for files: ••
same as Procedures.
.325 Labels for records:
same as Procedures .
. 326 Labels for variables: same as Procedures.
Local Labels:
.4
DATA
. 41
Constants
none.
.411 Maximum size constants
Integer
Decimal: . . . . .
no restriction, except may
not exceed 99 decimal
digits if referenced by a
general purpose macro
(limited by size of core).
Octal: . . . . . . .
none.
12/63
.
.
.42
Working Areas
.421 Data layout
Implied by use:
Specified in
program:
.422 Data type: ..
.423 Redefinition:
.43
Universal Labels
.33
.411 Maximum size constants (Contd.)
Fixed numeric
Decimal: ...... no restriction, exc.ept may
not exceed 99 decimal
digits if referenced by a
general purpose macro.
Octal: · ......
none.
Floating numeric
Decimal:
10 decimal digits (2 for exponent and 8 for mantissa).
Octal: ·
none.
limited only by size of core
Alphameric:
storage.
.412 Maximum size literals
Integer
52 characters. (35 if referDecimal:
enced by a macro).
Octal: · .
none.
Fixed numeric
Decimal:
52 characters (35 if referenced by a macro).
Octal: ·
none.
Floating numeric
Decimal:
12 characters.
none.
Octal: · ..
52 characters (35 if referAlphabetic:
enced by a macro).
52 characters (35 if referAlphameric:
enced by a macro) .
yes.
specified in description
statement.
yes; TRANS pseudo.
In2ut-Out2ut Areas
. 431 Data layout:
.432 Data type:
explicit layout .
specified in description
statement.
no.
.433 Copy layout:
.5
PROCEDURES
.51
Direct 02eration Codes
.511 Mnemonic
Existence:
Number:
Example:
. 512 Absolute: .
.52
yes .
mandatory.
113 .
MPY = multiply.
not usable .
Macro-Codes
.521 Number available
Input-output .
Arithmetic: . .
Math functions:
Error control:.
Restarts: . . . . .
Assembly control:
Data transmission:.
Data testing: ...
Program branch
control: . . . . .
Address modification:
Table: . . . . . .
Miscellaneous:
.
more than 60 (in IOCS80).
14 .
none.
none .
none.
5.
6.
3.
9.
7.
5.
3.
417: 171.522
MACHINE ORIENTED LANGUAGE: AUTOCODER
§
.65
171.
.651 File labels:
.522 Examples
Simple:
Elaborate:
. 523 New macros: . . . . .
. 53
Interludes: . . . . . .
.54
Translator Control
.541 Method of control
Allocation counter:
Label adjustment:
Annotation:
...
.542 Allocation counter
Set to absolute:
Set to label:
Step forward:
Step backward:
Reserve area: ..
..
.543 Label adjustment
Set labels equal:
Set absolute value:
Clear label table:.
.544 Annotation
Comment phrase:.
Title phrase: ..
.545 Other
Relative assignment: . . . . . . . .
EJECT pseudo: .
Flag characters:
ENTSO.
NDVHX roURCEI t:l
SOURCE2t:lRESULTl:C
OVERFLOW U U 250011l .
yes; separate run (same
run if needed only for that
program).
pseudo-operations.
pseudo-operations.
pseudo-operations and
special cards.
LASN, SASN, LITOR,
SUBOR, RASN, SUBRO.
LASN, SASN, LITOR,
SUBOR, RASN, SUBRO.
LASN, SASN, RASN.
LASN, SASN, RASN.
RCD, NAME, CON, RPT,
FPN.
TRANS pseudo.
TRANS pseudo.
no.
special cards, or comments
section of any line.
TITLE pseudo.
permits assembly at one
location to be treated as
though it were at another
location.
causes listing to skip to a
new page.
special instructions to Processor.
Boecial Arithmetic
...
Sorting:
.67
Diagnostics
· ..
· .....
.672 Tracers: ·
.673 Snapshots:
......
......
.7
LIBRARY FACILITIES
.71
Identity:
.72
Kinds of Libraries:
.73
Storage Form:
.74
Varieties of Contents:
· ......
·.
· ......
see Paragraph 417:151.13.
.751 Insertion of new
item: ·
.752 Language of new
item:
.753 Method of call:
.....
· ......
·.
.76
Automatic Decimal Point
macros:
ABSX, ADDS, etc. (see
Paragraphs 417:171. S1
and 417 :151. 22) .
macros, or use of DATGEN
Arithmetic Language.
§Ilecial Functions: .
none.
.63
Overlay Control:
no special facilities; incorporated by own coding.
.64
Data Editing
none.
none.
not required due to hardware editing capability.
see Paragraph 417:151.15:
Memory Print routine and
Snap Print routine.
none.
DME instruction and Memory Record function of
roCSSO .
70S0 Compiling System
Libraries (70S0 Processor Library, IOCS System Library, COBOL
Processor Library) .
expandable master.
card image records on
magnetic tape.
subroutines and macro
instruction routines.
Mechanism
special library run.
70S0 Autocoder.
INCL pseudo and LINK
macro.
Insertion in Program
.761 Open routines exist:
.762 Closed routines
exist:
.763 Open-closed is optional:
.764 Closed routines
appear once:
· .......
. 62
.641 Radix conversion:
. 642 Code translation: .
.643 Format control:
.66
.75
.61
.612 Method of call:
Reel labels:
Blocking: . . . . . . .
Error control: . . . .
Method of call:
. 671 Dumps: . . . . .
SPECIAL ROUTINES AVAILABLE
. . ....
.652
.653
.654
.655
handled by Input-Output
Control System, IOCSSO.
handled by roCSSO.
handled by roCSSO.
handled by roCSSO.
macros.
none.
.6
.611 Facilities:
In]2ut-Out]2ut Control
yes.
yes.
yes .
yes.
.S
MACRO AND PSEUDO TABLES
.SI
Macros
Code
SPEED:
•••
I
•••••
ENTSO:
· ......
ENTIP:
· ......
© 1964 Auerbach Corporation and Info, Inc.
Description
minimizes either size of
execution time of generated program.
causes subsequent macros
to generate instructions
in 7080 mode.
causes subsequent macros
to generate instructions
in 7080 interrupt mode.
12/63
IBM 7080
417:171.810
§
.81
171.
.81
Macros (Contd.)
Code
LEVIP:
LEV80: •...•..
ASU: . . . . . . . . .
CASU: . . . . . . . .
MOVE: . . . . . . .
BLANK: . • . . . .
ZERO: . . . . . . .
DEC: . . . . . . . .
COMP: . . . . . . . .
RANGE:
.•..•.
IFNUM:
SETON: . . . . . .
SETOF: . . . . . . .
IFON: . . . . . . . .
ALTTR:
...•.•
ALTNP:
..... .
FTTR:
...... .
FTNOP:
FTTRB:
..... .
FTNPB: . . . . . . .
ABSX: • . . . • . . .
NABSX:
ADDX:
SUBX: . • . . . . . .
12/63
Macros (Contd.)
Code
DIMX: . . . . . • • .
Description
causes subsequent macros
to generate instructions
in 7080 non-interrupt
mode.
causes subsequent macros
to generate instructions
in 7080 secondary mode.
sets an ASU and communicates its length and
availability.
sets a CASU and communicates its length and
availability.
moves data from one
field to another.
blanks fields.
places zeros in fields.
replaces defined decimal
point and/or fixed
dollar sign in previously
blanked RPT field.
compares data in two
fields and transfers on
low, equal, or high.
determines whether data
lies within a range and
transfers accordingly.
determines whether data
is numeric or not and
transfers accordingly.
sets SWitches ON.
sets switches OFF.
tests a switch and transfers according to an
ON or OFF condition.
alternately transfers and
continues in line, in
that order.
alternately continues in
line and transfers, in
that order.
transfers initially and
continues in line subsequently.
continues in line initially
and transfers subsequently.
transfers initially on a
bit and continues in
line subsequently.
continues in line initially
on a bit and transfers
subsequently.
obtains absolute value of
data in a numeric field
and places it in a numeric or RPT field.
same as ABSX (fornega,tive absolute value).
adds data in two numeric
fields and puts results
in a numeric or RPT
field.
same as ADDX, except
subtracts.
INCRX: ..••...
DECRX:
•..•..
MPYZ: . . . . • . .
NMPYX: .•.•...
DIVX: . . . . . . . .
NDIVX: •....••
DVHX:
.•.....
NDVHX: •.•....
TESTX: . . . . . . .
ADDA: • . . . . . . •
SUBA: . . . . . . . .
INCRA: . . . . . . .
DECRA:
INITA:
•.....
...... .
MOVEA: . . • . . . .
COMPA: ..•..•.
TBCTL: .••.•..
SERCH:
ADITM:
DLITM:
RPITM:
LINK: . . . . . . . .
STOP: . . . . . . . .
TyPE: . . . . . . . .
TRIN: . . . . . . . .
Description
same as SUBX if result
positive; if negative,
zeros are placed.
increments data in a
numeric field.
decrements data in a
numeric field.
multiplies data in 2 numeric fields and puts
result in a numeric or
RPT field.
same as MPYX, except
sign of product is reversed.
same as MPYX, except
divides.
same as DIVX, except
sign of quotient is reversed.
divides, provides automatic protection, halts
and transfers for a
zero divisor.
same as DVHX, except
sign of result is reversed.
tests data and transfers
according to negative,
zero, or positive condition.
replaces one field reference with another to
which an increment has
been added.
same as ADDA, but
decrement has been
subtracted.
increments a field reference.
decrements a field reference.
creates an address constant and replaces a
field reference with it.
replaces one field reference by another.
compares two field references and transfers on
low, equal, or high.
enables BSRCH and
SSRCH to search a
binary or serial table.
searches either a binary
or serial table.
places a new item in a
binary or serial table.
deletes an item from a
binary or serial table.
replaces the function of
an item in a binary or
serial table.
links to one or more
subroutines.
causes a "dead-end" halt.
types a message.
transfers indirectly.
MACHINE ORIENTED LANGUAGE: AUTOCODER
§
417:171.811
171.
. 81
.81
Macros (Contd.)
Macros (Contd.)
1414 Model 6 I/O Synchronizer
defines I/o Control System.
DDF: . • . . . . . .
defines data file.
OPEN: . . . . . • .
opens a file.
CLOSE: ••.....
closes a file.
puts a message out.
PUT: . • . . • . . .
LEVRT: . . . • . . .
leaves routine.
mocs:. . . . . . . .
IOCS80 Macros
Code
729 Tape Units
lOCS: . . . . . . . .
lOFTA, IOFTB,
IOFTC: . . . . . .
IOTA, lOTS:
...
lOGET, lOPUT: ..
lORD, 10WR: ...
lORDS, lOWRS: ..
IODMP: . . . . . . .
10BSD, lOPOS,
lOFSP, lOBSP,
10FSF, 10BSF: .
IOCLS: . . . . . . . .
lOMFO: . . . . . • .
10MFC: ..••..•
lOFER:
lORWD:
10RUN:
IOTYP:
IODEC:
IOHLD:
IOLNK:
IODCH, 10IOF,
10ION, IOMIP,
10MOP, IORET:
Description
precedes all other macros
in the assembly.
describe files of object
program; used to generate file table.
des~ibe tape drives used
by object program; used
to generate the Tape
Table.
move data between work
areas and I/O areas;
may also issue read or
write requests.
used to read or write a
tape record into or from
an area specified in the
file table; record length
checking may be performed.
used to read or write a
tape record into or from
an area specified in the
macro.
used to write a record on
tape from an address
specified in the macro
through to the end of the
octant.
7750 Programmed Transmission Control
DDF, OP
OPEN, PUT:.
same as 1414, above.
CLIP: . . . . . . . .
clears input area pool and
closes file for input
operations.
CLOSE: . . . . • . .
same as Qutput.
LOAD: . . . . . • . .
performs dynamic loading
of 7750 stored program.
ENDLD: . . . . . • .
releases 7750 from load
status.
DUMP:
unloads 7750 storage to
7080 memory.
IOCTL:
gives command to 7750
stored program.
leaves real-time routine.
LEVRT: • . . . . .
mocs,
1301 Disk Storage
DDF,
OPEN, CLOSE:.
GET, GETS, GETR:
mocs,
PUT, PUTS, PUTR:
MVRSA: ••.••••
tape movement macros.
terminates operations on
an input or output tape.
starts operations on a
new file of a multifile
output tape.
terminates operations on
an output tape without
rewinding.
forces an end-of-reel
condition.
rewinds a tape.
rewinds and unloads a
tape.
types a message.
sets up a message and
enters a waiting loop.
insures that all I/O operations on a particular
file have been completed
and checked.
with OPEN operand, reopens an input file.
links to specific routines in
IOCS other than CSDTS.
used to modify existing
specifications or conditions, in 10CS or the
object program.
© 1963
ENTDR: . . . . • .
FSEQP: . . . • . . .
LEVDR: •...•..
.82
same as 1414, above.
get data record; get
single reference; get
random record.
put data record; put
single-reference; put
random record.
moves record to stacking
area.
enters disk routine.
forces sequential processing.
leaves disk routine.
Pseudos
Code
LASN:
SASN:
RASN: ••••....
SUBRO: ••••..•
SUBOR: •••••••
LITOR: •••..•.
TCD: . . • . • . . .
INCL: . . . . . . . .
Auerbach Corporation and Info, Inc.
Description
. sets a location counter to
a specific location.
sets the Blank counter to
an actual address or to
a previously assigned
location.
allows assembly at a location to be treated as if
assembled in another
location.
assigns subroutines within
macros.
assigns library routines.
assigns literals.
causes a temporary halt
in loading of the object
program, and execution
of the portion just loaded.
designates a library subroutine that is to be inserted in the object
program.
12/63
417:171.820
§
IBM 7080
171.
.82
.82
Pseudos (Contd.)
Code
TRANS: • . . . . .
MODE: ...•.•.
EJECT:
TITLE:
RCD:
CON:
FPN:
.•...•.•
RPT:
NAME: . . • . . . .
CHRCD:
BITCD: . . • . . . .
SWT: . . . . . . . . .
SWN:
ALTSW:
12/63
.....
Description
-equates the operand of an
instruction into an actual
location derived from the
operand of the TRANS.
indicates a change in the
language of the source
program.
advances the listing to a
new page.
places lines or paragraphs
of descriptive information in listing.
defines a field in which
a record block, individual record, or a portion
of a record will be
placed.
defines a field which will
contain constant data
and provides a constant
itself.
defines a field for a constant floating point
number.
defines a field for numeric
data for a report and
specifies the print
format.
identifies a series of adjacent fields as a unit
and specifies the final
digit or digits of the
address of a field.
indicates a charactor code
switch.
indicates a bit code switch.
defines a program switch
which will be on initially.
defines a program switch
which will be off initially.
designates a console alteration switch.
Pseudos (Contd.)
Flags
Code
@:
A:
B:
C:
D:
F:
G:
H:
M:
R:
S:
T:
Z:
1:
.
Description
causes output produced
from corresponding
entry to start a new card.
reduces location assignment phase assembly
time.
causes Processor to scan
the entry from Right to
Left.
designates that the entire
card is a comment.
all diagnostic messages
for the entry are to be
deleted.
denotes beginning and end
of a chain of macros.
causes a change entry to
be treated as a generated
entry.
causes entry to be listed
on the Halts page of the
Operator's Notebook.
causes entries to be printed on the Operator's
Notebook with blank operands.
resets location counter.
program cards not to be
produced.
removes entry when not
test-assembling.
causes replacement of the
standard "00" card by
a TCD "00" card.
weights inner macro as
one so that accuracy of
the Frequency Table is
improved.
417:171.900
MACHINE ORIENTED LANGUAGE: AUTOCODER
§ 171.
.9
CODING FORM
FOR M
X28-1636
7080 PROCESSOR
CODING FORM
o
Page _ _ of _ _ _ __
0
Program or Macro _ _ _ _ _ _ _ _ _ __
Date _ _ _ _ _ _ __
Programmed by _ _ _ _ _ _ _ _ _ _ __
,
IDENT
I
PGlIN
1
INCLUSION TEST
TAG
56
,
OPERATION NUM
5 16
CODING ON BACK
80
75
OPERAND
v ,
M1
M2
M3
51
\2
,,
,G
-
COMMENTS
3940
20 21 22 23
-
6263
65
67
69
71
737
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
J
L
L
,I
~
~
I
I
I
I
I
I
I
I
I
I
I
I
-'
I
L
I
1
I
I
I
I
I
I
I
J
I
I
I
I
I
I
I
I
I
I
I
L
I
I
I
I
I
I
I
I
I
I
I
I
I
J
I
I
I
I
I
L
I
J_
© 1963
Auerbach Carparation and Info, Inc.
12/63
417:181.100
IBM 7080
Program Translator
7080 Processor
PROGRAM TRANSLATOR:
§
.22
lSI .
.1
GENERAL
• 11
Identity: ..
.12
Description
. . . . . . 70S0 Processor.
The 70S0 Processor module of the 70S0 Compiling
System is used to translate IBM 70S0 Autocoder,
FORTRAN, Report/File, Decision, Arithmetic,
and Table Creating languages (or any reasonable
combination of these) into 70S0 machine language
object programs. Operation of the 70S0 Processor
requires an IBM 7080 computer with 80,000
positions of core memory, 10 tape units, and a
printer. Use of a card reader is optional and
permits changes in assembly sequence and functions. If additional memory is available, it will
be used to optimize macro generation time and
increase the storage area for input-output operations and tables.
Documentation facilities are numerous and flexible.
A variety of assembly runs is possible. If a
160, OOO-position core memory is available, a
frequency count table is used to optimize macro
generation time during assembly. The Input/
Output Control System is a supplement to the
70S0 Processor that effectively eliminates the
need for detailed programming of standardized
input-output routines. Thirty-five macro instructions are provided with IOCS80 for use
with 729 Tape Units. Routines to perform
functions such as execution of input-output
operations, error correction, tape control,
recording of memory at various checkpoints,
GET/PUT operations (optional), and housekeeping
are made available by the use of IOCS macros.
A description of these macros can be found in
Section 417:171. Sl.
. 13
Originator: . . . . . . . . IBM Corp .
. 14
Maintainer: . . . . . . . . as above .
. 15
Availability:
.2
INPUT
.21
Language
.... released December, 1962.
7080 PROCESSOR
Form
.221 Input media: . . • . . . . card images on magnetic
tape plus optional control
cards; or cards .
. 222 Obligatory ordering: .. must be in correct sequence
according to coding sheet
page and line numbers.
.223 Obligatory grouping: .. none.
.23
Size Limitations
.231 Maximum number of
source statements: .. no practical limit.
.232 Maximum size source
statements: . . . . . . . no practical limit (can be
continued from card to
card).
. 233 Maximum number of
data items:
... no practical limit.
.234 Others
Maximum input
blocking factor: ... 31 SO-character records.
.3
OUTPUT
· 31
Object Program
.311 Language name: . . . . . IBM 70S0 machine language.
.312 Language style: . . . . . machine.
.313 Output media: . . . . . . card images on magnetic
tape; maximum of 65
columns of instruction
and/or constant entries
per card.
· 32
Conventions
.321 Standard inclusions:
. IOCS linkages (IOCS
routines may be preassembled or assembled
along with the program) .
. 322 Compatible with: . . . . IBM 7.080 Supervisory
Control System (Section
417:191) .
70S0 Autocoder library .
· 33
Documentation
Subject
. . . . . . . . . . IBM 7080 Autocoder,
FORTRAN, Report/File,
Decision, Arithmetic,
and Table Creating
languages, or any combinations of these.
.212 Exemptions: . . . . . . . none.
Provision
.211 Name:
© 1963
Source program:
.
Object program: ....
Storage map: . . . . . .
Restart point list:
.
Language errors: ..
Auerbach Corporation and Info. Inc.
listing.
listing.
listing.
none.
listing and messages
during assembly.
12/63
417:181.330
§ 181.
.33
IBM 7080
.41
Section 5:
Documentation (Contd.)
Operator's Notebook list of halts, titles,
location assignments, switches,
all TAG NOT
REQUIRED
messages, statements having
blank operands,
and those statements indicated
for separate
listing of flags: .. listing (optional).
Symbolic Analyzer entries defining or
referencing a tag
and operand
modifiers and
character adjustment are included: listing (optional).
Conditional listing
(if assembly could
not proceed): . . . . listing.
Note: Listings are written on magnetic tape
for off-line printing.
.4
TRANSLATING PROCEDURE
.41
Phases and Passes
Section 1 *: . . . . . . . . Processor housekeeping reads control cards,
processes them, and
loads and transfers to
next section. (If a
COBOL run is to be
made, the COBOL
Processor sections
will be placed after
this section).
Section 2: . . . . . . . . Processor input - reads
source program and
determines which section
is to be executed next.
Loads and transfers to
that section.
Section 3: . . . . . . . . . Higher language - executed
only if FORTRAN,
Report/File, DeCision,
Arithmetic, or Table
Creating languages are
present in the program.
Statements in these
languages are translated,
the next section is determined, and a transfer is made.
Section 4: . . . . . . . . . Processor librarian executed only if a
librarian assembly
has been requested.
Prepares macros for
use during the current
assembly.
*
"Section" refers to one or more related phases.
12/63
Phases and Passes (Contd.)
Section 6:
Section 7:
· 42
Optional Mode
.421
· 422
• 423
· 424
. 425
Translate: . . .
Translate and run:
Check only:
Patching:
Updating: ..
· 43
Special Features
· Processor edit - determines whether macroheaders are present. If
so, next section is loaded
and transfer is made. If
not, next section is 7.
· Macro-generation - produces one-for-one
statements for each
macro-header. Loads
next sections and transfers to it.
· Location assignment
through output (assembly
section) - assigns locations, inserts constants
and addresses, and
produces final listing
and program cards. If
multiple assembly is
requested, housekeeping
section is loaded and
assembly is repeated.
If not, as sembly ends
here.
·
·
·
·
·
yes.
no.
yes.
yes .
no.
.431 Alter to check only: . no.
· 432 Fast unoptimized
translate: . . . . . . . . no.
.433 Short translate on
restricted program: . yes; "high-speed assembly"
run bypasses all generator phases, so only onefor-one entries can be
inserted, deleted, or
replaced.
.44
Bulk Translating: .
· 45
Program Diagnostics
.451 Tracers: ..
.452 Snapshots:
.453 Dumps:
· 46
. yes; multifile assembly
control.
· none.
· none.
· manual dump or standard
memory print routines.
T flag allows certain
instructions to be included or deleted during
an assembly run.
Translator Library
.461 Identity: ••••...••• 7080 Processor library
(Autocoder library),
.462 User restriction: •••• none.
PROGRAM TRANSLATOR: 7080 PROCESSOR
417:181.463
.53
s 181.
OQtimizing Data (Contd.)
Implicit:
.463 Form
Storage medium: ... magnetic tape.
Organization: . . . . . . strings .
. 464 Contents
Routines: . . . . . . . . macro instruction routines
and subroutines.
Functions: • . . . . . . for FORTRAN Processor.
Data descriptions: .. none.
•
Labels should be left-justified to minimize
shifting operations during assembly.
" Unnecessary labels should not be used because
assembly time is lengthened.
.54
Object Program
Performance: . . . . . essentially unaffected (i. e. ,
same as hand coding);
efficiency will generally
decrease if macros are
used extensively.
.465 Librarianship
Insertion: ..
· ... during special library run.
Amendment:
· ... during special library run.
Call procedure: · ... INCL pseudo instruction
and LINK macro.
.5
TRANSLATOR PERFORMANCE
.6
COMPUTER CONFIGURATIONS
.51
Object Program SQace
.61
Translating Computer
.511 Fixed overhead
Name
IOCS:
Space
averages
over
20,000
positions
Checkpoint routine:
2,000
positions
SCS80:
approx.
3,200
positions
Comment
depends upon
facilities
used (number of
channels,
etc. )
used if no
work tape
is specified.
if operating
system is
used.
.512 Space required for
each input-output
file: . . . . . . . . . . . . as specified by programmer.
.513 Approximate expansion
of procedures: . . . . . one-to-one (except macros,
which are widely variable).
. 52
Translation Time:
. 53
Optimizing Data
•
•
. IBM 7080 computer with
80, 000 core storage
positions.
10 magnetic tape units
(729 or 7340).
.612 Larger configuration
advantages.: . . . . . . . extra tape unit or card
reader permits multiple
runs (bulk translation).
extra memory will be used
to store macro components, thus shortening
macro-generation time;
size of I/O areas and
tables are increased.
up to 3 more tapes can
receive output.
. 62
Target Computer
.621 Minimum
configuration: . . . . . 40K IBM 705 mor any
IBM 7080 computer .
.622 Usable extra facilities: all.
?
Explicit:
•
.611 Minimum
configuration:
.7
The SPEED macro instruction may be used
to optimize either space or time in the coding
generated from macro-headers.
ERRORS, CHECKS AND ACTION
Error
Missing entries:
Unsequenced entries:
Duplicate names:
Improper format:
Incomplete entries:
Target computer
overflow:
Inconsistent program:
Improper operation:
Flag 1 causes inner macro instructions to
be called once and enables the Frequency
Count Table (table of macros used) to reflect
more accurately the frequency of macros.
Efficiency of macro generation is increased
and processing time is reduced as a result.
Flag A used within Class B subroutines
reduces location assignment phase assembly
time.
.8
Check or
Interlock Action
check
check
check
check
check
noted
noted
noted
noted
noted
in
in
in
in
in
listing.
listing.
listing.
listing.
listing.
check
check
check
noted in listing.
noted in listing.
inserts NOP;
noted in listing.
ALTERNATIVE TRANSLATORS
Automatic:
•
A Frequency Count Table is generated which
contains the name of every macro instruction
used and (if 160K core is available) the number
of times it is used. This table is used during
macro generation to increase efficiency and
reduce processing time.
© 1963
Computer: . . . . . . . •
Identity: . . . . . . . . . .
Date:
....
Comment: . . . . . . . . .
Auerbach Carporotion and Info, Inc.
IBM 7080.
7058 Processor.
released October, 1960.
routines are provided to
produce complete
compatibility with the
7080 Processor.
12/63
417:1B2.100
IBM 70BO
Program Translator
70BO COBOL Processor
PROGRAM TRANSLATOR:
13 182.
.1
GENERAL
. 11
Identity: ..
. 12
Description
. . . . . . 7080 COBOL Processor .
The 7080 COBOL Processor, a module of the 7080
Compiling System, converts source programs
written in IBM 7080 COBOL (Section 417:161) into
7080 Autocoder language. The Autocoder statements
are then automatically assembled into a machine
language program by the 7080 Processor, the other
module of the 7080 Compiling System. The phases
of the 7080 Processor are described in Section
417:181. The COBOL Processor phases will be
performed after the 7080 Processor Housekeeping
Section, and will be followed by the 7080 Processor
Input Section.
The addition of the COBOL Processor module to
the 7080 Compiling System does not alter the minimum machine requirement: 80,000 positions of
core storage and 10 tape units. The use of the console card reader is optional. Additional core storage allows flexibility in the modification of Processor phases. Multiple assemblies are permitted
when an additional tape unit is available.
Two types of 7080 COBOL processor runs are
allowed:
•
Analysis Run - the COBOL source program is
analyzed, and diagnostic messages are provided
for any COBOL statement errors. No conversion to Autocoder is performed.
•
Compilation Run - a joint run of the COBOL
and 7080 Processors. The former converts the
COBOL source statements into a series of Autocoder statements, and the latter converts these,
plus any Autocoder source statements that the
program may contain, into a 7080 machine
language program. A partial compilation run
(i. e. , one that terminates at the end of the
COBOL phases) is also permitted. A preassembled Input/Output Control System can be used
with the COBOL source program. This results
in a saving in compilation time because the coding for the required IOCS functions does not need
to be generated.
70BO COBOL PROCESSOR
.13
Availability: . . . . . . . released February, 1%3.
.2
INPUT
· 21
Language
. 211 Name: ..
. 212 Exemptions: .•
· 22
© 1963
Form
· 2in Input media: . . . . . . . punched cards or single
reel of magnetic tape.
· 222 Obligatory ordering: .. Identification Division.
Environment Division.
Data Division.
Procedure Division.
.223 Obligatory grouping: •. by division, section, and
paragraph.
· 23
Size Limitations
.231 Maximum number of
source statements: ••
· 232 Maximum size of
source statements: •.
.233 Maximum number of
data items: . . . . • . .
.234 Others
Maximum computational item size: . . .
Maximum alphameric literal size: .
.3
OUTPUT
• 31
Object" Program
no practical limit.
no practical limit.
no practical limit.
18 digits.
120 characters.
(If a numeric literal contains a decimal point,
there cannot be more than
99 digits on either side of
the point.)
.311 Language name: •. . . . IBM 7080 Autocoder.
.312 Language style: ..
• symbolic (converted to final
machine language by the
7080 Processor; see
Section 417:181).
• 32 Conventions
.321 Standard inclusions: .• IOCS80.
.322 Compatible with: ..•• 7080 Autocoder library •
. 33
The 7080 COBOL Processor checks for a variety of
source program errors. Typewriter diagnostic
messages are produced upon the detection of certain errors. Compilation may be halted, an operator decision may be required, or compilation may
be continued, depending upon the seriousness of the
error. Other diagnostic messages are produced on
the output listing tape.
... IBM 7080 COBOL .
. see "Deficiencies with respect to Required COBOL
61" in Paragraph
417:161.142.
Documentatidn
Provision
Subject
listing tape.
Source program:
Object program:
listing tape.
Storage map: .•...• listing produced by 7080
Processor.
Restart point list: ..
none.
Language errors: ••
listing tape.
Auerbach Corporation and Info, Inc.
12/63
IBM 7080
417:182.400
§
182.
.41
.4
TRANSLA TING PROCEDURE
.41
Phases and Passes
.. controls the operation of
subsequent COBOL phases.
Phases 2, 3, 4: •
. . scan source program and
converts division entries
to Autocoder entries, internal records or statements, or definitions.
Phase 5: • . . . • . . • . produces General I/O and
File Name generator requests.
Phases 6,7: . . . . . . . sort the modifier records
into the same sequence as
the DDF definitions.
Phase 8: . • . . . • . • . . combines the modifiers with
the DDF definitions and
writes DDF and GRF onto
separate tapes.
Phase 9: • . . . . . . . . . reduces programmer names
to 5-character identifieraddresses; splits off
clauses with "CORRESPONDING. "
Phase 10: •••....•. reduces statements with
"CORRESPONDING" to
lowest level clauses.
Phase 11: . • . . . . . . . transfers characteristics
from DDF to GRF; creates
tags for reference names.
Phase 12: . . . . . . . . . converts DDF definitions to
Autocoder NAME, RCD,
etc; creates glossary used
only for compilation.
Phase 13: . . . . . . . . . generates Autocoder for GO,
STOP, PERFORM; generates necessary TR and
TRANS entries for IF, AT
END, and ON SIZE
ERROR.
Phase 14: .•...• ; .• splits off statements to be
processed by phases 15
thru 21.
Phases 15, 16, 17, 18,
19, 20: •.. " . . . . . . combine Environment and
I/O commands and generate Autocoder I/O macros.
Phase 21: ..•••..•. generates Autocoder for
ADD, DISPLAY, REDEF,
and SUBTRACT.
Phase 22:
.. expands and completes
conditional expressions.
Phase 23:
.. scans and pre-edits arithmetic and logic statements.
Phase 24: ..•••••.• optimizes and generates
Autocoder for arithmetic
and logic statements.
Phase 25: • • . . . . . . . converts source and
internally-created MOVE
statements to Al\tocoder
macros.
Phase 26: . . . . . . . • . sorts subscripts (from
prior generators) into
source program order.
Phase 27: ••...•.•. optimizes and generates
Autocoder macros to
calculate subscripts.
Phase 28: . . . . . . . . . generates PERFORM subroutines, and creates and
optimizes object constants
and work areas requested
by other generators.
Phase 29: . . . . . . . . . edits, splits, and blocks the
entire file of generated
Autocoder entries.
Phase 30: . . . . . . . . . replaces the actual value of
literals split off by
phase 4.
Phases 31, 32, 33,
34, 35: . . . . . . . . . sorts edits, combines, and
prepares card image or
list output.
Phase 36: . . . . . . . . . sets up file tables and constants for next phase of
7080 Processor.
Phase 1: . . . . . .
12/63
Phases and Passes (Contd.)
Note:
GRF
DDF
Generator Request File
Data Definition File
.42
OQtional Mode
.421
.422
. 423
.424
.425
... yes.
Translate:
Tra:J.slate and run: ... no.
Check only: .
· yes.
Patching:
· no.
Up-dating:
· no.
· 43
SQecial Features
...
.431 Alter to check only: .. no .
. 432 Short translate on
restricted program: . partial translation is
possible.
.44
Bulk Translating: . . . . yes.
.45
Program Diagnostics
.451 Tracers: ..
.452 Snapshots: .
· 453 Dumps: .•.
· no.
· no.
· automatic when serious
error is detected; can be
taken manually at any time
(no restart facility).
.46
Translator Library:
. 7080 Processor library;
see Paragraph 417: 181. 46.
·5
TRANSLA TOR PERFORMANCE
.51
Object Program Space
.511 Fixed overhead: . . . . . usually over 20,000 positions; depends upon 10CS
facilities used.
417:182.520
PROGRAM TRANSLATORS: 7080 COBOL PROCESSOR
§
182.
.52
Translation Time: ... no data available.
. 53
Optimizing Data
.612 Larger configuration
advantages:
. 1 more magnetic lape unit
permits bulk translalinp;.
larger corc storage permit;;
modification of COBOL
phases.
Explicit:
•
Storage size and optional features of the target
computer are specified in the OBJECTCOMPUTER paragraph of the Environment
Division.
•
Up to seven alternate input-output areas ean be
specified for magnetic tape files.
•
Magnetic tape records may be of variable length.
•
Multiple redefinitions of record item areas are
permitted.
. 62
Target Computer
.621 Minimum configuration: ..
. 622 Usable extra
facilities: .•..
. IBM 7080 Central Processor
with sufficient core storage
to hold object program,
data, and subroutines.
.. magnetic tape units, card
reader, card punch,
printer, typewriter, and
larger core storage.
Automatic:
Because of the 7080's alphamerie mode of operation, conversions between the DISPLAY and
COMPUTATIONAL modes are not required.
•
Autoeoder macros, arithmetic and logic statements, object constants, and work areas are
optimized by the compiler.
Implicit:
•
Do not compute anything within each pass through
a PERFORM loop that ean just as well be computed only once, before entering the loop.
•
Do not compute the same sub-expression in two
or more different COMPUTE statements if the
practiee can be avoided by evaluating the subexpression in a separate statement.
•
Check input data for reasonability.
. 54
Object Program
Performance: . . . . . no data available.
.6
COMPUTER CONFIGURATIONS
.61
Translating Computer
.611 Minimum configuration: . . .
... IBM 7080 with 80,000 core
storage positions.
10 magnetic tape units (729
or 7340).
console card reader
(optional) .
© 1963
.7
ERRORS, CHECKS AND ACTION
Error
Check or
Interlock
Action
Missing entries:
check
Unsequenced
entries:
Duplicate names:
Improper format:
Incomplete
entries:
Target eomputer
overflow:
Inconsistent
program:
Excessive item
size:
Data descriptor
conflicts:
Invalid editing:
Invalid
subscripting:
Class contradiction:
Tape assignment
for assembly:
check
type message (division missing).
tape listing message.
check
check
check
tape listing message.
tape listing message.
tape listing message.
check
tape listing message.
check
tape listing message.
check
tape listing message.
check
check
check
tape listing message.
tape listing message.
tape listing message.
check
tape listing message.
check
tape listing message.
Note: When a "critical" error which makes further
processing impractical is detected, compilation IS halted.
Auerbach Corporation and Info, Inc.
12/63
417:191.100
IBM 7080
Operating Environment
Supervisory Control System
OPERATING ENVIRONMENT: SUPERVISORY CONTROL SYSTEM
§ 191.
.213 Data: . . . . . . . . . ..
•1
GENERAL
.11
Identity:.........
. 12
Description
. 214 Master routines: ...
IBM 7080 Supervisory
Control System (SCS80).
The IBM 7080 Supervisory Control System (SCS80)
consists of two major parts: the Object Time
Routine and the Librarian. The functions of the
Librarian are to:
.22
Library Subroutines:
on "module tape" in Program Tape Library .
. 23
Loading Sequence:
manually controlled
(determined by order of
control cards in the console card reader); or can
be program controlled
(a program can request
the loading of another
program).
.3
HARDWARE ALLOCATION
.31
Storage
o Create and maintain a library of user's object
programs (Program Tape Library); and
o Create a current (program) tape, to be used at
object program time, by selection of programs
on the Program Tape Library which are
scheduled to be executed.
The Object Time Routine (which is used when the
object programs are executed) can:
o Search the program tape to locate the program
to be run (indicated by control cards);
III
Verify the settings of the console switches;
•
Load the program; and
o Transfer control to it.
.13
Availability: . . . • ..
December, 1961.
• 14
Originator: • . . . . ..
IE M Corp.
• 15
Maintainer:.......
IBM Corp.
.16
Reference:.......
IBM Publication
J28-6176.
.2
PROGRAM LOADING
.21
Source of Programs
.311 Sequencing of program
for movement
between levels: ... implemented by programmer (transfer to @0009
or use of an Execute control. card causes a new
segment to be loaded).
.312 Occupation of working
storage: . . . . . . . . assigned by programmer;
routines must start in
an address ending in 0 or
5 above @0499. @00040159 and 2,700 other core
positions are reserved
for the Object Time
Routine .
.32
Input-Output Units
. 321 Initial assignment:
.322 Alternation: . . . • . .
. 211 Programs from online libraries: . . . .
Librarian phase creates a
Program Tape Library
which contains all of
user's programs.
The current (Program)
tape is created from the
library to contain those
programs which are to
be run.
.212 Independent programs: must be added to Program
Tape Library by File
Maintenance Routine
(Librarian Phase).
© 1963
on tape with program or on
control cards .
on the Distribution Tape .
. 323 Reassignment: . . . . .
specified in program .
of program tape - by control card.
of tape used by object program - specified in
IOCS entries •
speCified by means of control cards.
.4
RUNNING SUPERVISION
.41
Simultaneous Working: as incorporated in user's
program.
. 42
Multi-running: . . . ..
no provision .
. 43
Multi-sequencing: ..
no provision .
Auerbach Corporation and Inlo, Inc.
12/63
417:191.440
IBM 7080
§ 191.
.44
Loading input
error:
Allocation
impossible:
In-out error
single:
In-out error
persistent:
Storage overflow:
check
check
check
check
check
Invalid instructions: none.
Program conflicts: none.
Arithmetic overflow:
check
Underflow:
check
Invalid operation:
check
Improper format:
check
Invalid address:
Reference to
forbidden area:
check
halt; manually
rewind PGM
tape and depress Load key
to restart.
.6
OPERATOR. CONTROL
. 61
Signals to Operator
Restarts: . . • . . .
halt and message; press
Start to try
again.
halt and message; resubmit as PATCH.
enter nonstop
mode.
halt and message; reassign
as PATCH.
enter nonstop
mode.
halt and message; recreate program
(current) tape.
enter nonstop
mode.
.51
Dynamic
message on console typewriter.
message on console typewriter.
.63
Operator's Signals
.631 Inquiry: . . . . . . . . .
.632 Change of normal
progress: . . . . . . .
•7
LOGGING: . . . . . . .
.8
PERFORMANCE
. 81
System Requirements
.811 Minimum
configuration:
• 812 Usable extra
facilities:
.813 Reserved equlpment:
none .
DMP instruction in object
program; program images
are written on tape by
IOCS.
not required; necessary
information is specified
on control cards.
Operator's Decisions: console keyboard.
operands of IOCS macros.
Memory Record phase of
IOCS80 and Memory
Restore System will
accomplish restart.
Operator may start a program at the beginning of a
load segment.
PROGRAM DIAGNOSTICS
manual, linkage to st:U1dal'd
routine, or linkage to
Snap Print routine.
.62
see Memory Restore System (CSMRS), Paragraph
417:151.17.
.5
. 511 Tracing: . . • . • . . .
. 512 Snapshots: . . • . . . .
. 611 Decision required by
operator: . . . . . . .
halt and message; .612 Action required by
reaSSign with
operator: .••....
PATCH.
.613 Reporting progress
of run: . . . . . . . . .
repeat operation.
none.
.451 Establishing restart
points: . . . . . . • . •
.452 Restarting process:
12/63
Post Mortem: . . . ..
Errors, Checks, and Action
Check or
Interlock
.45
.52
none.
methods are available to
abandon a run and to
alter the sequence of
routines by changing the
order of control cards.
console typewriter .
Object Time Routine 1 program tape,
1 console card reader,
1 console typewriter,
3,200 pOSitions of core.
File Maintenance Routine 4 magnetic tape units.
1 console card reader,
80,000 positions of core,
IOCS80.
Production-of-Current-Tape
Routine 5 magnetic tape units
(2 channels),
1 console card reader,
80,000 positions of core .
all.
core area reserved for
Object Time Routine
(3,200 positions).
console card reader.
console typewriter.
tape unit for Program Tape.
417:191.820
OPERATING ENVIRONMENT: SUPERVISORY CONTROL SYSTEM
§ 191.
· 82
.84
System Overhead
. 821 Loading time: . . . . .
· 822 Reloading frequency:
· 83
Program Space
Available: . . . . . • .
less than 1 second.
Object Time Routine is
loaded with each new
program.
all of core storage except
the 3,200 positions occupied
by the Object Time
Routine.
© 1963
.85
Program Loading
Time: . . . . . • . . .
limited by speed of input
tape unit .
Program Performance: use of SeS80 involves no
significant overhead during
the execution of a user's
program since SeS80
only "directs program
traffic. "
Auerbach Corporation and Info, Inc.
12/63
/
417:201.001
IBM 7080
System Performance
SYSTEM PERFORMANCE
§
201.
Generalized File Processing (417:201.1)
These problems involve updating a master file from information in a detail file
and producing a record of the results of each transaction. This application is one of the
most typical of commercial data processing jobs and is fully described in Section 4:200. 1
of the Users' Guide.
In the graphs for Standard File Problems A, B, C, and D, the total time required
for each standard configuration to process 10,000 master file records is shown. Times for
cases using both unblocked and blocked records in the detail and report files are shown by
means of solid and dashed lines, respectively.
The 7080 is shown in two standard System Configurations, both fully tape-oriented
and including an off-line IBM 1401 for input-output processing. These configurations are
shown in Section 417:031 and defined in the Users' Guide, Section 4:030. Only the times for
the on-line 7080 processing, with all input and output files on magnetic tape, are considered
here.
The graphs for Configuration VII B, for both blocked and unblocked Files 3 and 4,
show that tape time is usually the controlling factor on total time required over the activity
range. The only exception is for blocked files in Problem D, where central processor time
is controlling at high activity.
For the more powerful Configuration VIII B, the controlling factors in Problems A
and D for unblocked Files 3 and 4 are tape time of the master file near zero activity, central
processor time at low activity, and tape time of the report file at moderate and high activity.
Problem B shows that central processor time is controlling at low activity and tape time of
the report file at moderate and high activity. The controlling factors in Problem C are tape
time for the master file at low activity and for the report file at moderate and high activity.
The graphs for blocked Files 3 and 4 indicate that central processor time is usually the controlling factor. The exceptions are where tape time of the master file is controlling: at
low activity in Problems A, C, and D.
Sorting (417:201. 2)
The standard estimate for sorting 80-character records by straight forward merging on magnetic tape was developed from the time for Standard File Problem A according to
the method explained in the Users' Guide, Paragraph 4:200.213, using a three-way merge.
Timing tables for the IBM standard sort routines were used to draw the second
sorting graph (Page 417:201. 220). A comparison of the two graphs indicates the advantage of
using large blocking factors and sophisticated routines such as these for sorting operations on
large-scale computers.
© 1963
Auerbach Corporation and Info, Inc.
12/63
IBM 7080
417:201.011
SYSTEM PERFORMANCE
s 201.
WORKSHEET DATA TABLE 1
CONFIGURA TION
WORKSHEET
1
(Unblocked
(mocked
VIIB
Files
VIIB Files
3 and 4)
3 and 4)
ITEM
Char/block
Records/block
K
msec/block
File 1
=
1,100
10
10
10
File 2
24.9
24.9
10.7
10.7
--
File 2
msec/block
msec/record
msec/detail
msec/work
msec/report
msec/block
for C. P.
Standard
Problem A
F
= 1.0
~1 = File 2
-- t--- File 4
~
----~
I-- b5 + b9
b7 + b8
a2
f-
~
~
--
---
and
a3 K
dominant
File 1 Master In
column.
21.1*
0
f-....
~f-
1. 37 I - _ _ O'~
2.01
0.20
0 •.2
~
-- ~
--
2.01
0.2
-
0.2
5.3
23.7
1.8
2.0
36.2
-0.5
'--
1-0
1-0
C.P.
0.2
0.2
--- -5.3
-
0.2
-
f- - f--5.3
-
23.7
f-23.7
1.8
--- 1.8
- - 24.9
- - 1.4
--- 2.0
- - 27.3
92.2
24.9
117.1
36.2
52.2
1.8
1.8
f-....-
2.0
36.2
-
4:200.1132
1-0
0.2
' - - f--
5.3
- - - --23.7
f---
f--
1.4
C.P.
1-0
C.P.
4:200.112
0
- -0 -
_ _ 1.77 __
-- -
0
-'---
- -0.5- i- - - 0.5
I-0.1
0.1
0.1
0.1
--t-- --- I - - - 1.1
1.1
1.1
- f--- - - - - - - -1.2
- I- - - 1.1
1.2
1.2
1.2
--- -1.8
File 2 Master Out
- File 3 Details
- 1.4
--- -
Unit of Measure
--
0.2
-0.5
-
t----
-
-
-~
4:200.114
1.8
--1.4
-86.0
- - ---21.1
2.0
f--
86.0
36.2
21.1
(character)
Std. routines
~ed
~cks
20,000t
-
___
lto23)
-
6 (Blocks 24 to 48)
Files
Working
Total
*
20,000t
-,-
0
700
2,000
4,880
300
27,880
10 records per block.
t Estimated IOCS space requirements.
12/63
8.6
0
o~
-0.20
File 4 Reports
Standard
Problem A
Space
~2_ f-
27.3*
f-- _ _1.~
Total
4
f--~* f - -
0
C.P.
3
--- --- --- -_-_16~
I-
--- -I- --- - I- - - -- 0
- -00 - - - ----- 00 0
File 4
I--
File 3
Central
Processor
Times
--
File 3
msec penalty
2
I---
8.6
- - t--- - -
REFERENCE
1,10'0
1,100
10
9.2
=
(mocked
Files
3 and 4)
1,100
1---
File 1
VillB
(File 1)
File 4
msec/switch
(Unblocked
Files
3 and 4)
(File 1)
File 3
InputOutput
Times
VillB
0
--- 700
I-
2,000
8,680
35,000 t
- -0 -700
-2,000
--- f - - 4,880
-500
-- --300
31,e80
42,580
_ _3_5,0~
f----
~
700
2,~
8,680
500
46,880
4:200.1151
417:201.100
IBM
7oBo
System Perform once
SYSTEM PERFORMANCE
f:j
. 114 Graph: • . . . . • • . ..
201.
·1
GENERALIZED FILE PROCESSING
• 11
Standard File Problem A
see graph below.
. 115 Storage space required
Configuration VII B
(unblocked files 3
and 4): • '. . . . . . . •
Configuration VII B
(blocked files 3
and 4): . . • . • . . . .
Configuration VIII B
(unblocked files 3
and 4): • • . • . . . . .
Configuration VIII B
(blocked files 3
and 4): . • • . . • . . .
· 111 Record sizes
Master file: •..•.. 108 characters (expanded
to 110 to facilitate internal
data movement).
Detail file: . . • . . . . 1 card.
Report file: •••.•.. 1 line.
. 112 Computation: ••....• standard.
• 113 Timing basis: •.•... using estimating procedure
outlined in Users I Guide,
4:200.113
27,880 characters.
31,880 characters.
42,580 characters .
46,880 characters.
100.0
7
4
2
10.0
7
Time in Minutes to
Process 10,000
Master File Records
4
2
VI~
1.0
7
4
2
,.-
--- ---~
,-
~
---""
~
-VUB
..."
-
""""""
VIII B
./
VIII B
-
--- ---"",",
~--
0.1
0.0
0.1
0.33
1.0
Activity Factor
Averag.e Number of Detail Records Per Master Record
Note: Dashed lines denote blocked Files 3 and 4;
Roman numerals denote standard system
configurations.
© 1963
Auerbach Corporation and Info, Inc.
12/63
IBM 7080
417 :201.120
§
201.
. 12
122 Computation: . . . . . . . standard.
Standard File Problem B
.123 Timing basis: • . . . . . using estimating procedure
outlined in Users I Guide,
4:200.12.
. 121 Record sizes
Master file: . . . . . . 54 characters.
Detail file: • • . . . . . 1 card.
Report file: • . . . . . . 1 line.
. 124 Graph: . . . . . . . . ..
see graph below.
1,000.0
7
4
2
100.0
7
4
2
Time in Minutes to
Process 10,000
Master File Records
10.0
7
4
2
-------------------VII B
1.0
7
~
4
!-"
/
2
0.1
~
.~
VIII£"'__
..,.,.-/--- V-7.
........
~
VIIL_-
-
-
.-;-
-VmB
,/
0.0
0.33
0.1
Activity Factor
Average Number of Detail Records Per Master Record
Note:
12/63
--
Dashed lines denote blocked Files 3 and 4;
Roman numerals denote standard system
configurations.
1.0
-'.
SYSTEM PERFORMANCE
§
417:201.130
201.
. 13
Standard File Problem C
. 131 Record sizes
Master file: . . . . . . . 216 characters.
Detail file: . . . . . . . 1 card.
Report file: . . . . . . . 1 line.
. 132 Computation: ...
. standard .
. 133 Timing basis: ..
. using estimating pl'ol'l'dul'l'
outlined in Users' Guillt' .
4:200. 13.
.134 Graph: . • . . . . . . . . . see graph below.
1,000.0
7
4
2
100.0
7
4
2
Time in Minutes to
Process 10,000
Master File Records
10.0
7
4
2
----
VII B
7
VIII
4
------ --VII B
--
1.0
1Y
.-
.."
po.--
--
-
~
VIII B
2
0.1
0.0
0.1
0.33
1.0
Activity Factor
Average Number of Detail Records Per Master Record
Note:
© 1963
Dashed lines denote blocked Files 3 and 4;
Roman numerals denote standard system
configurations.
Auerbach Corporation and Info, Inc.
12/63
417:201.140
§
IBM 7080
201.
. 14
. 142 Computation: . . . . . . . trebled.
Standard File Problem D
.143 Timing basis: . . . . . . using estimating procedure
outlined in Users I Guide
4:200.14.
. 141 Record sizes
Master file:. • . . . . . 108 characters.
Detail file: . . . . . . . 1 card.
Report file: . . . . . . . 1 line.
.144 Graph: . . . . • . . . . . . see graph below.
1,000.0
7
4
2
100.0
7
4
2
Time in Minutes to
Proccss 10,000
Master File Hecords
lO.O
7
4
2
-
1.0
,.
7
4
2
""..
~-
~
--AI!!!
VII B
~
--~
OIl!!!::
~
~-
- ...
-
VIII B
./V
t-""'"
o. 1
0.0
0.33
0.1
Activity Factor
Average Number of Detail Records Per Master Record
Note:
12/63
Dashed lines denote blocked Files 3 and 4;
Roman numerals denote standard system
configurations.
1.0
SYSTEM PERFORMANCE
417:201.200
20l.
.2
SORTING
f;J
.21
.212 Key size: . . . . . . . . . 8 characters .
.213 Timing basis: •.••.. using estimating procedure
outlined in Users' Guide,
4:200.213; three-way
merge .
. 214 Graph: • . . • . . . . . . . see graph below •
Standard Problem Estimates
. 211 Record size: •..•... 80 characters.
100
7
4
2
V~
.J
10
1
Jl
;;
,
2
JI
~
{/
1
v
IJ'
/
f
4
Time in Minutes to
Put Records Into
Required Order
"
/'
"I
#
~
LB
II
1
//
VllBh
4
b
/
/'
1/ 7
2
/
0.1
1
4
2
0.01
100
2
71,000
2
7 10,000
2
7100,000
Num ber of Records
Roman numerals denote standard System Configurations.
© 1963
Auerbach Corporation and Info. Inc.
4/64
IBM 7080
417 :201.220
§ 201.
.22
Configuration VmB: • IBM 70S0 Generalized
Sorting Program, using
timing data in IBM Publication C2S-6324; 7340
Hypertape Drives are used
for input, output, and
merging.
Merge order = 3.
mocking factor = 112.
IBM Sort Routine Times
.221 Record size: •••...• SO characters.
.222 Key size: .•••.•••• S characters.
• 223 Timing basis:
Configuration VrrB: • Sort SO, using timing data
in IBM Publication C2S6125.
Merge order = 3.
Blocking factor = 29.
. 224 Graph: ••••••.•••• see graph below.
100
7
4
2
II
10
7
f7
1/
4
/
2
Time in Minutes to
Put Records Into
Required Order
/
V
1
7
/
II
II
~
"
#'
"
V
4
.I
2
VII7
/
/
I
V
0.1
/VIIIB
~
4'
7
7'
./
4
1/
2
0.01
100
2
4
7 1,000
2
4
710,000
2
Number of Records
Roman numerals denote standard System Configurations.
4/64
4
7 100,000
417:211.101
PHYSICAL CHARACTERISTICS
§
211.
Weight,
pounds
Power,
KVA
BTU
per hr.
69
69
70
1,850
2,600
1,500
0.86
5.83
12,700
7,350
15,420
30
58
69
44
2,200
I, 100
1. 04
0.09
9,850
1,270
29
56
29
74
34
30
60
30.5
69
69
48
70
1,200
2,200
1,350
1,000
1.6
0.29
4.0
-
3,900
2,500
12,000
3,400
7908 Data Channel
38
72
70
500
0.62
5,800
1301 Disk Storage, Modell
1301 Disk Storage, Model 2
7631 File Control
33
33
38
85.5
85.5
32
68.5
68.5
70
3,625
3,825
500
7.5
9.0
2.5
16,700
20,000
4,800
1414-61-0 Synchronizer
1009 Data Transmission Unit
1011 Paper Tape Reader
1014 Remote Inquiry Station
73.5
29
32
24
31. 5
31
38
29
70
40
60
35
1,100
500
530
175
1.3
0.3
1.8
0.2
4,450
1,000
4,100
500
7502 Console Card Reader
40
27
41
200
-
740
7622 Signal Control
35
33
59
800
1. 82
Unit
7804
7102
7302
7305
Power Cabinet
Arithmetic & Logic Unit
Core Storage
Central Storage and
1-0 Control
7153 Control Console
729
7621
7340
7640
Magnetic Tape Unit
Tape Control
Hypertape Drive
Hypertape Control
Width,
inches
Depth,
inches
68
68
80
30
40
36
68
101
Height,
inches
4,500
General Reguirements
Temperature: .• . . . . . . . . . . . . . . . . . . • . . . . . . ..
Relative humidity: • . . . . . . . . . . . . . . . . . . . . . . . ..
Power: . . . . . . . . . . . • . . . • . . . . . . . . . . . . . . . . .
© 1963
60 to 80°F.
40 to 70%.
208V, 60 cycle, 3-phase, 4-wire
for all units except the
following:
•
1009 and 1014 require 115/208V,
60 cycle, I-phase.
•
1011 and 7631 require 208V,
60 cycle, I-phase.
Auerbach Corporation and Info, Inc.
12/63
417:221.101
IBM 7080
Price Dolo
PRICE DATA
§
221.
IDENTITY OF UNIT
CLASS
i\]odl'l
Name
;';0.
PRICES
Monthly
Rental
$
CENTRAL
PROCESSOR
7102
$
725.00
685,000
Central Storage and 1-0 Control
Two channels
Four channels
7,300
8,400
235.00
275.00
345,000
395,000
7153
Console Control Unit
1,500
41. 50
75,000
6091
Optional Feature
Real Time Clock
175
2.75
8,000
7B04
Power Unit
1,600
32.75
60,000
17,500
10,000
580.00
3BO.00
840,000
4BO,000
Arithmetic & Logic Unit
7302
Modell
Model 2
Core Storage
160,000 characters
BO, 000 characters
1301
Modell
Model 2
Disk Storage
One disk array
Two disk arrays
2, 100
3,500
13B.OO
23B.00
115,500
185,500
1302
Modell
Model 2
Disk Storage
One disk array
Two disk arrays
5,600
7,900
?
?
252,000
355,500
10
?
400
7950
7631
Model 2
Model 3
Model 4
3213
INPUTOUTPUT
Purchase
$
14,500
7305
Modell
Model 2
INTERNAL
STORAGE
Monthly
Maintenance
729
729
729
729
II
IV
V
VI
1302 Attachment for 7631 File
Control
File Control
For this system
For this system and a 1410
For this system and another
7000 series system (except 7072).
835
1,035
30.00
35.00
42,000
52,000
1,035
35.00
52,000
25
1. 00
1,250
700
900
750
950
116.00
12B.00
122.00
134.00
36,000
41,250
37,200
42,450
Optional Feature
Cylinder Mode
Magnetic
Magnetic
Magnetic
Magnetic
Tape
Tape
Tape
Tape
© 1963
Unit
Unit
Unit
Unit
Auerbach Carporation and Info, Inc.
12/63
417:221.102
IBM 7080
PRICE DATA (Contd.)
8221.
PRICES
IDENTITY OF UNIT
CLASS
Model
No.
INPUTOUTPUT
(Contd. )
7621
Model 2
Model 4
7830
7908
Modell
Model 2
Model 3
Model 4
Model 5
Model 6
Model 7
Model 8
Model 9
3221
3221
Optional Feature
Tape Switch
Data Channel
One low speed channel
Two low speed channels
Three low speed channels
Four low speed channels
Two high speed channels
Two high speed and one low speed
channels
Two high speed and two low speed
channels
Two high speed and three low
speed channels
Two high speed and four low speed
chalmels
Data Channel Attachment for 7305
(for all models)
Data Channel Attachment for 7102
(for models 5 - 9)
Hypertape Drive, Modell
Hypertape Control
1284
Optional. Feature
Automatic Cartridge Loader
1011
1014
7864
7871
7875
7502
7622
I-a Synchronizer, Model 6
Data Transmission Unit and 3228
Adapter
Paper Tape Reader and 5514 Adapter
Remote Inquiry Unit and 6136
Adapter
Telegraph I-a Feature
Telegraph Input Feature
Telegraph Output Feature
Console Card Reader and 2265
Attachment
Signal Control
Purchase
$
$
3,300
3,470
233.00
233.00
147,000
153,800
85
6.50
4,400
1,825
2,075
2,450
2,700
2,850
38.00
48.00
58.00
65.00
79.00
73,000
83,000
115,900
125,000
130,800
3,100
90.00
137,600
3,350
101. 00
144,400
3,600
112.00
151,200
3,850
123.00
158,000
25
-
$
1,000
250
2.00
11,800
1,300
3,400
130.00
102.00
78,000
218,000
125
16.00
7,500
850
12.00
43,350
700
600
15.'00
44.00
37,400
26,150
400
500
110
125
18.25
11. 00
3.25
3.25
22,500
30,500
6,750
7,750
375
18.25
18,750
1,500
22.00
26,000
,-
Note:
12/63
Tape Control
For 729 II, IV
For 729 II, IV, V, VI
7340
7640
1414
100D
~
Monthly
Monthly
Maintenance
Rental
Name
Monthly maintenance prices shown here are
applicable for first 36 months after installation.
,
I
\
IBM 1130
International Business Machines Corp.
AUERBACH INFO, INC.
PRINTED IN U. S. A.
"--.
IBM 1130
International Business Machines Corp_
/'
AUERBACH INFO, INC.
PRINTED IN U. S. A.
418:001. 001
IA
I
I
\
AUERBACH
STANDARD
EDP
IBM 1130
REPORTS
CONTENTS
"--
CONTENTS
Introduction . • • • • • . • . • . . . • • • • • • • . • . • • • . • • . • • • • . • . • • • . .
Data Structure . . . . • • • . . . . . . . • • . . . • . • • • • . . . . . • . . . • . . . . . .
System Configuration Typical Card System . • . • . • • • • • • . . • • . . • . . . • . . . . . .
I
IX
Desk Size Scientific System . • . . . • . . • • . . . . • . . . . . . . . .
Internal Storage Core Storage . . . . • . . . • . . . . . . . . . . . . . • • . . . . . . . . . . . . . .
Disk File . . . . . • . . . • . . . • . . . • . • • . . . . . . . • . . . . . . • . . • .
Central Processor . . . • . . . . . . • . . . . • • . . . • • . . . . • . • • • • • • • . • .
Console . • . • • . . . • . . . . • . . . . . . . . . . • • • • . . • . • • . • . • . • • . . . .
Input-Output: Punched Card and Tape 1442 Card Read Punch, Models 6 and 7 . . . . . • . . • • • . . . . . • . . .
1134 Paper Tape Reader . . . . . • • • . . • . . • . . . • • • • • . . • . . • . .
1055 Paper Tape Punch . . • . • . • . . • . . . . . . . . . . . . . . . . . • • .
Input-Output: Printer 1132 Printer . . • . . . . . . • . . . . • . . • . . . . . • . . . • . . . . . . . . . .
Input-Output: Other1627 Plotter, Models 1 and 2: . . . . . . . . . . . . . . . . . . . • . . . . . .
Console Printer/Input Keyboard . . . • . . . . . • • . . . • . • • . . . • . . .
Simultaneous Operations. . • . . • • . . • . . . . . . . . . • . • • • • • • • • • • • . .
Instruction List. . . • • . . • . . . • • . • • . . • . . . . . . . . . • . . . • . . . . . . .
Data Code Table • . • . . . • . . . . . . • • • • . . . . . . . • • . • • • • • • • • . • • .
Problem Oriented Facilities . . . . . • • • • • . . . . . • . • . • • • • • . . • . • . .
Process Oriented Language 1130 FORTRAN ..••••.•..•••••.•...•••..•......••••
Machine Oriented Language 1130 Assembler . • • . • . . . • . . . • . . • . • . • . . . . • • • • . • • • • . .
Program Translators 1130 Assembler .••••••.•...••.••••.••••.•••••••••••.
1130 FORTRAN .••••••..•.•..•••..•••.•.••••••••••.
Operating Environment 1130 Monitor System • • • • . . . . . . . • • . • . . . . • . . . . . . • • . . . .
System Performance ••.••••••.•••.•.•.••.•••••••.•••••..
Physical Characteristics . . • . . . . . . . . . . . • • • • • . • • . • • . . • • • • • • .
Price Data .••••..••..•..•.••.•••••••..••••••..••.•••.
418:011
418:021
418:031.1
418:031. 2
418:041
418:042
418:051
418:061
418:071
418:072
418:072
418:081
418:101
418:102
418:111
418:121
418:141
418:151
418:161
418:171
418:181
418:182
418:191
418:201
418:211
418:221
,
(
\
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
IBM 1130
418:001.002
/'
IBM 1130 Computing System, showing the disc cartridge (lower right) and the
1442 Card Read Punch (at rear).
5/66
fA.
AUERBACH
@
-1-
fA.
-
AUERBAC~
•
418:011. 100
STINDARD
EDP
IBM 1130
INTRODUCTION
REPORTS
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.
o
1131 Model 2A Central Processor - 4,096 core memory locations
and 512,000 disc storage locations.
o
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 Central 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.
o
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 Plotter 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 don~ 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
Apri11966.
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
'"
/
.1
418:021. 100
SlIm"
/AEDP
AUER.AC~
-
IBM 1130
"'ORTS
DATA STRUCTURE
DATA STRUCTURE
.1
STORAGE LOCATIONS
Size
Purpose or Use
16 bits
plus parity
basic addressable unit (data
or instruction) of core
storage.
Word (double
length):
32 bits
Character:
8 bits
Card column:
12 bits
Sector:
321 words
double-precision arithmetic
operands.
alphanumeric data (printers
and paper tape reader and
punch).
alphanumeric data
(card reader-punch).
basic addressable unit of
disc storage.
Name of Location
Word:
.2
IN FORMA TION FORMATS
Type of Information
Representation
Operand:
Instruction:
1 or 2 words.
1 or 2 words.
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
-&..
418:031.001
mNDARD
/AEDP
AUERBAC~
__
IBM 1130
SYSTEM CONFIGURATION
BEPBITS
"'~-----J
SYSTEM CONFIGURATION
Every IBM 1130 system includes one of four 1131
Central Processor Units. Each processor model
includes a console typewriter and core memory;
the core memory cycle time is 3.6 microseconds.
The Central Processor Units differ in core memory
size and auxiliary disc memory combinations as
follows:
Each peripheral device requires a special centra!
processor attachment feature. If input from
paper tape only is selected, a special feature,
the 1134 Loader, is required on the paper tape
reader to allow the loading of programs; this
feature is not available if a 1442 Card Read
Punch Unit is incorporated.
•
1131 ModellA contains 4,096 words of core
storage (no disc storage).
•
1131 Model 1B contains 8,192 words of core
storage (no disc storage).
•
1131 Model 2A contains 4, 096 words of core
storage and a 512, OOO-word disc drive.
•
1131 Model 2B contains 8,192 words of core
storage and a 512, OOO-word disc drive.
In addition, a special Storage Access Channel is
available for attachment of non-IBM peripheral
devices to any model Central Processor Unit. Non-.
standard peripheral devices, such as IBM magnetic tape units, can be connected to the Storage
Access Channel on a Request Price Quotation
(RPQ) basis. The Storage Access Channel can
accommodate a peripheral device operating at up
to 270,000 words per second; each word transferred can be up to 16 bits in length.
IBM offers a limited line of peripheral equipment
for the 1130. One of each of the following devices
can be connected to an 1131 Processor (any model):
•
1442 Card Read Punch, Model 6 or 7;
•
1132 Printer;
•
1134 Paper Tape Reader, Modell or 2;
•
1055 Paper Tape Punch, Modell; and
•
1627 Plotter, Model 1 or 2.
Standard Configurations
The IBM 1130 system is shown in the following
Standard System Configurations, as defined in the
Users' Guide, page 4:030.100:
•
Configuration I: Typical Punched-Card
System, and
•
Configuration IX: Desk-Size Scientific
System.
(Contd. )
5/66
A
AUERBACH
'"
418:031. 100
SYSTEM CONFIGURATION
.1
TYPICAL CARD SYSTEM; CONFIGURATION I
Deviations from Standard Configuration: . . . . . . • . . core storage capacity is 18% greater.
card reader is 60% slower.
card punch is 30% slower.
card read-punch does not have
separate card paths for read and
punch operations.
printer is 92% slower.
Equipment
Rental
4,096 words of Core Storage
1131 (Model 1A) Central Processor**
$
625*
Console Typewriter/Printer
1132 Printer:
82 lpm
260
1442 Model 7 Card Read Punch
380
TOTAL RENTAL:
*
$1,275
Price includes attachments for printer and card read-punch.
** A replaceable-cartridge disc drive capable of storing 512,000 words can be included
with the Central Processor (Model 2A) for an additional $200 per month. The disc drive
allows use of the 1130 Monitor System.
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
418:031. 200
.2
IBM 1130
DESK-SIZE SCIENTIFIC SYSTEM; CONFIGURATION IX
Deviations from Standard Configuration: •••.••.•. core storage is 20% greater.
paper tape punch is 48% faster.
paper tape reader is 300% faster.
Equipment
Rental
8,192 words of Core Storage
1131 (Model1B) Central Processor**
$
825*
Console Typewriter/Printer
1134 Tape Reader, Model 2:
60 char/sec
60
1055 Tape Punch:
14.8 char/sec
40
TOTAL RENTAL:
*
**
$
925
Price includes attachments for paper tape reader and punch units, and the 1134 Loader feature.
A replaceable-cartridge disc drive capable of storing 512.000 words can be included with the
Central Processor (Model 2B) for an additional $200 per month. The disc drive allows use
of the 1130 Monitor System.
/-
5/66
A
AUERBACH
'"
418:041. 100
~
AUERBACH
•
SUMOARD
EDP
IBM 1130
INTERNAL STORAGE
CORE STORAGE
REPORTS
INTERNAL STORAGE: CORE STORAGE
.1
GENERAL
.3
DATA CAPACITY
.11
Identity: . . . • . • • • . . in 1131 Models 1A and 2A
Central Processor Units
(4,096 words).
in 1131 Models 1B and 2B
Central Processor Units
(8,192 words).
.31
Module and System Sizes
Minimum
Storage
Identity:
Model1A
or 2A
Words:
4,096
Characters:
8,192
Instructions Short format:
4,096
Long format:
2,048
Modules:
1
. 12
Basic Use: . . . . . . . . working storage .
.13
Description
Core storage is housed in the 1131 Central Processor Unit and consists of 4, 096 word locations in
Models 1A and 2A and 8,192 locations in Models 1B
and 2B. Each word location consists of sixteen data
bits and one parity bit. An arithmetic data word
uses the high-order position of the word as a sign
bit. Single- and double-length load, store, and
arithmetic operations can be performed. Memory
cycle time is 3.6 microseconds per one-word
access.
.4
CONTROLLER:
.52
Simultaneous
Operations:
.53
Maximum
Storage
Model1B
or 2B
8,192
16,384
8,192
4,096
1
.. no separate controller.
. . • . . . none.
Access Time Parameters and Variations
Availability: . • . . . . . 17 months.
.531 For uniform access Access time: . . . • . • ?
Cycle time: . . . . . • . 3.6 JJsec.
For data unit of: •.. one 16-bit word •
.15
First Delivery: . . . . . November 1965.
.6
.16
Reserved Storage
• 14
Purpose
Number of
Locations
Locks
Index registers:
Arithmetic registers:
Logic registers:
Interrupt addresses:
Printer image output:
3
none.
none.
6
8
no.
PHYSICAL FORM
.21
Storage Medium: .••. magnetic core.
. 23
Storage Phenomenon: . direction of magnetization.
. 24
Recording Permanence
.241 Data erasable by
instructions: . . • . . .
• 242 Data regenerated
constantly: . . • . . . .
• 243 Data volatile: . . . . . •
.244 Data permanent: . . . .
.245 Storage changeable: ••
.7
PERFORMANCE
.71
Transfer Load Size: .. one or two 16-bit words.
. 73
Effective Transfer Rate
With self . Using straight-line
coding: . . . . . . . . . 92,500 words/sec.
Using programmed
loop: . . . . . . . . . . 53,200 words/sec.
no.
no.
.2
CHANGEABLE
STORAGE: . . . • . . . none.
.8
ERRORS CHECKS AND ACTION
Error
Check or
Interlock
Invalid' address:
Receipt of data:
none.*
parity check
Recording of data:
record parity
bit .
parity check halt or continue,
based on setting of manual
switch.
send parity
bit.
none possible.
yes.
no.
no.
no.
no.
.28
Access Technique: •.• coincident current.
.29
Potential Transfer Rates
Recovery of data:
Dispatch of data:
.292 Peak data rates Cycling rates: ...•• 278,000 cps.
Unit of data: . . . . • . 16-bit word.
Data rate: . . . . • • . 278,000 words/sec.
Timing conflicts:
Action
halt or continue,
based on setting of manual
switch •
* Addresses beyond limits of core storage are
"wrapped around" modulo memory size.
© 1966 AUERBACH Corporation and AUERBACH Info, Inc,
5/66
1.
418:042.100
STlN.m
AEDP
IBM 1130
INTER NAL STORAGE
DISK FILE
AUERBAC~
•
REPORTS
INTERNAL STORAGE: DISK FILE SYSTEM
.1
GENERAL
. 11
identity: . . . . . . . ... Disk File (in 1131 Central
Processor, Models 2A
and 2B only).
. 12
Basic Use: . . . . . . . . auxiliary storage .
.13
Description
transfers the device indications to the accumulator
register in the central processor. The following
conditions can be checked: disc busy, disc not
ready, arm positioned at track zero, data error,
and operation complete.
Programming of disc operations is facilitated
by the use of standard IBM subroutines. Three
versions of the disc subroutine are provided for
programming in the 1130 Assembler language.
One provides for reading or writing a maximum of
one sector per subroutine call. The second allows
reading or writing any number of consecutive
sectors up to the maximum number of sectors on
the disc. The third subroutine is basically the
same as the second but reduces the chance of
losing a revolution between sectors.
The 11;30 Disk File is an integral part of the Model
2A and 2B 1131 Central Processing Units. It
allows random or sequential access to a moderate
amount of storage. The recording medium is an
oxide-coated disc in an interchangeable cartridge.
Each cartridge can store 512,000 words of 16
data bits and 3 check bits each. Although only
one disc drive is available in an 1130 system, unlimited off-line storage capacity is provided via
the replaceable cartridge feature.
The desired subroutine is automatically inserted
into the user's program by the 1130 Assembler
when called. The calling sequence for the disc
subroutine specifies the name of the desired
subroutine and parameters indicating the operation to be performed. Included in the parameter
list are the core memory addresses of the input
or output area and the user's error routine. The
number of 16-bit words that are to be transferred
must be loaded by the programmer into the first
word of the data area; the second word of the data
area must be loaded with the beginning disc
sector address.
The disc is divided into 200 "cylinders;" each
cylinder consists of one track on the upper surface
and one track on the lower surface. A magnetic
head is provided for each disc surface; both heads
move horizontally in unison to one of 200 discrete
positions. Each track is divided into four sectors,
each having a data storage capacity of 321 words.
Normally the sector number is stored in the first
word location. All reading and writing operations begin at the first word of a sector; partial
sectors can be transferred. The eight sectors
in a cylinder are numbered consecutively, allowing a continuous read or write operation of up
to 2,568 words without repositioning the head
mechanism.
When an interrupt is generated as a result of a
disc operation, the subroutine checks for succ~ss
ful completion of the operation. If an error IS
detected the subroutine loads the accumulator
register' with status information and transfers
control to the user's error subroutine. After an
analysis of the error information, the user's
routine can either cause a haIt or cause the
operation to be retried.
The Assembler disc subroutine features a fileprotect technique which is dependent upon sector
numbering. The user must prerecord the sector
number in the first word location of each sector of
the disc, and then insert into the disc subroutine
the address of the first unprotected sector.
Every sector up to the specified sector is protected against overwriting by a "write-with-readback" instruction; any other write instruction will
write into a protected area.
The average time required to position the head
mechanism for randomly placed data is 750 milliseconds. An additional 20 milliseconds is required to allow the head mechanism to stabilize.
The rotational speed is 1,500 revolutions per
minute, which corresponds to 40 milliseconds per
revolution. Thus, the average total access time
for randomly placed data is 790 milliseconds. The
relatively slow access time will limit the usefulness of this unit for random processing.
Check bits are appended to each word recorded
to make the number of "1" bits in the word divisible by four. When reading, each word (including
check bits) is checked to verify that it is evenly
divisible by four (modulo-four check). A testable
indicator is set if an error is detected. An interrupt is generated at the completion of a read
or write operation.
Three reserve cylinders (24 sectors) are included
on the magnetic disc. If a cylinder is found to be
defective (i. e., a sector cannot be read or written
after 10 consecutive attempts), the Assembler
disc subroutine automatically assigns the address
of the defective cylinder to one of the reserve
cylinders. The reserve cylinders are not directly
accessible by the programmer.
The tracks or cylinders are not directly addressable. Instead, the disc command specifies the
distance to be moved in terms of cylinders. The
user's program must keep track of the position of
the access mechanism. A device status indicator
is set When the heads are positioned at the first
cylinder; this provides a reference point. The
user's program can check'the condition of device
status indicators by issuing a device status instruction. The execution of this instruction
5/66
A simplified disc subroutine is provided for
FORTRAN programming. Data is transferred to
the disc one sector (320 words) at a time. The
FORTRAN disc subroutine performs a read-back
fA
AUERBACH
'"
(Contd. )
,/
418:042.130
INTERNAL STORAGE: DISK FILE
. 13
.14
Description (Contd.)
.3
DATA CAPACITY
. .. 17 months.
.31
... November 1965.
Module and System
Sizes: . . . . . . . . • . 512,000 16-bit words
-(based on 320 words
per sector). Additional
storage is available with
extra cartridges, but only
one cartridge can be online at a time.
.15
First Delivery:
Reserved Storage: ... none .
.2
PHYSICAL FORM
. 21
Storage Medium: . . . . magnetic disc .
.22
Physical Dimensions
. 222 DiscDiameter:. • . . . . . . ?
Number on shaft: . . . 1.
· 23
Storage Phenomenon: . direction of magnetization.
. 24
Recording Permanence
· 241 Data erasable by
instructions: . . . . . .
.242 Data regenerated
constantly: . . . . . . .
. 243 Data volatile: . . . . . .
. 244 Data permanent: . . . .
.245 Storage changeable: ..
yes.
no.
no.
no.
yes.
word.
16 data bits plus 1
space bit plus 3 check
bits (20 bits total) per
word.
Data rate: . . . . . . . . 35,000 words/sec.
.4
CONTROLLER: . . . . . no separate controller .
· 44
Data Transfer Control
· 441 Size of load: ..
.442 Input-output
area: . . . . • .
• 443 Input-output area
access: . . • . . . . .
.444 Input-output area
lockout: . . . . . . . .
.445 Synchronization: ...
.447 Table control: . . . . .
.448 Testable conditions: .
Data Volume per Band of 1 Track*
Words: . . . . . . . . . . .
Characters: . . . . . . .
Digits: . . . . . . . . . . .
Instructions Short format: . . . . .
Long format: . . . . . .
1,280.
2,560.
2,560.
1,280.
640.
.26
Bands per Physical
Unit: . . . . . . . . . . . 200 per disc side (400 total
-per disc).
· 27
Interleaving Levels: .. 1.
.28
Access Techniques
.281 Recording method: ... two magnetic heads
which move horizontally
in unison on a single
arm .
. 283 Type of access Description of stage
Possible starting stage
Move heads to selected track: •.... yes, if different track is
selected.
Wait for beginning
of selected sector: . yes, if head is positioned
on selected track.
Read or write
immediately: ..•.. no.
16-bit word.
no.
automatic .
none .
arm at track zero, disk
busy, operation complete, data error .
ACCESS TIMING
Arrangement of Heads
.511 Number of stacks Stacks per system: ..
Stacks per yoke: ...
Yokes per system: ..
.512 Stack movement: . . . .
.513 Stacks that can access
any particular
location: . • . . . . . . .
.514 Accessible locations:
By single stack With no movement:.
With all movement:
By all stacks With no movement:.
2.
2.
1.
horizontally across disc,
to one of 200 discrete
positions.
1.
1 track.
200 tracks.
2 tracks (1 cylinder).
.52
Simultaneous
Operations: . . . . . . none.
.53
Access Time Parameters and Variations
Stage
*
Total:
Based on 320 words per sector.
. core storage.
·5
.291 Peak bit rates Cycling rates: . . . . 1,500 rpm.
Potential Transfer Rates
. 1 to 320 words (1 sector).
· 51
Move heads to
selected track:
Head stabilization:
Wait for designated
sector:
Transfer data:
.29
varies with location .
varies with location.
700,000 bits/sec/track.
Availability: ..
. 16
. 25
Track/head speed: ..
Bits/inch/track: ...
Bit rate per track: ..
.292 Peak data rates Unit of data: . . . . . .
Conversion factor: ..
check following a write operation. If an error is
detected, the write operation is repeated. In a
similar manner, read operations in which an
error was detected are repeated. If the read or
write operation has not been successfully completed
after 10 attempts, the system halts with an error
display on the console. Subroutines written in the
Assembler language for incorporation into a
FORTRAN program must use the FORTRAN I/O
subroutines.
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
Variation,
msec
Average,
o to 1,500
20
750
20
o to 40
10 msec per
sector
20
10
ill.§gg
800 msec
5/66
418:042.600
IBM 1130
.6
CHANGEABLE STORAGE
.61
Cartridges
.73
.611 Cartridge capacity: .. 512,000 words (1 disc).
.612 Cartridges per drive: . 1.
• 613 Interchangeable: . . . . yes.
· 62
Loading Convenience
.621 Possible loading While computing system is in use: .•.. no.
While storage system is in use: . . . . no.
· 622 Method of loading: ... by operator.
· 623 Approximate change
time: . . . . . . . . . . . 1 minute, plus device haIt
and start-up times of
approximately 1. 5 minutes
each .
. 624 Bulk loading: . . . . . . . no.
.7
PERFORMANCE
. 72
.74
Update Cycle Rate
With core storage: ... 1.15 references/second.
Note: Based on random access of one SO-word
record (160 characters); reading, updating,
and rewriting that record; and rereading
for verification.
.75
Read-Only Reference Cycle Rate
With core storage: ... 1. 26 references/second .
Transfer Load Size
Note: Based on random access and reading of one
SO-word record (160 characters) with no
updating or rewriting .
With core storage: ... 1 to 320 words per write or
read command (1 sector).
.8
Effective Transfer Rate
With core storage Bulk mode: . . . . . . . 7,100 words/sec; based on
random access and transfer of S, 192 words .
Cylinder mode: ..•. 2,900 words/sec; based on
random access and transfer
of 2,560 words (one
cylinder).
ERRORS, CHECKS, AND ACTION
Error
Check or Interlock
Invalid address:
Invalid code:
Recording of data:
subroutine check
all codes valid.
record modulo-4
check bits.
modulo-4 check
none possible.
Recovery of data:
Timing conflicts:
Reference to protected
area:
Device not ready:
Read error remaining
after 10 attempts
at recovery:
Write error remaining
after 10 attempts
at writing:
subroutine check.
check
Action
set bit in accumulator register.
set indicator bit.
set bit in accumulator
register.
set indicator bit.
subroutine check
set bit in accumulator
register.
subroutine check
set bit in accumulator
register.
./
./
5/66
A
AUERBACH
'"
418:051. 100
1.
SI ......
.
/AEDP
AUERBAC~
IBM 1130
REPORTS
CENTRAL PROCESSOR
L..----,..;...------J
CENTRAL PROCESSOR
.1
GENERAL
. 11
Identity: ..
.12
Description
tents of one of three index registers specified in
the instruction to produce a final memory address.
If no index register is specified, the contents of
the instruction counter are algebraically added to
the address field. The short instruction address
field consists of eight bits, which include one sign
bit. This allows memory addressing of up to 127
locations in either direction from the base address.
. 1131 Central Processing
Unit, Models lA, 1B, 2A,
and 2B.
The IBM 1130 system offers a choice of four processor models. The models differ only in respect
to memory capacity, as described below:
•
1131 Model 1A Central Processor - 4,096
core memory locations; no disc storage.
•
1131 ModellB Central Processor - 8,192
core memory locations; no disc storage.
•
1131 Model 2A Central Processor - 4, 096
core memory locations; 512,000 disc storage
locations.
•
1131 Model 2B Central Processor - 8,192
core memory locations; 512,000 disc storage
locations.
The long instruction format requires two contiguous
memory locations per instruction. One location
contains information pertaining to the operation,
indexing, and indirect addressing. The second
location is a 16-bit address field which can be
used with or without indexing and/or indirect
addressing. Average execution time is approximately 11 microseconds for single-word instructions and approximately 15 microseconds for
double-word instructions.
Although the instruction repertoire does not contain direct comparison instructions, comparisons
can effectively be accomplished by using arithmetic
instructions in combination with branch-on-condition
instructions.
Core storage and the Disk File (with 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 16-bit word.
The Central Processor provides an interrupt
facility which recognizes six priority levels. Interrupts are initiated only by input-output devices.
The program cannot inhibit interrupt requests, but
interrupts assigned the same or a lower priority
level than the interrupt currently being processed
must wait until its completion before being serviced.
Higher-priority interrupts can interrupt the processing of a lower-level interrupt condition.
All models are single-address, binary processors
with an essentially fixed word-:-Iength. The instruction repertoire consists of 29 instructions, most of
which can include indirect addressing and indexing
functions. Instructions for binary addition and
subtraction in both single-word and double-word
precision are provided. Overflow is recorded by
the processor, and conditional branch can be executed based on the status of the overflow indicator.
The arithmetic operations are algebraic, and
negative results are represented in two's complement form. The signed words allow further
computation without programmed recomplementing.
Binary, fixed-point multiply and divide instructions
are also included in the standard instruction set.
There are no optional processing facilities.
Upon initiation of an interrupt, the Central Processor automatically saves the contents of the
Instruction Address register in a pre-assigned
location. The interrupt program must save all
pertinent registers and data and restore this information upon exit from the interrupt routine. The
interrupt subroutine requests information to determine the device(s) which caused the interrupt and
the specific cause of the interrupt. A special
branch instruction, used to exit from the subroutine,
resets the interrupt indicator and allows processing
of the next interrupt. Subroutines are furnished by
IBM to handle the input-output interrupt processing;
see page 418:111.200.
Excluding shift operations, all instructions can
be written in either a "short" or "long" instruction
format.
The short instruction format, which requires one
memory location per instruction, allows indexing
but not indirect addressing. The short instruction
address field, or "displacement" field as it is
called by IBM, is algebraically added to the con.2
.21
Availability: ..
.14
First Delivery: ...•. November 1965.
PROCESSING FACILITIES
Operations and Operands
Operation and
Variation
\
. 17 months, as of April
1966.
. 13
.211 Fixed point Add/ subtract:
Double-word
add/ subtract:
Provision
Radix
Size
automatic
binary
15 bits plus sign.
automatic
binary
31 bits plus sign.
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
IBM 1130
418:051.211
.211 Fixed point - (Contd.)
Multiply Short:
Long:
Divide No remainder:
Remainder:
none.
automatic
binary
31 bits.
none.
automatic
binary
16 bits.
.212 Floating point Standard range
add/ subtract:
Extended range
add/ subtract:
Standard multiply:
Extended multiply:
Standard divide:
Extended divide:
subroutine
binary
8 and 23 bits.
subroutine
subroutine
subroutine
subroutine
subroutine
binary
binary
binary
binary
binary
8
8
8
8
8
.213 Boolean AND:
Inclusive OR:
Exclusive OR:
automatic
automatic
automatic
binary
binary
binary
16 bits.
16 bits.
16 bits.
.214 Comparison: . . . .
. none. *
.215 Code translation: . . . . code translations between
BCD, hexadecimal,
binary, and output format
codes are performed by
subroutines. The subroutine timings and the
specific code translations
available are shown in
Table I.
. 216 Radix conversion: ... radix converions between
BCD and binary forms
are performed by sub~
routines. The subroutines
and related timings are
shown in Table I.
· 217 Edit: . . . . . . . . . . . . none.
.218 Table look-up: . . . . . . none.
. 22
Special Cases of Operands
. 221 Negative numbers: ... two's complement.
.222 Zero: . . . . . . . • . . . . positive zero only.
.223 Operand size determination: . . . . . . . . normally 16 bits; 32 bits
in double-word add and
subtract instructions.
.23
Instruction Formats
.231 Instruction structure:. one-address.
· 232 Instruction layout:
Displacement
· 233 Instruction parts Name
Purpose
OP: • . . . . . . . . . . . operation code.
F: . . . . . . . . . . . . . instruction format
designation.
*
Comparisons can be made indirectly by using
the Subtract instruction followed by a Branch
and Skip on Condition instruction.
5/66
A
and
and
and
and
and
31
23
31
23
31
bits.
bits.
bits.
bits.
bits.
T: . . . • . . . . . . . . . tag to specify index
register or instruction
address.
Displacement: . . . • . value used to modify
contents of a specified
index register to produce
an effective address; also
to specify number of bit
positions to be shifted.
IA: . . . . . . . . • • . . indirect addressing flag.
Address: . • . . . . . . designates operand address .
Spec: . . . . . . . . . . . used only by MDX instruction.
.234 Basic address
structure: .•••.•.. I-address.
.235 LiteralsArithmetic: . . • . . . . none.
Comparisons and
tests: ...•..•.•• none .
Incrementing
modifiers: . . . . . . 16 bits .
.236 Directly addressed operands Internal storage
type: • • . • . . . . . . core storage.
Minimum size: . . . . one 16-bit word.
Maximum size: . . . . two 16-bit words.
Volume accessible: . 255 locations using short
instruction format; total
core memory using long
instruction format.
· 237 Address indexing .2371 Number of methods: . one .
. 2372 Name • . . . . . . . • • . normal indexing.
.2373 Indexing rule: . • . . . tag fields of both long and
short format instructions
specify contents of an
index register to modify
instruction operand or
displacement field value.
· 2374 Index specification: .. bits 6 and 7 of instruction
word.
· 2375 Number of potential
indexers: . . . . • • . . 3 .
. 2376 Addresses which can
be indexed: . . . . . . operand addresses in
arithmetic, logical, load,
and store instructions.
· 238 Indirect addressing .2381 Recursive: . . . . . . . no; single stage only.
(Contd.)
AUERBACH
®
/
4 t 8:051. 2382
CENTRAL PROCESSOR
TABLE I: STANDARD IBM SUBROUTINES FOR CODE AND RADIX CONVERSIONS
Conversion
Per
Conversion
Per Reverse
Conversion
16-bit word.... 6-word
field containing one
sign word and 5 words
of one card-code decimal digit per word.
1130
1111
IBM Card Code
Hexadecimal
16-bit word-4 words;
one hexadecimal character per word.
620
760
IBM Card Code
Paper Tape
two words, 1 card-code
character per wordone 16-bit, 2-character
word in paper tape code.
1404
1394
IBM Card Code
Printer
EBCDIC Subset
(80 characters)
two characters, one per
16-bit word ...... one 16bit word containing 2
EBCDIC characters.
1251
1300
IBM Card Code
Printer
EBCDIC
(256 characters)
same conversion as
HOLEB, above, except
faster conversion.
270
394
IBM Card Code
Typewriter code
two card-code characters
...... one 16-bit word with
two typewriter-code
characters.
1311
-
Paper Tape
Typewriter code
two paper tape characters
per word- two typewritercode characters per word.
1577
-
Paper Tape
Printer
Two paper tape characters in one I6-bit word
_two EBCDIC characters
in one 16-bit word.
1466
1446
two EBCDIC characters
in one 16-bit wordtwo typewriter-code
characters.
1347
-
From
To
Description
Binary
IBM Card Code
Binary
EBCDIC Subset
(80 characters)
EBCDIC
Subset
*
Typewriter, code
The times in these columns are to perform the conversion descril:!ed or '.'1e ~ of
this conversion, as indicated. Except for the conversions from or to binary form,
these times do not include the subroutine initialization times which apply to each field
converted; the initialization times vary between 250 and 600 !lsec.
.2382 Designation:
. a one in bit 8 of long
instruction format only.
· 239 Stepping.2391 Specification of
increment: . . . . . . .
· 2392 Increment sign: . . . .
· 2393 Size of increment Short instruction: ..
Long instruction: ••
· 2394 End value: . . . . . . . .
.24
Execution Time, !lsec'
within instruction.
+ or -.
+127 to -128.
+32,767 to -32,768.
index or memory value
changes sign or becomes
zero.
Special Processor Storage
Category of Number of
storage
locations
Size in
~
Register:
1
16
Register:
1
16
Core
Memory:
3
16
Pro~am
usage
single-precision
arithmetic.
double-precision
arithmetic and
extended right
and left shifts.
Index registers.
.3
SEQUENCE CONTROL FEATURES
.31
Instruction Sequencing: sequential.
.32
Look-Ahead: . . • . . . . none.
· 33
Interruption
.331 Possible causes:
• conditions which can cause
interrupts vary for the
different types of I/O
devices but include normal
end of operation, abnormal
end, and inability to
respond to an I/O instruction.
· 332 Control by routine: ... no program control of
either the priority order
of interrupts or inhibiting
interrupt action.
· 333 Operator Control: ..• a manual interrupt can be
initiated by the operator
via a console keyboard
request.
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
418:051. 334
IBM 1130
.334 Interruption
.42 Processor Performance in Microseconds
conditions: ...•.• ,there are six interrupts, divided into six priority levels; .421 For random addresses Fixed Point **
each interrupt is associated
Single
Double
with one peripheral device
word
word
or a group of devices. The
23.2
c = a -I- b:
34.6
interrupt hardware delays
23.2
32.4
b = a + b:
execution of interrupts of
32.6N
23.2N
Sum N items:
equal or lower priority than
44.5
c = ab:
the one currently being proc98.4
c = alb:
essed. A higher priority
Floating Point*
interrupt request will cause
Standard
Extended
immediate interruption of a
precision
precision
lower-level interrupt in
460
440
c = a + b:
process.
460
440
b = a + b:
440N
460N
Sum N items:
.335 Interruption process 560
790
c = ab:
Disabling inter2060
766
c = alb:
ruption: . . . . • . . . no interrupt disable facility
Fixed Point**
.422 For arrays of data is provided.
Double
Single
Registers saved: ... contents of Instruction
word
word
Register.
~
63.7
ci = ai + b j :
Destination: . .. .. core memory locations
75.2
63.7
bj = ai + bj:
0008 through 00013, de37.7N
45.5N
Sum N items:
pending upon cause of
100.8
c
=
c
+
aibj:
interrupt.
Floating Point *
Standard
Extended
.336 Control methods precision
precision
Determine cause: ... own coding.
460
440
ci = ai + bj:
Enable interruption: . own coding.
460
440
b j = ai + bj:
Sum N items:
440N
460N
.34 Multiprogramming: .. no special provisions.
1230
1020
c = c + aibr
. 35
Muiti-Sequencing: .. ; none.
.4
PROCESSOR SPEEDS
. 41
Instruction Times in Microseconds
. 411 Fixed point Add/ subtract:
Double-word
add/ subtract:
Multiply
Divide:
.412 Floating point* Add:
Subtract:
Multiply:
Divide:
• 413 Additional allowance
for -
. 414
. 415
.416
. 417
. 418
5/66
* Subroutines .
** Short instruction format .
Short
format
Long
format
12.2
25.7
76.0
15.3
29.3
79.6
8.0
Standard
precision
460
560
560
766
Short
format
\.,
. 423 Branch based on comparison Numeric data: . . . . . 77.1.
. 424 SWitchingUnchecked: . . . . . . . 26.6.
Checked: . . . . . . . . 33.7.
List search: • . . . . . 45 .
.425 Format control: . . . . . see Table I.
.426 Table look-up, per comparison For a match: . . . . . . 67.3.
For least or
greatest: . • • . . . . . 68. O.
For interpolation
point: • . . . . . • . . . 54. 2
.428 Moving, average time per word Using single-word
load and store
instructions: . . . . . 45.3.
Using double-word
load and store
instructions: . . . . . 25.9 .
.5
ERRORS, CHECKS, AND ACTION
iT:""2
Extended
precision
440
490
790
2060
Long
format
Indexing:
~
~
Indirect addressing: 4.0
4. 0
Recomplementing:
0
0
ControlCompare:
no comparison instructions.
Branch:
'3.6
7.2
Counter control....;
Step and test:
4.5
18.5
Edit: . . . . . . . . • . . . no edit instructions.
Convert: . . . • . . . • . . see Table I.
Shift: . • . . • . . . . . . . 3.6 + O. 45(N-3), where N
equals number of positions
shifted.
Check or
Interlock
Action
Overflow or zero
divisor:
check
set programtestable indicator .
Invalid data:
Invalid operation
none.
check
automatic processor
halt .
Arithmetic error:
none.
Invalid address:
none. *
Receipt of data:
none .
Dispatch of data:
none .
* Addresses outside of memory range are resolved
modulo memory size.
A
AUERBACH
@
(
-1.
418:061. 100
Sl"',,"
AEDP
AUERBAC~
-
•
IBM 1130
REPaRTS
CONSOLE
CONSOLE
.1
GENERAL
.11
Identity: . . . . . . . . . . 1130 Desk Console.
.12
Associated Unit: . . . . input keyboard and output
printer are included with
the 1131 Central Processing Unit.
.13
Switches on the console provide the following
facilities:
o
Stopping or continuing on memory parity
error condition,
'"
Loading the instruction register with address
data from the console entry switches,
III
Automatic loading of the first card or paper
tape record into memory,
•
Stepping through instructions or parts of
instructions under manual control,
o
Implementation of a program trace by causing
an interrupt after the execution of each program instruction,
<:I
Displaying of core storage addresses, and
o
Storing of console entry switch information
into core storage.
Description
The Console is an integral part of all models of
the 1131 Central Processing Unit. It includes
an input keyboard, a console printer, console entry switches, control switches, and a display
panel. This compact unit allows an operator a
complete view of the status of an 1130 system.
The photograph on page 418: 001. 002 provides an
overall view of the Console.
The keyboard allows an operator to enter data and
instructions into core storage under program control. The console printer provides for printed
output from core storage at a peak speed of 15.5
characters per second; see Report Section 418: 102.
In addition to the keyboard option, console entry
switches can be used for storing data and instructions into core memory under either manual or
program control.
The display panel permits operator observation of
machine "Ready" and "Run" status, carry, overflow, memory parity, and interrupt level conditions. The contents of the instruction and operand
address registers and the contents of registers
used in arithmetic computations are displayed by
means of lights. The lights are marked off in
groups of fours to represent hexadecimal format.
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
-1. """"
418:071. 100
~\\EDP
IBM 1130
INPUT-OUTPUT
1442 CARD READ PUNCH
AUERBACH
R[PORTS
~_...;..,,......--_.J
INPUT-OUTPUT: 1442 CARD READ PUNCH
.1
GENERAL
. 11
Identity:......... 1442 Card Read Punch,
Models 6 and 7.
. 12
cards to be punched can be. either blank or
pre-punched. High-speed skipping is not
pOSSible; each blank column requires a full
punch cycle .
Des~ription
Card punching speeds depend upon .the number of
consecutive columns punched in each card. Model
6 punches at 80 columns per second and Model 7
at 160 columns per second;' To the punching times
must be added the time to feed and position each
card: 210 milliseconds for Model 6 and 160 milliseconds for Model 7. Punching speeds for various
numbers of columns are shown in Table I. When
reading and punching are being done on the same
card, the reading is performed during the punch
positioning time, and the overall read-punch operation proceeds at the same speed as punching alone.
The 1442 Card Read Punch is a combination inputoutput unit for standard 80-column punched cards.
From a single 1, 200-card hopper, the cards are
fed serially by column past a single photoelectric
reading station, past a single punching station,
and into a 1, 300-card radial stacker. A second
stacker, which may be program-selected, is
standard on both models. Cards can be loaded
and removed without stopping the unit. All format
control is by the stored program; there is no
plugboard. Only one 1442 Card Read Punch can be
connected to an 1130 system; the 1442 Attachment
feature (4454) is required in the processor.
The user's program can check the condition of the
device status indicators by issuing a device status
instruction. The execution of this instruction transfers device indications to the accumulator register
in the central processor. The following conditions
can be checked: read or punch errors, device not
available, device busy, operation complete, last
card, and read or punch interrupt.
A single 1442 can handle either an input or an output file. Alternatively, it can read data from and
punch results into the same card, or results can
be punched into trailer cards in the same file.
When cards are being read continuously and no
punching is being done, maximum speeds are 300
cards per minute for Model 6 and 400 cards per
minute for Model 7. Each column is read twice
and the results are compared. Checks are mad-e
for invalid characters, improper registration, and
malfunctions of the light-sensing mechanism. No
automatic conversion from the card code is performed by the 1442; conversions. must be performed by programming. Standard· IBM subroutines are provided to execute this function.
Assembly-language programming of card read-punch
operations is facilitated by the use of a standard
IBM subroutine. The subroutine is automatically
inserted in the user's program by the 1130 Assembler when called. The calling sequence for .the card
subroutine specifies the name of the desired subroutine and parameters indicating whether a read,
punch, operation-complete test, feed, or stacker
selection operation is to be performed. Included
in the parameter list are core memory addresses
of the input or output data area and the user's
error routine (optional). The number of columns
of card data that are to be transferred must be
loaded by the programmer into the first word of
the data area. Any number of columns can be
specified. Each 16-bit memory location contains
one card column of data in the 12 high-order bit
positions.
To maintain peak reading rates, the next read
command must be given within 35 milliseconds
(Model 6) or 25 milliseconds (Model 7) after the
end-of-operation interrupt. If these timing limits
are not met, the peak reading speed without punching drops to 285 cards per minute for the Model 6
and 375 cards per minute for the Model 7.
Card punching, like reading, is performed serially
by column, beginning at column 1, and terminated
after any number of columns from 1 to 80. The
TABLE I: IBM 1442 PUNCHING SPEEDS
Last Column
Punched
1
10
20
30
40
50
60
70
80
5/66
Total Punch
Cycle Time (msec)
Model 6
Model 7
229
341
466
591
716
841
966
1,091
1,216
169
226
288
351
413
476
538
601
663
A
Cards per Minute
Model 6
Model 7
202
176
127
102
84
71
62
55
49
355
265
208
171
145
126
112
100
91
(Contd.)
.,
AUERBACH
INPUT- OUTPUT:
. 12
418:071. 120
1442 CARD READ PUNCH
Description (Contd.)
Data read from the card reader can be converted
into a number of other codes by standard conversion
subroutines provided by IBM. See Table I in
Section 418:051, Central Processor.
Two versions of the Assembler card subroutine
are available. One requires operator intervention upon detection of a card error condition by
the subroutine; the second subroutine allows
automatic transferral to a user's error routine.
Checks are made for card read-punch hardware
conditions (e. g., last card, read or punch check,
etc.) and for validity of the calling sequence
specifications. When an interrupt is generated
as a result of a card operation, the subroutine
checks for successful completion of the operation.
If an error is detected, the first subroutine causes
a halt to await operator action; the second subroutine loads the accumulator register in the central
processor with coded status information and transfers control to the user's error routine. After an
analysis of the error, the user's routine can either
cause a halt or cause the operation to be retried.
In general, simplified I/O subroutines are sup-
plied for use in FORTRAN-coded programs .
These subroutines do not allow any overlapping
of a peripheral operation with non-I/O processing
or with the operation of another device. Data to
be punched must be stored in unpacked (one
character per word) EBCDIC format. The subroutine converts this to the IBM Card Code before punching. Data input from the card reader
is expected to be in mM Card Code and is converted by the subroutine to unpacked EBCDIC format. Special subroutines are provided to convert
between the unpacked EBCDIC format and the
data codes and formats of the various peripheral
devices.
The FORTRAN I/O subroutines utilize the Accumulator, the Accumulator Extension, and Index
Registers 1 and 2. Prior to entry into a subroutine, the contents of those registers must be
saved, if required, and replaced with constants
whose values depend on'the I/o operation. When
an error condition or malfunction is detected, the
subroutine causes a program halt to allow the
operator to take corrective action; the program
can be continued if the operator can correct the
condition.
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
418:072.100
&.
ST"""
/AEDP
IBM 1130
INPUT-OUTPUT
PUNCHED PAPER TAPE
EQUIPMENT
AUERBAC~
REPORTS
~
INPUT-OUTPUT: PUNCHED PAPER TAPE EQUIPMENT
.1
GENERAL
. 11
Identity:....
. 12
Description
register in the central processor. The following
conditions can be checked: punch not ready, punch
busy, reader not ready, reader busy, punch-initiated
interrupt, and reader-initiated interrupt. The status
condition of the reader should be checked before issuing a read command. If the reader is busy or not
available, erroneous data may be transferred to
memory without producing an error indication.
.. 1134 Model 1 and 2 Paper
Tape Reader.
1055 Modell Paper Tape
Punch.
One 1134 Paper Tape Reader and/or one 1055
Paper Tape Punch can be connected to an IBM 1130
system. The 1134 Reader requires the Paper Tape
Reader Attachment (3624); the 1055 Punch requires
the Paper Tape Punch Attachment (7923). In a
paper tape-oriented system (i. e., one with no 1442
Card Read Punch), the 1134 Loader feature (3624)
is required to permit program loading.
Programming of punched tape operations in 1130
Assembler language is facilitated by the use of standard IBM subroutines; two subroutines are available. One routine is used if simultaneous reading
and punching are desired; the other routine is used
to operate one device at a time in sequential fashion.
The paper-tape subroutines check for paper-tape
hardware conditions (e.g., device not ready), and
also for validity of the calling sequence specifications.
A subroutine is automatically inserted in the user's
program when called. The calling sequence for the
paper-tape subroutine specifies the name of the
desired subroutine and parameters indicating whether
read, punch, or a device busy check operation is to be
performed. Included in the parameter list is the core
memory address of the output data area and the user's
error routine. The number of words (characters)
that are to be transferred must be loaded by the programmer into the first word of the data area. Any
number of words can be specified that is within
range of the available core memory. Each 16-bit
memory location contains one character. Normally,
the IBM PTTC /8 code is used, but data can be in
either binary or character code form, as an entire
8-bit binary image is transferred by the subroutine.
The 1054 Paper Tape Reader is no longer offered
with IBM 1130 systems.
.121 1134 Paper Tape Reader
The 1134 Paper Tape Reader reads one-inch, fullypunched, 8-track paper tape at a maximum rate of
60 characters per second. Modell reads strips
only; Model 2 reads strips or spooled tape. The
reel capacity of Model 2 is 1,000 feet.
Each read command transfers one punohed-tape
character to one core memory location. An interrupt occurs after each character is read into core
memory. To obtain maximum reading speed, read
commands must be issued within 60 milliseconds
after an interrupt is generated.
In a paper tape-oriented 1130 system, the 1134
Loader feature allows automatic loading of instructions into core storage. When the Program Load
switch on the console is activated, data from the
1134 Paper Tape Reader is loaded into core storage beginning at location zero. Only four channels
are read from the tape, and four 4-bit characters
are stored in each 16-bit word location. Reading
is halted·when a punch in the fifth channel is
sensed, and a branch is made to the instruction
just loaded in the first word of memory. Subsequent action is under control of the stored program.
In general, simplified I/O subroutines are supplied
for use in FORTRAN-coded programs. These subroutines do not allow any overlapping of a peripheral
operation with non-I/O processing or with the operation of another device. Data to be output via the
paper tape punch must be stored in unpacked (one
character per word) EBCDIC format; the subroutine
converts this to the PTTC/8 code before punching.
Data input from the paper tape reader is expected
to be in PTTC/8 code and is converted by the subroutine to unpacked EBCDIC format. Special subroutines are provided to convert between the unpacked
EBCDIC format and the data codes and formats of
the various peripheral devices.
. 122 1055 Modell Paper Tape Punch
The 1055 Modell Paper Tape Punch fully punches
one-inch, 8-track paper tape at a maximum rate of
14.8 characters per second. Each write command
transfers the'contents of one core memory location
containing one paper tape character to the punch.
An interrupt occurs after each character is transferred. Delete codes (punches in all 8 channels)
and blank codes (no punches) can be punched manually by the operator.
The FORTRAN subroutines for punched tape operations restrict input or output records to a maximum
size of 80 characters, not including case shift characters. The FORTRAN I/O subroutines utilize the
Accumulator, the Accumulator ExtenSion, and Index
Registers 1 and 2. Prior to entry into a subroutine,
the contents of these registers must be saved, if required, and replaced with constants whose values
depend on the I/O operation. When an error condition
or malfunction is detected, the subroutine causes a
program halt to allow the operator to take corrective
action; the program can be continued if the operator
can correct the condition.
.123 Programming
The user's program can check the condition of the
device status indicators by issuing a device status
instruction. The execution of this instruction
transfers the device indications to the accumulator
5/66
IA
AUERBACH
(j)
- 1.
418:081. 100
STAHCA"
/AEDP
AUERBAC~
-
~
IBM 1130
INPUT-OUTPUT
1132 PRINTER
IIEPORTS
INPUT-OUTPUT: 1132 PRINTER
.1
GENERAL
. ll
Identity: ..
.12
---
.. 1132·Line Printer.
Description
The IBM 1132 Line Printer uses a typewheel
(drum) printing mechanism. The printer prints at
a maximum rate of llO single-spaced lines per
minute when printing numeric data only, and 82
single-spaced lines per minute when printing
alphanumeric data. The print line consists of
120 print positions, horizontally spaced at 10
characters per inch. Each print position is capable of printing 48 different symbols: 26 alphabetic,
10 numeric and 12 special characters. Only one
1131 Printer can be used in an ll30 system; it
requires a Printer Attachment (#3616), and an
Expansion Adapter (#3859) on the ll31 Central
Processor.
Form width can vary from 4.75 inches to 16.75
inches (edge to edge). Maximum form length is
22 inches at a line spacing of six lines per inch
and 16.5 inches at eight lines per inch.
Spacing and skipping are controlled by a i2-channel paper tape carriage mechanism which is activated by the stored-program printer instructions.
Skipping speed is approximately 10 inches per
second. The user's program can check the condition of the device status indicators by using a device status instruction. Execution of this instruction transfers the device indications to the accumulator register in the central processor. The
following conditions can be checked: forms position, printer busy, forms loaded, carriage busy,
skipping completed, and a read-emitter response.
Eight words of core storage are reserved to hold
the results of a scan operation and indicate the
print positions to be printed during each "cycle" or
character position of the typewheels. The'programming of print operations for the ll32 Printer
is relatively complex and requires close timing
by the programmer for optimum performance.
Each scanning, printing, and skipping operation
must be initiated by a separate command. Printing
and skips of greater than one line must be stopped
by an instruction.
\
Assembly-language programming of printer operations is facilitated by the use of a standard
IBM subroutine. The subroutine is automatically
inserted in the user's program by the 1130 Assembler when called. The calling sequence for the
printer subroutine specifies the name of the desired subroutine and parameters indicating whether
the operation to be performed is an alphanumeric
print, a numeric-only print, a skip under control
of the carriage tape loop, or a test for previous
operation complete. Included in the parameter list
are the core memory addresses of the output data
area and the user's error routine. The number of
IS-bit words that are to be transferred must be
loaded by the programmer into the first word of the
data area. A maximum of 60 words (120 characters) can be specified for each printer subroutine
call. Each 16-bit memory location must contain
two EBCDIC characters in the printer output area.
Conversion to EBCDIC characters can be performed
from a number of other codes by standardized
conversion subroutines provided by IBM. See Table
I in Section 418:051, Central Processor.
The Assembler printer subroutine checks printer
hardware status conditions (e.g., forms check), and
also the validity of the calling sequence specifications.
When an interrupt is generated as a result of a
printer operation, the subroutine checks for successful completion of the operation. If an error is detected, the subroutine loads the accumulator register
with status information and transfers control to the
user's error subroutine. After an analysis of the
error information, the user's routine can either cause
a halt or cause the operation to be retried.
In general, simplified I/o subroutines are supplied
for use in FORTRAN-coded programs. These
subroutines do not allow any overlapping of a peripheral operation with non-I/O processing or with
the operation of another device. Data to be output
to the printer must be stored in unpacked (one
character per word) EBCDIC format. Special
sub;routines are provided to convert between the
unpacked EBCDIC format and the data codes and
formats of the various devices. The FORTRAN
I/O subroutines utilize the Accumulator, Accumulator Extension and Index Registers 1 and 2. Prior
to entry into a subroutine, the contents of these
registers must be saved, if required, and replaced
with constants whose values depend on the I/O
operation. When an error condition or malfunction
is detected, the subroutine causes a program halt
to allow the operator to take corrective action; the
program can be continued if the operator can
correct the condition.
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
-&
418: 101. 100
STANDARD
AEDP
IBM 1130
INPUT-OUTPUT
1627 PLOTTER
AUERBAC~
•
REPORTS
INPUT-OUTPUT: 1627 PLOTTER
.1
GENERAL
. 11
Identity: . . . . • . . • . . IBM 1627 Plotter Model 1.
--IBM 1627 Plotter Model 2.
.12
Description
mand is issued, loss of information will probably
occur without an error indication to the program .
Assembly-language programming of plotter operations is facilitated by the use of a standard
IBM subroutine. The subroutine is automatically
inserted in the user's. program by the 1130 Assembler when called. The calling sequence for the
plotter subroutine specifies the name of the desired
subroutine and parameters indicating whether a
write or test status operation is to be performed.
Included in the parameter list are the core memory
addresses of the output data area and the user's
error routine. The number of points to be plotted
must be loaded into the first word of the output data
area. The plotter subroutine checks for plotter
hardware status conditions (e. g., plotter not
ready), and also for validity of the calling sequence
specifications.
When an interrupt is generated as a result of a
Plotter operation, the subroutine checks for
successful completion of the operation. If an
error is detected, the subroutine loads the
accumulator register with status information
and transfers control to the user's error subroutine. After an analysis of the error information, the user's routine can either cause a halt
or cause the operation to be retried.
The 1627 Plotter uses a Calcomp-designed drum
plotting mechanism to convert tabulated digital data
into graphic form. Models 1 and 2 are basically
the same, except for operating speeds and form
dimensions. The differences are summarized in
Table 1.
The 1627 plotters are two-axis recorders for plotting one variable against another. Each plotter consists of a ballpoint pen mounted on a carriage and
a bidirectional recording drum. Output words
from the computer direct pen carriage movement
and drum rotation, as well as the movement of the
pen against or away from the recording surface.
The pen carriage moves in the Y axis (horizontally),
and the drum moves in the X axis (vertically). All
movements are one step at a time, and are independent; i. e., diagonal movements are possible in one
operation.
The user's program can check the condition of the
device status indicators by issuing a device status
instruction. The execution of this instruction
transfers the device indications to the accumulator
register in the central processor. The following
conditions can be checked: operation complete,
device not ready, and device busy. The busy condition should be checked before issuing a plotter
write command; if the plotter is busy when a com-
IBM does not currently provide a subroutine to
control the plotter for FORTRAN-coded programs.
Users desiring to use a plotter in connection with
FORTRAN coding will have to write their own subroutines.
TABLE I: CHARACTERISTICS OF THE IBM 1627 PLOTTERS
Characteristic
Speed
X, Y increments,
steps/minute
Raise or lower pen,
operations/minute
Modell
Model 2
18,000
12,000
600
600
0.01
Increment Size, inches
Chart
Paper
5/66
Width, inches
Plotting width,
inches
Length, feet
Sprocket -hole
dimensions
12
11
120
O. 130 inch dia.
on 3/8-inch
centers
A
AUERBACH
<>
0.01
31
29.5
120
0.188 inch dia.
on 1-inch
centers
/
·418:102.100
/&.
AUERBACH
STAN~ARD
EDP
IBM 1130 SYSTEM
INPUT-OUTPUT
CONSOLE PRINTER
REP~RTS
INPUT -OUTPUT: CONSOLE PRINTERIINPUT KEYBOARD
.1
GENERAL
.11
Identity: ..
• 12
. IBM 1130 System Console
Printer/Input Keyboard.
Description
The console keyboard and printer are an integral
part of every 1131 Central Processing Unit. These
two units are independent; i. e., the keyboard keys
are not mechanically linked to the printing mechanism. Data entered via the keyboard is stored in
core storage, and a separate, programmed operation is required to print this data on the console
printer.
.121 Console Printer
The console printer is a modified IBM Selectric
Typewriter that operates at a peak speed of 15.5
characters per second. Eighty-eight different
characters can be printed. The characters include
the 26 alphabetic characters (both upper and lower
case), 10 numeric digits, and 26 symbols which
include FORTRAN symbols. There are seven
printer control characters: carriage return, tabulate, space, backspace, shift to red, shift to
black, and line feed. The maximum printed line
width is 120 character positions; horizontal spacing
is 10 characters per inch and vertical spacing is
6 lines per inch.
Information to be printed is stored in core storage,
one character per memory word, in Console Code;
see Section 418:141. An automatic interrupt occurs after the console printer has printed each
character or completed each functional operation
(carriage return, line feed, etc.)
The user's program can check the mechanical
availability of the console printer and whether the
printer is busy. The busy indicator should be
checked by the program before issuing a write order
to the printer. If a write order is issued when the
printer is in busy status, the transferred information will be lost; no program indication is produced
as a result of this error.
. 122 Keyboard
The console keyboard provides 46 keys, including
special function keys. The character set consists
of the upper-case alphabetics, the numerics, and
19 punctuation and special symbols. In addition,
special codes are generated for three keys which
are normally used to indicate backspace, end of
field, and message deletion.
Information entered from the keyboard is stored in
core storage, one character per word, in the IBM
Card Code; see Section 418:141. The processor
must issue a command to the keyboard prior t.o the
entry of each character of data. The initial command can be issued in response to an interrupt
generated by the operator at the keyboard or by
coding in the object program. An interrupt is generated when a character key is depressed, indicating
that a character of data is ready to be transferred
to the processor. The operator-initiated interrupt
is caused by depressing a particular key; this
causes a branch to a specified location which must
contain a user-CQded routine for handling the interrupt conditi~n.
. 123 Programming
Assembly-language programming of the console
keyboard/printer operations is facilitated by the
use of two standard IBM subroutines. One subroutine allows printer output only; the second subroutine allows keyboard input and automatic printing of the input data on the console printer. The
second subroutine also provides backspacing, endof-message indication, and message deletion as
initiated by the special keys mentioned in Paragraph .122.
The called subroutine is automatically inserted in
the user's program by the 1130 Assembler. The
calling sequence specifies the name of the desired
subroutine and parameters indicating whether a
print, read-print, or device-busy check operation
is to be performed. Included in the parameter list
are core memory addresses of the input or output
data area and of the user's error routine. The
number of words that are to be transferred must
be loaded by the programmer into the first word of
the data area. Each 16-bit memory location to be
printed contains two characters in Console Printer
Code. Data can be converted into console code from
other codes by standardized conversion subroutines
provided by IBM; see Table I in Section 418:051,
Central Processor.
When an interrupt is generated as a result of a
console printer operation, the subroutine checks
for successful completion of the operation. If an
error is detected, the subroutine loads the accumu1ator register with status information and transfers
control to the user's error subroutine. After an
analysis of the error information, the user's routine
can either cause a halt or cause the operation to be
retried.
A different subroutine is used for keyboard data input and console printer output in FORTRAN-coded
source programs. This subroutine permits a keyboard data entry of up to 80 characters and/or a
print-out of up to 120 characters per subroutine
call. Data to be printed must be stored in unpacked
(one character per word) EBCDIC format; the subroutine converts this to Console Printer Code before
printing. Data entered from the keyboard is converted from IBM Card Code to the unpacked EBCDIC
format. All data entered from the keyboard is
printed on the console printer. The FORTRAN subroutines do not permit overlapped operation of the
keyboard and console printer.
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
418:111.100
&
SIAN""
AEDP
AIJERBAC~
,
-
IBM 1130
SIMULTANEOUS
OPERATIONS
REPORTS
SIMULTANEOUS OPERATIONS
.1
GENERAL
. 11
Input-Output Channels
Three input-output channels are used in IBM 1130 systems to transfer data between the
peripheral devices and the central processor:
• Direct Program Control Channel - a standard inclusion in all models of the
1131 Central Processor Unit.
• Disk Channel - a standard inclusion in the 1131 Model 2A and 2B Central
Processor Units.
•
Storage Access Channel - an optional channel available for all models of the
1131 Central Processor Unit.
The Direct Program Control Channel, with appropriate adapters, is used for
connection of all standard peripheral devices available with the 1130 except
the Disk File. See Section 418:031, System Configuration, for a list of the
peripheral devices available and for configuration rules. In general, multiple
devices can operate simultaneously over the Program Control Channel, subject to
program timing considerations.
The Disk Channel is used with the Disk File, an integral part of the 1131 Model 2A
and 2B Central Processor Units.
The Storage Access Channel permits the connection of a non-IBM peripheral device.
Magnetic tape units such as the IBM 2415 can also be connected to an 1130 system
via this channel on a special request (RPQ) basis. The Storage Access Channel
can accommodate peripheral devices operating at up to 270,000 words per second .
. 12
Input-Output Process
Data transfers between all standard peripheral devices (except the 1132 Printer)
and the central processor operate on a character or word basis. The output process
for the 1132 Printer is relatively complex: see Section 418:081. Except with the Disk
File, an interrupt is generated for each character of data transferred. Whole blocks
of data (one or more sectors) are transferred between the Disk File and the central
processor, and an interrupt is generated only at the completion of an operation. The
interrupt facilities provide the 1130 with capabilities, at the hardware level, for overlapped operations .
.2
DEMANDS ON THE PROCESSOR
There are three factors to be considered in evaluating the effective demands imposed
upon the processor by the operations of the peripheral devices:
• Interference - the time required to transfer data, during which the processor is
interlocked against any type of processing.
• Interrupt processing - the time required to process the interrupts associated
with each character or block of data, during which the processor cannot perform
any non-I/O functions.
• Code translation and radix conversion processing - the time required to convert
between the external data codes and the internal forms required for computation.
In general, only Dne core storage cycle (3.6 microseconds) is actually required to
transfer one character or word of data between the peripheral device and the central
processor. Frequently, this time is small in comparison with the time required
for the interrupt processing and code conversion.
IBM provides a wide range of subroutines to control I/O operations and to accomplish
code translation and radix conversion. The I/o control subroutines provided for
Assembler-language programming return control to the user's program after the
completion of interrupt processing, thus making possible a degree of overlapping of
internal processing with input-output operations. The subroutines provided with the
FORTRAN compiler do not allow any overlapping. This is a significant limitation
since most user programming will probably be done in FORTRAN. In most
applications then, the 1130 will appear as a sequential system capable of performing
only one operation at a time.
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
IBM 1130
418:111. 200
Typical times for interrupt processing are presented in Table I for most of the basic
I/O operations. A list of the code translation and radix conversion subroutines is
presented in Section 418:051, Central Processor.
T ABLE I: IBM 1130 I/O INTERRUPT SERVICING TIMES
Processor Time msec (1)
Per Unit (3)
Fixed (2)
Device and Operation
1442 Card Read Punch, Model 6 or 7 Read one card
Punch one card
14.9
0.76
0.04
0.19
1132 Printer Print one line (alphameric)
Print one line (numeric)
Skip 1, 2 or 3 lines
44.1
31. 8
0.42
3.0
1.4
0.21
Console Keyboard/Printer Print one line
0.23
0.73
0.43
0.81
0.4~
0.68
1134 Paper Tape Reader Read one block
1055 Paper Tape Punch Punch one block
1627 Plotter Plot one point
Disk File Read one sector
Write one sector
Seek
(1)
1. 12
-
1.5
1.8
1.1
-
These times represent the use of the most straightforward of the standard
IBM subroutines; they do not include the actual data transfer times (see
Paragraph.2). The more sophisticated subroutines, which provide
additional facilities, usually require more time.
(2) These times are required once each time the subroutine is entered (i. e. ,
for each card read, each line printed, each point plotted, etc.).
(3)
These times apply to each character of data transferred, when applicable.
/
5/66
A
AUERBACH
'"
418: 121. 100
A
STANDARD
EDP
IBM 1130
INSTRUCTION LIST
RfP!lRTS
AUERBACH
~
INSTRUCTION LIST
Mnemonic
OP Code
(Hexadecimal Representation) 1
Instruction
COO
Load
Load
Load
Load
Load and Store
LD
LDD
LDX
lOS"
STO
STD
STX
STS
C80
600
200
DOD
D80
680
280
Accumulator
Double
Index
Status
Store Accumulator
Store Double
Store Index
Store Status
Arithmetic
A
800
AD
FOO
Add
Add Double
Subtract
Subtract Double
Multiply
Divide
And
Or
Exclusive Or
400
480
484
700
Bronch and Store Instruction Counter
Branch or Skip Conditionally
Branch Out or Skip Conditionally
Modify Index and Skip
100
108
18C
Shift Left Accumulator
Shift Left Accumulator and Q Reg.
Shift Left and Count Accumulator and Q Reg.
Shift Left and Count Accumulator
Shift Right Accumulator
Shift Right Accumulator and Q Reg.
Rotate Right
080
Execute I/O
100
300
No Operotion
Wait
880
900
980
S
SD
M
ADO
D
A80
EOO
AND
OR
E80
EaR
Branch
BSI
BSC
BOSC 2
MDX
Shift
SLA"
SLT*
SLC*
SLCA"
SRA"
SRT*
RTE*
IOC
104
180
188
Input/Output
XIO
Miscellaneous 3
NaP"
WAIT"
"Valid in short format only.
1. The hexadecimal representation of the machine operation code is derived from the instruction format in the manner shown below.
Bits 5, 6, 7, 10, and 11 are assumed to be zeros because they do not enter into the makeup of any operation codes.
2. Same as BSC with Bit 9 set to one.
3. An operand should nat be specified.
Hexadecimal Characters
o
1 2 3 4 5 6 7 8 9
10
11
12
13
14
15
OPcOde-----.J~tT
J
Format (F)
Index Tag bits (T)
IA bit, part of
displacement I or extension
ofOP code __________________________________________
~
Port of displacement,
or extension of OP code __________________________________--'
Condition "indicators,
or part of displacement __________________________________________~___.J
Reprinted from IBM 1130 Assembler Language, Form C26-5927-0, published by IBM.
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
418: 141. 100
A
AUERBAC~
•
DATA CODE TABLE
EBCDIC
Ref
No.
Binary
0123
4567
0
I
2
3
4
5'
6'
7*
8
9
10
II
12
13
14
15
0000
0000
0001
0010
0011
0100
0101
0110
0111
1000
1001
1010
16
17
18
19
20'
21'
22'
23
24
25
26
27
28
29
30
31
0001
32
33
34
35
36
37*
38'
39
40
41
42
43
44
45
DOlO
47
48
49
50
51
52
53'
54'
55
56
57
58
59
60
61
62
63
DOli
CD
Hex
Rows
12
1100
1101
1110
III I
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
0000
0001
DOlO
0011
0100
0101
0110
Olll
1000
1001
1010
lOll
1100
1101
1110
III I
10
II
12
13
14
15
16
17
18
19
IA
IB
IC
lD
IE
IF
12
0000
0001
0010
001l
0100
0101
0110
oIII
1000
1001
1010
lOll
1100
1101
IIiO
III I
20
21
22
23
24
25
26
27
28
29
2A
2B
2C
20
2E
2F
0000
0001
0010
DOli
0100
0101
0110
Dill
1000
1001
1010
lOll
1100
1101
lllO
III I
30
31
32
33
34
35
36
37
38
39
3A
3B
3C
3D
3E
3F
NOTES: Typewriter Output
Tobulote
@ Shift to block
1132
IBM Cord Code
Hex
00
01
02
03
04
05
06
07
08
09
OA
DB
DC
00
DE
OF
1011
46
II
12
0 9 8 7-1
0 9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
II
II
II
II
II
II
II
II
II
II
II
II
Il
Il
II
II
II
II
Graph i C5 and Control
Names
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
0 9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
I
I
2
3
4
5
6
7
I
2
3
4
5
6
7
I
I
2
3
4
5
6
7
I
2
3
4
5
6
7
I
I
2
3
4
5
6
7
I
2
3
4
5
6
7
I
I
2
3
4
5
6
7
I
2
3
4
5
6
7
B030
9010
8810
8410
8210
8110
8090
8050
8030
9030
8830
8430
8230
8130
80BO
8070
0030
5010
4810
4410
4210
4110
4090
4050
4030
5030
4830
4430
4230
4130
40BO
4070
7030
3010
2810
2410
2210
2110
2090
2050
2030
3030
2830
2430
2230
2130
20BO
2070
F030
1010
0810
0410
0210
0110
0090
0050
0030
1030
0830
0430
0230
0130
OOBO
0070
PF
HT
LC
DEL
EBCDIC
Subset Hex
U -Upper Case
L-Lower Case
Punch Off
Hariz.Tob
Lower Cose
Delete
RES
NL
BS
IDL
Restore
BYP
LF
EOB
PRE
Bypcss
PN
RS
UC
EOT
Punch On
Reoder Stop
Upper Cose
End of Trons.
New Line
Backspace
Idle
Line Feed
End of Block
Prefix
®
@4CarrierReturn
(1)
PTTC/8
Hex
Printer
Console
Printer
Hex
Not(!s
NUL
60~
41
(j)
05
81
II
@
Q)
6E 5
7F 5
$
4C
DD
5E 5
3D ~
3E 5
03
00$
OE 5
09
The Same in Either Cose
Shift to red
* Recognized by all Conversion subroutines.
Codes that are not asterisked are
recognized only by the SPEED subroutine.
Reprintedfrom IBM 1130 Subroutine Library, Form C26-5929-1, published by IBM.
5/66
ST .... RD
/AEDP
IBM 1130
DATA CODE TABLE
A
AUERBACH
®
(1)
REPIIRTS
418: 141. 101
DATA CODE TABLE
DATA CODE TABLE (Contd.)
EBCDIC
Ref
No.
64*
65
IBM Card Code
Binary
0123
4567
Hex
0100
0000
0001
0010
0011
0100
OJOI
0110
01 I I
1000
1001
1010
1011
1100
1101
1110
1111
40
41
42
43
44
45
46
47
48
49
4A
4B
4C
4D
4E
4F
0000
0001
0010
0011
0100
0101
0110
0111
1000
1001
1010
1011
1100
1101
1110
1111
50
51
52
53
54
55
56
57
58
59
5A
5B
5C
5D
5E
5F
0000
0001
0010
0011
0100
0101
0110
0111
1000
1001
1010
1011
1100
1101
1110
1111
60
61
62
63
64
65
66
67
68
69
6A
6B
6C
6D
6E
0000
0001
0010
0011
0100
0101
0110
0111
1000
1001
1010
1011
1100
1101
1110
1111
70
71
72
73
74
75
76
77
78
79
7A
66
67
68
69
70
71
72
73
74*
75*
76*
77*
78*
79*
80*
81
82
83
84
85
86
87
88
89
90*
91*
92*
93*
94*
95*
0101
96*
97*
98
99
100
101
102
103
104
105
106
107*
108*
109*
110*
111*
0110
112
113
114
115
116
117
118
119
120
121
122*
123*
124*
125*
126*
127*
0111
Rows
12
II
12
12
12
12
12
12
12
12
12
12
12
1.2
12
12
12
no punches
o 9
o 9
o 9
o 9
o 9
o 9
o 9
o 9 8
8
8
8
8
8
8
8
12
12
12
12
12
12
12
12
12
7D
7E
7F
9
9
9
9
9
9
9
9 8
8
8
8
8
8
8
8
I
2
3
4
5
6
7
I
2
3
4
5
6
7
I
2
3
4
5
6
7
I
2
3
4
5
6
7
II
II
11
11
II
II
II
II
12
12
12
12
12
12
12
12
12
12
I
2
3
4
5
6
7
0
0
0
0
0
0
0
0
0
9
9
9
9
9
9
9 8
8
0
0
0
0
0
8
8
8
8
8
3
4
5
6
7
9
9
9
9
9
9
9
9 8
8
8
8
8
8
8
8
I
2
3
4
5
6
7
1
II
II
J1
11
11
11
J1
11
11
11
Graphics and Control
Names
0 9 8 7-1
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
6F
7B
7C
Hex
0
0
0
0
0
0
0
0
0
1
2
3
4
5
6
7
0000
BOlO
A810
MIO
A210
AIIO
A090
A050
A030
9020
8820
8420
8220
8120
80AO
8060
8000
DOlO
C810
C410
C210
CliO
C090
C050
C030
5020
4820
4420
4220
4120
40AO
4060
4000
3000
6810
6410
6210
6110
6090
6050
6030
3020
COOO
2420
2220
2120
20AO
2060
EOOO
FOIO
E810
E410
E210
EIIO
E090
E050
E030
1020
0820
0420
0220
0120
OOAO
0060
1132
Printer
EBCDIC
Subset Hex
(space)
*
~
. (period)
<
4B
PTTC/8
Hex
U-Upper Case
L...Lower Case
Console
Printer
Hex
IO®
21
20 (U)
6B (L)
02 (U)
19 (U)
70 (U)
3B (U)
02
00
DE
FE
DA
C6
(
+
I (logical OR)
4D
4E
&
50
70 (L)
44
5B
5C
5D
5B (U)
5B (L)
08 (U)
IA(U)
13 (U)
6B (U)
42
40
D6
F6
D2
F2
60
61
40 (L)
31 (L)
84
BC
6B
3B (L)
15 (U)
40 (U)
07 (U)
31 (U)
80
06
BE
46
86
04 (U)
OB (L)
20 (L)
16 (U)
01 (U)
OB (U)
82
CO
04
E6
C2
E2
!
$
*
)
;
-,
(logical NOT)
- (dash)
/
, (comma)
%
_ (underscore)
>
?
iI
@
' (apostrophe)
=
"
7D
7E
* Any code other than those defined wi II be interpreted by PRNTI as a space.
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
418:141. 102
IBM 1130
DATA CODE TABLE (Contd.)
EBCDIC
Ref
No.
5/66
Binary
IBM Card Code
Hex
Rows
0123
4567
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
1000
0000
0001
0010
0011
0100
0101
0110
0111
1000
1001
1010
1011
1100
1101
1110
1111
80
81
82
83
84
85
86
87
88
89
8A
8B
8C
80
8E
8F
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
1001
0000
0001
0010
0011
0100
0101
0110
0111
1000
1001
1010
1011
1100
1101
1110
1111
90
91
92
93
94
95
96
97
98
99
9A
9B
9C
90
9E
9F
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
1010
0000
0001
0010
0011
0100
0101
0110
0111
1001
1010
1011
1100
1101
1110
1111
AO
Al
A2
A3
A4
A5
A6
A7
A8
A9
AA
AB
AC
AD
AE
AF
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
1011
0000
0001
0010
0011
0100
0101
0110
0111
1000
1001
1010
1011
1100
1101
1110
1111
BO
BI
B2
B3
B4
B5
B6
B7
B8
B9
BA
BB
BC
BD
BE
BF
iooo
12
11
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
9 8 7-1
8
8
8
8
8
8
8
8
2
3
4
5
6
7
8
1
1
2
3
4
5
6
7
8
9
a
a
a
a
a
a
a
Ii a
11 a
11 a
11 a
11 a
11 a
11 a
11 a
11 a
11 a
11 a
11 a
11 a
11 a
11 a
11 a
11 a
11 0
11 a
11 a
11 a
11 a
11 a
11 a
11
1
1
2
3
4
5
6
7
9
11
11
11
11
11
11
11
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
Hex
0
Graphics and Control
Names
8
8
8
8
8
8
2
3
4
5
6
7
8
1
1
2
3
4
5
6
7
8
9
8
8
8
8
8
8
2
3
4
5
6
7
8
1
1
2
3
4
5
6
7
8
9
8
8
8
8
8
8
2
3
4
5
6
7
B020
BOOO
A800
A400
A200
Al00
A080
A040
A020
A010
A820
A420
A220
A120
AOAO
A060
0020
0000
C800
C400
C200
C100
C080
C040
C020
COlO
C820
C420
C220
CI20
COAO
C060
7020
7000
6800
6400
6200
6100
6080
6040
6020
6010
6820
6420
6220
6120
60AO
6060
F020
FOOO
E800
E400
E200
El00
E080
E040
E020
E010
E820
E420
E220
E120
EOAO
E060
A
AUERBACH
'"
a
b
c
d
e
f
9
h
i
i
k
I
m
n
a
P
q
r
s
t
u
v
w
x
Y
z
1132
Printer
EBCDIC
Subset Hex
PTTC/8
Hex
U -Upper Case
L-Lower Case
Console
Printer
Hex
418:141. 103
DATA CODE TABLE
DATA CODE TABLE (Contd.)
EBCDIC
Ref
No.
Binary
IBM Card Cade
Raws
Hex
0123
4567
12
192
193*
194*
195*
196*
197*
198*
199*
200*
201*
202
203
204
205
206
207
1100
0000
0001
0010
0011
0100
0101
0110
0111
1000
1001
1010
1011
1100
1101
1110
1111
CO
Cl
C2
C3
C4
C5
C6
C7
C8
C9
CA
CB
CC
CD
CE
CF
208
209*
210"
211*
212*
213*
214*
215"
216*
217*
218
219
220
221
222
223
1101
0000
0001
0010
0011
0100
0101
0110
0111
1000
1001
1010
1011
1100
1101
1110
1111
DO
01
D2
D3
D4
05
D6
D7
D8
D9
DA
DB
DC
DD
DE
DF
224
225
226*
227*
228*
229*
230*
231*
232*
233*
234
235
236
237
238
239
1110
0000
0001
0010
0011
0100
0101
0110
0111
1000
1001
1010
1011
1100
1101
1110
1111
EO
El
E2
E3
E4
E5
E6
E7
E8
E9
EA
EB
EC
ED
EE
EF
240*
241*
242*
243*
244*
245*
246*
247*
248*
249*
250
251
252
253
254
255
1111
0000
0001
0010
0011
0100
0101
0110
0111
1000
1001
1010
1011
1100
1101
1110
1111
FO
Fl
F2
F3
F4
F5
F6
F7
F8
F9
FA
FB
FC
FD
FE
FF
11
0 9 8 7-1
0
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
Hex
1
2
3
4
5
6
7
8
0
0
0
0
0
0
11
11
11
11
11
11
11
11
11
11
11
11
11
8
8
8
8
8
8
2
3
4
5
6
7
1
2
3
4
5
6
7
8
9
9
9
9
9
9
9
11
11
11
11
11
11
11
11
9
9
9
9
9
9
9
0
11
11
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
8
8
8
8
8
8
2
3
4
5
6
7
8
2
1
2
3
4
5
6
7
9
8
9
9
9
9
9
9
9
8
8
8
8
8
8
2
3
4
5
6
7
0
1
2
3
4
5
6
7
8
12
12
12
12
12
12
11
11
11
11
11
11
Graphics and Control
Names
0
0
0
0
0
0
9
9
9
9
9
9
9
8
8
8
8
8
8
2
3
4
5
6
7
AOOO
9000
8800
8400
8200
8100
8080
8040
8020
8010
A830
A430
A230
A130
AOBO
A070
(+ zero)
6000
5000
4800
4400
4200
4100
4080
4040
4020
4010
C830
C430
C230
C130
COBO
C070
(- zero)
2820
7010
2800
2400
2200
2100
2080
2040
2020
2010
6830
6430
6230
6130
60BO
6070
2000
1000
0800
0400
0200
0100
0080
0040
0020
0010
E830
E430
E230
EI30
EOBO
E070
1132
Printer
EBCDIC
Subset Hex
PTTC/8
Hex
U-Uppcr Case
L-Lower Case
Consale
Printer
Hex
Cl
C2
C3
C4
C5
C6
C7
C8
C9
61 (U)
62 (U)
73 (U)
64 (U)
75 (U)
76 (U)
67 (U)
68 (U)
79 (U)
3C or:lE
18 or lA
lC or IE
30 or 32
34 or 36
10 or 12
14 or 16
24 or 26
20 or 22
P
Q
R
Dl
D2
D3
D4
D5
D6
D7
D8
D9
51
52
43
54
45
46
57
58
49
7C or 7 E
58 or 5A
5C or 5E
70 or 72
74 or 76
50 or 52
54 or 56
64 or 66
60 or 62
S
T
U
V
W
X
Y
Z
E2
E3
E4
E5
E6
E7
E8
E9
32 (U)
23 (U)
34 (U)
25 (U)
26 (U)
37 (U)
38 (U)
29 (U)
98 or 9A
9C or 9E
BO or B2
B4 or B6
90 or 92
94 or 96
A4 or A6
AO or A2
0
I
2
3
4
5
6
7
8
9
FO
Fl
F2
F3
F4
F5
F6
F7
F8
F9
lA (l)
01 (l)
02 (l)
13 (l)
04 (l)
15 (l)
16 (L)
07 (l)
08 (l)
19 (l)
C4
FC
D8
DC
FO
F4
DO
D4
E4
EO
A
B
C
D
E
F
G
H
I
J
K
l
M
N
0
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
(U)
(U)
(U)
(U)
(U)
(U)
(U)
(U)
(U)
5/66
418:141. 104
IBM 1130
IBM 1130 DATA CODE UTILIZATION (1)
Peripheral
Device
Code (2)
Input
1134 Paper Tape
Reader (3)
1055 Paper Tape
Punch (3)
.
1442 Card
Read Punch (3)
1132 Printer
Console Printer/
Keyboard
Output
PTTC/8
Card code
Card code
-
Characters
per Word
1
PTTC/8
1
Card code
1
EBCDIC
subset
2
Console
printer
1
(1)
See Table I in Section 418:051, Central Processor, for a
listing of the standard subroutines provided by IBM for
code translation.
(2)
This code indicates the form in which the data appears
in core storage immediately after an input operation or
before an output operation.
(3)
Since a binary image is transferred between these devices
and core storage, different meanings can be assigned to
each code if appropriate programming considerations are
made.
./
5/66
A
AUERBACH
'"
r--.
1.
A
AUERBAC~
•
418:151. 100
SIlK''''
EDlP
IBM 1130
PROBLEM ORIENTED
FACILITIES
REPORTS
PROBlEIVI ORIENTED FACiliTIES
.1
UTILITY ROUTINES
· 11
Simulators of
Other Computers: .. none.
· 12
Simulation by
Other Computers: .. none.
.13
Data Sorting and
Merging: . . . . . . . . no facilities.
.14
Report Writing:
· 15
Data Transcription
subroutines, data code translation subroutines,
radix conversion subroutines, arithmetic and functional subroutines, and selective dump subroutines.
In addition, subroutines for disc-oriented systems
permit overlay control of both Assembler and
FORTRAN object programs.
The Assembler and FORTRAN I/O control subroutines
are discllssed in the report sections on the individual
peripheral devices; see Sections 418:042 and 418:071
through 418:102.
. no facilities.
The data code translation and radix conversion subroutines are discussed in Section 418:051, Central
Processor.
1130 Utility Routines
Reference: . . . . . . . . IBM 1130 Card/Paper Tape
Programming System
Operators Guide, Form
C26-3629.
Date available: . . . . . March 1966.
Description:
Three straightforward data transcription routines
are included in the IBM 1130 Utility Routine
package. The Input/Output Routine can accept data
from the'1442 Card Read Punch or the 1134 Paper
Tape Reader and list it on the Console Typewriter
or 1132' Printer; this routine can also convert data
between punched card format and punched paper
tape format and can perform a listing on either the
Console Typewriter or Printer at the same time.
The Keyboard Routine allows an operator to prepare
source documents on punched cards or punched
paper tape from data entered on the Console Keyboard. The Card Reproducing Routine permits an
operator to reproduce a deck of cards; the columns
to be reproduced are selected by appropriate
punches in the desired columns of a header card.
The exact facilities available at a particular, installation depend on its peripheral equipment configuration. These routines utilize the standard inputoutput and data-code conversion subroutines contained in the 1130 Subroutine Library.
. 16
Selective dump subroutines are provided to allow
selected areas of core storage to be printed on the
Console Typewriter or the 1132 Printer in either
hexadecimal (four 4-bit characters per word) or
decimal (five 3-bit characters plus sign per word)
format.
.18
Arithmetic and Functional Subroutines
A wide range of subroutines is provided for use in
FORTRAN or Assembler coded programs. Table I
illustrates the range and execution times of these
subroutines. Standard -precision floating-point
format provides a fraction of approximately 6 significant digits; extended -precision floating-point format
provides a fraction of approximately 9 significant
digits.
.19
Applications Programs
Two categories of applications programs are available
for use on the IBM 1130 system:
o
Programs designed for specific applications,
such as civil engineering, and
o
Routines providing mathematical techniques
which can be used in many diverse applications •
File Maintenance
There are no file maintenance facilities for cardor paper tape-oriented 1130 systems. In discoriented systems utilizing the IBM 1130 Monitor
System, the Disk Utility Program provides facilities
for inserting or deleting subroutines and object
programs in the Disk File. (See also Section
418:191, IBM 1130 Monitor System).
· 17
The arithmetic and functional subroutines are discussed in Paragraph . 18 of this report section.
IBM 1130 Subroutine Library
Reference: .•••.... IBM 1130 Subroutine
Library, Form C26-5929.
Availability: . • . . . . . March 1966.
Description:
An extensive array of subroutines is provided by
IBM to aid users of an IBM 1130. The same general
facilities are available for card, paper tape, or
disc-oriented systems. The principal types of subroutines included in the Library are: assembler
I/O control subroutines, FORTRAN I/O control
In general, the use of these programs requires an
1130 system configuration with an 1131 Model 2B
Central Processor Unit, a 1442 Card Read-Punch,
and, in some cases, a 1627 Plotter.
Programs are provided for applications in the following specific fields:
o
Petroleum exploration and engineering: These
programs can be used to analyze seismic, magnetic, and gravity data, and to solve drilling,
production and reservoir problems.
o
Type composition: This program aids a type
compositor in the transcription of textual
material to a form required by line -casting
machines for setting type. The use of this
program requires an 1131 Central Processor
Model 2B and either an 1134 Paper Tape Reader
and 1055 Paper Tape Punch or a 1442 Card
Read Punch. Optional special features (RPQ)
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
IBM 1130
418: 151. 190
.19
Applications Programs (Contd.)
analYSiS, and fluid flow. These routines are
provided in addition to the standard library
functions of 1130 FORTRAN. The system requirements for the use of Mathpak vary depending on the size of the subroutines used.
are available to connect multiple 6-channel
paper tape units to an 1130 Central Processor.
•
Civil engineering: A simplified programming
system, Civil Engineering Coordinate Geometry
(COGO), is provided to allow civil engineers to
solve geometrical problems using civil engineering terminology.
The following routines are provided for general
scientific applications:
•
Mathpak: A set of 20 FORTRAN -coded subroutines which can be applied to problems in
electric field theory, elastiCity, harmonic
•
Statistical System: A set of four programs
providing the following statistical techniques:
step-wise linear regression, analysis of variance' polynomial curve fitting, and factor
analysis.
•
Numerical Surface Techniques and Contour Map
Plotting: A set of nine programs providing
techniques for describing and operating on geometric surfaces.
TABLE I: ARITHMETIC AND FUNCTIONAL SUBROUTINES
Symbolic Name
Execution Time, Microseconds
SUBROUTINE
Standard Precision
Extended Precision
Add/Subtract
Multiply
Divide
Load/Store F AC
*FADD/*FSUB
*FMPY
*FDIV
*FLO/*FSTO
*EADD/*ESUB
*EMPY
'EDIV
*ELD/*ESTO
460/560
560
766
180/180
440/490
790
2,060
160/170
Trigonometric Sine/Cosine
Trigonometric Arctangent
Square Root
Natural Logarithm
Exponential (eX)
Hyperbolic Tangent
FSINE/FCOSN
FATN, FATAN
FSQR, FSQRT
FLN, FALOG
FXPN, FEXP
FTNH/FTANH
ESINE/ECOSN
EATN, EATAN
ESQR, ESQRT
ELN, EALOG
EXPN, EEXP
ETNH/ETANH
3,000/3,400
5,200
4,500
5,100
2,000
4,300
5,400/5,900
8,900
10,400
8,000
4,400
8,100
Floating-Point Base to an Integer Exponent
Floating-Point Base to a Floating-Point Exponent
Floating-Point to Integer
Integer to Floating-Point
'FAXI
*FAXB
I FIX
FLOAT
NORM
FBTD/FDTB
*EAXI
*EAXB
IFIX
FLOAT
NORM
FBTD/FDTB
3,800
8,000
140
330
260
40,000/20,000
4,700
13,300
140
330
260
40,000/20,000
FARC
FARC
*FIXI
XSQR
XMDS
XMD
XDD
*FIXI
XSQR
XMD
XDD
465
550 (avg)
260
520
1,760
*FSBR
*FDVR
SNR
FAVL, FABS
lABS
*ESBR
*EDVR
SNR
EAVL, EABS
lABS
650
1,090
80
50
100
740
2,520
80
60
100
FGETP
EGETP
330
320
Standard Precision
Extended Precision
Floating-Point
Normalize
Floating Binary to Decimal/Floating Decimal
to Binary
Floating-Point Arithmetic Range Check
60
60
Fixed-Point
Integer Base to an Integer Exponent
Fixed-Point Square Root
Fixed-Point Fractional Multiply (short)
Fixed-Point Double Word Multiply
Fixed-Point Double Word Divide
-
465
550 (avg)
-
-
-
Special Function
Floating-Point Reverse Subtract
Floating-Point Reverse Divide
Floating-Point Reverse Sign
Floating-Point Absolute Value
Integer Absolute Value
Miscellaneous
Get Parameters
* By adding an X to those names prefixed with an asterisk, the user can cause the contents of Index Register 1 to be added to the address of the argument
specified in the subroutine calling sequence to form the effective argument address. For example, FADDX would be the modified form of FADD.
5/66
A
.,
AUERBACH
418:151. 100
.&..
A
ElD>IP'
AUERBAC~
STANDA"
-
IBM 1130
PROBLEM ORIENTED
FACILITIES
REPORlS
~
PROBLEM ORIENTED FACILITIES
.1
UTILITY ROUTINES
.11
Simulators of
Other Computers: .. none.
.12
Simulation by
Other ComQuters: .. none.
. 13
Data Sorting and
Merging: . . . . . . . . no facilities.
.14
Report Writing:
. 15
Data Transcription
subroutines, data code translation subroutines,
radix conversion subroutines, arithmetic and functional subroutines, and'selective dump subroutines.
In addition, subroutines for disc-oriented systems
permit overlay control of both Assembler and
FORTRAN object programs.
The Assembler and FORTRAN I/O control subroutines
are discussed in the report sections on the individual
peripheral devices; see Sections 418:042 and 418:071
through 418:102 .
. no facilities.
The d.ata code translation and radix conversion subroutines are discussed in Section 418:051, Central
Processor.
1130 Utility Routines
Reference: . . . . . . . . IBM 1130 Card/Paper Tape
Programming System
Operators Guide, Form
C26-3629.
Date available: . . . . . Mal'ch 1966.
Description:
Three straightforward data transcription routines
are included in the IBM 1130 utility Routine
package. The Input/Output Routine can accept data
from the'1442 Card Read Punch or the 1134 Paper
Tape Reader and list it on the Console Typewriter
or 1132' Printer; this routine can also convert data
between punched card format and punched paper
tape format and can perform a listing on either the
Console Typewriter or Printer at the same time.
The Keyboard Routine allows an operator to prepare
source documents on punched cards or punched
paper tape from data entered on the Console Keyboard. The Card Reproducing Routine permits an
operator to reproduce a deck of cards; the columns
to be reproduced are selected by appropriate
punches in the desired columns of a header card.
The exact facilities available at a particular, installation depend on its peripheral equipment configuration. These routines utilize the standard inputoutput and data-code conversion subroutines contained in the 1130 Subroutine Library.
. 16
Selective dump subroutines are provided to allow
selected areas of core storage to be printed on the
Console Typewriter or the 1132 Printer in either
hexadecimal (four 4-bit characters per word) or
decimal (five 3-bit characters plus sign per word)
format.
.18
Arithmetic and Functional Subroutines
A wide range of subroutines is provided for use in
FORTRAN or Assembler coded programs. Table I
illustrates the range and execution times of these
subroutines. Standard-precision floating-point
format provides a fraction of approximately 6 significant digits; extended .,.precision floating-point format
provides a fraction of approximately 9 significant
digits.
'
.19
Applications Programs
Two categories of applications programs are available
for use on the IBM 1130 system:
o
Programs designed for specific applications,
such as civil engineering, and
•
Routines providing mathematical techniques
which can be used in many diverse applications •
File Maintenance
There are no file maintenance facilities for cardor paper tape-oriented 1130 systems. In discoriented systems utilizing the IBM 1130 Monitor
System, the Disk Utility Program provides facilities
for inserting or deleting subroutines and object
programs in the Disk File. (See also Section
418:191, IBM 1130 Monitor System).
.17
The arithmetic and functional subroutines are discussed in Paragraph . 18 of this report section.
IBM 1130 Subroutine Library
Reference: . . . • . . . . IBM 1130 Subroutine
Library, Form C26-5929.
Availability: . . . . . . . March 1966.
Description:
An extensive array of subroutines is provided by
IBM to aid users of an IBM 1130. The same general
facilities are available for card, paper tape, or
disc-oriented systems. The principal types of subroutines included in the Library are: assembler
I/O control subroutines, FORTRAN I/O control
In general, the use of these programs requires an
1130 system configuration with an 1131 Model 2B
Central Processor Unit, a 1442 Card Read-Punch,
and, in some cases, a 1627 Plotter.
Programs are provided for applications in the following specific fields:
o
Petroleum exploration and engineering: These
programs can be used to analyze seismic, magnetic, and gravity data, and to solve drilling,
production and reservoir problems.
o
Type composition: This program aids a type
compositor in the transcription of textual
material to a form required by line-casting
machines for setting type. The use of this
program requires an 1131 Central Processor
Model 2B and either an 1134 Paper Tape Reader
and 1055 Paper Tape Punch or a 1442 Card
Read Punch. Optional special features (RPQ)
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
IBM 1130
418:151. 190
Applications Programs (Contd.)
.19
analYSiS, and fluid flow. These routines are
provided in addition to the standard library
functions of 1130 FORTRAN. The system requirements for the use of Mathpak vary depending on the size of the subroutines used.
are available to connect multiple 6-channel
paper tape units to an 1130 Central Processor.
•
Civil engineering: A simplified programming
system, Civil Engineering Coordinate Geometry
(COGO), is provided to allow civil engineers to
solve geometrical problems using civil engineering terminology.
The following routines are provided for general
scientific applications:
•
•
•
Mathpak: A set of 20 FORTRAN-coded subroutines which can be applied to problems in
electric field theory, elastiCity, harmonic
Statistical System: A set of four programs
providing the following statistical techniques:
step-wise linear regression, analysis of variance, polynomial curve fitting, and factor
analysis.
Numerical Surface Techniques and Contour Map
Plotting: A set of nine programs providing
techniques for describing and operating on geometric surfaces.
TABLE I: ARITHMETIC AND FUNCTIONAL SUBROUTINES
Symbolic Name
Execution Time, Microseconds
SUBROUTINE
Standard Precision
Extended Precision
Standard Precision
Extended Precision
Add/Subtract
Multiply
Divide
Load/Store F AC
*FADD/*FSUB
*FMPY
'FDIV
'FLD/*FSTO
*EADD/*ESUB
*EMPY
*EDIV
*ELD/*ESTO
460/560
560
766
180/180
440/490
790
2,060
160/170
Trigonometric Sine/COSine
Trigonometric Arctangent
Square Root
Natural Logarithm
Exponential (eX)
Hyperbolic Tangent
FSINE/FCOSN
FATN, FATAN
FSQR, FSQRT
FLN, FA LOG
FXPN, FEXP
FTNH/FTANH
ESINE/ECOSN
EATN, EATAN
ESQR, ESQRT
ELN, EALOG
EXPN, EEXP
ETNH/ETANH
3,000/3,400
5,200
4,500
5,100
2,000
4,300
5,400/5,900
8,900
10,400
8,000
4,400
8,100
Floating-Point Base to an Integer Exponent
Floating-Point Base to a Floating-Point Exponent
Floating-Point to Integer
Integer to Floating-Point
Normalize
Floating Binary to Decimal/Floating Decimal
to Binary
Floating-Point Arithmetic Range Check
*FAXI
*FAXB
IFIX
FLOAT
NORM
FBTD/FDTB
'EAXI
*EAXB
IFIX
FLOAT
NORM
FBTD/FDTB
3,800
8,000
140
330
260
40,000/20,000
4,700
13,300
140
330
260
40,000/20,000
FARC
FARC
'FIXI
XSQR
XMDS
XMD
XDD
'FIXI
XSQR
XDD
465
550 (avg)
260
520
1,760
'FSBR
'FDVR
SNR
FAVL, FABS
lABS
'ESBR
*EDVR
SNR
EAVL, EABS
lABS
650
1,090
80
50
100
740
2,520
80
60
100
FGETP
EGETP
330
320
Floating-Point
60
60
Fixed-Point
Integer Base to an Integer Exponent
Fixed-Point Square Root
Fixed-Point Fractional Multiply (short)
Fixed-Point Double Word Multiply
Fixed-Point Double Word Divide
XMD
465
550 (avg)
-
-
Special Function
Floating-Point Reverse Suhtract
Floating-Point Reverse Divide
Floating-Point Reverse Sign
Floating-Point Absolute Value
Integer Absolute Value
Miscellaneous
Get Parameters
• By adding an X to those names prefixed with an asterisk, the user can cause the contents of Index Register 1 to be added to the address of the argnment
specified in the subroutine calling sequence to form the effective argument address. For example, FADDX would be the modified form of FADD.
5/66
A.
AUERBACH
- 1.
418:161. 100
IA
AUERBAC~
STANDARD
IBM 1'130
PROCESS ORIENTED
LANGUAGE
FORTRAN
EDP
REPORTS
~
PROCESS ORIENTED LANGUAGE: IBM 1130 FORTRAN
.1
GENERAL
.11
Identity: . . . . . . . . . IBM 1130 FORTRAN.
. 12
Origin:.......... IBM Corporation.
.1S
Reference: . . . . . . . IBM Systems Reference
Library Form C26-59SS-S.
. 14
Description
IBM 1130 FORTRAN is an implementation of the
Basic FORTRAN Language as proposed by the
XS. 4. S FORTRAN Group of the American Standards Association, and as published in the Communications of the ACM, October 1964. Extensions to the Basic FORTRAN language include
machine indicator tests (by subroutines), mixedmode expressions (real and integer forms), disc
handling statements (Monitor System only), and
automatic input-output conversions. In the l1S0
FORTRAN language the normally ambiguous statement A**B**C is acceptable and is evaluated as .
A**(B**C).
In essence, the IBM l1S0 FORTRAN language is a
restricted version of the FORTRAN IV language as
implemented for the IBM 7090/7094. A detailed
description of the IBM 7090/7094 FORTRAN IV
language is presented in Section 408:162. The
l1S0 FORTRAN language is also a subset of the
IBM System/360 Basic Programming Support
FORTRAN IV language (see Section 420:163) with
the exception of the statements used in conjunction
with the 1130 disc drive.
The principal restrictions of the IBM 1130
FORTRAN language include the absence of double
precision, complex, and logical capabilities. A
source program written in 1130 FORTRAN language can be compiled using the IBM System/360
BPS FORTRAN IV compiler if the reserved word
conventions of the System/S60 compiler are observed and the disc handling statements are not
used.
.141 Restrictions Relative to IBM 7090/7094
FORTRAN IV
(1)
No facilities are provided for COMPLEX,
DOUBLE PRECISION, or LOGICAL arithmetic
operations.
(2)
The ASSIGN statement is not provided.
(3)
The maximum size of integer constants is
215 -1 in 1130 FORTRAN as compared to
235 -1 in 7090/7094 FORTRAN.
.142 Extensions Relative to IBM 7090/7094 FORTRANIV
(1) The range of REAL numbers in 1130 FORTRAN
is approximately 10- 3 9 to 1038 as compared
to 10 135 in the 7090/7094 FORTRAN.
(2)
The following statements which apply to the
1130 Monitor System (see Section 418:191):
CALL EXIT
DEFINE FILE
Disk READ, WRITE and FIND
CALL LINK
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
&
418:171. 100
-
SImARD
AEDP
AUERBAC~
IBM 1130
MACHINE ORIENTED
LANGUAGE
1130 ASSEMBLER
•
REPORTS
MACHINE ORIENTED LANGUAGE: 1130 ASSEMBLER
.1
GENERAL
• 11
Identity: .•..•..•.. 1130 Assembler Language .
• 12
Origin:
.1.3
Reference: . . • . . . . . IBM 1130 Assembler
Language, Form C265927-0, published byIBM.
.14
Description
Operand and Remarks: specifies unmodified operand address or literal.
A blank after this information indicates to the assembler that the remainder of the line is a comment.
Identification: .••.•. for program identification
and statement sequence
numbers.
. . . . . • . . . . IBM Corporation •
The IBM 1130 Assembler language is a straightforward, one-to-one symbolic language which provides:
•
A symbolic representation of the entire 1130
instruction repertoire; and
•
A set of pseudo-operations for reserving
memory areas, changing the program origin,
and calling subroutines from the subroutine
library.
.23
Corrections: •.•..•. no special provisions;
i. e., manual correction
and reassembly are
required.
.24
Special Conventions
.241 Compound addresses: •. symbols maybe modified
by + or -.
.242 Multi -addresses: .•.. none.
.243 Literals: . . . . . . . . . * causes address of that
instruction to be inserted
in address field; load index
register instruction uses
literal in address field.
The Assembler language can be used to generate
subroutines for FORTRAN programs and can call
FORTRAN subroutines, input-output subroutines,
and utility routines. The Assembler program
translator is described in Section 418:181.
.2
LANGUAGE FORMAT
. 21
Diagram: .•..••... refer to coding form below .
.3
LABELS
Legend
.31
General
Label: . . . . • . . . . . . sym bol(s) identifying the
address of an instruction
or data item. An asterisk
in column 21 of the Label
Field indicates the line is
a comment.
Operation: ....••.. instruction code.
F: . . . • • • • . . . . . . . format specification indicating either short (1word) or long (2-word)
instruction.
T: ..•••••••...•• tag specification indicating
address modification by
index register 1, 2, or 3
or no index modification.
.311 Maximum number
of labels: .•.•..•. maximum of 550 labels using
4,096 core memory locations; maximum of 1,915
labels using 8,192 core
memory locations.
.312 Common label
formation rule: .•.. yes.
.313 Reserved labels: . . . . none •
. 314 Other restrictions: .. first symbol must be alphabetic; maximum of five
characters.
.315 Designators: .••.•• none •
. 316 Synonyms permitted: . EQU pseudo-instruction
equates two symbols.
.22
IBM
IBM 1130 Assembler
Ceding Perm
P'ogrofl"l _ _ _ _ _ _ _ _ __
Dote _ _ _ __
Progro,,,,ned by _ _ _ _ _ _ _ _ _ _ _ _ _ _ __
Page No. _ _ of _ _
label
Operotion
F T
Operand~
8. Remorb
!
5/66
A
I
,
Identification
!
!
,
I
I
!
I
,
(Contd.)
AUERBACH
'"
418: 171. 320
MACHINE ORIENTED LANGUAGE: 1130 ASSEMBLER
.32
Universal Labels
· 321 Labels for procedures Existence: . . . • . . . mandatory if referenced by
other sections of the
program.
Formation rule First character: ... alphabetic.
Others: . . . . . . . . alphanumeric.
Number of characters: . . . . . . . 1 to 5.
· 322 Labels for library
routines: . . . . . . . . same as procedures.
. 323 Labels for constants: . same as procedures.
. 324 Labels for files: . . . . same as procedures.
• 325 Labels for records: .• same as procedures.
.326 Labels for variables: . same as procedures.
. 33
Local Labels: • . . . . . no provision for local labels.
.4
DATA
.41
Constants
.411 Maximum size constants Integer Binary: . . • • . . . . 5 decimal digits (65,536
maximum) or 4 hexadecimal digits.
Alphanumeric (8
bits per character): 1 character, or up to 35
characters.
Fixed numeric: ..•. none.
Floating numeric (binary) Standard precision:. 8-decimal-digit fraction
plus sign (16,777,215
maximum); 3-decimaldigit exponent plus sign
(128 maximum).
Extended precision: 10 -decimal-digit fraction
plus sign (4.295 x 10 9
maximum); 3-decimaldigit exponent plus sign
(255 maximum).
.412 Maximum size literals Binary integer: . . . . 5 decimal digits (65,536
maximum); used only for
incrementing or loading
an index register.
. 42
Working Areas
.421 Data layout: . . . • . . . specified in program.
.422 Data type: . . . . • . . . tabulated in program.
.423 Redefinition: .•••..• yes.
.43
Input-Output Areas
.431 Data layout: . . • . . . . . specified by subroutine
parameters.
. 432 Data type: . . • . . • . . . specified by subroutine
used •
. 433 Copy layout: ••••... no.
.5
PROCEDURES
.51
Direct Operation Codes
· 511 Mnemonic Existence: .•.••.. mandatory.
Number: . . . . • . . . 29.
Example: . . . . . . . . A = Add.
· 52
Macro-Codes: .•..•. none.
.53
Interludes: •..•.•.. none.
• 54
Translator Control
.541 Method of control Allocation counter: ..
Label adjustment: ..
Annotation: •.•...
· 542 Allocation counter Set to absolute: .•..
Set to label: . . . • . .
Step forward: . . • . .
Step backward: . . . .
Reserve area: .•.•.
· 543 Label adjustment Set labels equal: ..•
Set absolute value: ..
Clear label table: ..•
· 544 Annotation Comment phrase: ...
pseudo-operation.
pseudo-operation.
pseudo-operation.
ORG.
asterisk in address field.
ORG .
ORG .
BSS, BES .
EQU.
EQU .
none .
notes following space after
address field, or separate
line with asterisk in
column 21.
.6
SPECIAL ROUTINES AVAILABLE
.61
Special Arithmetic
.611 Facilities: . . . . . . . . . double-word fixed-point
arithmetic; standard and
extended precision floating-point add, subtract,
multiply, divide; conversion between fixed and
floating point.
.612 Method of call: . . . . . . CALL and DC pseudoinstructions.
.62 Special Functions
.621 Facilities: . . . . • . . . standard and extended
precision floating
point square root, sine,
cosine and arctangent
functions.
· 622 Method of call: •.••• LIEF and DC pseudoinstructions.
.63
Overlay Control: . . . . no special provisions .
.64
Data Editing
.641 Radix conversion: ... by subroutine; see Table I,
Section 418:051, for a
listing •
. 642 Code translation: .•.. by subroutine; see Table I,
Section 418:051, for a
listing.
· 643 Format control: ..••. none .
.65
Input-Output Control: . handled by subroutines.
.66
Sorting: ...••..•.• no facilities .
• 67 Diagnostics
.671 Dumps: .•..••.••. no direct facilities; core
memory dump program
requires separate load
operation; core address
limits specified via console entry switches.
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
IBM 1130
418:171.672
Description
Code
ORG: ••••...•.• sets memory location
assignment counter to a
specified value.
END: ••••...••• indicates last statement
of source program.
DC: .•••••••.•. generates a data constant
in memory.
DEC: ••.•••.••. specifies in decimal notation a value to be expressed in binary when placed
in memory.
XFLC: •.•..••.. specifies an extended precision floating-point
constant.
EBC: ••.....•.• used to generate two 8-bit
characters (one I6-bit
word) in Extended BCD
Interchange code.
BSS: .•••••.... reserves a specified number
of core memory locations
starting with an evennumbered core location.
BES: ..•...•.•• reserves a specified number of core memory locations starting with an oddnumbered core location.
EQU: . . . • . • . . . . equates two different symbolic addresses.
ENT: .
• . • • . defines an entry point
(label) in a user-written
subroutine.
CALL:
..••... generates linkages to load
a specified subroutine at
program execution time.
LINK* ••...•.•• causes loading of a new
program and initiation of
its execution.
EXIT* ..••...• returns control to the 1130
Monitor System.
DSA* . • . . • • . . . . allows symbolic resequencing of an unknown absolute
address on disc.
.672 Tracers: .•.•.•... none.
.673 Snapshots: •••••.•. none.
.7
liBRARY FACIliTIES
.71
Identity: •••...•.•. IBM Subroutine Library
(card- or .paper tapeoriented system).
disc library (discoriented system).
.72
Kinds of Libraries: .. expandable master.
.73
Storage Form:
.74
Varieties of Contents:. standard IBM subroutines;
user-coded subroutines;
user-coded programs
(with Monitor System
only) .
.75
Mechanism
...•. punched cards, paper tape,
or disc.
.751 Insertion of new item: • physically place new item
in library deck; or separate
run using Disc Utility Program (Monitor System only).
.752 Language of new item:. machine language.
.753 Method of call: .•..• liBF or CALL and DC
pseudo-instructions.
.76
Insertion in Program
• 761 Open routines exist: ..
. 762 Closed routines exist: .
.763 Open-closed
is optional: .••.•••
.764 Closed routines
appear once: . . . . • .
no .
yes •
no.
yes.
.8
MACRO AND PSEUDO TABLES
.81
Macros: .••..••..• none.
.82
Pseudos
Code
Description
ABS: ••••.••••• indicates source program
is to be assembled in
absolute format.
5/66
* Available with Monitor System only.
A
AUERBACH
'"
.&.
418: 181. 100
"ANDARD
.
AEDP
AUERBAC~
-
IBM 1130
PROGRAM TRANSLATOR
1130 ASSEMBLER
REPORTS
PROGRAM TRANSLATOR: 1130 ASSEMBLER
.1
GENERAL
.22
· 11
Identity: •..••...•. IBM 1130 Assembler.
· 221 Input media: •.
· 12
Description
The IBM 1130 Assembler translates source coding
written in the 1130 Assembler language (see
Section 418: 171) into machine language.
Versions of the Assembler are available for 1130
systems using either card or punched tape inputoutput. Assembly in non-disc systems requires
two passes of the source coding. With the card
assembler, the generated coding is punched
directly into the first 19 columns of the source
cards during the second pass, producing a "list"
deck. A separate run using a compressor program is required to generate a loadable objectprogram deck. A similar procedure is followed
in punched tape-oriented systems; the output of
the second pass is a tape containing both source
coding and object coding in a format similar to
that for cards. In a disc-oriented system, the
Assembler is kept on the disc as part of the
Monitor System; the object coding is retained on
the diSC, and a list deck or tape can be produced
if desired. Disc-oriented assemblies require two
passes if the list deck or tape is to be produced.
The source coding is checked for errors in format
and syntax. A record is kept of the first and last
error detected in each statement, and two coded
symbols are punched into the output list deck or
tape. In card or paper tape systems, a listing of
errors and coding can be obtained when the compression operation is performed. In disc-oriented
systems, the errors can be listed as they are detected.
In a disc-oriented system utilizing the Monitor
System, automatic and programmed overlay procedures are available in a manner similar to
those of the FORTRAN compiler; see Paragraph
418:182.12.
• 13
paper tape or punched
cards.
· 222 Obligatory ordering: .• END card must be last
statement in source deck.
.223 Obligatory grouping: .. none.
.23
Maintainer: .•••.••. IBM Corporation.
· 15
Availability:
INPUT
OUTPUT
.31
Object Program
· 311 Language name: ..••• machine language.
.313 Output media: .•.••• punched cards, paper tape,
or disc (with Monitor
System only).
.32
• 33
· 21
Language
.211 Name: . • • • . . . . . . . IBM 1130 Assembler
Language; see Section
418:171.
• 212 Exemptions: ••.••.. pseudo-instructions for
run-to-run linkages are
available only with
Monitor System, which
requires a disc drive.
••••
Size Limitations
.3
Card and Paper Tape
Assembler: .•.•.•• March 1966.
Disc Assembler (with
the Monitor System):. April 1966.
•2
0
.231 Maximum number of
source statements: .. not specified.
.232 Maximum size of
source statements: .• 50 characters; comments
can be extended.
.233 Maximum number of
data items: .••••.• see next entry.
.234 Maximum number of
labels Card or paper tape system With 4, 096 words
of core: . . . . . . 550.
With 8, 192 words
of core: . . . . . . 1,915.
Disc-oriented
system (with
Monitor System) With 4,096 words
of core: ..••. approx. 3,550.
With 8, 192 words
of core: .•••. approx. 4, 900.
Originator: ..•••••. IBM Corporation.
.14
Form
Conventions: . . . . . . . I/O, data code conversion,
and arithmetic subroutines
can be used with any model
1131 Central Processor.
Monitor System can be used
only with disc system;
permits program and subroutine library to be maintained on disc.
Documentation
Subject
Provision
Source program: ••••
Object program: ••••
Storage map: " ••••••
Language errors: .•••
separate transcription run.
separate transcription run.
none •
optional listing.
.4
TRANSLATING PROCEDURE
.41
Phases and Passes
Pass 1: .•••••..•. generates symbol table for
use in pass 2 .
Pass 2: .•••..•••• generates object code and
punches it into cards or
paper tape. (Object code
is stored on disc when
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
418:181. 410
.41
• 44
.45
IBM 1130
Phases and Passes (Contd.)
Monitor System is used.)
Additional run is required
to produce loadable object
deck or tape.
Bulk Translating: .••• sequential assemblies can
be performed without
need to reload Assembler.
Program Diagnostics
.451 Tracers: ••••••.•. none.
• 452 Snapshots: .••••••• none.
• 453 Dumps: .• ; ••••••• none; separate dump
routines are available.
.46
• 513 Approximate expansion
of procedures: ••.•• one 1130 instruction (1 or
2 words) per source
statement.
· 52 Translation Time
.521 Normal translatingWith 300 card/min
reader: ••••••••• 66 to 77 statements per
minute.
With 400 card/min
reader: .••••••• 92 to 102 statements per
minute •
NOTE:
Translator Library
.461 Identity: •••••••••• IBM 1130 Subroutine
Library.
• 462 User restriction: •••• none •
. 463 FormStorage medium: .•• punched cards, paper tape,
or disc.
Organization: ••.•• subroutines must be assembled in relocatable
code; programs can be
assembled in either
absolute or relocatable
code.
.464 ContentsSubroutines: .••••• closed.
Data descriptions: •. none.
Programs: ••••••• machine-language
object coding.
.465 LibrarianshipInsertion: .••••••• yes (separate run required).
Amendment: ••.••• no.
Call procedure: . • . • subroutine linkages are
generated at assembly
time. Referenced subroutines are transferred
to memory at program
load time.
The above data is for a non-diSC, cardoriented system; these ratES also apply
to compression. No data was available
for other configurations.
.53
Optimizing Data: •••• none •
· 54
Object Program
Performance: ••••. essentially unaffected (i. e. ,
same as hand machinelanguage coding).
.6
COMPUTER CONFIGURATIONS
.61
Translating Computer
• 611 Minimum configuration: .•••.•••..• 1131 Central Processor
Model 1A (4,096 words of
core storage capacity)
and either a 1442 Card
Read Punch or an 1134
Paper Tape Reader and
a 1055 Paper Tape Punch.
.612 Larger configuration advantages: ••.• greater core memory
capacity allows larger
symbol table; a disc drive
allows use of the 1130
Monitor System, which
provides automatic runto-run supervision.
.5
TRANSLATOR PERFORMANCE
• 62
.51
Object Program Space
.621 Minimum configuration: ••••.•••••• any valid 1130 system.
.622 Usable extra facilities: .•••••••••• all.
• 511 Fixed overhead*
Name
I/O Control subroutines Card:
Paper tape:
Disc:
Printer:
Keyboard/Printer:
Plotter:
Radix conversion:
Code translation:
Space, words
Comment
.7
275.
290.
350 to 800.
400.
300.
235.
270.
1,586
includes all
subroutines
for code
translation.
ERRORS, CHECKS, AND ACTION
Error
Check or
Interlock
Action
Missing entries:
Unsequenced entries:
Duplicate names:
Improper format:
Incomplete entries:
Symbol table overflow:
Inconsistent program:
none
none
check
check
check
check
none
*
*
*
*
*
*
*
* In general, the Assembler ignores or makes an
assumption about the statement in error and
continues assembly. Up to two error codes per
statement are punched into the list deck or tape.
In card or paper tape oriented systems, a separate transcription run is required for a listing;
with the Monitor System, a listing can be made
during assembly.
* This information applies to all configurations;
detailed information about the additional overhead due to use of the Monitor System is not
available to date.
.512 Space required for each
input-output file: •.. as programmed.
5/66
Target Computer
A
AUERBACH
'"
418: 182.100
/&•
AUERBACH
STANDARD
EDP
IBM 1130
PROGRAM TRANSLATOR
1130 FORTRAN
REPORTS
PROGRAM TRANSLATOR: 1130 FORTRAN
.1
GENERAL
. 11
Identity: ..
.12
entered at object load time, that multiple subroutines share the same core storage area on an
overlay basis. The subroutines are stored on
disc storage and loaded into the core overlay
area when called at execution time; only one
such subroutine at a time can actually be in
core storage.
.. IBM 1130 FORTRAN
compiler.
Description
The IBM 1130 FORTRAN compiler converts
source coding written in 1130 FORTRAN language
(see Section 418:161) into a machine-language program. The FORTRAN compiler is available on
punched cards or paper tape, or it can be placed
on the disc in an 1130 using the 1130 Monitor
System (Section 418: 191). Input of a FORTRAN
source program is via punched cards or punched
paper tape. Translation is continuous from the
loading of the source program to the generation of
an object program.
The object programs produced by the punched
card or paper tape FORTRAN compiler can be
loaded by a relocatable loader; loading requires a manual operation.
The disc-oriented FORTRAN compiler does not
directly produce punched card or paper tape output; the object program resides in disc storage.
The loading of the source program, the compilation, and the subsequent transfer of control to the
program can be automatically handled by the
Monitor System (Section 418:191). Alternatively,
the object program can be transferred to the
library area of the disc or punched into cards or
paper tape.
Minimum requirements for use of the 1130
FORTRAN translator are a Model lA Central Processor (4, 096-word core memory) and either
paper tape input-output or a card read-punch.
The FORTRAN compiler for the 1130 Monitor System cannot be operated independently of the
Monitor system. The disc-oriented FORTRAN
language contains control statements to define
record size and to specify automatic transfer of
other programs for disc to core storage for subsequent execution.
A disc-oriented 1130 System using the Monitor
System provides automatic overlay of subroutines
used in FORTRAN programs that require more
core memory than is available at load time. Automatic overlay occurs in one of two ways, depending on whether the disc is used only for residence of the Monitor system or additionally used
as a data input-output device. If the disc is used
for Monitor residence only, the allocated core
memory overlay area contains either the arithmetic and functional subroutines or the FORTRAN
input-output subroutines. If the disc is used for
data storage 1 the allocated core memory overlay
area contains one of three subroutine groups:
arithmetic and functional, non-disc input-output,
or disc input-output. The user can alter the
grouping of standard subroutines and can include his
own subroutines in these groupings.
In addition to the automatic overlay facility, the
programmer can specify, via control records
Program execution time can be increased by as
much as eight times when one of the overlay
techniques is used. The exact increase in program
execution time depends on the relative placement .
of the overlay segments and data on the disc (which
affects disc arm motion time) and the number of
overlay operations required.
The input-output subroutines called by the 1130
FORTRAN compiler are not the same subroutines
called by the 1130 Assembler. The 1130 FORTRAN
subroutines do not permit simultaneous operation
of the various input-output devices, nor of an
input-output device and the central processor;
i. e., FORTRAN input or output operations are
performed in a sequential order, and each operation inhibits the execution of non-input-output
instructions in the central processor.
.13
Originator: . . . . . . . . IBM Corporation.
· 14
Maintainer: . . . . . . . . IBM Corporation.
.15
Availability
Card-oriented compiler: . . . . . . . . . . March 1966.
Paper-tape oriented
compiler: . . . . . . . March 1966.
Disc-oriented compiler (with Monitor
System): . . . . . . . . April 1966.
.2
INPUT
· 21
Language
.211 Name: . . . . • . . . . . . IBM 1130 FORTRAN;
see Section 418:161.
.212 Exemptions: . . . . . • . disc-handling statements
are usable only with the
Monitor System, which
requires a disc drive.
.22
Form
.221 Input media: . . . . . . . punched cards or punched
paper tape.
· 222 Obligatory ordering: .. order of Specification
statements must be:
Type, EXTERNAL,
DIMENSION, COMMON,
and EQUIVALENCE; an
array variable must be
preceded by either a
DIMENSION, COMMON,
or Type statement.
· 223 Obligatory
grouping:
. . . . none.
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
IBM 1130
418:.182.230
· 23 Size Limitations
• 231 Maximum number of
source statements: .. see Table I.
. 232 Maximum size of
source statements: .. 396 characters.
· 233 Maximum number of
data items: ....••. depends on storage available; floating-point data
items require two words
(standard precision) or
three words (extended
precision); integer data
items normally require
two words if standard
precision arithmetic is
used and three words if
extended precision arithmetic is used (integer
items require only one
word if this is specified at
compile time).
TABLE I: APPROXIMATE MAXIMUM NUMBER OF
SOURCE STATEMENTS*
Memory size, words
System Type
4,096
8,192
Card or Paper Tape
120
400
Disc No overlay
One overlay
Two overlays
65
105
140
365
395
435
* Based on inclusion of normal floating-point
arithmetic subroutines and all overhead; see
Paragraph.511. Inclusion of other subroutines
such as trigonometric or exponentiation functions
would reduce these numbers.
.3
OUTPUT
.31
Object Program
• 311 Language name: .••.. machine language .
.313 Output media: . . . • . . punched cards, paper tape,
or disc (Monitor System
only).
.32
Conventions: •..•... compiled programs can be
stored on disc to form
part of the user's library;
there are no program
library facilities for
punched card or paper tape
oriented systems.
. 33
Documentation
Subject
Provision
Source program: . . . .
Object program: ..••
Storage map: ••••••.
Language errors: ••.•
.4
optional printed listing.
none.
optional printed listing.
printed listing at end of
compilation; object code
not output if error
detected.
TRANSLATING PROCEDURE
. 41
Phases and Passes
· 42
Optional Mode
.421
.422
.423
.424
.425
Translate: . . . . . . . .
Translate and run: ...
Check only: . . . . . . . .
Patching: . . . . . . . . .
Updating: . . . . . . . . .
.43
Special Features: .•.. none.
· 44
Bulk Translating: . . . . Monit"r System permits
sequential translation of
multiple source programs with no operator
intervention required.
· 45
Program Diagnostic s
.451 Tracers: . . . . . . . . . linkages to trace routines
can be generated at
compile time. The trace
routines print the value
of each assigned variable
and/or the value of
each IF expression or
computed GO TO index.
Tracing can be controlled
by statements in the
source program and by a
console switch .
. 452 Snapshots: . . . . • . . . none.
.453 Dumps: . . . . . . . . . . no direct facilities; all
systems include provisions
for dumping core storage
and disc storage, if
included, by separate
routines.
.46
Translator Library
. 461 Identity: . . . . . • . . . . 1130 Subroutine Library .
. 462 User restriction: . . . . none .
.463 Form Storage medium: ... punched cards, paper
tape, or disc.
Organization: . . . . . relocatable binary.
.464 Contents Routines: . . . . • . . . closed.
Functions: . . . . . . . standard and user-coded.
Data descriptions: .. none.
.465 Librarianship Insertion: . . . . . . . . manual insertion into card
library deck; punched
onto paper tape library
file; automatic insertion
onto disk with Monitor
System.
Amendment: ... .. not possible.
. use of function name
Call procedure: .
in an arithmetic statement causes inclusion
of appropriate subroutine
at load time; subroutine
referenees by LIBF
or CA..... L statement
are included at load time.
.5
TRANSLATOR PERFORMANCE
.51
Object Program Space
. 511 Fixed overhead
The following table shows the fixed overhead
for FORTRAN object programs. Most of the
groupings are composed of several subroutines,
each of which can be called and loaded individually.
Phase 1: ..••...•.• read in and compress
source coding.
Phase 2: . . . • . . . . . . generate object coding.
5/66
yes.
only with Monitor System .
yes.
no special provisions.
no special provisions.
fA
AUERBACH
'"
(Contd. )
,/
PROGRAM TRANSLATOR;
1130 FORTRAN
418:182.512
Space,
words
Usual floating-point
subroutines:
Trigonometric subroutines:
Exponentiation
subroutines:
Non-disc I/O subroutines:
604
includes addition, subtraction,
multiplication, and division.
280
includes sine/cosine and arctangent.
422
includes integer and real exponentiation, and hyperbolic tangent.
1,550
Non-disc I/O format
interpreter:
850
Disc subroutines:
450
Supervisor
(Monitor System):
I/O buffer area:
Loader:
400.
121
40 (core
image);
450 (relocatable)
.512 Space required for
each input-output file: included in buffer area; see
table above.
.513 Approximate expansion of procedures: . 14 to 1; i. e., an average
of 14 words of 1130
coding per FORTRAN
source statement.
· 52
Comment
Translation Time
.521 Normal translating With listing: . . . . . . 1. 5S + C + 90 sec.
With no listiJ;lg: .•.. O. 45S + O. 2C + 87 sec.
Where S equals the number of non-comment
source cards and C equals the number of comment
source cards.
Note: The above data is for a card-oriented system including the 400 card/min. 1442
Model 7 Card Read Punch; no data was
available for compilation times under other
conditions.
· 53
Optimizing Data: . . . . none.
.54
Object Program
Performance: . . . . . no definitive data available
to date.
.6
COMPUTER CONFIGURATIONS
· 61
Translating Computer
· 611 Minimum configuration: . • . • . . . . . . . 1131 Model1A Central
Processor (4,096 words
of core memory) and
either a 1442 Card Read
includes subroutines for all peripheral devices except disc.
required if any I/O operations are
performed; includes decimal/
binary conversion facility.
includes buffer area and format
interpreter.
not required with Monitor system.
can be completely overlaid by data
storage.
Punch or an 1134 Paper
Tape Reader and a 1055
Paper Tape Punch.
· 612 Larger configuration
advantages: . . . . . • . greater core memory
capacity allows larger
symbol table; disc and
Monitor System minimize operator intervention and provide
automatic run-to-run
control.
· 62
Target Computer
· 621 Minimum configuration:. . . . . .
. . any valid 1130 System.
· 622 U sable extra
facilities: . . . . • . . . all.
.7
ERRORS, CHECKS, AND ACTION
Error
Check or
Interlock
Missing entries:
Unsequenced entries:
Duplicate names:
Improper format:
Incomplete entries:
Target computer
overflow:
Inconsistent program:
*
.8
Action
check
check
check
check
check
*
*
*
*
*
check
check
*
*
Error messages are printed upon conclusion
of compilation; if errors are detected, no
object program is produced.
ALTERNATIVE
TRANSLATORS: ... none.
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
418: 191. 100
A•
STAKDARD
EDlP'
IBM 1130
OPERATING ENVIRONMENT
MONITOR SYSTEM
REPORTS
AUERBACH
OPERATING ENVIRONMENT: 1130 MONITOR SYSTEM
.1
GENERAL
.11
Identity:
The Disk Utility Program inserts or deletes programs or subroutines in the disc library. As programs are deleted, the remaining programs are
"packed" to yield the maximum temporary work
storage on the disc. The Disk Utility Program
maintains an updated address table (map) on disc
storage of the IBM subroutines, users' program,
and data areas currently on the disc. A running
program can symbolically call any program on the
disc; i. e., it is not necessary to know the absolute
address of a program stored on the disc; the
Utility Program automatically finds and loads the
referenced routine. The Disc Utility Program can
unload any of the library programs or subroutines
onto punched cards, paper tape, or printer, and
can move a program from the library area to the
disc work area.
. IBM 1130 Monitor System.
Supervisor;
Disk Utility Program;
Subroutine Library;
1130 FORTRAN Compiler;
1130 Assembler.
In IBM 1130 systems without a disc drive, each
program or routine is loaded individually, and
operator intervention is required to initiate each
activity. With the addition of the disc drive contained in the Model 2A and 2B Central Processing
Units, facilities become available to reduce the
amount of operator attention required; IBM has
taken advantage of these facilities in the Monitor
System.
The 1130 Monitor System is a group of related
routines and programs including:
o
o
The IBM 1130 Subroutine Library includes:
o
A supervisor routine to interpret control information and direct the sequencing of programs;
The Disk Utility Program to insert or delete
programs and subroutines in the disc library;
Q
Arithmetic and functional subroutines,
FORTRAN and Assembler input-output
subroutines,
o
Data-code conversion subroutines,
o
User-written subroutines, and
Data transcription routines.
o
The IBM 1130 Subroutine Library;
o
o
The 1130 FORTRAN compiler; and
o
The 1130 Assembler.
The disc-oriented FORTRAN compiler provides the
same language facilities as the non-disc compiler
with the addition of disc handling statements. The
1130 FORTRAN language is described in Section
418:161, and the compiler in Section 418:182. The
disc-oriented 1130 Assembler provides the same
facilities as the non-disc Assembler; see Section
418:171 for a description of the 1130 Assembler
language and Section 418:181 for a description of
the translator.
Program control information, read from punched
cards or punched tape or entered via the console
keyboard directs the Monitor operations. Programs an'd data can be loaded from either punched
cards or punched tape. The Monitor System occupies 19 percent (97,300 words) of the disc capacity, not including the 1130 Subroutine Library.
The Assembler and FORTRAN compiler can be
deleted from extra disc cartridges, thereby decreasing the Monitor area on those discs to 10
percent of the total capacity (51,200 words).
The supervisor routine performs the following
functions:
o
Interprets control information which specifies
the sequence of program operations.
o
Loads the Assembler, FORTRAN compiler,
input-output subroutines, or the Disc Utility
Program from the disc.
I»
Loads data from punched cards, paper tape,
or disc.
o
Initiates execution of a program in core
storage.
o
Allows control information to be entered via
. the console typewriter.
o
Provides for an automatic program halt for
anticipated operator intervention at the conclusion of the execution of a program.
.13
Availability:
.14
Originator:.
. . . . . . April 1966.
. IBM Corporation.
.15
Maintainer:
. IBM Corporation.
.2
PROGRAM LOADING
.21
Source of Programs
.211 Programs from
on-line libraries: .
. from disc; programs can
be added from card reader
or paper tape reader.
.212 Independent programs: from cards or paper tape.
.213 Data: . . . . . . • .
. as incorporated in user's
program.
.214 Master routines:
. from disc.
.22
Library Subroutines: .. from disc.
.23
Loading Sequence: ... specified by operator via
control statements from
card reader or paper
tape reader.
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
IBM 1130
418: 191. 300
.3
.31
.512 Snapshots: ••.••.•. none .
HARDWARE
ALLOCATION: . . . . as incorporated in user's
program.
.52
Post Mortem: .•.••• core memory dump routine
can be loaded from disc,
card reader , or paper
tape reader; dump output
can be to ptinter, console
typewriter, or punched
cards. Disc dump can be
called by operator; dump
output can be to printer,
console typewriter, or
punched cards.
.6
OPERATOR
CONTROL: ..•...• as incorporated in user's
program; program stop
switch on console allows
manual intervention to
alter program progress
or to abort program.
.7
LOGGING: . . . . . . . . . as incorporated in user's
program.
Storage
.311 Sequencing of program
for movement
between levels: . . . . automatic segmentation of
certain standard and usercoded subroutines into two
or three groups if program cannot fit into core
at object load time; programmer can also specify
a separate subroutine
overlay; see also Paragraph 418:182. 12.
.312 Occupation of working
storage: . • . • . . . . . in the automatic overlay,
the particular subroutine
group is loaded into the
core overlay area when
one of the included subroutines is called; in the
programmer-specified
overlay, only one subroutine at a time can be
in the overlay area.
.32 Input-Output Units: .•. fixed assignment.
.4
RUNNING SUPERVISION
. 41
Simultaneous Working: incorporated in Assembler
program by translator; in
some cases, it depends on
the choice of the I/o subroutine used; in FORTRAN
programs, no processing
or I/O operations can be
overlapped.
.42
Multiprogramming: ... none.
.43
Multi-Sequencing: ... none.
.44
Errors, Checks, and
Action: ••.•.•...• the supervisor routine only
--checks for loading errors,
such as insufficient core
storage and input errors;
detected errors cause the
system to halt. Detection
of program errors such as
arithmetic overflow must
be programmed. Standard
I/O subroutines detect
many peripheral device
errors and cause a branch
to a user's error routine.
• 45
PERFORMANCE
.81
System Requirements
.811 Minimum configuration: . • • . . . . • . . . 1131 Central Processor
Model 2A (includes disc);
card read punch or paper
tape reader and punch •
.812 Usable extra
facilities: •..•.•. none.
.813 Reserved equipment: •. 244 words of core storage
are required for skeleton
supervisor (does not include control routine for
disc operations); approximately 97, 300 words of
disc storage are required,
which includes the Assembler and FORTRAN compiler; 51,200 words of
disc storage are required
when the Assembler and
FORTRAN compiler are
excluded.
• 82
PROGRAM DIAGNOSTICS
. 51
Dynamic
.511 Tracing:
System Overhead
.821 Loading time: . . . • . . 10 to 15 seconds (average,
estimated by IBM).
.822 Reloading frequency: •• loaded at beginning of job
stack only.
Restarts: •..••...• as incorporated in user's
program.
.5
5/66
.8
• 83
Program Space
Available: ..••..• all except space requirements specified in .813 •
• 84
Program Loading
Time: ..•...•••. limited by speed of input
-device .
· 85
Program Performance: . . . . . • . no running overhead other
than I/O control; the 1130
Monitor only handles runto-run supervision.
..•.•.. can be included in
FORTRAN programs;
see Paragraph
418:182.451.
fA
AUERBACH
'"
418:201. 001
A.
STANDARD
ED]?
IBM 1130
SYSTEM PERFORMANCE
R£I'(IRTS
AUERBACH
SYSTEM PERFORMANCE
Because of the relatively low speed of the available peripheral devices, the performance of the
IBM 1130 has been evaluated only for the scientifically-oriented standard benchmark problems.
MATRIX INVERSION (418: 201. 300)
The standard estimate for inverting a non-symmetric, non-singular matrjx was computed by
the simple method described in Paragraph 4:200. 312 of the Users' Guide, using the standard
IBM 1130 floating-point subroutines. Estimates were made for both the standard precision
(6-digit precision) and the extended precision (9-digit precision) floating-point format.
GENERALIZED MATHEMATICAL PROCESSING (418:201.400)
Standard Mathematical Problem A is an application in which there is one stream of input data,
a fjxed computation to be performed, and one stream of output results. Two variables are
introduced to demonstrate how the time for a job varies with different proportions of input,
computation, and output. The factor C shows the effect of variations in the amount of computation per input record. The factor R indicates the ratio of input records to output records.
The procedure used to evaluate performance on the Standard Mathematical Problem is fully
described in Paragraph 4:200.2 of the Users' Guide. The standard problem was evaluated for
both Standard Configuration I and Configuration IX. Computation was performed in the extended
preciSion floating-point format (9-digit precision) using standard IBM 1130 subroutines.
Standard IBM 1130 input-output, code translation, and radjx conversion subroutines were used
to handle the complexities of getting data into and out of the system and of converting between
the various data codes, radices, and formats required for computation. To more realistically
portray the performance of the 1130 in an actual operating environment using FORTRAN programming, minimum overlapping of I/O operations with each other or with processing was
assumed. The standard 1130 FORTRAN subroutines do not permit any overlapping of I/O
operations with each other or with non-I/O processing. In estimating the performance of the
1130 on Standard Mathematical Problem A, only that portion of the I/O subroutines dealing
with processing of input-output interrupts was considered to be overlapped with the input or
output operation. Input and output operations were considered to be sequential.
Because the processor is delayed during I/O operations, both configurations are processorlimited for all conditions of the standard problem. The effect of the heavy requirements for
code translation and for radjx and format conversions, coupled with the sequential, non-overlapped input-output operations, is graphically illustrated by the significant differences in the
curves for various values of R, the output ratio (see the graphs on pages 418 :201. 400 and
418:201. 415). In Configuration I, the requirements for inputting and outputting predominate
for low and moderate values of C, the computation ratio. These requirements predominate
throughout the whole range of C for Configuration IX.
WORKSHEET DATA TABLE 2
CONFIGURATION
ITEM
REFERENCE
I
Fixed/floating point
5
Unit name
~.---
Size of record
~,---
msec/hlock
~
msee penalty
~....:I::L-
output
Standard
Mathernati-
cal
Problem
A
r---
output
output
Floatinl! Point*
1442 Model 7
IX
Floatinl! Point*
1134 Paper Tape Reader
1132 Printer
Console Typewriter
80 char
100 char
120 char
120 char
T1
165
1670
T2
744
8762
162
T4
579
1663
8707
msee/record
TS
259
280
msec/5 lODES
T6
msee/report
T7
*
output
60.8
434
-
4:200.413
-
60.8
299
Using standard IBM subroutines.
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
418:201. 300
IBM 1130
.3
MA TRIX INVERSION
.31
Standard Problem Estimates
.312 Timing basis: . • . . . . using estimating procedure
outlined in User's Guide,
4:200.312 •
• 311 Basic parameters: ... general, non-symmetric
• 313 Graph: .•.••..•.•• see graph below •
matrices, using floating
point to at least 8
decimal digits precision.
7-+-----+--~-+-+~++++----_+--~_+_+~+++r----_r--;__+_r;_rrH
2
10
7
II
if
4
IIJ
/1
2
Time in Minutes
for Complete
Inversion
1
7
II
4
2
O. 1
1/
1/
7
III
4
2
0.01
2
1
4
2
7
4
10
7
2
100
4
7
i.non
Size of Matrix .
/
5/66
A
(Contd. )
AUERBACH
.~
SYSTEM PERFORMANCE
418:201.400
.4
GENERALIZED MATHEMATICAL PROCESSING
.41
Standard Mathematical Problem A Estimates
formed in extended precision floating-point mode
(9-digit precision) .
. 413 Timing basis: • . . . . . using estimating procedure
outlined in Users' Guide,
4:200.413, and standard
IBM subroutines •
.414 Graph: . . . . . . . . . . see graph below for Configuration 1.
.411 Record sizes: . . . • • . 10 signed numbers; average
size 5 digits, maximum
size 8 digits.
• 412 Computation: . . • • . . . 5 fifth-order polynomials;
5 divisions and 1 square
root; computation is per-
CONFIGURATION I
100,000
7
4
2
10,000
7
~
t...-o'
4
f::J~
-
2
R
Time in
1,000
Milliseconds
per Input Record 7
4
-
"",'"
1.0
-R
0.1
-R
0.01
..-
..."
lEe
",.
~
\l'"\.~
~
~"
~"\.
1-
\l'
"<5:-
--- ....
2
100
7
4
2
10
2
0.1
4
7
2
1.0
4
7
2
4
10.0
7
100.0
C, Number of Computations per Input Record
(R = number of output records per input
~·ecord.
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
)
5/66
IBM 1130
418:201. 410
.41
STANDARD MATHEMATICAL PROBLEM A (Contd.)
.415 Graph: . . • . . . . . . . . see graph below for Configuration IX.
CONFIGURA TION IX
100,000
1
4
2
R
R
1.0
1.0
10,000
1
4
.~
~
~~
~
io""
I;'
~f'
-1'-""
~
t-R 0.1
R - 0.01
•..
2
Time in
1,000
Milliseconds
per Input Record 1
4
2
100
1
4
2
10
2
0.1
4
2
7
4
2
7
10.0
1.0
C, Number of Computations per Input Record
(R '" number of output records per input record.)
5/66
A
AUERBACH
@
4
1
100.0
&
.
418:211. 101
STAN .. RD
AEDP
AUERBAC~
IBM 1130
PHYSICAL CHARACTERISTICS
REPORTS
~_---_...J
PHYSICAL CHARACTERISTICS
Width,
inches
Unit
1131 Central Processor
Model lA or lB
(without disc storage)*
Model 2A or 2B
(with disc storage)*
?
Depth,
inches
?
Height,
inches
?
Weight,
pounds
Power,
KVA
?
?
BTU
per hr.
?
5S.2
29
31.5
500
1.4
3,SOO
1442 Card Read-Punch,
Model 6 or 7
43
24
49
525
0.46
1,500
1134 Paper Tape Reader,
Modell or 2
?
?
?
7.75
?
?
?
26
0.56
335
1055 Paper Tape Punch,
Modell
51.37
17.1
1132 Printer
47
29
44
500
0.61
1,600
1627 Plotter:
Modell
Model 2
lS
40
15
15
10
10
33
55
0.12
0.12
375
375
* Includes core storage.
General Requirements
Temperature: . . . . . . . . . . . • . 60 to 90°F.
Relative humidity: . . . . . . . . . . 10 to SO%.
Power: . . . . . . . . . . . . . . . . .. 115 volt, single-phase,
3-wire, 60 cycle; or 195, 220, 225 volt,
single-phase, 3-wire, 50 cycle.
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
r--.
1.
418:221. 101
Sf .... "
/AEDP
AUERBAC~
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):
lA
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)
IB
2A
2B
2315
Disk Cartridge
Peri12heral 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, Model 1
(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
IBM· SYSTEM/360
~...
International Business' Machines Corp.
c
c·
AUERBACH INFO, INC.
PRINTED IN U. S. A.
IBM SYSTEM/360
International Business Machines Corp.
AUERBACH INFO, INC.
PRINTED IN U. S. A.
420:001.001
IBM System/360
Contents
CONTENTS
Report 420: IBM System/360 - General
Introduction
..•.•.•..•............•...................
Data Structure . . . . . ' . . . . . . . . . . . . . • . • . . . . • . . . . . • . . . . . . . . . .
System Configuration (general) . . . . . . . . • . . . . . . . . . . . . • . . . • • . . . .
Internal Storage Main (processor) Core Storage . . . . . . . . . . • . . . . . . . . . . . .
2361 Core Storage (bulk) . . . • . . • . . . . . • . . . . . . . . . . . . . . . . . .
2302 Disk Storage . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2311 Disk Storage Drive . . . . . . . . . . . . . . . . . . . . . . . . . ". . . . . .
2321 Data Cell Drive . . . . . . . . . • . . . . . . . . . . • . . . . . . . . . . . . .
7320 Drum Storage. .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2301 Drum Storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2314 Direct Access Storage Facility . . . . . . . . . . . . . . . . • . . . . . .
Central Processors (general) . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Console . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
System Control Panel . . . . . . . . . . . . . . . . . . . . . . . . . .
2150 Console. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1052 Printer-Keyboard . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1050 Data Communication System . . . . . . . . . . . . . . . . . . . . . . . . .
Input-Output; Punched Card and Tape 2540 Card Read P'unch . . . . • • • • . . . • • . . . • . . . . . . . . . . . . . . . .
1442 Card Read Punch . . . . . • . . . • • • • . • . • • • • • . . . • . . . . . . . .
2671 Paper Tape Reader . . . . . . . . • . . . . . . . • . . . . . . . . . . . . . .
2501 Card Reader . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2520 Card Read Punch . . . . . • . . . . . . . . . • . . . • . . . . . . . . . . . . .
2560 Multi-Function Card Machine . . . . . . . . . . . . . . . . . • . . . . . .
Input-Output; Printers 14Q.3 Printer . . . . . . . . . . . • . . . • . . . . . . . . . . . . . . . . . . . . . . .
1404 Printer . . . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . . . . .
1443 Printer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1445 Printer (MICR) . . . . • . . . . . . . . . . . • . . . . . . . . . . . . . . . . .
2203 Printer • . . . . . • . • . . . . . . . . . . . . . . . . . . . • . . . . . . . . . .
Input-Output; Magnetic Tape 2400 Series Magnetic Tape Units at 800 bpi . . . . . . . . . . . . . . . . . .
7340 Hypertape Drive . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2415 Magnetic Tape Unit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2400 Series Magnetic Tape Units at 1600 bpi . . . . . . . . . . . . . . . . . .
Input-Output; Others "
2250 Display Unit. . . . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . .. .
2260 Display Station . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7770 Audio Response Unit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7772 Audio Response Unit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1412
1419
1418
1428
1231
1285
420:011
420:021
420:031
420:041
420:042
420:043
420:044
420:045
420:046
420:047
420:048
420:051
420:061
420:061. 131
420:061.132
420:061.133
420:061. 134
420:071
420:072
420:073
420:074
420:075
420:076
420:081
420:082
420:083
420:084
420:085
420:091
420:092
420:093
420:094
420:101
420:102
420:103
420:104
Magnetic Character Reader . . . . . . . . . . . . . . . . . . . . . . . . . .
Magnetic Character Reader . . . . . . . . . . . . . . . . . . . . . . . . . .
Optical Character Reader . . . . . . . . . . . . . . . . . . . . . . . . . . .
Alphameric Optical Reader . . . . . . . . . . . . . . . . . • . . . . . . . .
Optical Mark Page Reader . . . . . . . . . : . . . . . . . . . . . . . . . .
Optical Reader . . . . . . . . . . . . . . . . . . . . . • . . . . . . . . . • • .
420:105.121
420:105.122
420:105.123
420:105.124
420:105.125
420:105.126
Data Adapter Unit • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Data Communication System . . . . . . . . . . . . . . . . . . . . . . . . .
Data Communication System . . . . . . . . . . . . . . . . . . . . . . . . .
Process Communication System . . . . . . . . . . . . . . . . . . . . . . .
Input Station . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Printer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Bell System 83B2 and 83B3 Service . . . . . . . . . . . . . . . . . . . .
Western Union Plan 115A . . • . • . . '. . . . . . . . . . . . . . . . . . . .
1009 Data Transmission Unit . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1013 Card Transmission Terminal . . . . . . . . . . . . . . . . . . . • . . . .
7701 Magnetic Tape Transmission Terminal . . . . . . . • . . . . . . • . . .
420:106
420:106.121
420:106.122
420:106.123
420:106.124
420:106.124
420:106.125
420:106.126
420:106.127
420:106.128
420:106.129
2701
1050
1060
1070
1031
1033
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
IBM SYSTEM/360
420:001.002
7702
7710
7711
7740
7750
2740
2741
Magnetic Tape Transmission Terminal . . . . . . . . . . • . . . . . . •
Data Communication Unit . . . . • . . . . . . . . . . . . . . . . . . . . . .
Data Communication Unit . . . . . . . . . . . . . . . . . . . . . . . . . . .
Communication Control System . . • . . . . . • . . . . . . . . . . . . . .
Programmed Transmission Control . • . . . . . . . . . . . . . . . . . .
Communications Terminal . . . . . . . . . • • . • . . . • . . • • . . . . .
Communications Terminal . • . . . . . . . • . . . . . . . . . . . . . . . .
420:106.129
420:106.130
420:106.131
420:106.132
420:106.133
420:106.134
420:106.135
2702 Transmission Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2703 Transmission Control . . . . . . . • . . . . . . . . . . . . . . . . . . . . .
2712 Remote Multiplexor . • . . . . . . . . . . . . . . • . . • . . . . . . . . . . .
420:107
420:108
420:108
2280 Film Recorder . . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . . . .
2281 Film Scanner . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2282 Film Recorder/Scanner . . . . . . . . . . . . . . . . . . • . . . . . . . . .
420:109.121
420:109.122
420:109.123
Simultaneous Operations . . . . • . . . . . . . . . . . • . . . . . . . . . . . . . . . . . .
Multiplexor and Selector Channels . . . . . . • • . . • • . • . . . . • . . . . . .
Input- Output Control. • . . . . . . . . . . . . . . . . . . . . . . . . • • . • . . . .
Instruction List . • . . . • . • . . • . • • . . . . . . . . . . . . . . . . . . . . . . . . . . .
Compatibility With IBM 1401/1440/1460 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
With IBM 1410/7010 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
With IBM 7070/7074 . . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . . . .
With IBM 7080 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
With IBM 7040/7090 Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
With IBM 1620 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Data Codes (ASCII-8 and EBCDIC) . . . . . . . . . . . . . . . . . . . • . . . . . . . . .
Problem Oriented Facilities Operating System/360 . . . . . . . . . . . . . . . . . • . . . . . . . . . . . . . . .
Basic Operating System/360. . . . . . . . . . . . . . . . • . . . . . . . . . . . .
Basic Programming Support . . . . . . . . . . . . . . . . . . . . . . . . . . . .
/,.--
420:111
420:111. 2
420:111.3
420:121
420:131
420:132
420:133
420:134
420:135
420:136
420:141
420:151
420:152
420:153
Process Oriented Languages Operating System/360 FORTRAN . . . . . . . . . . . . . . . . . . . . . . . . .
Basic Operating System/360 FORTRAN . . . . . . . • . . . . . . . . . . . . .
Basic Programming Support FORTRAN . . . . . • . . . . . . . . . . . . . . .
Operating System/360 COBOL . . . . . . . . . . . . . . . . . . . . . . . . . . .
Basic Operating System/360 COBOL . . . . . . . . . . . . . . . . . . . . . . .
Operating System/360 P L/I . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
420:161
420:162
420:163
420:164
420:165
420:167
Machine Oriented Languages Operating System/360 Assembler . . . . . . • . . . . . . . . . . . . . . . . . .
Basic Operating System/360 Assembler . . . . . . . . . . . . . . . . . . . . .
Basic Programming Support Assembler . . . . . . . . . . . . . . . . . . . . .
420:171
420:172
420:173
Operating Environment Operating System/360 . . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . . .
Basic Operating System/360. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Basic Programming Support . . . . . . . . . . . . . . . . . . . . . . . . . . . .
System Performance (general) • • . . • . • . . . . . . . . . . . . . . . . . . . . . . . .
Physical Characteristics . . . . • • . . • . . . . . • . . • . . • . . • . • . . . . . . . . .
Price Data . . . . . . • . . . . . • . . . . . • • . . . . . . . . . . . . . • . . . • . . . . . .
420:191
420:192
420:193
420:201
420:211
420:221
Report 422: IBM System/360 - Model 20
Introduction. . . . . . . . . . . •.. • . . . . . . . . . . . . . . . • • . . . . . . • . . . . . .
System Configuration . . . • • . • . • • . . . . . . . . . . • . . • . . . . . . . . . . . . .
Central Processor . • • . . . . • • . . . . . . . • • • • . . • • . . • . . . . . . . . . . . .
Simultaneous Operations . . . . • . • . • . • . . . . . • . . . . . . . . . . . . . . . . . .
Software . . . . . . . . • . . . . . . . • • • • " . • . . . . . . . . • . . . . . . . . . . . . . .
System Performance . . . . . . . . • • • . . . . • • . • . . . . • . • . . . . . . . . . . . .
Price Data . . • • . . . • . • . . . • . • . . . . . • . . . . . . . . . . • . . . . . . . . . . •
422:011
422:031
422:051
422:111
422:151
422:201
422:221
Report 423: IBM System/360 - Model 30
Introduction. • . . . . . • . . . . . . . . . . . . • . . . . . . . • . . . . . . . . . . . . . . .
System Configuration . . • • . . . . . • . . . . . . . . • . . . • . . . . . . . . . . . . ..
Central Processor . . • • • . . . • . • • . . . . • . . • • . . • . . . . . . . . . . . . . . .
Simultaneous Operations . . • . . . . . . . . . • . • . . . . • . . . . . . . . . . . . . . .
System Performance . . . . • . . . . . . . • . . • . . • • . . . . • . . . . . . . . . . . . .
423:011
423:031
423:051
423:111
423:201
(Contd.)
8/65
420:001.003
CONTENTS
Report 424: IBM System/360 - Model 40
Introduction . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
System Configuration . . . . . . . . . . . . . . . . . . . • • . . • . . . . • . . . . . . . .
Central Processor . . . . . • . . • • • • • . • • . . . . . • • . . • . . . . • . . . . . . . •
Simultaneous Operations . . . • . . • . • . • . • • • • . • . . • . . . • • . • • • . . . . .
System Performance .•••.•.•••.•.•••.•..•••.•..••••...••.•
424:011
424:031
424:051
424:111
424:201
Report 425: IBM System/360 - Model 50
Introduction. • . • • . . • . . . . • . . . . . . • . • . . . . . • . . . . . . . . . . . . . . . .
System Configuration . • . • . • • . • • • . • • • . . . . • . . . . . . • . . . . . . . . . .
Central Processor . • . • • . . • . • • . • . . . • . • . • . . • . • • . • . . . . .••..•
Simultaneous Operations . . . . . . . • . . . . . . . . . . . . . • • . . . . . . . . . . . .
System Performance . . • . . . • . . • . . . • . . • • • . . • . . . . . . . . . . . . . . . .
425:011
425:031
425:051
425:111
425:201
Report 426: IBM System/360 - Model 65
Introduction. . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . . . . . .
System Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Central Processor . . . . . . . . -. . . . . . . . . . . . . . . . . . . . . . .
Simultaneous Operations . . . . . . . . . . . . . . . . . . . . . . . . . . .
System Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
426:011
426:031
426:051
426:111
426:201
Report 427: IBM System/S60 - Model 67
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
System Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Central Processor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Simultaneous Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Programming Languages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Operating Environment (Time-Sharing Monitor) . . . . . . . . . . . . . . . . . . . .
System Performance . . . . . . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Price Data . . . . . . . . . . . . • . . . . . • . . . . . . . . . . . . . . . . . . . . . . . . .
427:011
427:031
427:051
427:111
427:161
427:191
427:201
427:221
Report 428: IBM System/S60 - Model 75
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . . .
System Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
Central Processor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Simultaneous Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
System Performance . . . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
428:011
428:031
428:051
428:111
428:201
8/65
420:011.1 00
IBM System/360
Introduction
INTRODUCTION
1
SUMMARY
System/360 is the "brand name" for IBM's extensive third-generation family of central
processors, storage modules, peripheral devices, and supporting software. Noteworthy
characteristics of the System/360 include:
•
The "universality" concept - a single line of equipment designed to
handle widely varying types and sizes of computer workloads.
•
The high degree of program compatibility, both upward and downward,
among most of the processor models.
•
The wide range of input-output and storage devices.
•
The numerous arithmetic modes and data formats, and the resulting
complexity of machine-language coding.
•
The emphasis upon software support through integrated operating
systems, now offered at three different levels.
•
Solid-Logic Technology, IBM's name for the "hybrid" electronic
circuitry used in the System/360, which is a compromise between
earlier solid-state techniques and true monolithic integrated circuits.
The System/360 constitutes the "third generation" of equipment from the leading computer
manufacturer. As such, it is now the primary standard for comparison in most computer
selection studies, and it is important for every computer user to develop a good understanding of its characteristics, performance, strenogths, and weaknesses. This comprehensive report will help you to gain that basic understanding and will serve as a continuing
reference source.
The format of this report is designed to present and analyze all the facts about the System!
360 in a way that will make it easy for you to locate and study the material you require,
while placing proper emphasis upon the similarities and differences among the various
models. This coverage consists of a general Computer System Report (behind Tab 420)
which analyzes the concepts, hardware, and software that are common to all System/360
models, and individual subreports (behind Tabs 422 through 428) which report the characteristics, performance, and pricing of computer systems using each of the System/360
processor models. (System/360 Models 20 and 67 make use of specialized software,
which is therefore described within the individual subreports for these models.)
For the same purposes of clarity and reader convenience, this Introduction is divided into
six independent sections, each of which describes and (where pertinent) analyzes some
particular facet of the System/360. Each section is independent and can be read as your
needs and interests warrant. The six sections are:
.1
.2
.3
.4
.5
.6
.2
Summary
System/360 - the First Year
Data Structure
Hardware
Software
Compatibility .
SYSTEM/360 - THE FIRST YEAR
As announced on April 7 , 1964, the IBM System/360 consisted of 6 program-compatible
central processors spanning a 50-fold range of internal processing speeds, 44 new and
previously-announced peripheral devices, and a comprehensive package of language processors, utility routines, and control programs called the Operating System/360. IBM
announced that "the System/360 marks the achievement of a truly all-purpose computer
that can solve any type of data-handling problem with greater speed and efficiency than
ever before. "
©
1965 AUERBACH Corporotion ond AUERBACH Info, Inc.
8/65
IBM SYSTEM/360
420:011.200
.2
SYSTEM/360 - THE FIRST YEAR (Contd.)
Unquestionably the System/360, as originally announced, did offer an unprecedented range
of processing speeds, storage capacities, and input-output equipment, and it spanned a
broader range of potential applications than any previous computer system. Nevertheless,
it soon became apparent that there were some significant weaknesses in the origiriallyannounced line of hardware and software, as noted in last year's AUERBACH standard EDP
Reports analysis of the System/360. A brief chronology of the announcements that have
changed the complexion of the System/360 since April 1964 will show how IBM has endeavored to correct these weaknesses and to fill out and strengthen its overall product line.
•
August, 1964: IBM announced the System/360 Model 92, an ultra-highperformance computer "more powerful than any computer now available. "
IBM said it would enter into special contracts to build Model 92 computers
based on customers' particular needs. Although Model 92's instruction
repertoire and data format are similar to those of the smaller System/360
models, it will not be program-compatible with them because Model 92
lacks facilities for decimal arithmetic.
•
October, 1964: The first public demonstration of a working System/360
(a Model 40) was featured at the Business Equipment Exposition and Conference in Los Angeles.
•
October, 1964: IBM announced a series of new Compatibility Features hardware-software combinations called "emulators" - to permit various
models of the System/360 to execute programs written for the following
older IBM computers: 1410, 7010, 7070, 7074, 7080, 709, 7040, 7044,
7090,7094,7094 II. (More recently, the 1620 was added to the list.)
The only previously-available CompatibilitY Features enabled the smaller
System/360 models to execute IBM 1401, 1440, or 1460 programs. The
new emulators represented IBM's answer to widespread complaints
from users of its other second-generation computers about the difficulties
involved in reprogramming for the System/360.
•
November, 1964: IBM announced the System/360 Model 20, a small-scale,
business-oriented computer designed primarily for small companies that
are considering a step upward from conventional punched-card accounting
machines. Announced along with Model 20 were the 2560 Multi-Function
Card Machine, a 500-card-per-minute punch, and several other new
peripheral devices. Model 20 extends the System/360 range downward
into new marketing areas, but its degree of compatibility with the larger
System/360 models is limited by its much smaller instruction repertoire,
its limited core storage capacity, and its different method of handling
input-output operations.
•
December, 1964: IBM advanced the scheduled date for initial customer
deliveries of Model 30 and 40 systems from the third quarter to the second
quarter of 1965. Delivery dates for Models 50 through 70 were also moved
up. The advanced delivery dates, "made possible by accelerated production
at IBM manufacturing facilities," were IBM's response to vigorous complaints
about the long lead time between announcement and scheduled deliveries of
the System/360.
•
December, 1964: IBM reduced the extra-usage rental rate for most
System/360 components from 30% to 10% of the hourly rate for prime-shift
use. The reduced extra-usage rate applies to all units with model numbers
in the 2000 series (and the 1302 Disk Storage Unit was concurrently redesignated the 2302). This very significant reduction, which can have a major
effect upon rent-versus-buy decisions, was IBM's response to the elimination
or great reduction of extra-shift rental charges in several competitive computer lines.
•
January, 1965: IBM announced a 33 per cent increase in internal processing
speed of the System/360 Model 30 through reduction of its core storage cycle
time from 2.0 to 1. 5 microseconds. Concurrently, the 2400 Series Magnetic
Tape Units were speeded up from 22,500 to 30,000 bytes per second (Modell)
and from 45,000 to 60,000 bytes per second (Model 2); the 90,000 bytes-persecond speed of the Model 3 units remained unchanged. These speed increases helped to keep the performance of the System/360 in line with that
of the program-compatible RCA Spectra 70 computer family, announced in
December, 1964.
(Contd.)
8/65
420:011.201
INTRODUCTION
.2
SYSTEM/360 - THE FIRST YEAR (Contd.)
•
February, 1965: The IBM 1130 was announced as a desk-size computer
designed primarily for individual use by scientists and engineers. The introduction of the 1130, which bears little resemblance to the System/360, was
IBM's first clear indication that the System/360, even in extended or restricted versions, is not practical for every type and size of computer application.
•
February, 1965: IBM announced a complete restructuring of software support
for the System/360. To meet complaints that the Operating System/360 required too much core storage and peripheral equipment to perform its impressive functions, while the facilities of the Special Support System (the only
previous alternative) were far too restricted, IBM committed itself to the
gigantic task of producing three different levels of software support: the
Operating System/360, Basic Operating System/360 (BOS), and Basic Programming Support (BPS). Table V shows the facilities offered at each level
and their scheduled delivery dates.
•
March, 1965: IBM announced two more additions to the System/360
line: the time-sharing Models 64 and 66. Models 64 and 66 featured
an associative memory to facilitate dynamic relocation of programs
and a channel controller to permit flexible interconnections among the
system components. The two time-sharing systems used the standard
System/360 instruction repertoire plus additional instructions to
direct the time-sharing features. Announced as non-standard models
to be offered only through special proposals, the time-sharing systems represented IBM's response to the success of General Electric
and other manufacturers in winning contracts for multi-console, timesharing applications, where the System/360 as originally announced
had been weak.
•
March, 1965: The 2870 Multiplexor Channel was announced, providing
the capability to connect a large number of low-speed input-output
devices to the larger System/360 models. Previously, the lack of
Multiplexor Channels for the larger models had seriously restricted
upward compatibility and made it almost mandatory to use a Model 30,
40, or 50 processor in conjunction with the larger processors for
control of punched-card, printer, and/or data communications operations.
•
March, 1965: IBM announced the 2260 Display Station, a low-cost,
buffered, cathode-ray-tube terminal for remote or local displays of
alphameric data. An optional keyboard permits convenient man/
machine communication. Concurrently, the more expensive 1015
Inquiry Display Terminal, which featured a self-storing dark-trace
cathode ray tube, was dropped from the System/360 product line.
•
March, 1965: IBM demonstrated the 1401 Compatibility Feature for
the System/360 Model 30 at its Endicott, New York facility. A wide
variety of user-submitted 1401 programs were run on the System/360
with relatively few difficulties and, in most cases, at significantly
higher speeds than on the original 1401. The practicality of the allhardware, stored-logic approach to 1401 compatibility used in the
Model 30 was convincingly demonstrated.
•
April, 1965: IBM completely restructured the upper half of the Sys:tem/360
line by adding three new models and dropping five others. Model 65 superseded original Models 60 and 62, Model 75 superseded Model 70, and
Model 67 superseded the just-announced, time-sharing Models 64 and 66.
Models 65 and 75 offer significantly higher processing speeds at lower
prices than their predecessors, indicating that the principal purpose of
the restructuring was to bring the price/performance ratios of the larger
System/360 models more closely into line with the offerings of competitors.
(Meanwhile, IBM indicated that the design of tJ:1e Model 92 was being "reevaluated, " and that no performance details would be released until redesign
of the Model 92 - or its successor - had been completed. )
•
April, 1965: IBM announced the 2314 Direct Access Storage Facility, the
2415 Magnetic Tape Unit, and the 2540 Card Read Punch. The 2314, a multidrive, replaceable-cartridge disc storage unit, is the seventh distinct type
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
IBM SYSTEM/360
420:011.202
.2
SYSTEM/360 - THE FIRST YEAR (Contd.)
of auxiliary storage in the System/360 line. The 2415, a low-speed, economymodel tape unit, provides magnetic tape capabilities for the System/360 Model.
20 and makes it IBM's lowest-priced tape system. The 2540 supersedes the
widely-used 1402 Card Read Punch and provides increased punching speed (300
cards per minute) and a number of detail improvements.
.3
•
April, 1965: Initial customer deliveries of the System/a60 were made.
IBM announced that more than 1,000 System/360' s will be delivered by the
end of 1965, and that deliveries will reach a rate of 35 systems per day in
mid-1966.
•
July, 1965: IBM underlined the steadily increasing importance of data communications applications by announcing eight new communications devices.
Most significant are the 2703 Transmission Control, which links up to 176
communications lines to a System/360, and the 2712 Remote Multiplexor,
which can multiplex data from as many as 14 remote, low-speed terminals
over a single high-speed line to a computer.
•
August, 1965: IBM made doubled data rates available for all of the 2400
Series Magnetic Tape Units through a recording technique called "phase
encoding, " which permits 1600 bytes per inch to be recorded on standard
half-inch tape.
•
August, 1965: IBM announced the System/360 Model 44, a processor
especially designed for scientific and process control applications.
Model 44 features high-speed binary arithmetic, a built-in single-disc
storage drive, and up to 131,072 bytes of core storage; it cannot be
equipped with decimal arithmetic facilities or Selector Channels .
DATA STRUCTURE
The basic unit of data storage in the System/360 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 System/360 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.
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 essential for efficient operation of the
larger central processors, which access up to eight 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 System/360 line of processors ,the Model 20 can perform arithmetic
operations on two basic types of operands: fixed-point binary, using the 16-bit halfword
as the standard operand size, and variable-length decimal. The larger System/360
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 used by all models except Model 20 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 characteristic occ'tfies 7 bits in both formats and permits
representation of numbers ranging from 10- 7 to 10 75 .
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.
IBM's adoption of the 8-bit data code for use with the System/360 is higbly significant.
The 8-bit byte structure has certain basic advantages-over the currently-popular 6-bit
(Contd.)
8/65
INTRODUCTION
.3
420:011.300
DATA STRUCTURE (Contd.)
data format: decimal digits can be packed more conveniently, the new standard 7 -bit
ASCII code and the Extended BCD Interchange Code can be used, and today's familiar
character sets can be conveniently expanded. These advantages, coupled with IBM's wide
dissemination of the 8-bit code, will probably induce other computer manufacturers to
adopt the 8-bit byte as the standard data structure in their future equipment.
.4
HARDWARE
.41
Central Processors
Nine central processor models currently form the nucleus of the System/360. Five of the
central processors are program-compatible and suitable for)! broad range of business
and scientific applications: Models 30, 40, 50, 65, and 75. Model 20 is a scaled-down
version of the other System/360 models, designed primarily for use in small data-processing
installations or as a remote terminal linked to larger System/360 processors. Model
44 is oriented toward high performance in scientific and process control applications.
Model 67 is specially designed for use in time-sharing and multi-processor environments.
The ultra-large-scale System/360 Model 90 Series will be offered to satisfy customdesign requests from large and diversified information-processing installations. Models
20, 44, 67, and the 90 Series will each function with an individualized set of software.
Comparative arithmetic execution times for the various System/360 central processors
are illustrated in Table I. Table II shows the various core storage capacities that can be
obtained with each of the basic central processors.
TABLE I: ARITHMETIC EXECUTION TIMES FOR THE IBM SYSTEM/360
PROCESSOR MODELS
CENTRAL PROCESSOR MODEL
TASK
Mod 30 Mod 40 Mod 50
Mod 20
Mod 65
Mod 67
Mod 75
Fixed Point Binary
c=a+b
c = ab
c = alb
180
78
296
481
36
113
216
12
40
44
96*
395*
767*
64*
178*
349*
#
#
#
107*
295*
350*
#
#
#
161*
874*
1,717*
#
#
3.5
7.0
11
4.2
7.7
12
2.3
5.1
9.0
35
86
97
9.0
32
47
9.7
33
49
7.3
25
33
43*
105*
157*
14
29
30
4.7
6.1
9.3
5.4
6.8
10
2.4
3.6
5.4
62*
294*
511*
21
49
81
4.8
9.7
16
5.5
10
17
2.4
5.6
8.6
Fixed Point Decimal
c=a+b
c = ab
c = alb
658
3,968
5,162
Floating Point - Short
c=a+b
c= ab
c = alb
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 floatingpoint times are based on both the short-form (32 bits) and the longform (64 bits) binary operands.
#
Facility not available in Model 20.
*
With optional feature.
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
IBM SYSTEM/360
420:011.410
.41
Central Processing (Contd.)
AU of the central 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. The Model 20 central processor includes eight
16-bit general registers; the other System/360 models have sixteen 32-bit general registers and four 64-bit floating-point registers. The general registers can be used as fixedpoint accumulators or as index registers.
The remainder of this discussion of Central Processors refers to the processors
associated with System/360 Models 30, 40, 50, 65, and 75.
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 !'bas~ 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.
All three parts of an address (base, displacement, and index) are treated as unsigned,
positive binary integers and are added together with overflows ignored. Since every address includes a base, the sum is always 24 bits long; this provides a logical capability
for addressing up to 16,777,216 bytes, although the direct part of the address (the 12-bit
displacement) permits direct addressing of only 4,096 bytes. The base-register technique of address formation facilitates program relocation and segmentation, at the expense of increased programming complexity.
The basic arithmetic mode of the System/360 is· fixed-point binary, using 32-bit operands
and twofs-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 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 Standard Instruction Set includes 86 instructions which perform fixed-point 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 by
most input-output devices and the two-digits-per-byte format used for decimal arithmetic.
The other instructions in the standard set are quite conventional in form and function.
The Floating-Point Arithmetic feature is optional in Models 30 and 40 and standard in the
larger models. It provides 44 additional instructions for addition, subtraction, multiplication, division, loading, storing, and comparison of both "short" (32-bit) 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.
The Decimal Arithmetic feature is optional in Models 30 and 40 and standard in the larger
models. It provides eight 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 can be up to 16 bytes (31 digits and sign) in length.
The length of each decimal field is specified in the instruction referencing it; the word
mark concept used throughout the IBM 1400 series has been abandoned. Two-address
(6-byte) instructions of the storage-to-storage type are used for aU decimal operations;
the general and floating-point registers are not utilized.
Decimal arithmetic in the System/360 is considerably slower than binary arithmetic. It is
designed for processes which require relatively few computational steps between input and
output, so that radix conversions and use of fast-access registers for temporary storage
of results are not justified.
(Contd.)
8/65
INTRODUCTION
.41
420:011.411
Central Processors (Contd.)
A comprehensive interrupt system permits System/360 processors to respond to a variety
of special conditions arising within the processor and in peripheral units. The basis for
the interrupt system is the Program Status Word (PSW), a double (64-bit) word that indicates the operational status of a program. When an interrupt condition arises, the active
PSW is automatically stored in a fixed location whose address depends upon the class of
the interruption. The processor then fetches a new PSW from another fixed location, and
this new PSW governs entry to a routine that services the interrupt condition. After the
interrupt condition has been serviced, the PSW of the interrupted program is restored to
the active pOSition, thereby resetting the processor to the status it had just before the interrupt occurred. There are five basic classes of interrupts: input-output, program error,
supervisor-call, external signal, and machine check. SpeCific bits in the PSW can be used
to mask off (i. e., inhibit) certain interrupt conditions.
The System/360 offers powerful processors that include many valuable features. But because these processors incorporate a design that is meant to satisfy a vast group of computer users involved in diverse applications, internal processor complexity has resulted.
For example, 14 different Add instructions, each with its own format rules, can be selected for use within the same program. Due in part to the intricacies and the sheer number
of available instructions, System/360 programming at machine and assembly-language
levels, and program checking, patching, and similar operations, are relatively complex
and conducive to human errors. This complexity results also from the indirect, base-plusdisplacement addressing system, the operand placement rules (which vary for each of the
four types of operands), the interrelationships between operand types and machine instruction types, and the unusual (and therefore initially"unfamiliar) data codes and index register
methods. These points are discussed in more detail in the description of the Central
Processors in Section 420:051.
. 42
Internal Storage
The most outstanding characteristic of the System/360's internal storage scheme is the wide
variety of capacities and speeds in which both core storage and auxiliary storage can be
supplied. Table IT indicates the range of core storage sizes and speeds available with various models of the System/360.
TABLE II: SYSTEM/360 MAIN CORE STORAGE CHARACTERISTICS
SYSTEM MODEL
Core Storage
Capacity, Bytes
4,096
8,192
16,384
32,768
65,536
131,072
262,144
524,288
1,048,576
Cycle Time,
Mod 20
B20
C20
D20
-
-
IJ sec
Mod 30
-
C30
D30
E30
F30
-
Mod 40
Mod 50
Mod 65
Mod 75
-
-
-
-
D40
E40
F40
G40
H40
-
F50
G50
H50
-
-
H65
-
165
J65
H75
175
J75
3.6
1.5
2.5
2.0
0.75*
0.75*
Bytes Accessed
per Cycle
1/2
1
2
4
8
8
Cycle Time per
Byte, /J,sec
7.2
1.5
1. 25
0.50
0.094*
0.094*
* Effective cycle time in Models 65 and 75 is somewhat faster due to
interleaved accessing of core storage.
The Storage Protection feature (optional in Models 30 and 40, standard in the larger
models) can prevent 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. {Only the Model
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
420:011.421
.42
IBM SYSTEM/360
Internal storage (Contd.)
67's extended Storage Protection feature offers both read and write protection.) Users
of large-scale, multiprogrammed computers may find this protection scheme to be inadequate, since only 15 different programs cim be protected against overwriting at anyone
time. Another disadvantage lies in the fact that if any input data has been inadvertently
assigned a 4-bit "protection key" of zero, or if any program has a 4-bit "storage key" of
zero, no data in core-storage can be positively safeguarded against overwriting.
Seven different types of auxiliary storage devices are available in the form of magnetic
cores, drums, discs, and strips. The storage capacity of these devices ranges from less
thanonemillionbytes to over three billion bytes under one control (and more than one
control can be used in a system). Similarly, access times can be chosen to suit the system, with average times ranging from 8 microseconds to 450 milliseconds. Table III
lists the various auxiliary storage devices with their principal functional characteristics.
Noteworthy among the auxiliary storage devices offered with the System/360 is the 2361
Core Storage unit - an addressable bulk store that can act as a direct extension of the
main core storage and can perform identical functions. The 2361 offers users of Models
50, 65, 67, and 75 up to 8,388,608 bytes of directly -addressable supplementary core
storage. Cycle time is eight microseconds, a figure that is considerably slower than that
of the main core storage; but when the 2361 is considered as a random access storage device, its eight-microsecond access time is far superior to that of any disc or drum unit.
The 2361 Core Storage unit will make feasible many system applications which were
previously impractical due to the inability to gain rapid random access to sufficiently
large volumes of data.
Six additional types of random access storage equipment are offered. Three types are
disc-based, and two of these (the 2311 and the 2314) use removable Disk Packs. Two
types are drum-based and have the same drum revolution time of 17 milliseconds, but
the newer 2301 reads and writes 4 bits in parallel to achieve an extremely high data transfer rate: 1,200,000 bytes per second. The sixth random access device is the highcapacity 2321 Data Cell Drive. Each Data Cell holds forty million bytes and canbe removed
and stored independently. Ten cells can be mounted on a single Data Cell Drive at one
time, yielding an on-line capacity of 400 million bytes per drive. Access is slow by comparison to other units: nearly half a second is required to gain access to a randomlyplaced record or file. The main features of the random access units (now called "direct
access" units in IBM documentation) are compared in Table m.
TABLE m: SYSTEM/360 AUXILIARY STORAGE UNITS
DEVICE
CAPACITY RANGE
(Millions of
Bytes Per
Control Unit)
2361 Core Storage
lto8
2302 Disk Storage
112 to 897
2311 Disk Storage
AVERAGE
ACCESS TIME
DATA TRANSFER
RATE
(Bytes/sec)
REPORT
REFERENCE
500,000 to
2,000,000
420:042
165 msec
156,000
420:043
7 to 58
98 msec
156,000
420:044
2314 Direct Access Storage
26 to 207
75 msec
312,000
420:048
2321 Data Cell Drive
40 to 3,200
450 msec
55,000
420:045
8 !lsec
7320 Drum Storage
0.8to6.6
9.6 msec
135,000
420:046
2301 Drum Storage
4.1 to 16.4
8.6 msec
1,200,000
420:047
. 43
Sequential Input-Output Units
A wide variety of sequential input-output equipment is offered for the System/360.
./
A new line of medium-performance magnetic tape equipment, the 2400 Series, reads
and records data in nine tracks (8 data and 1 parity) on half-inch tape at a density of
800 or 1600 bytes per inch. The nine-track recording is not compatible with any
previous IBM magnetic tape equipment, though a 729 Compatibility feature is available
to enable a 2400 Series tape drive to read and write in the seven-track IBM 729 mode.
(Contd.)
8/65
420:011.430
INTRODUCTION
.43
Sequential Jnput-Output Units (Contd.)
The new 7340 Model 3 Hypertape Drive features the highest recording density yet
announced for a commercial tape unit - 3,022 bytes per inch - and a data transfer
rate of 340,000 bytes (or 680,000 packed decimal digits) per second. The cartridge
loading feature of Hypertape facilitates tape handling and reduces the likelihood of
tape contamination or damage.
The punched card, paper tape, and printing equipment available with the System/360
closely resembles previous IBM offerings, with a few notable exceptions. The 2520
Card Read Punch provides punching speeds of up to 500 cards per minute, or 200
cards per minute faster than any card punch previously offered by a major computer
manufacturer. The 2560 Multi-Function Card Machine, currently available only in
Model 20 systems, is a highly unconventional unit that combines many of the facilities
of a card reader, card punch, collator, and interpreter. Several models of the 1403
Printer can now be equipped with a Universal Character Set feature that permits up
to 240 different characters to be printed; the relationships between internal codes and
printed characters can be changed at any time by the user.
Table IV summarizes the capabilities of representative models of System/360 sequential input-output devices.
TABLE IV: SYSTEM/360 SEQUENTIAL INPUT-OUTPUT UNITS
UNIT
MODEL
Punched Cards
1442 Card Read Punch
PEAK SPEED
REPORT REFERENCE
N1
Reads 400 cpm
420:072
2501 Card Reader
B1
B2
Reads 600 cpm
Reads 1, 000 cpm
420:074
420:074
2520 Card Read Punch
B1
Reads 500 cpm
Punches 500 cpm
420:075
2540 Card Read Punch
-
Reads 1,000 cpm
Punches 300 cpm
420:071
2560 Multi-Function Card Machine
-
Reads 500 cpm
Punches 91 to 260 cpm
Prints 140 cpm
420:076
-
1,000 char/sec
420:073
Punched Pal2er Tal2e
2671 Paper Tape Reader
MagnetiC Tal2e
2400 Series Units
1
2
3
4
5
6
2415 Magnetic Tape Unit
1,2,3
4,5,6
7340 Hypertape Drive
Printers
1403 Printer
1
3
2, 7,
3, N1
30,000
60, 000
90, 000
60,000
120,000
180,000
bytes/sec
bytes/sec
bytes/sec
bytes/sec
bytes/sec
bytes/sec
15,000 bytes/sec
15,000/30,000 bytes/
sec
170, 000 bytes/sec
340,000 bytes/sec
600lpm
l,100lpm
420:091
420:091·
420:091
420:094
420:094
420:094
420:093
420:094
420:092
420:092
420:081
420:081
600 lpm (forms) or
800 lpm (cards)
420:082
N1
200 to 600 lpm
420:083
1445 Printer (MICR)
N1
190 to 525 lpm
420:084
2203 Printer
-
300 to 750 lpm
420:085
1404 Printer
2
1443 Printer
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
420:011.440
IBM SYSTEM/360
TABLE IV: SYSTEM/360 SEQUENTIAL INPUT-OUTPUT UNITS (Contd.)
. UNIT
MODEL
Optical Readers
1231 Optical Mark Page Reader
1285 Optical Reader
PEAK SPEED
REPORT REFERENCE
N1
2,000 sheets/hr.
42.0:105
-
2,6.00 lpm
420:1.05
1418. Optical Character Reader
1
30.0 to 420 documents/min.
42.0:105
1428 Alphameric Optical Reader
1
40.0 documents/
min.
420:1.05
Magnetic Character Readers
1412 Magnetic Character Reader
1
950 documents/
min.
420:1.05
1419 Magnetic Character Reader
1
16.0.0 documents/
min .
42.0:1.05
. 44
Display Equipment
Display devices are a means for presenting information either to a camera or directly
to men. 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.
IBM is offering several types of display devices for use with the System/36.o (although
none is currently available for use with the small-scale Model 2.0).
The 225.0 Display Unit can provide dynamic visual displays of charts, graphs, and
alphameric characters on a 12-inch-square screen on the face of a 21-inch cathoderay tube. Up to 3,848 characters can be displayed. Optional features provide for
buffering, keyboard data entry, and use of a "light pen." The 225.0 is designed for
local use only (within 2,.0.0.0 cable-feet of a System/36.o).
The 226.0 Display Station can be locally connected to a System/36.o via an input-output
channel or remotely connected via a 27.01 Data Adapter Unit. The 226.0 provides alphanumeric display, on a 4-inch-by-9-inch screen, of up to 96.0 characters at a peak rate
of 2,56.0 characters per second when connected locally. Remote transmission is at
either 12.0 to 24.0 characters per second, depending on the data set used. The 226.0
cannot be used to display charts or graphs.
The 228.0 Film Recorder, 2281 Film Scanner, and 2282 Film Recorder/Scanner provide the capability to prepare or read charts, graphs, and alphameric data on 35-millimeter unsprocketed film. Alphameric information can be written directly, but a complex program is necessary to read such data. The recorders contain complete facilities for recording data on film, developing an image, and viewing the processed image.
The film units share a controller with, and can work in conjunction with, the 225.0 Display Unit. In combination, these units provide the capability for displaying non-alphameric input and output data. Information can be entered or extracted directly in piCtorial or graphical form, giving engineers, scientists, and other interested personnel
an opportunity to work directly with the computer •
• 45
Data Communications Equipment
The IBM 27.01 Data Adapter Unit makes possible direct connection of a wide variety
of data communications equipment to a System/360. Devices that can be connected
include the 1.03.0 Data Collection System, the 1.05.0 and 1.06.0 Data Communication
Systems, the 1.07.0 Process Communication System, telegraph terminals, and the
new IBM 274.0 and 2741 Communications Terminals. Up to eight 27.01 Data Adapter
Units can be connected to either a Selector or Multiplexor Channel. Each 27.01 can
control a maximum of four start/stop communication lines or two synchronous communication lines. Line speeds up to 4.0,8.0.0 bits per second can be accommodated.
The 27.01 operates in a half-duplex mode. Numerous special adapters are available
for connecting the 27.01 to the various types of communication facilities and terminal
equipment. These adapters provide the necessary bit-byte conversions, interfaces,
and control circuits.
(Contd.)
8/65
420:011.450
INTRODUCTION
.45
Data Communications Equipment (Contd.)
The 2702 Transmission Control must be connected to a Multiplexor Channel, permitting on-line connection of various low-speed communication terminals to a
System/360 via private or common-carrier transmission facilities. A 2702 can
handle up to 31 half-duplex lines operating at speeds up to 600 bits per second;
inclusion of lines with speeds above 180 bits per second reduces the total number of
lines that can be connected. The 2702 operates in a start/stop mode, and data transmission is serial by bit. The 2702 accomplishes all necessary bit-byte conversion,
data control, and interfacing functions. One 8-bit buffer is provided for each line.
Characters from incoming messages are interleaved and assembled in processor
storage, so the 2702 imposes no restrictions upon message length.
The characteristics of the recently-announced 2703 Transmission Control are similar
to those of the 2702. The two units differ only in the number of terminal lines that
can be connected and in certain other hardware features. A 2703 can handle up to
176 half-duplex lines operating at speeds up to 180 bits per second, or up to 76 halfduplex lines operating at speeds up to 600 bits per second. A four-character buffer
is provided for each line.
Two Audio Response Units - the 7770 and the 7772 - provide recorded human-voice
responses to telephones inquiries. The 7770 has a limited vocabulary (32 to 128
words) and can service up to 48 communication lines. The newer 7772 uses a different technique to generate the audio responses, has a virtually unlimited vocabulary, and can service a maximum of 8 lines .
. 46
System Configuration
The System/360 is highly modular, largely upward- and downward-compatible, and
allows a broad range of peripheral devices and supporting control units. Although the
overall configuration rules are quite complex, most of the input-output devices can be
connected to most of the System/360 processor models. Only the small-scale Model
20 is significantly restricted in its configuration possibilities.
Processor core storage capacities can range from 4, 096 bytes in Model 20 to over
one million bytes in Models 65 and 75, as shown in Table II. In System/360 Models
50, 65, 67, and 75, up to 8 million bytes of supplementary 8-microsecond 2361 Core
Storage can be added.
Peripheral devices and their controllers are connected to Model 30 and larger systems
through input-output channels of various types and capacities. The standard Multiplexor Channel is included in every Model 30, 40, and 50 Processing Unit and can control up to 256 low-speed I/O devices. High-Speed Multiplexor Channels, capable of
controlling up to eight high-transfer-rate devices, are also available for these models.
Selector Channels can also be used with Models 30, 40, and 50 to provide direct control of one high-speed input-output operation at a time; the maximum number of
Selector Channels is two in Models 30 and 40, and three in Model 50. Models 65 and
75 use specialized Multiplexor Channels (controlling up to 196 devices) and Selector
Channels (up to six per processor). Report Section 420:031 includes a detailed list
of System/360 peripheral devices and the system configurations required to support
them .
• 47
Simultaneous Operations
An IBM System/360 Central Processor (except for the small-scale Model 20) can
concurrently execute:
•
One machine instruction; and
•
Up to six high-speed input-output operations (one per Selector Channel);
and
•
Multiple slower input-output operations via a Multiplexor Channel.
Detailed information on the number of channels of each type that can be connected,
and their data rate capacities, is presented in Report Section 420:11L
In general, the relationships between System/360 peripheral devices and data channels
are determined at installation time and cannot be altered under program control,
although optional features permit switching magnetic tape and random access units
between two channels. Because it is not possible to assign any free data channel to
serve any peripheral device (as in some currently-available systems), the number of
operations that can actually occur simultaneously will in many cases be considerably
less than the theoretical maximum.
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
420:011.500
.5
IBM SYSTEM/360
SOFTWARE
The current organization of software for use with the IBM System/360, Models 30,
40, 50, 65, and 75, is grouped around three fully-integrated operating systems.
Each system offers a wide choice of control program services and language translators,
providing virtually tailor-made software for all levels of System/360 computer configurations. Users of System/360 models that include at least 32K bytes of core storage
:in a direct-access-device environment can use the facilities of the Operating System/360.
The Basic Operating System/360 can be selected by installations that have at least 16K
bytes of core storage or, alternatively, 8K bytes when direct-access devices are available. The Basic Programming Support package can be used by any installation that includes at least 8K bytes of core storage.
The control facilities of the Operating System/360 and the other executive systems are
modularly designed, permitting supervisory services of greatly varying sophistication
to be selected. When the Operating System/360 is restricted to an environment of 32K
bytes of core storage, it offers such services as input-output control, interrupt supervision, stacked-job processing, remote inquiry control, and operator communications.
In an environment of 256K bytes of core storage, the Operating System/360 can provide
many additional services, such as full multiprogramming control, multiprocessing
capabilities, remote control of the computer installation, and supervision of time-shared
operations. Report Section 420:191 summarizes the Operating System/360's various
control program services and specifies the main and aUxiliary storage requirements for
these services.
Multiprogramming in the System/360 is facilitated by the hardware interrupt system
and the control programs of the operating systems. Various degrees of multiprogramming are possible, and the core storage requirements increase for each higher
form. In the simpler forms, data transcription routines can be overlapped with the
processing of a main program, and remote inquiries can interrupt the main processing
program in order to obtain possibly-unrelated on-line information. More complex
multiprogramming, such as running two or more main programs at the same time, is
also possible, but the requirements in terms of peripheral devices and core storage are
extensive.
Within each of the three available operating systems, several different versions of
many of the language and utility programs are offered. These versions are constructed
at various "program design levels," according to. the amount of core storage that the
programs themselves require. However, the actual operation of a given program will
require more main storage than the amount indicated by the program design level because nearly all System/360 programs share core storage with resident control routines.
Table V indicates the range of support programs offered by IBM for use with each of
the three operating systems. The year and quarter in which delivery is expected are
listed for each program; e. g., 4Q-65 means "fourth quarter of 1965." Note that software support for the System/360 Models 20, 44, 67, and the 90 Series is not considered
in this listing of program availability, since these models and their respective software
support differ considerably from the principal, program-compatible group of System/360
models.
Of the many System/360 software facilities, the new PL/I language has stirred the
most comment. PL/I (originally called NPL, and then MPPL) represents a joint
IBM/SHARE effort to develop a programming language that will be suitable for both
business and scientific applications. As described in the IBM language manual, PL/I
offers programming facilities that are more powerful and extensive than those of any
other currently-available computer compilation language. Significant additions to this
new language include facilities to check out untested programs, to process selfidentifying input-output files, to select various random-access processing methods,
to modify and parameterize programs Immediately before compilation, to perform
asynchronous "off-line" processing, and to dynamically allocate storage areas for
program segments and data blocks. The true value of the PL/I language and its imposing facilities cannot be accurately assessed until individual P L/I compilers become
available and their operational characteristics become kIi.own.
The IBM System/360 is a powerful but complex computer system. Its software shares
these characteristics. Programming at the assembly-language level is complex, especially if the full power of the System/360 is to be utilized. The newly-developed PL/I
language could facilitate programming by assuming the burden of adapting simplified
coding to advanced hardware facilities; but the price of the powerful compiler services
in terms of decreased compiler and object program performance could be high. The
COBOL and FORTRAN languages offer the standard programming facilities and more;
(Contd.)
8/65
420:011.501
INTR ODUcr ION
TABLE V: SOFTWARE AVAILABLE FOR THE IBM SYSTEM/360
BASIC OPERATING
SYSTEM/360
OPERATING
SYSTEM/360
PROGRAMS
TO BE SUPPLIED
Program Design Level
12-18K
44K
Supervisory Control
Functions
4Q-65
FORTRAN IV
4Q-65
COBOL
4Q-65
2Q-66
-
-
1Q-66
Assembler
4Q-65
Sort/Merge
4Q-65
Report Writer
2Q-66
PL/r
-
Autotest
TESTRAN
4Q-65
Data Transcription
1Q-66
.5
Program Design Level
BASIC PROGRAMMING SUPPORT
Program Design Level
8-10K
200K
4K
10K
4K
-
2Q-66
3Q-65
4Q-65
3Q-65
4Q-65
-
2Q-66
4Q-65
1Q-66
-
3Q-65
3Q-66
-
4Q-65
2Q-66
3Q-65
4Q-65
1Q-65
-
-
3Q-65
4Q-65
3Q-65
3Q-65
4Q-65
4Q-65
-
3Q-65
4Q-65
-
-.
-
-
4Q-65
4Q-65
-
4Q-65
3Q-66
3Q-65
-
SOFTWARE (Contd.)
but the restricted versions at the smaller design levels will tempt a move to higher,
more powerful program levels at the expense of increased hardware costs. In summary,
System/360 software offers extremely comprehensive coverage of the standard programming support functions; but the very extent of its modular facilities can lead the incautious user to buy more hardware than he really needs .
•6
COMP ATIBILITY
.61
Program Compatibility Within the System/360
IBM emphasizes the high degree of program compatibility, in both the upward and
downward directions, among the following models of the System/360: Models 30, 40,
50, 65, and 75. Among these five 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.
The compatibility rule does not apply to "invalid programs" (programs that violate the
specifications in the programming manual) or to the handling of machine malfunctions.
These limitations seem to mean 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 possible 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.
For the reasons indicated elsewhere in this Introduction, Models 20, 44, 67, and 92
will not be fully program-compatible with any of the other System/360 processors at
the machine-language level. In many cases, however, it will be possible to achieve
program compatibility through reassembly or recompilation, with little or no need
for manual alterations of the programs.
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
420:011.620
.62
IBM SYSTEM/360
Program Compatibility With Earlier IBM Computers
IBM now offers an extensive series of Compatibility Features, or "emulators," that
enable certain models of the System/360 to run programs written for second-generation
IBM computer systems. The earlier IBM computers whose programs can be run
by each model of the System/360 (when properly equipped) are as follows:
System/360 Model
Model
Model
Model
Model
Model
20:
30:
40:
50:
65:
..•••••••...••.•
••••.•.•.••••••.
••••••••••.•••..
•.....•..••..•••
••••.••...•.•...
Model 75: . • . . . • . . . . . . . . . .
Systems Emulated
none.
1401*1440/1460, 1620.
1401/1460, 1410/7010.
1410/7010, 7070/7074.
7070/7074, 7080, 709/7040/7044/
7090/7094/7094 IT.
none.
Emulation, in general, requires a System/360 with an equivalent array of peripheral
equipment, more proce~sing power, and more core storage than the second-generation
system to be emulated. The functions of most of the common peripheral devices (e. g. ,
card readers and 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 several types of disc
storage) cannot. Time-dependent programs and programs not written in accordance
with'IBM programming manuals, when emulated, may yield results which differ from
those obtained on 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 Compatibility Feature. Nevertheless, it is likely that most users
of second-generation IBM computers will be able to run most of their programs on a
System/360 with little or no need for immediate reprogramming. For details on the
capabilities, performance, and limitations of each Compatibility Feature, please refer
to Sections 420:131 through 420:136 of this report.
The principal value of the Compatibility Features is that they enable users of secondgeneration IBM computers to spread the task of reprogramming for the System/360
over an extended period of time. In nearly every case, the emulation mode will involve
,significant additional equipment costs and will fall far short of fully utilizing the performance capabilities of the System/360. * Therefore, for maximum efficiency, most
users will want to recode all of their principal applications for the System/360 as soon
as possible. Unfortunately, the cost of the additional core storage and features required for emulation must be borne until all of the user's programs have been recoded.
* A significant exception is the 1401/1440/1460 Compatibility Feature for Model
30, which utilizes an all-hardware simulation technique and achieves relatively
efficient use of the Model 30's capabilities, with no need for additional core
storage beyond the requirements of the programs being run. All of the other
Compatibility Features use a combination hardware-software approach and
require additional core storage (up to 262K bytes) to hold the associated simulation routines.
8/65
,/
420:021.100
IBM System/360
Data Structure
DATA STRUCTURE
·1
STORAGE LOCATIONS
.2
Name of Location Size
Byte:
8 data bits +
1 parity bit
Word:
4 bytes
Halfword:
Double word:
General registers:
2 bytes.
8 bytes.
Floating-point
registers:
Representation
Purpose or Use
Type of Information
basic addressable
storage unit; holds
1 character or 2
"packed" decimal
digits.
basic fixed-point
binary operand
length.
Alphameric character: 1 byte.
Decimal digit: . . . . . . 4 bits; packed 2 digits per
byte.
Fixed-point binary
operand: . . . . . . . . 1 word (or 1 halfword in most
instructions) .
Floating-point operand
(short): . . . . . . . . . 1 word; 24-bit fraction* and
7-bit hexadecimal exponent.
Floating-point operand
(long): . . . . . . . . . . 1 word; 56-bit fraction* and
7-bit hexadecimal exponent.
Decimal operand:. . . . 1 to 16 bytes (i. e., i to 3 i
digits plus sign).
Instruction: . . . . . . . . 2, 4, or 6 bytes (specifying
0, 1, or 2 core storage
addresses, respectively).
32 bits each
fixed-point accumulators, baseaddress registers,
or index registers.
64 bits each
floating-point accumulators.
Row (magnetic
tape):
8 data bits
Column (punched
cards):
12 positions
INF ORMATION FORMATS
holds 1 byte.
holds 1 character.
Note: 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, mustbe a multiple of the
length of the field in bytes. This restriction
is essential for efficient operation of the
larger Processing Units, which access up
to eight bytes in parallel, and the same restriction has been applied to the smaller Processing Units 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.
* The unusual floating-point representation used
in the System/360 (hexadecimal rather than
binary exponents) makes the effective precision
three bits shorter than the actual length of the
fractional part.
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
420:031.100
IBM System/360
System Configuration
SYSTEM CONFIGURATION
The overall configuration rules of the System/360 are quite complex. Every system
has at least one Processing Unit and at least one bank of main (processor) core storage, which
is an integral part of the Processing Unit in Models 20, 30, 40, and 50, and a separate unit in
the larger models. The processor storage capacities available for the various System/360
models are summarized on page 420:041.100. In addition, up to 8,388,608 bytes of 8-microsecond core storage are available for Models 50, 65, and 75 (see Section 420:042, 2361 Core
Storage).
Peripheral devices are connected to Model 30 and larger systems through various
types of input-output channels. The maximum number and types of channels available for each
of these models are listed in Table I. The peripheral devices currently available for System/360
and the rules governing their connection to each model are presented in Table II. In Table In,
the minimum permissible configurations for Model 30 and larger systems are shown. These
minimum configurations are necessary, according to IBM, to permit standard diagnostic
routines to be used for maintenance and system checkoutpurposes.
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. The
peripheral devices available for Model 20 are also listed in Table II. Model 20 configurations
are subject to the following restrictions:
•
•
Only one peripheral device of any given type can be connected.
A 2560 Multi-Function Card Machine and a 2520 Card Punch cannot
be used in the same system.
•
The maximum configuration possible is three card read stations, two
punch stations, one document print station, one magnetic character
reader, one line printer, and one 2415 Magnetic Tape Unit and Control
(any model) .
For diagrams and prices of representative standard configurations of the System/360,
as defined in Section 4:030 of the Users' Guide, see the System Configuration sections of the
subreports on the individual models:
Model 20: •.••.•...•••...•.••.. Section 422:031.
Model 30: ...•...•••••••..•.... Section 423:031.
Model 40: ••••.•..••.........•• Section 424:031.
Model 50: • • • • • • . . . . . . . . . . . . . . . Section 425:031.
Model 65: •.•....•...•....••..• Section 426:031.
Model 75: .•••••..•.•••..•.•••. Section 428:031.
Configuration rules and representative system configurations for the large-scale
System/360 Model 67 are shown in Section 427:031.
TABLE I: SYSTEM/360 I/O CHANNELS
Standard Multiplexor Channel
Model
30
Model
40
Model
50
Model
65
Model
75
1
1
1
-
-
2
1
2
1
-
-
3
1
1
Optional" Channels
Standard Selector (1)
High Speed Multiplexor (1)
High Speed
2860 Selector
2870· Multiplexor
-
--
-
-
-
-
-6
-6
1
1
196
-
Maximum Number of Subchannels
per MultiElexor Channel
Standard Multiplexor (2, 3)
32, 96, or 224
High Speed Multiplexor (2)
2870 Multiplexor (3, 4)
4
. (1)
(2)
(3)
(4)
-
16, 32, 64, or 64, 128, or 256
128
4 or 8
4
-
-
196
A High Speed Multiplexor Channel pre-empts one standard Selector Channel.
When a High Speed Multiplexor Channel is included in a system, the maximum
number of multiplexor subchannels is reduced by either four or eight.
The Standard Multiplexor Channel with 128 or fewer subchannels includes 8
shared-path subchannels. The Standard Multiplexor Channel with more than
128 subchannels, and the 2870 Multiplexor Channel, include !!Q. shared-path
subchannels.
Four of these subchannels are optional Selector Subchannels, each capable of
handling 8 control units (maximum of 16 peripheral devices), one at a time.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
420:031.101
IBM SYSTEM/360
TABLE II: SYSTEM/360 PERIPHERAL DEVICES
Cooneetion to I/O Channel (11)
Peripheral
Device
Controllers
per System,
System/360 Model
Controller
20
30
40
50
Maximwn
65
Multiplexor
Subehannels
per
Controller (1)
Peripheral Devices per
Controller. Maximum
75
Random Access
2302 Disk Storage
The basic 2841 can control up to eight
2311 Disk Drive
2311 access mechanisms. Optional
features permit the 2841 to control up
2B41
2321 Data Cell Drive
-
ms
ms
ms
sx
sx
8/eh
IS
to eIght access mechanisms in any
combination of the devices listed, and an
additional eight 2302 access mechanisms.
7320 Drum
See the individual report s ectioDs for the
number of access mechanisms per device.
2301 Drwn
2820
-
-
-
h
s
s
8(2)
2314 Direct Access
Storage
selfcontained
-
-
(3)
s
s
s
-
ms
ms
ms
msx msx
4 2301 Drums.
8/eh
-
8/eh
1
The 28211s available in several models
and, with special features, can control
up to three 1403 printers, one 2540 and
8 on-line disc drives per 2314 unit.
Punched Card
2540 Card Read
Punch, Model 1
2821
up to three 1403 printers (any model). or
one 2540 and one 1404 printer.
1442 Card Read
Punch, Model N1
1442 Card Punch:
Model 5
Model N2
selfcontained
selfcontained
self2520 Card Read
contained
-
ms
a
-
-
-
-
-
ms
ms
ms
msx msx
Punch:
Model Al
Model B1
-
-
-
-
-
ms
ms
ms
msx msx
selfcontained
self-
selfcontained
Model B1, B2
selfcontained
2560 MFCM
-
a
contained
2501 Card Reeder:
Model AI, A2
msx msx
contained
contained
Model B2, B3
ms
selfself-
2520 Card Punch:
Model A2, A3
ms
self-
-
a
-
-
-
-
-
ms
ms
ms
msx msx
-
8/eh
1
1
1
-
1
8/ch
1
1
1
-
1
Bleh
1
1
1
-
1
B/eh
1
1
a
-
-
-
-
1
-
1
-
ms
ms
ms
msx msx
8/eh
1
1
a
-
-
-
-
1
-
1
contained
-
Printers
1403, Models 2, 7.
2821
a (4) ms
ms
ms
msx msx
8/eh
Nl
1403, Model 3
1
See 2540 Card Read Punch.
2821
-
ms
ms
ms
msx msx
8/eh
1
See 2540 Card Read Punch.
1404, Model 2
2B21
-
ms
ms
ms
8/eh
1
See 2540 Card Read Punch.
1443, Model N1
self-
-
ms
ms
ms
msx msx
8/eh
1
1
-
ms
-
-
-
-
8/eh
1
1
a
-
-
-
-
-
1
-
1
2822
-
ms
ms
ms
-
-
8/ch
1
1
2803, 2804,
-
ms
ms
rns
sx
sx
8/ch (5)
IS/ch
8
ms(6) ms(6) ms(6 s
s
S/ch
18
S
- -
8/ch
IS
6
contained
1445, Model N1
selfcontained
2203, Model Al
selfcontained
-
-
Punched PaEer TaEe
267l Paper Tape
Reader, Model 1
M!E!!etic
T~e
2400 Series
2403, 2404
7240 Model 3
2802
-
2415
self-
r (7) rns
contained
-
-
(Contd. )
7/65
SYSTEM CONFIGURATION
420:031.102
TABLE II: SYSTEM/360 PERIPHERAL DEVICES (Contd.)
Connection to I/O Channel (11)
Controllers
Peripheral
Device
per System,
System/360 Model
Controller
20
30
40
50
Maximum
Multiplexor
Subchannels
per
Controller (1)
Peripheral Devices per
Controller, Maximum
75
65
DlsEla:.: Units
2250 Display Unit,
Modell
-
selfcontained
ms
rns
ms
sx
sx
8/ch
1
1
The 2840 Control, with the approprIate
2250 Display Unit,
Model 2
number of multiplexor features, can control
up to 8 2250 Model 2 Units or up to 4 Film
2280 Film Recorder
2840
-
rns
rns
ms
ax
} 2848
-
ms
ms
rns
rnsx rnsx
2281 Film Recorder
sx
8/ch
IS
8/ch
17 or IS
Units (any type). The maximum nwnber of
Display Units that can be connected in combtnatlon with 1, 2, or 3 Film Units (any
type) is 5, 4, or 1, respectively.
2822 Film Recorder /
Scanner
2260 Display Unit:
Modell
Model 2
Model 3
9 or IS
25 or 2S
24
16
8
Audio ResEonse Units
7770
7772
selfcontained
self-
contained
-
m
m
m
m
m
B/ch
48 max.
up to 4B lines.
-
m
rn
rn
m
rn
B/ch
8 max.
up to Slines.
QEticai Readers (B)
1231 Optical Mark
Page Reader,
Model N1
selfcontained
-
rns
-
-
-
-
1
1
1
1285 Optical Reader,
Modell
selfcontained
-
m
m
-
-
-
B
1
1
141B Optical Charaeter Reader I
Models 1. 2, 3
self-
-
m
-
-
-
-
1
1
1
142B Alphameric Optical Reader,
Models 1, 2, 3
selfcontained
-
m
-
-
-
-
1
1
1
1412 Magnetic Charaeter Reader I
Modell
selfcontained
-
m
-
-
-
-
1
1
1
1419 Magnetic Charaeter Reader.
Modell
selfcontained
r
rns
ms
-
-
-
2(9)
1
1
selfcontained
-
ms
ms
ms
rns
ms
B/ch
4 max.
up to 4 lines.
2702 Transmission
Control
selfcontained
-
m
m
m
m
m
B/ch
31 max.
up to 31 lines.
2703 Transmission
Control
selfcontained
-
rn
rn
m
m
m
B/ch
176 max.
up to 176 lines.
-
-
-
-
1
contained
MmCR Readers (B)
Data Communications
2701 Data Adapter
Unit
2073 Communications
Adapter
-
r(10) (1)
(2)
(3)
(4)
(5)
(6)
\
(7)
(B)
(9)
(10)
(11)
©
-
1 line.
Applicable only when controller is connected to a Multiplexor Channel. The
symbol liS" indicates a shared-path subchannel.
One 2820 per Model 50 systemj two per 2860 Modell (1 Selector Channel);
three per 2B60 Model 2 (2 channels); four per 2B60 Model 3 (3 channels).
Can be connected to first Selector Channel only.
Connection of these models of the 1403 Printer to a Model 20 system requires
only a special attachment device, and not the 2821 Controller. Only one
printer can be connected to a Model 20 system.
Dual-channel controllers require one control unit position on each of two channels.
When a 7340 Hypertape unit is connected to a Multiplexor Channel (or to the
Selector Channel of a Model 30), only the 170,000 bytes/sec transfer rate can
be accommodated.
Only one 2415 (any model) can be connected to a Model 20 system.
Each of these devices, except the 1419 when attached to a Model 20 and the 1231 J
requires a special System/360 adapter.
6nly one 1419 is allO\ved in a Model 20 system.
This adapter cannot be incorporated in a Model 20 system with the Universal
Character Set Feature for a 1403 printer.
The symbols in theBe columns indicate the possible methods for connecting each
controller to each System/360 Model.
a - directly to any Model 20 Processor, by means of a special attachment.
r - directly to Model 20 Processor, Model B2, C2 J or D2, by means of a
special attachment.
m - via a standard Multiplexor Channel or a 2870 Multiplexor Channel.
s - via a standard Selector Channel or a 2860 Selector Channel.
x - via a selector subGhannel of a 2870 Multiplexor Channel.
h - via a High Speed Channel.
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360
420:031.103
TABLE III: SYSTEM/ 360 MINIMUM CONFIGURATIONS (1)
INPUT
DEVICES
OUTPUT
DEVICES
7/65
Model
30
(2)
Model
40
(2)
Model
50
(3)
Model
65
(4)
Model
75
(4)
1442
2501
2520
2540
Card Read Punch
Card Reader
Card Read Punch
Card Read Punch
x
x
x
x
x
x
x
x
2400
2415
7340
2311
Series Magnetic Tape
Magnetic Tape
Hypertape Drive
Disk storage Drive
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
Console Typewriter
1403 Printer
1404 Printer
1443 Printer
1445 Printer - MICR
x
x
x
(1)
Each processing unit in a system coniiguration requires access to at least one of
the inPut devices and at least one of the output devices marked with "x" in the
above table. Access can be direct (to unlta conneoted to the processor) or indirect
(to units connected to another processor).
(2)
When access to a minimWl1 I/o unit is indirect only, aU processing units in the
conilgu~ation must have at least 32,768 bytes of processor storage.
(3)
Model 50 must have direct access via a Selector Channel to the minimum input unit.
(4)
When access to the minimum input unit is indirect only, the processing units must
be interconnected via the Channel-to-Channel Adapter.
420:041.1 00
IBM System/360
Internal Storage
Processor Storage
INTERNAL STORAGE: PROCESSOR STORAGE
.1
GENERAL
.11
Identity: . • • • • • • • . . • . . . . • • • . . . • . • . . . . processor storage for Model 20 series:
contained in 2020 Processing Unit,
Models B, C, D.
processor storage for Model 30 series:
contained in 2030 Processing Unit,
Models C, D, E, F.
processor storage for Model 40 series:
contained in 2040 Processing Unit,
Models D, E, F, G, H.
processor storage for Model 50 series:
contained in 2050 Processing Unit,
Models F, G, H, 1.
processor storage for Model 65 series:
contained in 2365 Processor Storage,
Models 1, 2.
processor storage for Model 75 series:
contained in 2365 Processor Storage,
Model 3 •
• 12
Basic Use: •••••••••••••.•••.••...... working storage .
• 13
Description
The currently available models of System/360 differ primarily in main (processor) core
storage capacity, speed, and number of bytes accessed per cycle. Main storage characteristics of the available models can be summarized as follows:
System/360 Model
Capacity, bytes
Model
20
4,096
8,192
16,384
32,768
65,536
131,072
262,144
524,288
1,048,576
B20
C20
D20
Cycle Time, /1-sec
3.6
-
Bytes Accessed per Cycle 1/2
Cycle Time per Byte,
/-Isec
7.2
Model Model Model Model Model
75
30
40
50
65
-
C30
D30
E30
F30
-
-
-
D40
E40
F40·
F50
040 ~ G50
H:~,O ...
H50
-
-
065
H65
165
J65
-
H75
175
J75
-
-
1.5
2.5
2.0
0.75* 0.75*
1
2
4
8
1.5
1.25
0.50
0.094* 0,094*
8
* Effective cycle time in Models 65 and 75 is somewhat faster due to interleaved
accessing of core storage.
Main storage is physically integrated with the Processing Unit in Models 20, 30, 40, and
50, and is housed in separate cabinet modules in Models 65 and 75. The logical structure
of the system is the same in either case.
Each byte consists of eight data bits and one parity bit, and each byte is directly addressable. Internal storage addressing is binary (though decimal or symbolic addresses
© 1965 AUERBACH Corporation and AUERBACH Infa, Inc.
7/65
420:041.130
IBM SYSTEM/360
· 13
Description (Contd.)
are used in the assembly programs to facilitate coding). The eight data bits of a byte
may represent binary, alphameric, or packed decimal (two decimal digits per byte) data.
The Storage Protection feature (optional in Models 30 and 40, standard in the larger
models) prevents the contents of specified 2, 048-byte blocks of core storage from being
altered as a result of program errors or misguided input data. A 4-bit "storage key"
is associated with each 2, 048'-byte block, and a 4-bit "protection key" is supplied with
the data to be stored. Detection of a mismatch between the two keys results in a program interrupt. As many as 15 independent programs can be protected at anyone time.
Each protected program can occupy any number of blocks, and the blocks do not need
to be contiguous. Storage Protection is an essential safeguard wherever more than
one program is to be loaded and run at the same time.
The Shared Storage feature (currently available only for Model 50) permits the main
storage units of two Model 50's to be shared and addressed by either Processing Unit
as a single main storage.
When blocks of data must be moved from one area of main storage to another, this can
be most effiCiently accomplished by means of the Move instruction, which operates in
the storage-to-storage mode and can move up to 255 bytes per application of the instruction. An alternative method is to use a loop with multiple-register load and store instructions. Paragraph. 73 shows the effective transfer rates for both methods .
. 14
Availability: •..•....•••.•••..••..••.. see Central Processor
section, Paragraph 420:051.13 .
. 15
First Delivery: .•••..•.•.••...•....... see Central Processor
section, Paragraph 420:051.14 .
• 16
Reserved Storage
Purpose
Number of locations
•2
Model 20Internal processor
control: ••..••••••.•••••••••...... 144 bytes.
Model 30 Index/arithmetic/logic registers: ..•••..... 64 bytes.
Arithmetic registers: . • . • • . . . • . . • . . . . . . 32 bytes (for floating-point
registers) .
Models 30,40, 50, 65, 75 Initial program reading data:: • • • . . . • • . • . • • 24 bytes.
Program control: •••.•••••..•........• 68 bytes.
I/O control: • • . . . • . . . • • . • . . . • . . . . . . . 28, bytes.
Unused: . . . • . . . • . . . • • . . • • . . . . . . . . . . 8 byt~s.. _
Diagnostic scan-out area: ••.••......... :variable.
£"} /'
PHYSICAL FORM
/:-' .
.21
Storage Medium: ••...•..••...•...•.... magnetic core .
• 23
Storage Phenomenon: . . . • . . • . • . . . . . • . . . . direction of magnetization .
. 24
Recording Permanence
.241
. 242
• 243
• 244
• 245
Data erasable by instructions: . . . • . • . . . . . • .
Data regenerated constantly: . . • . . . . . . . • . . .
Data volatile: •••••...••••••••.••.•....
Data permanent: ..•••••••.•.•.•.•••....
Storage changeable: ••••••••••••..••••..
• 28
Access Techniques
(r
yes .
no .
no.
no .
no.
• 281 Recording method: •••••.••••.••.•.•••.. coincident current.
• 283 Type of access: • • • • • • • • • • • . . • • . . . . . . . . uniform.
(Contd. )
7/65
420:041.290
INTERNAL STORAGE: PROCESSOR STORAGE
.29
Potential Transfer Rates
.2.92 Peak data rates -
System/360 Model
20
30
40
50
65
75
Unit of data,
bytes/access
1/2
byte
1
byte
2
bytes
4
bytes
8
bytes
8
bytes
Cycling rate,
cycles/second
278,000
667,000
400,000
500,000
1,333,000*
1,333,000*
Conversion factor,
data bits/byte
8
Data rate, bytes/sec
8
0.14
million
0.67
million
8
8
2.0
million
0.80
million
8
8
10.7
million*
10.7
million*
* Effective cycle time in Models 65 and 75 is somewhat faster due to interleaved accessing
of core storage •
.3
DATA CAPACITY
.31
Module and System Sizes
System/360
Model
B20
C20
D20
C30
D30
E30
F30
D40
E40
F40
G40
H40
F50
G50
H50
G65
H65
165
J65
H75
175
J75
Processing· Unit
No.
Model
2020
2020
2020
2030
2030
2030
2030
2040
2040
2040
2040
2040
2050
2050
2050
2065
2065
2065
2065
2075
2075
2075
B
C
D
C
D
E
F
D
E
F
G
H
F
G
H
G
H
I
J
H
I
J
Processor Storage
Unit
No.
Model
No. of
Storage
Modules
Included in Processing Unit
2365
2365
2365
2365
2365
2365
2365
1
2*
2*
2'"
3*
3**
3**
1
1
2
4
1
2
4
Total Storage
(bytes)
4,096
8,192
16,384
8,192
16,384
32,768
65,536
.W,3l:S4
32,768
65,536
131,072
262,144
65,536
131,072
262 144
131,072
262,144
q24,288
1 048 576
262,144
524,288
1,04.8,576
Words or
4-byte Instructions
1 024
2,048
8,192
2,048
4,096
8.192
16,384
4,096
8,192
16,384
32,768
65,536
16,384
32,768
65 536
32,768
65,536
131,072
262 144
65,536
131,072
262,144
* Storage words are interleaved by pairs to improve sequential access rate.
** Storage words are four-way interleaved to improve sequential access rate .
. 32
Rules for Combining Modules: . . . . . . • . . • . . . see chart above.
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
420:041.400
IBM SYSTEM/360
.4
CONTROLLER: ••••..••••.•••••••••.•. no separate controller; all required
control facilities are included
in processor and/or storage
modules.
.5
ACCESS TIMING
.52
Simultaneous Operations: ••••••••.••••••• none •
• 53
~ccess
Time Parameters and Variations
• 531 For uniform access System/360 Model
20
30
40
50
65
75
Cycle time,
microseconds
3.6
1.5
2.5
,
2.0
0.75
0.75
Unit of data,
bytes/access
1/2
1
2
4
8
.6
CHANGEABLE STORAGE:
•7
PERFORMANCE
.72
Transfer Load Size
8
•.••...•...••.• none .
With self: •••••••..••••••••••••.•.••. 1 to 255 bytes using Move instruction.
Variable amount using multiple register
instructions in a loop (not Model 20) •
. 73
Effective Transfer Rate (With Self)
(All transfer rates are in bytes per second.)
System/360
Model
20:
30:
40:
50:
65G:
65H, I, J:
75:
.8
Using Multiple Register
Instructions in a Loop
62,500
321,000
390,000 v:
851,000
4,760,000
4,780,000
130,000
307,000''-760,000
3,430,000
3,560,000
?
?
ERRORS, CHECKS, AND ACTION
Invalid address:
Invalid code:
Receipt of data:
Recording of data:
Recovery of data:
Dispatch of data:
Reference to locked
area:
Reference to protected
area:
*
7/65
Using Move
Instruction
Check or Interlock
Action
check
all 8-bit codes valid.
parity check
record parity bit.
parity check
transmit parity bit.
program interrupt. *
program interrupt. *
program interrupt. *
no locked area.
check, if Storage Protection feature is
present
program interrupt.
Errors in a Model 20 system cause the processor to halt.
420:042.100
IBM System/360
Internal Storage
2361 Core Storage
INTERNAL STORAGE: 2361 CORE STORAGE
.1
GENERAL
.11
Identity:
.12
Basic Use:
.13
Description
applications. It will probably be used mainly for
storage of:
2361 Core Storage,
Models 1 and 2 (LCS).
Q
large-capacity auxiliary
core storage.
e Small, frequently-referenced files in
real-time systems where fast response
is essential.
Two models of IBM 2361 Large-Capacity Core
Storage provide up to 8,388,608 bytes of directlyaddressable core storage for System/360 Models
50, 65, and 75. Model 1 stores 1,048,516 bytes
and Model 2 stores 2,097,152 bytes. Each byte
location holds eight data bits and one parity bit.
Cycle time is eight microseconds - which is considerably slower than the main storage cycle times
in these System/360 models, but many times faster
than any of the mechanical direct-access storage
devices in IBM's line.
An optional interleaved addressing scheme between
2361 modules permits the overlapping of read/write
storage cycles in sequential operations. Thus, the
effective cycle time for sequential accesses can be
reduced to four microseconds. One Model 1 or one
to four Model 2 units can be used in systems without the interleaving feature. Two Model 1 units or
two to four Model 2 units can be used in systems
with the interleaving feature. In a Model 50 system, interleaving can be specified only if the 2361
Core Storage is shared with a Model 65 or 75
system.
Addressing of the increased storage is by direct
extension of the addressing system used for main
(processor) core storage. Each reference to the
2361 Core Storage obtains eight bytes, the amount
obtained by Models 65 and 75 Processing Units
from their main core storages. Model 50 uses only
four of the eight bytes per access, which is the
amount obtained from its main storage. Each byte
is directly addressable, and read access time is
three microseconds. IBM states that all operating
features of the 2361 Core Storage except cycle time
are the same as those of the main (processor) storage of the system.
Storage Protection is a standard feature of the 2361
Core Storage. The Shared storage special feature
permits two ProceSSing Units to address the same
2361 Core Storage units. A Model 50 Processing
Unit can share 2361's with a Model 50, 65, or 75
which has equal or greater processor (main) storage, and the larger Processing Units can share
2361's with any of the larger units having equal
processor storage.
The cost of IBM 2361 Core Storage will rule out its
use for master-file storage in most data processing
©
Directories, to permit rapid accessing
of information stored in slower directaccess storage devices (discs, drums,
or data cells).
•
Q
•
Frequently-used subroutines.
Segments of active programs which must
be temporarily moved out of main storage
in multiprogramming environments.
Large matrices, to extend the size of
scientific problems that can be handled
without the need to resort to tape or
direct-access devices for intermediate
storage.
.14
Availability: . . . . .. ?
.15
First Delivery: . . . . ?
.16
Reserved Storage:
.2
PHYSICAL FORM
.21
Storage Medium: . ..
magnetic core.
.23
Storage Phenomenon:.
direction of magnetization.
.24
Recording Permanence
.241 Data erasable by
instructions: . . . . .
.242 Data regenerated
constantly:
.243 Data volatile: .
.244 Data permanent:
.245 Storage changeable:
........
......
.. .
.28
yes.
no.
no .
no .
no.
Access Techniques
· 281 Recording method: ..
. 283 Type of access: . . . .
· 29
none.
coincident current.
uniform .
Potential Transfer Rates
· 292 Peak data rates Unit of data: . . . . .
Cycling rate: ... .
Conversion factor: .
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8 bytes.
125, 000 cycles per second.
8 data bits and 1 parity bit
per byte.
7/65
420:042.292
IBM SYSTEM/360
.292 Peak data rates (Contd.)
Data rate Model 50 Processing Unit: . . . . . . . 500,000 bytes per second*
(Model 50 uses only 4 bytes
per access) .
Models 65 and 75
Processing Units: .. 1,000,000 bytes per second*.
* The data rate can be increased by as much as
100% for sequential operations if the interleaving
feature is incorporated.
.3
DATA CAPACITY
.31
Module and System Sizes
Rules for Combining
Modules: . . . . . . . . one Modell or one to four
Model 2 units per system
without interleaving.
two Model 1 units or two to
four Model 2 units per sys..:
tern with interleaving.
.8
.5
ACCESS TIMING
. 51
Simultaneous
Operations: . . . . . . . if interleaving is specified,
sequential read/write
storage cycles can be
overlapped, reducing the
effective cycle time to 4
p,sec.
Access Time Parameter and Variations
. 531 For uniform access Read access time: •. 3.0 J,lsec.
Cycle time: . . . . . . . 8.0 J,lsec.
For data unit of: ... 8 bytes.
.6
CHANGEABLE
STORAGE: . . . . . . . none.
.7
PERFORMANCE
The manufacturer states that the number of bytes
obtained per storage access, and all other features
of large-capacity storage except cycle time, are the
same as those of the main (processor) storage of
the system.
ERRORS, CHECKS AND ACTION
Error
Check or Interlock
Action
Invalid address:
check
program
interrupt.
Invalid code:
all-8-bit codes are valid.
Receipt of data:
parity check.
Recording of
data:
Recovery of data:
record parity bit.
parity check.
Dispatch of data:
Reference to
locked area:
Reference to protected area:
7/65
CONTROLLER: ••••• no separate controller.
53
Maximum
Minimum
Storage
Storage
Identity: 2361 Model 1 2361 Model 2 four Model2's
1,048,576
2,097,152
8,388,608.
Bytes:
262,144
524,288
2,097,152.
Words:
4.
Modules: 1
1
.32
.4
program
interrupt.
program
interrupt.
transmit parity bit.
no locked area.
check.
program
interrupt.
420:043.100
IBM System/360
Internal Storage
2302 Disk Storage
INTERNAL STORAGE: 2302 DISK STORAGE
.1
GENERAL
. 11
Identity:
.12
Basic Use: .••••... auxiliary storage.
.13
Description:
. . • . . . 2302 Disk Storage Models
3 and 4.
IBM 1302 Disk Storage was announced late in 1963
as a faster, higher-capacity version of 1301 Disk
Storage; see Section 417 :043 for a comparison of the
1301 and 1302 as used in the IBM 7080 system.
IBM has renamed the 1302 Disk Storage for use
in System/360; it is now the 2302. The 2302's
comb-like access mechanisms and "cylinder" mode
of data organization makes it suitable for either
random or sequential processing applications.
Two models of 2302 Disk Storage are available for
the System/360. Model 3 contains one storage
module and Model 4 contains two storage modules.
Each module has two independent access mechanisms and a capacity of 112.14 million bytes (or
224.28 million packed decimal digits and signs).
Record length and number of records per track are
variable and user-defined. Each track has a maximum data capacity of 4, 984 bytes.
Eight random access mechanisms (four 2302 storage modules) can be connected to a 2841 Storage
Control Unit, and up to eight control units can be
connected to each System/360 input-output channel.
Each 2841 Storage Control Unit used with 2302 Disk
Storage requires a 2302 Attachment. An additional Storage feature for the 2841 permits it to control a total of sixteen 2302 access mechanisms
(eight storage modules).
Each module contains 25 discs; 45 of the 50 disc
surfaces are normally used for data storage. Each
module is served by a comb-like access mechanism
than moves horizontally between the discs. The
access mechanism contains a separate read-write,
head for eaqh of the disc surfaces. The 45 tracks,
one on each disc surface, that can be read or recorded upon when the access mechanism is in any
given position, are referred to as a "cylinder."
There are 500 tracks on each disc surface, defining
500 cylinders per module. Each of the 500 cylin- '
ders can store up to 224,280 data bytes, or 448,560
packed decimal digits. One additional disc surface
is available for use as an alternate in case difficulty
is encountered in reading or recording on any data
track.
Repositioning time for the 2302 access mechanisms
ranges from 50 to 180 milliseconds and averages
165 milliseconds. Seek operations on all mechanisms are independent, and they can be overlapped
with the single read or record operation permissible
©
in the input-output channel used with the 2302.
Disc rotation time is 34 milliseconds. Peak data
transfer rate between 2302 Disk Storage and core
storage is 156,000 bytes per second.
Each of the two access mechanisms in the 2302
operates in the same fashion as the single mechanism of the 2301; i. e., just 250 of the 500 tracks
on each disc surface are assigned to each mechanism. The two mechanisms can operate independently and both can be in motion simultaneously, but
since each is restricted to motion within its own
zone of operation. one access mechanism cannot
read a track written by the other access mecha'"
nism.
Each record may contain a key portion; a separate
"count" byte in each record is used to specify the
key length, which can be as long as 255 bytes. The
number of bytes in the data portion of the record is
specified by a two-byte count code. Thus each
record can have a count area, a key (optional), and
a data area. An additional area is included at the
beginning of each track to specify the track address,
and a one-byte code is used to indicate possible unusable -recording portions of the track.
When multiple records are present on a track, the
total amount of data that can be recorded is decreased markedly. For example, a track which
stores records consisting of a 10-byte key and 150
data bytes can hold 20 records. or only 3,000 data
bytes. In the form of a single record with a 10byte key, a single track can contain 4,954 bytes of
data.
.
A series of file commands permits any or all of the
three areas (count, key, and data) of a record in a
random access file to be searched, read, or written.
With the use of chained commands, it is possible
to read or write up to 45 successive tracks (cylinder
mode) during 45 successive disc rotations. Selection
of a track is initiated by transferring a 6-byte seek
address from the Processing Unit to the 2841. The
optional File Scan Function feature permits an
automatic search for a specific identifier or key.
The Record Overflow feature permits a single
record to overflow from one track to another track
within a cylinder. The Two-Channel Switch feature
enables a 2841 Storage Control to be switched from
. one 'channel to another under program control.
The 2841 storage Control is used to control the
2311 Disk Storage, 7320 Drum Storage, and 2321
Data Cell Drive as well as 2302 Disk Storage. The
2841 interprets and executes file commands, provides a data path and performs the necessary data
format conversions between the Processing Unit
and the direct access devices, performs checks
to ensure accurate data transfers, and furnishes
operation status information to the Processing Unit.
The 2841 strips the parity bit off each byte to be
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360
420:043.130
. 13
Description (Contd.)
.25
recorded in random access storage. The validity of
the recorded information is checked by generating a
string of 16 "cyclic check bits" and appending it to
the end of each disc record. When the.data is read,
the check bits are regenerated and compared; an unequal comparison results in a data error signal.
Parity bits are restored as the data is transferred
back into core storage.
.14
Availability: . . . . . . . ?
.15
First Delivery: .••. , 2nd quarter, 1965 (1302 with
IDM 7080 Data Processing
System).
· 16
Reserved Storage
Purpose
Number of Locations
Spare (data):
1 disc surface per
module
Format and
clocking:
Data Volume per Band of 1 Track
Characters: . . . . . . • . see "Bytes".
Digits: . . • . • . . . . . • 4,984 in zoned byte format.
9,968 in packed format.
Words: . • • . • . . . . . . 1,246.
Bytes 1 record per track: 4,984.
Multiple records per
track: .•.••. , .•• N =(4964 - L
+ 1) L
81 + 1.049L
where N = number of bytes
per track, and L = length
of data plus key in each
record, in bytes.
· 26
Bands per Physical
Unit: .. , • , , . . . . . 500 per disc surface.
.27
Interleaving Levels: . ,1; i, e., no interleaving.
· 28
Access Techniques
Locks
not normally
addressable.
· 281 Recording method: . , , magnetic heads which move
horizontally in unison on
a comb-like access arm
mechanism (2 access
mechanisms per module) ,
,283 Type of access -
4 disc surfaces per not addressmodule
able.
.2
PHYSICAL FORM
.21
Storage Medium:....• multiple magnetic discs.
Description of Storage Possible Starting Stage
.22
Physical Dimensions
Move heads to selected
track (cylinder):. , . , if new cylinder is selected.
Wait for beginning of
selected track:; .. '.. if same cylinder was previously selected.
Transfer data: .. ,.•.. no .
.222 DiscDiameter: .•.••.• " ?
Number on shaft: .•. 25 discs per module.
• 23
Storage Phenomenon: . magnetization.
,29
.24
Recording Permanence
.291 Peak bit rates Cycling rates: •.• , .1, 790,rpm (34 msec/rev).
Bit rate per track: . ,1,248,.000 bits/sec/track.
.292 Peak data rates Unit of data: '" • , , byte.
Conversion factor: . , 8 bits per byte.
Data rate: ... , . . . . 156,000 bytes per second .
312,000 packed decimal
digits per second •
· 241 Data erasable by
instructions: .....•
· 242 Data regenerated
constantly: ...••..
. 243 Data volatile: . . . . . . .
.244 Data permanent: . . . . .
• 245 Storage changeable: .:
yes.
no,
no.
no .
no.
.3
DATA CAPACITY
.31
Module and System Sizes
Potential Transfer Rates
Minimum
Storage
Identity:
Physical Units:
Modules:
Access mechanisms:
Discs:
Tracks:*
Cylinders:
Words:*
Bytes:
Packed digits:
2302-3
1
1
Maximum Storage
per Control
2302-4
1
2302-4
2
2
2
4
4
8
23 data
22,500
500
28,035,000
112,140,000
224,280,000
46 data
45,000
1,000
56,070,000
224,280,000
448,560,000
92 data
90,000
2,000
112,140,000
448,560,000
897,120,000
2302-4
4.
8.
16.
184 data.
180,000.
4,000.
224,280,000.
897,120,000.
1,794,240,000.
/'
* All data capacities are based on 45 tracks per cylinder.
(Contd. )
7/65
INTERNAL STORAGE: 2302 DISK STORAGE
420:043.320
.32
Rules for Combining
Modules: . . . . . . . • any number up to 4 modules
(standard) per 2841 Control Unit.
any number up to 8 modules
(using Additional Storage
special feature) per 2841
Control Unit.
•4
CONTROLLER
.41
Identity: ..•.••••.• IBM 2841 Storage Control
Unit with 2302 Attachment .
.445 Synchronization: .•.• automatic.
.447 Table control: ..••.. yes;· scatter-read and
gather-write facilities
are available at programmer's option. See
Section 420:111, Simultaneous Operations.
.448 Testable conditions: .. available.
busy .
not operational.
performing operation with
interruption pending.
. 42
Connection to System
.5
ACCESS TIMING
.51
Arrangement of Heads
.421 On-line: .••.•••••. see Section 420:031, System Configuration •
. 422 Off-line: .••••••••. none.
• 43
Connection to Device
.431 Devices per
controller: •.•••.• 1 to 8 access mechanisms,
using 2302 Attachment.
1 to 16 access mechanisms,
using 2302 Attachment
and Additional Storage
special feature.
.432 Restrictions: .•••.•• maximum of 8 storage access mechanisms per 2841
except when using Additional Storage special feature
described above, which permits an additional 8 2302
access mechanisms. See
follOWing table listing the
number of access mechanisms per module.
Device
No. of Access Mechanisms
2302 Disk Storage,
Model 3
2302 Disk Storage,
Model 4
2311 Disk Storage
Drive
2321 Data Cell Drive,
Modell
7320 Drum Storage
.44
2
4
1
• 511 Number of stacks Stacks per system: .. 90 to 360 (4 modules) .
90 to 720 (8 modules, using
Additional Storage special
feature).
Stacks per module: .. 90 (2 access mechanisms).
Stacks per yoke: •.• 45 (1 access mechanism).
Yokes per module: •• 2 (each has access to half
the tracks on a disc) .
.512 Stack movement: ..•• horizontally across· disc
surface to one of 250
tracks.
.513 Stacks that can access
any particular
location:. . . . . . . . . 1.
. 514 Accessible locations
By single stack With no movement: 1 track.
With all movement: 250 tracks.
By all stacks With no movement: 90 tracks per module (2
cylinders) .
90 to 720 tracks per system
(using 8 modules and
Additional Storage special
feature).
.52
Simultaneous
Operations: •.••.•• a read, write, or seek
operation using one access
mechanism can be overlapped with seek operations
taking place in any or all
other access mechanisms.
Only one read or write
operation can take place at
a time on each Selector
Channel.
.53
Access Time Parameters and Variations
1
1
Data Transfer Control
.441 Size of load 1 record per track: .• 1 to 4,984 bytes.
Multiple records
per track: ..••... N = (4964 - L
+ 1) L
81 + 1. 049L
where N = number of bytes
per track, and L = length of
data plus key in each record,
in bytes.
Cylinder mode: .•.. 1 to 45 tracks (up to
224,280 bytes).
.442 Input-output area: •••. main core storage.
.443 Input-output area
access: .••..••.. each byte.
.444 Input-output area
lockout: •••.•.•.. blocks of 2, 048 bytes can
be protected (optional on
Model 30 and 40 Processing Units, standard on
larger models).
©
.532 Variation in access time Stage
Move head to
selected track
(cylinder) :
Wait for beginning
of selected
track:
Transfer data:
Total:
1965 AUERBACH Corporation and AUERBACH Info, Inc.
Variation,
msec
Average,
msec
o or 50 to 180
165.
o to 34
34 per track
17.
34.
216 msec.
7/65
420:043.600
.6
CHANGEABLE
STORAGE: .••••.. none.
.7
PERFORMANCE
. 72
Transfer Load Size
IBM SYSTEM/360
.75
With no overlapping
of seek times: ••••• 5.4 references/sec .
With maximum overlapping of seek times: 53.5 references/sec .
Note: Based on random accessing and reading of
one 150-character record, with no updating or
rewriting.
With core storage Single track: ..•••• 1 to 4,984 bytes.
Cylinder: .•.••... up to 224,280 bytes (45
tracks) per cylinder.
.8
.73
Effective Transfer Rate
With core storage Cylinder mode,
I-way transfer: •.• 132,000 bytes per second
(including random access
and rotational delay times).
• 74
Read-Only Reference Cycle Rates
ERRORS, CHECKS, AND ACTION
Error
Check or Interlock
Invalid address:
program check on program
recorded address
control.
all 8-bit codes
are valid.
compute cyclic
record
check code
check
code.
none, except programmed read.
compute cyclic
error
check code and
signal.
compare with recorded check
code.
check
attaches
parity
bit to
each byte.
interlock
wait.
program check on program
track address
control.
Invalid code:
Receipt of data:
Update Cycle Rates
With no overlapping
of seek times: ...•• 4.0 references/sec.
With maximum overlapping of seek times: 11. 5 references/sec.
Note: Based on random accessing of one 150character record; reading, updating, and rewriting
that record; and rereading for verification of recording accuracy.
Recording of data:
Recovery of data:
Dispatch of data:
Timing conflicts:
Wrong track:
Action
/
7/65
420:044.100
IBM System/360
Internal Storage
2311 Disk Storage Drive
INTERNAL STORAGE: 2311 DISK STORAGE DRIVE
.1
GENERAL
.11
2311 Disk Storage Drive.
1316 Disk Pack.
. 12
Basic Use: . . . . . . .
.13
Description
random-access auxiliary
storage.
The new mM 2311 Disk Storage Drive is an upgraded version of the mM 1311 Disk Storage Drive
used in the mM 1401, 1410, 1440, and 1620 Data
Processing Systems. The 2311 is significantly
better than the 1311 with respect to data capacity,
access time, and data transfer rate. Like the
1311, the 2311 features rapid interchangeability of
the removable "Disk Pack" storage cartridges.
The 2311's multiple-head access mechari.ism and
"cylinder" mode of data organization make it
suitable for either random or sequential processing
methods. The 10-pound Disk Pack on each 2311
drive unit can be removed and replaced with
another in less than one minute.
The table below shows the similarities and the
differences between the 2311 and the older 1311.
1311 Disk
2311 Disk
Storage Drive
Storage
(used with 1401) Drive
Storage capacity of
1 Disk Pack:
Discs per pack:
Recording surfaces
per pack:
Tracks per disc
surface:
Data rate:
Rotation time:
Average positioning
time:
Maximum storage
capacity:
2.0 million
characters
(20 sectors
per track)
2. 98 million
characters
(1 record per
track)
7.25 million
8-bit bytes
(1 record
per track).
6
6.
10
10.
100
77,000 char /
sec.
40 msec
200.
156,000
bytes/sec.
25 mseC.
150 or 250 msec 85 msec.
20 to 30 million
characters
per system
58 million
bytes per
control unit.
Eight random access mechanisms (eight 2311 storage modules) can be connected to a 2841 Storage
Control Unit, and up to eight control units can be
connected to each System/360 input-output channel.
©
Each 2311 module holds one Disk Pack containing
6 discs, and 10 of the 12 disc surfaces are used for
data storage. Each module is served by a single
comb-like access mechanism that moves horizontally
between the discs. The access mechanism contains
a separate read-write head for each of the disc surfaces. The 10 tracks, one on each disc surface,
that can be read or recorded upon when the access
mechanism is in any given position, are referred
to as a "cylinder."
.
There are 200 tracks on each disc surface, defining
200 cylinders per module. Each of the 200 cylinders
can store up to 36,250 data bytes, or 72, 500 packed
decimal digits. Three additional tracks per surface
are available for use as alternates in case difficulty
is encountered in reading or recording on any data
track in a cylinder.
Repositioning time for the 2311 access mechanism
ranges from 30 to 145 milliseconds and averages
85 milliseconds. Seek operations on all mechanisms
are independent, and they can be overlapped with
the single read or record operation permissible in
the input-output channel used with the 2311. Disc
rotation time is 25 milliseconds. Peak data transfer rate between 2311 Disk Storage and core storage
is 156;000 bytes per second.
Record length and number of records per track are
variable and user-defined; each track has a maximum data capacity of 3,625 bytes. Each record
may contain a key portion; a separate "count" byte
in each record is used to specify the key length,
which can be as long as 255 bytes. The number of
bytes in the data portion of the record is specified
by a two-byte count code. Thus each record can
have a count area, a key (optional), and a data area.
An additional area is included at the beginning of
each track to specify the track address, and a onebyte code is used to indicate possible unusable
recording portions of the track.
When multiple records are present on a track, the
total amount of data that can be recorded is decreased markedly. For example, a track which
stores records consisting of a lO-byte key and
150 data bytes can hold 14 records, or only
2,100 data bytes. In the form of a single record
with a 10-byte key, a single track can hold 3,595
bytes of data.
A series of file commands permits any or all of the
three areas (count, key, and data) of a record in a
random access file to be searched, read, or written.
With the use of chained commands, it is possible to
read or write up to 10 successive tracks (cylinder
mode) during 10 successive disc rotations.
Selection of a track is initiated by transferring a
6-byte seek address from the Processing Unit to
the 2841. The optional File Scan Function feature
permits an automatic search for a specific identifier or key. The Record Overflow feature permits
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
420:044.130
.13
IBM SYSTEM/360
Description (Contd.)
a single record to overflow from one track to
another track within a cylinder. The Two-Channel
Switch feature enables a 2841 Storage Control to be
switched from one channel to another under program
control.
.243 Data volatile: .•••••• no .
• 244 Data permanent: .•••• no.
.245 Storage changeable: •• yes.
.25
Characters: ..••.•• see "Bytes".
Digits: .••••••.•.. 3,625 in zoned byte format.
7,250 in packed format.
Words: ••.••..•.•. 906.
Bytes (3615 _ L
)
1 record per track: •. N = 81 + 1. 049L + 1 L
where N = number of bytes
per track, and L = length
of data plus key in each
record, in bytes.
The 2841 storage Control is used to control the
2302 Disk Storage, 7320 Drum Storage, and 2321
Data Cell Drive as well as 2311 Disk Storage
Drives. The 2841 interprets and executes file
commands, performs the required conversions
between serial-by-bit and parallel-by-bit data
modes, checks the validity of the data being transferred, and furnishes status information about the
random access file units to the Processing Unit.
The 2841 strips the parity bit off each byte to be
recorded in random access storage. The validity
of the recorded information is checked by generating
a string of 16 "cyclic check bits" and appending it
to the end of each disc record. When the data is
read, the check bits are regenerated and compared; an unequal comparison results in a data
error signal. Parity bits are restored as the data
is transferred back into core storage.
While the 1316 Disk Packs are physically interchangeable among drive units of 2311 and 1311
Disk Storage, data recorded by a 1311 cannot be
read by a 2311 (even when the 2311 is used in the
mM 1401 Compatibility Mode) because of differences
in recording density.
. 26
Bands per Physical
Unit: ....•••.•.. 200 per disc surface.
.27
Interleaving Levels: .. 1; i. e., no interleaving.
.28
Access Techniques
.281 Recording method: ... magnetic heads which move
horizontally in unison on
a comblike access arm
mechanism.
. 283 Type of access Description of Stage
.29
Availability:
?
· 15
First Delivery: . . . .
?
• 16
Reserved Storage: ••. none. (Note that 3 tracks
per disc surface are held
in reserve in case of
recording problems.)
Possible Starting Stage
Move heads to selected track
(cylinder): . . . . . • . . if new cylinder is selected;
Wait for beginning
of selected track: ... if same cylinder was
previously selected.
Transfer data: . . . . . . no.
The removable 1316 Disk Packs are 14 inches in
diameter and 4 inches high. When a Disk Pack is
not mounted on a drive, the pack and its cover
combine to form a sealed container that can be
conveniently stored and transported.
· 14
Data Volume per Band of 1 Track
Potential Transfer
Rat~s
.291 Peak bit rates Cycling rates: .•...
Bits/inch/track: . . . .
Bit rate per track: ..
.292 Peak data rates Unit of data: . . . . . .
Conversion factor: ..
Data rate: .•..•..•
2,400 rpm (25 msec/rev).
1,100 approximately.
1,250,000 bits/sec/track .
byte.
8 bits per byte.
156,000 bytes per second .
•2
PHYSICAL FORM
.3
DATA CAPACITY
.21
Storage Medium: .•.• multiple magnetic discs.
.31
Module and System Sizes
.22
Physical Dimensions
.222 DiscDiameter: .•.•••.. 14 inches o. d.
Thickness: .•••••• thin discs, at half-inch
spacing.
Number on shaft: •.. 6.
Identity:
Modules:
Access mechanisms:
Discs:
• 23
Storage Phenomenon: • magnetization.
• 24
Recording Permanence
Tracks:
Cylinders:
Words:
Bytes:
Packed digits:
Cartridges on
line (Disk
Packs):
• 241 Data erasable by
instructions: ...••• yes.
• 242 Data regenerated
constantly: .••.... no.
Minimum
Storage
Maximum Storage l2er Control
12311
1
82311's.
8.
1
6 (10 active
surfaces)
2,000
200
1,812,500
7,250,000
14,500,000
8.
48 (80 active
surfaces) .
16,000.
1,600 .
14,500,000.
58,000,000.
116,000,000.
1
8.
(Contd. )
7/65
INTERNAL 'STORAGE: 2311 DISK STORAGE DRIVE
. 32
Rules for Combining
Modules:. • • . . . ..
.4
CONTROLLER
. 41
Identity:.........
.42
Connection to System
any number up to eight 2311
Disk Storage Drives per
2841 Storage Control Unit.
IBM 2841 Storage Control
Unit
.421 On-line: . . • . . . . . . . see Section 420:031,
System Configuration.
.422 Off-line: .••••..... none.
. 43
Connection to Device
.431 Devices per controller: . . . . . . . . . . . . 1 to 8.
.432 Restrictions :........ maximum of.s storage access
mechanisms per 2841; see
following table.
Device
·5
ACCESS TIMING
· 51
Arrangement of Heads
· 511 Number of stacks Stacks per system: ..
Stacks per module: ..
Stacks per yoke: . . . .
Yokes per module: ..
.512 Stack movement: ....
· 513 Stacks that can access
any particular
location: . . . . . . . . .
.514 Accessible locations
By single stack With no movement:
With all movement:
By all stacks With no movement:
1.
1 track .
200 tracks.
10 tracks per module.
80 tracks per system of
8 modules.
Simultaneous Operations: . . • . . . . . . . a read, write, or seek oper--ation using one access
mechanism can be overlapped with seek operations
taking place in any or all
other access mechanisms.
Only one read or write
operation can take place
at a time on each. Selector
Channel.
.53
Access Time Parameters and Variations
2
4
1
1
1
10 to 80 .
10.
10 .
1.
horizontally across disc
surface to one of 200
tracks.
· 52
No. of Access Mechanisms
2302 Disk Storage,
Model 3
2302 Disk Storage,
Model 4
2311 Disk Storage
Drive
2321 Data Cell Drive,
Modell
7320 Drum Storage
.44
420:044.320
.532 Variation in access time -
Data Transfer Control
.441 Size of load 1 record per track: . 1 to 3, 625 bytes.
Multiple records
(3615 _ L
).
per track: . . . • . . . N = 81 + 1.049L + 1 L
where N = number of bytes
per track, and L = length
of data plus key in each
record, in bytes.
Cylinder mode: ..•• 1 to 10 tracks (up to 36,250
. bytes) .
.442 Input-output area: ..•• main core storage.
.443 Input-output area
access: . . . . . . . . . each byte.
. 444 Input-output area
lockout: ......•... blocks of 2,048 bytes can
. be protected (optional on
Model 30 and 40 Proccessing Units, standard on
larger models).
.445 Synchronization: ...•. automatic.
.447 Table control: ....•. yes; scatter-read and
gather-write facilities
are available at programmer's option.
See Section 420:111,
Simultaneous Operations .
. 448 Testable conditions: .. available.
busy.
not operational.
performing operation with
interruption pending.
©
Stage
Variation,
msec
Move head to selected
track (cylinder):
30 to 145
Wait for beginning of
selected track:
o to 25
Transfer data:
25 per track.
Total:
.6
CHANGEABLE STORAGE
.61
Cartridges (Disk Pack)
Average,
~
85.0
12.5
25.0
122.5 msec.
.611 Cartridge capacity: .. 7,250,000 bytes (maximum) •
· 612 Cartridges per
module: . • . • . . . . . 1 .
· 613 Interchangeable: .•... yes, between a 2311 Disk
Storage Drive and any
other 2311. Disk Packs
written by a 2311 cannot
be read by any mM 1311
Disk Storage Drives, and
the converse is also true.
.62
Loading Convenience
.621 Possible loadiIig While computing system is in use: •... yes.
While storage system
is in use: . • . . . . . yes, if the particular
module is not addressed.
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360
420:044.622
.622 Method of load:ing: ... operator.
· 623 Approximate change
time: . . . . • . . . • . . 1 m:inute.
· 624 Bulk loading: .•.•... no; 1 cartridge of 6 discs
at a time.
.7
PERFORMANCE
· 72
Transfer Load Size
With core storage Single track: .••••. 1 to 3,625 bytes.
Cyl:inder: .••.•••. up to 36,250 bytes (10
tracks) per cyl:inder.
.73
Note:
.8
Based on .random accessing and reading
of one 150-character record, with no
updating or rewriting.
ERRORS, CHECKS, AND ACTION
Invalid address:
Invalid code:
Receipt of data:
Recording of
data:
Update Cycle Rates
Recovery of
data:
With no overlapping of
seek times: •...•.. 6.7 references/sec.
With maximum overlapping of seek times: 15.5 references/sec.
Dispatch of data:
Note:
7/65
With no overlapping of
seek times: •••••.. 10.1 references/sec.
With maximum overlapping of seek times: 70. 0 references/sec.
Effective Transfer Rate
With core storage Cyl:inder mode,
1-way transfer: ... 104,000 bytes per second,
based on random accessing
and transferring of one
cylinder (36, 250 bytes) of
data.
.74
.75 . Read-Only Reference Cycle Rates
Based on random access:ing of one 150character record; reading, updat:ing, and
rewrit:ing that record; and rereading for
verification of recording accuracy.
Timing conflicts:
Wrong track:
Check or
Action
Interlock
program check' program control.
.
on recorded
address
all 8-bit codes
are valid.
compute cyclic record check
check code
code.
none, except
programmed
read.
compute cyclic
check code
and compare
with recorded
check code
check
interlock
program check
on track
address
error signal.
attaches parity
bit.
wait.
program control.
420:045.100
IBM System/360
Internal Storage
2321 Data Cell Drive
INTERNAL STORAGE: 2321 DATA CELL DRIVE
.1
GENERAL
.11
Identity:
.. 2321 Data Cell Drive,
Modell.
.12
Basic Use:
.. random-access auxiliary
storage.
.13
Description
IDM's new 2321 Data Cell Drive provides economical on-line random access storage for. extremely
large volumes of data in applications where relatively slow access times can be tolerated. Each
2321 drive stores up to a maximum of 400 million
bytes (or 800 million packed decimal digits and
signs) in 10 removable, mterchangeable Data Cells
with a capacity of 40 million bytes each.
Data is recorded on magnetic strips which are held
in Data Cells mounted vertically around the circumference of a cylinder or "tub file" which can be
rotated. Each of the 10 Data Cells is divided into
20 subcells, and each subcell contains 10 magnetic
strips. A bidirectional rotary positioning system
positions the selected subcell beneath an access
station. The selected strip is withdrawn from the
Data Cell, placed on a separate rotating drum, and
moved past the read/write head assembly, where
reading or recording takes place. The strip is
returned to its original location in the Data Cell if:
(1) a Restore instruction is issued;
(2) a Seek instruction references a new strip; or
(3) 800 milliseconds elapse between successive·
Data Cell instructions. (This is a safeguard
to protect the flexible magnetic strips from
unnecessary wear.)
Each magnetic strip is 13 inches long, 2.25 inches
Wide, and 0.005 inch thick; has an iron-oxide coating
on one side and a carbon anti-static coating on the
other; has a pair of coding tabs to ide~tify its position in the cell; and provides 100 addressable recording tracks. Each track has a maximum data
capacity (based on one record per track) of 2, 000
bytes. Record length and number of records per
track are variable and user-defined.
The physical components of each Data Cell Drive
are arranged in an L-shaped cabinet whose sides
measure about four feet by six feet in length. The
components include an electronics section and
pneumatic, hydraulic, and mechanical equipment.
The read/write head assembly contains 20 heads and
can be moved to any of 5 discrete positions in order
to serve the 100 data tracks on each strip. Recording is serial by bit, strip velocity is 250 inches
per second, and data transfer rate is about 55,000
©
bytes per second. With the use of chained commands,
it is possible to read or Write up to 20 successive
tracks (cylinder mode) during 20 successive read/
write drum rotations without repositioning the heads.
When a previously addressed strip is on the drum,
access time to data on a different strip varies from
375 to 600 milliseconds. When no strip is on the
drum, access time varies from 175 to 400 milliseconds. When the proper strip is already on the
drum, access time averages 95 milliseconds if repositioning of the read/write head assembly is
required. Drum rotation time is 50 milliseconds,
and an entire data track passes under the heads in
41.8 milliseconds. Only 100 microseconds are
required for head switching.
Each Data Cell can be removed and interchanged
with any other Data Cell in any 2321 Data Cell Drive.
.A combination handle-cover facilitates removal and
protects the magnetic strips during handling. A
covered Data Cell containing 200 strips weighs only
about 5 pounds. One Data Cell can be removed and
replaced by another in less than 30 seconds. When
less than a full complement of 10 Data Cells is required, ballast cells are used to balance the rotating
array.
The 2841 Storage Control is required to control the
2321 Data Cell Drive. Up to eight 2321 Data Cell
Drives can be connected to a 2841 Storage Control,
and up to eight control units can be connected to
each System/360 input-output channel. Each 2841
Storage Control used with Data Cell Drives requires
a 2321 Attachment unit.
The 2841 Storage Control is used to control the 2311
Disk Storage, 2302 Disk Storage, and 7320 Drum
Storage as well as the 2321 Data Cell Drive. The
2841 interprets and executes file commands, performs the required conversions between serial-bybit and parallel-by-bit data modes, checks the
validity of the data being transferred, and furnishes
status information about the random access file units
to the Processing Unit.
The 2841 strips the parity bit off each byte to be
recorded !n random access storage. The validity of
the recorded information is checked by generating a
string of 16 "cyclic check bits" and appending it to
the end of each record. When the data is read,
the check bits are regenerated and compared; an .
unequal comparison results in a data error signal.
Parity bits are restored as the data is transferred
back into core storage.
Record format may be single or multiple records
per track, and each record may contain a key portion. A separate "count" byte in each record is
used to specify the key length, which can be as long
as 255 bytes. The number of bytes in the data
portion of the record is specified by a two-byte
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360
420:045; 130
· 13
Description (Contd. )
Bytes 1 record per track:. . 2, 000 bytes.
Multiple records
per track: •••••..
_(1984-L
N - 100 + 1. 049L
1 L
where N = number of bytes
per track, and L = length
of data plus key in each
record, in bytes.
count code. Thus each record can have a count area,
a key (optional), and a data area. An additional
area is included at the beginnin'g of each track to
specify the track address, and a one-byte code is
used to indicate possible unusable recording portions
of the track.
When multiple records are present on a track, the
amount of data which can be recorded is decreased
markedly. For example, a track which stores
recQrds consisting of a 10-byte key and 100 data
bytes holds 9 records, or only 900 data bytes. In
the form of a single record with a 10-byte key, a
single track can contain 1,974 bytes of data.
A series of file commands permits any or all of the
three areas (count, key, and data) of a record in a
random access file to be searched, read, or written.
Movement of an access mechanism is initiated by
transferring a 6-byte seek address from the Processing Unit to the 2841. The optional File Scan Function feature permits an automatic search for a
specific identifier or key. The Record Overflow
feature permits a single record to overflow from
one track to another track within a cylinder. The
Two-Channel Switch feature enables a 2841 Storage
Control to be switched from one channel to another
under program control.
.14
Availability: ..
?
· 15
First Delivery:
?
. 16
Reserved Storage: ... none .
.2
PHYSICAL FORM
. 21
Storage Medium: .
.22
Physical Dimensions
+)
Records: . . . . . . . . . variable.
· 26
Bands per Physical
Unit: . . . . . . . . . . . 100 per strip.
.27
Interleaving Levels: .. 1; i. e., no interleaving.
.28
Access Techniques
.281 Recording method: •.. magnetic strip passes by
fixed heads.
· 283 Type of access Description of Stage
Position subcell be':'
neath access station
and withdraw strip: .. yes, when a different strip
is required.
Position head block to
selected track: . . . . . yes, if done within 800
milliseconds after completion of previous Data
Cell instruction.
Wait for start of
yes, if done within 800
strip (data): ...
milliseconds after completion of previous, Data
Cell instruction, and if
track is available to head
block without repositioning.
Transfer data: . . . . . . no.
Restore strip to subcell location: . . . . . . yes (automatic operation, if
necessary, when Seek instruction is given) .
. magnetic strips.
.223 Magnetic stripLength: . . . . . . · .. 13 inches.
Width: . . . . . . . · .. 2.25 inches.
Thickness:
• O. 005 inches.
Number: . . . . . . · .. 200 per Data Cell.
. 23
Storage Phenomenon: . direction of magnetization.
. 24
Recording Permanence
· 241 Data erasable by instructions: . . . . . . .
· 242 Datil regenerated
constantly: . . . . . . .
.243 Data volatile: . . . . . . .
· 244 Data permanent: . . . . .
. 245 Storage changeable: ..
• 25
yes.
no.
no.
no.
yes, in units of 200 strips
(1 Data Cell).
Data Volume per Band of 1 Track
Words: ....••...•• 500.
Characters: .••.•... see "Bytes."
,Digits: .•.••.••... 2,000 in zoned byte format;
4,000 in packed format.
Possible Starting Stage
.29
Potential Transfer Rates
.291 Peak bit rates ---'
CYCling rates: . . . . .
Track/head speed: ..
Bits/inch/track: . . . .
Bit rate'per track: ..
· 292
1,200 rpm (50 msec/rev).
250 inches/sec .
1,750.
438,000 bits/sec/track.
Pe~
data rates Unit of data: . . . . . . byte.
Conversion factor: .. 8 bits per byte.
Data rate: ....•.•. 54,700 bytes per second, or
109, 400 packed decimiLl
digits per second.
(Contd. )
7/65
./
420:045.300 .
INTERNAL STORAGE: 2321 DATA CELL DRIVE
.3
DATA CAPACITY
.31
Module and System Sizes
Minimum
Storage
.32
Maximum Storage
per Control
Identity:
1 Data Cell
1 Data Cell
Drive
8 Data Cell Drives.
Data Cells:
Strips:
Bands:
Cylinders:
Bytes:
VVords:
Packed digits:
Modules:
1
200
20,000
1,000
40,000,000
10,000,000
80,000, 000
10
2,000
200,000
10,000
400,000,000
100,000,000
800,000,000
80.
16,000.
1,600,000.
80,000.
3,200,000,000.
800, 000, 000.
6,400,000,000.
1
1
8.
Rules for Combining
Modules: . . . . . . . . . any number up to 8 Data
Cell Drives per 2841
Control Unit.
·4
CONTROLLER
.41
Identity: . . . . .
· 42
Connection to System
. IBM 2841 Storage Control
Unit with 2321 Attachment.
.421 On-line: .••••••.•• see Section 420:031, Sys-
tem Configuration.
. 422 Off-line: . . . . . . . . . . none.
.43
Connection to Device
. 431 Devices per control-
ler: . . . . . . . . . . . . 1 to 8, using 2321 Attachment.
.432 Restrictions: . . . . . . . maximum of 8 storage
access mechanisms per
2841; see following table.
No. of Access
Mechanisms
Device
2302 Disk Storage, Model 3
2302 Disk Storage, Model 4
2311 Disk Storage Drive
2321 Data Cell Drive, Modell
7320' Drum Storage
· 44
2
4
1
1
1
Data Transfer Control
· 441 Size of load
1 record per track: .. 1 to 2,000 bytes.
Multiple records per
(1984 _ L
track: •••••••.•• N = 100 + 1. 049L
\
)
+1 L
where N = number of bytes
per track, and L = length
of data plus key in each
record, in bytes.
Cylinder mode: . . . . 1 to 20 tracks (up to 40,000
bytes).
.442 Input-output area: . . . . main core storage;
. 443 Input-output area
access: . . . . . . . . . . each byte.
©
· 444 Input-output area
lockout: . . . . . . . . . blocks of 2, 048 bytes can
be protected (optional on
Model 30 and 40 Processing Units, standard on
larger models) .
. 445 Synchronization: . . . . . automatic .
.447 Table control: . . . . . . yes; scatter-read and
gat.lJ.er-write facilities
are available at programmer's option. See
Section 420:111, Simultaneous Operations.
.448 Testable conditions: .. available .
busy.
not operational.
performing operation with
interruption pending .
·5
ACCESS TIMING
· 51
Arrangement of Heads
· 511 Number of stacks -
Stacks per module: .. 1 per 2321 Data Cell Drive.
Heads per stack: ... 20.
· 512 Stack movement: . . . . across strip width to any
one of 5 positions.
· 513 Stacks that can access
any particular
location: . . . . . . . . . 1.
.514 Accessible locations
By single stack VVith no movement: . 20 tracks of the strip on
read/write drum (1
cylinder).
VVith all movement: . 100 tracks of the strip on
read/write drum.
· 52
Simultaneous Operations: . . . . . . . . . . . a read, write, or seek
-operation on anyone Data
Cell Drive can be overlapped with seek operation
on other drives. Only one
read or write operation
can take place at a time on
, each Selector Channel.
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360
420:045.530
.53
• 612 Cartridges per
module: •••.•.•.. 10 •
· 613 Interchangeable: . . . . . yes.
Access Time Parameters and Variations
. 532 Variation in access timeVariation,
Stage
Average,
msec
Strip on drum and
only head switching
required Switch head and
wait for start of
strip: . . . . . . . . O. 1 to 8.2 (given
between end and
beginning of
strip), or
4.1, or
0.1 to 50 (given at 25.0
random time)
Transfer data: ... 41. 8 per track
41.8
45.9 or
66.8.
• Strip on drum and
head repositioning
required Move head: . . . . . 95
95.0
Transfer data: ... 41. 8 per track
41.8
136.8.
• No strip on drum Position subcell
and withdraw
strip: . . . . . . . . 175 to 400
350.0
Transfer data: ... 41. 8 per track
41.8
391. 8
• Strip on drum and
new strip addressed Deposit strip,
position new
subcell, and
withdraw strip:
375 to 600
550.0
Transfer data:
41. 8 per track
41.8
591.8.
•
.6
CHANGEABLE STORAGE
.61
Cartridges (Data Cells)
.62
Loading Convenience
.621 Possible loadingWhile computing system is in use: ••.. yes.
While storage system is in use: •••• yes (if individual Data Cell
Drive unit is free).
.622 Method of loading: ... operator procedure.
· 623 Approximate change
time: •••......•. 30 seconds.
.624 Bulk loading: . . . . . . . no; 1 Data Cell at a time.
.7
PERFORMANCE
.72
Transfer Load Size
With core storage Single track: • . . . . . 1 to 2, 000 bytes.
Cylinder: . . . . . . . . up to 40,000 bytes (20
tracks per cylinder).
.73
Effective Transfer Rate
With core storage Cylinder mode, 1 way
transfer: . . . • . . . . 25,800 bytes per second;
based on making initial
random selection of strip
and transferring one
cylinder (40,000 bytes)
of data.
• 74
Update Cycle Rate
Reference to strip already
on drum: ••••..•. 5.1 references/sec.
Reference to new
strip: ...•..••.•. 1. 5 references/sec •
Note:
. 611 Cartridge capacity: ... 40, 000,000 bytes (maximum).
.8
ERRORS, CHECKS AND ACTION
Error
Invalid address:
Invalid code:
Receipt of data:
Recording of data:
Recovery of data:
Dispatch of data:
Timing conflicts:
Wrong strip:
Wrong track:
7/65
Based on random accessing of one 150character record; reading, updating,
and rewriting that record; and rereading
for verification of recording accuracy .
Check or
Interlock
program check on
recorded address
all 8-bit codes are
valid.
compute cyclic check
code
none, except programmed read.
compute cyclic check
code and compare
with recorded check
code
attach parity bit to each
byte.
interlock
program check on
track address
program check on
track address
Action
program control.
record check
code.
error signal.
wait.
program control.
program control.
/
420:046.100
IBM System/360
Internal Storage
7320 Drum Storag&
INTERNAL STORAGE: 7320 DRUM STORAGE
.1
GENERAL
.11
Identity: ..•...•.•• 7320 Drum Storage, Model 2.
.12
Basic Use: ..••.•.. fast random-access auxiliary storage.
.13
DesGription
Processing Unit to the 2841. The optional File
Scan Function feature pE;lrmits an automatic search
for a specific identifier ,or key. The Record
Overflow feature permits a single record to overflow from one track to another track within a
cylinder. The Two-Channel Switch feature enables
a 2841 Storage Control to be switched from one
channel to another under program control.
The 2841 Storage Control is used to control the
2311 Disk storage, 1302 Disk Storage, and 2321
Data Cell Drive as well as the 7320 Drum Storage.
The 2841 interprets and executes file commands,
performs the required conversions between serialby-bit and parallel-by-bit data modes, checks the
validity of the data being transferred, and furnishes
status information about the random access file
units to the Processing Unit.
mM 7320 Drum Storage was developed for the
SABRE airline reservation system and has previously been offered for use in mM 7090/7094 systems. The drum offers relatively fast access (8.6
milliseconds average) to a moderate amount of storage (830, 000 bytes per drum). Up to eight drums
can be connected to an mM 2841 Storage Control
Unit, and up to eight control units can be connected
to each System/360 input-output channel. Each 2841
Storage Control Unit used with 7320 Drum Storage
requires a 7320 Attachment unit.
The 2841 strips the parity bit off each byte to be
recorded in random access storage. The validity
of the recorded information is checked by generating
a string of 16 "cyclic check bits" and appending it
to the end of each disc record. When the data is
read, the check bits are regenerated and compared;
an unequal comparison results in a data error
signal. Parity bits are restored as the data is
transferred back into core storage.
Typical applications for the 7320 are storage of
program segments or subroutines, fast-access
directories for larger-capacity storage units, and
extension of core memory for large problems.
Each drum can store 830,000 bytes or 1. 66 million
packed decimal digits. There are 400 standard and
40 alternate data tracks. "Each track is served by
its own fixed head and is capable of storing a maximum of 2,075 bytes (based on one record per track).
Recording is serial by bit. Drum rotation speed is
approximately 3,490 revolutions per minute, so the
rotation time is 17.2 milliseconds. The peak data
transfer rate is 135,000 bytes per second.
.14
Availability: . . . • . . . ?
.15
First Delivery: ....• currently is use with mM
7090/7094.
.16
Reserved Storage
Purpose
Record length and number of records per track are
variable and user-defined, and each record may
contain a key portion. A separate "count" byte
in each record is used to specify the key length,
which can be as long as 255 bytes. The number of
bytes in the data portion of the record is specified
by a two-byte count code. Thus each record can
have a count area, a key (optional), and a data
area. An additional area is included at the
beginning of each track to 'specify the track address,
and a one-byte code is used to indicate possible
unusable recording portions of the track.
When multiple records are present on a track, the
amount of data Which can be recorded is decreased
markedly. For example, a track which stores
records consisting of a 10-byte key and 150 data
bytes holds 7 records, or only 1,050 data bytes.
Each record contains a total of 278 bytes in this
case. In a form of a single record with a 10byte key, a single track can contain 2,047 bytes
of data. A series of file commands permits any
or all of the three areas (count, key, and data)
of a record in a random access·file to be searched,
read, or written. Selection of a track is initiated
by transferring a 6-byte seek address from the
©
Number of Locations
Alternate data tracks: 40 tracks.
Clock tracks: .••... 3 tracks.
.2
PHYSICAL FORM
.21
storage Medium: .... magnetic drum.
· 22
Physical Dimensions
.222 DrumDiameter: . . . . • . . . approximately 12 inches.
Length: . . . . . . . . . approximately 12 inches.
Number on shaft: ... 1.
• 23
Storage Phenomenon: . magnetization.
· 24
Recording Permanence
· 241 Data erasable by
instructions: ...•..
· 242 Data regenerated
constantly: . . . . . . .
.243 Data volatile: .•.•..
· 244 Data permanent: ..•..
.245 Storage changeable: ..
1965 AUERBACH Corporation and AUERBACH Info, Inc.
yes.
no.
no.
no.
no.
7/65
IBM SYSTEM/360
420:046.250
.25
Data Volume per Band of 1 Track
.42
Records: •••.•••.•
Words: •••••.•...•
Characters: ••••••••
Digits: •••.••••.•.
.421 On-line: . . . . • . • . . . see Section 420:031,
System Configuration.
.422 Off-line: .•.••.•... none.
variable.
518.
see "Bytes".
2,075 in zones byte format;
4,150 in packed format.
Bytes 1 record per track:. • 2, 075 bytes.
Multiple records
per track: ••..•.• N = (2057-L
+ 1) L
118 + 1.049L
whereN = number of bytes
per track, and L = length
of data plus key in each
record, in bytes.
.26
• 43
Interleaving Levels: •• 1; i. e., no interleaving .
.28
Access Techniques
No. of access mechanisms
2302 Disk Storage,
Model 3
2302 Disk Storage,
Model 4
2311 Disk Storage
Drive
2321 Data Cell Drive,
Modell
7320 Drum Storage,
Model 2
.281 Recording method: ••• fixed heads; 1 per track .
• 283 Type of access .44
Description of Stage
Potential Transfer Rates
.291 Peak bit ratesCycling rates: ...•. 3,490 rpm (17.2 msec/rev).
Bit rate per track: .• 1,080,000 bits/sec/track.
.292 Peak data ratesUnit of data: ...•.• byte.
Conversion factor: .• 8 bits per byte.
Data rate: .••••..• 135,000 bytes per second.
270,000 packed decimal
digits per second.
.3
DATA CAPACITY
.31
Module and System Sizes
Identity:
Bands:
Bytes:
Packed digits:
Words:
Modules:
Cylinders:
.32
Minimum
Storage
Maximum Storage
per Control
1 drum
400
830,000
1,660,000
207,500
1
10
8 drums.
3,200.
6,640,000.
13,280,000.
1,660,000.
8.
80.
Rules for Combining Modules: ••••• any number up to 8 Drum
Storage units per 2841
Storage Control Unit.
.4
CONTROLLER
.41
Identity: . . . . . . . • • . IBM 2841 Storage Control
Unit with 7320 Attachment.
7/65
2
4
1
1
1
Data Transfer Control
Possible Starting Stage
Wait for drum rotation: • • • • • . • . • • yes.
Transfer data: .••••. no.
.29
Connection to Device
.431 Devices per controller: . . . . . • . . . . . . 1 to 8, using 7320 Attachment.
.432 Restrictions: .•.•..• maximum of 8 storage access mechanisms per
2841; see following table.
Bands per Physical
Unit: . • • • • . • • • • . • 400 data bands.
-40 alternate bands.
3 clock bands.
• 27
Connection to System
.441 Size of load 1 record per track: 1 to 2, 075 bytes.
Multiple records
(2057 _ L
)
per track: . . . . . . . N = 118 + 1. 049L + 1 L
where N = number of bytes
per track, and L = length
of data plus key in each
record, in bytes.
. 442 Input-output area: . . . . main core storage .
.443 Input-output area
access: ...•...• :each byte.
.444 Input-output area
lockout: . . . . . . . . . blocks of 2,048 bytes can
be protected (optional on
Model 30 and 40 Processing Units, standard
on larger models).
.
.445 Synchronization: ..••. automatic.
.447 Table control: •..•.. yes; scatter-read and
gather-write facilities
are available at programmer's option.
See Section 420:111,
Simultaneous Operations.
.448 Testable conditions: .. available.
busy.
not operational.
performing operatio:t;l with
interruption pending.
.5
ACCESS TIMING
.51
Arrangement'of Heads
. 511 Number of.,stacks Stacks per drum: ... 400 .
Stacks per cylinder: . 400.
Heatis'per stack: .•. 1.
(Contd. )
INTERNAL STORAGE: 7320 DRUM STORAGE
420:046.512
• 512 Stack movement: •••• none •
• 513 Stacks that can access
any particular
location: • • • • • • • • • 1.
.514 Accessible locations By single: •••••••• 1 track.
By all stacks: ••••• 400 tracks per drwn.
.52
• 53
Simultaneous Operations: •••••••..• maximum of one 7320 Drwn
--Storage operation per
Selector Channel.
Update Cycle Rate: •• 22.2 references/sec •
Note:
.75
Stage
Variation I
~
Wait for start of
addressed band:
Transfer data:
Total:
Oto17.2
17.2 per track
.6
CHANGEABLE
STORAGE: .•••••• none.
.7
PERFORMANCE
.72
Transfer Load Size
Average.
Note:
~
8.6
17.2
25.8insec.
With core storage Single track: • • • • • • 1 to 2. 075 bytes.
Cylinder: •••••••• up to 830,000 bytes (400
tracks per cylinder).
Effective Transfer Rate
With core storage Cylinder mode. 1way transfer: ••••• 119,000 bytes per second;
based on transferring 40
tracks (83,000 bytes) of
data; includes rotational
delay time.
©
.8
Based on'random accessing of one 150character record; reading, updating, and
rewriting that record; and rereading for
ver:lfication of recording accuracy.
Read-Only Reference
Cycle Rate: ••••••• 93.3 references/sec.
Access Time Parameters and Variations
.532 Variation in access time -
.73
.74
Based on random accessing and reading of
one 150-character record. with no updating
or rewriting •
ERRORS, CHECKS. AND ACTION
Error
Check or Interlock
Invalid address: program check
on recorded
address
Invalid code:
all 8-bit codes
are valid.
Receipt of data: compute cyclic
check code
Recording of
data:
none. except
programmed
read.
Recovery of
data:
compute cyclic
check code and
compare with
recorded check
code
Dispatch of data: attach parity bit
to each byte.
Timing conflicts: interlock
Wrong track:
program check
on track
address
1965 AUERBACH Corporation and AUERBACH Info, Inc.
Action
program control.
record check
code.
error signal.
wait.
program control.
7/65
420:047.100
IBM System/360
Internal Storage
2301 Drum Storage
INTERNAL STORAGE: 2301 DRUM STORAGE
.1
GENERAL
.11
Identity:
.12
Basic Use: ..••.•.. fast random-access auxiliary storage.
.13
Description
A series of file commands permits any or all of the
three areas (count, key, and data) of a record in a
random access file to be searched, read, or written. Selection of a track is initiated by transferring a 6-byte seek address from the Processing
Unit to the 2820. The optional File Scan function
and Record Overflow feature permitted on the 2841
Storage Control are not available for the 2820
Storage Control.
. . . . . . • . . 2301 Drum Storage.
mM's new 2301 Drum Storage unit provides random
access storage for approximately 4 million bytes
(or 8 million packed decimal digits and signs) and
an unusually high data transfer rate of 1.2 million
bytes per second. The 2301 differs from the bitserial 2302, 2311, 2321, and 7320 file units in that
it reads and records four bits of information in
paralleL The 2301 drum has been improved relative to mM's older 7320 drum by doubling the
track density (twice as many tracks) and by 'approximately doubling the recording density. It offers
relatively fast access (8.6 milliseconds average)
to a moderate amount of storage (4.1 million bytes
per drum). Typical applications for the 2301' are
storage of program segments or subroutines, fastaccess directories for larger-capacity storage
units, function tables, compiler lists, and extensions of core memory for large problems.
The 2301 Drtim Storage cannot be used with System/
360 Model 30 or 40 'Processing Units because of
its high data transfer rate, but it is available for
Models 50, 65, arid 75, Up to four 2301 drums
can be connected to an mM 2820 Storage Control
(a special control unit which can be used only for
2301 Drum Storage). One 2820 Storage Control
can be connected to a Model 50 Processing Unit,
and up to eight 2820's to a Model 65 or 75, depending upon the number and models of Selector
Channels used. Each drum can store 4,096,600
bytes or 8,193,200 packed decimal digits. There
are 800 data tracks, read or recorded in groups of
4 tracks (one band) at a time. Each of the 200
bands is capable of storing a maximum of 20,483
bytes (based on one record per band). Drum rotation speed is approximately 3,490 revolutions
per minute, so the rotation time is 17. 2 milliseconds. The peak data transfer rate is 1. 2
million bytes per second.
Record length and number of records per band are
variable and user-defined, and each record may
contain a key portion. A separate "count" byte in
each record is used to specify the key length, which
can be as long as 255 bytes. The number of bytes
in the data portion of the record is specified by a
two-byte count code. Thus each record can have
a count area, a key (optional), and a data area. An
additional area is included at the beginning of each
band to specify the track address, and a one-byte
code is used to indicate possible unusable recording
portions of the band. When multiple records are
present on a band, the 3.)llount of data that can be
recorded is decreased.
7/65
The Two-Channel Switch feature enables a 2820
Storage Control to be switched from one channel
to another under program control.
The 2820 Storage Control interprets and executes
file commands, performs the required data-mode
conversions, checks the validity' of the data being
transferred, and furnishes status information
about the random access file units to the Processing
Unit.
The 2820 strips the parity bit off each byte to be
recorded'in random access storage. The Validity
of the recorded information is checked by generating a string of 16 "cyclic check bits" and appending it to the end of each drum record. When
the data is read, the check bits ar'e regenerated
and compared; an unequal comparison results in
a data: error signaL Parity hits are' restored as
the data is trll?sferred back into core storage.
• 14
Availability: . . . . . . . ?
.15
First Delivery: . . . . . ?
.16
Reserved Storage: ... 80 alternate data tracks
and 4 clock tracks.
.2
PHYSICAL FORM
.21
Storage Medium: . . . . magnetic drum.
· 22
Physical Dimensions
.222 DrumDiameter: ..••.••. approximately 10.7 inches.
Length: . • . • . . . . . . approximately 12 inches.
Number on shaft: ... 1.
· 23
Storage Phenomenon: . magnetization.
· 24
Recording Permanence
· 241 Data erasable by
instructions: . . . . . .
.242 Data regenerated
constantly: . . . . . . .
· 243 Data volatile: . . . . . . .
.244 Data permanent: .•...
.245 Storage changeable: "
yes.
no.
no.
no.
no.
INTERNAL STORAGE: 2301 DRUM STORAGE
. 25
Data Volume per Band of 4 Tracks
.422 Off-line: . • . . . . . . . . none .
Records: .•.....••• variable.
Words: .•••••••••• 5,120.
Characters:. . • • • • • • see "Bytes."
Digits: .••••••.••• 40,966.
Bytes 1 record per band: •• 20,483.
Multiple records
(20 483 )
per band: " ••••• N = L ~ 186 L
where N = number of bytes
per band, and L = length
of data plus key in each
record, in bytes.
.43
• 26
Bands per Physical
Unit: " • • . • . • • • . • 200 data bands.
.27
Interleaving Levels: .• 1; i. e., no interleaving.
. 28
Access Techniques
· 281 Recording method: • • • fixed heads.
· 283 Type of access Description of Stage
Possible Starting Stage
Wait for drum rotation: •• • • • • • • . • yes.
Transfer data:. • • . • • no.
.29
Potential Transfer Rates
• 291 Peak bit rates Cycling rates: .•.•• 3,490 rpm (17.2 msec/rev).
Track/head speed: .• 2,000 inches/sec, approx.
Bits/inch/track: .••. 1,250 approx.
Bit rate per track: .• 2,400,000 bits/sec/track .
. 292 Peak data rates Unit of data: .•••.• byte.
Conversion factor: .• 8 bits per byte.
Gain factor: •••.. ".• 4 tracks per band.
Loss factor: •.•.•. 2 four-bit groups per byte.
Data rate: .••••.•• 1,200,000 bytes per second.
.3
DATA CAPACITY
• 31
Module and System Sizes
Identity:
Bands:
Bytes:
Packed digits:
Words:
Modules:
I
• 32
\
420:047.250
Minimum
Storage
Maximum Storage
per Control
2301 drum
200
4,096,600
8,193,200
1,024,150
4 2301 drums.
800.
16,386,400.
32,772,800.
4,096,600.
4.
1
Rules for Combining
Modules: ••.•••.. any number up to 4 drums
per 2820 Storage Control.
.4
CONTROLLER
.41
Identity:
.42
Connection to System
. . • . . . . . . mM 2820 Storage Control.
.421 On-line: .••..•.... see Section 420:031,
System Configuration.
©
Connection to Device
.431 Devices per controller: . • • • • . . . . • . . 1 to 4 drums.
.432 Restrictions: . • . . . . . only 2301 drums can be
connected to the 2820
Storage Control.
.44
Data Transfer Control
· 441 Size of load1 record per track: .. 20,483 bytes.
Multipl.e records per
_ (20,483 )
track. . . . . . . . . . . N - L + 186 L
where N = number of bytes
per band, and L = length
of data plus key in each
record, in bytes .
• 442 Input-output area: .••. main core storage .
.443 Input-output area
access: ....••...• each byte.
.444 Input-output area
"
lockout: . . . • . . . . . blocks of 2, 048 bytes can
be protected.
.445 Synchronization:. . . • • automatic .
. 447 Table control: ... "... yes; scatter-read and
gather-write facilities
are available at programmer's option.
See Section 420:111,
Simultaneous Operations .
. 448 Testable conditions: .. available.
busy.
not operational.
track address compare.
•5
ACCESS TIMING
· 51
Arrangement of Head
· 511 Number of stacks Stacks per drum: ... 200.
Stacks per module: .. 200 .
Heads per stack: . . . 4 .
. 512 Stack movement: .... none; fixed heads .
· 513 Stacks that can access
any particular location: . . . . . . . . . . 1.
· 514 Accessible locations By single stack: . • • • 1 band of 4 tracks.
By all stacks: ..••. 200 bands of 4 tracks each.
.52
Simultaneous Operations: •••••••••• one 2320 Drum operation
--at a time. Other inputoutput channels are
locked out during drum
transmission time.
.53
Access Time Parameters and Variations
.532 Variation in access time Variation,
stage
~
Wait for start of
addressed band: Oto17.2
Transfer data:
17 . 2 per track
Total:
1965 AUERBACH Corporation and AUERBACH Info, Inc.
Average,
~
8.6
17.2
25.8 msec.
7/65
IBM SYSTEM/360
420:047.600
.6
CHANGEABLE STORAGE: .•••••••...• none.
.7
PERFORMANCE
.72
Transfer Load Size
With core storage Single track: •••••• 1 to 20,483 bytes.
Cylinder: .••••.•• 4,096,600 bytes.
.73
Effective Transfer Rate
With core storage I-way transfer: •••• 1,134,000 bytes/sec,
based on transfer of
204,830 bytes (10 tracks);
includes rotational delay
time.
• 74
.75
Update Cycle Rate: •.. 22.7 references/sec .
Note: Based on random accessing of one record; reading, updating, and rewriting
that record; and rereading for verification of recording.
Read-Only Reference
Cycle Rate: .•••••• 116 references/sec.
Note:
7/65
Based on random accessing and reading of
one record with no updating or rewriting •.
.8
ERRORS, CHECKS, AND ACTION
Errors
Check or Interlook
Action
Invalid address: program check program control.
on recorded
address
Invalid code:
all 8-bit codes
are valid.
Receipt of data: compute cyclic record check
check code
code.
Recording of
none; except prodata:
grammed read.
Recovery of
data:
compute cyclic error signal.
check code and
compare with
recorded
check code.
Dispatch of data: attach parity bit
to each byte.
Timing conflicts: interlock
wait.
Wrong track:
program check program control.
on track address
420:048.100
IBM System/360
Internal Starage
2314 Direct Access Starage Facilit.y
INTERNAL STORAGE: 2314 DIRECT ACCESS STORAGE FACILITY
rea,d, the check bits are regenerated and compared;
an unequal comparison results in a data error
signal. Parity bits are restored as the data is
transferred back into core storage .
.1
GENERAL
.11
Identity: . . . . . . • . . . 2314 Direct Access Storage
Facility.
. 12
Basic Use: . . . . . . . . random-access, changeablecartridge auxiliary
storage.
. 13
Description
The removable 2316 Disk Packs are 14 inches in
diameter and 6 inches high. When a Disk Pack is
not mounted on a drive, the pack and its cover
combine to form a sealed container that can be
conveniently stored and transported.
The IDM 2314 Direct Access Storage Facility is a
new changeable-cartridge disc storage unit that
provides increased storage capacity and a higher
data transfer rate than the 2311 Disk Storage Drive
described in Section 420:044. The 2314 was announced in April, 1965. Note that the 2316 Disk
Pack used with the 2314 is not the same as the 1316
Disk Pack used with the 2311 and 1311 Disk Storage
Drives'. Disk Packs cannot be interchanged between
a 2314 and a 2311 or 1311 drive.
Repositioning time for the 2314 access mechanism
is a maximum of 140 milliseconds ·and averages 75
milliseconds. Seek operations on all access
mechanisms are im;:lependent, and they can be
overlapped with the single read or write operation
permissible in the input-output channel used with
the 2314. Disc rotation time is 25 milliseconds, so
the average rotational delay is 12.5 milliseconds.
Peak data transfer rate between 2314 Disk Storage
and core storage is 312,000 bytes per second.
Each 2314 unit contains nine disc drives; eight
drives are used on-line, and the ninth is a spare. for
use in case one of the other eight becomes inoperable. The control circuits required to connect a
2314 to' a System/360 input-output channel are
included in the 2314, so no external control unit
is required: Each of the active disc drives holds
one 2316 Disk Pack containing 11 discs; 18 of the
22 disc surfaces are used for data storage. Each
disc drive is served by a single comb-like access
mechanism that moves horizontally between the
discs. The access mechanism contains a separate
read-write head for each of the disc surfaces. The
18 tracks, one on each disc surface, that can be
read or recorded upon when the access mechanism
is in any given position, are referred to as a
"cylinder. "
Because of the high data transfer rate, there are
restrictions on the connection of the 2314 to some
models of the System/360. See Section 420:031,
System Configuration, for the rules governing the
connection of the 2314 to a particular System/360
configuration.
There are 200 tracks on each disc surface, defining
200 cylinders per disc drive. Each of the 200
cylinders can store up to 133,384 data bytes, or
266,768 packed decimal digits. The storage capacity of a 2314 is up to 25.87 million bytes, depending on the record format. Three additional
tracks'per surface are available for use as alternates in case difficulty is encountered in reading or
recording on any data track.
The control section of the 2314 interprets and
executes file instructions, performs the required
conversions between serial-by-bit and parallel-bybit data modes, checks the Validity of the data being
transferred, and furnishes status information about
the random access file units to the Processing Unit.
The parity bit is stripped 'off each byte to be recorded in random access storage. The validity
of the recorded information is checked by generating
a string of 16 "cyclic check bits" and appending it
to the end of each·disc record. When the data is
©
Record length and number of records per track are
variable and user-defined; each track has a maximum
data capacity of 7, 188 bytes. When multiple records
are present on a track, however, the total amount
of data that can be recorded is markedly decreased.
The data format is identical with that of the devices
controlled by the 2841 Storage Control Unit. Each
record may contain a key portion; a separate "count"
byte in each record is used to specify the key
length, which can be as long as 255 bytes. The
number of bytes in the data portion of the record is
specified by a two-byte count code. Thus each
record can have a count area, a key (optional), and
a data area. An additional area is included at the
beginning of each track to specify the track address,
and a one-byte code is used to indicate possible
unusable recording portions of the track.
A series of file commands permits any or all of
the three areas (count, key, and data) of a record
in a random access file to be searched, read, or
written. With the use of chained commands, it is
possible to read or write up to 18 successive tracks
(cylinder mode) during 18 successive disc rotations.
Selection of a track is initiated by transferring a
6-byte seek address from the Processing Unit to
the control section of the 2314. The File Scan
Function feature, which permits an automatic
search for a specific identifier or key, is standard
on the 2314. Also standard is the Record Overflc;>w
feature, which permits a single record to overflow
from one track to another. The optional TwoChannel Switch feature enables a 2314 to be switched
from one channel to another under program control.
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
420:048.140
IBM SYSTEM/360
.14
Availability: ••.•••• ?
.15
First Delivery: ••••• 1st quarter, 1967.
.4
CONTROLLER: ••••. no separate controller;
all necessary control
electronics are included
in the 2314 •
• 16
Reserved Storage: ••• none.
•2
PHYSICAL FORM
.5
ACCESS TIMING
.21
Storage Medium: .••. multiple magnetic discs.
.53
Access Time Parameters and Variations
• 22
Physimil Dimensions
.532 Variation in access time -
.222 DiscDiameter: .••••••• 14 inches o. d.
Number on shaft: •.• 11.
• 23
Storage Phenomenon: • magnetization.
.24
Reco'rding Permanence
• 241 Data erasable by. instructions: .•.•.••
• 242 Data regenerated
constantly: ••• • • • .
• 243 Data volatile: •••••••
• 244 Data permanent: .••••
• 245 Storage changeable: .•
.25
Variation,
~
Move head to
selected band:
Wait for beginning
of band:
Transfer data:
Total:
yes.
no.
no.
no •
yes.
Data Volume per Band
Characters: •••.•••• 7,188 (maximum capacity,
with 1 record per track).
Digits: ••••••..••• 7,188 in zoned byte format.
14,376 in packed format.
up to 140
o to 25
25 per band
.6
CHANGEABLE STORAGE
.61
Cartridges
Average,
msec
75
12.5
25
112.5
.611 Cartridge capacity: .•• 25,876,800 bytes (maximum).
• 612 Cartridges per
module: .•••.••.• 1.
· 613 Interchangeable: •.••. yes.
• 62
Loading Convenience
.26
Bands per Physical
Unit: ••••••••.•. 200 per disc surface.
• 27
Interleaving Levels: •• 1; i.e" no interleaving.
.28
Access Teclihique: ••. magnetic heads which move
in unison on a comb-like
access arm mechamism.
• 29
Potential Transfer
Rates: ••••••••••• peak rate is 312,000
-bytes/sec.
.621 Possible loading While computing system
is in use: .•.•.•. yes .
While storage system is in use: ••.. yes, if the particular drive
is not addressed.
.622 Method of loading: ••• operator.
• 623 ApprOximate change
time: •••••••..•. 1 minute .
.624 Bulk loading: •••••.. no; 1 cartridge of 11 discs
at a time .
.3
DATA CAPACITY
•8
.31
Module and System Sizes
Identity:
Disc drives:
Discs:
Tracks:
Cylinders:
Words:
Bytes:'
Packed Digits:
Cartridges online (Disk
Packs):
2316
11 (18
active
surfaces)
3,600
200
6,469,200
25,876,800
51,753,600
2314.
8.
88 (144
active
surfaces).
28,800.
1,600.
51,753,600.
207,014,400.
414,028,800.
8.
.32 Rules for Combining
Modules: ....••.. each 2314 contains 8 disc
drives (one additional
drive is present, but
is only used when one
of the first eight becomes
inoperable) •
7/65
ERRORS, CHECKS, AND ACTION
Error
Check or
Interlock
Invalid address:
Program check
on recorded
address
all 8-bit codes
are valid.
compute cyclic
check code
Invalid code:
Receipt of data:
Recording of
data:
Action
program control.
record check
code.
none except
programmed
read.
Recovery of data: compute' cyclic error signal.
check code and
compare with
recorded check
code
Dispatch of data: attach parity bit
to each byte.
Timing conflicts: interlock
wait.
Wrong track:
program check program control.
on track address
420:051.100
IBM System/360
Central Processors
CENTRAL PROCESSORS
.1
GENERAL
.11
Identity:
. 12
Description
... mM System/360 Processing Units; Models
2030, 2040, 2050, 2065, and 2075. t
The IBM System/360 has been designed to provide, in a single package, nearly all the
features required by any business or scientific computing installation. IBM has, therefore, found it necessary to make available:
(1)
a large amount of directly addressable storage;
(2)
many types of coding; and
(3)
four types of arithmetic.
To meet these demanding requirements, the System/360 has a basic two-address instruction format that provides for direct addressing of over 16 million core positions; uses an
8-bit data code that can represent any of the more familiar 4, 5, or 6-bit codes; and provides fixed-point binary, single and double precision floating-point binary, and variablelength decimal arithmetic. Moveover, there are well-developed facilities for input and
output editing, radix conversions, code translations, supervisory control, and running
more than one program at a time.
However, as in almost all such plans, there are costs involved. In this case, the costs
are shown in the complexity of the instruction repertoire*, in the alignment rules for
operands, and in a lack of uniformity in handling instructions and operands which make
the System/360's repertoire one of the most complex currently existing.
Within such an ambitious central processor design, there are many aspects which deserve
attention; in particular:
•
The basic processor design.
•
The instruction repertoire as a programming tool.
•
The interrupt system.
•
The facilities for multiprogramming.
•
The treatment of errors and special cases.
•
The compatibility question.
•
The available special features.
In this Description, each of these aspects is considered, along with its implications for
the user, in a separate section. Each section is independent, so the sections can be
read in any desired sequence •
• 121 Basic Design
The System/360 Processing Units 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
* There are 22 different Load instructions, 14 different Add instructions, etc.
t See Section 422:051 for details of the 2020 Processor used in System/360 Model 20
systems.
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360
420:051.121
.121 Basic Design (Contd.)
between main storage and peripheral devices. There are, at present, five programcompatible Processing Units, and they are used in the following System/360 models:
•
2030 Processing Unit -
System/360 Model 30; see also Section 423:051.
•
2040 Processing Unit - System/360 Model 40; see also Section 424:051.
•
2050 Processing Unit - System/360 Model 50; see also Section 425:051.
•
2065 Processing Unit - System/360 Model 65; see also Section 426:051.
•
2075 Processing Unit - System/360 Model 75; see also Section 428:051.
Processor Registers
Each Processing Unit has 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. They are specified by the 4-bit R, B, or X fields in many System/360 instructions.
Some operations use two adjacent registers coupled together to provide a 64-bit capacity.
Addressing
Main storage addresses are formed by adding a 12-bit "displacement" (contained in the D
field of every System/360 instruction that references main storage) to a 24-bit "base address"
(contained in a general register specified by the 4-bit B field in the same instruction). The
addresses in many instructions (including most binary arithmetic and logical instructions)
can be further modified by adding a 24-bit "index" contained in a general register specified
by the 4-bit X field in the instruction; this effectively provides a double indexing capability.
All three parts of an address (base, displacement, and index) are treated as unsigned,
positive binary integers and are added together with overflows ignored. Since every
address includes a base, the sum is always 24 bits long; this provides a logical capability
for addressing up to 16,777,216 bytes, although the direct part of the address (the 12-bit
displacement) permits direct addressing of only 4,096 bytes. The base-register technique
of address formation facilitates program relocation and segmentation.
Instruction Format
Instructions can be two, four, or six bytes in length. A 2-byte instruction causes no reference to main storage. A 4-byte instruction causes one reference to main storage, while a
6-byte instruction causes two storage references. There are five basic instruction formats:
•
Type RR - Register to Register (2 bytes)
•
Type RX - Register to Indexed Storage (4 bytes)
Op
•
Type RS - Register to Storage (4 bytes)
Op
•
Type SI - Storage and Immediate Operand (4 bytes)
Op
•
12
I I
B1
/
D1
Type SS - Storage to Storage (6 bytes)
Op
L1
I L2 I B1 I
D1
B2
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-bit operand register specification
X = 4-bit index register specification
7/65
D2
(Contd. )
420:051.1211
CENTRAL PROCESSORS
.121 Basic Design (Contd.)
Fixed-Point Arithmetic
The basic arithmetic mode of the System/360 is fixed-point binary, using 32-bit operands
and two's-complement notation. Most operations can alternatively specify the use of 16bit halfword operands to improve storage utilization. Most 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 4 bytes long when they specify an operand address in main storage (type
RS or RX) and 2 bytes long when both operands are in registers (type RR).
The Standard Instruction Set includes 86 instructions which perform fixed-point arithmetic, comparison, branching, moving, loading, storing, shifting, radix conversion, code translation,
packing, unpacking, and Boolean operations. These 86 instructions are present in all System/360
Processing Units. 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-perbyte format used by most input-output devices and the two-digits-per-byte format used for decimal
arithmetic.
The other instructions in the standard set are quite conventional in form and function, as
shown in the Instruction List (Section 420: 121). However, in addition to performing their
explicit functions, many instructions take action to ensure that valid operations are
being performed upon acceptable operands, and also to set a Condition Indicator that can
subsequently be tested to control conditional branching. These additional processor
functions occur on most, but not all, instructions. Where a check fails (e. g., an invalid
operand, result, or instruction is noted), the program is interrupted and a forced transfer is made to an appropriate routine, with proper linkages being set up to effect a return
to the original program.
Floating-Point Arithmetic
The Floating-Point Arithmetic feature is optional in Models 30 and 40 and standard in the
larger models. It provides 44 additional instructions that provide for addition, subtraction, multiplication, division, loading, storing, and comparison of both "short" (32-bit)
and "long" (64-bit) floating point numbers. The fractional part occupies 24 bits in the
short format and 56 bits in the long format. The characteristic occupies 7 bits in either
format, represents the power of 16 by which the fractional part is to be multiplied, and
permits representation of numbers ranging from 10- 78 to 1075 .
In this type of floatingpoint representation, a "normalized" fraction may contain up to three leading zeros; the
resulting precision is either 6 or 16 decimal digits.
.
There are four 64-bit floating-point registers in the Processing Unit. 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 instructions are of type RX (4 bytes) or RR (2 bytes). Addition and
subtraction may be either normalized or unnormalized.
Decimal Arithmetic
The Decimal Arithmetic feature is optional in Models 30 and 40 and standard in the larger
models. It provides 8 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 L field of the instruction referencing it; the
word mark concept used throughout the IBM 1400 series has been abandoned. Two-address
instructions of the storage-to-storage (SS) type are used for all decimal operations; the
general and floating-point registers are not utilized.
Decimal arithmetic in the System/360 is considerably slower than binary arithmetic. It is
designed for processes which require relatively few computational steps between input and
output, so that radix conversions and use of fast-access registers for temporary storage of
results are not justified.
The editing instruction is part of the Decimal Arithmetic feature. A packed decimal field
is unpacked and edited under control of a pattern, and the edited result replaces the pattern.
Editing can include sign and punctuation control and the suppression and protection of leading
zeros.
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360
420:051.122
.121 Basic Design (Contd.)
The additional cost of the Decimal Arithmetic feature on the Model 30 and 40 Processing
Units is small, and the power of the editing instruction may make the feature worthwhile
even where the decimal arithmetic itself is of little or no direct value .
. 122 The Instruction Repertoire As A Programming Tool
The basic instruction form of the System/360 is two-address (i. e., an operation code
followed by two operands). This address structure, which is familiar because of its
use in the IBM 1400 series, often leads to the destruction of one of the two operands;
thus "ADD A to B" causes the destruction of operand B. If this is not desired, B must
be previously copied into some safe location.
In the System/360 there are two types of machine-language addresses: register addresses and main storage addresses. There are 16 general registers, so 4 bits can
specify a register address; and there may be up to 8 million characters of core storage, requiring at least 23 bits to specify a particular storage address (24 bits are
actually used). There are three possible arrangements of addresses in an instruction:
two main storage addresses, one main storage address and one register address, or two
register addresses. Because a register address is much shorter than a main storage
address, there are three instruction lengths: 2, 4, and 6 bytes (i. e., 16, 32, and 48
bits) . The 24-bit main storage addresses are prohibitively long to be incorporated
directly into the System/360 instructions, so all main storage addresses are in the form
of a 12-bit "displacement" (which is written in the instruction) and a reference to one of
the 16 general registers which contains a 24-bit "base address."
Operand lengths also vary. Fixed-point binary operands have a length of 32 bits, and 16bit "halfword" operands can be used with many instructions. Short floating-point operands
have a mantissa of 24 bits and an exponent (based on powers of 16, not 10) of 7 bits. Long
floating-point operands have the same form of exponent but have 56 bits in the mantissa.
Decimal arithmetic is based on variable-length fields, whose lengths are specified in the
instructions. Character operations are based on 8-bit bytes, and the number of bytes
mayor may not be specified in the instructions.
These different operand lengths are probably unavoidable in such a complex system, but"
in addition to their lengths, there are restrictions (called "alignment ru1es") on the placing of operands within core storage. BaSically, an operand must begin "at a byte position
whose address is an exact mu1tiple of its length (an "integral boundary"). Thus, a 32-bit
(4-byte) fixed-point operand must be placed starting at byte position 800, or 804, or 808 but not 801, 802, 805, 806, etc. Variable-length (decimal) fields can start at any byte
position in core storage.
These two variable factors - instruction length and operand length - will tend to complicate programming of the System/360 at the machine-language level and patching of
machine-language programs.
The Condition Indicator provides a form of running commentary on the program being ex-"
ecuted. The indicator is set by a variety of conditions, depending upon the instruction,
and its setting can subsequently be tested to control conditional branching. Thus, after
an Add instruction the Condition indicator denotes positive, negative, zero, or overflow
values of the sum. Mter a Compare instruction the indicator denotes which of the two
operands is greater, or equality. Mter a Translate and Test instruction the indicator
denotes whether all the function bytes were zero, whether the last function byte was
non-zero, or whether a non-function byte occurred before the first operand field was
exhausted.
The Condition Indicator is a usefu1 feature, requiring no additional execution time and no
specific instructions, but the irregu1arities in its usage, once again, will tend to cause
programmer errors. For example, while "Add A to B" sets the indicator, "Mu1tiply
A by B" leaves the indicator unchanged but still testable. Thus, a program may appear
to be logically correct although it functions incorrectly (e. g., a programmer may think
he is testing the resu1t of a Mu1tiply instruction when, in fact, he is testing the resu1t of
a preceding Add instruction) .
. 123 Interrupt System
A powerful interrupt system permits System/360 Processing Units to respond to a variety
of special conditions arising within the processor and in peripheral units. The basis for the
interrupt system is the Program Status Word (PSW), a double (64-bit) word that indicates
the operational status of a program. When an interrupt condition arises, the active PSW
7/65
(Contd. )
420:051.123
CENTRAL PROCESSORS
. 123 Interrupt System (Contd.)
is automatically stored in a fixed location whose address depends upon the cause of interruption. The Processing Unit then fetches,a: new PSW from another fixed location, and this
new PSW governs entry to a routine which services the interrupt condition. After the
interrupt condition has beep. serviced, the PSW of the interrupted program is restored to
the active position, thereby resetting the Processing Unit to the status it had just before
the interrupt occurred.
There are five classes of interrupt conditions:
•
•
•
•
•
Input-Output - occurs upon termination of an I/O operation,
when an I/O error occurs, or when an I/O device requires
attention.
Program - occurs when an unusual condition is encountered
in the execution of a program (e. g., overflow, invalid address,
invalid operation code, violation of "integral boundary" rules
for fixed-length operands, reference to a protected storage
area, etc.)
Supervisor-Call - switches system status from the program
state to the supervisor state when the Supervisor-Call instruction is executed. (In the program state, all input-output and
some control instructions are invalid.)
External - occurs upon receipt of a signal from the timer, the
interrupt key on the console, or one of six external interrupt
lines provided with the optional Direct Control feature.
Machine-Check - occurs upon detection of a machine malfunction.
Specific bits in the Program Status Word can be used to mask off certain interruption
conditions. When masked off, an interrupt signal may either be ignored or remain pending.
The 64-bit Program Status Word holds enough information about a running program so
that, provided the contents of the 16 general registers are also preserved, it is possible
to interrupt and restart a program without risk. The Program Status Word, whose contents form a running commentary on the operational status of a program, provides the
basic mechanism for both overlapped input-output operations and lUultirunning of independent programs. Both require facilities that can place a running program in "cold
storage" and then bring it back and continue operations as though no interruption had
taken place. In many systems this is not possible because, during the interruption process, some intermediate results might be over-written, flags might be changed, etc.
In the System/360 all intermediate results are held in the 16 general registers or in the
4 floating-point registers. All flags, the location counter, and other necessary data are
held in the Program Status Word. Whenever it be-::omes necessary to change from one
program to another, an interrupt signal automatically stores the present PSW at a fixed
position and then enters the new program with a new PSW stored for this contingency.
The new PSW provides a new location counter setting, new interrupt conditions, and new
flags, all without losing the corresponding data of the preceding program.
Five levels of interruption are provided in the System/360. Each level has its own particular standard position to store a prior Program Status Word and to enter an ana:lysis
routine to determine the cause of the interrupt. Thus, at anyone time, up to six routines
can be ready for immediate execution in the processor: one main program; one routine
for handling an input or output operation; one routine for handling overflow, illega:l addressing, or some other fault in the program being executed; one routine for handling any
reference to the supervisory routine from the main program; one routine for handling any external
interruption (operator, real-time clock, certain types of external equipment, etc.); and
one routine for handling machine faults. * Other main programs can also be in core storage, but they will not become functionally active until the supervisory routine determines
that one of them has the highest priority of any program ready to run, resets its Program
Status Word and register contents, and restarts it as the "active" main program.
*
The System/360 can attempt to recover from many transient machine faults by means of a
specia:l diagnostic instruction. Existing documentation of this feature is vague, but its existence could be of considerable importance in real-time systems.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360
420:051.124
. 124 Multiprogramming Facilities
The capability to run more than one program at a time requires effective solutions to two
major hardware problems. These are the sequencing problem (i. e., providing automatic
switching between programs to maximize overall throughput) and the safety problem (i. e. ,
safeguarding each program from all the others). In the System/360 the necessary functions are performed by a supervisory routine in conjunction with the interrupt system
(described in the preceding section) and two special sets of instructions.
Whenever an interrupt occurs, the running program is safeguarded, and a special routine
is entered which determines the cause of the interrupt and then transfers control to the
supervisory routine. The Supervisor Call instruction, which deliberately causes a further
interrupt, switches the mode of operation of the computer to permit the use of a small
group of "privileged" instructions. These instructions permit changing Storage Protection keys, altering the channel controls, and initiating input-output instructions. Thus,
the partnership between hardware and softw:;tre provides organized sequence control and
a safeguard for programs. This safeguard, the Storage Protection feature, is designed
to prevent one program from overwriting another.
The Storage Protection feature cannot prevent one program from referring to another; any
program can read data from any area of core storage, so the protection is against destruction rather than in favor of privacy. A 4-bit "storage key" is assigned to each 2, 048-byte
block of core storage, and a 4-bit "protection key" is associated with each program (in the
Program Status Word) and with each input and output operation In the case of input operations, the key can be read in with the data itself. Whenever an attempt is made to write
data into core storage, the storage key associated with the block concerned is compared
with the protection key associated with the data to be written. If the two keys match, or if
either key is zero, the data is written into storage. If not, the operation is aborted and
a special interrupt occurs.
A significant loophole in the Storage Protection facility is the inability of the programmer
to be sure that no other programmer has used a protection key of zero on some data (in
which case he could not positively safeguard his program against overwriting) .
. 125 Errors and Special Cases
Errors in the System/360 are handled through the interrupt system in the following ways:
•
megal operation codes and addresses - handled by the supervisory routine,
usually leading to abortion of the program.
•
Input-output errors - handled by the supervisory routine, usually leading
to attempts to repeat the input or output operation successfully.
•
Machine malfunctions - handled by the supervisory routine, leading to
special diagnostic routines that attempt to repeat the instruction and check
on its functioning and on the ability of the program to continue .
. 126 The Compatibility Question
Intermodel Compatibility
mM promises that System/360 Models 30, 40, 50, 65, and 75 will be.-atrictly program
compatible, both upward and downward. This means that any valid program which runs
on System/360 configuration A will run on System/360 configuration B and produce the
same results if:
(a)
configuration B includes the required amount of main storage,
all required input-output devices, and all required special
features; and
(b)
the program is independent of the relations between instruction execution times and input-output rates.
The compatibility rule does not apply to "invalid programs" (programs that violate the
specifications in the programming manual) or to the handling of machine malfunctions.
(Contd. )
7/65
420:051. 126
CENTRAL PROCESSOR
.126 The Compatibility Question (Contd.)
Although the System/360 Processing Units have the same logical structure to provide compatibility, their physical structures differ significantly for engineering reasons. The most
significant physical differences (none of which need to be considered by the programmer)
are summarized in the following table:
Model:
30
40
50
65
75
Main Storage Cycle,
/Lsec:
Bytes Accessed/Cycle:
Register Type:
Register Cycle, /Lsec:
Control Technique:
Control Cycle, /Lsec:
1.5
1
MS
1.5
RO
0.75
2.5
2
CA
1.25
RO
0.625
2.0
4
CA
0.50
RO
0.50
0.75
8
TR
0.75
8
TR
RO
0.25
CC
CA
CC
MS
RO
TR
=
core array
= conventional circuits
= main storage
=
=
read-only storage
transistor registers.
Compatibility with Earlier IBM Systems
IBM will provide optional features for various models of the System/360 that will make it
possible to run programs written for most second-generation IBM computer systems with
little or no change. IBM computers that can be "emulated" on one or more models of the
System/360 include the 1401, 1440, 1460, 1620, 7010, 7040, 7044, 7070, 7074, 7080, 7090,
and 7094. See the report sections on "Compatibility" (Sections 420:131 through 420:136)
for details of the capabilities, equipment requirements, limitations, and performance of
these compatibility features .
•127 Special Features
The Storage Protection feature (optional in Models 30 and 40, standard in the larger models)
prevents accidental alteration of the contents of specified 2, 048-byte blocks of main storage
(see Paragraph .124).
The Interval Timer feature (optional in Model 30, standard in the larger models) occupies
a full word of main storage, holds a signed binary integer, is counted down at the rate of
60 cycles per second, and initiates an interrupt when the value goes from positive to
negative.
The Direct Control feature (standard in Model 75, optional in the others) provides six
external interruption lines and instructions that can transfer a single byte at a time of control information between a special external device and main storage.
The Decimal Arithmetic and Floating-Point Arithmetic features (both optional in Models
30 and 40, standard in the larger models) are described in Paragraph .121 of this report
section.
The Channel-to-Channel Adapter permits high-speed data transmission between two
System/360 Processing Units of any model by linking together two input-output channels one from each of the two Processing Units. When this is done, each Processing Unit can
treat the other as a standard peripheral unit. The adapter can be used on either Multiplexor or Selector Channels and requires one control unit position on each channel.
.13
Availability: ••••.•••.••• ?
.14
First Delivery:
Model
Model
Model
Model
Model
*
30:
40:
50:
65:
75:
••••••••••••••
•.••••••••••••
••••••••.•.•..
•••••••••••••.
•••••••••••••.
2nd quarter,
2nd quarter,
3rd quarter,
1st quarter,
4th quarter,
1965.
1965.
1965.
1966. *
1965. *
Customers who ordered Model 60 or 62 systems will get them, beginning in 3rd quarter,
1965; these will be upgraded to Model 65 systems in 1st quarter, 1966. Customers who
ordered Model 70's will get Model 75's instead.
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360
420:051.200
.2
PROCESSING FACIIJTIES
• 21
Operations and Operands
Operation
and Variation
.211 Fixed pointAdd- subtract:
Multiply Short:
Long:
Divide No remainder:
Remainder:
Provision
Radix
Size
automatic
automatic
binary
decimal (optional
in Models 30 and 40)
full or halfword.
variable: 1 to 31 digits.
automatic
automatic
automatic
binary
binary
decimal
(optional in Models
30 and 40)
halfword (32-bit product) •
full word (64-bit product).
variable: 1 to 15 digits.
none
automatic
automatic
binary
decimal
(optional in Models
30 and 40)
.212 Floating point (optional in Models 30 and 40) Add-subtract: t
automatic
binary
Multiply:
automatic
binary
Divide:
automatic
binary
full word (64-bit dividend).
variable: 3 to 31 digits.
24
56
24
56
24
56
and
and
and
and
and
and
7
7
7
7
7
7
bits
bits
bits
bits
bits
bits
(short).
Oong).
(short).
(long).
(short).
(long).
t Both normalized and unnormalized .
• 213 BooleanAND:
Inclusive OR:
Exclusive OR:
.214 Comparison Numbers:
automatic
automatic
automatic
binary
binary
binary
1 word.
1 word.
1 word.
automatic
fixed point binary:
fixed point decimal:
floating point binary:
32
up
32
32
32
32
or 16 bits.
to 32 digits. *
or 64 bits. *
bits or up to 256 bytes.
bits or up to 256 bytes.
bits or up to 256 bytes.
Absolute:
automatic
Letters:
automatic
Mixed:
automatic
* Optional in Models 30 and 40.
Collating sequence ASCn code: .•••••••••• specials, numbers, letters.
Extended BCD code: ..••• specials, letters, numbers.
(See Data Code Tables, Section 420:141.)
Provision
From
To
Size
.215 Code translation:
automatic**
automatic
automatic
any 8-bit
code
binary
decimal
1 to 256 bytes.
.216 Radix conversion:
any 8-bit
code
decimal
binary
/'
15 digits + sign.
31 bits + sign.
** Special code tables must be provided to use the translate instructions.
Provision
.217 Edit format Alter size:
Suppress zero:
Round off:
Insert point
Insert spaces
Insert fill
character:
Protection:
Float dollar
sign:
7/65
generally make larger
automatic
none
automatic
automatic
Comment
2 to 256 bytes.
can edit multiple
fields with one
instruction
automatic
automatic
semi-automatic
must use instruction
first to indicate position
at which symbol is to be
inserted.
(Contd. )
420:051.218
CENTRAL PROCESSOR
.218 Table look-up: none •
• 219 OthersBinary shift:
Decimal shift:
• 22
Provision
Comment
automatic
semi-automatic
binary
must use Move with
Offset and Logical
Move instructions
Special Cases of Operands
.2372 Names: . . . . . . . . . . (1)
. 221 Negative numbers Binary: . . • . . . . . . . 2's complement and sign
bit.
Decimal: ..••...•. sign in least significant
byte.
. 222 ZeroBinary: .•.•..•.. only positive zero.
Decimal: . . . . . . . . . positive or negative zero;
treated as equal in comparisons.
.223 Operand size determination Binary: •.•.•..•.. fixed size: halfword (16
bits), full word (32 bits),
or double word (64 bits) ,
implied by instruction
used.
Decimal (and certain
logical operations): variable size, indicated by
operand length fields in
instruction.
.2373
.2374
· 2375
. 2376
.23
Instruction Formats
.231 Instruction structure:
1, 2, or 3 halfwords (16,
32, or 48 bits), depending
on number of main storage
addresses necessary.
. 232 Instruction layout and
parts: •.•...•.•.. see "Instruction Format"
in Paragraph 420:051.121.
.234 Basic address structure: . . • • . . . . . . . 2 + 0: variations in instruction length are due to the
fact that either operand
address may be either a
main storage address or
a register address.
.235 LiteralsArithmetic (logical): . 1 byte.
Comparisons and
tests (logical): .•.• 1 byte.
Incrementing modifiers: .•...•...• none; increment is either
-lor contained in a
register .
• 236 Directly addressed operands Internal storage Minimum Maximum Volume
~
size
size
accessible
Core storage:
General
registers:
1 byte
256 bytes
1 register
16 registers
16,777,216
bytes*
16 oneword
registers
* If base registers are used for relative addressing, a maximum of 4,096 bytes are
accessible via each register so allocated.
• 237 Address indexing.2371 Number of methods: . 2.
32 digits.
indexing using the base
register addresses .
(2) indexing using the X
field (in instruction
format RX only) ;
permits double indexing if used with
method (1) .
Indexing rule: . . . . . . base address and index field
are treated as 24-bit
positive binary integers;
displacement is treated as
a 12-bit positive binary
integer. All these are
added to form a 24-bit
binary integer, ignoring
overflows.
Index specification: .. base address (B) field and
index (X) field both specify
the number of a register.
Number of potential
indexers: . . . . . . . 16 .
Addresses which can be indexed Type of address
Application
Storage reference:
all can be indexed by base
register contents.
Storage address in
RX instruction
format: ...•..... can have double indexing (by
base register and index
register).
.2377 Cumulative indexing: . via double indexing and
Execute instruction.
· 2378 Combine index and
step: . • . . . . . . . . none .
. 238 Indirect addressing: .. none.
Note:
The Execute instruction permits one instruction which is not in the direct sequence of instructions to be modified and
executed, followed by an automatic return
to the next instruction in the original sequence.
· 239 Stepping
· 2391 Specification of increment: •..••.... always minus one for Branch
on Count; for Branch on
Index, the increment is
found in a register .
. 2392 Increment sign: .•.. minus for Branch on Count;
minus or plus for Branch
on Index.
.
.2393 Size of increment: •.. always one for Branch on
Count; 32 bits for Branch
on Index.
.2394 End value: . . . . . . . . implied as zero for Branch
on Count; for Branch on
Index, the value is in a
storage location specified
by the instruction .
.2395 Combined step and
test: . . . . . . . . . . . yes.
© 1965 AUER.BACH Corporation and AUERBACH Info, Inc.
7/65
420:051.240
.24
IBM SYSTEM/360
Special Processor storage
.241 Category of
storage
Number of
locations
Size in
bits
Program
usage
General registers:
16
32
indexing,
base addresses,
and accumulators,
Floating-point
registers:
4
64
floatingpoint operations.
Program Status
Word:
1
64
holds location counter and
various
flags.
Channel Control
Word:
64
1
· 332
holds I/O
control
information.
Note: The physical characteristics of the above
registers are summarized in Paragraph
420:051.126.
.3
SEQUENCE CONTROL FEATURES
. 31
Instruction Sequencing
.333
· 334
.311 Number of sequence
control facilities: "
11 Program Status Words
(PSW), only one of which
is active at a time.
.312 Arrangement: . . . . . . one PSW for initial program
loading, two for each of
the 5 types of interruption: I/O, program, supervisor-call, external, and
machine check.
. 313 Precedence rule: . . . . priority of interrupts:
(1) machine check
(2) program or supervisorcall
(3) external
(4) I/O.
.314 Special sub-sequence
counters: . . . . . . . . the length of variable-size
operands is held in decimal
arithmetic instructions.
. 315 Sequence control step
size: . . . . . . . . . . . halfword.
.32
Look-Ahead: . . . . . . . none.
.33
Interruption (see also Paragraph 420:051.123)
. 331 Possible causesInput-output units: .. unit available.
unit ceased transmission.
unit malfunction before
transmission starts.
Input-output controllers: . . . . . . . controller available.
controller ceased transmission (perhaps because of
error noted during transmission).
· 335
.336
controller malfunction before
transmission starts.
Processor errors: .. illegal operation code.
operation code and data incompatible.
overflow, underflow, or
divide error.
all-zero floating-point
result.
operand incorrectly aligned.
violation of storage protection.
Other: . . . . . . . . . supervisory routine violation.
system malfunction.
external action from console
or another system.
Control by routine Individual control: . . . . . . . . . . acceptance or non-acceptance
of I/O interrupts can be
controlled by channel.
length checks are controlled
by specific instruction.
Method: . . . . . . . . specific bits in Program
Status Word. (These bits
are normally controlled by
the system, not by the
user's program. )
Operator control: ... operator may only initiate a
request for an external
interrupt .
Interruption conditions: . . . . . .
(1) interruption condition
signalled.
(2) interruption condition
attains the necessary
priority on a channel
to be forwarde-d to the
computer interface
(I/O interrupts only).
(3) this interrupt not masked
out by program or system masks .
(4) processor in a mode of
operation in which this
type of interrupt is
allowed to occur.
Interruption process Interruption action: present PSW (Program
Status Word) is stored
and replaced by a standby
PSW .
Registers saved: .. none of the 16 general registers are saved automatically (they can be saved
conveniently by the Store
Multiple instruction). Most
of the necessary operational
data is saved in the old PSW .
Destination: •••••. contained in ·standby PSW;
one of 5 locations corresponding to the 5 levels of
interrupt conditions.
Control methods Determine cause: •• analysis of flags by appropriate standard routines.
Enable interruption: •••..•..•• by setting of bits in the PSW
or an I/O control word.
(Contd. )
7/65
420:051.340
CENTRAL PROCESSORS
.34
Multiprogramming
Model
Model
Model
Model
Model
(Multiprogramming is the process of intermingling
instructions from several independent programs. )
.341 Method of control: ••• handled by the Operating
System/360; see Section
. 420:191.
.342 Maximum nwnber of
programs: ••••.•• undefined to date; probably
15 .
• 343 Precedence rules: ..• undefined to date.
• 344 Program protectionStorage: .•.•..••• each 2, 048-byte block can
be protected by a "storage
key" of 0 through 15. Any
attempt to store data with
a different "protection
key" - unless either key
is 0 - causes an interrupt.
Input-output units: •• no hardware protection; it
will be provided by the
standard control programs.
• 35
Multisequencing
(Multisequencing is the process of intermingling
instructions from several different sets which have
precedence interlocks; i. e., from semi-independent
sequences of a program. Multisequencing usually
implies the use of two or more separate central
processors. )
Facilities for multisequencing, using two or more
interconnected Processing Units, are available in
the Operating System/360.
.4
.5
30
40
50
65
75
The performance of each System/360 Processing
Unit, in terms of both basic instruction times and
speeds on our standard measures of processor
performance, is shown in the Central Processor
section of the appropriate subreport:
©
Section 423:051Section 424:051Section 425:051Section 426:051Section 428:051.
ERRORS, CHECKS, AND ACTION
Error
Check or Action
Interlock
Overflow:
check
forced transfer to
program interrupt
routine .
Underflow:
(floatingpoint):
check
forced transfer to program interrupt routine.
Zero
divisor:
check
illegal data:
check
forced transfer to program interrupt routine.
forced transfer to program interrupt routine .
Forbidden
operation:
check
forced transfer to
supervisor interrupt
routine.
Unavailable
operation:
check
forced transfer to program interrupt routine.
Illegal storage
address:
check
forced transfer to
program interrupt
routine.
parity
check
forced transfer to
machine error or
input/output interrupt routine.
Receipt of
data:
PROCESSOR SPEEDS
-
Dispatch of
data:
send parity
bit.
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
420:061.100
IBMSystem/360
Console
CONSOLE
.1
GENERAL
. 11
Identity: ..
.12
Associated Units: . . . . 2150 Console (optional)
with Models 65 and 75
Processing Units.
1052-Printer-Keyboard
(optional).
units of an IBM 1050 Data
Communication System
(optional) with Model 30
only.
2250 Display Unit (optional)
with Models 65 and 75
only.
IBM System/360 System
Control Panel *
• Wait Light - indicates that the processor is not
performing operations, but is able to respond
to interrupt signals or to initial program loading.
• Manual Light - indicates that the processor is
in the stopped state and will respond to start
key or initial program loading.
• System Light - indicates that processor meter
or customer-engineering-meter is running. This
light is on when processor is running or when
an input-output operation is in the operating
state.
• Load-Unit Switches - three rotary switches
which provide the 11-bit address of the device
to be used for initial program loading. They are
set to an awkward three-group code; i. e., 0-7,
0-9, A-F.
. 13
Description
. 131 System.Control Panel
A system control panel built into each System/360
Processing Unit provides the switches, keys, and
lights needed to operate and control the system.
Both the equipment and the supervisory programs
are designed to reduce operator errors by minimizing the need for manual intervention. The system control panel enables the operator to reset all
system components, to store and display information in main storage and processor registers, and
to load initial program information. The system
controls are divided into three sections: operator
control, operator intervention, and customer engineering control. Each section is separately described below.
Operator Control
The operator control section contains only the controls needed when the system is operating under
full supervisory program control. The same controls are provided for all System/360 Processing
Units. The main functions provided by the operator control section are the control and indication
of power, the indication of system status, operatorto-machine communication, and initial program
loading. The major keys, switches and lamps are
listed below.
• Emergency Pull Switch - removes all system
power, and latches in the out position.
•
Power-On Key - lights when power comes on;
system reset function is performed.
• Interrupt Key - used to initiate an external program interrupt.
*
.
• Load Key, Load Light - used in initial program
loading .
This report section does not cover the Model 20's
console facilities; see page 422:011. 100.
Operator Intervention
The operator intervention section provides the
controls needed to enable the operator to intervene
into the normal operation of the system. This includes the system reset and store-and-display
functions.
The intervention controls vary in form and placement from model to model, but functional compatibility is maintained. The major keys and switches
are listed below.
e System Reset Key - resets the Processing Unit,
input-output channels, and all on-line, nonshared (by another processor) control units and
input-output devices. The processor is set to
the stopped state. System reset occurs also
when initial program loading or power-on sequence occurs.
•
Stop Key, Start Key.
•
Rate Switch - selects run or step mode of instruction operation.
e Storage-Select and Address Switches - select
storage type (main or registers) and actual
address of location for I-word display or entry.
•
Data Switches - used to set up binary word for
manual entry.
•
Store Key, Display Key.
•
Set IC Key- - permits entry of an address into
the instruction-address portion of the current
program Status Word.
•
Address-Compare Switches - used to select
mode for processor to stop at address set up
on Address Switches.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
420:061.131
.131 System Control Panel (Contd.)
Customer Engineering Control
The customer engineering section is different for
each processor model and contains controls intended
for use only by customer engineers. Some System/
360 storage, channel, and control units contain separate customer engineering controls.
.132 2150 Console
The optional IBM 2150 Console is a free-standing
console desk that duplicates the System/360 operator controls at a distance from the processor. The
2150 is usable only with the Model 65 and 75 Processing Units, and uses either an available position of a
Selector Channel, or one available position and one
subchannel of a Multiplexor Channel. A single 2150
Console can contain one or two independent operator control panels and control each processor in
a two-processor system. (In this case, one of the
processors can be a Model 50.) The 1052 PrinterKeyboard (next paragraph) can be connected to
the 2150 Console.
"BM SYSTEM/360
each other. The keyboard is used for input to the
system, and the printer accepts output from the
system. Printing speed is 14.8 characters per
second. An I/O interruption can be initiated from
the keyboard.
A 1052 can be connected to a.Model 40, 50, or 65
system via a 1052 Adapter in the Processing Unit,
or to Model 65 or 75 via a 2150 Console. The 1052
Adapter uses one available position and one subchannel of a Multiplexor Channel, or one available
position of a Selector Channel (Model 65 only). The
1052 can be connected to a Model 30 system via a
1051 Control Unit and a 1051 Attachment feature.
.134 1050 Data Communication System
The IDM 1050 Data Communication System, usable
with Model 30 only, can include the 1052 PrinterKeyboard, a' separate printer, a card reader or
punched-tape reader, and a card punch or tape
punch. It requires the 1051 Control Unit, which
transfers data at 14.8 characters per second to or
from any of the 1050 system units.
.135 Display Console
.133 1052 Printer-Keyboard
The optional IBM 1052 Printer-Keyboard provides
two-way communication between a System/360 and
its operator. The typewriter-style keyboard and
the Selectric typing mechanism are independent of
7/65
A 2250 Modell Display Unit can be used as a
console for Model 65 or 75 Processing Units. As
in the 2150 Console, up to two operator control
panels can be incorporated. See Section 420:101
for a detailed description of the 2250 Display Unit.
420: 071. 100
IBM System/360
Input.Output
2540 Card Read Punch
INPUT·OUTPUT: 2540 CARD READ PUNCH
.1
GENERAL
. 11
Identity:
.12
Description
into the EBCDIC internal code. Invalid codes and
hole count errors cause an indicator to be set.
. . . . . . . . . 2540 Card Read Punch,
Modell.
The 2540 Card Read Punch, Modell, is a new version of the 1402 Card Read Punch; it will replace
the 1402 Model N1 in all applicable orders for a
System/360. The 2540 Modell consists of a 1,000card-per-minute :..'eader and a completely independent 300-card-per-minute punch housed in the
same cabinet. The most significant differences
from the 1402 Model N1 are increased punching
speed (from 250 to 300 cards per minute), increased capacities of the hoppers and stackers,
the addition of a pre-stacker station in the reader
section and different styling. The changes in
styling ~ermit the loading and unloading operations
to be performed more easily.
The addition of a card station between the second
read brushes and the stackers in the reader section
increases the usefulness of the selectable-stacker
feature. When reading, selection of cards to be
separated is normally based on information in the
cards themselves. Without the separate prestacker card station, .the decision to select must be
made between the time the card is read and the
time the card arrives at the stackers. This arrangement in many cases precludes the use of
multiple input areas for the card reader, particularly in a multiprogramming environment, due to
the uncertain amount of time between the reading
of the card and the processing of the data read
from the card. The pre-stacker card station in
the 2540 eliminates this problem since the stackerselect decision is made at the time of the next card
feed operation.
The 2540 is controlled, buffered, and connected to
an available position on a System/360 channel by a
2821 Control Unit. The rules governing the connection of 2540 to a particular System/360 configuration of the 2540 to a particular System/360 configuration are presented in Section 420:031, System Configuration.
. 121 Reader Unit
Cards are fed from a 3, 100-card hopper, in rowby-row fashion, past two sets of reading brushes
and into one of three 1, 350-card stackers where
they are stacked on end. Cards go to the "normal"
stacker (R1) unless they are directed by the program to stacker R2 or RP3. Stacker RP3 is shared
by the punch unit and is normally assigned to either
the reader or the punch during a single run. The
first row of brushes provides a hole count, and the
second row reads the data and checks the hole
count. Data is automatically translated from the
Extended BCD Interchange (EBCDIC) card code
The card cycle time, at full speed, is 60 milliseconds. Three clutch points are provided as a
standard feature. Data is read into any area of
core storage specified by the programmer.
Since the cards are stacked on end, there are limitations on the types of cards which can be used
with the reader; the manufacturer should be con- .
·sulted on this point.
Optional Reader Features
The 51-column Interchangeable Read Feed provides
for the reading of either 51- or 80-column cards.
When this feature is present, reading speed is permanently reduced to 800 cards per minute, and
the capacity of the normal stacker and stacker
number 1 are permanently reduced to 800 cards.
Selection of 51- or 80-column reading is made by
the operator.
The Column Binary feature permits reading of
binary cards; the contents of each card column
are inserted into two byte positions in core storage, with the two high-order bits of each byte
position automatically set to zero.
. 122 Punch Unit
Cards are fed from a 1, 350-card hopper, in rowby-row fashion, past a punch station and a read
station and into one of three 1,350-card stackers
where they are stacked on end. Cards go to the
"normal" stacker (PI) unless they are directed by
the program to stacker P2 or RP3. Stacker RP3
is shared with the reader unit. Reading brushes in
the read station are used for hole-count checks
upon the punched data; they cannot be used to read
data into the system. Unless the Column Binary
feature is incorporated, use of prepunched cards
in the punch unit causes incorrect hole-counts,
which set a check indicator. Data is 'automatically
translated from the internal Extended BCD Interchange Code into the EBCDIC card code.
The card cycle time, at full speed, is 200 milliseconds. Four clutch points are provided. Data
can be transferred to the punch from any area of
core storage specified by the programmer.
Optional Punch Features
The Punch Feed Read feature provides a read station ahead of the punch station, which reads
pre~ously-punched data from the card into core
storage. Additional data can then be punched into
the same card at the punch station.
The Column Binary feature permits punching of
binary cards by punching the low-order six bits of
two adjacent byte positions into each card column.
© J 965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
420:072.100
IBM System/360
Input-Output
1442 Card Read Punch
INPUT-OUTPUT: 1442 CARD READ PUNCH
.1
GENERAL
. 11
Identity: . . . . . . . . . . 1442 Card Read Punch,
Models N1, N2, and 5.
. 12
Description
The 1442 Card Read Punch is a combination inputoutput unit for standard 80-column punched cards.
It is cheaper and slower than the 2540 Card Read
Punch, but the 1442 does not have independent
punch and read facilities.
The three models (N1, N2, and 5) described here
are functionally identical except that Model N1 has
two 1, 200-card radial stackers and Models N2 and
5 have one 1, 300-card radial stacker. Rules governing the connection of the various 1442 models to
a particular System/360 configuration are presented
in Section 420:031, System Configuration.
From a single 1, 200-card hopper, the cards are fed
serially by column past a single photoelectric reading station, past a single punching station, and into
a radial stacker. In Model N1, cards go to stacker
1 unless they are program-directed to stacker 2.
Cards can be loaded and removed without stopping
the unit. All format control is by the stored program: there is no plugboard.
Peak reading speed is 400 cards per minute. Punching rate is 160 card columns per second, resulting
in speeds ranging from 91 cards per minute when
all 80 columns are punched to 360 cards per minute
when only 1 column is punched in each card. Similar and slower versions of the 1442 Card Read
Punch are used with the IBM 1440 system.
A single 1442 can handle either an input or an output file. Alternatively, it can read data from and
punch results into the same cards, or the results
can be punched into trailer cards in the same file.
When cards are being read continuously and no
punching is being done, maximum speeds are 400
cards per minute, with a corresponding card cycle
time of 150 milliseconds. Each column is read
twice and the results are compared. Checks are
made for invalid characters, improper registration,
off-punching, and malfunctions of the light-sensing
mechanism.
These models of the 1442 normally use an extended
card code that permits any of 256 code combinations
to be punched into each card column. These 256
combinations include the common Hollerith punch
combinations, automatically translated from the
internal EBCDIC 8-bit code. The optional Card
Image feature, available for Models N1 and N2,
"packs" the low-order 6 bits of 160 consecutive
bytes of main storage into a single card, 2 bytes
per column. When data punched in the Card Image
mode is read back into core storage, the two highorder bits of each byte position are automatically
set to zero.
7/65
Reading begins at column 1 and can be terminated
after 1 to 80 columns have been read by a count
specified in the instruction. Decreasing the number of columns read has no effect on the input card
rate, but it does increase the available computing
time per card.
Card feeding is asynchronous; there are no discrete
clutch points as in the: IBM 1402 Card Read-Punch.
This means that if processing time slightly exceeds
overlap time, card input speed will be reduced
proportionately in the 1442; with a 1-point clutch,
the input speed under the same circumstances
would be cut in half. When the peak card reading
speed cannot be maintained, the card input cycle
time is increased from 150, to 160 milliseconds.
The increase is caused by the clutch pickup time
when feeding on demand.
Card punching, like reading, is performed serially
by column, always begins at column 1, and can be
terminated after any number of columns from 1 to
80. The cards to be punched can be either blank
or prepunched. Skipping is not possible; each
blank column requires a full punch cycle. An echo
check is made to insure that the correct dies have
been actuated for each character.
Card punching speeds depend upon the number of
consecutive columns punched in each card; punching
proceeds at 160 columns per second. To the punching times must be added the time to feed and position each card: 160 milliseconds. Resulting
punching speeds can be summarized as follows:
Number of Columns
Punched, P
1:
10:
20:
40:
80:
Formula:
(cards per minute)
punchin, Rate,
Cards min.
360
270
210
146
91
60,000
6. 25P + 160
When reading and punching ar'e done on the same
card, the reading occurs during the punch positioning time of 160 milliseconds, and the overall
read-punch operation proceeds at the same speed
as punching alone.
The 1442 contains all necessary control circuits,
so no external control unit is required. However,
a special adapter is required to connect Model 5 to
a Model 20 Central Processor.
Optional Feature
Card Image (Models N1 and N2 only): provides
direct input to and output from core storage of any
card code, as described above; the code translation
and character validity check are bypassed. The
feature is useful for reading and punching column
binary cards or any cards that use non-standard
codes.
420:073.100
IBM System/360
Input-Output
2671 Paper Tape Reader
INPUT-OUTPUT: 2671 PAPER TAPE READER
.1
GENERAL
. 11
Identity: . . . . . .
.12
Description
2671 Paper Tape Reader.
The punched tape can be 5-track telegraphic code
(11/16 inch wide); 6- or 7-track (7/8 inch); or 8track (1 inch). The peak speed of 1, 000 characters
per second (when reading strips) is reached after
an acceleration time of about 18 milliseconds. When
using the optional spooling facilities, the data rate
varies between 500 and 1, 000 characters per second
The IBM 2671 Paper Tape Reader is a new unit that
reads fully-punched tape photoelectrically at a peak
speed·of 1,000 characters per second. The 2671 is
normally mounted on top of its controller, the IBM
2822 Paper Tape Reader Control Unit. See Section
420:031, System Configuration, for the rules governing the connection of this device to a particular
System/360 configuration.
Parity checking is done by the 2822 control unit.
Code translation, when required, is performed by
the processor. The control unit provides status
information, end-of-record signals, 'and end-of-tape
signals to the processor.
The basic reader accepts only strips of punched
tape, in any length conveniently handled by an ope~
ator. With optional supply features, reels 10.5
inches in diameter can be used (either center-fed
or outside-fed). The reels can be rewound automatically.
Buttons and switches on the reader are used to set
particular end-of-record codes, and to specify
tape code, width, parity, and delete-code recognition. Lamps provide the operator with status
information.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
420:074.100
IBM System/360
Input-Output
2501 Card Reader
INPUT-OUTPUT: 2501 CARD READER
.1
GENERAL
.11
Identity: . . . . . . . . . . 2501 Card Reader.
Models A1, A2, B1, and
B2.
. 12
Description
The 2501 Card Reader is one of a group of new peripheral devices developed specifically for the mM
System/360. Models A1 and B1 are functionally
identical, as are Models A2 and B2. Models A1 and
B1 read standard 80-column cards at a peak rate of
600 cards per minute. Models A2 and B2 read 80column cards at a peak rate of 1, 000 cards per
minute. In all four models, reading is performed
by photo-electric cells which read one column at a
time. See Section 420:031, System Configuration,
for the rules governing the connection of the various
2501 models to a particular System/360 configuration.
Cards are fed from a 1, 200-card hopper past a read
station and into a single 1, 300-card stacker. Each
column is read twice by the 12 photoelectric cells
at the read station. The character code sensed by
the first reading is sent to a check register; data
from the second reading is sent to the buffer register. The contents of the two registers are then
compared; failure of this comparison sets the unit
check status indic ator .
The normal mode of reading for the 2501 uses an extended card code which has 256 valid code combinations. The card codes are automatically translated
into the internal EBCDIC 8-bit code.
The optional Card Image feature, available for
Models B1 and B2, allows a second mode of reading,
under program control. In the Card Image mode,
7/65
the information in each card column is read into the
low-order six bits of two consecutive byte pOSitions
of core storage. The high-order two bits of each
byte position are automatically set to zero. This
method of column binary reading is compatible with
the method used by other System/360 card readers
and punches .
Reading begins at column 1 and can be terminated,
after 1 to 80 columns have been read, by a count
specified in the instruction. Reducing the number
of columns read does not affect the card reading
rate, but it does increase the available computing
time per card.
To maintain the peak reading speed of 600 cards per
minute in Models A1 and B1, the next read command
must be given within 47.4 milliseconds after completion of the previous read operation. The corresponding time to maintain the peak reading rate of
1,000 cards per minute in Models A2 and B2 is 7.4
milliseconds. If these limits are exceeded, the
reader will be delayed for one full cycle (100 milliseconds for Models A1 and B1, 60 milliseconds
for Models A2 and B2), since the 2501 has only a
Single clutch point.
The 2501 Card Reader contains all necessary control circuits, so no external control unitis required.
However, a special adapter is required to connect
Model A1 or A2 to aSystem/360ModeI20Processor.
Optional Feature
Card Image (Models B1 and B2 only): Provides
direct input to core storage of any card code, as
described above; the code translation and character
validity checks are bypassed. This feature is useful for reading column binary cards or any cards
using a non- standard code.
420:075.100
IBM System/360
Input-Output
2520 Card Read Punch
INPUT-OUTPUT: 2520 CARD READ PUNCH
.1
GENERAL
.11
Identity:
.
2520 Card Read Punch,
Models Al and B1.
2520 Card Punch, Models
A2, A3, B2, and B3.
.12
Description
The 2520 is one of a group of new peripheral devices designed specifically for the System/360.
Models Al and Bl are functionally identical and
provide facilities for both card reading and card
punching in a single card path. Reading only,
punching only, or combined reading and punching
can proceed at a maximum rate of 500 cards per
minute. The punching rate of 500 cards per minute is the highest punched-card output rate announced to date by a maj or manufacturer.
Models A2 and B2 incorporate only the punching
facilities of Models Al and BI, respectively, and
have a peak output rate of 500 cards per minute.
Models A3 and B3 are slowed-down versions that
incorporate only punching facilities and have a peak
output rate of 300 cards per minute.
The rules governing the connection of the various
models of the 2520 to a particular System/360 configuration are presented in Section 420:031, System
Configuration.
Cards are fed in parallel (row-by-row) from a
1, 200-card hopper, serially (column-by-column)
through a solar-cell read station (in Models Al and
BI only), in parallel through a punching station,
and then into one of two 1, 300-card stackers, selected under program control. In the card read
section (models Al and Bl only), checks are made
for improper functioning of the sensing mechanism,
invalid character codes, off-register punching,
and mispositioned cards. Punching accuracy is
checked by comparing a signal generated as each
hole is punched with the data in the output area.
Error cards are automatically directed to Stacker 2.
In the normal mode of operation, an extended card
code is used which has 256 valid code combinations.
Automatic translation takes place between the card
code and the internal EBCDIC 8-bit code.
An optional Card Image feature, available for
Models Bl, B2, and B3, allows the translation circuitry to be bypassed. In this mode the low-order
6 bits of two byte positions in core storage correspond to one card column. When reading, the two
high-order bits of each byte are set to zero. This
method of column binary reading and punching is
compatible with the method used by other System/
360 punched card devices.
In Models Al and BI, cards are automatically
halted after reading and prior to punching. Cards
can be advanced from the pre-punch station and
the input hopper independently or simultaneously.
Commands are available to load the punch buffer
(contained in core storage) without advancing the
cards; punching is initiated by the next read instruction, which always advances the cards at
both stations.
To maintain peak operating speeds, the following
timing considerations must be observed:
•
Read-only operations (500 cpm) - the next
read instruction must be received within 41. 7
milliseconds after completion of the previous
operation.
•
Punch-only operations (500 cpm) - the next
punch instruction must be received within 9
milliseconds after completion of the previous
operation.
o Combined read and punch operations (500 cpni)
the next read instruction must be received within
45 milliseconds, or the next punch instruction
must be received within 9 milliseconds, after
c.ompletion of the previous operation.
o Punch-only operations (300 cpm) - the next
punch instruction must be received within 18
milliseconds after completion of the previous
operation.
All models of the 2501 contain the necessary control
circuits, so no external control unit is required.
However, a special adapter is required to connect
Model AI, A2, or A3 to a System/360 Model 20
Processor.
Optional Feature
Card Image (Models Bl, B2, and B3 only): provides direct input to and from core storage of any
card code, as described above; the code translation and character validity checks are bypassed.
This feature is useful for reading and punching
column binary cards or any cards using a nonstandard code.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
42:076.100
IBM System/360
Input-Output
2560 MFCM
INPUT-OUTPUT: 2560 MULTIPLE FUNCTION CARD MACHINE (MFCM)
.1
GENERAL
.U
Identity: . . . . . . . . . . 2560 Multiple Function
Card Machine (MFCM).
. 12
Description
The 2560 is a unique punched-card input-output
unit developed especially for the IBM System/360
Model 20. Equipped with two 1, 200-card feed
hoppers, a reading station, a punching station, an
optional printing station, and five 1, 300-card
radial stackers, the 2560 MFCM combines many
of the facilities of a card reader, gang punch,
summary punch, collator, interpreter, and card
document printer in a single unit under storedprogram control. In conjunction with the Model 20
Processing Unit, the 2560 can also perform the
functions of a calculator or accounting machine.
Cards can be fed independently from either the
primary or secondary hopper; they follow separate
paths tbrough pre-read, read, and pre-punch stations. The cards are read serially (column-bycolumn) by means of solar cells, at a maximum
speed of 500 cards per minute. (This speed is
achieved only when no punching or printing is being
done on the cards.) The sensing mechanism (which
serves both the primary and secondary feed paths)
is checked for proper functioning during each read
cycle. Checks are also made for invalid character
codes, off-register punching, and mispositioned
cards. Sixty-four characters of the Extended BCD
Interchange Code can be recognized.
.
Upon leaving the separate primary and secondary
prepunch stations, the cards merge into a single
feed path tbrough the punch, pre-print, and print
stations. (The print station functions only when the
optional Card Print feature, described below, is
installed.) 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.
The effective speed depends (as in the IBM 1442)
upon the position of the last column punched in each
card:
7/65
Last column punched, P
10
~
60
80
Speed, cards per minute
260
M5
112
91
Formula (cards per minute):
60,000
6.UP + 170
Punching accuracy is checked by comparing a signal generated as each hole is punched with the data
in the output area in core storage. Either blank or
prepunched cards can be punched.
When multiple operations are being performed
simultaneously, the overall tbroughput rate is equal
to the rate of the slowest operation.
Personnel familiar with unit record machines having multiple card paths and card stations, such as
a collator, should not experience undue difficulty in
programming input-output operations using the
2560 MFCM.
Optional Feature
Card Print: Provides a printing unit that can print
two, four, or six lines of information on any or allcards passing through the MFCM. Each line can
hold up to 64 printed characters, spaced 10 to the
inch. Each of the 2, 4, or 6 print heads can be
manually positioned to print in anyone of 25 line
pOSitions, which extend from above the 12-punch
. row to below the 9-punch row. Each head can print
any of 63 characters, or blanks, as transmitted
from core storage under program control. Rated
printing speed is 140 character positions per second, regardless of the number of print heads that
are simultaneously activated at each position. As
in the punch station, actual printing tbroughput
depends upon the location of the last position to be
printed on the longest line of each card:
Last position printed, P
Speed, cards per minute
10
40
60
Formula (cards per minute):
296
143
106
60,000
7. 23P + 130
420:081.100
IBM System/360
Input-Output
1403 Printer
INPUT-OUTPUT: 1403 PRINTER (MODELS 2, 3,7, AND N1)
.1
GENERAL
.11
Identity:.........
. 12
Description
1403 Printer, Models 2, 3,
7, and N1.
Four models of the mM 1403 Printer are available
for u~e with the System/360. See Section 420:031,
System Configuration, for the rules governing the
connection of the various models to a particular
System/360 configuration.
Model 2 operates at a peak speed of 600 lines per
minute and has 132 printing positions. This model
is also used in several mM 1400 and 7000 Series
data processing systems, where its horizontal-chain
printing mechanism has earned it a reputation for
high-quality printing. The Numerical Print special
feature, which permits speeds of up to 1,285 lines
per minute when printing all-numeric data in the
older systems, is not available for use in System/
360. The 1403 Model 2 is analyzed in detail in
Section 401:081 of the mM 1401 report.
The 1403 Model 3 is a newer, faster version of
Model 2, which also has 132 printing positions. It
is rated at 1,100 alphameric lines per minute and
uses a "train" of type slugs (three characters per
slug) moving through a horizontal channel instead
of the horizontal-chain printing mechanism used in
Model 2. The 1403 Model 3 is analyzed in detail in
Section 402:082 of the mM 1410 report.
Model 7 is an economy version of Model 2. The
size of the print line is reduced to 120 printing
positions in Model 7, and the dual-speed skipping
feature is not incorporated. Several of the optional
features available for Model 2 are not available for
Model 7 (see Table I). Peak printing speed remains
at 600 lines per minute .
The 1403 Model N1 replaces the 2201 Model 3
Printer which was previously offered with some
models of the System/360. Model N1 is functionally
identical to the 1403 Model 3 except that Model N1
has only 120 printing positions. The most significant physical difference between Model N1 and
Model 3 is the addition of a hydraulically-operated
acoustical cover to Model N1 in an effort to reduce
the noise problem.
Table I summarizes the main characteristics of
each model.
The dual-speed carriage of Models 2, 3, and N1
enables skips of more than 8 lines to be performed
at 75 inches per second (600 lines per second at
6 lines per inch). Skips of fewer than 8 lines on
these three models, and all skips on Model 7, are
performed at 33 inches per second. The printing
speeds of the models described here, as a function
of average line spaCing, are shown in the graph
on page 420:081.121.
Line spacing is 6 or 8 lines per inch, set by the
operator. Continuous marginally-punched forms
from 3.5 to 18.75 inches j.n width can be used.
The minimum form depth is 1 inch, while the
maximum is 22 inches at 6 lines per inch or 16. 5
inches at 8 lines per inch.
TABLE I: CHARACTERISTICS OF 1403 PRINTER, MODELS 2, 3, 7, AND N1
Model 2
Print cycle time at single
spacing, msec:
Peak speed (with full
character set), lines/min:
Speed at I-inch line spacing,
lines/min:
Number of printing positions:
Number of printable characters:
Skipping speed, inches/sec:
Printing mechanism:
100
600
480
132
48
33/75
Horiz.
chain
Model 3
54.5
1,100
755
132
48
33/75
Horiz.
train
Model 7
100
600
480
120
48
33
Horiz.
chain
Model N1
54.5
1,100
755
120
48
33/75
Horiz.
train
Special features available Numerical Print:
Preferred Character Set:
Universal Character Set:
Selective Tape Listing:
Interchangeable Chain
Cartridge Adapter:
Auxiliary Ribbon Feeding:
No
No
Yes
Yes
No
No
Yes
Yes
No
No
No
No
No
No
Yes
Yes
Yes
Yes
Standard
Standard
Yes
Yes
Standard
Standard
© 1965 AUERBACH Corporation and AUERBACH Info, InC!.
7/65
420:081.120
. 12
Description (Contd.)
The printers receive a line of edited output data
(one character per byte) from the computer and
store it in a line buffer. Each printer makes a
timing/echo check on the actuation of the print
hammer magnets and compares the data printed
with the data in the print buffer.
mM 1403 Printers are connected to a System/360
(except Model 20) data channel via a 2821 Control
Unit. Several models of this control unit are
available for controlling up to three 1403 printers
(in any combination of models) and one 2540 Card
Read Punch. A special 1, 100-LPM Printer Adapter
is required for each 1403 Model 3 or N1 Printer
attached. One or more adapters, depending on the
special features incorporated in the printer, are
required to connect a 1403 Printer to a System/360
Model 20.
Optional Features
Type font changes in all models can be made by the
operator, using removable chain or train cartridges;
Models 2 and 7 require the optional Interchangeable
Chain Cartridge Adapter, whereas the facility is
standard in Models 3 and N1.
The optional Selective Tape Listing feature, available for Models 2, 3, and N1, permits up to 13
characters to be printed on each of up to eight
1. 5-inch tapes. The tapes can be individually
advanced under program control, but no forms
skipping can be done while the feature is in use.
Two 1. 5-inch tapes can be replaced by one 3.1inch tape, on which 29 characters can be printed.
Any character set normally available for these
models can be used. Primary use of this feature
has been with magnetic ink character recognition
equipment, in providing batch control over checks.
For satisfactory utilization of polyester film ribbons,
the Auxiliary Ribbon Feeding feature (optional on
Models 2 and 7, standard on Models 3 and N1) is
recommended. This type of ribbon provides improved print quality and is normally required if the
printed documents are to be read by the 1418 or
1428 Optical Character Readers.
The Universal Character Set feature, available
for Models 2, 3, and N1, makes it possible for a
1403 to print any set of graphics containing up to
240 different characters. The graphics can be
arranged in any desired sequence on the print
IBM SYSTEM/360
chain (in Model 2) or train (in Models 3 and N1),
and selected "preferred" characters can be repeated as many as 15 times around the chain or
train to improve the effective printing speed.
The Universal Character Set feature uses a special
240-character read/write storage unit within the
2821 Control Unit. Each position in the storage
unit corresponds to one graphic on the print chain
or train; the code contained in each position defines
the internal data code that will cause the corresponding graphic to be printed. The,contents of the
240-character storage unit can/be changed, by
means of data loaded from punched c.ards, in order
to change the relationships between internal data
codes and printed graphics. If a code in the data
record being printed fails to match any of the codes
in the 240-character storage unit, an error indication will be sent to the processor.
Any chain or train configuration previously /
announced for mM 1400 Series systems is aVailable for a System/360 equipped with the Universal
Character Set feature. In addition, mM offers six
new chain/train arrangements which are designed
for efficient performance in functions such as
commercial and scientific text printing, commercial applications of FORTRAN and COBOL, and
use of the PL/I language. The "TN" arrangement,
for example, contains 120 graphics, including both
upper and lower case alphabetics, digits, superscripts, and a variety of special symbols; the full
120-character set occurs twice on the 240-character chain or train. The "QN" arrangement contains 60 graphics; 45 "preferred" graphics are
repeated five times around the chain or train,
while the other 15 graphics occur only once. The
Universal Character Set feature also permits users
to design chain or train arrangements tailored to
their own special needs.
Speeds attained with the Universal Character Set
feature depend upon the frequency of appearance
upon the print chain or train of the characters to
be printed. The chain or train always moves past
the printing positions at the same speed, so the
character sets containing the largest number of
different characters yield the slowest printing
speeds. When a 120-character text printing
arrangement is used, nominal speeds for singlespaced printing are 273 lines per minute for Model
2 and 563 lines per minute for Models 3 and N1.
In no case can the printing speed exceed 750 lines
per minute in Model 2 or 1,400 lines per minute
in Models 3 and N1.
(Contd. )
7/65
INPUT-OUTPUT: PRINTERS
420:081.121
EFFECTIVE SPEEDS OF SYSTEM/360 PRINTERS
6,000
5,000
4,000
3,000
2,000
roo...
1,000
900
800
""
"-
K.
.......... """-
700
600
Printed
Lines
Per
Minute
i"""
500
"""-
400
r-..;
I- r-...
200
1403 MLel 2*
~I
r-... roo...
300
- ~ --r---.
deis 3, N1*
""'" t"-o ""'"
~
f'..
l' ....
.......
'"
~
100
90
80
70
~
~
~
-
---r------
~
r--. ~ r-----.- r-----I---
60
50
40
30
20
o
1/2
1
2
3
4
5
Interline Spacing in Inches
* With standard 48-character chain/train.
** With 52-character typebar.
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
420:082.100
IBM System/360
Input-Output
1404 Printer
INPUT-OUTPUT: 1404 PRINTER
.1
GENERAL
.11
Identity: . . . . . . . . . . 1404 Printer, Model 2.
. 12
Description
The 1404 Printer uses the same horizontal-chain
printing mechanism, tape-controlled carriage,
and continuous-form feeding and stacking system
as the 1403 Printer, Model 2, used with the System/360 and other mM systems. Like the 1403
Model 2, it has 132 printing positions and a peak
speed of 600 lines per minute. In/addition, the
1404 has a feeding and stacking system for card
forms. By unlocking a single knob, tlie entire
printing assembly can be moved laterally to print
on either continuous forms or cards (but not on
both during the same run) . The card transport
mechanism can handle single 51- or 80-column
cards, 80-column cards with 80-column stubs,
or two 51- or 80-column cards fed side by side
("two-up"). Peak speed, when feeding two cards
at a time and printing one line per card, is 800
cards per minute. Up to 25 lines can be printed
on a standard mM card. Section 401:082 of the
mM 1401 report contains a more complete description of the 1404 Printer. Deliveries of the
1404 started in 1962.
Cards can be printed on either side depending on
how they are loaded into the hopper; however, cards
are stacked in reverse of the order in which they
are loaded into the hopper. This condition will present file sequencing problems if additional processing is required.
7/65
With the optional Read-Compare special feature,
up to 30 columns of punched information (specified
by plugboard wiring) read from the card itself can
be printed on the card during the same pass (or
used for computation by the system, with the
results printed on the same card during the pass).
The Interchangeable Chain Cartridge Adapter
special feature is available, as on the 1403 Printer,
to adapt the printer for convenient changing of type
fonts for special printing jobs.
The variations in card feeding and printing rates,
as the number of printed lines per card varies,
are shown in the following table.
Lines printed per card
1
2
4
8
16
Speed, cards per minute
(one-up!two-up)
400/800.
240/480.
136/272.
72/144.
37/74.
Printing speed on continuous paper forms, as a
function of average line spacing, is shown by the
curve for the 1403 Model 2 in the graph on Page
420:081. 121.
A single 1404 is controlled, buffered, and connected to a System/360 channel by a 2821 Control
Unit, Model 4. See Section 420:031, System
Configuration, for the rules governing the connection of a 1404 Model 2 Printer to a particular
System/360 configuration.
420:083.100
IBM System/360
Input-Output
1443 Printer
INPUT-OUTPUT: 1443 PRINTER
.1
GENERAL
• 11
IdentitY:.........
• 12
Description
1443 Printer, Model N1.
The 1443 Printer is a low-speed line printer originally developed for use with the mM 1440 system.
The 1443 utilizes a horizontal oscillating typebar
printing mechanism whose chief advantage is rapid
interchangeability of character sets. Peak speed,
using the standard 52-character set, is 240 singlespaced or double-spaced lines per minute. Using
the optional character sets, peak speeds range from
200 to 600 single- or double-spaced lines per
minute:
•
13-character set
600 lines per minute
•
39-character set
300 lines per minute
•
52-character set
240 lines per minute
•
63-character set
200 lines per minute.
See the graph on page 420:081.121 for the effective
speeds of the ,1443 Printer when multiple-line skips
are made between printed lines.
The standard printed line is 120 characters long,
at a horizontal spacing of 10 characters to the inch.
An additional 24 printing positions are available as
an optional feature. Line spacing can be set at six
or eight lines per inch by an external switch. Continuous fanfold, sprocket-punched forms can vary
from 4 to 16.75 inches in overall width. Maximum
form length is 22 inches at 6 lines per inch and 16. 5
inches at 8 lines per inch. Skipping speed is
©
approximately 15 inches per second, and skipping
and spacing are controlled by a 12-channel paper
tape loop.
All editing and format control is performed by the
stored program; there is no plugboard. A "sync"
check is made to insure that each print hammer was
actuated at the proper time. A print error causes
the setting of a testable indicator.
The 1443 Model N1 contains its own control circuits
and storage buffer, so no separate control unit is
required for use in the System/360. See Section
420:031, System Configuration, for the rules
governing the connection of Model N1 to a particular
System/360 configuration.
Operational Features
Selective Character Set feature: Required for use
of any of the optional typebars described below.
Character Sets: Can be interchanged with the
standard typebar in approximately 15 seconds.
A dial must be manually set to the proper number of characters. The following sets are available:
*
•
13-character: 0-9 and specials . -
•
39-character: 0-9, A-Z, and specials. , $
•
52-character: 0-9, A-Z, and 16 specials
•
63-character: 0-9, A-Z, and 27 specials.
Additional Print Positions: Provides 24 additional
print positions, expanding the print field from 120
to 144 positions. Printer operation is unchanged.
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
420:084.100
IBM System/360
Input-Output
1445 Printer
INPUT-OUTPUT: 1445 PRINTER
.1
GENERAL
.11
Identity: . . . . . . . . . . 1445 Printer, Model Nl.
. 12
Description
The 1445 Printer, Model Nl, is a modified version
of the 1443 Printer (Section 420:083) that can print
both the A. B.A.-approved font E-13B characters
and conventional characters in magnetic ink. The
standard character set contains 56 characters: 26
alphabetics, 10 numerics, 6 special symbols, and
the 14 font E-13B symbols. The magnetic-ink ribbon feed can be r.eplaced by a conventional ribbon
feed for conventional printing requirements. The
peak printing speed, using the standard character
set, is 190 lines per minute. The 1445 Printer,
Model Nl, is presently available for use with the
System/360 Model 30 only.
7/65
There are 113 print positions on a line, spaced at
8 characters per inch across a print span of 14
inches. Vertical spacing can be set at six or eight
lines per inch, as in the 1443. Forms feeding and
sizes are the same as for the 1443 printer .
The Selective Character Set special feature
provides a choice of two· non- standard typebars.
The 14-character (numeric plus specials) typebar
allows a peak printing rate of 525 lines per minute.
The 42-character alphameric type bar (alphabetics, numerics, and six specials) allows printing
speeds up to 240 lines per minute when the Font
E-13B symbols are not required.
The 1445 Printer, Model Nl, contains its own control circuits and storage buffer, and can be connected directly to a control unit position of either
a Multiplexor or Selector channel.
First customer delivery of a 1445 Printer was
made in May, 1965.
420:085.100
IBM System/360
Input-Output
2203 Printer
INPUT-OUTPUT: 2203 PRINTER
.1
GENERAL
. 11
Identity:.........
. 12
Description
2203 Printer, Model AI.
The 2203 Printer is a speeded-up version of the
1443 Printer. Currently, the 2203 is available
only for the Model 20, and it requires a special
attachment on the processor, as do other Model 20
peripheral devices. The peak speed with the 52character set, the one most commonly used, is
350 single-spaced lines per minute. Using optional
character sets, peak speeds range from 300 to 750
single-spaced lines per minute:
and spacing are controlled by a 12-channel paper
tape loop. Two sets of forms can be printed
simultaneously or independently, under program
control, with the Dual Feed Carriage optional
feature .
All editing and format control is performed by the
stored program; there is no plugboard. A "sync"
check is made to insure that each print hammer was
actuated at the proper time.
Optional Features
Character Sets: Can be interchanged with the standard typebar in approximately 15 seconds. A dial
must be manually set to the proper number of characters. The following sets are available:
•
13-character set
750 lines per minute.
•
39-character set
425 lines per minute.
•
13-character: 0-9 and specials
•
52-character set
35.0 lines per minute.
•
39-character: 0-9, A-Z, and specials. , $
•
63-character set
300 lines per minute.
•
52-character: 0-9, A-Z, and 16 specials
•
63-character: 0-9, A-Z, and 27 specials.
The effective speeds for larger interline spacings
with the 52-character set are shown in the graph
on page 420:081.121.
The standard printed line is 120 characters long, at
a horizontal spacing of 10 characters to the inch.
An additional 24 printing positions are available as
an optional feature. Line spacing can be set at six
or eight lines per inch by an external switch. Continuous fanfold, sprocket-punched forms can vary
from 4 to 16.75 inches in overall width. Maximum
form length is 22 inches at 6 lines per inch and
16.5 inches at 8 lines per inch. Skipping speed is
approximately 15 inches per second, and skipping
©
.-*
Additional Print Positions: Provides 24 additional
print positions, expanding the print field from 120
to 144 positions. Printer operation is unchanged.
Dual Feed Carriage: Two tape-controlled paper
drive units are provided, one at each end of the
platen. Two sets of forms can be printed upon
and advanced individually or in unison under program control. The main carriage has a 12channel tape; the other carriage tape has 6
channels standard, with an additional 6 optional.
The 13-character set cannot be used in combination with this feature.
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
420: 091. 100
IBM System/360
Input-Output
2400 Series Magnetic Tape Units
INPUT-OUTPUT: 2400 SERIES MAGNETIC TAPE UNITS
.1
GENERAL
.11
Identity:
Note:
. 12
. 2401 Magnetic Tape Unit,
Models 1, 2, 3.
2402 Magnetic Tape Unit,
Models 1, 2, 3.
2403 Magnetic Tape Unit and
Control, Models 1, 2, 3.
2404 Magnetic Tape Unit and
Control, Models 1, 2, 3.
2803 Tape Control.
2804 Tape Control.
2816 Switching Unit,
Modell.
This report section does not cover the
2415 Magnetic Tape Unit; see Section
420:093.
on each character rather than a bit-by-bit comparison of the recorded character code with the
correct one.
The methods used to check that the data has been
read correctly from magnetic tape into the System/
360 computers are distinctly different from the
checking methods used in the older IBM magnetic
tape units. The parity of each individual character
and the parity of each tape row are checked, as
before. In addition, a new check called the Cyclic
Code Check is performed upon each tape block,
or physical record. This check is based upon an
eight-bit character which is computed from the
data characters during the write operation and
recorded at the end of the tape block. During the
read operation this check character is recomputed,
and the neWly-computed version is compared with
the check character that was originally recorded .
Description
The new IBM 2400 Series Magnetic Tape Units
read and record data in 9 tracks (8 data, 1 parity)
on half-inch tape at a density of 800 bytes per inch.
Each byte holds one alphameric character or two
decimal digits and is transferred to or from core
storage without translation, packing, or unpacking.
The ability to read backward is a standard feature
that will help to speed tape sorting operations.
The 2400 Series includes the following tape drives,
tape controls, and combination units:
•
2401 Magnetic Tape Unit (one drive).
•
2402 Magnetic Tape Unit (two drives in a
single cabinet).
•
2403 Magnetic Tape Unit and Control (one
drive and a read or write control).
•
2404 Magnetic Tape Unit and Control (one
drive and a simultaneous read-write control).
•
2803 Tape Control (read or write) .
•
2804 Tape Control (simultaneous read-write).
Each of the four tape drives (2401, 2402, 2403,
and 2404) is available in three models, whose
characteristics are summarized in Table 1.
See Section 420:031 for the rules governing the
connection of these magnetic tape units to a particular System/360 configuration.
Checking that correct information has been recorded
upon the tape is handled, as in the 7-track IBM 729
tape units, by reading back each character after it
has been recorded. With the increased number of
tracks, the degree of assurance provided by this
check is appreciably reduced because the readafter-write check consists of a simple parity check
© 1965
The 2400 Series magnetic tape equipment has the
ability to reconstruct small amounts of data lost
through any cause. The first bit lost from any
tape block can always be recovered, and in favorable
circumstances it may be possible to reconstruct
additional bits. The data reconstruction occurs
during an automatic repeat read operation which
edits the data being read through the use of special
circuitry.
Each tape control can control up to eight tape
drives of any model. The simultaneous readwrite controls connect to two channels and permit
a read operation on one tape drive to be overlapped
with a write, erase, or write tape mark operation
on another drive connected to the same control.
Tape controls can be connected to either Selector
or Multiplexor Channels. When a Multiplexor
Channel is used, the tape control always operates
in the "burst mode" (i. e., no other I/O device can
use the channel simultaneously and, in Models 30
and 40, no overlapped internal processing can
occur).
The basic IBM 2816 Model 1 Switching Unit permits
four 2400 Series tape drives to be addressed by
either of two single-channel tape controls under
program control. Various adapters are available
to permit up to four single-- channel control units
to address a common pool of up to eight tape
drives. With the 16-Drive Addressing feature and
a second 2816 Model 1 Switching unit, up to 16
tape drives can be addressed by 2, 3, or 4 controls.
Any 2400 Series tape units and controls except the
dual-channel models 2404 and 2804 can be included
in the tape-SWitching pool. The control units in this
pool can be connected to separate channels of the
same processing unit or to channels of different
processing units. In general, there can be as many
simultaneous magnetic tape data transfer operations
as there are control units in the pool. Thus, in
the maximum configuration, any 4 of the 16 tape
drives in the pool can be simultaneously transferring
AUERBACH Corporation and AUERBACH Info, Inc.
7/65
420:091.120
. 12
IBM SYSTEM/360
Description (Contd.)
data. Note that each control unit must be addressed
individually by the program; i. e., an I/O command
to a magnetic tape unit in this pool that is directed to
a busy controller will not be automatically switched
to a non-busy controller.
The optional Seven-Track Compatibility feature
enables a 2400 Series tape drive and control to
read and write seven-track tape at a density of
200, 556, or 800 characters per inch, thereby
achieving tape compatibility with the IBM 729 and
7330 Magnetic Tape Units used in most current
IBM systems - and with many competitive tape
units. Conversions between the seven-track BCD
tape code and the System/360 code are performed
automatically. Tapes written in seven-track format by a System/360 can be read backward, but
tapes written by other systems cannot. This compatibility feature also permits the System/360 to
read tapes recorded on IBM 7701 and 7702 Magnetic
Tape Terminals. A tape drive can have either a
nine-track or a seven-track head installed, but not
both.
.213 Feed drive: . . . . .
. 214 Take-up drive: ..
· 22
· 13
Availability:.......?
. 14
First Delivery: . . . . . April, 1965.
.2
PHYSICAL FORM
· 21
Drive Mechanism
. 211 Drive past the head: ..
· 212 ReservoirsNumber: . . . . . . . .
Form: . . . . . . . . . .
Capacity: . . . . . . . .
7/65
pinch roller friction.
2.
vacuum columns.
about 7 feet.
Sensing and Recording Systems
.221 Recording system: ... magnetic head.
.222 Sensing system: . . . . . magnetic head.
· 223 Common system: . . . . two-gap head provides readafter-write checking.
.23
Multiple Copies: . . . . none.
· 24
Arrangement of Heads
Use of station: . . . . . . recording.
Stacks: . . . . . . . . . 1.
Heads/ stack: . . . . . . 9.
Method of use: . . . . . 1 row at a time.
Use of station: . . . . . . sensing.
Distance: . . . . . . . . . 0.3 inch after recording
head.
Stacks: . . . . . . . . . . . 1.
Heads/ stack:. . . . . . . 9.
Method of use: . . . . . 1 row at a time.
Use of station: . . . . . . 2-stack unit, as above, for
recording and sensing 7track tapes (optional
feature; replaces 9-track
heads).
Stacks: . . . . . . . . . . . 1 for recording, 1 for
sensing.
Heads/ stack:. . . . . . . 7, as special compatibility
feature, in place of
standard 9-head stack.
Method of use: . . . . . 1 row at a time.
Installations using the Seven- Track Compatibility
feature may also add the Data Conversion feature,
which converts three 8-bit System/360 bytes to
four 6-bit tape characters or vice versa. The Data
Conversion feature facilitates the handling of binary
data on seven-track tapes. These binary tapes
can be read only in the forward direction.
All operations performed by the tape controls and
tape units result from the four basic System/360
input-output instructions: Start I/O, Test Channel,
Test I/O, and Halt I/O. The Test Channel instruction permits general testing of channel status, while
the Test I/O instruction presents more detailed information about channel and device status and the
presence of errors. The Halt I/O instruction provides for. an arbitrary termination of an operation.
The Start I/O instruction initiates an input-output
operation based on one of six possible input-output
commands found in a Channel Command Word located in main storage: Read, Write, Read Backward, Control, Sense, and Transfer In Channel.
These general commands are interpreted for the
device specifically addressed. The first three
transfer data to or from main core storage. The
Sense instruction permits more detailed interrogation of status and error conditions of the device,
and the Control command initiates various control
operations of the device (backspace, reWind, etc.).
Transfer In Channel causes a transfer to a different
area of main storage for obtaining subsequent
Channel Command Words.
motor .
motor.
.3
EXTERNAL STORAGE
.31
Form of Storage
.311 Medium:. . . . . . .
plastic tape with magnetizable coating.
.312 Phenomenon: . . . . . . . magnetization.
.32
Positional Arrangement
.321 Serial by: . . • . . • . . . 12 to N rows at 800 rows
per inch; N limited by
available core storage.
.322 Parallel by: . . . . . . . 9 tracks, standard.
7 tracks, using SevenTrack Compatibility
special feature.
.324 Track useData: . . . . . . . . . . . 8.
Redundancy check: . . 1.
Timing: . . . . . . . . . o (self-clocking).
Control signals: ... . o.
Unused: .. .
O.
Total: . . . . . . . . . . 9 .
With optional Seven-Track
Compatibility feature Data: . . . . . . . . . . . 6 .
Redundancy check: . . 1.
Timing: . . . . . .
0 (self- clocking).
Control signals:.
o.
Unused: . .
o.
Total: . . . . . . .
7.
(Contd. )
INPUT-OUTPUT: 2400 SERIES MAGNETIC TAPE UNITS
.325 Row useData: • • • . • . . . . . .
Redundancy check: ..
Timing: . . . . . . .
Control signals: . . . .
Unused: . . . . . . . . .
Gap:
O.
O.
O.
O. 6 inch (nominal).
See Data Code Tables,
Section 420:141.
Com~atibility:
.34
Format
. 35
Physical Dimensions
with IBM 729 and 7330 tape
units when optional SevenTrack Compatibility
feature is used.
.351 Overall width: . . . .
.352 Length:
0.5 inch.
50 to 2,400 feet per reel.
........
.4
CONTROLLER
. 41
Identity: . . .
.42
2403 Magnetic Tape Unit
and Control (single
channel).
2404 Magnetic Tape Unit
and Control (simultaneous
read/write capability).
2803 Tape Control (single
channel).
2804 Tape Control (simultaneous read/write capability) .
Connection to System
.421 On-line 2403, 2803: . . . .
each requires one available
control unit position on a
System/360 Multiplexor or
Selector Channel.
2404, 2804: . . . . . . . each requires one available
control unit position on
each of two System/360
Multiplexor or Selector
Channels.
.422 Off-line: . . . . . . . . . none.
.43
The control units used are the single-channel units
(2403, 2803). The basic 2816 switch provides
switching between two tape controls and four tape
drives. The largest system permits switching
between 4 tape control units and 16 tape drives.
1 to N.
2.
...........
Coding: . . . . . . . . . . one 8-bit byte per tape row .
.33
420:091.325
Connection to Device
.431 Devices per controller: 1 to 8 tape drives total, in
any combination of 2401's
or 2402's, unless 2816
Model 1 Switching Units
are connected. In this
case, the single-channel
control units (2403, 2803)
can address up to 16 tape
drives.
.44
.441 Size of load: . . . . . . .
.442 Input-output areas: ...
.443 Input-output area
access: . . . . . . . . .
. 444 Input-output area
lockout: . . . . . • . . .
•
Permits a control unit to address up to 16 tape
drives (using two 2816's, Modell).
©
each byte .
.5
PROGRAM FACILITIES AVAILABLE
.51
Blocks
.511 Size of block: . . . . . . 1 to N bytes, limited by
available core storage.
.512 Block demarcation Input: . . . . . . . . . . . gap on tape or count in
command.
Output: . . . . • . . . • . count in command.
.52
Input-Output Operations
.521 Input: . . • • . . . . . • . . read data from tape, forward or backward, into
core storage, with input
stopped by count or gap.
Data read backward is
placed in descending
.order in main storage.
· 522 Output: . • . . . • . . . . . write one block forward
from core storage.
· 523 Stepping:
none.
skip forward and erase
· 524 Skipping:
defective tape areas.
skip one block forward or
backward during a read
operation, with data
transfer inhibited.
· 525 Marking: . . . . .
inter-block gap.
end-of-file mark (tape
mark) .
. 526 Searching: . . . . . . . . none.
.53
Code Translation: ... matched codes (i. e., no
translation), except when
using the optional SevenTrack Compatibility
feature. In that case,
translation between internal 8-bit bytes and
6-bit BCD tape codes is
automatic. With 7track recording, the optional Data Conversion
feature permits three
8-bit bytes to be recorded
as four 6-bit tape characters.
.54
Format Control: . . . . none.
The 2816 Switching Unit:
Permits a group of tape drives to be addressed
by anyone of up to four control units, under
program control.
1 to N bytes.
core storage.
blocks of 2, 048 bytes can
be protected (optional on
Model 30 and 40 Processing Units, standard
on larger models) .
.445 Table control: . . . . . . yes, using data chaining in
Channel Command Words .
. 446 Synchronization: . . . . automatic.
2816 Switching Unit, Modell
•
Data Transfer Control
1965 AUERBACH Corporation and AUERBACH Info, Inc.
IBM SYSTEM/360
420:091.550
. 55
Control Operations
.8
Request interrupt: ... automatic.
Select format: . . . . . . yes (using optional Data
Conversion feature with
7 - track recording).
Select code: . . . . .
see "format" above.
yes.
Rewind: . . . . . . .
yes.
Unload: . . . . . . . .
.56
Error
Check or Interlock
Recording:
read-after-write parity check
lateral, longitudinal,
and cyclic parity
checks
set indicator.
byte count check
Storage Protection
set indicator.
program interrupt.
Reading:
Testable Conditions
Disabled: . . . . . . . . .
Busy device: . . . . . . .
Output lock: . . . . . . .
Nearly exhausted: ...
Busy controller: . . . .
End of medium marks:
File protect condition: .
.6
PERFORMANCE
. 61
Conditions: . . . .
.62
ERRORS, CHECKS AND ACTION
yes.
yes.
yes.
no.
yes.
yes.
yes.
standard operation of
2400 Series tape drives,
except where use of the
optional Seven-Track
Compatibility feature is
indicated.
Input area
overflow:
Output block
size:
byte count check
Invalid code: all 8-bit codes are
valid.
Imperfect
see recording check,
medium:
above.
Timing coninterlock
flicts:
Invalid
command: check by channel
Speeds
.621 Nominal or peak speed (bytes/sec):
Standard
Operation
Modell:
30,000
Seven-Track Compatibility
Feature
7, 500 at 200 bits/inch
20, 800 at 556 bits/inch
30,000 at 800 bits/inch
Model 2:
60,000
15,000 at 200 bits/inch
41,700 at 556 bits/inch
60,000 at 800 bits/inch
Model 3:
90,000
22, 500 at 200 bits/inch
62,500 at 556 bits/inch
90,000 at 800 bits/inch.
.622 bnportant parameters: see Table I.
.623 Overhead: . . . . . . . . . see Table 1.
.624 Effective speeds: ..•. see Table 1.
(See also graph of effective
speed, Page 420:091.900.)
.63
Demands on System: .. varies with System/360 model
and type of channel used;
see Simultaneous Operations, Section 420:111.
.7
EXTERNAL F ACIIJTIES
.73
Loading and Unloading
.731 Volumes handled Reel: . . . . . . . . . . . 2,400 feet; holds a maximum of 15,600,000 bytes
when in blocks of 1, 000
bytes.
.732 Replenishment time: .. 1. 0 to 1. 5 minutes; tape
unit must be stopped.
. 734 Optimum reloading period Modell: . • . . . . . . . 12.9 minutes.
Model 2: . . . . • . . . . 6. 5 minutes.
Model 3: . . . • . . . . . 4.3 minutes.
Invalid data
address:
check
Data or command chaining error: check
Channel control signal
parity
error:
check
I/O interface
invalid
signal (timing, parity,
wrong bit
combinations, format):
check
Input
rate too
high for
address
resolution: check
Note:
set indicator and
prepare for rereading with
attempted reconstruction of bad
data.
set indicator.
wait, or set
indicator and
interrupt.
program-check
or interruption condition
set.
terminate operation; set indicator; interrupt.
set indicator.
set indicator;
possible terminate operation.
terminate operation, set
indicator.
terminate operation, set
indicator.
These error indications and other status
information are tranEmitted from the Tape
Control to core storage in response to a
Sense Command .
(Contd. )
INPUT-OUTPUT: 2400 SERIES MAGNETIC TAPE UNITS
420:091.900
EFFECTIVE SPEED
2400 SERIES TAPE UNITS
1,000,000
7
4
2
100,000
ModelS
7
4
Bytes per Second
~
10,000
7
./
4
1,000
.~
LL
V
~""
~
-----
~
",; 1-0-""
Model 2
I
MJde~ 1
~
./
~
!;i'
~
I'"
V
""
~ ""
./
2
2
---
~
~
/
i..-
)1
V V
~
7
4
2
100
10
2
7
4
2
100
4
7
2
4
1,000
7
10,000
Bytes per Block
TABLE I: CHARACTERISTICS OF IBM 2400 SERIES MAGNETIC TAPE UNITS
Model
No.
Tape
Recording
Speed,
Density,
inches
bits per
per sec inch
Peak
Speed,
bytes
per sec
Interb~~ck
Gap Lengths
nominal)
incbes
msec (1)
Efficiency,
% (3)
bytes (2)
100-byte
blocks
1,000-byte
blocks
Demand on
Core
Storage,
%
Full
Rewind
Time,
minutes
1
37.5
800
30,000
0.6
16.0
480
17.2
67.5
(4)
2
75.0
800
60,000
0.6
8.0
480
17.2
67.5
(4)
1.4
3
112.5
800
90,000
0.6
5.3
480
17.2
67.5
(4)
1.0
(1)
(2)
(3)
(4)
3.0
Time in milliseconds to traverse each interblock gap when reading or writing consecutive blocks.
Number of byte positions occupied by each interblock gap.
Effective speed at the indicated block size, expressed as a percentage of peak speed.
Varies widely with System/360 model and type of channel used; see Simultaneous Operations, Section 420:111.
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
420:092: 100
IBM System/360
Input-Output
7340 Hypertape Drive
INPUT-OUTPUT: 7340 HYPERTAPE DRIVE
.1
GENERAL
. 11
Identity: .
.12
Description
The IBM 2816 Model 2 Switching Unit permits four
or eight Hypertape Drives to be switched to any
one of up to four 2802 controls, under program
control. With the 16-Drive Addressing feature
and a second 2816 Model 2 Switching Unit, up to
16 drives can be switched between 2, 3, or 4
controls.
7340 Hypertape Drive,
Model 3.
The 7340 Model 3 is a new version of IBM's cartridge-loaded Hypertape Drive that operates at
either of two recording densities: 1,511 or 3,022
bytes per inch. At the 3,022 byte-per-inch densitythe highest density yet announced for a commercial
magnetic tape drive - a data transfer rate of
340,000 bytes per second (or 680, 000 packed decimal digits per second) will be achieved. At the
1,511 byte-per-inch density, the 7340 Model 3 can
read tape written by the 7340 Modell or 2 Hypertape
Drives used in other IBM systems, with character
.code translations performed by the System/360
program. According to IBM, tapes written by
Model 3 can be read by Models 1 and 2, despite the
fact that Model 3 tapes have a shorter interblock
gap. The data transfer rate at the lower density is
170,000 bytes (or 340,000 digits) per second. Density is selected under program switch control.
Ten tracks are recorded on one-inch-wide tape.
Eight tracks are used for data (one byte per tape
row), and two tracks are used for double odd
parity checking. This enables the Hypertape
Drive to detect all single and double bit read
and write errors and, during reading, to correct
all single-bit errors and 33 of the possible 45
double-bit errors, including all double-bit errors
in adjacent tracks.
The read-write head is located in a port at the
side of the ri.ght-hand vacuum column. The surface
of the head is shaped in such a way that, as the
tape moves past the head for reading and recording,
an air film is created which lifts the tape slightly
off the head. IBM states that the tape is in contact
with the head only when the tape is stationary.
A major feature of the Hypertape Drive is its cartridge loading technique, which facilitates tape
handling and reduces the posibility of tape contamination or damage. Supply and take-up reels
holding 1,800 feet of tape are enclosed in l\. sealed
cartridge that measures about 17 by 10 by 2 inches
and weights about 8 pounds. The operator loads
a reel of tape by simply raising the top cover,
sliding the cartridge into place, lowering the cover,
and depressing the load-unload button. Then the
tape reels move backward to engage the hubs, the
tape is lowered into the vacuum columns, and the
read-write head moves into position. Unloading is
accomplished by reversing the procedure. It is
not necessary to rewind the tape before unloading.
The file protection device on each cartridge can be
turned on manually or by program control; resetting must be done manually when the cartridge
is not loaded on the drive.
Except for the addition of the 3, 022 byte-per-inch
density, the 7340. Model 3 is quite similar to the
7340 Modell as analyzed in Section 417:093 (IBM
7080). Tape speed is 112. 5 inche~ per second,
"nominal read/write access time" is 3.5 milliseconds (compared to 4.2 in Modell), and rewind
time is 1. 5 minutes per full reel. Hypertape
Drives can read backward as well as foward. Interblock gap spacing is reduced to 0.38 inches, nominal, from 0.45 inches in ModelL (The 7340
Model 2 Hypertape Drive is used in the IBM 1401,
1410, and 1460 systems. Model 2 operates at a
tape speed of 22.5 inches per second, which reduces the data rate to 34, 000 bytes per second.)
Up to eight 7340 Model 3 Hypertape Drives can be
connected to a 2802 Hypertape Control, which requires an available control unit position on a MuItiplexor or a Selector Channel. See Section 420:031,
System Configuration, for the ruIes governing the
connection of Hypertape Drives and Controls to a
particuIar System/360 configuration.
The 2802 Hypertape Control Unit can control either
a read or write operation, but not both at one time;
no dual-channel controller is yet available for Hypertape in the System/360. A MuItiplexor Channel (or
a Selector Channel when used with a Model 30
processor) can operate only with the Hypertape
Drive set at "low" density (170,000 bytes per second).
Data chaining cannot be used with Hypertape Drives
on Model 30 and 40 Processing Units. Note also that
the 2802 Hypertape Control carmot be attached to a
?ystem/360 Model 30 or Model 50 Multiplexor
Channel if the Additional Multiplexor Subchannels
feature is installed.
An optional Automatic Cartridge Loader unloads
one Hypertape cartridge and loads another in less
than 30 seconds under manual or program control.
.13
Availability:...
?
.14
First Delivery:
? (7340 Model 2's were
delivered in December,
1963. )
·2
PHYSICAL FORM
· 21
Drive Mechanism
· 211 Drive past the head: . . tape is drawn past readwrite head by vacuum
action of reservoirs.
(Contd. )
7/65
AUERBAC~
INPUT-OUTPUT: 7340 HYPERTAPE DRIVE
.212 ReservoirsNumber: . . . . . • . . .
Form: . . . . . . . . . .
Capacity: . . . . . . . .
. 213 Feed drive: . . . . . . . •
.214 Take-up drive: . . . . .
.22
2.
vacuum columns.
about 4 feet .
motor.
motor.
Sensing and Recording Systems
. 221 Recording system: ... magnetic head .
.222 Sensing system: . . . . . magnetic head
.223 Common system: . . . . two-gap head provides readafter-write checking.
.... none.
.23
Multiple Copies:
.24
Arrangement of Heads
. .
. ..
Use of station:
Stacks: . . . . . . . . . . .
Heads/ stack: . . . .
.
Method of use:
.
recording .
1.
10.
1 row at a time .
sensing.
Use of station: ...
Distance: . . . . . . . . . 0.150 inch after recording
head.
Stacks: . . . . . . . . . . . 1.
Heads/ stack: . . . . . . . 10.
Method of use: . . . . . . 1 row at a time.
Use of station: . . . . . . erase.
Distance: . • . . . . . . . precedes read-write head
and is located at reverse
side of tape.
Stacks: . . . • . . . . . . . 1.
Heads/ stack:. . . . . . . 1.
Method of use: . . . . . . all tracks at a time.
420:092.212
.34
Format Compatibility: only with 7340 Modell or 2
Hypertape Drives at 1,511
bytes per inch.
. 35
Physical Dimensions
. 351 Overall width: .
. 352 Length: . . . . .
.4
CONTROLLER
.41
Identity: . . . . . . . .
.42
Connection to System
1. 0 inch .
1,800 feet per cartridge.
IBM 2802 Hypertape Control.
.421 On-line:. . . . . . . ..
one 2802 Control per available control position on
Multiplexor or
Selector Channel.
Multiplexor Channels in any
System/360 and Selector
Channels in a System/360
Model 30 will operate
only at the 170,000 byteper-second (low density)
rate.
.422 Off-line: . . . . . . . . . none.
.43
Connection to Device
.431 Devices per controller: . . . . . .
1 to 8 tape drives, unless
2816 Model 2 Switching
Units are connected. In
this case, a 2802 Control
can address up to 16 tape
drives.
2816 Switching Unit, Model 2
.3
EXTERNAL STORAGE
The 2816 Switching Unit:
.31
Form of Storage
•
Permits a group of tape drives to be addressed
by anyone of up to four control units, under
program control.
•
Permits a control unit to address up to 16 tape
drives (using two 2816's, Model 2) .
. 311 Medium: . . . . . . .
plastic tape with magnetizable surface.
. 312 Phenomenon: . . . . . . • magnetization.
. 32
Positional Arrangement
. 321 Serial by: ..
1 to N rows at 1,511 or 3,022
rows per inch; N is limited
by available core storage .
. 322 Parallel by: . . . . . . . 10 tracks.
.44
.324 Track useData: . . . . . . • . . . . 8.
.441
Redundancy check: . . 2.
.442
Timing: . . • • . . . . • o (self-clocking).
.443
Control signals: ... . o.
Unused: . . . . . . . . . O.
.444
Total:
10.
. 325 Row use Data: . . . . . . . . . .
Redundancy check: .
Timing: . . . . . . . .
Control signals: ...
Unused: . . . . . . . .
Inter-block gap: ..
• 33
Coding:........
.
.
.
.
.
1 to N.
O.
O.
O.
O.
0.38 inch (nominal).
one 8-bit byte per tape row.
The basic 2816 switch provides switching between
two tape controls and four tape drives. The largest
system permits switching between 4 tape control
units and 16 tape drives.
Data Transfer Control
Size of load: . . . . . . . 1 to N bytes.
Input-output areas:. . . core storage.
Input-output area
access: . . . . . . . . . each byte.
Input-output area lockout: . • • . . . . . . . . . blocks of 2, 048 bytes can be
protected (optional on
Model 30 and 40 Processing Units, standard on
larger models).
.445 Table control: . . • . . . yes, using data chaining in
Channel Command Words.
(Models 30 and 40 permit
no data chaining using the
7340.)
.446 Synchronization: ..•. automatic.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360
420:092.500
.5
PROGRAM FACILITIES AVAILABLE
. 51
Blocks
· 511 Size of block:
Note: data presented is for the common tapelimited condition, in which case the tape
unit passes the gaps at full tape speed .
Under compute-limited conditions, the gap
can increase in size by an approximate
maximum of 0.2 inches, and the interblock time can increase by an approximate
maximum of 2.5 milliseconds. The tape
drive will continue to pass tape at full
speed if an order is given to it within 1
millisecond after detection of the gap
terminating the previous read or write
operation.
1 to N bytes; limited by
available core storage.
· 512 Block demarcationInput: ..•..••...• gap on tape.
Output: . . • • . . . • . . count in command.
.52
Input-Output Operations
.521 Input: . . . . . . . . . . . . read data from tape, forward or backward, into
core storage, with input
stopped by count or gap .
• 522 Output: ..•.•.••..• write one block forward.
.523 Stepping: . • . . . . . . . none.
· 524 Skipping: . . . . . . . . . skip one block forward
(space) or backward
(backspace) .
skip forward or backward to
tape mark or end-of-tape
marker.
skip forward and erase (to
skip defective tape areas).
• 525 Marking: . . . . • • . . . interblock gap.
end-of-file mark (tape
mark).
• 526 Searching: . • . . . • • • none.
.53
Code Translation: ... matched codes, from tape
rows to bytes in core
storage.
.54
Format Control: . . . . recording density (1,511 or
3,022 bytes per inch).
.55
Control Operations
Disable: • . . • . . . . .
Request interrupt: ...
Select format: . . . . . •
Select code: .•...•.
Rewind: ... ' ... .
Unload: . . . . . . . . . .
Enter fUe-protect
status: . . . . . . . . .
.56
disabled after unloading.
automatic.
density only.
no.
yes.
yes.
.62
Speeds
· 621 Peak speedI: . . . . . . . . . . . . . 170,000 bytes/sec.
II: . . . . . . . . . . . . . 340,000 bytes/sec.
· 622 Important parameters Recording density: .. 1,511 rows/inch at low
density.
3,022 rows/inch at high
density.
Tape speed: . . . . . . 112.5 inches/sec.
Start time: ..•• ' ... 3.0 msec maximum .
Stop time: . . . . . . . 3.0 msec maximum.
Full rewind time: ... 1. 5 minutes.
Rewind speed: . . . . . 225 inches/sec.
Inter-block gap: . . . . 0.38 inch .
. 623 Overhead: . . . . . . . . . 3.5 msec/block, plus 0.25
to 0.5 msec for synchronization characters in record .
. 624 Effective speeds (in
tape-limited applications)I: . . . . . . . . . . . . . 170, OOON/(B + 655) bytes/sec.
II: . • . • • . . . . . . . . 340, OOON/ (B + 1, 2·28) bytes/
sec. where B = number of
bYtes per block. (See
graph on Page 420:092.900).
· 63
Demands on System: .. varies widely with System/
360 model; see Simultaneous Operations, Section
420:111.
Testable Conditions
.7
EXTERNAL FACILITIES
Disabled: . . • • . . . . .
Busy device: .••....
Output lock: • . . . . . .
Nearly exhausted: ...
. 71
Adjustments:. ; . . . . . none .
.72
Other Cont:rols
yes.
yes.
yes.
yes.
yes; perforations 50 feet from
end-of- tape mark.
Busy controller: ••.. yes.
End of medium marks: yes; perforations 15 feet
from physical beginning
and end.
Operator attention
required: . . . . . . . . yes.
Correction occurred: yes.
.6
PERFORMANCE
.61
Conditions
I: • . . . . . . . .
. . low density (1,511 bytes
per inch).
II: . . . . . . . . . . . . . . high density (3,022 bytes
per inch).
7/65
Function
Form
Address
selection: dial
Load-unload: button
File protection:
Rewind:
Comment
selects unit address 0
through 9.
lowers tape into reservoirs and moves head
into position.
manual or program
button and
slide on recontrol.
verse side of
cartridge
button.
(Contd. )
INPUT-OUTPUT:
.73
7340 HYPERTAPE DRIVE
Loading and Unloading
420:092.730
.8
.731 Volumes handledCartridge: . . . . . . . 1,770 feet (1,800 feet less
15 feet unavailable at each
end); data capacity is:
• 19,400,000 bytes at low
density, using 1,000byte blocks.
• 27,700, 000 bytes at
high density, using
1, O~~-byte blocks.
• 38,700,000 bytes at
high density, using
2, O~~-byte blocks.
.732 Replenishment time: .. 0.3 to 0.5 minute; tape drive
needs to be stopped. Cartridge can be removed without rewinding tape. (Automatic Cartridge Loader
special feature automatically unloads one cartridge
and loads another, under
program or manual initiation. )
.734 Optimum reloading
period:. . . • . . • . • • 3. 1 minutes.
ERRORS, CHECKS AND ACTION
Error
Check or Interlock
Action
Recording:
read-after-write
dual parity
dual parity
set indicator.
byte count check;
storage protection
set indicator.
Reading:
Input area
overflow:
Output block
size:
byte count check
Invalid
operation
code:
check
Exhausted
medium:
reflective marker
or tape mark
Imperfect
medium:
none.
Timing
conflicts: overrun check
Excessive
skew:
check
Circuit
failure:
check
Note:
set indicator; correct error if
possible.
set indicator.
'set indicator.
set indicator.
set indicator.
set indicator.
set indicator.
These error indications and other status
information are transmitted from the
Hypertape Control to core storage in
response to a Sense command.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
420 :092.900
IBM SYSTEM/360
EFFECTIVE SPEED
7340 HYPERTAPE DRIVE, MODEL 3
10,000,000
7
4
2
1,000,000
7
4
3,022 Bytes/inch io"""
~
2
V
~
100,000
Bytes
per Second
...
~I--"""'"
1, 511 Bytes/inch
7
/.
~~
4
77
IP'
2
~
10,000
7
4
2
~~~
II
Vlr
1,000
7
4
2
100
10
2
4
7
2
4
7
100
Bytes per Block
7/65
2
1,000
4
7
10,000
420:093.100
IBM System/360
Input-Output
2415 Magnetic Tape Unit
INPUT-OUTPUT: 2415 MAGNETIC TAPE UNIT
.1
GENERAL
.11
Identity: • . . . . . . . . • 2415 Magnetic Tape Unit,
Models 1, 2, 3.
.12
Description
the 2400 Series Magnetic Tape Units. The 2415 is
similar except for performance, which is presented below, and for the restrictions mentioned
above.
Deliveries of the 2415 Magnetic Tape Unit are
scheduled to begin in the third quarter of 1966.
The 2415 Magnetic Tape Unit is a new, low-speed,
9-track tape unit that is compatible with the higherperformance 2400 Series units announced previously.
The peak speed of all models of the 2415 is 15,000
bytes per second. See Section 420:031 for the rules
governing the connection of the 2415 to a particular
System/360 configuration.
The 2415 is available in three models, each containing a single-channel controller and the indicated
number of tape drives:
•
Model 1 - two tape drives.
•
Model 2 - four tape drives.
•
Model 3 - six tape drives.
The physical characteristics of the 2415, aside
from speed, are similar to those of the 2400 Series
units described in Section 420:091, except that in
the 2415, the diagonal parity check code is calculated
and recorded during writing but is ignored during
reading. Thus, the error-correction feature of the
other 2400 Series tape units is not incorporated, but
compatibility is maintained.
The programming characteristics of the 2415 are
also similar to those of the other 2400 Series units,
except when the 2415 is used in a System/360 Model
20 system. Input-output operations in the Model 20
are handled in a different manner than in the larger
models. See Section 422:111, Simultaneous Operations, for a description of how the Model 20 inputoutput operations are handled. In general, errors
and malfunctions cause the Model 20 processor to
halt with a console panel display.
Please refer to Section 420:091 for a description of
the characteristics and special features provided by
©
.6
PERFORMANCE
.61
Conditions: ••.••••. standard operation of 2400
Series tape drives, except
where use of the optional
Seven- Track Compatibility
feature is indicated.
.62
Speeds
.621 Nominal or peak speed (bytes/sec) Standard operation: .• 15,000.
Seven- Track Compatibility Feature: . 3,750 at 200 bits/inch;
10,425 at 556 bits/inch;
15,000 at 800 bits/inch.
.622 Important parameters Density: . . • . • . . • . 800 bytes/inch.
Tape speed: . • . . . . . 18.75 inches/sec.
Full rewind time: . . . 4.0 minutes.
Interblock gap: .•••. O. 6 inches.
Interblock gap time: . 32.0 msec (nominal).
.624 Effective speed: .•.•• 15, OOOB/(B + 480) bytes
per second, where B is
number of bytes per
block (See also graph
of effective speed, Page
420:092.900) .
.63
Demands on System: •. varies with System/360
model and type of channel;
see Simultaneous Operations, Section 420:111.
The Model 20 processor is
delayed for the entire duration of a tape read or write
operation.
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360
420:093.900
EFFECTIVE SPEED: 2415 TAPE UNIT
1,000,000
7
4
2
100,000
7
4
Effective Speed,
char/sec
2
-----
10,000
7
~
~
I'
4
/'
2
~V
1,000
""
.,
7
4
I~
"
./
IL
2
100
2
10
4
7
100
2
4
7
2
1,000
Characters per Block
7/65
4
7
10,000
420:094.100
IBM System/360
Input.Output
1600 bpi Magnetic
Tape Units
INPUT·OUTPUT: 1600 BPI MAGNETIC TAPE UNITS
.1
GENERAL
• 11
Identity: • . . . . . . . • • 2.401 Magnetic Tape Unit,
Models 4, 5, 6.
2402 Magnetic Tape Unit,
Models 4, 5, 6.
2403 Magnetic Tape Unit
and Control, Models 4,
5, 6.
2415 Magnetic Tape Unit
and Control,' Models 4,
5, 6.
• 12
Description
In August 1965, IBM announced new models of the
nine-track 2.400 Series Magnetic Tape Units that
provide doubled peak data transfer rates through
doubling of the recording density to 1600 bytes per
inch. IBM states that the higher density is made
practical by adoption of the "phase encoding"
method of recording in place of the. "NRZI" (NonReturn to Zero, IBM) method used in the IBM 729
Series and in the SOO-bpi 2400 Series tape units •.
The phase· encoding method, which has been in use
for several years in IBM's 7340 Hypertape Drives
and UNIVAC's Uniservo mc tape units, uses a
magnetic flux reversal in phase with a uniform
reference voltage to represent a O-bit and a flux
reversal out of phase with. the reference voltage to
represent a 1-bit. With the NRZI method, a flux
reversal represents a 1-bit, while the absence of
a flux reversal is interpreted as a O-bit. Because
the phase encoding technique records a flux reversal in every track position of every properly-recorded. data-. frame, it has two inherent advantages:
over the NRZI. technique:
• The absence of a flux reversal indicates an
error condition, permitting in,..flight correction
of single-track errors.
Each track is self-clocking,. so the chances of
errors due to· skewed recording are greatly reduced._
The older SOO'-bpi models of the 2401, 2402, 2.403
and 2404 Magnetic Tape Units are described in
detail in. Section 420:091 of this report, and the
SOO-bPi models of the 2415 are described in
Section 42.0:093 .. The newly-announced models are
essentially the same tape units with 1600-bpi, ninetrack, phase encoding read/write circuitry in
place of (or in addition to) the SOO-bpi, nine,...track,
NRZI read/write circuitry in the. earlier models.
Table 1. summarizes. the fUnctional characteristics
of the. new models, and the Price Data on page
420:221. 104 will help to clarify the picture. The
_1600-bpi density is. standard in all of the new
models, and optional features are required to read
and record atSOO bytes per inch. Note that IBM
has not announced a 1600-bpi version of the 2404
Magnetic Tape Unit and Control, which combines
a two-channel controller and a single tape drive.
The increased recording density doubles the peak
data transfer rate of each tape drive and enables
each reel of tape to hold considerably more data.
Because the inter block gap length (0. 6 inch) remains unchanged, its effect upon overall performance is greater at the increased density, as the
"Efficiency" figures in Table I clearly show. This
means that the effective data transfer rate, upon
data blocks of a given length, will by no means be
doubled when an installation shifts from SOO-bpi to
1600-bpi. tapes. For example, when converting
from 2401 or 2402 Model 2 to Model 4 tape units,
the peak data transfer rate goes from 60, 000 to
120,000 bytes per second; the effective data transfer rate on 1, OOO-byte blocks, however, increases
by only 51%, from 40,500 to 61,200 bytes per
second. The use of long data blocks, which tend
to decrease the performance degradation due to the
inter block gap, is therefore more desirable than
ever with the new 1600-bpi tape drives.
Error Detection and Correction
When recording at 1600 bytes per inch, the 2400
Series Magnetic Tape Units will append a vertical
check. bit to each byte, but the longitudinal parity
check and the cyclic· redundancy check used at SOO
bpi will be omitted; IBM maintains that the phase
encoding technique eUminates the need for these
two additional. check characters at the end of each
tape block. The usual read-after-write check will
be made while recording.
When reading at 1600 bytes per inch, the tape units
will continuously monitor the signals from all ninetracks. As soon. as anyone track fails to provide a
flux reversal in any data frame, that track will be
disabled for the remainder of the block. Using the
vertical parity check bit for each frame, the information bits in the disabled track will then be regenerated automatically. This scheme will provide
automatic detection and in-flight correction of all
errors which are confined to a single track.
Multiple-track errors will be. detected and recognized
as non-correctable, and standard error recovery
procedures (backspace and reread) will have to be
used in. these cases •.
Controllers
The 2415 Magnetic Tape Unit and Control consists
of a single-channel controller and two, four, or six
tape drives in. Moder 4, 5, or 6, respectively; no
external controller is therefore· required, and no
additional tape drives can be connected.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/6-5
IBM SYSTEM/360
420:094.120
• 12
Description (Contd.)
Also required, on each 2401/2402/2403 Modei 4,
5, or 6 tape drive to be operated at SOO bpi, is the
Dual Density feature.
The 2403 Magnetic Tape Unit and Control consists
of a single-channel controller and one tape drive;
a 2403 Model4, 5, or 6 can control its own drive
plus up to seven more drives in any combination of
2401 and/or 2402 Models 1 through 6. (Each 2402
consists of two tape drives in a single cabinet.)
The 2S03 and 2S04 Tape Controls are free-standing
controllers for up to eight 2401 and/or 2402 tape
drives. The 2S03 is a single-chamiel controller,
while the 2S04 connects to two channels and permits
simultaneous reading and writing. The newlyannounced 2S03 Model 2 and 2S04 Model 2 can each
control up to eight drives in any combination of
2401 and/or 2402 Models 1 through 6. With the
16-Drive Addressing Feature and two 2S16 Model 1
Switching Units, a 2S03 can control up to 16 tape
drives.
Seven and Nine-Track Compatibility: A control unit
feature that permits 1600-bpi tape drives to read
and write at SOO-bpi in both the'nine-track and
seven-track (729-compatible) modes, as well as in
the 1600-bpi mode; the Dual Density feature is
also required on each tape drive to be operated at
SOO-bpi.
Mode Compatibility: Required on each 2401/2402
Modell, 2, or 3 Magnetic Tape Unit to be connected to any of the following SOO/1600-bpi controllers: 2403 Model 4, 5, or 6, 2803 Model 2, or
2S04 Model 2.
Optional Features
Other optional features available for the 2400 Series
Magnetic Tape Units are described in Report Section
420:091; also see the Price Data on page
420:221. 105.
Nine-Track Compatibility: A control unit feature
that permits 1600-bpi tape drives to read and
write in the nine-track, 800-bpi, NRZI mode, as
well as in the 1600-bpi, phase encoding mode.
Deliveries of most of the 1600-bpi tape units are
scheduled to begin in the second quarter of 1966,
with 2415 Magnetic Tape Unit deliveries scheduled
for the third quarter of 1966.
TABLE I: CHARACTERISTICS OF IBM 1600-BPI MAGNETIC TAPE UNITS
Model
Number
Tape Recording Peak
Interblock Gap Lengths
Speed, Density,
Speed,
(nominal)
inches
bytes
bits per
per sec inches' msec (1) bytes (2)
per sec
inch
Full
Rewind
Time,
100-byte 1000-byte minutes
blocks
blocks
9.4
51. 0
3.0
17.2
67.5
Efficiency,
%(3)
2401/2402/2403
Model 4
37.5
1600
SOO*
60,000
30,000
0.6
0.6
16.0
16.0
960
4S0
2401/2402/2403
Model 5
75.0
1600
SOO*
120,000
60,000
0.6
0.6
S.O
8.0
960
480
9.4
17.2'
51.0
67.5
1.4
2401/2402/2403
Model 6
112.5
1600
800*
180,000
90,000
0.6
0.6
5.3
5.3
960
480
9.4
17.2
51. 0
67.5
1.0
1600
800*
30,000
15,000
0.6
0.6
32.0
32.0
960
480
9.4
17.2
51.0
67.5
4.0
2415 Models
4, 5, 6
*
18.75
Optional features are required for operation at 800 bpi.
(1) Time in milliseconds to traverse each inter block gap when reading or writing
consecutive blocks.
(2) Number of byte positions occupied by each interblock gap.
(3) Effective data transfer rate at the indicated block size, expressed as a
percentage of peak speed.
/
8/65
420:101.100
IBM System/360
Input-Output
2250 Display Unit
INPUT-OUTPUT: 2250 DISPLAY UNIT
.1
GENERAL
. 11
Identity: . . . . . . . . . 2250 Display Unit,
Models 1 and 2.
. 12
Description
The IBM 2250 Display Unit can provide dynamic
visual displays of charts, graphs, and alphameric
characters on a 12-inch-square screen on the face
of a 21-inch cathode ray tube. The display area
contains a 1024-by-1024-point matrix, providing
over one million discrete points that can be addressed by their X and Y coordinates. Within this
area, 3,848 "basic size" characters (52 lines of
74 characters each) or 1,715 "1-1/2 size" characters (35 lines of 49 characters each) can be displayed.
When a buffer is used, the keyboard can control
the location of a movable electronic marker used
to identify data locations on the display screen.
The keyboard-buffer combination permits entering a message into the buffer and displaying it
for validity checking before it is transmitted to
the computer.
•
Buffer - 4, 096 or 8, 192 bytes of core storage,
useful for holding addresses and data so that
displays can be regenerated without tying up the
processor. Maximum buffer loading speed
from the processor is 475,000 bytes per second.
•
Character Generator - translates a byte code
into the analog signals required to trace a char. acter on the face of the tube. The two character
sizes are program-selectable.
Both models of the 2250 use the same type of
cathode ray tube, which requires regular regeneration of the display. Optional features provide for
buffering, keyboard data entry, and use of a light
pen (a hand-held electronic pointer which permits
program detection of specific lines, characters,
and symbols that are displayed on the face of the
tube). The prospective user should keep in mind
that effective application of this type of display unit
requires good system design and considerable programming effort.
•
Absolute Vectors - permits displaying a straight
line between the present beam position and any
other point in the matrix by specifying only the
X and Y coordinates of the end point.
•
Programmed Function Keyboard - consists of
32 keys, indicator lights, and 8 overlay code
senSing switches. Each key initiates a subroutine associated with the respective overlay
program. Up to 256 different overlays can be
identified by means of the overlay switches.
See Section 420:031, System Configuration, for the
rules governing the connection of these devices to
a particular System/360 configuration.
• Light Pen - a hand-held electronic device used
by the operator to identify to the program a
particular point or character on the display
screen.
Modell
•
The 2250 Model 1 has a self-contained control unit
that connects directly to a System/360 input-output
channel; it is designed for applications where only
one display unit is required. (IBM 2250 Model 2's
are used in multiple displays, and up to eight
Model 2's can be connected to a single IBM 2840
Display Control. )
Model 1 without optional features has only the basic
capability of displaying dots at any addressed
points in the 1024-by-1024-point matrix and of
plotting horizontal, vertical, or 45-degree lines.
Curves are drawn as a series of short straight
lines. The matrix address of each point must be
supplied by the computer. In order to plot curves,
graphs, symbols, or alphameric characters, computer subroutines or tables must provide the necessary matrix addresses. When no buffer is used,
the display must be regenerated about 30 times
each second by the computer to keep it visible.
Model 1 Optional Features
•
Alphameric Keyboard - a keyboard similar to
that on a 1052 Printer-Keyboard, which permits
an operator to enter messages into the computer.
©
Operator Control Panel, First and Second these panels duplicate the on/off and program
load facilities on the control panel of the Processing Unit of a first or second Model 50, 65,
or 70. The first of these two features is a prerequisite for the second.
Model 2
One or two 2250 Model 2 Display Units can be connected to the basic 2840 Display Control. The Display Control requires an available control position
on a System/360 data channel. Optionally, up to
three more pairs of 2250 Model 2 Display Units can
be connected to the Display Control through the use
of up to three Display Multiplexors.
The 2840 Display Control provides common circuits, control, and buffer storage for the 2250
Model 2 Display Units connected to it. Standard
features in the 2840 are a character generator and
8,192 bytes of core storage, which are timeshared by all of the connected 2250 Model 2's.
Buffer areas are addressable, permitting seli9cted
lines or characters to be changed without the need
to reload the entire buffer. The 2250 Model 2
Display Units can be located up to 2,000 feet away
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
420: 101.120
. 12
IBM SYSTEM/360
Description (Contd.)
from the controlling 2840. The 2840 also serves
as the control for IBM 2280, 2281, and 2282 Film
Units. The 2250 Model 2 and the Film Units can
be intermixed on the same 2840 Control. See.
Section 420:031, System Configuration, for the
rules governing the connection of these two types
of devices to the 2840.
Model 2 Optional Features
•
•
•
The Alphameric Keyboard, Absolute Vectors,
Light Pen, and Programmed Function Keyboard, as described under "Modell Optional
Features. "
Buffer - an additional 8,192 bytes of core
buffer storage.
Display Multiplexer - up to three of these
permit up to three pairs of additional Model 2
Display Units to be connected to a single 2840
Display Control.
Note: the Operator Control Panel features are
not available for use with a 2250 Model 2
Display Unit.
Programming
The visible display on the face of the cathode ray
tube is produced by the action of an electron beam
hitting a phosphor coating, which causes the coating to glow briefly. Since the glow normally fades
in a few milliseconds, the display must be regenerated (redisplayed) about 30 times each second in
order for the human eye to observe a steady visible
image. In 2250 units equipped with the buffer,
this regeneration is performed automatically. For
units lacking a buffer, the regeneration must be
programmed; i. e., the complete display must be
sent to the 2250 unit by the processor about 30
times each second for as long as the display is
required.
The electron beam of the cathode ray tube can be
deflected in either of two basic programming
modes: graphic mode and character mode. A twobyte prefix code always precedes the display data
to identify the mode of operation and the particular
operation to be performed.
The graphic mode is used for point plotting, for
beam positioning, and for plotting hOrizontal,
vertical, or diagonal lines under program control.
Points plotted four or more raster units apart can
be distinguished by the human eye as discrete
points. (A raster unit is the distance between two
adjacent addressable locations in the matrix.)
Four bytes are required to hold the coordinates of
7/65
each point addressed. Operations available in the
graphic mode include movement of the beam from
one point to any other point, displaying the end
point, and displayiJ;lg the straight-line path between
two points. (Note that display of a diagonal path at
a slope other than 45 degrees requires the Absolute
Vectors special feature.) Alphanumeric characters
can be displayed in this mode only through complex
programming; i. e., the outline of each character
must be traced under program control.
The character mode (optional with Modell; standard with Model 2) can be used to reduce the time
required to display alphanumeric data. The twobyte prefix code specifying the character mode also
specifies which of the two character sizes shall be
displayed. Each byte following the prefix code
then represents one character to be displayed. The
beam must initially be positioned to the first display position by means of a graphic mode operation.
Horizontal spacing from character to character is
automatic. A new line can be started by a special
New Line character in the data. A new line is also
begun automatically after encountering the end of
the available display line. Alternatively, another
beam-positioning operation can be used to begin a
new line anywhere in the display area.
The time required to display a point ora line is
. the largest of the data acquisition time, the vertical positioning time, or the horizontal positioning
time. These three functions are overlapped. The
data acquisition time is 16.8 microseconds. A
deflection of up to 136 raster units can be made
during this time. The positioning time for deflections larger than 136 raster units is given by the
formula:
.
POSitioning Time = 0.0894 (N-136) + 16.8
microseconds, where N is the vertical or
horizontal deflection, whichever is larger,
in raster units.
The average time to display alphanumeric data
using the Character Generator feature is 15 microseconds per "basic size" character and 17 microseconds per "1-1/2 size" character.
The only demand imposed upon the System/360
Processing Unit by a 2250 display operation is the
time required to send the necessary data and control codes to the 2250. Note that for the 2250
Modell without the buffer feature, this data must
be sent repetitively, about 30 times each second,
which can represent a very significant demand
upon the Processing Unit. See Section 420:111,
Simultaneous Operations, for the demand imposed
on the system by data transfers via the various
data channels.
420: 102.1 00
IBM System/360
In put-Output
2260 Display Station
INPUT-OUTPUT: 2260 DISPLAY STATION
.1
GENERAL
. 11
Identity:....
2260 Display Station.
2848 Display Control,
Models 1, 2, 3.
. 12
Description
The 2260 Display Station is a cathode ray tube unit
for displaying alphameric data. It is controlled
and buffered by the 2848 Display Control. Up to
960 characters can be displayed on the face of the
tube, depending on which model of the 2848 Display Control is used. Multiple Display Stations
can be connected to the 2848; each station can be
located up to 2,000 cable feet away from the 2848.
Display Station
Characters are presented on a 4-by-9-inch display
area; each character is formed by a 5-by-7-dot
matrix. Sixty-four different symbols can be displayed; 26 letters, 10 numeric digits, and 28
special symbols. Two of the special symbols
cannot be transferred between the processing unit
and the display station; these are the parity error
and cursor symbols.
A Display Adapter is required for each pair of
2260 Display Stations. If more than two Display
Adapters are connected to the 2840 Modell, or
more than one to either the Model 2 or the Model 3,
one or two Expansion Units are required. In
addition, one 1053 Printer can be attached to the
2848 through a special adapter for each model.
The 2848 can be connected directly to a System/360
processor through an input-output channel. See
Section 420:031, System Configuration, for the
rules governing the connection of the 2848 Display
Control to a particular System/360 configuration.
Alternatively, the 2848 can be remotely connected
to a System/360 through a 2701 Data Adapter Unit
(see Section 420: 106) and an appropriate data set.
The 2848 operates in a half-duplex mode over a
common-carrier, leased, 4-wire, private-line
telephone system, on Schedule 4A lines or better.
One keyboard, either alphameric or numeric, can
be attached to each Display Station, permitting the
2260 to function as an input station.
The maximum data rate of the 2848 when connected
to an input-output channel is 2,560 characters per
second. When connected through a communications
data set, the maximum data rate is either 120
characters per second (1,200 bits per second) or
240 characters per second (2, 400 bits per second),
depending on the data set used. Operation at 120
characters per second requires a Western Electric
Data Set 202D or the equivalent; operation at 240
characters per second requires a Western Electric
Data Set 201B or the equivalent.
Display Control
Optional Features
The 2848 Display Control provides the control
logic and buffer storage unit for 2260 Display
Stations. Three models of the 2848 are available;
they differ in the number of characters that can be
displayed per .station and the number of stations
that can be connected to each control. The characteristics of the three models are as follows:
The following special features are available for the
2848 Display Control.
•
Model 1 - can control the display of up to 6
lines of 40 characters each (240 characters
total) on up to twenty-four 2260 Display
Stations.
•
Model 2 - can control the display of up to 12
lines of 40 characters each (480 characters
total) on up to sixteen 2260 Display Stations.
•
Model 3 - can control the display of up to 12
lines of 80 characters each (960 characters
total) on up to eight 2260 Display Stations.
Line Addressing: Permits the computer program
to specify the starting locations of incoming data.
The starting location can be the first character of
any row.
.
Non-Destructive Cursor: Allows the operator to
move the cursor symbol anywhere on the display
without changing the displayed information. A
special Non-Destructive Adapter is also required
on each Display Adapter.
1053 Adapter: contains a 960 -character buffer
and logic for controlling a 1053 Printer. Data can
be printed from the 2848 buffer or directly from the
computer system. Only one 1053 Printer can be
connected to a 2848 Display Control.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
420; 103.100
IBM System/360
Input-Output
7770 Audio Response Unit
INPUT -OUTPUT: 7770 AUDIO RESPONSE UNIT
.1
GENERAL
.11
Identity:
.12
Description
7770 Audio Response Unit,
Model 3.
The IBM 7770 Audio Response Unit, Model 3, is a
buffered data communication unit that accepts
telephoned digital requests and relays them to a
System/360 Processing Unit, which processes the
data and returns a coded reply to the 7770. The
7770 interprets the reply, selects the proper words
from its stored ~ocabulary, and transmits these
words as a human-voice'response (optionally male
or female) back to the inquirer. See Section
420:031, System Configuration, for the, rules
governing the connection of a 7770 to a particular
System/360 configuration.
The Audio Response Unit is composed of three
sections: inquiry, control, and audio output.
The inquiry section accepts dialed or keyed digital
inquiries transmitted from the connected inquiry
terminals through common-carrier communication
facilities. Transmission of the inquiry between the
common-carrier receiver and the 7770 is accomplished in a 3-out-of-14 or 2-out-of-8 inquiry code.
The 7770 receives the inquiry from the receiver
terminal in serial-by-character, parallel-by-bit
form. An inquiry is assumed to be complete
when the 7770 receives no data within a fivesecond period. The inquiry is then translated and
transferred to the Processing Unit.
The digital control section controls the data flow
between the Processing Unit and the 7770. A Read
command causes the inquiry to be transferred from
the 7770 to the processor. When the proc'essor
completes its interpretation of the inquiry, it composes a coded response message. The processor
then issues a Write command, and the response
message is transferred to the 7770. The response
me'ssage is composed of vocabulary word locations
(located on the 7770's magnetic drum), and is sent
one word location at a time to the 7770.
The audio output section provides the actual audio
response to the inquirer. These words, recorded
on the magnetic drum in analog form as an audio
signal, are amplified and transmitted to the terminal originating the inquiry.
The vocabulary is recorded on the removable
magnetic sleeve of a drum 4 inches in diameter and
10 inches long. There are 128 tracks on the drum,
each having a unique address. In general, each
track contains one "word" which can be an alphabetic letter, a numeral, or a short word from the
master list.
Some polysyllabic words require more than one
track and consequently more than one word-location
specification in the response message. Vocabularies
requiring 64 or fewer tracks are recorded twice on
the drum. The drum revolution time is 500 milliseconds. The first track always contains 500
milliseconds of "silence, " permitting a programmer to insert pauses in the voice response.
The prerecorded drum vocabulary can consist of 32,
48, or 64 words and can be increased to BO, 96,
112, or 12B words by a special feature. The vocabulary is chosen from a master vocabulary,
provided by IBM, which consists of a list of fre,quently used industry words, the numbers from
zero through nine, and the letters of the alphabet.
The vocabulary can be changed, by exchanging the
analog recorded cylinder, at any time by the
()ustomer. (Changes or additions to the vocabulary
recorded on,a cylinder can only be made at the
factory.)
The 7770 operates in a half-duplex mode over toll,
leased, or privately-owned voice-grade lines. The
basic 7770 handles up to four lines. The line capacity can be expanded, in 4-line increments, to a
maximum of 48 lines. Random inquiries on any or
all input-output lines can be serviced simultaneously.
The 7770 operates with any of the following inquiry
terminal devices, provided the proper arrangements
have been made by the customer with the common
carrier: 1001 Data Transmission Terminal, rotary
dial' telephone with associated pushbutton manual
dialing device, rotary dial telephone with pushbutton card dialer device, pushbutton manual dialing
telephone, and pushbutton type card dialer telephone.
The 7770 Model 3, used in the System/360, differs
from the 7770 Modell (used in IBM 1401, 1440, or
1460 systems) and Model 2 (used in 1410 or 7010
systems) in that the inquiry and response messages
are transmitted character by character between the
7770 Model 3 and the Processing Unit, whereas
Models 1 and 2 contain internal message buffers.
Analysis of 7770 throughput (number of inquiry
calls processed per hour) is complex and depends
on such factors as the number of lines, the length
of the inquiry messages, the amount of internal
processing required on each message, the length
of the response messages, and the number of inquiries outstanding. An approximation by IBM of
the throughput rate of a 7770 for short calls (less
than one minute) is presented in the following
formula;
C = N(10B-T)-70
where C = the number of calls per hour.
N = the number of lines.
T = the average length of a call in seconds.
(Contd. )
7/65
INPUT-OUTPUT: 7770 AUDIO RESPONSE UNIT
. 12
420:103.120
Description (Contd.)
• Queueing time (0.3 sec/4 lines) .
The length of a call is the total time after the
telephone connection to the computer has been
made until the connection is broken. The call
length depends on the sum of the following factors:
•
Processor time (about O. 1 second).
•
Number of words in response (0. 5 sec/word).
•
Inquirer disconnect (5 seconds).
•
Number of digits in the inquiry (rotary dial, 1. 3
sec/digit; pushbutton dial, 0.4 sec/digit; card
dial, 0.2 sec/digit).
•
Time-out delay (5 seconds).
©
The numbers in parentheses show the time for each
operation. The queueing, processor, and disconnect times are approximate since they vary with
the application. The minimum length for a call is
approximately 12 seconds.
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
420:104.100
IBM System/360
Input-Output
7772 Audio Response Unit
INPUT-OUTPUT: 7772 AUDIO RESPONSE UNIT
.1
GENERAL
.11
Identity: . . . . . . . . . • 7772 Audio Response Unit,
Model 3.
. 12
Description
The IBM 7772 Audio Response Unit is currently
for use with the System/360 only. It
performs the same general function as the IBM
7770 Audio Response Unit described in the previous rep0rt section (420:103) and features a
larger vocabUlary; conversely, it cannot control
as many communications lines as a 7770. See
Section 420:031, System Configuration, for the
rules governing the connection of a 7772 unit to a
particular System/360 configuration.
availabl~
The 7772 is a buffered data communications unit
that accepts telephoned digital requests through
data sets over appropriate communications
facilities. It transmits the inquiry messages to
a System/360 Processing Unit, under program
control. The processor processes the input
message and composes a response, selecting
the desired words in proper sequence from a
vocabulary .stored in digitally-coded voice form.
The response message consists of a list of
addresses that refer to a group of words digitally
encoded in a random access storage unit. The
digital information stored for each word is used
to activate a set of band filters that cover the
total telephone frequency band. The sum of the
band-filter outputs constitutes the audio response.
Approximately 2,400 bit poSitions (300 bytes) of
random access storage are required to store the
information for generating one second of audio
output. Each syllable, which can be an alphabetic
letter, numeral, part of a multi-syllable word,
or a pause, requires approximately 0.5 seconds
of audio output. Thus, about 250 simple words
or syllables can be stored on one "cylinder"
(36,250 bytes) of it 2311 Disc storage Drive.
The 7772 Audio Response Unit's vocabulary is
chosen from a library of frequently-used words
associated with commercial and industrial applications and is presented in a male voice. The
vocabulary itself is provided to the user in digitally-coded form on either punched cards or magnetic tape. The customer may have any number
of words from the available library for entry into
his system.
The 7772 operates in a half-duplex'mode over toll,
leased, or privately-owned voice-grade lines. The
basic 7772 handles up to two lines. The line capaCity can be expanded, in 2-line increments, to
a maximum of 8 lines. Random inquiries on any
or all lines can be serviced simultaneously.
The 7772 operates with any of the follOwing
inquirr terminal devices, provided the proper
7/65
arrangements have been made by the customer
with the common carrier: 1001 Data Transmission Terminal, rotary dial telephone with
associated pushbutton manual dialing device,
rotary dial telephone with pushbutton card dialer,
pushbutton manual dialing telephone, pushbutton
type card dialer telephone, mM 1093 Programmed
Keyboard, or IBM 1094 Line Entry Keyboard.
The functional differences between the two Audio
Response Units can be summarized as follows:
•
The 7772 generates audio responses from
a vocabulary prerecorded in digitally-coded
form on an external disc file, whereas the
7770 generates audio responses from a
vocabulary prerecorded in analog form on
a magnetic drum within the 7770.
•
Input from the ProceSSing Unit to the 7772
is the digitally-coded voice response itself,
whereas the 7770 receives a string of drum
addresses of the words comprising the
response.
• The 7772 offers a virtually unlimited vocabulary with a small number of communication lines (2 basic, 8 maximum), whereas
the 7770 has a limited vocabulary (32 words
basic, 128 maximum) and can handle a: large
number of lines (4 basic, 48 maximum).
Analysis of 7772 throughput (number of inquiry
calls processed per hour) is complex and depends
on such factors as the number of lines, the length
of the inquiry messages, the amount of internal
processing required on each message, the length
of the response messages, the time required to
access the digitally-encoded response words, and
the number of inquiries outstanding. The length
of a call is the total time after the telephone connection to the computer has been made until the
connection is broken, and depends on the sum of
the following factors:
•
Number of digits in inquiry (rotary dial, l. 3
sec/digit; pushbutton dial, 0.4 sec/digit;
card dial, 0.2 sec/digit).
•
Time-out delay (5 seconds).
• Queuing time (0.3 sec/2 lines).
•
Processor time (about 0.5 sec).
• Number of words in response (0.5· sec/word
plus access time).
• Inquirer disconnect (5 seconds) ..
The numbers in parentheses show the time
for each operation. The queuing, processor,
response, and disconnect times are approximate
since they vary with the application. The minimum length for a call is approximately 15 seconds.
420: 105.100
IBM System/360
Input-Output
Optical and Magnetic Readers
INPUT-OUTPUT: OPTICAL AND MAGNETIC READERS
.1
GENERAL
.11
Identity:
.12
IBM 1412 Magnetic Character Reader, ModelL
IBM 1419 Magnetic Character Reader, ModelL
IBM 1418 Optical Character
Reader, Models 1, 2, 3.
IBM 1428 Alphameric Optical Reader, Models 1, 2,
3.
IBM 1231 Optical Mark Page
Reader, Model N1.
IBM 1285 Optical Reader.
Use of the IBM System/360 1412 Input/Output
Control Program or an equivalent routine is required. Operation of the 1412 is not included
under the System/360 1401, 1440, or 1460
Compatibility Features.
A detailed description of the 1412 appears in Section 401:104 of the IBM 1401 report.
Optional Features
Multiple Column Select-Sort Suppress:
•
Description
Six different types of optical and magnetic readers
can be connected to a System/360. These units are
described below.
The two magnetic character readers offer two
speeds at which documents encoded with Font E-13B
magnetic ink characters can be read. The optical
readers permit reading of more conventional type
styles, such as IBM 407 Accounting Machine and
1403 Printer output (except for the output of the
faster 1403 Model 3 Printer). The IBM 1285
Optical Reader is a new unit that reads data from
journal tapes. The 1231 Optical Mark Page Reader,
Model Nl, is used for automatic reading of pencil
marks from 8-1/2 by 11 inch test sheets, questionnaires, etc.
.121 1412 Magnetic Character Reader, Model 1*
The 1412 Magnetic Character Reader reads 950
documents (6 inches long) per minute and sorts
them into any of 13 pockets. Documents up to 8-3/4
inches long can be handled at proportionately lower
rates, and the documents can be of intermixed sizes
and thicknesses. Document length can range from
6 to 8.75 inches, width from 2.75 to 3.67 inches,
and thickness from 0.003 to 0.007 inches.
The 1412 is functionally similar to the later Model
1419 Magnetic Character Reader, whose rated
speed is 1,515 documents per minute (see Paragraph .122 below). Characters printed in magnetic
ink character Font E-13B (adopted as standard by
the American Bankers' Association) are recognized
by the reader head. Only the 10 numerical characters and 4 special symbols comprising Font E-13B
can be read, and all magnetic ink imprinting must
be on a single line within five-eighths of an inch of
the bottom edge of the document. The 1412 can be
used on-line for document sorting and reading, or
off-line for sorting.
* See Section 420:031, System Configuration, for
the rules governing the connection of this device
to a particular System/360 configuration.
©
Permits the sorting to pocket A or B of documents with specific digit values in up to four selected digit positions.
o Permits the routing to pocket A or B of documents
that do not have specified numbers in designated
column positions of a predetermined field.
a
Permits the routing to pocket A or B of documents that contain a specific code in a predetermined field, while sorting all other documents.
Document Counter: counts all documents passing
through the 1412 and maintains a running total.
Self-Checking Number Verification: provides verification of account numbers. Available in either
Modulo 10 or Modulo 11; the latter is more effective and more expensive.
Electronic Accumulator and Sequence Checking:
permits accumulated totals of the amounts read
from documents to be printed as lists, and checks
selected positions for correct document sequencing.
(Multiple Column Select-Sort Suppress is a prerequisite. )
Endorser: imprints a full endorsement at the normal speed of the 1412.
.122 1419 MagnetiC Character Reader, Model 1*
The 1419 Magnetic Character Reader is a faster
version of the 1412 reader described in the preceding paragraph. It reads 6-inch documents at
a speed of 1,515 per minute, compared to 950 for
the 1412, and can also be used for off-line sorting.
A System/360 Adapter unit is required. A detailed
description of the 1419 appears in Section 401: 103
of the IBM 1401 report.
Use of a 1419 involves more restrictions than use
of a 1412. The use of the IBM System/360 1419
Input/Output Control Program or an equivalent
routine is required. Operation of the 1419, like
the 1412, is not included under the System/360
1401 or 1460 Compatibility Features. Reprogramming of 1419 operations will therefore be required
when a System/360 replaces an IBM 1400 Series
installation.
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360
420: 105.122
.122 1419 Magnetic Character Reader, Modell (Contd.)
Minimum processing time, including pocket selection, is 32.2 milliseconds. Since the 1419 has only
one operating speed, if the required processing
time exceeds this figure, the data will have to be
transcribed to magnetic tape for processing in a
subsequent run.
Optional Features
Features in addition to those offered for the 1412
are:
•
•
51-Column Card Sorting: permits intermixed
feeding of 51-column cards and standard-size
documents. Feeding rate of 51-column cards
by themselves is approximately 1,875 per minute.
Split Field: the first ABA dash symbol following
a digit will separate any field into two elements;
either element can be treated as a separate
field.
.123 IBM 1418 Optical Character Reader; Models 1, 2, 3*
The 1418 reads numerical characters printed in IBM
407 -style type at the rate of 413 six-inch documents
per minute. The standard model reads one line of
data, which may be positioned anywhere on the
document. Acceptable input data for the reader can
be produced by an IBM 407, 408, or 409 Accounting
Machine, by a 1403 (except Model 3) or 1404
Printer, or by a typewriter equipped with the 407
type font. The 1418 can alt€rnatively be specified
to read documents printed by plastic plate imprinters with a type style resembling the 407 type
style elongated to 0.114 inches high.
Models 1 and 3 have three stackers and are designed
for on-line use only. Model 2 is equipped with
13 stackers and can be used off-line for singledigit sorting on any digit position. Models 1 and 2
handle document lengths from 5.875 to 8.75 inches
and widths from 2.75 to 3.67 inches, and Model 3
handles a broader range of document sizes. All
models have an "overflow stacking" mode that
permits unloading the stackers without stopping
the reader when no sorting is required.
Use of the IBM System/360 1418 Input/Output
Control Program or an equivalent routine is required. Operation of the 1418 is not included under
the System/360 1401 or 1460 Compatibility
Features.
Thirteen characters can be re(;ognized by the
standard 1418 reader: the digits 0 to 9, the lozenge,
the dash, and a solid vertical line (preprinted on
the forms) . A detailed description of the 1418
appears in Section 401:102 of the IBM 1401 report.
* See Section 420:031, System Configuration, for
the rules governing the connection of this device
to a particular System/360 configuration.
Optional Features
Additional Read Station: permits reading data from
two lines on each document in a single pass.
Switching between the two read stations, which are
4.25 inches apart, is under program control.
Mark Reading Station: permits the reading of up to
37 columns of numerical data marked on a standard
IBM card with an ordinary black pencil or ink; cannot be used when the Additional Read Station or the
Mark Reading Station, Slanted Mark, is installed.
Mark Reading Station, Slanted Mark: permits the
reading of slanted marks, made with certain types
of pens or pencils, on Models 1 or 2 only; cannot
be used when the Additional Read Station or Mark
Reading Station is installed.
.124 IBM 1428 Alphameric Optical Reader, Models
1, 2, 3*
The 1428 reader can read upper case alphabetic
characters, numeric characters, and several
special characters printed on card and paper documents in the 1428 type font, which is available on
IBM 1403 Model 1, 2, 4, and 5 Printers and on
certain IBM electric typewriters. The reading
rate is approximately 400 six-inch documents per
minute - about 4 per cent slower than the reading
rate of the 1418 Optical Character Reader described in the preceding paragraph.
.
The 43 characters which can be read are as follows:
upper case letters A through Z, digits 0 through 9,
dollar sign ($), diagonal (/), asterisk (*), dash (-),
decimal point (.), comma (,), and a solid vertical
line preprinted on the documents. Optional features
permit the reading of pencil marks or slanted
marks, as described for the 1418 above.
The 1428 Alphameric Optical Reader, Model 1, 2,
or 3, is the same in all other respects, including
the optional features, as the 1418 Optical Character Reader described above. A detailed description of the 1428 appears in Section 401: 105 of the
IBM 1401 report.
.125 IBM 1231 Optical Mark Page Reader, Model Nl*
The Optical Mark Page Reader, Model Nl, reads
ordinary pencil marks (not printed characters)
from 8-1/2 by 11 inch data sheets directly into a
System/360 Data Processing System.
The 1231 (or its off-line counterpart, the 1232) will
be useful in organizations that use standardized
forms for such functions as surveys, orders, applications, medical records, payroll time records,.
inventory listings, and sales analyses. The 1231's
chief advantage is the elimination of much of the
key punching and verifying normally associated
with the preparation of input for automatic data
processing. In a single pass of the pencil-marked
data sheets through the 1231, the marks are read
and the data is transferred to the computer. (The
1232, working off:-line, converts the marked information into punched cards.)
Documents are read at varying rates of speed, depending upon the mode switch settings. When set to
(Contd. )
7/65
420: 105.125
INPUT-OUTPUT: OPTICAL AND MAGNETIC READERS
.125 IBM 1231 Optical Mark Page Reader, Model Nl.
(Contd. )
"continuous", feeding is at a constant speed of 2,000
documents per hour. When set to "demand",
feeding is controlled by the computer program with
speeds varying up to 1,600 documents per hour.
The appropriate feeding mode depends upon the
computer program control method used.
.126 IBM 1285 Optical Reader*
The IBM 1285 Optical Reader is a recentlyannounced unit that can optically read numeric
digits and several upper case alphabetic characters
printed in the 1428 type font from continuous rolls of
paper (journal tapes) prepared by adding machines,
cash registers, and similar devices. The 1285
is available for use with the IBM 1401, 1440, and
1460 as well as the IBM System/360.
Data sheets are fed from a pneumatically-controlled
hopper (600-sheet capacity) through the reading
area and directed to one of two stackers. The main
stacker holds 600 sheets. Sheets with detected
errors are directed to a separate stacker (50-sheet
capacity) . Documents are stacked in reverse sequence in both stackers.
The 1285 reads the digits 0 to 9 and the upper case
letters C, N, S, T, X, and Z. Up to 25 characters
per line can be read from a tape whose width can
range from 1- 5/ 16 to 3-1/2 inches and whose length can
vary from 3 to 200 feet. Printing color should be
black for optimum operation of the 1285.
All marks read from a data sheet are stored as
bits in a sonic delay line storage unit until they
are transferred to the computer by execution of a
Read instruction. Less than 10 milliseconds are
required to transfer the data read from one data
sheet from the sonic delay line to the computer
storage.
The reading station contains an electronic flying
spot scanner consisting of a cathode ray tube, an
optical system, and photomultipliers. Initially
unreadable characters are automatically rescanned,
followed by either a halt for on-line correction by
the operator or by automatic marking of the unreadable line and an advance to the next line.
The user may have up to 1, 000 mark positions on
each side of a sheet (2, 000 on both sides). Mark
positions are pre-printed in rows of 20 positions.
Each row is divided into two 10-position groups.
Each 10-position group is called a "word" for the
purpose of defining a marking area. Each word
can be divided into two 5-position segments. Data
words and segments can be grouped in various
combinations to form fields.
First deliveries of the 1231 were made in the third
quarter of 1964.
Reading rates of the 1285 are in the range of 1, 500
to 2,500 lines per minute. The formula for determining the speed is:
LPM
= 1000
(
600
)
13N + 518/L + 14
where LPM = reading rate in lines per minute
N = characters per line
L = lines per inch.
Marking of lines which contain unreadable characters does not appreciably reduce the throughput.
Optional Feature
Master Mark: a master data sheet, containing up
to 10 words of marked data, can be read and stored
in the delay line for transfer to the computer. The
master sheet is identified by a special preprinted
mark and contains data that is to be associated
with all subsequent data sheets until a new master
sheet is read. Thus, data common to a series of
data sheets need be recorded and read only once.
* See Section 420:031, System Configuration, for
the rules governing the connection of this device
to a particular System/360 configuration.
©
Acceptable type styles are the IBM 1428 font and
the NCR Optical Type Font. The 1428 Alphameric
Optical Reader is described in paragraph . 124
above.
Optional Feature
NCR Optical Type Font: permits reading the "NOF"
type style developed by NCR; reading of this type
style and IBM 1428 type style is interchangeable
under operator switch control.
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
420:106.100
IBM System/360
Input-Output
2701 Data Adapter Unit
INPUT-OUTPUT: 2701 DATA ADAPTER UNIT
.1
GENERAL
• 11
Identity: •••••••.•• IBM 2701 Data Adapter Unit*
.12
Description
The IBM 2701 Data Adapter Unit (and the 2702 and
2703 Transmission Controls described in the next
two report sections) make possible direct connection
of a wide variety of data communications equipment
to'an IBM System/360. Numerous adapters are
available for connecting the 2701 to various types
of communications facilities and terminal equipment.
These adapters provide the necessary bit-byte
conversions, interfaces, and control circuits.
Devices that can be connected include the 1030 Data
Collection System, the.1050 and 1060 Data Communication Systems, the 1070 Process Communication
System, remote 2260 Display stations, telegraph
terminals, and the new 2740 and 2741 Communication
Terminals. See Section 420:031, System Configuration, for the rUles governing the· connection .of the
2701 to a particular System/360 configuration.
The adapters are contained on one or two gates.
Each gate can have two adapters for start/stop
communications lines or one adapter for a synchronous line. Thus, a maximum of four lines can
be controlled by a 2701. Some of the communications adapters can be equipped with an automatic call
feature. When this feature is used with start/stop
adapters, the total number of adapters that can be
included in the 2701 is reduced by one. One or
both of the adapters on the second gate can be connected to a different input-output channel from the
channel used by the other adapters by means of
the Second Channel feature .. This effectively converts a single 2701 into two logically independent
data communications controllers.
Input-output operations involving a 2701 are programmed in much the same way as operations
involving other System/360 peripheral devices.
See Section 420:111, SimUltaneous Operations, for
a general description of the System/360 inputoutput process. In addition to the normal I/O
commands, special diagnostic commands have also
been implemented. These commands vary according to the particUlar communications adapter,
but in general they permit sending data to a special
one-character register and reading it back into
core storage. This permits many of the functions
of the adapter to be checked.
The 2701 operates in a half-duplex mode (transmission in either direction, one way at a time), and
can use the following types of lines:
* See Paragraphs .121 through .135 of this report
section for descriptions of the numerous devices
and systems that can communicate with a 2701
Data Adapter Unit.
7/65
•
Common-carrier switched telephone network
at 150, 1200, or 2000 bits per second .
•
Common-carrier leased private-line telephone'
service, at up to 2400 bits per second.
•
Common-carrier leased private line telegraph
service at 45, 55, or 75 bits per second, depending on service.
•
Teletypewriter Exchange (TWX) Network at
150 bits per second.
•
Telephone company schedUle 3A Data Channel
(150 bits per s~cond).
• Western Union Class D Channel (180 bits per
second).
•
Western Union Class E Channel (600 or 1200
bits per second).
• Western Union Class F Channel (2400 bits
per second).
•
Privately-owned or· leased communication networks.
•
Common-carrier broadband communication'
service at 19,200 or 40,800 bits per second
(Telpak A).
Special functions that can be provided by the 2701
Data Adapter Unit and its various adapters include:
•
Parallel data transmission - the Parallel Data
Adapter provides a data path from 16 to 48 bits
wide (in 8-bit increments) that permits parallel
half-duplex data transfers at high speeds (limited by the computer configuration) between the
2701 and anyone of up to 8 connected devices.
Odd parity checking is provided.
•
Synchronous serial data transmission using 4out-of-8 checking code - provides for connection of IBM Synchronous Transmit-Receive
(STR) communication terminals or other terminals satisfying a standard industry interface
specification (RS 232). Such IBM terminals
are the 1009, 1013, 7701, 7702, 7710, 7711,
and the 7740 and 7750 program-controlled
terminals. These units are described more
fUlly on -the followiilg pages.
•
Automatic calling - permits processor-initiated
automatic dialing on common-carrier switched
telephone or 150-bits-per-second TWX networks.
(Automatic answering of incoming calls can be
provided by use of the proper common-carrier
data set.)
(Contd. )
420: 106.120
INPUT-OUTPUT: 2701 DATA ADAPTER UNIT
. 12
Description (Contci.)
•
Programming systems under the Basic Operating System/360 will be provided for 1030,
1050, 1060, and 1070 systems, for 2740 Terminals, and for Telegraph Adapters.
•
The 2741 Communications Terminal will be
supported by the Administrative Terminal
System/360 application program.
Specific adapters enable the 2701 to communicate
over appropriate communication facilities with the
following terminals:
• IBM 1050 Data Communication Systems - 14.8
or 8.33 char/sec.
•
IBM 1060 Data Communication Systems - 14.8
char/sec.
•
IBM 1070 Process Communication Systems 14.8 or 66.6 char/sec.
• IBM 1031 Input Stations - 60 char/sec.
• IBM 1033 Printers - 14.8 char/sec.
•
Bell System 83B2 or 83B3 Selective Calling
Terminals - 45, 57, or 75 bits/sec.
• Western Union Plan 115A Outstations - 57 or
75 bits/sec.
•
Common Carrier TWX Stations using 8-level
code - 110 bits/sec. only.
•
1009 Data Transmission Units - 150, 250, or
300 char/sec.
•
1013 Card Transmission Terminals - 150, 250,
or 300 char/sec.
•
2740 Communications Terminal- 14.8 char/sec.
•
2741 Communications Terminal- 14.8 char/sec.
•
7701 Magnetic Tape Transmission Terminals 150 char/sec.
•
7702 Magnetic Tape Transmission Terminals 150, 250, or 300 char/sec.
•
7710 or 7711 Data Communication Units - 150,
250, 300, 2400, or 5100 char/sec.
•
7740 Communication Control Systems - 150,
250 or 300 char/sec.
•
7750 Programmed Transmission Controls - 250
char/sec.
•
System/360 Model 20 equipped with a Communication Adapter - 150, 250, or 300 char/sec.
•
2848 Display Control with 2260 Display Stations 120 or 240 char/sec.
•
Another System/360 equipped with a 2701 Data
Adapter Unit - 150, 250, 300, or 5100 char/sec.
IBM programming support for 2701 users will be
as follows:
• Programming systems under the Operating
System/360 will be provided for 1030, 1050,
and 1060 systems, for 2740 Terminals, and
for Telegraph Adapters.
©
Descriptions of most of the devices and systems
that can communicate with a 2701 Data Adapter
Unit are presented in the following paragraphs.
.121 IBM 1050 Data Communication System
The 1050 is a multi-device, general-purpose
communication system that operates and transmits
data at 14.8 or 8.33 characters per second over
toll, leased,. or privately-owned half-duplex telephone, telegraph, or TWX circuits. Input can be
from punched tape, punched cards, or keyboard;
and output can be printed or punched on tape or
cards. Devices which are not on-line at any particular time can'be used for off-line data recording
or preparation.
Every 1050 system includes a 1051 Control Unit,
which provides the basic connection between the
input-output units and the modulator-demodulator
equipment for the transmission line. A 1050 system can also include one 1052 Printer-Keyboard
and one or two of each of the following units: 1053
Printer, 1054 Paper Tape Reader, 1055 Paper
Tape Punch, 1056 Card Reader, 1057 Card Punch,
and 1058 Printing Card Punch. The only configuration restrictions are that neither the total number
of readers nor the total number of punches may
exceed two.
The associated remote transmission control units
can provide communication with the System/360,
the 1240 Bank Data Processing System, and nearly
any IBM 1400 or 7000 series system.
.122 .IBM 1060 Data Communication System
The 1060 System is designed for on-line or off-line
banking and savings and loan applications, and
utilizes 1061 Control Units and 1062 Teller Terminals.
When used on-line with a data processing system, the
1060 operates in a half-duplex mode over commoncarrier leased telephone or sub-voice-grade lines
or private lines, at a speed of 14.8 characters
per second.
The associated remote tr.ansmission control units
can provide communication with the System/360,
the 1240 Bank Data Processing System, and nearly
any IBM 1400 or 7000 series system. (see also
Section 414:102. )
.123 IBM 1070 Process Communication System
The 1070 is a new IBM communication system designed to handle two-way transmissions of data and
control information between remote process locations and a central System/360 equipped with a
2701 Data Adapter Unit or a 2702 Transmission
Control (or an IBM 1050 Data Communication System). In an on-line process control application,
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360
420:106.123
. 123 mM 1070 Process
Comm~cation
System (Contd.)
a 1070 system can address up to 300 input-output
points, and can perform input-output functions
involving digital/analog conversion, contact operation and sensing, BCD and decimal transfers,
thermocouple sensing, data display, and printing.
The 1071 Terminal Contror Unit controls selection
and conversion of process signals in either a random
or sequential mode for up to 50 points. Additional
system units provide a variety of other functions.
Transmission is in half-duplex mode at 14.8 or 66. 6
characters per second, depending upon the communication facilities used.
. 124 IBM 1031 Input Station and 1033 Printer
These units are elements of an IBM 1030 Data
Collection System. The 1030 can operate as an
independent privately-wired system for in-plant
data collection, or it can transfer data in either
direction over communication facilities. Associated IBM data processing systems can be the IBM
System/360, the 1440 or 1460 systems, the 1240
Bank Data Processing System, and the 1410 or 7010
systems via a 1440.
The 1031 Input Station transmits data at 60 characters per second from cards, badges, data cartridges, or manually-set slides.
The 1033 Printer provides alphameric printed copy
at remote locations at a rated speed of 14.8 characters per second. Used in conjunction with the
1031 Input Station, the 1033 Printer provides twoway on-line inquiry and reply capabilities.
.125 Bell System 83B2 and 83B3 Teletypewriter
Selective Calling System
The 83B2 is a private line. service that sends and
receives messages over one or more telegraphspeed circuits. It provides for automatically selecting in sequence ("polling") automatic punched
tape transmitters at a central office and up to 39
remote stations. The messages contain calldirecting codes which determine the message routing. Messages are transmitted at speeds of 60, 75,
or 100 words per minute (6, 7.5, or 10 characters
per second). The central or switching office
punches each message as received, and then transmits the message over the required output line as
directed by the address code when the line becomes
free.
The 83B3 service is similar to the 83B2 but
features a higher polling rate.
The 1009 is able to transmit or receive data at
line speeds of 75, 150, 187.5, 250, or 300 characters per second when using appropriate line
facilities. The terminal at the remote end of the
line can be: another 1009, a Magnetic Tape Transmission Terminal (a 7701 at 75 or 150 char/sec or
a 7';'02 at 150, 250, or 300 char/sec), a 1013 Card
Transmission Terminal (at 150, 250, or 300 chari
sec), a 7710 or 7711 Data Communication Unit (at
150, 250, or 300 char/sec), a 7740 Communication
Control System (at 150, 250, or 300 char/sec), a
7750 Programmed Transmission Control (at 150
char/sec). or the 2701 Data Adapter Unit (at 150,
250, or 300 char/sec) .
Data is transferred in a serial-by-bit synchronous
(timed) mode which does not require start-stop
bits. Characters are transmitted in 8-bit groups,
using either 4-out-of-8 checked BCD codes or
binary characters.
.128 IBM 1013 Card Transmission Terminal
The 1013 is an independent, synchronous communication terminal used for transmission and reception
of punched card data with another 1013, with other
IBM synchronous terminals, or with data processing
systems via appropriate control units. The 1013
can transmit 1 to 80 columns per card of Hollerithcoded data only, via a buffer which stores up to
329 characters per transmission record. Data from
up to 7 cards can be sent per transmission record,
with some fixed data being stored in the buffer as
part of the 329 characters if desired.
Transmission takes place at line speeds of 150,
250, or 300 characters per second over appropriate
dialed or leased communication facilities, resulting
in card transmission rates of 50 to over 400 cards
per minute (depending on the number of characters
per card, cards per record, line speed, and type
of receiving terminal). Reception of data by a
1013 permits card punching at 160 columns per
second, which will limit the card rate to 30 fullypunched cards per minute at best. Transmission
is in the half-duplex mode.
Alternative receiving terminals can be the IBM
1009, 7701, 7702, 7710, 7711, 7740, or 7750, as
well as the 2701 Data Adapter Unit of a System/360.
These units permit the high~r card transmitting
rates to be'achi,eved and are described elsewhere
in this report section.
. 129 IBM 7701 and 7702 Magnetic Tape Transmission
Terminals
. 126 Western Union 'Plan 115A:
Plan 115A is a Western Union service which is
generally similar to the Bell System 83B2 service
described above.
.127 IBM 1009 Data Transmission Unit
The 1009 is a data transmission control unit used
to connect a number of different IBM data processing systems to communication lines. These
systems are the 1401, 144,0, and 1460, as well as
the 1410, 7010, 7040, 7044, 7070, 7074, 7080,
7090, 7094, and 7094-11 when connected to a 1414
Input-Output Synchronizer, Model 4, 5, or 6.
7/65
Both the 7701 and 7702 are independent, synchronous
communication terminals used for transmission and
reception of magnetic tape data with another 7701
or 7702, with other IBM synchronous terminals, or
with data processing systems via appropriate control units. The 7702 is a newer tape transmission
terminal than the 77 01 and can transmit or receive
at line speeds of 150, 250, or 300 characters per
second oyer appropriate dialed or leased communi·cation facilities. The 7701 operates at 75 or 150
characters per second. Tape records of any size
can be coded in either binary or BCD format. Recording density is 200 characters per inch. Polyester or heavy duty tape is recommended.
(Contd.)
INPUT-OUTPUT: 2701 DATA ADAPTER UNIT
420: 106.129
.129 IBM 7701 and 7702 Magnetic Tape Transmission
Terminals (Contd.)
The table below shows the terminals which can
communicate with the 7701 and 7702, and the
usable line speeds.
Line Speed (characters
per second)
Unit
7701
7701 Magnetic Tape Trl:!l1smission Terminal:
7702 Magnetic Tape Transmission Terminal:
1009 Data Transmission
Unit:
1013 Card Transmission
Terminal
7710 Data Communication
Unit:
7711 Data Communication
Unit:
7740 Communication Control System:
7750 Programmed Transmission Control:
7702
75/150
150
150
150/250/300
75/150
150/250/300
150
150/250/300
150
150/250/300
150
150/250/300
150
150/250/300
150
150
All of the listed units are described elsewhere in
this report section.
Data is read-checked by means of transverse and
longitudinal parity checking. The receiving terminal performs the same checks, and magnetic tape
recording terminals also perform a read-afterwrite data check. The units will stop and signal
the operator only after three unsuccessful attempts
to record a: valid record. Both the 7701 and the
7702 will operate in either the half-duplex or fullduplex mode.
. 130 IBM 7710 Data Communication Unit
The 7710 is a synchronous data transmission control unit used with an IBM ;I.401 Data Processing
System to permit data transmission or reception
at 150, 250, or 300 characters per second over
suitable voice-grade communication services and
at 5,100 characters per second over broad-band
communication services. Two 1401's can communi-cate with one another by means of 7710's under 1401
program control, sending variable-length records
in binary or BCD formats. Transmission code is
the 4-out-of-8 checking cQde, and a longitudinal
redundancy check character is included. Retransmission is performed under program control.
A 1401-7710 combination can communicate with the
IBM 1009, 1013, 7701, 7702, 7711, 7740, or a
2701 Data Adapter Unit used with a System/360.
. 131 IBM 7711 Data Communication Unit
The 7711 is an independent, synchronous communication terminal used for transmission and
reception of magnetic tape data with another 7711,
with other IBM synchronous terminals, or with
data processing systems via appropriate control
units. The 7711 controls either an IBM 729 Series
or 7330 Magnetic Tape Unit, unlike the 7701 and
7702 which have self-contained tape transports. A
©
200-character buffer is standard with the 7711, and
its capacity can be expanded to 2,400 characters.
Line speeds are the same as those of the 7710 Data
Communication Unit: 150, 250, 300, or 5100
characters per second. Data checking is done as
in the 7710. The usable associated units and data
processors are the same as those listed in the 7710
description above.
.132 IBM 7740 Communication Control System
The 7740 is a stored-program computer designed
to provide message storage and switching, using
common-carrier lines and its own switching facilities. By means of appropriate communication
services, the 7740 system is capable of data transmission and reception using telegraph-speed equipment and higher-speed (up to 300 characters per
second) synchronous data transmission terminals.
The 7740 system can stand alone as an independent
message switching system, but will more typically
be used as a switching facility for an IBM data
processing system. As an independent system, the
7740 can connect to up to five IBM 1311 Disk Storage Drives, each capable of storing over two mil~
lion characters. (Magnetic tap!,!s are not available
to the 7740.) As a system serving a "host" computer, the 7740 transfers messages to the·host
computer for disc or magnetic tape storage. The
7740 can be connected directly to any of the following
IBM computers: 1401, 1410, 1440, 1460, 7010,
7040, 7044, 7070, 7074, 7080, 7090, 7094.
From 1 to 84 low-speed half-duplex lines (45 to 200
bits per second) and 1 to 4 high-speed half-duplex
lines (1200 to 2400 bits per second) can be connected to the 7740 system. The maximum number
of each type of line depends on the combination of
lines desired, and on Whether or not a disc file is
connected. Pairs of half-duplex connections can be
used to form full-duplex connections .
Communication lines can be leased (private) lineS
or toll lines to exchanges. Telegraph or voicegrade lines or exchanges can be used by the system.
The 7740 has facilities for automatic dialing and
for automatic answering of dialed input messages.
Transmission terminals and systems that can be
connected to a 7740 include:
•
Telegraph terminals, operating at speeds up to
180 bits per second (approximately 24 Baudot
characters per second).
•
IBM 65/66 Card Data Transceivers.
•
IBM 1009 Data Transmission Units.
•
IBM 1013 Card Transmission Terminals.
•
IBM 1050 Data Communication Systems .
•
IBM 1060 Data Communication Systems.
•
IBM 7701 and 7702 Magnetic Tape Transmission
Terminals.
•
IBM 7710 Data Communication Systems.
•
Other IBM 7740 systems.
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360
420:106.132
• 132 IBM 7740 Communication Control System (Contd.)
•
IBM 7750 Programmed Transmission Controls
(~n~ect~~ to other IBM Data Processing Systems).
•
IBM System/360.
The total number of characters per second that the
system can handle varies greatly with the amount
of processing performed on header and message
characters, and on overall message processing.
Typical gross throughput values range from 300
to 1,000 characters per second.
A symbolic assembly program is available for
assembling 7740 programs; it runs on an IBM 1401
or 1410 computer. Card or magnetic tape assembler
output can be loaded into the 7740 from an associated computer. Assembler output on an IBM 1311
Disk storage Drive can be used for direct loading
into a 7740.
For further details on the IBM 7740, refer to
Section 414:106.
. 133 IBM 7750 Programmed Transmission Control
The 7750 is a stored-progrartJ. computer designed
to provide message storage and switching using
common carrier lines and -its own switching facilities. It performs the input-output function for a
directly-connected "host" coinputer, which can
be an IBM 1410 Data Processing System or any
7000 series system except the 7072. Unlike the
IBM 7740 Communication Control System, the 7750
has no provision for connection to its own disc
storage file. Using appropriate communication
services, the 7750 is capable of data transmission
and reception using telegraph-speed equipment and
higher-speed (up to 150 characters per second)
synchronous data transmission terminals.
The 7750's stored program is generally used to
perform code conversion, editing, formatting,
network monitoring, checking, etc., and to control the assembly and distribution of messages.
. Data processing can be performed in the host computer. From 4 to 112 low-speed lines (45 to 200
bits per second) or 1 to 16 high-speed lines (voice
quality, 1,200 bits per second) or a combination of
the two types can be connected. Communication
lines can be leased (private) lines or toll lines to
exchanges. Telegraph or voice-grade facilities
can be used.
Transmission terminals and systems that can be
connected to a 7750 include:
•
7/65
Telegraph terminals, operating at speeds up to
180 bits per second (approximately 24 Baudot
characters per second).
•
IBM 65/66 Card Data Transceivers •
•
IBM 1009 Data Transmission Units.
• . IBM 1013 Card Transmission Terminals.
•
IBM 7701 and 7702 Magnetic Tape Transmission Terminals.
•
IBM 7740 Communication Control Systems.
•
Other IBM 7750 systems.
•
IBM System/360.
A symbolic assembly program using mnemonic
operation codes and symbolic addresses ·is available for assembling 7750 programs. This assembly
program includes a Data Control Package and an
Input-Output CQntrol System. After assembly,
programs are loaded into the 7750 via the host
computer.
For further details on the IBM 7750, refer to
Section 402:105.
.134. 2740 Communications Terminal
The 2740 is a new IBM remote terminal designed
around the IBM Selectric Typewriter; it provides
manual keyboard input and printed output. Data
is transmitted or received in a half-duplex mode at
14.8 characters per second over common-carrier
switched telephone, TWX, or leased or private
communications facilities capable of operating at
133. 2 bits per second. Optional features permit
dial-up operation over common-carrier dial networks, vertical and longitudinal parity checking, and
multistation operation (multiple stations interconnected by a single co~munications line).
.135 2741 Communications Terminal
The 2741 is another newly-announced remote terminal designed around the IBM Selectric Typewriter.
It is similar to the 2740 (see above) but can operate
only in a point-to-point mode; 1. e., one terminal
per line .. The .2741 is primarily intended for use
as a remote inquiry station for a System/360.
Dial-up operation over the common-carrier dial
telephone network is available as an optional
feature when used with the apprOPriate data set.
When connected to a System/360 via a 2702 or
2703 Transmission Control (see Sections 420:107
and 420:108), but not via a ·2701, the 2746 can interrupt the computer while the computer is. transmitting data to it .. A full-duplex circuit is required
for this facility.
420:107.100
IBM System/360
Input-Output
2702 Transmission Control
INPUT-OUTPUT: 2702 TRANSMISSION CONTROL
.1
GENERAL
.11
Identity: .••.•••... IBM 2702 Transmission
Control.
. 12
Description
The 2702 Transmission Control permits on-line connection of various low-speed communication terminals to a System/360 via private or commercial
common-carrier transmission facilities. See
Section 420:031, System Configuration, for the
rules governing the connection of the 2702 to a
particular System/360 configuration.
The 2702 differs from the IBM 2701 Data Adapter
Unit in transmission speeds and in the number of
lines it can handle. The basic unit provides for
attachment of up to 15 half-duplex communication
lines, all of which can transfer data simultaneously
at speeds up to 200 bits per second per line. Each
line has ,an eight-bit buffer used for data transfers
to or from the Processing :Unit. Optional features
enable a 2702 to handle up to 31 lines, at 200 bits
per second each, or the basic 15 lines at up to 600
bits per second each.
The 2702 operates in a start/stop mode, and data
transmission is serial by bit. The 2702 accomplishes all necessary bit-byte conversion, data
control, and interfacing functions. Characters
from incoming messages are interleaved and assembled in processor storage, so the 2702 imposes
no restrictions upon message length. Optional
features permit processor-initiated automatic
dialing on common-carrier switched telephone or
150 bits-per-second TWX networks on a maximum
of 16 lines. Automatic answering of incoming calls
can be provided by use of the proper commoncarrier data set.
Input-output operations involving the 2702 are
programmed in much the same manner as other
System/360 peripheral operations; each line is
individually addressed. A general description of
the input-output process is presented in Section
420:111, Simultaneous Operations.
A special diagnostic command, Autowrap, is provided to allow the functioning of the 2702 to be
checked under program control. This command
causes the output of an addressed line to be
"wrapped" around to the input of line O. If line 0
is free, the data can be read into core storage and
checked. The test data can be inhibited from being
sent over the communications line.
The 2702 can utilize common-carrier switched
telephone or 150-bits-per-second typewriter Exchange (TWX) networks, leased private line telephone or telegraph service, Western Union Class
D or E (180 or 1200 bits-per-second) channels,
and privately-owned communication networks.
Specific terminal devices that can be connected to
a 2702 include:
•
1060 Data Communication System - 14.8 chari
sec .
•
1050 Data Communication System - 14.8 or
8.33 char/sec.
•
1031 Input Station - 60 char/sec.
•
1032 Digital Time Unit - 600 bits/sec.
•
1033 Printer - 14.8 char/sec.
•
1070 Process Communication System - 14.8
or 66.6 char/sec.
•
2740 Communications Terminal- 14.8 chari
sec.
•
2741 Communications Terminal sec.
•
Bell System 83B2 Selective Calling Terminals 45, 57, or 75 bits/sec.
•
Western Union Plan 115A Outstations - 57 or
75 bits/sec.
•
Common-carrier TWX stations using 8-level
code - 110 bits/sec. only.
14.8 chari
•
2712 Remote Multiplexor - up to 150 chari
sec; see description of the 2712 on page
420:108.100.
The 1032 Digital Time Unit can be connected
directly to a 2702 by a 20-foot cable; the 1032
provides the time of day in hours and hundredths
of an hour. It occupies one line position of the
2702. Two models are available: Modell contains a synchronous motor drive; Model 2 is
minute impulse self-regulatmg for attachment to
a user's master clock system.
Refer to the previous report section (420:106) for
descriptions of the functions of the other data
communications devices and systems listed above.
IBM Programming support for users of the 2702
Transmission Control will be as follows:
•
Programming Systems under the Operating
System/360 will be provided for 1030, 1050,
and 1060 Systems, for 2740 Terminals, and for
Telegraph Terminal Controls.
•
Programming Systems under the Basic Operating
System/360 will be provided for 1030, 1050,
1060, and 1070 systems, for 2740 Terminals,
and for Telegraph Terminal Controls.
•
The 2740 Communication Terminal is supported
by the Administrative Terminal application
program.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
~
420: 108.100
~
IBM System/360
Input-Output
2703 Transmissian Cantrol
INPUT-OUTPUT: 2703 TRANSMISSION CONTROL
.1
GENERAL
.11
Identity: •.••••..•• IBM 2703 Transmission
Control.
IBM 2712 Remote Multiplexor.
.12
Description
The 2703 Transmission Control is a new (July 1965)
addition to the IBM System/360 line of communications controllers. The 2703 is similar to the 2702
(Section 420:107) in the types of remote terminals
and communications facilities that it can accommodate (except that the 1032 Digital Time Unit cannot
be connected to a 2703); see Section 420:107 for a
list of these facilities. Descriptions of each of these
facilities are contained in Sections 420:106 and
420:107. See Section 420:031 for the rules governing
the connection of the 2703 to a particular System/
360 configuration.
The primary differences between the 2703 and the
3702 are in the number of lines that can be connected
and in certain hardware features. The 2703 can
accommodate up to 176 low-speed communications
lines (up to 180 bits per second) or up to 72 lines
operating at 600 bits per second. All lines can be
active simultaneously. The greater line capacity
of the 2703, as compared with the 2702, can free
control unit positions on the Multiplexor Channel,
thereby making it possible to connect more card
readers, punches, printers, etc.
Hardware features of the 2703 include automatic
polling (with no program interrupts when negative
responses are received) and a four-character
buffer for each line. The optional Two Processor
Switch permits the 2703 and the connected communications equipment to be switched between the
Multiplexor Channels of two different System/360
Processing Units.
Programming support for 2703 users will include
provisions for IBM 1030, 1050, and 1060 systems,
the 2740 and half-duplex telegraph terminals, and
the Auto-Poll capability. These capabilities will
be available with the Basic Operating System/360
(16K Disk) and the Operating System/360 in late
1967.
2712 Remote Multiplexor
The 2712 Remote Multiplexor is a new IBM unit
(also announced in July 1965) that permits
messages from multiple low-speed lines to be
multiplexed over a single voice-grade line. The
low-speed lines are terminated at the 2712 via data
sets in the normal manner. The 2712 is connected
to a System/360 through a 2702 or 2703 Transmission Control with the appropriate adapters and
data sets.
Two models of the 2712 are available. Modell
can handle up to 10 lines from IBM 1050, 1060,
2740, or 2741 units in any combination. Model 2
can handle up to 14 Teletypewriter lines. All
lines can be either single-station (point-to-point)
or multistation, and all can operate simultaneously
and independently. Use of a 2712 Remote Multiplexor has no effect on programming or logical
considerations. A 2702 Transmission Control (see
Section 420:107) can communicate with up to two
2712's; a 2703 Transmission Control can communicate with up to four 2712's. Each line among the
group of low-speed lines concentrated by a 2712
requires one line position of a 2702 or 2703.
,/
7/65
420: 109. 100
IBM System/360
Input-Output
Film Units
INPUT-OUTPUT: FILM UNITS
.1
GENERAL
There are four basic modes of data recording:
.11
Identity: . . . . . . . . . . 2280 Film Recorder.
2281 Film Scanner.
2282 Film Recorder/
Scanner.
•
Vector mode - In this mode a straight line can
be recorded between any two addressable points.
This mode can also be used to position the beam
at a desired location without recording the path.
. 12
Description
•
Point plot mode - Individual points are plotted
by positioning the beam at a specific location
and then turning on the beam.
•
Character mode - The character generator of
the 2840 Display Control can generate a series
of vector addresses corresponding to alphanumeric data in the buffer. These addresses are
converted into analog values and applied to a
special deflection yoke on the cathode ray tube.
The characters are written by a series of short,
high-speed strokes. Three program-selectable
character sizes of 14, 28, and 56 raster units
can be recorded in this mode. Spacing between
characters of up to 126 raster units can be specified in the command and is performed automatically by the circuitry.
These Film Units use unsprocketed 35-millimeter
film as the input-output medium for graphic and/
or alphanumeric information. A cathode ray tube
is used to record or scan images within a 4, 096by-4, 096-point matrix on a 1. 2-inch (nominal)
square frame of film.
The Film Units are connected to a 2840 Display
Control, as described in Section 420:101. Film
Units and 2250 Display Units can be connected to
the same 2840 Display Control. The buffer and
character generator of the Display Control are
time-shared by all attached units. Up to four Film
Units can be attached to a 2840 Display Control.
See Section 420:031, System Configuration, for the
rules governing the connection of Film Units and
2250 Display Units to a 2840 Display Control and to
a particular System/360 configuration.
. 121 2280 Film Recorder
The 2280 Film Recorder provides facilities for recording data on film, for developing the image,
and for viewing the proceS'Sed image. The standard
film produces dark lines on a light background; note
that this type of image is the reverse of the type
read by the 2281 and 2282 Film Scanner Units. The
film path consists of a supply cassette with a film
capacity of 400 feet, a recording station, a developing station, a projection station, and a take-up
c,assette. Vacuum columns between the stations
allow the various functions to proceed asynchronously by buffering film movements between the
stations. An exposed (recorded) image can be
processed and projected on the 22. 8-inch-square,
rear-projection screen within 48 seconds after the
exposure. Images previously projected can be
backspaced by the operator to the projection station.
Digital information from the buffer of the 2840
Display Control is used to generate analog signals
that control the positioning of the electron beam
from the cathode ray tube. All recorded images
are composed of straight lines and/or points. A
straight line can be recorded between any two addressable points. A point can be recorded at any
addressable location. A matrix of 4, 096 points can
be addressed within the 1. 2-inch-square frame of
film. , The distance between two adjacent addressable locations is called a raster unit (RU). Either
of two line densities and either of two line widths
can be selected under program control.
©
• Stroke mode - The analog signals are generated
as in the vector mode, but these ,signals are applied to the character yoke rather than to the
main deflection yoke of the cathode ray tube.
This mode provides the capacity for generating
short strokes within anyone of the standard
character sizes at a high rate of speed. Special
symbols not available from the character generator can be written in significantly less time
in the stroke mode than in the vector mode.
•
The film is advanced past the recording station
under program control in increments of O. 756,
1. 008, 1. 212, or 2. 004 inches. With proper programming, continuous images longer than the 1. 2inch frame size can be recorded.
The time required to record a line or pOSition the
beam in the vector mode is a function of the line
length. This time varies between 102 and 408
microseconds. The average time required to
write a character and space to the next character
position is 24 microseconds. The time required
to reposition the beam to begin a new line of data in
the character mode is 150 microseconds. Short
lines can be recorded in the stroke mode at the
rate of one every four microseconds. The time required to advance the film past the recording station varies between 147 and 321 milliseconds, depending upon the frame advance increment. The
exposed film can be processed at the developing
station at the rate of 40 inches per minute.
. 122 2281 Film Scanner
The 2281 Film Scanner converts negative images
(light lines on a dark background) on unsprocketed
35-millimeter film into digital data. Note that it
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360
420: 109.122
122. 2281 Film Scanner (Contd.)
Film can be advanced under program control in increments of O. 756, 1. 008, 1. 212, or 2.004 inches.
In addition, the film can be registered in the film
gate in increments of O. 012 inch under program control. A 3-inch-square viewing screen allows the
operator to view the image at the scan gate.
will not be practical for a 2281 to read the dark-onlight film images normally produced by the 2280 and
2282 Film Recorder units; this could be accomplished only by reading the light areas and interpreting (by program) the dark lines as those areas
which are not found to be light.
The operator can manually register the film image
at the film gate. IBM states that digital responses
can be obtained from high-contrast images with lines
as fine as 1/2000 of the image size and spaced
1/500 of the image size (center-to-center), if the
scan vector length does not exceed 2.5% of the
image size or if the scan stroke length does not exceed 24 raster units. If a 24-inch-square document
were reduced to the 1. 2-inch- square film format,
these limits of resolution would correspond to lines
0.012 inch thick and spaced O. 048 inch apart in the
original document.
Information can be retrieved from a 1. 15-inchsquare image that is registered in the film gate.
The beam from a cathode ray tube is directed along
two paths: one through the film to a photomultiplier
tube, and the other directly to a second photomultiplier tube. The amount of light passing through the
film is compared with the light intensity of the,direct
beam. A response which is above a preset threshold sensitivity is considered a "strike." Anyone
of 63 threshold values can be selected under program control, which provides the ability to analyze
film images of varying density and contrast.
There are two basic modes of operation, the scan
vector mode and the scan stroke mode.
In the scan vector mode, .the beam from the cathode
ray tube is moved in a straight line from one addressable matrix point to another. The same
4, 096-by-4, 096-point reference grid used in the
Film Reader Units is used by the Film Scanner
Units. The path of the beam, or vector, is divided
into 16 logical parts, and the circuitry determines
whether a strike occurs in each part. The 16 bits
of information from each programmed vector scan
are recorded in two data bytes in the buffer of the
2840 Display Control. The scan vector mode can
also be used to position the electron beam without
scanning. The time required to generate a vector
scan varies between 102 and 408 microseconds, depending on the length of the vector.
Negative images can be reduced to digital form with
the 2281, but note that a complex computer program
will be required to interpret the digital information,
particularly if the subject material is alphanumeric
data.
. 123 2282 Film Recorder/Scanner
The IBM 2282 Film Recorder/Scanner provides the
capabilities of both the 2280 Film Recorder and the
2281 Film Scanner in a single unit. The operations
of film recording and film scanning, however" cannot be performed simultaneously. To convert from
one mode of operation to the other, the film must
be changed, the proper mode switches must be set,
and the status of the film processor must be reset.
When switching from recording to scanning, the
film developer heads must be blotted free of
chemicals.
In the scan stroke mode, the character yoke of the
cathode ray tube generates high-speed scan strokes
within a 14, 28, or 56 raster-unit square, as specified by the program. A single response is recorded
for each stroke. Scan strokes can be generated at
a rate of one every 20 microseconds.
7/65
.13
Availability: . . . . . . . ?
. 14
First Delivery: . . . . . 2nd quarter, 1966.
420: 111.1 DO
IBM System/360
Simultaneous Operations'
SIMULTANEOUS OPERATIONS
.1
GENERAL
The important characteristics of a channel are:
An mM System/360 processor* can concurrently
execute:
•
The channel capacity, normally expressed
in bytes per second.
•
One machine instruction; and
•
•
Up to six fast input-output operations,
depending upon the number of Selector
Channels in the system (see Table I); and
The demand on the processor, or "interference," normally expressed in terms of
the percentage reduction in internal proceSSing capacity while data transmission is
taking place.
•
Multiple slower input-output operations via
a Multiplexor Channel, if incorporated in
the system; and
•
As many previously-initiated buffered inputoutput operations as have not yet been completed.
Details of each System/360 model's capabilities
for simultaneous operations are presented in the
Simultaneous Operations section of the subreport
on the appropriate model (e.g., Section 423:111
for Model 30).
The number of simultaneous operations should not
be confused with the number of concurrent programs. In the System/360 only one program can
be executed by a Processing Unit at anyone time,
but a number of other programs may be residing
in main storage and utilizing input-output equipment, although the programs themselves are
passive. This "multiprogramming" mode of
operation is made possible by the built-in interrupt system and the software Operating System/
360, as described in Section 420:191.
.2
CHANNELS
Each System/360 input-output channel directs
the flow of information between main storage and
the connected input-output devices. The channel
provides a standard interface that permits many
different types of input-output devices to utilize
the same Processing Unit instructions and channel
commands. The channel contains the common
facilities required to control input-output operations. All chanp.els are physically located within
the Processing Unit in Models 30, 40, and 50;
in Mod!'lls 65 and 75, however, the channels are
separate units.
The System/360 uses two basic types of inputoutput channels: Multiplexor Channels and
Selector Channels. Both types are described
in succeeding paragraphs. See Section 420:031
for details concerning the availability of channels
and the connection of the various peripheral devices for a particular System/360 configuration.
*Except for the Model 20, whose input-output operations
are handled differently from those of the larger models,
as described in Section 422:111.
©
.21
Channel Capacity
The measurement of channel capacity is based
upon the highest instantaneous gross data transmission rate that can be safely maintained. For
this purpose, it is necessary to consider the peak
rates of all the peripheral units that can simultaneously transmit data through the channel. In
the separate subreports on each of the System/360
models, the capacity of each channel is listed,
along with the overall system data capacity and
the peak data rate of each of the individual peripheral units where known.
A channel may have more than one capacity listed
in its specifications. In this case, the different
capacities will correspond to different servicing
requirements during data transmission. There
are two major considerations that can affect
channel capacities in the System/360 computer
family: one depends upon how the program is
written, and the other upon how the data channel
is being used.
Programming considerations are related to tlie
concept of "chaining." Between the transmission
of one byte and the next, it is possible that the
channel control system will have to change the
input-output area in use. This will involve
obtaining the address of the new area to be used,
performing various operations on it, and bringing
it into use. All of these operations must be safely
completed before the next byte can be accepted by
the channel, and therefore the rate at which bytes
can be safely accepted when the data is chained is
conSiderably slower than the rate at which they can
be accepted if it is known that chaining does not
occur in the program. Indiscriminate use of
chaining by programmers could have drastic
effects on the functioning of a system, particularly
when very high-speed devices, such as Hypertape
or some random access storage units, are included. Because of the possible effect on system
performance, careful consideration should be
given to the use of chaining prior to the selection
of peripheral devices and during the system analysis
for individual jobs.
Hardware considerations are related to the
Multiplexor Channel. On this channel, it is
possible either to control a number of simultaneous operations or to operate in "burst" mode,
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360
420:111.210
.21
operations through automatic interleaving of
accesses to main storage.
Channel Capacity (Contll.)
in which case the channel can control only a
single operation at a time. Operation in the
burst mode eliminates the need to scan and service the other connected devices after each byte
is transmitted on the single operating subchannel,
and this greatly increases the safe operational
speed of the channel.
. 22
In the burst mode, a single input-output device
monopolizes all the channel controls throughout
the data transfer operation. The advantage of the
burst mode is that it can handle significantly
higher data transfer rates than the multiplex
mode. Internal proceSSing cannot be overlapped
with operation in the burst mode in Models 30 and
40, though it can be overlapped in larger models .
Processor Demands
The measurements of processor demands, or
"interference," are based on the average data
transmission rates during specific peripheral
operations, rather than upon the peak rates.
The difference between these two rates can most
easily be seen by considering the operation of a
buffered printer. Typically, when a print order
is issued, a complete line of 132 characters is
sent from memory to the printer buffer, which
may, for example, be able to accept the entire.
line within 1.1 milliseconds. This defines the
peak data transmission rate: around 120,000
bytes per second in this example. Subsequently,
the data in the buffer will be printed, the forms
will be advanced one line, and the printer will be
ready for another operation. The entire print
cycle takes 48 milliseconds in a 1,250-line-perminute printer, and it is this period which is
used to define the average data transmission rate
as 2,750 bytes per second, as compared with the
peak rate of around 120,000 bytes per second.
The Selector-Subchannel optional features for the
2870 Multiplexor Channel allow one to four relatively high-speed input-output operations (up to
100,000 bytes/sec each) to proceed concurrently
with operations on the basic channel of the 2870.
Each Selector-Subchannel clin have up to 8 control
units (a maximum of 16 peripheral devices) connected to it, but only one can be active at a time.
A special High-Speed Multiplexor Channel containing up to eight subchannels, depending on the
processor model, is available for Models 30,
40, and 50. This channel permits relatively
high-speed data transfel's in a multiplexing mode,
and can be installed in place of one Selector
Channel.
.24
Selector Channels can be used in Model 30 and all
larger models of the System/360. The maximum
number of Selector Channels per system is two
in Models 30 and 40, three in Model 50, and six
in Models 65 and 75. Each Selector Channel
transmits data to or from only one device at a
time. Input-output operations on all Selector
Channels can occur simultaneously and can be
overlapped with internal processing, provided
that the maximum data-handling rate of the
Processing Unit is not exceeded. Selector
Channels are designed primarily for high- speed
input-output devices such as magnetic tape units
and disc files, but most low-speed devices can
also be connected.
In the Simultaneous Operations sections of the
subreports on each of the System/360 models,
the average and peak data transmission rates
of each of the peripheral units, and the resulting
processor demand, are listed where known.
.23
Multiplexor Channels
A Multiplexor Channel provides facilities, for
time-sharing ("multiplexing") a single channel
among a number of slower peripheral devices.
While the theoretical maximum number of units
that can operate simultaneously is large (up to
256 in some configurations), the actual limiting
factor will usually be the total data rate capacity
of the channel (see "Max Data Rate" in Table I).
Only one Multiplexor Channel can be incorporated
in a System/360 configuration; it is standard in
Models 30, 40, and 50 and optional (2870 Multiplexor Channel) in Models 65 and 75. A Multiplexor Channel can operate in two modes:
"multiplexuor "burst."
In the multiplex mode, the channel can be timeshared by a number of simultaneously operating
low-speed input-output devices such as prInters,
card readers, card punches, and communication
terminals. The channel is effectively divided into
a number of "subchannels." Each subchannel consists of a group of storage locations holding the
addresses, count, and status information associated with one input-output operation. Thus, the
number of simultaneous input-output operations
that the Multiplexor Channel can accommodate
in the multiplex mode is limited only by the number of subchannels and by the channel's maximum
gross data rate. Internal proceSSing can always
be overlapped with multiplexed input-output
7/65
Selector Channels
A special High-Speed Channel is available for the
Model 50 only. This channel permits the operation of one very high-speed (up to 1,200,000
bytes per second) peripheral device. The -ProceSSing Unit and other input-output channels are
locked out during an operation on this channel.
• 25
Channel Capabilities
The input-output channel capabilities of the System/360 models are summarized in Table I. It
is noteworthy that at press time, 15 months after
announcement of the System/360 , mM still had
not officially specified many of the pertinent
figures regarding channel data rate capacities and
processor demands. Estimated figures have been
inserted, and marked accordingly, wherever the
speCifications are sufficiently firm to allow
reasonable estimates to be made.
.26
Channel-to-Channel Adapter
The Channel-to-Channel Adapter permits highspeed communication between two System/360
ProceSSing Units of any model by linking together
(Contd. )
SIMULTANEOUS OPERATIONS
. 26
420:111.260
Channel-to Channel Adapter (Contd.)
two input-output channels - one from each of the
two Processing Units. When this is done, each
Processing Unit can treat the other as a standard
peripheral unit. The adapter can be used on either
Multiplexor or Selector Channels and requires
one control unit position on each channel.
.27
Control Units
A control unit, which may be a separate unit
or an integral part of an input-output device,
adapts the characteristics of each type of inputoutput device to the requirements of the standard
channel interface. Up to eight control units can
be connected to each System/360 channel, and
many of the control units can accommodate a
number of input-output devices.
.3
INPUT-OUTPUT CONTROL
The System/360 (except Model 20) uses only four
which can be executed only when
the Processing Unit is operating in the supervisor
state:
I/o instructions,
•
start I/O - initiates an I/O operation and
specifies the channel and I/O device to be
used.
•
Test Channel - places an indication of the
channel's status (available, busy, not
operational, or interrupt pending) into the
Program Status Word.
•
Test I/O - supplies the program with
information (in the form of a 64-bit Channel
status Word) about the status of the specified
I/o device and its channel.
•
Halt I/O - causes immediate termination
of an I/O operation (usually because the
channel is required for an operation of
higher priority).
When the Start I/O instruction is executed, the
specified channel fetches a 32-bit Channel
Address Word (CAW) from a fixed main storage
location. The CAW specifies the main storage
location where the channel program for the
desired I/O operation begins. The channel.
program consists of one or more 64-bit Channel
Command Words (CCW's), which are executed
by the channel independently of the Processing
Unit (except that in the smaller System/360
models, some circuits are shared between the
Processing Unit and channel functions).
There are six channel commands: Read, Write,
Read Backward, Control, Sense, and Transfer
in Channel. The Read, Write, and Read Backward commands initiate the corresponding data
transfer operations. Control commands initiate
functions peculiar to certain I/O devices, such
as rewinding a tape unit, advancing forms on a
printer, and seeking a particular disc record .
Sense commands provide the program with detailed status information peculiar to a particular
I/O device. The Transfer in Channel command
simply causes a branch to a specified location
during execution of the channel program.
An input-output area is described in each Channel
Command Word, along with the channel command
itself. Where more than one input-output area
is required to complete the input-output operation, additional Channel Command Words are
"chained" to the original CCW simply by placing
them in the next sequential storage locations.
This facility allows scatter-read and gatherwrite operations, and is called "data chaining"
to distinguish it from "command chaining. "
Command chaining occurs where more than
one type of channel command is required to
complete the input-output operation - perhaps
a Write followed by a Rewind. Command
chaining, like data chaining, is accomplished
through the use of a series of Channel Command
Words in successive locations. Flags are
provided in each CCW to indicate which type of
chaining, if any, is to take place.
Two additional facilities, Skip and ProgramControlled Interruption (PCI), add to the flexibility
(and complexity) of the input-output capabilities
of the System/360. Both are specified by setting
flags in the appropriate CCW. If the skip flag is set
on, write operations are unaffected; read operations proceed as usual except that the data read
from the input medium is not transferred to main
storage. Skipping combined with data chaining
permits selected portions of a data block to be
read into main storage.
If the PCI flag is set on, an interrupt request
will be generated when that CCW is fetched. This
interrupt does not affect the execution of the current instruction. If chaining occurs prior to
servicing this interrupt, the condition is carried
over to the new CCW. The PCI conditions are
not stacked; i. e., if a new CCW that has the
PCI flag set on is fetched prior to servicing the
previous PCI condition, only one interrupt occurs.
The PCI flag can be used to alert the program to
the progress of chaining during an input-output
operation.
Input-output interruptions are caused by termination of an I/o operation, by fetching a CCW with
the PCI flag set on, or by operator intervention
at an I/O device. See Paragraph 420:051.123 for
a description of the interruption process.
When a Selector Channel is used for a data
transfer operation (Read, Write, or Read Backward), the selected channel is monopolized from
the time the command is issued until the data
transfer is completed. Control and Sense commands cause no transfer of data and tie up the
channel for only a short period of time (on the
order of 100 microseconds).
\
'"
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360
420: 111.300
TABLE I: SYSTEM/360 INPUT-OUTPUT CHANNEL CAPABIliTIES
SYSTEM MODEL
65 & 75
?
696
1,200
768
?
1,792
1
128
1
256
-
128
31
3.1%
est.
256
40
1.6%
est.
-
1
200
(4)
1
200
0.6%
40
SYSTEM DETAILS
Maximum throughput, KB/sec (1)
Maximum number of addressable devices
450
736
STANDARD MULTIPLEXOR CHANNEL_
Number of channels per system
Maximum number of Bubchannels
1
224
Multiplexed Mode
Maximum number of simultaneous data
transmissions
Maximum total data rate, KB/sec (2)
Processor demand, per KB/sec (3)
224
31
::6.25%
Burst Mode
Maximum number of simultaneous data
transmissions
Maximum total data rate, KB/sec (2)
Processor demand, per KB/sec (3)
1
267
(4)
2870 MULTIPLEXOR CHANNEL
Number of channels per system
Number of subchannels
Number of Selector-Subchannels
0
-
-
0
--
0
-
--
-
--
Selector Subchannels
Maximum number of simultaneous data
transmissions
Maximum total data rate, KB/ sec (2)
Processor demand per KB/sec (3)
-
--
-
--
o or 1
o or 1
Multiplexed Mode
Maximum number of simultaneous data
transmissions
Maximum total data rate, KB/sec (2)
4
200
Processor demand, per KB/sec (3)
0.15%
Burst Mode
Maximum number of transmissions
Maximum total data rate, KB/sec (2)
1
200
Processor demand, per KB/sec (3)
0.15%
SELECTOR CHANNELS
Number of channels per system
Maximum data rate per channel, KB/sec (2)
Maximum total data rate, KB/ sec (2)
Processor demand, per KB/sec (3)
HIGH SPEED CHANNEL
Number of channels per system
Maximum total data rate, KB/sec (1)
Processor demand, per KB/sec (2)
o to 2
250
450
0.15%
0
-
4
4
200
est.
0.25%
est.
-
o or 1
-
0.2%
est.
Oto3
400
800
0.05%
0
o or 1
(2)
Kilobytes (thousands of bytes) per second; ~ data rate, based on
no chaining and with no Transfer in Channel commands. The permissible data rate will be reduced if data chaining, command
chaining, or Transfer-in-Channel commands are used, and, in
some cases, if other input-output channels are in operationa
(3)
Processor demand (or "interference") for each kilobyte of data
being input or output during each second. (This is a measure of
the percentage of the total processor cycles which is required to
store or access the data being input or output and/or to control
the I/O process, and which is therefore unavailable for computational
purposes.)
(4)
Processing Unit is inhibited for the entire duration of each data
transfer operation of this tYPe.
0
8
?
Oto2
400
600
0.13%
Kilobytes (thousands of bytes) per second; this is the maximum data
rate with the most advantageous combination and usage of channels.
4
400
0.01%
-
0.05%
est.
(1)
o or 1
8
1
200
est.
0.13%
est.
-
1
-
-
192
110
0.08%
-
4
-
192
Oto4
-
-
0
-
Multiplexed Mode, Basic Channel
Maximum number of simultaneous data
transmissions
Maximum total data rate, KB/sec (2)
Processur demand, per KB/sec (3)
HIGH SPEED MULTIPLEXOR CHANNEL
Number of channels per system
Maximum number of subchannels
7/65
50
30
1
?
1,200
(4)
o to 6
?
1,300
0.0094%
0
-
.-
420: 121.1 00
IBM System/360
Instruction list
INSTRUCTION LIST: MODELS 30, 40, 50, 65, AND 75
The Instruction List presented below can be more easily understood if the report section
on the IBM System/360 Central Processors, Section 420:051, is previously read, with particular attention to the paragraphs entitled Instruction Format and Interrupt System. The instruction Type heading below refers to the five basic classes of instructions: Register to
Register (RR), Register to Indexed Storage (RX), Register to Storage (RS), Storage and
Immediate Operand (81), and Storage to Storage (8S). The instruction Exceptions, which refer to instruction or data situations that cause program interrupts, are listed with their
symbols at the end of this Instruction List.
Standard Instruction Set
NAME
Add
Add
Add Halfword
Add Logical
Add Logical
AND
AND
AND
AND
Branch and Link
Branch and Link
Branch on
Condition
Branch on
Condition
Branch on Count
Branch on Count
Branch on Index
High
Branch on Index
Low or Equal
Compare
Compare
Compare Halfword
Compare Logical
Compare Logical
Compare Logical
Compare Logical
Convert to Binary
Convert to Decimal
Diagnose
Divide
Divide
Exclusive OR
Exclusive OR
Exclusive OR
Exclusive OR
Execute
Halt 110
Insert Character
Load
Load
Load Address
Load and Test
Load Complement
Load Halfword
Load Multiple
Load Negative
Load Positive
LoadPSW
Move
Move
Move Numerics
Move with Offset
Move Zones
Multiply
MNEMONIC
EXCEPTIONS
TYPE
IF
IF
IF
CODE
lA
5A
4A
IE
5E
14
54
94
04
05
45
AR
A
AH
ALR
AL
NR
N
NI
NC
BALR
BAL
RR C
RX C
RX C
RR C
RX C
RR C
RX C
SI
C
SS C
RR
RX
BCR
RR
07
BC
BCTR
BCT
RX
RR
RX
47
06
46
BXH
RS
86
BXLE
CR
C
CH
CLR
CL
CLC
CLI
CVB
CVD
RS
RR
RX
RX
RR
RX
SS
SI
RX
RX
SI
RR
RX
RR
RX
SI
SS
RX
SI
RX
RR
RX
RX
RR
RR
RX
RS
RR
RR
SI
SI
SS
SS
SS
SS
RR
87
19
59
49
15
55
D5
95
4F
4E
83
DR
D
XR
X
XI
XC
EX
HIO
IC
LR
L
LA
LTR
LCR
LH
LM
LNR
LPR
LPSW
MVI
MVC
MVN
MVO
MVZ
MR
©
C
C
C
C
C
C
C
A,S,
A,S,
A,S,
A,S
P,A
P,A
A,S
A,S
A,S
A
A
A,S,D, IK
P,A,S
M, A,S
S,
IK
A,S,
IK
C
C
C
C
A,S
P,A
P,A
A,S,
CM
A
A,S
C
C
A,S
A,S
C
C
LM, A,S
P,A
P,A
P,A
P,A
P,A
S
1D
5D
17
57
97
D7
EX 44
9E
43
18
58
41
12
IF 13
48
98
11
IF 10
82
92
D2
Dl
Fl
D3
lC
Multiply
Multiply Halfword
OR
OR
OR
OR
Pack
Set Program Mask
Set System Mask
Shift Left Double
Shift Left Single
Shift Left Double
Logical
Shift Left Single
Logical
Shift Right Double
Shift Right Single
Shift Right Double
Logical
Shift Right Single
Logical
Start 110
Store
Store Character
Store Halfword
Store Multiple
Subtract
Subtract
Subtract Halfword
Subtract Logical
Subtract Logical
Supervisor Call
Test Channel
TestI!O
Test Under Mask
Translate
Translate and Test
Unpack
OC
PACK
SPM
SSM
SLDA
SLA
RX
RX
RR
RX
SI
SS
SS
RR
SI
RS
RS
SLDL
RS
SLL
SRDA
SRA
RS
RS
RS
SRDL
RS
SRL
SIO
ST
STC
STH
STM
SR
S
SH
SLR
SL
SVC
TCH
TIO
TM
TR
TRT
UNPK
RS
SI
RX
RX
RX
RS
RR
RX
RX
RR
RX
RR
SI
SI
SI
SS
SS
SS
M
MH
OR
0
or
A,S
A,S
C
C
C
C
A,S
P,A
P,A
P,A
L
M, A
S,
C
C
C
C
IF
IF
S
80
S
89
8E
8A
S
8C
CM
P,A,S
P,A
P,A,S
P,A,S
C
C
C
C
C
CM
CM
C
C
5C
4C
16
56
96
D6
F2
04
80
8F
8B
A,S,
A,S,
IF
IF
IF
A,S
A
P,A
A
P,A
88
9C
50
42
40
90
IB
5B
4B
IF
5F
OA
9F
9D
91
DC
00
F3
Floating-Point Feature Instructions
NAME
Add Normalized
(Long)
Add Normalized
(Long)
Add Normalized
(Short)
Add Normalized
(Short)
AddUnnormalized (Long)
Add Unnormalized (Long)
Add Unnormalized (Short)
Add Unnormalized (Short)
Compare (Long)
MNEMONIC
TYPE
EXCEPTIONS
CODE
NADR
RRF,C
S,U,E,LS
2A
NAD
RXF,C
A,S,U,E,LS
6A
NAER
RRF,C
S,U,E,LS
3A
NAE
RXF,C
A,S,U,E,LS
7A
AWR
RRF,C
S, E,LS
2E
AW
RXF,C
A,S,
E,LS
6E
AUR
RRF,C
S,
E,LS
3E
AU
CDR
RXF,C
RRF,C
A,S,
S
E,LS
7E
29
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
420: 121.1 01
IBM SYSTEM/360
NAME
Compare (Long)
Compare (Short)
Compare (Short)
Divide (Long)
Divide (Long)
Divide (Short)
Divide (Short)
Halve Long
Halve (Short)
Load and Test
(Long)
Load and Test
(Short)
Load Complement
(Long)
Load Complement
(Short)
Load (Long)
Load (Long)
Load Negative
(Long)
Load Negative
(Short)
Load Positive
(Long)
Load Positive
(Short)
Load (Short)
Load (Short)
Multiply (Long)
Multiply (Long)
Multiply (Short)
Multiply (Short)
Store (Long)
Store (Short)
Subtract Normalized (Long)
Subtract Normalized (Long)
Subtract Norma\ized (Short)
Subtract Normalized (Short)
MNEMONIC
TYPE
EXCEPTIONS
CODE
RXF,C
RRF,C
RXF,C
RRF
RXF
RRF
RXF
RRF
RRF
A,S
LTDR
RRF,C
S
22
LTER
RRF,C
S
32
LCDR
RRF,C
S
23
LCER
LDR
LD
RRF,C
RRF
RXF
S
S
A,S
33
28
68
LNDR
RRF,C
S
21
LNER
RRF,C
S
31
LPDR
RRF,C
S
20
LPER
LER
LE
NMDR
NMD
NMER
NME
STD
STE
RRF,C
RRF
RXF
RRF
RXF
RRF
RXF
RXF
RXF
S
S
A,S
S,U,E
A,S,U,E
S,U,E
A,S,U,E
P,A,S
P,A,S
30
38
78
2C
6C
3C
7C
60
70
NSDR
RRF,C
S,U,E,LS
2B
NSD
RXF,C
A,S,U,E,LS
6B
NSER
RRF,C
S,U,E,LS
3B
NSE
RXF,C
A,S,U,E,LS
7B
CD
CER
CE
NDDR
NDD
NDER
NDE
HDR
HER
S
A,S
S,U,E,FK
A,S,U,E,FK
S,U,E,FK
A,S,U,E,FK
S
S
69
39
79
2D
6D
3D
7D
24
34
Subtract Unnormt.lized (Long)
Subtract Unnormalized (Long)
Subtract Unnormalized (Short)
Subtract Unnormalized (Short)
SWR
RRF,C
S,
E,LS
2F
SW
RXF,e
A,S,
E,LS
6F
SUR
RRF,e
S,
E,LS
3F
SU
RXF,C
A,S,
E,LS
7F
The scientific instruction set includes the instructions of both
the standard instruction set and the lIoating-point feature.
Decimal Feature Instructions
NAME
Add Decimal
Compare Decimal
Divide Decimal
Edit
Edit and Mark
Multiply Decimal
Subtract Decimal
Zero and Add
AP
CP
DP
ED
EDMK
MP
SP
ZAP
EXCEPTIONS
TYPE
MNEMONIC
SS
SS
SS
SS
SS
SS
SS
SS
T,e
T,C
T
T,C
T,e
T
T,e
T,C
Commercial Instruction Set
The commercial instruction set includes the instructions of both
the standard instruction set and the decimal feature.
Universal Instruction Set
The universal instruction set includes the instructions of the
standard instruction set, the lIoating-point feature, and the
decimal feature.
Direct Control Feature Instructions
NAME
Read Direct
Write Direct
TYPE
MNEMONIC
RDD
WRD
EXCEPTIONS
SI Y M,P,A
SI Y M, A
C
D
DF
DK
E
EX
F
FK
IF
IK
L
LS
M
N
P
S
T
U
Y
Z
Insert Storage Key
Set Storage Key
ISK
SSK
RRZ
RRZ
M, A,S
M, A,S
Addressing I'xception
Condition code is set
Data exception
Decimal-overllow exception
Decimal-divide exception
Exponent-overllow exception
Execute exception
Floating-point feature
Floating-point divide exception
Fixed-point ovcrllow exception
Fixed-point divide exception
New condition code loaded
Significance exception
Privileged-operation exception
Normalized operation
Protection exception
Specification exception
Decimal feature
Exponent-underflow exception
Direct control feature
Protection feature
Reproduced from IBM System/360 Principles of Operation, Appendix G.
7/65
CODE
85
84
Protection Feature Instructions
Meaning of symbols in TYPE and EXCEPTIONS columns
A
CODE
P,A, D, DF FA
F9
A, D
P,A,S,D, DK FD
DE
P,A, D
DF
P,A, D
FC
P,A,S,D
P,A, D, DF FB
P,A, D, DF F8
09
08
420: 131.1 DO
IBM System/360
Compatibility With
IBM 1401/1440/1460
COMPATIBILITY WITH IBM 1401/1440/1460
.1
GENERAL
System/360 Models 30 and 40 can, under certain
circumstances, execute machine-code programs
written for IBM 1401 and 1460 computers. Programs written for IBM 1440 computer systems
can be similarly run on a properly-equipped System/360 Model 30, but not on a Model 40. A
number of configuration requirements must be
met before a user can run IBM 1400 Series programs on his System/360 processor, and these
are summarized in Table I.
The basic purpose of the ability to run programs
written for older IBM systems on a System/360
is simply to allow an installation more time to
reprogram. Reprogramming is always necessary
in moving from any older IBM computer to a
System/360, because the machine code used by
the System/360 is fundamentally different from
that of any of IBM's older systems.
The basic method selected by IBM to allow programs written in "foreign" machine codes to run
on the System/360 Model 30 is different from the
method used on Model 40. In Model 30, the whole
simulation operation is performed by hardware,
with no software contribution whatsoever. This
means that the peripheral units in the new system
must appear to operate exactly like the peripheral
units in the older system, which in turn necessitates the use of speCialized adapters on most of
the new peripheral units. It also means that the
entire main core memory of the Model 30 is
available for simulating the program storage
of the old system, with no additional core storage
area being needed to hold routines that simulate
the old machine codes.
In contrast to the hardware method used in Model
30, a combined hardware/software method is
used in Model 40 systems. Here the operational
differences between the old and new peripheral
units are handled by the use of simple routines
held in core memory, thus obviating the need for
special hardware adapters on the peripheral units,
but requiring the use of some core storage to
hold the routines. This means that a larger
amount of core storage is used in Model 40 systems than was used in the systems being simulated.
I
\.
Although both Model 30 and Model 40 use special
hardware to duplicate the functions of the basic
1400 Series instructions, the overall performance
of the faster System/360 Model 40 computer will
generally be somewhat below the performance
of the slower Model 30 when simulating a 1400
Series system. A Model 30 can execute 1400
Series machine instructions at about 3.5 times
the speed of the 1401, while Model 40 generally
executes them at only about 3 times the speed
©
of the 1401. An exception is the important Move
Characters and Edit instruction, which is handled
by software simulation routines in Model 40 and
takes 20 times as long as in a 1401 computer.
The System/360 peripheral devices are not significantly faster than their 1400 Series counterparts,
but the System/360's greatly increased capabilities for simultaneous operations should lead to
significant increases in overall throughput in
most applications. The basic 1401/1440/1460
Compatibility Feature for Model 30, however,
does not enable it to take advantage of System/
360's capability to overlap magnetic tape reading
and/or writing with computation. An optional
Tape Overlap feature for Model 30 provides for
special buffering and reblocking in a 4, 096-byte
area of core storage; this permits read/compute
or write/compute simultaneity in systems with
one Selector Channel, and read/write/compute
simultaneity in systems with two Selector Channels. In Model 40, the configuration restriction
that all emulated tape units must be connected to
the same Selector Channel effectively prevents
simultaneous reading and writing.
Details of the operational efficiency of 1400 Series
programs when operating on a System/360 Model
30 or 40 are presented in Paragraph 420:131. 7
below, and the techniques used in these emulations
are described in Paragraph 420:131.6
IBM tends to use the term "compatibility" for the
hardware-only approach used in Model 30 and
the term "emulation" for the combination hardware/software approach used in Model 40. From
the user's point of view, the main difference
between the two approaches is that the extra
costs entailed by the compatibility approach are
entirely for special-purpose hardware; when the
last of the user's own programs has been rewritten in System/360 code, these special features
can be discontinued. Thus, the extra rental cost
involved in running programs written for the old
computer can be ended as soon as there are no
more unconverted programs to be executed. By
contrast, the extra cost/> involved in the emulation
technique will end only if and when the additional
main core storage needed for this approach is·
removed from the system. Historically, although
many systems have been purchased with additional
core storage for reasons of this type, this additional core storage has rarely been removed when
it has served its original purpose. Some programs
usually will have been written to use all of the
available core storage, even though there were
rio absolute requirements for using it. In such
cases, the.added cost of emulation will continue
to be paid as long as the System/360 is installed.
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
IBM SYSTEM/360
420: 131.200
TABLE I: CONFIGURATION REQUIREMENTS FOR EMULATION OF IBM 1401/1440/1460 SYSTEMS
System/360 Model:
Model 40
Model 30
Processors that can run 1401/
1440/1460 programs
Any Model 30 system with at least
8K bytes and 1401/1440/1460
Compatibility Features.
Any Model 40 system with at least
16K bytes and 1401/1460
Compatibility Features.
Peripheral units whose
operations can be
duplicated
Card unlts. magnetic tapes.
printers. and Console Inquiry
stations (replaced by System/3GO
equivalents with special adapters).
Card units. magnetic tapes.
printers. and Console Inquiry
stations (replaced by System/3GO
equivalents; no special adapters
required).
1311 Disk Drives are replaced by
2311 Disk Drives (with noncompatible recording mode).
No provision for use of 1311 or
other random access devices.
Core storage requirements during emulation
No extra core storage required
(i.e •• BK-byte Model 30 can
run BK-char 1400 programs;
16K-byte Model 30 can run
16K- char 1400 programs).
BK bytes required for emulator
software (I.e •• an 8K Model 40
cannot emulate any 1400 system;
a IGK Model 40 can emulate only
an 8K 1400; and a 32K Model 40
is required to emulate either
a 12K or a 16K 1400 system).
1400 Series features
emulated by standard
Compatibility Features
Advanced Programming.
Bit Test.
Expanded Print Edit.
High-lDw-Equal Compare.
Multiply/Divide.
Print Storage.
Additional Print Control.
Read Punch Release.
Space Suppression
Sense Switches.
(processing Overlap Is not
provided as such. but all
printer- and 1402 Card Read
Punch operations are
buffered.)
Advanced Programming.
Bit Test.
Expanded Print Edit.
High-lDw- Equal Compare.
Print storage.
Additional Print Control.
Processing Overlap.
Read Punch Release.
Space Suppression.
Sense Switches.
Additional 1400 Series
features that can be
emulated
Column Binary (1401).
Binary Transfer (1460).
51-Column Interchangeable Read
Feed. Punch Feed Read.
Column Binary (1401).
Binary Transfer (1460).
51-Column Interchangeable Read
Feed. Punch Feed Read.
Computers that can be
simulated by a single
System/360 installation
A single system can run both 1401
and 14GO programs; a card deck
resets parts of the simulating
hardware to handle either system.
A single system can run both
1401 and 1460 programs without
restriction.
1440 programs that use a 1442
or 1443 cannot be run on the
same System/3GO as 1401 or
1460 programs.
.1
GENERAL (Contd.)
8/65
The 1401/1460 and 1410/7010
Compatibility Features are
mutually exclusive.
or 1405 Disk Storage, etc. Lists of the nonsimulated 1400 Series peripheral devices and
special features are included in Paragraph
420:131.8.
Although IBM has officially placed limitations on
the scope of the compatibility and emulating
techniques, a user can normally expect to continue to patch and debug his old programs while
running them on the new system. IBM says it
anticipates that only operational, fully-tested
programs will be run this way, but IBM is in fact
supporting the provision of the standard assemblers and debugging aids for 1400 Series programs on the System/360.
IBM is not promising direct System/360 compatibility for all of the existing 1400 Series installations; only basic card, tape, and (on Model 30)
1311 Disk systems can be simulated at the present
time. This means that reprogramming will probably be necessary for any 1400 Series installation
that uses data communications equipment, optical
or magnetic character readers, paper tape, 1301
1440 programs cannot be run by
a Mode14O.
The Compatibility Features for Model 30 have bee...
in use at IBM's Endicott facility since January
.
1965, and they have already been used to run a
variety of 1400 Series programs. First deliveries
of the Compatibility Feature for Model 40 are
scheduled for September 1965.
.2
CONVERSION OF DATA
.21
Punched Card Files
Punched card data is handled exactly the same
way in the System/360 (using EBCDIC coding) as
it was in the older IBM system.
(Contd.)
COMPATIBILITY WITH IBM 1401/1440/1460
.22
Magnetic Tape Files
Either seven-track or nine-track tape can be used
with the 1400 Series Compatibility Features.
Seven-track tape requires use of the Seven-Track
Adapter Unit to perform conversions between the
BCD tape code and the EBCDIC internal code of
the System/360. To use nine-track tape, the
user must first convert his existing tape fiies
from the seven-track mode (used in all 1400
Series systems) to the nine-track mode by transcribing them on a System/360 equipped with tape
units of both the seven-track and nine-track types.
It should be noted that at least one seven-track
tape unit will be needed on the System/360 if the
user requires any tape communications between
his System/360 and any second-generation IBM
equipment.
.23
• 24
420: 131.220
Disc Files
Data presently on 1311 Disk Packs, in either the
load mode or the move mode, must be dumped
(on cards or tape) and then reloaded onto the new
2311 Disk Packs. In the System/360, load-mode
data can be used only by the Emulator and Simulator Programs; it must be translated before it
.can be used by any other System/360 program.
. 52
• 53
.4
The machine language of the System/360 and the
1400 Series systems are totally different, and all
personnel who will be required to use the System/
360 other than via emulation will need extensive
retraining. Operators will require familiarization courses before they start running emulated
programs on the System/360.
.5
OPERATION OF CONVERTED PROGRAMS
.51
Operation of Individual Programs
A simulated program requires the same general
operating procedures on the System/360 as on
the IBM 1400 Series system. Some differences
will be encountered, including the following:
Processing of error conditions will not be
identical.
Utilization of Program Libraries
Utilization of Operating Systems
Preparation of Operating Instructions and
Program Documentation
An Automatic Documentation and Flow-Charting
Program for the IBM 1401 allows an Autocoder
program to be listed and flowcharted automatically. This routine is useful where no other
documentation exists, or where there are doubts
as to whether the existing documentation is up
to date. The routine can also be used for IBM
1460 and 1440 programs, but it cannot be run on
an IBM 1440 system.
CONVERSION OF PROGRAMS
CONVERSION OF PERSONNEL
•
No standard, IBM-supplied operating systems
are presently in use with IBM 1401, 1440, or
1460 computers, so this problem will not normally
arise .
• 54
IBM expects a properly-equipped System/360 to
be able to run most 1400 Series programs with
little or no hand editing. This does not include
time-dependent programs and certain other types
of programs, which are listed in the "Limitations"
section below. It will, however, be possible to
use the normal 1400 Series programming support
packages such as assemblers, debugging aids,
compilers, and other IBM-provided software.
Most application packages will also be able to
operate in the compatibility mode, although IBM
has not committed itself to support all application
packages.
Card positions in the card reader and the
card punch may not be identical upon
termination of a program.
An installation requiring the use of a program
library must prepare its own library facilities,
either by emulating the appropriate 1400 Series
programs or by writing new procedures in System/360 language.
Some of the special symbols on the 1400 Series
printers and console typewriter are not available
on equivalent System/360 models.
.3
•
Programs that have been converted to System/360
coding will normally use completely new operational procedures.
Collating Sequence
Printed Records
Register contents at the various program
halts may not be identical.
A specially-prepared mask can be placed over the
console of the System/360 to help the operator set
sense switches, read indicator lights, etc.
The original 1400 Series collating sequence is
maintained in the System/360 EBCDIC code.
• 25
•
.6
SPECIAL TECHNIQUES
• 61
The Model 30 Compatibility Technique
Use of the Compatibility Features modifies the
internal organization of the System/360 Model 30
by adding a special 1400 Read-Only storage (ROS)
Control and an area of core storage called the
Auxiliary Storage. This area includes residence
for general-purpose, floating-point, and condition
registers, for the Decimal-to-Binary Conversion
Table, the 1400 EBCDIC-to-BCD Table, the Op
Code Table, the Unit Control Words, etc.
Auxiliary Storage cannot be addressed by a
System/360 programmer.
Sequences of the Read-Only Storage words, combined with timing pulses, define data paths in the
Model 30; these sequences - called "microprograms" - have been written so that the function
of each IBM .1401, 1460, or 1440 instruction is
effectively simulated.
Most of the main core storage is used to simulate
the 1400 Series core memory, on a byte-percharacter basis. Because the 1400 Series core
© 1965 AUERBACH Corporotion and AUERBACH Info, Inc.
8/65
420: 131.610
.61
The Model 30 Compatibility Technique (Contd.)
IBM SYSTEM/360
.7
storage capacities are exact multiples of 1,000
while the System/360 capacities are exact multiples
of 1,024, a few additional locations are available
in the System/360; these are used to detect wraparound and for other functional requirements. A
word-mark on a character in core storage is coded
into bit position 1 in the EBCDIC code; this bit
position of each byte is not.otherwise used in the
compatibility mode.
In operation, an instruction held in main core
storage is read by the ROS Control microprogram,
which examines the operation· code and branches
to the microprogram which simulates that specific
instruction. This microprogram, where appropriate, uses the Decimal-to-Binary Conversion
Table to convert the addresses of operands from
decimal mode to binary mode before using them
to control the actual execution of the instruction.
This decimal-to-binary conversion - which takes
approxImately six microseconds - occurs immediately before each operand address is used, and
the addresses in core storage are always held in
decimal form. This means that no complications
will occur when a programmer modifies his instructions during a program. The rest of the
microprogram duplicates the functions of the
appropriate 1400 Series machine instruction and
then fetches the next 1400 Series instruction to be
simulated.
. 62
Table II summarizes the operational efficiency
of IBM 1401 and 1460 programs when executed
in the compatibility mode on System/360 processors. No details are presently available
regarding the performance of IBM 1440 programs
on the System/360 Model 30, but it is anticipated
that the internal processing speeds will be about
3. 5 times as fast as in the 1440.
.8
LIMITATIONS
.81
1400 Series Peripheral Devices Which Cannot
Be Handled in the Compatibility Mode
IBM
IBM
IBM
IBM
IBM
IBM
IBM
IBM
IBM
IBM
IBM
IBM
IBM
IBM
IBM
IBM
IBM
IBM
IBM
IBM
The Model 40 Compatibility Technique
During emulation, certain registers and core
storage locations of the Model 40 are used to
represent the instruction counter, registers,
sense switches, and other facilities of the original
1400 Series computer system. The Model 40's
core storage is divided into two areas; one is used
to duplicate the original computer system's core
storage, on a byte-per-character basis, and the
other holds the Emulator Program. The latter
area presently occupies 8,192 bytes of core
storage.
The simulation of the 1400 Series program is
performed by two different, complementary procedures. Special hardware, which IBM calls the
1401/1460 Compatibility Feature, is used to decode
the instructions and simulate all the basic operations except those related to input-output and the
Move Characters and Edit instruction. The Compatibility Feature analyzes the operation code of
each 1400 Series machine instruction immediately
before its execution; controls the setting of the
Instruction Counter, the A and B storage Address
Registers, and the A Character Register; and
finally either executes the 1400 Series instruction
by direct simulation or transfers control to the
Emulator Program.
When an operation code is encountered that must
be handled by the Emulator Program, the Compatibility Feature transfers control to a specific
location within the Emulator Program. A software
routine simulates the 1400 Series instruction and
then returns control to the Compatibility Feature.
The cycle is repeated for each instruction in the
original 1400 Series program.
OPERATIONAL EFFICIENCY
1009 Data Transmission Unit.
1011 Paper Tape Reader.
1012 Paper Tape Punch.
1026 Transmission Control Unit.
1301 Disk Storage
1231 Optical Mark Page Reader.
1235 Optical Reader.
1404 Printer (when not using continuous forms).
1405 Disk Storage. 1409 Console Auxiliary.
1412 Magnetic Character Reader.
1418 Optical Character Reader.
1419 Magnetic Character Reader.
1428 Alphameric Character Reader.
1447-1050 system and its components.
1448 Transmission Control Unit.
7340 Hypertape Drive •
7641 Hypertape Control.
7710 Data Communication Unit.
7770 Audio Response Unit.
Note: It may be possible to emulate some of
the devices listed above through the use of
equivalent System/360 devices and special routines if there is enough additional core storage
available. IBM is not promising software to
support such configurations, so the responsibility
for .the programming involved (which can be
extremely complicated) will fall upon the user.
.82
1400 Series Features Which Cannot Be Handled
in the Compatibility Mode
Compressed Tape. *
Direct Data Channel.
Numeric Print Control.
Serial I/O Adapter.
. 83
Programming Restrictions and Limitations
The 1401/1440/1460 Compatibility Feature and
the 1401/1460 Emulator permit the execution of
programs written in accordance with IBM principles-of-operation manuals. The user must,
however, consider the following program restrictions and limitations:
•
Programs written for a 1401/1460 system
which depend on the absence of a particular
optional feature may not be properly emulated.
*A "standard" RPQ feature is available which
can handle Compressed Tape.
(Contd.)
8/65
420:131.830
COMPATIBILITY WITH IBM 1401/1440/1460
TABLE II: OPERATIONAL EFFICIENCY IN THE 1401/1460 COMPATIBILITY MODE
~
IBM 1460
IBM 1401
NEW COMPUTER
When available:
Now.
Now.
Processing:
Approx. 3.5 times as fast.
Approx. 1. 8 times as fast.
Input-output operations:
Depends upon the specific units
involved.
Depends upon the specific units
involved.
Average new performance:
Approx. 2. 5 times as fast.
Approx. 1. a times as fast.
Reserved hardware:
(In addition to simulated core
memory, at one byte/character)
None.
None.
Rental cost of special
features, reserved hardware,
etc.
$265 to $460.
$265 to $460.
When available:
Sept. 1965.
Sept. 1965.
Approx. 3.0 times as fast.
Approx. 1. 5 times as fast.
Approx. 20 times slower.
Depends upon the specific units
involved.
Approx. 2.4 times as fast.
Approx. 40 times slower.
Depends upon the specific
units involved.
Approx 1. 2 times as fast.
Reserved hardware:
(In addition to simulated core
memory, at one byte/character)
8K bytes*
8K bytes*
Rental cost of special
features, reserved hardware,
etc.
$500, or $1,400 if additional
core storage is obtained in
order to allow 1401
emulation.
$500, or $1,400 if additional
core storage is ob~ained in
order to allow 1460
emulation.
Performance, as compared with
original system -
IBM System/360
Model 30 with
at least BK
bytes of core
storage
Performance, as compared with
original system -
IBM System/a60
Model 40 with
at least 16K
bytes of core
storage
Processing, except for
Move Characters & Edit
instruction:
Move Characters & Edit
instruction:
Input-output operations:
Average new performance:
*The available Model 40 core storage sizes may, in fact, make it necessary to procure up to 20K additional
bytes beyond the requirements of the 1401/1460 programs.
. 83
Programming Restrictions and Limitations (Contd. )
•
Time-dependent programs may not yield results
identical to those obtained in a 1400 Series
system.
•
The operator cannot perform equivalent 1400
Series console operations, except for altering
the Instruction Counter and setting sense
switches.
•
The emulator does not check for effective
addresses beyond the limits of core storage.
Such an address, instead of causing a machine
stop as in the 1400 Series, produces a memory
wrap-around.
• Programs which contain undetected programming errors may not yield results identical to
those obtained in a 1400 Series system.
•
The System/360 device unit-!addresses "00"
and !IFF" may not be used.
©
Certain differences between the operation of the
System/360 emulators and the 1400 Series systems
will result from differences in the two sets of I/O
devices and different internal speeds. Three
examples of I/O differences are:
o The Read-Release and Punch-Release instructions do not cause an actual card-feed motion
in the IBM 2540 Card Read Punch. After a
System Stop operation, the cards in the hoppers
and stackers of the 2540 may not be in the same
relative positions as in a 1400 Series system.
o In a 1400 Series system, after a card is read,
about 10 milliseconds are available for starting
a Stacker Select operation. With the emulators,
a Stacker Select operation issued at any time
between two Read instructions affects the first
of the two cards.
o The channel-9 and channel-12 indicators,
used with the 1400 Series printer carriagecontrol tape, do not remain set until the next
tape punch is detected. The indicators are
reset at the end of the printer operation after
the one in which they are set.
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
420: 132.100
IBM System/360
Compatibility With
IBM 1410/7010
COMPATIBILITY WITH IBM 1410/7010
.1
GENERAL
System/360 Models 40 and 50 can, under certain
circumstances, execute machine-code programs
written for IBM 1410 and 7010 computer systems.
A number of configuration requirements - notably
those governing the core storage requirements of
the new system - must be met before a user can run
1410/7010 programs on a System/360 processor.
These requirements are summarized in Table I.
The basic purpose of the ability to run programs
written for older IBM systems on a System/360 is to
allow an installation more time to reprogram. Reprogramming is always necessary in moving from
any older IBM computer to a System/360, because
the machine code used by the System/360 is fundamentally different from that of any of IBM's older
systems.
Emulation by means of the hardware/ software approach adopted by mM is most practical where users
are trading up to significantly larger and more powerful systems. In such cases, the additional core
storage required on the new system to run the emulator programs can be economically used when emulation is not in process. These additional core requirements are considerable: in some cases more
than twice as much core storage is needed in the
emulating System/360 as was installed in the original 1410 or 7010 computer system.
The cost of this additional core storage may make
emulation impractical for users who wish to apply
third-generation computer technology to reduce their
computer rentals rather than to handle a vastly increased processing load. An estimate of the cost of
emulation from such a user's point of view can be obtained by comparing his present rental with the rental of an equally powerful modern computer and with
the rental of the least expensive system capable of
emulating his present system. Using a 6-tape IBM
1410 installation as a basis, the resulting figures are
compared below.
Installations whose workload includes a number of
IBM 1401 programs as well as 1410 or 7010 pro-
grams will need to recode either the 1401 programs
or the 1410/7010 programs before transferring to a
System/360. This is necessary because, although
a System/360 Model 40 can emulate either IBM 1401
or 1410/7010 systems, the two compatibility features
are mutually exclusive. Furthermore, the 1401
Compatibility Mode, which is a standard feature in
IBM 1410 and a standard option in IBM 7010 systems,
is not available when a System/360 is used to emulate !I. 1410 or 7010.
In both Model 40 and Model 50, IBM uses a combined
hardware/ software method of emulation. Operational
differences between the old and the new peripheral
units are resolved by the use of routines held in core
memory. This obviates the need for special hardware adapters on the peripheral units themselves,
but adds to the amount of core storage that is required. From the user's point of view, the main
difference between the all-hardware approach to emulation (used in Model 30) and this hardware/software
approach is that it is rarely practical to discontinue
the use of the additional core storage areas when
emulation of 1410/7010 programs is no longer required. This may lead to an installation's using a
more expensive system than it actually needs, and
effectively continuing to pay for the ease of transferring to the System/360 long after reprogramming
has been concluded.
Although IBM has officially placed limitations on the
scope of the emulating technique, a user can normally expect to continue to patch and modify his old programs as necessary during the emulation phase.
IBM disclaims responsibility for the proper operation of any programs other than fully-tested, properly-written, time-independent programs, but in fact
IBM is supporting the provision of the normal complement of assemblers and debugging aids for the
1410/7010 systems on the System/360.
Details of the operational efficiency of 1410/7010
programs when operating on a System/360 Model 40
or 50 are presented in Paragraph 420:132.7, and the
techniques used in these emulations are described
SYSTEM DETAILS*
6-tape, 20K IBM 1410
Present System
Equally Powerful System/360
6-tape, 32K System/360
Model 30
Minimum System That Can
Emulate User's Present
System
6-tape, 65K System/360
Model 40
Additional Cost Due To
Need To Emulate
APPROXIMA TE RENTAL
$12,240
$ 7,000
$10,300
plus $650 for the
Compatibility Feature.
$ 3,950
* Based on Standard Configuration III, as defined in the Users' Guide, Page 4:030.120.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
IBM SYSTEM/360
420: 132.101
TABLE I: CONFIGURATION REQUffiEMENTS FOR EMULATION OF IBM 1410/7010 SYSTEMS
System/360 Model:
.1
Model 50
Model 40
Processors that can run
1410/7010 programs
Any Model 40 system with at least
65K bytes, 1410/7010 Compatibility Feature, Decimal Arithmetic,
and 1052 Adapter.
Any Model 50 system with at least
65K bytes, 1410/7010 Compatibility Feature, and 1052 Adapter.
Peripheral units whose
operations can he
duplicated
Card units, magnetic tapes, printers,
and 1415 Console Printer are replaced by System/360 equivalent
units on a one-for-one basis.
Card units, magnetic tapes, printers, and 1415 Console Printer
are replaced by System/360
equivalent units on a one-furone basis.
1301 Disk Storage units are replaced by 2302 units, sometimes
with fewer units being required.
1301 Disk Storage units are replaced by 2302 units, with a·
single 2302 replacing up to
four 1301's.
Core storage requirements during emulation
25K bytes extra core storage
required (i. e., a 65K Model 40
can only emulate a 40K 1410 or
7010; a 131K Model 40 can
emulate an SOK 1410 or 7010).
25K bytes extra core storage
required (i. e., a 65K Model 50
can only emulate a 40K 1410 or
7010; a 131K Model 50 can
emulate an SOK 1410 or 7010).
1410/7010 features
that can be emulated
by standard
Compatibility Features
Processing Overlap,
Dual Synchronizer Adapter,
Priority.
Processing Overlap,
Dual Synchronizer Adapter,
Priority.
Additional 1410/7010
features that can be
emulated
51-Column Interchangeable Read
Feed.
51-Column Interchangeable Read
Feed.
Computers that can he
emulated by a single
System/360 installation
A single System/360 can run 1410
and 7010 programs; no other
Compatibility Feature can he
installed in the same Processor.
A single System/360 can run 1410
and 7010 programs; no other
Compatibility Feature can he
installed in the same Processor.
cations between his System/360 and any secondgeneration IBM equipment.
GENERAL (Contd)
in Paragraph 420:132. 6. mM is not promising
direct System/360 compatibility for all 1410/7010
installations; lists of the non-emulated equipment
and other restrictions on the use of the emulation
technique are shown in Paragraph 420:132.8.
.23
Data presently on 1301 Disk Storage Units, in either
load mode or move mode, can be simply dumped and
reloaded onto the new 2302 Disk Storage Units. In
the System/360, load-mode data can be used only in
the emulation mode; it must be translated before it
can be used by other System/360 programs.
First customer deliveries of the mM 1410/7010 emulators are scheduled for March 1966 on Model 40
systems and June 1966 on Model 50 systems.
.2
CONVERSION OF DATA
.21
Punched Card Files
Punched card data is handled exactly the same way
in the System/360 (using EBCDIC coding) as it was
in the older IBM system.
. 22
.24
Collating Sequence
The original collating sequence is maintained in the
System/360 EBCDIC code.
• 25
Printed Records
Some of the special characters on the 1415 Console
Typewriter are not available on the 1052 PrinterKeyboard used with System/360.
MagnetiC Tape Files
Either seven-track or nine-track tape can be used
with the 1400 Series Compatibility Features. Seventrack tape requires use of the Seven-Track Adapter
Unit to perform conversions between the BCD tape
code and the EBCDIC internal code of the System/
360. To use nine-track tape, the user must first
convert his existing tape files from the seven-track
mode (used in all 1400 Series systems) to the ninetrack mode by transcribing them on a System/360
equipped with tape units of both the seven-track and
nine-track types. It should be noted that 'at least
one seven-track tape unit will be needed on the
System/360 if the user requires any tape communi-
Disc Files
.3
CONVERSION OF PROGRAMS
IBM expects a properly-equipped System/360 to be
able to run most 1410/7010 programs with little or
no hand editing. Exceptions are time-dependent programs and certain other types of programs which are
listed in the "Restrictions" section below. It will be
possible to utilize the 1410/7010 programming support packages, including assemblers, debugging aids,
and other mM-supplied software. Most application
packages will also be able to operate in the compatibility mode, although IBM has not committed itselfto support all application packages.
(Contd.)
8/65
420: 132.400
COMPATIBILITY WITH IBM 1410/7010
.4
CONVERSION OF PERSONNEL
bytes of core storage, simulates all 1410/7010 inputoutput instructions and the instructions "Move Characters and Edit" and "Move Characters and Suppress
Zeros." All other 1410/7010 instructions are handled by direct hardware simulation.
The machine languages of the System/360 and the
1410/7010 systems are totally different, and all personnel who will be required to use the System/360
other than in the 1410/7010 compatibility mode will
need extensive retraining.
Operators will need special training to run the new
system in the compatibility mode, a special operating mode which is different from the normal operational methods for the System/360. A prior
familiarity with 1410/7010 operations will be very
desirable for operators running emulated programs.
•5
OPERATION OF CONVERTED PROGRAMS
• 51
Operation of Individual Programs
A more detailed description of this type of emulation
is included in Section 420:135, Compatibility with
IBM 7040/7090 Systems.
.62
In general the method used in Model 50 is the same
as the one used in Model 40. There are some detail
differences (e. g., in register allocation), but these
need not concern a programmer .
.7
.52
The required console operations may not be
available, or they may require modified operational procedures.
•
The handling of error conditions may differ.
•
Register contents in the processor and card
positions in on-line readers and punches may
not be identical when a program halts or terminates.
OPERATIONAL EFFICIENCY
Table III summarizes the estimated operational efficiency of IBM 1410 and 7010 programs emulated
by a System/360 Model 40 or Model 50.
An emulated program requires the same general
operating procedures on the System/360 processor
as on the IBM 1410/7010 system. Some differences
will be encountered, including the following:
•
The Model 50 Compatibility Technique
.8
LIMITA TIONS
.81
Peripheral Devices
All of the 1410/7010 peripheral devices whose
operations can be emulated in the 1410/7010 Compatibility mode are listed below. Users should not
assume that any other peripheral device can be
emulated.
IBM
IBM
IBM
IBM
IBM
IBM
IBM
Utilization of Program Libraries
An installation requiring the use of a program
library must prepare its own library facilities,
either by emulating the appropriate programs or by
writing new procedures in System/360 language.
.82
• 53
1410/7010 Features Which Cannot Be Emulated
utilization of Operating Systems
1401-1410 Compatibility Mode
7010 Fourth I/O Channel
Program Relocate and Storage Protect
Interval Timer
Floating-Point
Column Binary
The IBM 1410/7010 Operating System can be used
during emulation on a System/360.
.54
1402 Card Read Punch, Model 2
1442 Card Reader, Model 3
1403 Printer, Models 1, 2, or 3
729 II, IV, V, and VI Magnetic Tape Units
7330 Magnetic Tape Units
1415 Console Printer
1301 Disk Storage, Model 1 or 2.
Preparation of Operating Instructions and Program
Documentation
IBM has stated that, in general, it intends to provide special aids to assist in conversion and documentation of IBM 1400/7000 Series programs. No
specific programs for the IBM 1410/7010 systems
have yet been announced.
.6
SPECIAL TECHNIQUES
· 61
The Model 40 Compatibility Technique
Some special instructions are provided in the Model
40's Read-Only Storage to assist in the emulation.
Using these instructions, the hardware Compatibility Feature analyzes the operation codes of the
1410/7010 instructions which reside in the simulated
1410/7010 storage area in the Model 40's core memory. Then the Compatibility Feature controls the
settings of the simulated 1410/7010 registers and
either directly executes the functions required to
simulate the 1410/7010 instruction or transfers control to the appropriate routine in the Emulator Program. The Emulator Program, which uses 25K
. 83
Programming Restrictions and Limitations
The Compatibility Features permit execution of IBM
1410/7010 object programs with little or no reprogramming, if they have been written according to
IBM principles-of-operation manuals. The user
must, however, consider the following program
restrictions and limitations:
•
The 1410/7010 feature, 1401-1410 Compatibility Mode, cannot be emulated.
•
Time-dependent programs may not yield results
identical to those obtained in a 1410/7010
system.
•
The following functions or instructions are not
available with the Emulator:
a.
The 7010 diagnostic instructions, Branch
on C Bit and Branch on Tape Indicate J (I) K.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
IBM SYSTEM/360
420: 132.830
.83
Programming Restrictions and Limitations (Contd)
b.
g.
Address error detection for effective addresses beyond the limits of the user's
specified system storage.
•
c.
Address error detection of operand addresses that contain illegal special characters,
alphabetics, or word marks.
d.
Instruction checks caused by the absence
of a terminating word mark.
e.
Replacement of invalid characters with
asterisks during input operations.
l.
The underscore of invalid characters on
the printer-keyboard.
1410/7010 OFF NORMAL functions (e. g. ,
Reset and Restart, Inhibit Printout).
Console operations are limited to the following:
Address Set (IAR only)
Display
Alter
Console Read
Console Write
Computer Reset
Program Reset
Start
Stop
Inquiry
Release
Cancel.
TABLE II: OPERATIONAL EFFICIENCY IN THE 1410/7010 COMPATIBILITY MODE
~
IBM 1410
(10K, 20K, 40K, 60K, or SOK
characters)
IBM 7010
(40K, 60K, SOK, or lOOK
characters)
March 1966.
March 1966.
Processing, except for
Move Characters & Edit
and Move Characters &
Suppress Zeros
Twice as fast as 1410
SlighUy better than half the
speed of 7010.
Move Characters & Edit, and
Move Characters & Suppress
Zeros
20 times slower than 1410.
70 times slower than 7010.
Average new performance:
Depends upon the specific
units involved.
At least as fast as 1410.
Depends upon the specific units
involved.
Usually conSiderably slower than
7010, but in some cases may
approach 7010 speeds.
Reserved hardware:
(In addition to simulated core
memory, at one byte/ character)
2SK bytes in 65K Model 40;
51K bytes in 131K & 262K
Model 40.
25K bytes in 65K Model 40;
51K bytes in 131K & 262K
Model 40.
Rental cost of special
features, reserved hardware,
etc.
$1, 600 to $2,000.
$1,600 to $2,000.
When available:
June 1966.
June 1966.
Processing, except for
Move Characters & Edit, and
Move Characters & Suppress
Zeros
2 to 3 times as fast as 1410.
Approaches 7010 speeds.
Move Characters & Edit, and
Move Characters & Suppress
Zeros
14 times slower than 1410.
50 times slower than 7010.
Depends upon the specific units
involved.
Twice as fast as 1410.
Depends upon the specific units
involved.
Usually somewhat slower than
7010, but in some cases will
approach 7010 speeds.
Reserved hardware:
(In addition to simulated core
memory, at one byte/ character)
25K bytes.
25K bytes.
Rental cost of special
features, reserved hardware,
etc.
$1,400 to $1,600.
$1,400 to $1,600.
NEW COMPUTER
When available:
Performance, as compared with
original system -
IBM System/360
Model 40 with
6SK, 131K, or
262K bytes
Input-output operations:
Performance, as compared with
original system -
IBM System/360
Model SO with
at least
65K bytes
Input-output operations:
Average new performance:
(Contd. )
8/65
COMPATIBILITY WITH IBM 1410/7010
.83
420: 132.831
Programming Restrictions and Limitations (Contd.)
•
•
•
Printer channel-9 and channel-12 branch conditions terminate differently depending on whether
the 1410/7010 or the Emulator is used.
The Emulator cannot be loaded from 51-Column
cards; however, 51-column control cards are
permitted.
The Emulator issues a Single Space instruction
to the 1403 Printer before each printout, unless
©
forms motion has occurred before the print instruction.
•
A "data check" or "condition" resulting from a
Write-Line instruction may not inhibit printing.
•
Column binary card processing is not provided.
•
A number of programming limitations must be
observed in the emulation of 1301 Disk Storage
(e. g., no track may contain more than 24 records).
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
420: 133.100
IBM System/360
Compatibility With
IBM 7070/7074
COMPATIBILITY WITH IBM 7070/7074
.1
GENERAL
In Models 50, 65, and 67, IDM uses a combined
System/360 Models 50, 65, and 67 can, under
certain circumstances, execute machine-code programs written for IDM 7070 and 7074 computer
systems. A number of configuration requirementsnotably those governing the core storage requirements of the new system - must be met before a
user can run 7070/7074 programs on a System/360
processor. These requirements are summarized
in Table I.
The basic purpose of the ability to run programs
written for older IDM systems on a System/360
is to allow an installation more time to reprogram.
For efficient equipment utilization, reprogramming
is always necessary in moving from any older mM
computer to a System/360, because the machine
code used by the System/360 is fundamentally
different from that of any of the ffiM's older
systems.
Emulation by means of the hardware/software approach adopted by IDM is most practical where
users are trading up to Significantly larger and
more powerful systems. In such cases, the additional core storage required on the new system to
run the emulator programs can be economically
used when emulation is not in process. These
additional core requirements are considerable: in
some cases more than five times as much core
storage is needed in the emulating System/360
as was installed in the original 7070 or 7074 computer system.
The cost of this additional core storage may make
emulation impractical for users who wish to apply
third-generation computer technology to reduce
their computer rentals rather than to handle a
vastly increased processing load. An estimate of
the cost of emulation from such a user's point of
view can be obtained by comparing his present
rental with the rental of an equally powerful modern
computer and with the rental of the least expensive
system capable of emulating his present system.
Using a 6-tape, 5, OOO-word IDM 7070 installation
as a basis, the resulting figures are compared
below:
hardware/software method of emulation. Operational differences between the old and the new
peripheral units are resolved by the use of routines
held in core memory. This obviates the need for
special hardware adapters on the peripheral units
themselves, but adds to the amount of core storage
that is required. From the user's point of view,
the main difference between the all-hardware
approach to emulation (used in Model 30) and this
hardware/softwaI'e approach is that it is rarely
practical to discontinue the use of the additional
core storage areas when emulation of 7070/7074
programs is no longer required. This may lead
to an installation's using a more expensive system
than it actually needs, and effectively continuing
to pay for the ease of transferring to the System/360
long after reprogramming has been concluded.
Although IDM has officially placed limitations on
the scope of the emulating technique, a user can
normally expect to continue to patch and modify
his old programs as necessary during the emulation
phase. IDM disclaims responsibility for the proper
operation of any programs other than fully-tested,
properly-written, time-independent programs, but
in fact IDM is supporting the provision of the normal complement of assemblers and debugging aids
for the 7070/7074 systems on the System/360.
Details of the operational effiCiency of 7070/7074
programs when operating on a System/360 Model
50 or 65 are presented in Paragraph 420:133.7,
and the techniques used in these emulations are
described in Paragraph 420:133. 6. IBM is not
promising direct System/360 compatibility for all
7070/7074 installations; lists of the non-emulated
equipment and other restrictions on the use of the
emulation technique are shown in Paragraph
420:133.8.
First customer deliveries of the IDM 7070/7074
emulators are scheduled for the second quarter
of 1966.
SYSTEM DETAILS*
APPROXIMATE RENTAL
6-tape, 5, OOO-word IDM 7070
$19,800
Equally Powerful System/360 6-tape, 65K-byte System/360
Model 40
$10,300
Minimum System That Can
Emulate User's Present
System
$21,200
Present System
Additional Cost Due To
Need To Emulate
6-tape, 262K-byte System/360
Model 50
$10,900
*Based on Standard Configuration Ill, as defined in the Users' Guide, Page 4:030.120.
(Contd.)
8/65
420: 133.200
COMPATIBILITY WITH IBM 707017074
TABLE I: CONFIGURATION REQUIREMENTS FOR EMULATION OF IBM 7070/7074 SYSTEMS
System/360 Model:
Model 50
Processors that can run
7070/7074 programs
Model 50 system with at least
262K bytes, 7070/7074
Compatibility Feature, and
1052 Printer-Keyboard.
Model 65 or 67 system with
at least 262K bytes,
7070/7074 Compatibility
Feature, and either 2150
Console or 1052 PrinterKeyboard.
Peripheral units whose
operations can be duplicated
Card units, magnetic tapes,
printers, and console printer
are replaced by System/360
equivalent units on a one-forone basis.
Card units, magnetic tapes,
printers, and console printer
are replaced by System/360
equivalent units on a one-forone basis.
Core storage requirements
during emulation
262K bytes are used to emulate a
10K word (50K character)
7070/7074 system.
262K bytes are used to emulate
a 10K word (50K character)
7070/7074 system.
7070/7074 features that can
be emulated by standard
Compatibility features
Floating Decimal Arithmetic.
Floating Decimal Arithmetic.
Additional 7070/7074 features
that can be emulated
None.
None.
Computers that can be
emulated by a single System/
360 installation
7070/7074 Compatibility Feature
precludes installation of any
other Compatibility Feature on
same processor.
7070/7074 Compatibility Feature
precludes installation of any
other Compatibility Feature on
same processor.
.3
·2
CONVERSION OF DATA
· 21
Punched Card Files
· 22
Punched card data is handled exactly the same way
in the System/360 (using EBCDIC coding) as it
was in the older IBM system.
Magnetic Tape Files
Either seven-track or nine-track tape can be used
with the 7070/7074 Compatibility Features. Seventrack tape requires use of the Seven-Track Adapter
Unit to perform conversions between the BCD tape
code and the EBCDIC internal code of the System/
360. To use nine-track tape, the user must first
convert his existing tape files from the seventrack mode (used in all 7000 Series systems) to
the nine-track mode by transcribing them on a
System/360 equipped with tape units of both the
seven-track and nine-track types. It should be
noted that at least one seven-track tape unit will
be needed on the System/360 if the user requires
any tape communications between his System/360
and any second-generation IBM equipment .
'--
• 23
CONVERSION OF PROGRAMS
IBM expects a properly-equipped System/360 to
be able to run most 7070/7074 programs with
little or no hand editing. Exceptions are timedependent programs and certain other types of
programs which are listed in the "Limitations"
section. It will be possible to utilize the 7070/
7074 programming support packages, including
assemblers, debugging aids, and other IBMsupplied software. Most application packages
will also be able to operate in the compatibility
mode, although IBM has not committed itself to
support all application packages.
.4
CONVERSION OF PERSONNEL
The machine languages of the System/360 and the
7070/7074 systems are totally different, and all
personnel who will be required to use the System/
360 other than in the 7070/7074 compatibility mode
will need extensive retraining.
Disc Files
There are no present plans for emulation of the
Disk Storage Units used in IBM 7070/7074 systems.
• 24
Model 65 or 67
Collating Sequence
The original collating sequence is maintained in
the System/360 EBCDIC code.
Operators will need special training to run the new
system in the compatibility mode, a special operating mode which is different from the normal operational methods for the System/360. A prior
familiarity with 7070/7074 operations will be
very desirable for operators running emulated
programs.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
IBM SYSTEM/360
420: 133.500
.5
OPERATION OF CONVERTED PROGRAMS
.51
Operation of Individual Programs
instruction DIL - Do Interpretive Loop. The
DIL instruction fetches the simulated instruction
counter, converts its contents to the System/360
address of the next instruction to be executed,
fetches this instruction, updates and re-stores
the simulated instruction counter, performs
any indexing that may be required, converts the
resultant operand address to the corresponding
System/360 address, interprets the operation
code, and transfers control to the appropriate
sequence of System/360 instructions that simulates the original instruction's functions. Use
of the micro-instructions in this manner permits
more efficient utilization of the limited amount
of available Read-Only Storage. At present IBM
states that it is not possible to increase the available Read-Only Storage space by adding additional
modules.
An emulated program requires the same general
operating procedures on the System/360 processor
as on the IBM 7070/7074 system. Some differences
will be encountered, including the following:
• The required console operations may not be
available, or they may require modified
operational procedures.
• The handling of error conditions may differ.
• Register contents in the processor and card
positions in on-line readers and punches may
not be identical when a program halts or
terminates.
.52
• Software simulation. Conventional software
routines in main core storage are used to
simulate the older computer's machine instructions, input-output operations, interrupt
facilities, etc. The entire System/360 instruction repertoire, including the special
Compatibility Feature instructions described
in the preceding paragraph, is used in the
simulation process .
Utilization of Program Libraries
An installation requiring the use of a program
library must prepare its own library facilities,
either by emulating the appropriate programs or
by writing new procedures in System/360 language.
. 54
Preparation of Operating Instructions and Program
Documentation·
IBM has stated that, in general, it intends to provide special aids to assist in conversion and
documentation of IBM 1400/7000 Series programs.
No specific programs for IBM 7070/7074 systems
have yet been announced. IBM states that data
conversion programs for the 7070/7074 will be
announced in the first quarter of 1966.
.6
SPECIAL TECHNIQUES
. 61
The General Emulation Technique
In System/360 Models 40, 50,65, and 67, emulation
of IBM's second-generation computers is being
achieved by a combined hardware/software approac
that niakes use of three separate facilities:
• Special hardware. Special-purpose transistor
logic circuits are used where their high speed
is essential; e. g., for overflow detection,
character recognition, and address conversion.
Because of the relatively high cost of such
special-purpose hardware, however, its use
is held to a minimum.
• Special microprogramming. Models 50, 65,
and 67 use microprogram sequences in ReadOnly Storage to define the functions of each
System/360 machine instruction. In addition
to the Read-Only Storage space required to
define all of the standard and optional System/
360 instructions, some additional space is
available. This space can be, and is, used to
define special instructions for use only by the
Compatibility Features.
These special Compatibility Feature instructions
generally are not designed to simulate specific
instructions in the older computer's repertoire,
but rather to assist in the simulation of key
parts of the older computer's instruction execution process. An important example is the key
.62
Specific Techniques for 7070/7074 Emulation
IBM 7070/7074 systems use a 2-out:-of-5 internal
code to represent each decimal digit, which provides
for additional checking on all data transfers. In the
compatibility mode, the System/360 4-bit packed
decimal representation is used, and the checking
functions are handled by the standard parity checking
circuits .
The arrangement of the 262K-byte System/360 core
memory during emulation is shown in Figure 1.
The top half (131K bytes) is used to hold the emulated
7070/7074 core storage, using 8 System/360 bytes
for each emulated 10-digit word. This area is not
absolutely contiguous; some gaps have been left to
facilitate conversions between the. decimal addresses
used in 7070/7074 systems and the binary System/
360 addresses.
The System/360 Floating-Point Registers are used
to hold the simulated 7070/7074 Instruction Counter
and the Indicator Image. AU other indicators are
simulated in the Emulator Program area.
Thirty-five special System/360 instructions are
provided with th(3 7070/7074 Compatibility Feature.
These include the DIL instruction (see .61 above),
which takes 3.8 microseconds, and Field Definition,
Field Store, and Branch If instructions.
262K
EMULATED IBM 7070/7074 CORE STORAGE
(10K WORDS)
1:!lK
EMULATOR PROGRAM, SUBROUTINES, AND
I/O BUFFERS
o
Figure 1: System/3GO storage Map During 7070/7074 Emulation
(Contd.)
8/65
COMPATIBILITY WITH IBM 7070/7074
420: 133.620
TABLE II: OPERATIONAL EFFICIENCY IN THE 7070/7074 COMPATmILITY MODE
~
NEW COMPUTER
mM 7070
IBM 7074
(5,000 or 9,990 10digit words)
(5,000 or 9,990 10digit words)
2nd quarter 1966.
2nd quarter 1966.
Processing:
Approx. 2.0 times
as fast as 7070.
Approx. 0.4 times as
fast as 7074.
Input-output
operations:
Average new
performance:
DependR upon specific units involved.
Approx. 1. 5 times
as fast as 7070.
Depends upon specific
units involved.
Approx. 3 times slower
than 7074.
131K bytes.
131K bytes.
$4,500 (for 5,000word emulations);
$10,700 (for 9,990word emulations) •
$4,500 (for 5, OOO-word
emula,tions) ;
$10,700 (for 9, 990-word
emulations) •
December 1965.
December 1965.
Approx. 5.0 times
as fast as 7070.
Approx. same speed
as 7074.
When available;
Performance, as compared with original
system mM System/360
Model 50
with at least
262K bytes
of core storage
Reserved hardware:
(in addition to simulated
core memory, at 8
bytes per word)
Rental cost of special
features, reserved
hardware, etc.
When available:
Performance, as compared with original
system Processing:
IBM System/360
Model 65 with at
least 262K bytes
of core storage
Input-output operations:
Average new
performance:
Reserved hardware:
(in addition to simulated core memory,
at 8 bytes per word)
Rental cost of special
features, reserved
hardware, etc.
I
Depends upon specific Depends upon specific
units involved.
units involved.
No estimate available. No estimate available.
131K bytes.
131K bytes.
$3,750 (for 5,000word emulations);
13,050 (for 9,990word emulations).
$3,750 (for 5,000word emulations);
13,050 (for 9,990word emulations).
\
©
I 965 AUERBACH Corporation, and AUERBACH Info, Inc.
8/65
IBM SYSTEM/360
420: 133.700
.7
OPERATIONAL EFFICIENCY
.82
Table ill summarizes the estimated operational
efficiency of IBM 7070 and 7074 programs emulated
by a System/360 Model 50 and Model 65.
.8
LIMITATIONS
.81
7070/7074 Peripheral Devices Which Cannot Be
Emulated
Additional core storage beyond 10,000 words.
Interval Timer.
Check disable, address stop, single-cycle storage
test, and unit-record priority controls.
The ability to read and write tape from storage
locations 9990 through 9999.
Halt mode for sign change, field overflow, and
accumulator overflow.
.83
1301 Disk Storage
1302 Disk Storage
7631 File Control
7340 Hypertape Drives
7740 Communication Control
7750 Programmed Transmission Control
1414 I/O Synchronizer, Model 6
7090 Inquiry Stations.
8/65
7070/7074 Features Which Cannot Be Emulated
Programming Restrictions and Limitations
IBM has provided only general indications of the
programming restrictic;ms and limitations that may
be encountered in the 7070/7074 Compatibility mode:
time-dependent programs may not yield identical
results, the handling of certain error conditions will
differ, and diagnostic (customer engineering) operations will not be simulated.
420: 134.100
IBM System/360
Compatibility With
IBM 7080
COMPATIBILITY WITH IBM 7080
.1
In both Model 65 and Model 67, IBM uses a combined
GENERAL
System/360 Models 65 and 67 can, under certain
circumstances, execute machine-code programs
written for IBM 70BO computer systems. A number
of configuration requirements - notably those
governing the core storage requirements of the new
system - must be met before a user can run 70BO
programs on a System/360 processor. These requirements are summarized in Table I..
The basic purpose of the ability to run programs
written for older IBM systems on a System/360 is
to allow an installation more time to reprogram.
For efficient equipment utilization, reprogramming
is always necessary in moving from any older IBM
computer to a System/360, because the machine
code used by the System/360 is fundamentally different from that of any of IBM's older systems.
Emulation by means of the hardware/software approach adopted by IBM is most practical where
users are trading up to significantly larger and
more powerful systems. In such cases, the additional core storage required on the new system to
run the emulator programs can be economically
used when emulation is not in process. These additional core requirements are considerable: in some
cases more than three times as much core storage
is needed in the emulating System/360 as was installed in the original 70BO computer system.
The cost of this additional core storage may make
emulation impractical for users who wish to apply
third-generation computer technology to reduce
their computer rentals rather than to handle a vastly increased processing load. An estimate of the
cost of emulation from such a user's point of view
can be obtained by comparing his present rental
with the rental of an equally powerful modern computer and with the rental of the least expensive system capable of emulating his present system. Using an B-tape, BO, OOO-character IBM 70BO installation as a basis, the resulting figures are compared
below:
hardware/ software method of emulation. Operational
differences between the old and the new peripheral
units are resolved by the use of routines held in core
memory. This obviates the need for special hardware
adapters on the peripheral units themselves, but adds
to the amount of. core storage that is required. From
the user's point of view, the main difference between
the all-hardware approach to emulation (used in Model
30) and this hardware/software approach is that it is
rarely practical to discontinue the use of the additional
core storage areas when emulation of IBM 70BO programs is no longer required. This may lead to an
installation's using a more expensive system than it
actually needs, and effectively continuing to pay for
the ease of transferring to the System/360 long after
reprogramming has been concluded.
Although IBM has officially placed limitations on the
scope of the emulating technique, a user can normally
expect to continue to patch and modify his old programs as necessary during the emulation phase. IBM
disclaims responsibility for the proper operation of
any programs other than fully-tes·ted., properlywritten, time-independent programs, but in fact
IBM is supporting the provision of the normal complement of assemblers and debugging aids for the
70BO system on the System/360.
Details of the operational efficiency of IBM 70BO
programs when operating on a System/360 Model 65
are presented in Paragraph 420:134.7, and the techniques used in these emulations are described in
Paragraph 420:134.6. IBM is not promising direct
System/360 compatibility for all 70BO installations·
lists of the non-emulated equipment and other re- '
strictions on the use of the emulation technique are
shown in Paragraph 420:134. B.
First customer deliveries of the IBM 70BO emulator
are scheduled for the second quarter of 1966.
SYSTEM D ETAILS*
APPROXIMA TE RENTAL
Present System
B-tape, BOK IBM 70BO
$ 45, BOO
Equally Powerful System/360
8-tape, 131K System/360
Model 40
$14,000
Minimum System That Can
Emulate User's Present
System
B-tape, 262K System/360
Model 65
$ 30, 700
Additional Cost Due To
Need To Emulate
$16,700
* Based on Standard Configuration VIIB, as defined in the Users' Guide, Page 4:030.120.
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
B/65
IBM SYSTEM/360
420: 134.200
TABLE I: CONFIGURATION REQUffiEMENTS FOR EMULA TION OF IBM 7080 PROGRAMS
Model 65 or 67
System/360 Model:
Processors that can
run 7080 programs
Model 65 or 67 system with at least 262K
bytes, 7080 Compatibility Feature,
and either 2150 Console or 1052
Printer-Keyboard.
Peripheral units whose
operations can be
duplicated
Card units, magnetic tapes, printers,
and console printer are replaced by
System/360 equivalent units on a
one-for-one basis.
Core storage requirements during emulation
262K bytes are required to emulate
an 80, OOO-character 7080; 524K
bytes to emulate a 160K 7080.
7080 features that can be
emulated by standard
Compatibility Features
None.
Additional 7080 features
that can be emulated
None.
Computers that can be
emulated by a single
System/360 installation
7080 Compatibility Feature precludes
installation of any other Compatibility
Feature on same Processor. (IBM
705 programs can be run because the
705 mode of the 7080 is emulated.)
.2
CONVERSION OF D A T A . 3
CONVERSION OF PROGRAMS
.21
Punched Card Files
IBM expects a properly-equipped System/360 to be
able to run most 7080 programs with little or no
hand editing. Exceptions are time-dependent programs and certain other types of programs which
are listed in the "Limitations" section. It
will be possible to utilize the 7080 programming
support packages, including assemblers, debugging
aids, and other IBM-supplied software. Most application packages will also be able to operate in the
compatibility mode, although IBM l:J.as not committed
itself to support all application packages.
Punched card data is handled exactly the same way
in the System/360 (using EBCDIC coding), as it was
in the older IBM system.
· 22
Magnetic Tape Files
Either seven-track or nine-track tape can be used
with the 7080 Compatibility Feature. Seven-track
tape requires use of the Seven-Track Adapter Unit
to perform conversions between the BCD tape code
and the EBCDIC internal code of the System/360.
To use nine-track tape, the user must first convert
his existing tape files from the seven-track mode
(used in all 7000 S!,!ries systems) to the nine-track
mode by transcribing them on a System/360 equipped
with tape units of both the seven-track and ninetrack types. It should be noted that at least one
seven-track tape unit will be needed on the System/
360 if the user requires any tape communications
between his System/360 and any second-generation
IBM equipment.
· 23
· 24
.4
CONVERSION OF PERSONNEL
The machine languages of the System/360 and the
7080 system are totally different, and all personnel
who will be required to use the System/360 other
than in the 7080 compatibility mode will need extensive retraining.
Operators will need special training to run the new
system in the compatibility mode, a special operating mode which is different from the normal operational methods for the System/360. A prior familiarity with 7080 operations will be very desirable
for operators running emulated programs.
Disc Files
There are no present plans for emulation of the
Disk Storage Units used in IBM 7080 systems.
.5
OPERATION OF CONVERTED PROGRAMS
Collating Sequence
.51
Operation of Individual Programs
The original collating sequence is maintained in the
System/360 EBCDIC code.
An emulated program requires the same general
operating procedures on the System/360 processor
(Contd.)
8/65
420: 134.510
COMPATIBILITY WITH IBM 7080
.51
instruction fetches the simulated instruction
counter, converts its contents to the System/360
address of the next instruction to be executed,
fetches this instruction, updates and re-stores
the simulated instruction counter, performs any
indexing that may be required, converts the resultant operand address to the corresponding
System/360 address, interprets the operation
code, and transfers control to the appropriate
sequence of System/360 instructions that simulates the original instruction's functions. Use of
the micro-instructions in this manner permits
more efficient utilization of the limited amount of
available Read-Only Storage. At present IBM
states that it is not possible to increase the available Read-Only Storage space by adding additional
modules.
Operation of Individual Programs (Contll..)
as on the IBM 7080 system. Some differences will
be encountered, including the following:
,52
•
The required console operations may not be
available, or they may require modified operational procedures.
•
The handling of error conditions may differ.
•
Register contents in the processor and card
positions in on-line readers and punches may
not be identical when a program halts or terminates.
Utilization of Program Libraries
An installation requiring the use of a program
library must prepare its own library facilities,
either by emulating the appropriate programs or
by writing new procedures in System/360 language.
· 53
Utilization of Operating Systems
The Supervisory Control System and Input/Output
Control System currently in use on IBM 7080 systems can be transferred to and emulated by the
System/360.
· 54
u Software simulation: Conventional software routines in main core storage are used to simulate
the older computer's machine instructions, inputoutput operation, interrupt facilities, etc. The
entire System/360 instruction repertoire, including the special Compatibility Feature instructions
described in the preceding paragraph, is used in
the simulation process.
. 62
Preparation of Operating Instructions and Program
Documentation
IBM has stated that, in general, it intends to provide special aids to assist in conversion and documentation of IBM 1400/7000 Series programs. A
data conversion package for the 7080 is scheduled
for the first quarter of 1966.
.6
SPECIAL TECHNIQUES
· 61
The General Emulation Technique
In System/360 Models 40, 50, 65, and 67, emulation of IBM's second-generation computers is being
achieved by a combined hardware/software approach
that makes use of three separate facilities:
•
Special hardware: Special-purpose transistor
logic circuits are used where their high speed is
essential; e. g., for overflow detection, character
recognition, and address conversion. Because
of the relatively high cost of such special-purpose
hardware, however, its use is held to a minimum.
•
.7
Special microprogramming: Models 65 and 67
use microprogram sequences in Read-Only Storage to define the functions of each System/360
machine instruction. In addition to the ReadOnly Storage space required to define all of the
standard and optional System/360 instructions,
.8
some additional space is available. This space
can be, and is, used to define special instructions
.81
for use only by the Compatibility Features.
These special Compatibility Feature instructions
generally are not designed to simulate specific
instructions in the older computer's repertoire,
but rather to assist in the simulation of key parts
of the older computer's instruction execution
process. An important example is the key instruction DIL - Do Interpretive Loop. The DIL
Specific Techniques for 7080 Emulation
The arrangement of the System/360 core memory
during 7080 emulation is shown in Figure 1. The
160, OOO-character simulated 7080 memory is held
in the center 262K bytes of a 524K-byte System/360
core memory, thus allowing a 262K byte Model 65
to conveniently emulate an 80, OOO-character 7080.
Eight System/360 bytes are used to simulate the
five 6-bit characters of each 7080 word. The words
themselves are packed 80K 7080 characters to 131K
System/360 bytes, with some gaps between consecutive simulated words to ease the problem of address
conversions between the decimal 7080 and binary
System/360 addressing modes.
The System/360 General-Purpose Registers are
used to hold the simulated SPC and SAR registers.
Additional hardware is provided with the Compatibility Feature to handle decimal-to-binary conversions, .incrementing or decrementing of 7080 addresses by 5, and a number of special 7080 instructions which are emulated on a direct, one-for-one
basis because it would be difficult to simulate their
highly specialized functions by means of sequences
of standard System/360 instructions.
OPERA TIONAL EFFICIENCY
Table ITI summarizes the estimated operational
efficiency of IBM 7080 programs emulated by a
System/360 Model 65.
LIMITA TIONS
7080 Peripheral Devices Which Cannot Be
Emulated
1301 Disk Storage
1302 Disk Storage
7631 File Control
7340 Hypertape Drives
7622 Signal Control
7908 Data Channel.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
IBM SYSTEM/360
420: 134.830
524K
524K SYSTEM/360
393K
262K SYSTEM/360
262K ~------------------------~
EMULATED 160K 7080
CORE STORAGE
EMULATED 80K 7080
CORE STORAGE
131K
131K r-------------------------~
EMULATOR ROUTINES
EMULATOR ROUTINES
7080 STORAGE IMAGE
7080 STORAGE IMAGE
I/O BUFFERS
I/O BUFFERS
01.---------------------------1
o
Figure 1: System/360 Storage Maps During 7080 Emulation
TABLE II: OPERATIONAL EFFICIENCY IN TIlE 7080 COMPATIBILITY MODE
~
NEW COMPUTER
.
When available:
IBM 7080
(80K or 160K characters)
March 1966
Performance, as compared with
original system -
IBM System/360
Model 65 or 67
with 262K or
524K bytes of
core storage
Processing:
Approx. as fast as 7080.
Input-output operations:
Depends upon specific units
involved.
Reserved hardware:
128K bytes in addition to
simulated 7080 core memory
at 128K bytes per 80K 7080
characters.
Rental cost of special
features, reserved hardware,
etc.
.82
7080 Features Which Cannot Be Emulated
7080 systems without card or magnetic tape I/O.
The Select 6001 and Select 6002 instructions.
The Set Bit Redundant instruction.
The Control 0005 instruction.
The TMC, TEC, TAR, and SUP instructions.
The RD 03, RD 04, RD 05, WR 03, WR 04, and
WR 05 instructions.
. 83
•
The Automatic Restart Indicator is not turned
on by execution of instructions exceeding the
2-second limit for execution time.
•
The Instruction Check Indicator is not turned on
by the 4/9 check or the Data Register VRC Check.
•
The 1442 Model N1 Card Read Punch treats the
7080 Reader Storage Mark as an illegal character.
"
Time-dependent programs may not yield identical
results .
•
Handling of error conditions will differ.
•
Customer-engineering diagnostic operations will
not be simulated.
Programming Restrictions and Limitations
In addition to the non-emulated instructions listed
above, the following differences in operation should
be considered:
8/65
$3, 975 per month.
420: 135.100
IBM System/360
Compati bi Ii ty Wi th
IBM 704017090 Series
COMPATIBILITY WITH IBM 704017090 SERIES
.1
GENERAL
In Models 65 and 67, IDM uses a combined hardware/
System/360 Models 65 and 67 can, under certain
circumstances, execute machine-code programs
written for IDM 7040/7090 Series computer systems
(the 709, 7040, 7044, 7090, 7094, and 7094 Il).
A number of configuration requirements - notably
those governing the core storage requirements of
the nc;Jw system - must be met before a user can
run 7040/7090 Series programs on a System/360
processor. These requirements are summarized in
Table I.
The basic purpose of the ability to run programs
written for older IDM systems on a System/360
is to allow an installation more time to reprogram.
Reprogramming is always necessary in moving from
any older IDM computer to a System/360, because
the machine code used by the System/360 is fundamentally different from that of any of IDM's older
systems.
Emulation by means of the hardware/software approach adopted by IDM is most practical where
users are trading up to significantly larger and more
powerful systems. In such cases, the additional
core storage required on the new system to run the
emulator programs can be economically used when
emulation is not in process. These additional core
requirements are considerable: in some cases up
to 16 times as much core storage is needed in the
emulating System/360 as was installed in the
original 7040/7090 Series computer system.
The cost of this additional core storage may make
emulation impractical for users who wish to apply
third-generation computer technology to reduce
their computer rentals rather than to handle a
vastly increased processing load. An estimate of
the cost of emulation from such a user's point of
view can be obtained by comparing his present
rental with the rental of an equally powerful modern
computer and with the rental of the least expensive
system capable of emulating his present system.
Using a 8-tape, 32K-word IDM 7090 installation as
a basis, the resulting figures are compared below:
software method of emulation. Operational differences
between the old and the new peripheral wilts are
resolved by the use of routines held in core memory.
This obviates the need for special hardware adapters
on the peripheral units themselves, but adds to the
amount of core storage that is required. From the
user's point of view, the main difference between
the all-hardware approach to emulation (used in
Model 30) and this hardware/software approach is
that it is rarely practical to discontinue the use of
the additional core storage areas when emulation of
7040/7090 Series programs is no longer required.
This may lead to an installation's using a more
expensive system that It actually needs, and in
effect continuing to pay for the ease of transferring
to the System/360 long after reprogramming has been
concluded.
Although IBM has officially placed limitations on
the scope of the emulating technique, a user can
normally expect to continue to patch and modify
his old programs as necessary during the emulation
phase. IDM disclaims responsibility for the proper
operation of any programs 'other than fully-tested,
properly-written, time-independent programs, but
in fact IDM is supporting the provision of the normal complement of assemblers and debugging aids
for the 7040/7090 Series systems on the System/360.
Details of the operational efficiency of 7040/7090
Series programs when operating on a System/360
Model 65 is presented in Paragraph 420:135.7,
and the techniques used in these emulations are
described in Paragraph 420:135. 6. IDM is not
promising direct System/360 compatibility for all
7040/7090 Series installations; lists of the nonemulated equipment and other restrictions on the
use of the emulation technique are shown in
Paragraph 420:136.8.
First customer deliveries of the IDM 7040/7090
Series emulators are scheduled for December
1965.
SYSTEM DETAILS*
APPROXIMA TE RENTAL
Present System
8-tape, 32K-word IDM
7090
$60,700
Equally Powerful System/360
8-tape, 262K-byte
System/360 Model 50
$24,600
Minimum System That Can
Emulate User's Present
System
8-tape, 524K-byte
System/360 Model 65
$42,600
Additional Cost Due To
Need To Emulate
$18,000
* Based on Standard Configuration VIm, as defined in the Users' Guide, Page 4:030.120.
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
IBM SYSTEM/360
420: 135.200
TABLE I: CONFIGURATION REQUIREMENTS FOR EMULATION OF
IBM 7040/7090 SERIES SYSTEMS
System/360 Model:
Model 65 or 67
Processors that can run 7040/7090
Series programs
Model 65 or 67 system with at least
524K bytes, 7040/7090 Compatibility
Feature, and either 2150 Console or
1052 Printer-Keyboard.
Peripheral units whose
. oPerations can be duplicated
Card units, magnetic tapes, printers,
and console printer are replaced by
System/360 equivalent units on a
one-for-one basis.
Core storage requirements during
emulation
524K bytes are required for all 7040/
7090 Series emulations.
Computers that can be emulated by
a single System/360 installation
7040/7090 Compatibility Feature
precludes installation of any other
Compatibility Feature on same
Processor. The 7040/7090 Compatibility Feature can be modified
by control cards to handle IBM 709,
7040, 7044, 7090, 7094, or 7094 II
programs .
.2
CONVERSION OF DATA
· 21
Punched Card Files
7040/7090 Series programming support packages,
including assemblers, debugging aids, and other
IBM-supplied software. Most application packages
will also be able to operate· in the compatibility
mode, although IBM has not committed itself to
support all application packages.
Punched card data is handled exactly the same way
in the System/360 (using EBCDIC coding) as it
was on the older IBM system.
· 22
· 23
Magnetic Tape Files
CONVERSION OF PERSONNEL
The machine]anguages of the System/360 and the
7040/7090 Series systems are totally different,
and all personnel who will be required to use the
System/360 other than iri the 7040/7090 compatibility mode will need extensive retraining.
Either seven-track or nine-track tape can be used
with the 7000 Series Compatibility Features.
Seven-track tape requires use of the Seven-Track
Adapter Unit to perform conversions between the
BCD tape code and the EBCDIC internal code of the
System/360. To use nine-track tape, the user must
first convert his existing tape files from the seventrack mode (used in all 7000 Series systems) to the
nine-track mode by transcribing them on a System/
360 equipped with tape units of both the seven-track
and nine-track types. It should be noted that at
least one seven-track tape unit will be needed on
the System/360 if the user requires any tape
communications between his System/360 and any
second-generation IBM equipment.
.5
OPERATION OF CONVERTED PROGRAMS
Disc Files
. 51
Operation of Individual Programs
There are no present plans for emulation of the
Disk Storage Units used in IBM 7040/7090 Series
systems.
· 24
.4
Collating Sequence
The original collating sequence is maintained in the
System/360 EBCDIC code.
.3
An emulated program requires the same general
operating procedures on the System/360 processor
as on the IBM 7040/7090 Series systems. Some
differences will be encountered, including the
following:
•
The required console operations may not be
available, or they may require modified
operational procedures.
•
The handling' of error conditions may differ.
•
Register contents in the processor and card
positions in on-line readers and punches
may not be identical when a program halts
or terminates.
CONVERSION OF PROGRAMS
IBM expects a properly-equipped System/360 to be
aqle to run most 7040/7090 Series programs with
little or no hand editing. Exceptions are timedependent programs and certain other types of
programs which are listed in the "Limitations"
section. It will be possible to utilize the
8/65
Operators will need special training to run the new
system in the compatibility mode, a special
operating mode which is different from the normal
operational methods for the System/360. A prior
familiarity with 7040/7090 Series operations will
be very desirable for operators running emulated
programs.
./
420: 135.520
COMPATIBILITY WITH IBM 704017090 SERIES
. 52
IBM states that it is not possible to increase
the available Read-Only Storage space by
adding additional modules.
Utilization of Program Libraries
An installation requiring the use of a program
library must prepare its own library facilities,
either by emulating the appropriate programs or
by writing new procedures in System/360 language.
.53
" Software simulation. Conventional software
routines in main core storage are used to
simulate the older computer's machine instructions, input-output operations, interrupt facilities, etc. The entire System/360
instruction repertoire, including the special
Compatibility Feature instructions described
in the preceding paragraph, is used in the
simulation process.
Utilization of Operating Systems
IBM-supplied operating systems can be used during
emulation.
. 54
Preparation of Operating Instructions and Program
Documentation
.62
IBM has stated that, in general, it intends to
provide special aids to assist in conversion and
documentation of IBM 1400/7000 Series programs.
No specific programs for the IBM 7040/7090 systems have yet been announced.
.6
SPECIAL TECHNIQUES
. 61
The General Emulation Technique
Specific Techniques for 7040/7090 Series Emulations
The arrangement of the 524K System/360 core
memory during emulation is shown in Figure 1.
The top half (262K bytes) is used to hold the emulated core storage, using eight System/360 bytes
(64 bits) for each emulated 36-bit word. This
allows the Model 65 or 67 to access a complete
7040/7090 "word" during a single memory cycle .
524K
In System/360 Models 40, 50, 65, and 67, emulation
of IBM's second-generation computers is being
achieved by a combined hardware/software
approach that makes use of three separate
facilities:
•
•
EMULATED IBM 7040/7090
CORE STORAGE
(32KWORDS)
Special hardware. Special-purpose transistor
logic circuits are used where their high speed
is essential; e. g. for overflow detection,
character recognition, and address conversion.
Because of the relatively high cost of such
special-purpose hardware, however, its use
is held to a minimum.
262K
EMULA TION SUBROUTINES
EMULA TION PROGRAM
AND
I/O BUFFERS
Special microprogramming. Models 40, 50, 65,
and 67 use microprogram sequences in ReadOnly Storage to define the functions of each
System/360 machine instruction. In addition
to the Read-Only Storage space required to
define all of the standard and optional System/
360 instructions, some additional space is
available. This space can be, and is, used to
define special instructions for use only by the
Compatibility Features.
These special Compatibility Feature instructions generally are not designed to simulate
specific instructions in the older computer's
repertoire, but rather to assist in the simulation of key parts of the older computer's
instruction execution process. An important
example is the key instruction DIL - Do
Interpretive Loop. The DIL instruction
fetches the simulated instruction counter,
converts its contents to the System/360 address
of the next instruction to be executed, fetches
this instruction, updates and re-stores the
simulated instruction counter, performs any
indexing that may be required, converts the
resultant operand address to the corresponding
System/360 address, interprets the operation
code, and transfers control to the appropriate
sequence of System/360 instructions that simulates the original instruction's functions. Use
of the micro-instructions in this manner permits
more efficient utilization of the limited amount
of available Read-Only Storage. At present
,
245K
o
Figure 1: System/360 Storage Map
During 7040/7090 Series
Emulation
The System/360 General-Purpose Registers are
used to hold the simulated Accumulator, MultiplierQuotient Register, and seven simulated index registers. The 7040/7090 Compatibility Feature adds
a new hardware register to the Model 65 to hold
the simulated Instruction Counter.
The DIL instruction (see .61 above), which initiates
the simulation process for each instruction, takes
0.8 microseconds to execute.
The double-preCision floating-point multiply and
divide instructions may produce different results
on the System/360 than they did on the IBM 7094.
A 27-bit adder was used to produce these results
in the 7094, and this introduced inaccuracies. By
contrast, a 54-bit adder is used in System/360;
the new results are stated by IBM to be more accurate whenever there is any difference.
.7
OPERA TIONAL EFFICIENCY
Table III summarizes the estimated operational
efficiency of IBM 7090 programs emulated by a
System/360 Model 65.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
420: 135.800
IBM SYSTEM/360
TABLE II: OPERATIONAL EFFICIENCY IN THE 7090 COMPATmlLITY MODE
~
NEW COMPUTER
,When available:
mM 7090*
December 1965.
Performance, as compared with
original system ApproX. 2. 0 times as fast as
7090.
Depends upon the specific units
involved.
Processing:
mM System/360
Model 65
with at least
524K bytes
of core storage
Input-output operations:
262K bytes, in addition to
simulated 7090 core memory at
8 bytes per 36-bit word.
Reserved hardware:
Rental cost of special
features, reserved hardware,
etc.
$10,125.
* No details are available to date regarding performance when emulating other 7040/7090
Series systems.
.8
.81
LIMITATIONS
Peripheral Devices Which Cannot Be Emulated
7340 Hypertape Drives
All Disk Storage Units
All Drum Storage Units
All Tele-Processing Equipment
1401 Adapter
Direct Data Adapter
Direct Couple.
• 82
Features Which Cannot Be Emulated
No details are presently available.
.83
Programming Restrictions and Limitations
Few details are yet available. mM states that
time-depend,ent programs may not yield identical
results, that I/O operations dealing with very large
records may have reduced performance, that the
VLM and VDP instructions may not produce the same
results as in the original system for counts greater
than 35 bits, that the SCH ... X instruction will
not detect a channel state in which there is a
partially-completed command, and that the RDC ...
instruction will not halt magnetic tape in the middle
of a record.
/
8/65
420: 136. 100
IBM System/360
Compatibility With
IBM 1620
COMPATIBILITY WITH IBM 1620
.1
GENERAL
A System/360 Model 30, when suitably equipped,
can execute machine-code programs written for
IBM 1620 systems at approximately the speed of a
1620 Model 2. The configuration requirements
for 1620 emulations are summarized in Table 1.
System/360 compatibility with IBM 1620 systems is
programmed by a combination of special microprogrammed routines in the Model 30's Read-Only Storage and conventional simulation routines which occupy
3,500 bytes of main core storage. A 16K-byte
Model 30 can emulate IBM 1620 programs requiring
up to 20,000 storage positions; emulation of a 40K
or 60K 1620 system requires a Model 30 system
, with 32K or 64K bytes of core storage, respectively.
It is important to note that, because no paper tape
punch is currently offered as part of the System/
360 line, it is not possible to duplicate the paper
tape output function of the 1621 Paper Tape Unit -
and paper tape is the principal input-output medium
for many of the existing IBM 1620 installations.
Areas where exact compatibility with IBM 1620
operations is not maintained include the following:
•
In the System/360, all arithmetic operations
are performed without the use of tables;
therefore, if the user has modified the standard
1620 addition or multiplication table, he will
not get the same results in the System/360.
•
The ability to use "storage wraparound" is
limited to I/O operations and the Transmit
Record instruction.
•
Field lengths are limited to 256 IBM 1620
digits.
•
Console operating procedures are different.
First customer deliveries of the 1620 Compatibility
Feature are scheduled for June 1966.
TABLE I: CONFIGURATION REQUIREMENTS FOR EMULATION OF IBM 1620
System/360 Model:
Model 30
Processor that can run 1620 programs
Model 30 systems with at least 16,384
bytes, 1620 Compatibility Feature,
and 1052 Printer-Keyboard.
Peripheral units whose operations can
be duplicated
Console Typewriter (by 1051-1052);
1621 Paper Tape Unit (reader only, by
2671);
1622 Card Read Punch (by 2540; or by
1442, 2501, and/or 2520 in any combination that provides separate paths
for card reading and punching);
1311 Disk Storage Drives (by 2311's on
a one-for-one basis) ;
1443 Printer (by 1443 or 1403).
Core storage requirements during
emulation
16K bytes are required for emulation
of a 20K-digit 1620, 32K bytes for
a 40K 1620, and 64K bytes for a
60K 1620.
1620 features emulated by standard
Compatibility Feature
Additional Instructions,
Automatic Divide,
Index Registers,
Indirect AddreSSing,
Binary Capabilities,
Floating-Point Operations.
Additional 1620 features that can be
emulated
None.
Computers that can be emulated by a
single System/360 installation
1620 Compatibility Feature precludes
installation of any other Compatibility
Feature on same Processor.
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
/
420:141.100
IBM System/360
Data Code Tables
DATA CODE TABLES
The System/360 Processing Units can manipulate and translate any 8-bit character
code, although decimal arithmetic and editing operations assume the use of either the "Extended
BCD Interchange Code" (EBCDIC) or an 8-bit representation of the 7-bit American Standard
Code for Information Interchange (ASCIl-8). The code to be used is designated by the setting
of Bit 12 in the Program Status Word.
IBM states that it offers the user a choice of the EBCDIC or ASCII-8 code. In fact,
IBM is using EBCDIC exclusively in its currently-announced software systems and has no
immediate plans for similar usage of ASCII-B. Data files on tape, disc, and drum that are
organized in either code can be processed and/or sorted without intermediate code translation; such peripheral devices can perform direct data transmissions of any B-bit code. But
most character-set-sensitive System/360 input-output devices (printers, card readers, card
punches, typewriter, etc.) expect data in EBCDIC form in their normal modes of operation.
At present, only the 2260 Display Station uses ASCII-B in data transfers. The card readers
can operate in the binary mode, accepting any punched-card data code, but an internal translation to the EBCDIC or ASCII-B code will then be required in most cases. The card punches
can also operate in the binary mode. Several line printers available with the System/360 can
utilize the Universal Character Set optional feature and can be programmed to print virtually
any character (up to 240 graphic symbols) upon sensing any bit configuration in the data record.
The following tables shmy the graphics that have been defined to date for both codes.
Note that graphics have not yet been defined for many of the 256 bit combinations in each code,
so there is ample room for expansion of the character sets.
TABLE I: AMERICAN STANDARD CODE FOR INFORMATION INTERCHANGE (ASCll)
EXTENDED TO EIGHT BITS
LL-.:::::
Bit Positions---". 76
X5
00-----,
r----Ol----,
r----l0
I I
11-----,
10
11
0
@
P
!
1
A.
DC 2
2
B
R
b
r
DC3
C
5
c
s
DC4
5TOP
*$
3
4
D
T
d
t
WRU
ERR
%
5
E
U
e
u
0110
RU
5YNC
&
0111
BELL
LEM
1000
BKSP
50
1001
HT
51
1010
LF
52
1011
VT
53
1100
FF
54
1101
CR
55
1110
50
56
1111
51
57
4321
00
0000
NULL
DCO
0001
50M
DC l
0010
EOA
0011
EOM
0100
EOT
0101
01
10
11
00
D
blank
,
(
.
.)
+
/
01
10
11
00
01
00
Q
01
10
11
a
q
P
6
F
V
f
V
7
G
W
9
w
B
H
X
h
X
9
I
Y
i
,
Y
J
Z
i
z
;
K
C
k
'"
=
L
\
I
M
J
m
>
N
f
n
E5C
?
0
4-
a
DEL
\
Table reproduced from IBM System/360 Principles of Operation, Appendix F, p. 141.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
IBM SYSTEM/360
420: 141.101
TABLE II: EXTENDED BINARY-CODED DECIMAL INTERCHANGE CODE (EBCDIC)
(Note that both internal and punched card code representations are illustrated on these tables.)
c==c:J
Bit Positions 0, 1
01
~ I
00
r_
~ @ (3
Q;
I~
p
01 110
.~
[;2i
7;P
¥
,@
/
~
Bit Positions 2,3
III
-0
Q:; &. (!,
~:2':
~
~
~
,I~
p,
RE
BY
PN
I~
HT
NL
L'
RS
I~
LC
BS
E.
UC
~
I~
DL
IL
PR
ET
t--.!..
I~
Bit PosltionsO,1
II
10'
ti~ l,i!~"
I.
,',\>:-
:9
~
d
.29,
,~
12
0
:0
9
12
\I
9
11
0
9
12
II
0
ho.o------Zone Punches----~.~I
~~~_~_~-r·~lt~P~~'~'tI~.m~O,~1
~~00:.l~-L~..L~lIy.~lt~P~~;~tI~."=-.2"',=-3
1001
8-1
1010
,
5M
lOti
Q§
!
$
<
1100
.
%
1101
r
,
1110 .
+
;
>
I
~
?
1111
[I:
9"
I.
CD
®
®
@
9
9
II
1121 II
0
Zene Punches
®
12-ow9·8-1
12-11-9-8-1
@
11-0-9-8...1
12-11-0-9-8-1
0
®
-
~
1010
~
"-
1011
1100
rill
8-5
'.
--:-
~
~
8-6
'"
~
:
8-2
#
~
]
@l
-!:i.
~ j
,
-
---,
~:
I 1"(
.
!:L;
~
12
0
12
II
12
H
0
0
• • •
®
®
12-0
II
@
@
0-8-2
11-0
@
@
@
12
12
II,
Il
,0
'II
'0
0
"I.;O------Zcne Punches
No Punches
12
12-11"()
!i..',
\111
1:1
'11
~
:
----~....I
0-1
11-0-9-1
12-11
Control Characters
PF
Punch Off
HT
LC
DL
Horizontal Tob
Lower Case
Delete
RE
Restore
NL
New Line
BS
IL
BY
LF
Backspace
Idle
Bypass
Line Feed
E.
PR
End of Block
Prefix
PN
Punch On
RS
UC
ET
SM
SP
Reader Stop
Upper Case
End of Transmission
Set Mode
Space
>
Greater-than Sign
Question Mark
Colon
Number Sign
At Sign
Prime, Apostrophe
Equal Sign
Quotation Mark
Special Graphic Characters
<
(
+
I
&
I
$
~ent Sign
Period, Decimal Point
Less-than Sign
left Parenthesis
Plus Sign
Vertical Bar, logical OR
Ampersand
Exclamation Point
Dollar Sign
-,
/
%
Asterisk
Right Parenthesis
Semicolon
logical NOT
Minus Sign, Hyphen
Slash
Comma
Percent
Underscore
?
N
@
Table reproduced from mM Form N22-0155, p.2.
8/65
,/
420: 151.100
IBM System/360
Problem Oriented Facilities
Operating System/360
PROBLEM ORIENTED FACILITIES: OPERATING SYSTEM/360
Note: Software for the IBM System/360 is grouped
into three major categories according to the level
of operating system with which it functions: the
Operating System/360, the Basic Operating
System/360, and the Basic Programming Support.
TheProblem Oriented Facilities associated with
the Operating System/360 are described in this
section; the corresponding facilities supplied with
the Basic Operating System/360 and the Basic Programming Support are described in Sections
420:152 and 420:153, respectively. It should also
be noted that the System/360 features and routines
provided for "emulation" of older IBM computer
systems are analyzed in the Compatibility sections
of this report, Sections 420:131 through 420:136.
.1
UTILITY ROUTINES
.11
Simulators of Other Computers
IBM 7090/7094 and 7070/7074SimulationPrograms*
Date available: . . . . . . 1st quarter, 1966.
Description:
These programs will reproduce most of the functions of ffiM 7090/7094 and 7070/7074 Data Processing Systems on a System/360. Simulation of a 10K
7070/7074 requires a Model G (12SK) System/360;
simulation of a 32K 7090/7094 requires a Model I
(512K) System/360. ffiM states that programs may
require more than 10 times as long when run on a
System/360 in the simulation mode as they did when
run on the original machines. Therefore, it is clear
that these simulators are intended only as stop-gap
measures during the reprogramming period, or for
routines that are run infrequently.
IBM 1410/7010 and 70S0 Simulation Programs*
Date available: . . • . . . 2nd quarter, 1966.
Description:
\.
.12
Simulation by Other Computers
IBM 7090/7094 Support Package
Reference: ..•..•.• IBM Publication C2S-6501-2.
Date available: ..•.•. 3rd quarter, 1964.
Description:
The IBM 7090/7094 Support Package for the IBM
System/360 consists of three programs designed to
permit the assembly, testing, and execution on an
IBM 709, 7090, or 7094 of programs written for a
System/360. The 7090 Assembly Program will accept source programs written in System/360 Basic
Assembler Language and will produce object programs in System/360 machine language, ready for
execution. Such programs may be executed either
on an actual System/360 or by further use of the
7090/7094 Support Package. The 7090 Simulator
Program will accept the output of the Assembly
Program, or a System/360 machine-language program prepared by other means, and will simulate its
execution. Input to both 7090/7094 programs is
prepared originally in punched card form and is
converted to magnetic tape format by the 1401 Input
Program, the third of the Support Package programs.
The Support Package is capable of processing most
System/360 programs that are not specifically dependent on input-output timing considerations. The
Assembly Program can accept all of the machine
instruction and pseudo-instruction mnemonic codes
of the Basic Assembler Language, and the Simulator
can process the Standard, Floating-Point, and
Decimal Arithmetic instruction sets of the System/
360, and can handle the Direct Control and Storage
Protection hardware features. It also simulates
most of the input-output operations and interruption
procedures. In addition, it provides facilities to
aid the programmer in detecting and tracing errors
in an object program. However, only 64K bytes
of System/360 object program can be simulated
with the Support Package.
These programs will reproduce most of the functions
of IBM 1410/7010 and 70S0 Data Processing Systems
on a System/360. Simulation of a 40K 1410/7010 or
an SOK 70S0 requires a Model G (12SK) System/360.
A lOOK 1410/7010 or a 160K 70S0 can be simulated
on a Model H (256K) System/360. IBM states that
programs may require more than 10 times as long
when run on a System/360 in the simulation mode as
they did When run on the original machines. Therefore, it is clear that these simulators are intended
only as stop-gap measures during the reprogramming period, or for routines that are run infrequently.
The Assembly and Simulator Programs operate as
one system under a supervisor program called in
by the 7090/7094 Operating System Monitor, IBSYS.
* See. Sections 420:131 through 420:136 for detailed
descriptions of the "emulators" which will permit more efficient System/360 simulation of
these second-generation IBM systems through
the use of optional Compatibility Features.
The 7090/7094 Support Package requires the use of
an IBM 709 with data channel traps, or a 7090 Or
7094 with 32, 76S positions of core storage, a 711
Card Reader, a 716 Printer, one input-output unit
containing the Assembly and Simulator Programs,
The System/360 input-output devices whose operations can be simulated by the 7090/7094 Support
Package include the following: 1052 PrinterKeyboard, 2540 Card Read Punch, 1403 Printer,
1442 Card Read Punch, 144;3 Printer, 2311 Disk
Storage Drive, and the 2401, 2402, 2403, and 2404
Magnetic Tape Units. The tape Read Backward
feature is also simulated.
© 1965 AUERBACH Corporation and AUERBACH Infa, Inc.
8/65
IBM SYSTEM/360
420: 151.120
.12
The minimum machine requirements of the Sort/
Merge program include 12K bytes of core storage,
an input unit, three tape units or one direct access
device for intermediate storage, and one output
device. Any of the follOWing types of devices may
be used for input, output, and/or intermediate
storage: 2400 Series Magnetic Tape Unit, 7340
Hypertape Drive, 1302, 2311, and 2314 Disk Storage Drives, and 2301 Drum Storage Unit. In
addition, a 2540 or 1442 Card Read Punch can be
used to read the input file, and a 1403 Printer to
list messages.
Simulation by other Computers (Contd.)
and 2, 5, or 6 mM 729 Magnetic Tape units depending on whether simulation, assembly, or both are
to be performed. In addition, an IDM 1401 with
4,000 positions of core storage, Advanced Programming, Column Binary, !High-Low-Equal Compare
features, one 729 or 7330 Magnetic Tape Unit, a
1402 Card Read Punch, and a 1403 Printer, is
required.
• 13
Data Sorting and Merging
One input-output channel satisfies the minimum
requirements for the Sort/Merge program, but two
channels will greatly improve the efficiency of
the operation. Increased performance can also be
gained by making large blocks of core storage
available for use by the sorting procedures, arid by
allocating additional units of work storage.
Operating System/360 Sort/Merge
Reference: . . . . . . •
Record size: . . . . . . .
Block size: . . • . . . . .
Key size: . . . . . . . . .
IDM Publication C28-6543-0.
2,400 bytes for 32K systems.
2,400 bytes for 32K systems.
256 bytes, separated into up
to 12 control fields.
File size: . . . . . . . • up to one reel of tape
when three tapes are used
for intermediate storage;
up to 1. 8 tracks of a direct
access device for every
3 tracks allocated for
intermediate storage.
Number of tape/diSC
units: . . . . . . • . . . . a minimum of three tape
units or one direct-access
device.
Date available: . . . . . . December 1965 for 2400
Series tapes and 2311
Disk Storage.
2nd quarter 1966 for 7340
Series tapes, 1302 Disk
Storage, and 2301 Drum
Storage.
Description:
.14
Operating System/360 Report Program Generator
Reference: . . . . . . . mM Publication C24-3337-0.
Date available: . . . . . June 1966.
Description:
The Operating System/360 Report Program Generator (RPG) is an extended and improved version
of the standard mM report writers currently in
use. The power and flexibility of the RPG are
derived principally from its use of all the on-line
facilities of the Operating System/360. The RPG
is more than a report generator: object programs
are, as usual, generated from user-coded specification sheets, but an additional facility is provided
to process the input files immediately following the
compilation of the object program.
The Operating System/360 Sort/Merge program is a
generalized routine designed to perform tape or
disc file sorts and merges under the supervison
of the Operating System/360. The program can
provide up to 16-way merge passes to speed the
sorting operation, and can process the merging
of information from up to 16 input "data sets"
(logically related files of records) .
Interesting additions to the standard RPG features
include: the ability to perform Virtually any type
of arithmetic operation on input data, with conditional branchings depending upon the results of
the calculations; the use of numeric and alphabetic
literals as operands of arithmetic and compare
operations; the provision to read input files from
several devices during the same job; data chaining
operations; table lookup procedures; and the
provision to exit to user-supplied routines that
cannot be conveniently coded in RPG language. All
RPG features and job descriptions are contained on
six user-prepared specification sheets. The sheets
are edited for proper entries and for possible
illogical situations.
Information is supplied to the Sort/Merge program
by the user's control statements. Included in the
user-supplied information is a general description
of the records to be processed, specifications of
the sorting control fields, and any modifications
to be made to the Sort/Merge program prior to or
during its execution. The user's control statements
follow the Operating System/360's Job Control cards
in the input data stream. Job Control provides
the Supervisor with descriptive information concerning the job to be executed.
Features include .provisions for blocked or unblocked
fixed-length and variable-length records, use of
up to 12 control fieldS, sorting and/or merging
in either ascending or descending sequence (which
may be specified separately for each control field),
optional linkage to checkpoint and restart routines,
ability to handle label checking through use of the
Operating System's control functions, and the
capability to add routines to the Sort/Merge program
at any predetermined end-of-phase points.
Report Writing
A single RPG program is available with the Operating System/360. The minimum machine requirements for its use include 12K bytes of core storage
for the generator itself, one input device, four
magnetic tape units or one direct access device for
intermediate storage, and one output device. The
Decimal Arithmetic option is required in the
System/360 Processing Unit.
.15
Data Transcription
!
Operating System/360 Utility Programs
Reference: ....•.•. mM Publication C28-6519-0.
Date available: .•.•.. 1st quarter 1966.
(Contd. )
8/65
/
420: 151.150
PROBLEM ORIENTED FACILITIES: OPERATING SYSTEM/360
.15
and all are capable of printing the symbolic program names associated with the data that is tested
or displayed.
Data Transcription (Contd.)
Description:
These utility programs are designed to operate
under control of the Operating System/360. They
must be stored on the direct-access device that
serves as the program library for the Operating
System/360, and can be accessed by a series of
control cards or by a call from a user-written
program. The user can specify which portions of
fixed-length input records are to be transferred
and indicate where in the output records the selected
fields are to be written. The user can also specify
the blocking factor and format of output records.
The programs included are:
Card to Printer/Punch
Card to Tape
Card to Direct Access Storage
Tape to Card
Tape to Printer
Tape to Direct Access Storage
Tape to Tape
Tape Compare
Direct Access Storage to Tape
Direct Access Storage to Printer
Clear Direct Access Storage
Dump Direct Access Storage
Restore Direct Access Storage.
The Utility Programs are available at one design
level - 12K bytes of core storage. The input-output
units required are dependent upon the specific
application.
. 16
File Maintenance: . . . . routines and macroinstructions are available
in all of the Operating
System/360's language
translators to provide
for the organization and
maintenance of the tape
and direct-access files.
.17
~
Test Translator (TESTRAN)
Reference: . . . . . . . . IBM Publication C28-6541-0.
Date available:. . .
. DecemQer 1966.
Description:
The Test Translator is a group of diagnostic routines controlled by the Operating System/360 and
generated by macro-instructions in assembly
language. The TESTRAN routines must reside on
the direct-access device that is used as the program library for the Operating System/360.
TESTRAN performs its program diagnostics
during the execution of the object program. All
of the programmer-requested routines can reside
in core storage with the problem program; or
optionally, each routine can be called from its
direct-access device residence only as required.
All TESTRAN routines can print test results in
any data mode (e.g., binary, decimal, alphabetic),
©
Several varieties of dump and trace routines are
provided, and all can be entered and modified by
dynamic conditions during object program execution.
Dumps are provided for system tables and directories, general and floating-point registers, and
specified control sections of main storage. Traces
can be performed of program segment transfers,
subroutine calls, and reference to specific data.
Limits and counters can be set to avoid interminable
tracing of runaway program loops. All test results
are written on magnetic tape, disc, or drum, and
specified sections or classes of results can optionally be selected for printing.
The minimum amount of core storage required for
TESTRAN control is 3,600 bytes.
Conversion Aids
Reference: . . . .
Date available:.
Description:
.. IBM Publication C20-1S12-0.
. . ?
This is a series of routines designed to facilitate
the production and updating of documentation for
current user programs which are to be converted
to run on a System/360. The routines produce
printed documents that reflect the current status of
a user's production model object program compared
with the corresponding original source program.
The routines will flag any operation code the user
specifies. Thus, the listing can disclose object
program patches and selected operation codes that
will require attention during the manual portion of
the conversion process. Several automatic flowcharting programs are also available to produce
printed listings of the current logic design of each
program to be converted, together with any descriptive documentation that was incorporated in the
source program.
Individual conversion packages will be provided for
the IBM 1401, 1410/7010, 1460, 7040/7044, 7070/
7074, 7080, and 7090/7094 systems. The routines
will operate on current IBM equipment, primarily
the 1401.
Mathematical Programming System (MPS)
Reference: . . . . . . . . ?
Date available: . . . . . . ?
Description:
MPS is a general-purpose linear programming
system designed for use in conjunction with the
Operating System/360. MPS consists of a number
of routines - called "procedures" - that reside
on the direct-access device that houses the Operating
System's program library. Procedures are called
into storage as requested by user-supplied control
statements. The procedures included within MPS
include the MARVEL language processor and mathematical routines that can be used on both raw and
refined data to accomplish matrix generation,
solution analysis, and management report writing.
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
/
420: 152.100
IBM System/360
Problem Oriented Facilities
Basic Operating System/360
PROBLEM ORIENTED FACILITIES: BASIC OPERATING SYSTEM/360
Note: Software for the IBM System/360 is classified
into three principal groups, according to the operating system with which it is available and under
whose control it functions. The Problem Oriented
Facilities associated with the Basic Operating
System/360 are described in this section. Similar
routines related to the Operating System/360 and
the Basic Programming Support are described in
report Sections 420:151 and 420:153, respectively.
.1
UTILITY ROUTINES
.11
Simulation of Other Computers: . . • . . . . . . . see Paragraph 420:151.11.
.12
Simulation by Other
Computers: . . . . . . . see Paragraph 420:151.12.
. 13
Data Sorting and
Merging
The Disk Sort/Merge Program is a generalized
routine that provides the facility to sort fixed or
variable length tape or disc records in ascending
or descending sequence. Files on punched cards
can be converted to tape or disc storage during an
initialization program pass. The sort key can
consist of up to 12 control fields with a maximum
total length of 256 bytes. All sorting specifications,
including I/O file descriptions, machine configuration, and optional features, are included in userprepared control statements, which follow the
Basic Operating System's Job Control cards in
the input job stream. The Sort/Merge routines
are retrieved as needed from disc storage residence by the Basic Operating System. The sort
operation is performed in five phases.
The Sort/Merge Program includes such features
as the writing of program status checkpoint records
at specified intervals, thus permitting operatorinitiated program restarts at the beginning of any
sort-merge phase. In addition, the facility is provided to allow exits to user-supplied routines at
various points in the program, and to automatically
recognize and check standard tape and disc file
and volume labels. A possible problem area in the
design of the Sort/Merge Program lies in the fact
that tape input files containing checkpointed records
cannot be processed by this program.
Disk Sort/Merge Program
Reference: . . . . . . . . IBM Publication C24-3321-0.
Record size: . . . . . . . maximum of 640 bytes When
sorting with an SK-byte
processor, 1250 bytes
with a 16K-byte processor.
Block size: . . . . . . . . same maximum lengths as
for records, above.
Key size: . . . . . . . . . up to 12 control fields using
a total of up to 256 bytes.
File size: . . . . . . . . . up to 15,500 SO-character
records for each 2311
Disk Storage Drive available
for use, to a maximum of
6 Drives.
Number of tape/disc
units: . . . . . . . . . . . up to four 2400 Series
Magnetic Tape Units for
input-output, and up to
six 2311 Disk Storage
Drives 'for input-output
and intermediate work
storage for sorting operations.
Date available: ...•.• third quarter 1965 for SK
disc version;
fourth quarter 1965 for 16K
tape version.
Description:
Four Sort/Merge programs have been announced
for use with the Basic Operating System/360:
two tape sorts for systems that have 16K bytes
of core storage, and two disc versions for systems that have SK and 16K bytes of core storage,
respectively. Apart from their configuration
requirements, both tape versions and the 16K
disc version are expected to include features
similar to those of the SK disc version described
in the following paragraphs.
In order to use the Disk Sort/Merge Program, the
following minimum machine requirements must be
met: SK bytes of core storage; one 2311 Disk
Storage Drive; one 1403, 1404, or 1443 Printer;
and one 2540, 1442, 2501, or 2520 Card Reader.
Up to 65K bytes of core storage can be utilized
by the program to provide increased sorting
efficiency and handle larger record sizes.
IBM estimates that the Disk Sort/Merge Program
will be able to sort 10,000 100-character discstored records in 26 minutes, using one 2311 Disk
Storage Drive and a System/360 Model 30 Processing
Unit.
.
.14
Report Writing
Basic Operating System Report Program Generator
Reference: . . . . . . . . IBM Publications C2433S7-1 and C26-3570-0.
Date available: ..•••. third quarter 1965 for SK
Disk BOS;
fourth quarter 1965 for 16K
Disk or Tape BOS.
Description:
Two Report Program Generators have been announced for use with the Basic. Operating System/
360: a disc version at the 4K program design
level, and a disc or tape version at the 10K design
level. Aside from the differences in required
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
IBM SYSTEM/360
420: 152.140
.14
options. By means of the Field Select features,
specified fields within each input record can be
rearranged, deleted, or converted to zoned or
packed decimal data format before being transcribed
to an output file. In addition, input records can be
blocked or deblocked as specified. However, only
fixed-length records can be deblocked or fieldselected.
Report Writing (Contd.)
machine configurations, the higher-level RPG will
include features similar to those of the 4K disc
version described below.
The BOS Report Program Generator offers many
extensions and improvements to the standard mM
report writers currently iri use. The range and
flexibility of its facilities result primarily from its
use of the on-line services of the Basic Operating
System. All RPG features and job descriptions are
contained on five user-prepared specification sheets.
The RPG generates an object program from the
specification sheets; then, if desired, it ean immediately execute the generated program to process
specified input files, producing updated output files
and printed reports.
.
The minimum machine requirements to operate the
Disk Resident Utilities include: SK bytes of core
storage; one 2540, 1442, 2520, or 2501 Card Read
Punch or Reader; and one 2311 Disk Storage Drive.
Other input-output devices that are supported
include the 1403 and 1443 Printers and the 2400
Series Magnetic Tape Units. The printers and the
1052 Printer-Keyboard can also be used to record
logging and error messages provided by the Basic
Operating System/360.
Noteworthy features included within this improved
RPG are:
•
The ability to perform virtually any type of
arithmetic operation on input data, with conditional branching dependent upon the results
of these calculations.
•
The ability to use numeric and alphabetic
literals as operands of arithmetic and compare
operations.
•
The provision to read input files from several
devices during the same job.
•
The facility to perform data record chaining
and table lookup operations.
•
The utility programs provided include the following;
Tape-to-Tape
Tape-to-Disk
Tape-to-Card
Tape-to-Printer
Disk-to-Tape
Disk-to-Disk
Disk-to-Card
Disk-to-Printer
Card-to-Tape
Card-to-Disk
Card-to-Printer and/or Punch
Clear Disk.
Utility Programs for 16K Tape and Disk Resident
Systems
The provision to exit to user-supplied routines
that cannot be conveniently coded in RPG language.
Reference: . . . . . . . . ?
Date available: . . . . . . fourth quarter 1965.
Description:
The minimum machine requirements for use of the
4K-Ievel RPG include: SK bytes of core storage;
one 2540, 2520, or 1442 Card Read Punch (or
2501 Card Reader); one 1443, 1403, or 1404 Printer;
one 2311 Disk Storage Drive; and the Standard
Instruction Set with the Decimal Arithmetic feature.
. 15
An additional group of Utility Programs is available
for use with 16K tape and disk versions of the
Basic Operating System. Facilities that·have been
added to those of the SK Disk Utilities include
file-to-file capabilities for the 2321 Data Cell
Drive with printers, 2400 Series Magnetic Tape
Drives, and the 2311 Disk Storage Units. In
addition, a Tape Compare program will be supplied.
Data Transcription
Utility Programs for SK Disk Resident Systems
Reference: . . . . . . . . mM Publication C24-3409-0.
Date available: . . . . . . third quarter 1965.
Description:
The Utility Programs offered for use with the SK
Disk Resident Basic Operating System are generalized file-to-file routines that are controlled
and modified by Job Control and other user-supplied
specification cards. Eleven file-to-file routines
are available for use, as well as one Clear Disk
routine for the 2311 Disk Storage Drive.
Fixed or variable length tape and disc storage
records, as well as punched card files, can be
processed as input to these routines. Standard
System/360 tape and disc file labels can be recognized as such, checked, and updated if desired.
These programs are more than data transcription
routines; they als 0 provide us eful file maintenance
8/65
.16
File Maintenance: ... file maintenance facilities
are provided in the
utility programs described
in Paragraph 420:152.15,
above.
. 17
Other Facilities.
Autotest
Reference: . . . . . . . . mM Publication C24-337S-0.
Date available: ....'. third quarter 1965 for
Disk Resident Autotest;
fourth quarter 1965 for
Tape Resident Autotest.
Description:
Autotest is a set of diagnostic routines that provides debugging facilities for use during the
execution of object programs. Three versions of
(Contd. )
PROBLEM ORIENTED FACILITIES: BASIC OPERATING SYSTEM/360
.17
Other Facilities (Contd.)
Autotest have been announced by IBM for use with
the Basic Operating System/360: versions for the
16K Disk Resident BaS, for the 16K Tape Resident
BaS, and for the 8K Disk Resident BaS. The
principal difference between these three versions
lies in the source and method of Autotest initialization. Once the object program to be tested has
been loaded into core storage with the Autotext
control routines, the functional operations are
essentially the same for each version.
Autotest provides many useful facilities to assist
in the preparation of a production program.
Included in these facilities is Autopatch - a group
of rou~ines designed to patch object programs at
execution time by exchanging, adding, or deleting
instructions and constants by means of control
cards. Printed listings of the patched areas are
provided to file with the assembly listing. Several
other routines are provided to dump core storage
and/or the registers at specified points during the
execution of the test program. The dumps can also
be contingent upon dynamic conditions that develop
during execution, and upon abnormal job termination.
©
420: 152. 170
Autotest is also capable of generating variablelength tape records to serve as test data for the
program being tested. Card to tape, card to
print, and tape to print utility routines are included
among the Autotest routines.
Any program that has been assembled in relocatable
form by the Basic Operating System Assembler can
use the facilities of Autotest, provided sufficient
core storage is available. Since the relocatable
test program and theAutotest program are combined
into one load module by the Linkage Editor routine
of the BaS just prior to loading, both programs
occupy core storage concurrently. Thus, in order
to test a program, core storage must be allotted
for the following routines: Basic Operating System
Supervisor (4K to 6K bytes), an area to store
patch entries, a table containing all Autotest
requests, the Autotest program itself (1,300
bytes), and the problem program.
The minimum machine requirements' for the use of
Autotest with the Basic Operating System/360
include 8K bytes of core storage, one card reader,
one printer, and one 2311 Disk Storage Drive or
2400 Series Magnetic Tape Drive for residence of
the BaS and its program libraries.
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
g .-
420:153.100
~---------~-~l~~
IBM System/360
Problem Oriented Focillties
Basic Programming Support
PROBLEM ORIENTED F'ACILITIES: BASIC PROGRAMMING SUPPORT
the facility is provided to recognize and check
standard System/360 tape labels, and to permit
exits to user-supplied routines at specified exit
points in the Sort/Merge program.
Note: Problem Oriented Facilities that are
controlled by or associated with the Basic
Programming Support software are described
in this section. The facilities that are supervised by and supplied with the Operating System/
360 and the Basic Operating System/360 are
described in Sections 420:151 and 420:152,
respectively.
·1
UTILITY ROUTINES
· 11
Simulators of Other
Computers: . . . . . . . see Paragraph 420:151.11.
· 12
Simulation by Other
Computers: . . . . . . . see Paragraph 420:151. 12.
· 13
Data Sorting and Merging
The Sort/Merge program is a generalized routine
that is modified by control-card statements.
Records can be sorted in ascending sequence by
control fields, twelve of which are permitted per
record. A single input file of fixed or variable
length records can be sorted, and up to five
input files can be merged. All input-output files
are stored on magnetic tape. The record length
and general sorting efficiency are increased if
more than 8K bytes of core storage are available
for use. Up to 65K bytes of storage can be utilized
when available.
. 14
Report Writing
Tape Sort/Merge Program
Reference: . . . . . . . . C24-3320-2.
Record size: . . . . . . . for 8K systems using
3 tape units: 2200 bytes
on input, 1400 bytes on
output.
for 16K systems using 3
tape units: 7700 bytes
on input, 4100 bytes on
output.
Block size: . . . . . . . . up to 255 records per
block.
Key size: ... . . . . . . up to 12 control fields
using a total of up to
256 bytes.
File size: . . . . . . . . . a single reel of tape at
optimum record blocking.
Number of tapes: .... up to six 2400 Series
Magnetic Tape Units
(7- and/or 9-track)
can be used as work units.
Date available: . . . . . 3rd quarter 1965 for single
and dual I/o channel
versions.
Description:
Two Sort/Merge programs have been announced
for use with the Basic Programming Support
package. Both programs require a minimum
of 8K bytes of core storage, three 2400 Series
Magnetic Tape Units, one card reader, and one
line printer. One Sort/Merge is designed to
make use of one input-output channel, and the
other - for increased effiCiency in I/O data
transfers - is designed to utilize two channels.
The BPS Sort/Merge programs operate independently of any control programs.
The Sort/Merge Program includes such features
as the writing of program status checkpoint records
at specified intervals, thus permitting operatorinitiated program restarts at the beginning of any
pass of the External Sort Phase. In addition,
8/65
Basic Programming Support Report Program
Generator
Reference: . . . . . . . . IBM Publication C24-3374-0.
Date Available: . . . . . 4th quarter 1965 for card
and tape versions.
Description:
The IBM System/360's Basic Programming Support
offers a card version of the Report Program Generator that is capable of performing its reportwriting services independently of any supervisory
control program. A tape-oriented RPG is also
offered for use as a processing program under
control of the Tape Basic Programming Support
software system. Apart 'from differences in
configuration requirements, the features of
both RPG versions are essentially the same.
The BPS Report Program Generator offers many
extensions and improvements to the standard IBM
report writers currently in use. It not only
generates object programs from five userprepared specification sheets, but it can optionally continue the data processing task by
calling in input data, performing the generated
program steps, and producing output in various
specified forms.
Noteworthy features included within this improved
RPG are:
•
The ability to perform virtually any type
of arithmetic operation upon the input data,
with conditional branchings dependent upon
the results of these calculations.
•
The ability to use numeric and alphabetic
literals as operands of arithmetic and
compare operations.
•
The provision to read input files from up to
three devices during the same job.
(Contd. )
420: 153.140
PROBLEM ORIENTED FACILITIES: BASIC PROGRAMMING SUPPORT
.14
The minimum machine requirements to operate
the BPS utility Programs include 8K bytes of
core storage and one 2540, 1442, or 2501 Card
Reader. Other input-output devices that are
supported include the 2520 Card Punch, the 1403
and 1443 Printers, the 2311 Disk Storage Drive,
and the 2400 Series Magnetic Tape Units. The
printers and the 1052 Printer-Keyboard can also
be used to record logging and error messages
provided by the programs.
Report Writing (Contd.)
•
The facility for performing table lookup
operations.
•
The provision to exit to user-supplied
routines that cannot be conveniently coded
in RPG language.
The minimum machine requirements for use of
the Basic Programming Support RPG in card form
are listed below. The tape version of RPG designed for use with the BPS control programs has
the same basic requirements plus the addition of
three 2400 Series Magnetic Tape Units. The
minimum RPG requirements include:
• 15
•
8,192 bytes of core storage.
•
One card reader.
•
One card punch (if object program card
deck is desired).
•
One printer (if source-card speCification
diagnostics are desired).
•
The Decimal Arithmetic instruction set.
The data transcription routines provided with the
BPS utility Programs include the following:
Tape-to-Tape
Tape-to-Disk
Tape-to-Card
Tape-to- Printer
Disk-to-Tape
Disk-to-Disk
Disk-to-Card
Disk-to-Printer
Card-to-Tape
Card-to-Disk
Card-to-Printer and/or Punch.
. 16
File Maintenance: ... file maintenance facilities
are provided in the utility
programs described in
Paragraph 420:153.15,
above.
.17
Other Facilities
Data Transcription
Utility Programs for Basic Programming Support
Reference: . . . . . • . • . IBM Publications
C24-3363-1 and
C24- 3392-1.
Date available: . . . . . . 3rd quarter 1965.
Description:
The utility Programs designed for use with the
Basic Programming Support consist of eleven
generalized file-to-file data transcription routines. Each routine is self-loading and nonrelocatable, operating independently of any
control programs. The job descriptions and
any selected optional features are specified on
free-form user-prepared control cards.
The-specialized Multiple Utility Program provides
the capability to process up to three data transcription operations simultaneously. The routines
that can be performed with this program are
Card to Tape, Tape to Card, and Tape to Printer.
Any combination of these routines, up to a total
of three, can be run simultaneously.
Fixed or variable length tape and disc storage
records, as well as punched card files, can be
processed as input to these routines. Standard
System/360 tape and disc file labels can be
recognized as such, checked, and written if
desired.
These programs are more than data transcription
routines; they also provide useful file maintenance
options. By means of the Field Select features,
specified fields within each input record can be
rearranged, deleted, or converted to zoned or
packed decimal data format before being transcribed to an output file. In addition, input records
can be blocked or deblocked as specified. Only
fixed-length records, however, can be deblocked
or field-selected.
©
Basic Programming Support Dump Programs
Reference: . . . . . . . . IBM Publication C28-6557-0.
Date available: . . . . . . April 1965.
Description:
These program dump routines are supplied as
part of the Basic Programming Support package
for use as independently-controlled programs.
The Single-Phase Dump program can be assembled with the problem program by the BPS's
Relocatable Loader; or it can be assembled
separately and used as an independent program.
At points designated by the problem program,
the dump routine produces a listing of the registers
and/or core storage areas as requested. The
Two-Phase Dump program has these same characteristics, except that the dumped output can additionally be produced on cards or tape. If a problem
program being executed comes to an abnormal endof-job, the Self-Loading Dump program can be
loaded into lower core storage, after which a
complete dump of all main storage and the
registers is produced.
Autotest
Reference: . . . . . . . . C24-3343-1.
Date available: . . . . . 3rd quarter 1965.
Description:
Autotest is a set of diagnostic routines that provides
debugging facilities for use during the execution of
object programs. BRS Autotest can be used with
16K and larger tape systems.
Autotest provides many useful facilities to assist
in the preparation of a production program. Included in these facilities is Autopatch - a group
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65·
420: 153. 170
.17
Other Facilities (Contd.)
of routines designed to patch object programs at
execution time by exchanging, adding, or deleting
instructions and constants by means of control
cards. Printed listings of the patched areas are
provided for filing with the assembly listing.
Several other routines are provided to dump core
storage and/or the registers at specified points
during the execution of the test program. The
operation of the dumps can be contingent upon
dynamic conditions that develop during execution,
and upon abnormal job termination. Autotest is
also capable of generating variable-length tape
records to serve as test data for the program
being tested. Card-to-tape, card-to-print, and
tape-to-print utility routines are included among
the Autotest routines.
Any BPS object program in punched card form
can utilize the testing facilities of Autotest, provided sufficient core storage is available. In
order to test a program using the Autotest routines, core storage must be allotted for the BPS
Supervisor (averaging 2500 bytes), for Autopatch
value entries (about 8 bytes per value card), for
a table of Autotest control cards (about 14 bytes
per entry), for the Autotest control program
(a maximum of 1100 bytes), and for the problem
program. The BPS Autotest programs support
the use of up to 65K bytes of core storage for
program testing.
The minimum machine requirements for the use of
Autotest with the Basic Programming Support
include 16K bytes of core storage, one card
reader, one printer, one 9-track 2400 Series
Magnetic Tape Unit, and the standard instruction
set. The program also supports the use of the
1052 Printer-Keyboard for printed messages to
the operator.
IOCS for Magnetic Character and Optical Readers
. C24-3398-1.
Reference: •..
Date available: . . . . . 4th quarter 1965 for
1412, 1418, 1419,
1428, 1231, and 1285
IOCS;
1st quarter 1966 for 1419
Dual Operation IOCS.
Description:
The 1412/1419 Control Program is supplied by
IBM to control the input-output operations of the
1412 and-1419 Magnetic Character Readers. Provided in symbolic form, these routines can be
8/65
IBM SYSTEM/360
modified if desired, and then inserted into the
user's programs prior to assembly. The user
need only code the pocket-select routine necessary
to route the MICR document to a particular pocket
or stacker. The 1412/1419 Control Program includes coding to handle I/O requests, input buffering, operator communication, interrupt conditions, and device error situations. These
routines can be assembled with independentlyoperating problem programs or with programs
designed to operate in conjunction with the BPS
control routines. A minimum of 16K bytes of
core storage is required for the use of the
1412/1419 I/O program package with the Basic
Programming Support system.
Similar input-output support will also be provided
for the 1418 and 1428 Optical Character Readers
and for the 1231 and 1285 Optical Mark Page
Readers.
Communications Control Application
Reference: . . . .
Date available:.
Description:
?
?
The Communications Control package is a set of
programs written in System/360 Basic Assembler
language and designed to assist in the implementation of a computer-controlled data communications network. The basic functions of message
and network control are supplied with this package,
to which modifications can be made as necessary.
The Communications Control programs perform
functions such as: terminal polling, receipt and
transmission of all inquiries, logging and routing
of all inquiries, line and network status reporting,
queuing of inquiries and messages, and line and
terminal error detection and reporting. Information concerning symbolic names of terminals,
polling characters, and message header formats
must be included in the communications program
by the user.
The minimum machine requirements for using the
Communications Control package include 16K
bytes of core storage, a 1051 Control Unit, a
1052 Printer-Keyboard, a 2841 Storage Control
Unit with at least one 2311 Disk Storage Drive, a
2702 Transmission Control Unit, the Decimal
Arithmetic instruction set, and the Interval Timer.
420: 161.100
IBM System/360
Process Oriented Language
Operating System/360 FORTRAN IV
PROCESS ORIENTED LANGUAGE: OPERATING SYSTEM/360 FORTRAN IV
.1
GENERAL
.11
Identity:.........
.12
Origin:.. • . • • • . . . . .. IBM Corporation.
.13
Reference: ••...••... IBM Publications C28-65153 and C28-6513-0.
• 14
Description
Operating System/360
FORTRAN IV (200K);
Operating System/360
FORTRAN IV (E Level
Subset).
Mathematical processing on the System/360 will be
facilitated by the availability of two programming
languages -FORTRAN and Programming Language/
I (PL/I). FORTRAN is at present the most widely
used mathematical language in America. The
specifications for P L/I have only recently become
available, so whether it will eventually displace
FORTRAN as the mathematical language of the future cannot yet be predicted. Naturally, the relative emphasis placed upon each language by IBM
will play an important part in this matter.
Two FORTRAN compilers are being prepared for
use with the Operating System/360. Both are based
upon the FORTRAN IV language. One, a fairly full
version with most of the facilities of 7090,7094
FORTRAN IV and numerous extensions (e. g., subscripts to 7 levels and random processing control
with direct-access devices), is designed for large
systems which include direct-access storage and
have at least a Model H Processing Unit (262,144
bytes of core storage).
Users of smaller equipment configurations who want
a FORTRAN compiler and the control facilities of
the Operating System/360, but who cannot justify
more than 32K bytes of core storage, can utilize
IBM's FORTRAN IV "E Level Subset" language.
The E Level Subset is designed at a 13K program
level for use in the controlled, direct-access-device
environment of the Operating System/360. The restrictions of this limited FORTRAN IV language as
compared with the full 200K version are listed below.
Two additional FORTRAN IV languages/compilers
are provided with the IBM System/360: a 10K tape
or disc version for use with the Basic Operating
System/360, and a 10K card or tape version for use
with the Basic Programming Support. These versions of FORTRAN IV are described in Sections
420:162 and 420:163, respectively.
200K Version
The full version of Operating System/360 FORTRAN,
called the IBM System/360 FORTRAN IV (200K)
language, closely resembles IBM 7090/7094 FORTRAN IV as described in Section 408: 162. The
principal restrictions on the IBM System/360
FORTRAN IV (200K) language are the lack of fixed
symbolic input-output unit designations and the reduced maximum sizes of constants, as summarized below:
Type of
Constant
IBM
System/360
1-10 digits
Integer
1-7 digits
Real
Double Precision 1-16 digits
IBM 7090/7094
1-11 digits
1-9 digits
1-17 digits
Other restrictions of System/360's full FORTRAN
IV language are listed in Paragraph .142 below.
The most important extension of Operating System!
360's full FORTRAN (relative to 7090/7094 FORTRAN IV) is the ability to read and write FORTRAN
data records on direct-access devices in a random
order that is specified by the programmer. Toward
this end, two additional forms have been added to
the READ and WRITE statements, allowing the programmer to specify by means of an integer expression the relative position within the data set of the
record to be read or written. Also, an executable
FIND statement has been added to permit the retrieval of a relatively-referenced record while
computation proceeds. A new DEFINE FILE specification statement (describing the characteristics
of a data set to be placed on or retrieved from a
direct-access device) is required in order to utilize
these random processing facilities. It should be
noted that standard forms of READ and WRITE can
be used for sequential processing of data on directaccess storage devices.
Another important extension of IBM System/360
FORTRAN is the capacity to use up to seven levels
of subscripting. Other useful extensions are the
implementation of the "T" -specification in FORMAT
statements to indicate the print or punch position of
data simply by column number, an "V'-conversion
to specify logical variables, a "G"-specification for
generalized data formats, and increased possible
scaling of real and double precision constants to
16 63 (or about 1075~ for IBM System/360 FORTRAN,
as compared to 10 3 for 7090/7094 FORTRAN IV.
Other extensions of the Operating System/360's
200K FORTRAN with respect to 7090/7094 FORTRAN
include a "double-precision" concept applied to integer, real, and complex variables and constants.
All variables can be explicitly assigned a length
value other than their standard 4-byte or 8-byte
lengths. Also, by means of an IMPLICIT statement,
the type of each variable can be specified according
to the first letters of its name. Mixed-mode expressions (i. e., those that consist of constants and
variables of various types and lengths) are also
© '965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
420: 161.140
IBM SYSTEM/360
------====================~~~~~~~~~~~~~~~~~-------
.14
Description (Contd)
An IBM-supplied timing example shows the estimated
time requirements to perform a full compilation of a
FOR TRAN IV (200K verSion) main program of 200
statements and ten 100-statement, user-supplied
subprograms on a System/360 sample Configuration
C. ill this example, it is assumed that four arithmetic function library routines are linked into the
compiled program. The estimated time in seconds
to perform this FORTRAN IV compilation, reading
the source statements from disc storage and writing
the compiled program to disc storage, is 28.38
seconds.
permitted. 'The full FORTRAN IV compiler provides a variety of options to control and format input-output data transfers, as well as additional
facilities in the use of subprograms. The principal
FORTRAN language extensions are listed in Paragraph • 143 below.
Full FORTRAN IV for use with the Operating
System/360 requires a minimum machine configuration of 256K bytes of core storage, one directaccess device (disc or drum) for program residence
and intermediate storage, one input device, and one
output device (cards, tape, or disc for I/O). The
Standard illstruction Set with the Floating-Point
Feature is required in both the compiling and object
Processing Units.
.141 Availability:
Language specifications: ••.•••.
E Level Subset
The E Level Subset FORTRAN IV available with the
Operating System/360 is a restricted version of the
full 200K FORTRAN IV. This subset compiler requires the same machine configuration for its operation as the full version, except for the size of core
storage. The E Level Subset can function with a
minimum of 32K bytes of core storage.
Among the features of the Operating System/360's
full FORTRAN IV (200K) that have not been implemented in the E Level Subset are the following
statements:
ASSIGN
BLOCK DATA
Labeled COMMON
COMPLEX
DATA
Assigned GO TO
Logical IF
PRINT
PUNCH
READ b, list
END and ERR parameters in a READ
IMPLICIT
ENTRY
RETURNi
NAME LIST
PAUSE with literal.
Other full-version features not implemented in the
E Level Subset include: more than three dimensions
for array storage allocation, adjustable dimensions,
the ability to read FORMAT statements at execution
time, the "double-precision" concept for constants
and variables (although a DOUBLE-PRECISION type
specification is provided), and G and L FORMAT
codes.
The restriction and extensions of E Level FORTRAN IV with respect to 709017094 FORTRAN IV
are listed in Paragraphs . 144 and . 145 of this
report section.
Compilation Times
IBM has provided us with estimated FOR TRAN IV
compilation times based on three sample System/360
configurations. Table I describes these sample
configurations and indicates the several timing factors that must be summed to arrive at a total time
for a FORTRAN IV compilation performed in conjunction with the Operating System/360.
Compiler: •••••••
April 1964 and March 1965
for 200K FORTRAN;
March 1965 for E Level
Subset.
June 1966 for 200K FORTRAN;
December 1965 for E Level
Subset.
.142 Restrictions of 0l2erating System/360 200K FORTRAN IV Relative to IBM 709077094 FORTRAN IV
31-1
(1) illteger constants can range to 2
as
compared to 235-1 in 7090/7094 FORTRAN
IV; real (floating-point) constants can be
from 1 to 7 digits in length as compared to
a maximum of 9 digits for 7090/7094
FORTRAN IV; and double precision constants can range from 1 to 16 digits as
compared to a maximum of 17 digits for
7090/7094 FORTRAN IV.
(2) ill FORMAT statements: O-type conversions are not available; X-conversion fields
must be separated from other fields by
commas.
(3) Fixed symbolic input-output unit designation is not provided.
(4) The SSWTCH subroutine is not provided.
(5) ill statement functions, a maximum of 15
variables that appear within an expression
can be used as arguments of the function.
.143
(1) The processing of randomly-accessed records on direct-access devices is provided
by means of a FIND and two new forms of
READ and WRITE input-output statements.
(2) Up to seven levels of subscripts are permitted.
(3) In FORMAT state,ments, the T-specification
indicates the print position of the data; Lconversion specifies logical variables; Gspecification indicates a generalized format
for integer, real, complex, or logical forms
of data.
(4) Dumps can be in hexadecimal, logical, double-precision, real, integer, complex, or
literal format.
(Contd. )
8/65
----
PROCESS ORIENTED LANGUAGE: OPERATING SYSTEM/360 FORTRAN IV
420: 161.143
TABLE I: FORTRAN IV COMPILATION TIME FACTORS
Configuration Components
Sample Configura.tions
Configuration A
System/3GO Model 30 with 32K bytes of core
storage, one Multiplexor Channel, one Selector
Channel, one 2540 Ca.rd Read Punch, one 1403
Printer (Model 3). two 2311 Disk Storage Drives,
and the Operating System/3GO's sequential
scheduler. *
Configuration B
System/3GO Model 50 with 65K bytes of core
storage, two Selector Channels, two 2311 Disk
Storage Drives, and the Operating System/3GO's
priority scheduler. **
Configuration C
System/3GO Model 65 with 262K bytes of core
storage, two Selector Charmels. two 2301 Drum
Storage Units, four 2311 Disk Storage Drives,
and the Operating System/360's concurrent job
scheduler. **
Compilation Time in Seconds
Estimated Timing Factors
FORTRAN IV-
E Level,
Configuration A
For compilation initiation
For each compilation
For each user-supplied
subprogram
For each source card
For library initialization
For each library access
For each library routine used
For each byte of library
routine used
For each 2, 04B-byte control
section compiled
For each byte compiled
FORTRAN IV200K.
Configuration B
FORTRAN IV200K.
Configuration C
18.3
23.0
11.7
16.0
2.52
0.95
23.0
0.25
19.1
2.0
0.75
16.0
0.10
13.9
1.4
0.37
0.10
0.013
4.01
0.32
0.088
0.0004
0.00009
0.00002
0.46
0.00084
0.088
0.00018
0.045
0.0001
Source program input is assumed to be entered through the Card Read Punch, and object
program output is assumed to be \Vl"itten on disc storage.
Source-program input and object program output are assumed to be read and
disc storage.
.143 Extensions of Operating System/360 200K FORTRAN IV Relative to IBM 7090/7094 FORTRAN
IV (Contd.)
(5) The magnitude of real and double-precision
constants can range to 16 63 (about 10 75 ) as
compared to 10 38 for 7090/7094 FORTRAN
Literal constants (enclosed within quotation
marks) of up to 255 characters are permitted.
(7) In SUBROUTINE subprograms, exit to any
numbered statement in the calling program
is permitted.
(8) An IMPLICIT statement permits assignment of type-specification to the first letter
of variable names.
(9) A "double-precision" concept is applied to
integer, real, and complex constants and
variables, and is not considered to be a
separate type specification. Variables can
be optionally assigned specific lengths other
than the standard 4- or 8-byte lengths.
to
(13) An integer constant can be included in the
STOP statement.
(14) Integer constants and variables can have
real exponents.
(15)
The END statement is used to define the end
of a source subprogram, as well as main
program.
(16)
The NAME LIST statement can be used with
a READ or WRITE statement to eliminate
the need to enumerate list-entries after
such statements.
N.
(6)
~itten
(17) The parameters END and ERR can be used
with READ statements to indicate the statement number to which transfer is made when
end-of-file and error conditions are encountered.
(18) In the FORMAT statement, the length of Afields is not limited, and literal data can
be included.
(19) A "+" character can be used for carriage
control purposes, indicating "no advance."
(10) Mixed modes are permitted within expressions.
(20) Variables within EQUIVALENCE statements
can have multiple subscripts.
(11)
(21)
The ENTRY statement permits transfer to a
subroutine or function at a point other than
the first statement.
(22)
Eleven additional built-in functions and
mathematical subroutines are provided.
Subscripted variables can be used with
logical operators. Complex logical expressions are permitted.
(12) A literal constant can be used with the
PAUSE statement.
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
420: 161. 144
IBM SYSTEM/360
.144 Restrictions of Operating System/360 FORTRAN IV
(E Level Subset) Relative to IBM 7090/7094 FORTRAN IV
(1)
(2)
31-1
Integer constants can range to 2
as
35
1
compared to 2 - in 7090/7094 FORTRAN
IV; real (floating-point) constants can be
from 1 to 7 digits in length as compared to
a maximum of 9 digits for 7090/7094
FORTRAN IV; and double precision constants can range from 1 to 16 digits as
compared to. a maximum of 17 digits for
7090/7094 FORTRAN IV.
In FORMAT statements: O-type conversions are not available, and X-conversion
fields must be separated from other fields
by commas.
(3)
Fixed symbolic input-output unit designation is not provided.
(4)
The ATAN2 mathematical subroutine is not
provided.
(5)
The SSWTCH subroutine is not provided.
(6)
In statement functions, a maximumQf 15
variables that appear within an expression
can be used as arguments of the function.
(7)
Variables and constants cannot be of type
COMP LEX or LOGICAL; logical and ,tela~
tional expressions are not permitted.
(16)
The following statements are not implemented:
ASSIGN
BLOCK DATA
Assigned GO TO
Logical IF
PRINT
PUNCH
READ Q, list.
.145 Extensions of Operating System/360 FORTRAN IV
(E Level Subset) Relative to IBM 7090/7094
FORTRAN IV
(1)
In FORMAT statements, the T-specification indicates the starting print position
of the data.
(2)
Storage dumps can be printed in hexadeCimal, double-precision, real, or integer formats.
(3)
The magnitude of real and doubleprecision constants can range up to 16 63
(about 1075), as compared to 10 38 for
7090/7094 FORTRAN IV.
(4)
Literal data' ill FORMAT statements can
extend up to 255 characters in length.
(5)
Mixed modes are permitted in arithmetic
expressions and statements.
(6)
Integer constants and variables can have
real exponents.
(8)
FORMA T statements cannot be read at
object time.
(7)
An integer constant can be included in the
STOP statement.
(9)
Only one level of parentheses (in addition to
those enclosing the entire statement) is permitted in FORMAT statements.
(8)
The END statement is used to define the
end of a source subprogram, as well as a
main program.
(10)
Specification statements must precede the
first executable statement of the source program. In addition, all Statement Function
definitions must follow the speCification
statements and precede the executable program.
(9)
In FORMAT statements, the length of Afields is not limited, and literal data can
be included.
(10)
A "+" character can be used for carriage
control purposes, indicating "no advance."
(11)
Variables in EQUIVALENCE statements
can have multiple subscripts.
(11)
Adjustable dimensions are not permitted.
(12)
COMMON block names are not permitted.
(12)
The double-precision float (DFLOA T)
mathematical function has been included.
(13)
Thirty-nine words are reserved for use by
the compiler and cannot be used as names of
variables, arrays, or subprograms.
(13)
(14)
All complex mathematical functions and
common logarithm functions have been
excluded.
(15)
Literals cannot serve as arguments in a
CALL statement.
Special statements can optionally be provided for handling the input-output operations of direct-access storage devices
when data records are to be randomly
accessed. These include a DEFINE FILE
specification statement, new forms of the
READ and WRITE verbs, and a FIND
verb which permits record retrieval to
proceed -concurrently with computation.
8/65
420: 162. 100
IBM System/360
Process Oriented Language
BOS FORTRAN IV
PROCESS ORIENTED LANGUAGE: BASIC OPERATING SYSTEM/360 FORTRAN IV
.1
GENERAL
.11
Identity: ..•••.•..• Basic Operating System/360
FORTRAN IV.
. 12
Origin: •••••.••••. IDM Corporation .
. 13
Reference: . . . . • . . . not published to date .
.14
Description
A single FORTRAN compiler has been announced by
IBM for use in conjunction with the Basic Operating
System/360. The compiler is designed to operate
with the Tape or Disc Resident BOB, in a minimum
environment of 16, 384 bytes of core storage.
FORTRAN IV for the Basic Operating System/360 is
a proper, though very restricted, subset of the
Operating SYlltem/360 full (200K) FORTRAN IV
language, which is described in report Section
420:161. In fact, the BOS FORTRAN IV language
resembles to a high degree the Operating System/
360's E Level Subset FORTRAN language, except
that BOB FORTRAN IV is a slightly less flexible
language than the E Level Subset. Upward compatibility will definitely exist among these three compilers; i. e., any program successfully compiled by
the BOB FORTRAN IV compiler will produce identical results if subsequently compiled by the
Operating System/360's E Level Subset or full
200K FORTRAN compiler.
The language facilities of BOS FORTRAN IV are
identical to those offered with the Operating
System/360's E Level Subset FORTRAN IV, except
for the four restrictions listed below. (Report
Section 420:161 should be referred to for a description of the E Level Subset language.) The capability
to process data records stored on direct-access
devices, in a random order specified by the programmer, is a notable inclusion in the Disk BOB
FORTRAN IV language. The restrictions of the
BOB FORTRAN IV language with respect to the
Operating System/360's E Level Subset FORTRAN
IV include the following:
•
l
Embedded blanks are not permitted within
real and double-precision constants, all
statements, and FORTRAN key words.
•
In the F-specification of FORMAT statements,
,an extra space must be allotted for a digit to
the left of the decimal point.
•
The last statement in every DO loop must be
executable and cannot be STOP, pAUSE, or
RETURN.
•
Forty-one key words are reserved by the
compiler - two more than are reserved by the
E Level Subset compiler.
The BOS FORTRAN IV language can be obtained
in a slightly altered form that excludes the use
of all statements related to direct-access devices.
This version will be available to those users of the
Basic Operating System whose configurations include 2400 Series Magnetic Tape Units but no
direct-access storage' units.
Among the features of the Operating System/360's
full FORTRAN IV (200K) that have not been implemented either in the E Level Subset or in BOS
FORTRAN IV are the following statements:
ASSIGN
BLOCK DATA
Labeled COMMON
COMPLEX
DATA
Assigned GO TO
Logical IF
PRINT
PUNCH
READ b, list
END and ERR parameters in a READ
IMPLICIT
ENTRY
RETURN i
NAMELIST
P A US E with literal.
other full-version features not implemented in the
BOS FORTRAN IV subset include: more than three
dimensions for array storage allocation, adjustable
dimensions, the ability to read FORMAT statements
at execution time, the "double-precision" concept
for constants and variables (although a DOUBLEPRECISION type specification is provided), and G
and L FORMA T codes.
The minimum machine configuration required to
operate the Basic Operating System FORTRAN IV
compiler includes 16,384 bytes of core storage,
one card reader and punch, one printer, three
2400 Series Magnetic Tape Units or one 2311 Disk
Storage Drive for intermediate storage, and the
Standard Instruction Set with the Floating-Point
feature.
No performance figures have been released by
IBM to date relative to the estimated compilation speed or operating efficiency of BOB
FORTRAN IV. (See Section 420:161 for estimated compilation times for the Operating
System/360 FORTRAN compilers.)
In Paragraphs .142 and .143 that follow, BOS
FORTRAN IV is compared with IBM 7090/7094
FORTRAN IV, a language that has gained wide
acceptance for use in the solution of scientific
and engineering data processing problems.
IDM 7090/7094 FORTRAN IV is described in
detail in Section 408:162 of the IDM 7090 report.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
420: 162. 141
• 141 Availability
Language specifications: ?
Compiler: . . . . . . . . . 4th quarter 1965.
31-1
(1) Integer constants can range to 2
as
1
35
compared to 2 - in 7090/7094 FORTRAN IV;
real (floating-pOint) constants can be from 1 to
7 digits in length as compared to a maximum of
9 digits for 7090/7094 FORTRAN IV; and
double precision constants can range from 1
to 16.digits as compared to a maximum of 17
digits for 7090/7094 FORTRAN IV.
(2) In FORMAT statements: O-type conversions
are not available; X-conversion fields must
be separated from other fields by commas.
IBM SYSTEM/360
(17)
In the F-specification of FORMA T statements, an extra space must be allotted
for a digit to the left of the decimal point.
(18)
Embedded blanks are not permitted within
variable names, real and double-precision
constants, all statements, and FORTRAN
key words.
(19)
The following statements are not implemented:
ASSIGN
BLOCK DATA
Assigned GO TO
Logical IF
PRINT
PUNCH
READ Q; list.
.143 Extensions of Basic 02erating S;ystem/360
FORTRAN IV Relative to mM 709077094
FORTRAN IV
(1)
In FORMAT statements, the T-specification
indicates the starting print position of the
data.
(4) The A TAN2 mathematical subroutine is not
provided.
(2)
Storage dumps can be printed in hexadecimal,
double-precision, real, or integer formats.
(5) The SSWTCH subroutine is not provided.
(3)
The magnitude of real and double-precision
constants can range up to 16 63 (about 1075),
as compared to 10 38 for 7090/7094 FORTRAN
IV.
(4)
(7) Variables and constants cannot be of type
COMPLEX or LOGICAL; logical and relational
'expressions are not permitted.
Literal data in FORMAT statements can
extend up to 255 characters in length.
(5)
Mixed modes are permitted in arithmetic
expressions and statements.
(8) FORMAT statements cannot be read at object
time.
(6)
Integer constants and variables can have
real exponents.
(9) Only one level of parentheses (in addition to
those enclosing the entire statement) is
permitted in FORMAT statements.
(7)
An integer constant can be included in the
STOP statement.
(3) Symbolic input-out unit designation
is not provided.
(6) In statement functions, a maximum of 15
variables that appear within an expression
can be used as arguments of the function.
(8)
(10) Specification statements must precede the
first executable statement of the source
program. In addition, all Statement Function
definitions must follow the specification
statements and precede the executable program.
The END statement is used to define the
end of a source subprogram, as well as a
main program.
(9)
In FORMAT statements, the length of Afields is not limited, and literal data can
be included.
(11) Adjustable dimensions are not permitted.
(10)
A " + " character can be used for carriage
control purposes, indicating "no advance."
(11)
Variables in EQUIVALENCE statements
can have multiple subscripts.
(12)
The double-precision float (DFLOAT) mathematical function has been included.
(13)
Special statements can optionally be provided
for handling the input-output operations of
direct-access ~torage devices when data
records are to be randomly accessed. These
include a DEFINE FILE specification statement, new forms of the READ and WRITE
verbs, and a FIND verb which permits
record retrieval to proceed concurrently
with computation.
(12) Common block names are not permitted.
(13) Forty-one words are reserved for use by the
compiler and cannot be used as names of
variables, arrays, or subprograms.
(14) All complex mathematical functions and
common logarithm functions have been
excluded.
(15) Literals cannot serve as arguments in a
CALL statement.
(16) The last statement in every DO loop must be
executable and cannot be STOP, PAUSE, or
RETURN.
8/65
/
420: 163. 100
IBM System/360
Process Oriented Language
BPS FORTRAN IV
PROCESS ORIENTED LANGUAGE: BASIC PROGRAMMING SUPPORT FORTRAN IV
.1
GENERAL
.11
Identity:.........
Basic Programming Support
FORTRAN IV.
• 12
Origin:..........
IBM Corporation.
. 13
Reference:.......
IBM Publications C28-65042, C28-6583-0, and
C28-6584-0.
• 14
Description
Two versions of FORTRAN IV compilers have been
announced for users of the IBM System/360 Basic
Programming Support (BPS): a Tape FORTRAN IV
and a Card FORTRAN IV, both of which operate
independently of any control programs. Both BPS
FORTRAN IV compilers require a minimum of
16,384 bytes of core storage for their operation,
in addition to the Standard Instruction Set with the
Floating-Point feature. The tape version requires
two 2400 Series Magnetic Tape Units for intermediate storage of compiler results, and another
2400 Series Magnetic Tape Unit for compiler and
control program storage. In order to "compileand-go," a fourth magnetic tape unit is required.
The BPS Tape FORTRAN IV compiler is currently
available for use in the field.
The language specifications of BPS Card FORTRAN
IV have not been released by IBM to date. It is
assumed that the card version's language facilities
will be very similar to those of the BPS Tape
FORTRAN IV language, which are described
throughout the remainder of this report section.
Basic Programming Support FORTRAN IV is a
proper, though very restricted, subset of the
Operating System/360's full (200K) FORTRAN IV
language, which is described in report Section
420:161. In fact, BPS FORTRAN IV resembles very
closely the Operating System/360's E Level Subset
FORTRAN and the Basic Operating System's
FORTRAN. Upward compatibility will definitely
exist among the various System/360 FORTRAN
compilers; i. e., any program successfully compiled by the BPS FORTRAN IV compiler should
produce identical results if subsequently compiled
by the BOS or Operating System/360 FORTRAN IV
compilers.
The language facilities of Basic Programming
Support Tape FORTRAN IV are identical with those
of Basic Operating System FORTRAN IV (described
in Section 420: 162), except for the lack of any statements relating to direct-access storage devices in
the BPS version. Both of these languages are .
proper subsets of the Operating System/360's E
Level Subset FORTRAN IV.
Listed below are the only known language restrictions of BPS FORTRAN IV as compared with the
Operating System/360's E Level Subset FORTRAN
IV (described in Section 420: 161):
•
Embedded blanks are not permitted within
variable names, real and double-precision
constants, all statements, and FORTRAN
key words .
•
In the F-specification of FORMAT statements,
an extra space must be allotted for a digit to
the left of the decimal point.
•
The last statement in every DO loop must be
executable and cannot be STOP, PAUSE, or
RETURN.
•
Forty-one key words are reserved by the
compiler - two more than are reserved by
the E Level Subset compiler.
•
No statements related to the use of directaccess devices are permitted.
In Paragraphs . 142 and • 143 that follow, BPS
FORTRAN IV is compared with IBM 7090/7094
FORTRAN IV, a language that has gained wide
acceptance for use in the solution of scientific and
engineering data processing problems. IBM 7090/
7094 FORTRAN IV is described in detail in
Section 408:162 of the IBM 7090 report.
No performance figures have been released by IBM
to date relative to the estimated compilation speed
or operating efficiency of the BPS Tape FORTRAN
IV compiler. The core storage requirements for
BPS Tape FORTRAN IV program compilation and
execution are listed in Table I.
.141 Availability
Language
specifications: . • ..
Compiler: . •• • . • ••
April 1964 and March 1965
for BPS Tape FORTRAN
IV;
Not yet available for BPS
Card FORTRAN IV.
2nd quarter 1965 for BPS
Tape FORTRAN IV;
1st quarter 1966 for BPS
Card FORTRAN IV.
• 142 Restrictions of Basic Programming Support
FORTRAN IV Relative to IBM 7090/7094
FORTRAN IV
(1) Integer constants can range to 2 31 - 1 as compared to 2 35 - 1 in 7090/7094 FORTRAN IV;
real (floating-point) constants can be from
1 to 7 digits in length as compared to a maximum of 9 digits for 7090/7094 FORTRAN IV;
and double precision constants can range from
1 to 16 digits as compared to a maximum of
17 digits for 7090/7094 FORTRAN IV.
(2)
In FORMAT statements: O-type conversions
are not available; X-conversion fields must
be separated from other fields by commas ..
(3) Symbolic input-output unit designation is not
provided.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65·
420: 163.142
IBM SYSTEM/360
• 142 Restrictions of Basic Programming
Support FOI,tTRAN IV Relative to
IBM 7090/7094 FORTRAN IV (Contd.)
(19) The following statements are not implemented:
ASSIGN
BLOCK DATA
Assigned GO TO
Logical IF
PRINT
PUNCH
READ Q, list.
(4) The ATAN2 mathematical subroutine is not
provided.
(5) The SSWTCH subroutine is not provided.
(6) In statement functions, a maximum of 15
variables that appear within an expression
can be used as arguments of the function.
(7) Variables and constants cannot be of type
COMPLEX or LOGICAL; logical and relational
expressions are not permitted.
(8) FORMAT statements cannot be read at object
time.
(9) Only one level of parentheses (in addition to
those enclosing the entire statement) is permitted in FORMAT statements.
(10) Specification statements must precede the first
executable statement of the source program.
In addition, all Statement Function definitions
must follow the specification statements and
precede the executable program.
(11) Adjustable dimensions are not permitted.
.143 Extensions of Basic Programming Support
FORTRAN IV Relative to IBM 7090/7094
FORTRAN IV
(1) In FORMAT statements, the T-specification
indicates the starting print position of the data.
(2) Storage dumps can be printed in hexadecimal,
double-precision, real, or integer formats.
(3) The magnitude of real and double-precision
constants can range up to 16 63 (about 10 75 ),
as compared to 103 8 for 7090/7094 FORTRAN
IV ..
(4)
Literal data in FORMAT statements can extend
up to 255 characters in length.
(5)
Mixed modes are permitted in arithmetic
expressions and statements.
(12) Common block names are not permitted.
(I?) Forty-one words are reserved for use by the
. compiler and cannot be used as names of
variables, arrays, or subprograms.
(14) All complex mathematical functions and common logarithm functions have been excluded.
(6) Integer constants and variables can have real
exponents.
(7)
An integer constant can be included in the
STOP statement.
(15) Literals cannot serve as arguments in a CALL
statement.
(8) The END statement is used to define the end
of a source subprogram, as well as a main
program.
(16) The last statement in every DO loop must be
executable and cannot be STOP, PAUSE, or
RETURN.
(9) In FORMAT statements, the length of A-fields
is not limited, and literal data can be included.
(17) In the F-specification of FORMAT statements,
an extra space must be allotted for a digit to
the left of the decimal point.
(18) Embedded blanks are not permitted within
variable names, real and double-precision
constants, all statements, and FORTRAN key
words.
(10) A "+" character can be used for carriage
control purposes, indicating "no advance."
(11) Variables in EQUIVALENCE statements can
have multiple subscripts.
(12) The double-precision float (DFLOAT) mathematical function has been included.
TABLE I: BPS TAPE FORTRAN IV CORE STORAGE SIZE CONSIDERATIONS
System/360 Core Storage Size in Bytes
Average number of source statements
permitted per compilation
32K
65K
400
1,000
2,000
Core storage reserved at execution
time by FORTRAN control system, bytes
9K
9K
9K
Core storage available at execution time
for object program, bytes
7K
23K
55K
Average number of source statements
permitted for a compilation and immediate execution
8/65
16K
150
400
1,000
420: 164. 100
IBM System/360
Process Oriented Language
Operating System/360 COBOL
PROCESS ORIENTED LANGUAGE: OPERATING SYSTEM/360 COBOL
.1
GENERAL
. 11
Identity: ...
. 12
Origin: .. . . . . . . • • • IBM Corporation•
.13
Reference: ••...•.. IBM Publication
C28-6516-2.
.14
Description
Operating System/360 COBOL F
... Operating System/360
COBOL.
The IBM Operating System/360 COBOL language is
a subset of COBOL-61 Extended, the most widely
accepted pseudo-English common language designed for use in business-oriented data processing
applications. Operating System/360 COBOL includes all but six of the facilities prescribed for
implementation in Required COBOL-61 (see
Paragraph.142 below). It also includes two
valuable Extensions to Required COBOL-61: the
Report Writer and Sort functions. IBM has also
chosen to include many of the COBOL-61 Electives
in its Operating System/360 COBOL language. The
restrictions, extensions, and elective facilities of
Operating System/360 COBOL with respect to
Required COBOL-61 are tabulated for ease of
reference in Paragraphs. 142 through. 145.
Although Operating System/360 COBOL includes
most of the facilities officially suggested for inclusion in COBOL compilers - including the
features currently being studied by the American
Standards Association for COBOL - yet it still
falls short of being a really "common" language.
Conversion of existing COBOL source programs to
System/360 COBOL will not be an automatic process. Ignoring program differences that will result from environmental and data format considerations, IBM lists 64 Operating System/360 COBOL
language elements that are sensitive to conversion
operations from currently-operational IBM COBOL
source programs. Acknowledging the COBOL program conversion problem, IBM has offered its
1400 and 7000 Series COBOL users a COBOL
Language Conversion Program (LCP) that flags
items which are incompatible with System/360
COBOL and reconstructs them into compatible
statements wherever possible.
Operating System/360 COBOL is implemented at two
program design levels: "COBOL E" at 13K and
"COBOL F" at 44K bytes of core storage. Both
compiler versions operate in cooperation with and
under the supervision of the Operating System/360
in a direct-access device environment. Among the
language facilities of COBOL F that are not included
in its subset, COBOL E, are the random processing
and multiprogramming (asynchronous processing)
language extensions. A complete listing of differences between the two Operating System/360
COBOL language versions is supplied below.
Operating System/360 COBOL F requires a minimum
of 65, 536 bytes of core storage for its operation, in
addition to a direct-access device for system residence and another such device for intermediate
storage. Input-output devices can include card
readers, printers, 2400 Series Magnetic Tape Units,
and any available direct-access storage devices.
The Standard Instruction Set with the Decimal
Arithmetic optional featur!=l is also required as part
of the minimum equipment configuration for using
the COBOL F compiler.
Two different programming techniques are provided
through the use of certain Operating System COBOL
F facilities. One of these is the conventional technique of processing instructions sequentially and
handling data records in the same order in which
they are read. This technique is referred to as
"synchronous processing." The other technique "asynchronous processing" - permits the programmer to take full advantage of direct-access
storage devices in the performance of multiprogramming tasks.
By means of a newly-implemented PROCESS statement, a named direct-access-device-oriented
routine is entered and executed in an out-of-line
mode, while the in-line portion of the program proceeds in its sequential operations. Control over
this method of asynchronous processing is facilitated by the addition of a HOLD processing statement. The out-of-line program that is to be
processed in the asynchronous, multiprogramming
mode must be written in the USE FOR RANDOM
PROCESSING Section in the Declarations portion of
the Procedure Division. The primary function of
this programming technique is to allow the overlapping of random accessing time with the processing of other data records.
Operating System/360 COBOL F supports various
data organizations, record formats, and access
methods. The three available types of data organization include the standard sequential, the indexed, and the relative organization. Since the
number and type of control fields used to locate
logical records within a file differ, depending on
the data organization method used, incompatibility
between the methods results. Thus, records
created in a standard sequential file cannot also be
read as an indexed file. The indexed and relative
data organizations are, based on symbolic or actual
keys, supplied by the programmer, which direct
the positioning of direct-access devices in preparation for logical ;record processing. Special forms
of the READ and WRITE statements control these
sequential and random-access methods of proCeSsing logical data records. A complete list of the
clauses and statements that permit random processing of records stored on direct-access devices is
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
IBM SYSTEM/360
420: 164.140
. 14
Description (Contd.)
provided in Paragraph. 143, Extensions to Required COBOL-61.
Another useful extension included in Operating
System/360 COBOL (both E and F versions) is a
program debugging language. Debugging statements
can be included anywhere in the source program, or
they can be arranged in groups or "packets" according to program section-names referenced and
entered for compilation immediately following the
source program. A TRACE statment causes
specific messages to be written as the object program enters every program paragraph or section.
EXHIBIT produces formatted snapshots of any datanames listed in the statement, and (optionally)
inhibits the printing of the data-names until the
values contained therein are changed. Another control statement that regulates the execution of the
debugging entries is ON - a conditional statement
that permits the operation of specified diagnostics
only when given conditions are satisfied. If the debugging statements are grouped in packets, a
DEBUG statement must be used to indicate the beginning of each logical testing operation.
The Report Writer facility of Extended COBOL-6i
is also included in System/360 COBOL F. This
facility is implemented by entries in the Data
Division and by three new verbs. Report specifications in the Data Division are contained in the
File Description, Report Description, and Report
Group Description entries. The latter two entries
describe the format of the report page. A report
group describes a set of data that is to be considered as an individual unit (i, e., a detail line, a
set of constant report headings, or a series of
variable control totals). The INITIATE verb
initiates the processing of a particular report, the
GENERATE verb links the Procedure Division to
the Report Writer at object time, and the TERMINATE verb terminates the processing of a report.
Additional flexibility in controlling the Report
Writer is provided by the ability to enter control
parameters by means of the USE BEFORE
REPORTING declarative statement of the Procedure
Division.
Operating System/360 COBOL F also provides the
SORT feature of Extended COBOL-61. This facility
can be used for two purposes: to sort an intermediate file (intermediate data is created and then
sorted into some sequence for further processing);
and to process data before it is sorted and to
further process it after it has been sorted. The
SORT facility is implemented by a Sort Description
entry in the Data Division and by three new verbs.
The SORT verb controls the sequencing of records,
the RELEASE verb transfers records to the initial
phase of a sort operation, and the RETURN verb
obtains sorted records from the final phase of a
sort operation.
Corresponding to the flexibility of internal data formats inherent in the System/360 design, the COBOL
language permits data to be maintained in storage in
five different formats, as specified by the USAGE
clause of the record description entry. The five
USAGE entries and their corresponding data formats are as follows:
DISPLAY ......••• one character per byte •
COMPUTATIONAL ..• binary data item.
COMPUTATIONAL-1 . short (one-word) floating
point.
COMPUTATIONAL-2 . long (two-word) floating
point.
COMPUTATIONAL-3 . packed decimal (2 digits
per byte).
Noteworthy elective features included within
Operating System/360 COBOL F are the Segmentation Feature, the Source Program Library Facility,
and the COMPUTE verb. (A complete listing of
COBOL-61 electives implemented in Operating
System/360 COBOL is provided in Paragraph. 144. )
The Segmentation Feature of Elective COBOL is implemented in a nonstandard way in that the linking
mechanism between the main program and called
subprograms is not provided automatically by the
compiler. Instead, the System/360 programmer
must construct and control the program call-in
procedures according to his needs. The ENTER
statement, used in conjunction with CALL or
ENTRY statements, sets up the framework of communication between the COBOL object program and
one or more COBOL subprograms or subprograms
in other languages. Data-names describing the
subprograms to be linked to the main program are
listed in the ENTRY statement and defined in the
Linkage Section of the Data Division. The RETURN
VIA statement enables the restoration of whatever
registers were saved at a subprogram entry point,
and indicates the point of return in the main program.
The Source Program Library facility permits
source program entries in the System/360 program
library to be included in the COBOL program at
compile time. Thus, an installation can utilize
standard COBOL file dE;lscriptions, record descriptions, or procedures without having to program
them repetitively. These entries and procedures
are entered into the source program by means of a
COPY clause or an INCLUDE statement.
The COMPUTE verb is another valuable elective
incorporated into Operating System/360 COBOL F.
COMPUTE permits arithmetic operations to be
expressed in a concise formula notation similar
to that of FORTRAN. For example, the COBOL
operations:
SUBTRACT B FROM A GIVING T
DIVIDE C INTO T GIVING X
can alternatively be expressed as:
COMPUTE X
= (A-B)/C.
Operating System/360 COBOL F as announced by
IBM provides no direct language facilities to control Teleprocessing operations. However, the
Operating System/360's Queued Telecommunication
Access Method (QT AM) can be used to obtain
messages from and place messages on a message
queue by means of specialized GET and PUT
macro-instructions.
Operating System/360 COBOL E
Operating System/360 COBOL E is a proper subset
of COBOL F. It is designed at a 13K-byte program
(Contd. )
8/65
420: 164.141
PROCESS ORIENTED LANGUAGE: OPERATING SYSTEM/360 COBOL
. 14
Description (Contd.)
design level and requires a minimum of 32,768
bytes of core storage in order to perform compilations in conjunction with the Operating System/
360. A direct-access devicE:! is required for compiler and control program residence, and another
direct-access device or four 2400 Series Magnetic
Tape Units are required for intermediate storage
of compiler results. The Standard Instruction Set
with the Decimal Arithmetic feature is also required.
Among the features of Operating System/360
COBOL F that are not included in the COBOL E
subset are:
•
Asynchronous Processing, including the USE
FOR RANDOM PROCESSING sentence and the
HOLD and PROCESS statements.
•
The Report Writer Feature.
•
The Sort Feature.
•
The CORRESPONDING option of the ADD,
SUBTRACT, and MOVE statements.
•
Implied subjects and relations in compound
conditions.
•
Nested IF statements.
The following features cif Operating System/360
COBOL F will be available in COBOL E, but in
restricted forms:
figurations and indicates the several timing factors
that must be summed to arrive at a total time for
a COBOL compilation performed in conjunction with
the Operating System/360.
An IBM-supplied timing example shows the estimated time requirements to perform a full COBOL
E compilation of a program consisting of 500 source
statements on a System/360 sample Configuration
A. In this example, it is assumed that the compiled
program will be 10,000 bytes long. The estimated
time in seconds to perform this COBOL E compilation, reading the source statements from
punched cards and writing the compiled program to
disc storage, is 308 seconds.
· 141 Availability
Language: . . . . . . .
Compiler COBOL E: . . . . . . .
COBOL F: . . . . . . •
April and December 1964.
December 1965.
June 1966.
· 142 Deficiencies of Operating System/360 COBOL E
and F With Respect to Required COBOL-61
Environment Division:
•
•
•
The SOURCE-COMPUTER and OBJECTCOMPUTER paragraphs cannot be copied from
the COBOL library.
A SPECIAL-NAMES paragraph is not provided.
The OPTIONAL, RENAMING, and MULTIPLE
REEL options of the FILE-CONTROL paragraph have not been implemented.
Data Division:
•
The DEPENDING ON option of the OCCURS
clause.
•
The Sterling Currency conversion feature.
•
The EXHIBIT statement in the debugging
language.
Finally, the following features of Operating System/
360 COBOL F will eventually be made available in
COBOL E, according to IBM, although they will not
be available initially.
•
The Source Program Library facility that permits copying of entire source programs.
•
Random Processing capability with direct-access
devices. (These devices can still be used, but
only with sequential processing methods. )
/
Among the noteworthy language and compiler
features of Operating System/360 COBOL F that
are implemented in the smaller COBOL Eversion
are the Debugging Language, the Segmentation
Feature, a restricted Source Program Library
facility, the COMPUTE verb, and the use of five
different data formats by means of the USAGE
clause.
Compilation Times
IBM has provided us with estimated COBOL compilation times based on three sample System/360
configurations. Table I describes these sample con-
©
•
•
•
The recorddescription clauses SIZE, POINT,
CLASS, ZERO SUPPRESS, CHECK PROTECT,
and FLOAT DOLLAR SIGN are not allowed.
The JUSTIFIED LEFT option in the record
description section is not permitted.
No Constant Section is permitted.
· 143 Extensions of Operating System/360 COBOL With
Respect to Required COBOL-61.
Extensions Included in Both COBOL E and
COBOL F
•
•
A COBOL debugging language is provided
which includes TRACE, EXHIBIT, ON (conditional control), and DEBUG verbs.
Sterling currency conversion requirements can
be satisfied by defining sterling nonreport
items to be used internally, and sterling report
data items to be edited for printing. Facilities
will be provided to handle sterling nonreport
items in the MOVE, ADD, and SUBTRACT
verbs. The use of sterling nonreport items in
other arithmeti.c statements is permitted only
in COBOL.F. Other provisions for international use of COBOL are interchangeability
of commas and decimal points, ability to alter
character sets and the currency symbol, and
substitution of sentences to allow translation of
output messages into any non-English language.
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
420: 164.143
.IBM SYSTEM/360
TABLE I: COBOL COMPILATION TIME FACTORS
Sample Configurations
Configuration Components
Configuration A
System/360 Model 30 with 32K bytes of core storage, one Multiplexor Channel, one Selector Channel, one 2540 Card Read
Punch, one 1403 Printer (Model 3) two 2311 Disk Storage
Drives, and the Operating System/a60'S sequential scheduler. lie
Configuration B
Syatem/360 Model 50 with 65K bytes of core storage, two
Selector Channels, two 2311 Disk Storage Drives, and the
Operating System/360 I s priority scheduler.
Configuration C
System/360 Model 65 with 262K bytes of core storage, two
Selector Channels, two 2301 Drum Storage Units, four 2311
Disk Storage Drives, and the Operating System/360's concurrent job scheduler.
*'"
"'*
Compilation Time in Seconds
Estimated Timing Factors
COBOLE,
Configuration A
For compilation initiation
For each compilation
For each soUrce card
For library initiation
For "each library access
For each library routine used
For each byte of library
routine used
For each 2, 048-byte control
section compiled
For each byte compiled
1\1
**
21.1
30.0
0.45
19.1
2.0
0.75
16.1
13.0
0.054
13.9
1.4
0.37
0.00009
0.00002
0.088
0.00018
0.045
0.0001
Source program input is assumed to be entered through the Card Read Punch, and object
program output is assumed to be written on disc stor~ge.
Source program input and object program output are assumed to be read from and written
to disc storage.
A TRANSFORM verb is provided to alter characters according to a set transformation rule. The
rule is determined by the combination of FROM
and '1'0 options that is chosen. The format for
the statement is:
•
An extended Source Program Library facility
provides the option of attaching a complete
source COBOL program to the calling COBOL
program at compilation time.
•
The technique of asynchronous processing permits the COBOL programmer to take advantage
of direct-access storage devices in the performance of multiprogramming tasks. To implement this technique, . a USE FOR RANDOM
PROCESSING sentence and PROCESS and HOLD
verbs have been developed.
•
Clauses and statements are provided to handle
random processing of data stored on directaccess devices. These include: the
ORGANIZATION, ACCESS, SYMBOLIC KEY,
ACTUAL KEY, and ASSIGN to DffiECT-ACCESS
clauses; the RESTRICTED SEARCH OF integer
TRACKS option of the APPLY clause; the REWRITE statement; the 1-0 option of the OPEN
statement; and the INVALID KEY option of the
READ and WRITE verbs.
•
The Report Writing facility is implemented,
although with some restrictions. The following
clauses are not allowed in the Report Group
Description entries: CLASS, POINT, SIGNED,
SIZE, USAGE, ZERO SUPPRESS, CHECK,
FLOAT SIGN, and the SELECTED option of the
SOURCE clause. A PRINT-SWITCH option can
inhibit printing of ·specified report groups.
•
The SORT facility is also implemented in a
slightly restricted manner. In the Sort
Description (SD) entry, the FILE CONTAINS
optional clause is omitted. Also not included
are the FROM option of the RELEASE verb and
the INTO option of the RETURN verb.
TRANSFORM data .... name-3 CHARACTERS
TO
figurative-constant-1
non-numeric-literal-1
data-name-1
{
{
figurative-constant-2
non-numeric-literal-2
data-name-2
}
}
•
Floating-pOint literals and items (external and
internal) are permitted.
•
A NO REWIND option is available with the OPEN
verb.
Extensions Included in COBOL F Only:
•
3.92
7.0
0.013
4.01
0.32
0.088
0.046
0.00084
• The Source Program Library facility is included
to permit the automatic inclusion of catalogued
COBOL file descriptions, record descriptions,
and procedure statements into the source program at compilation time.
FROM
COBOLF,
Configuration C
0.0004
• 143 Extensions of Operating System/360 COBOL With
Respect to Required COBOL-61 (Contd.)
•
COBOL F,
Configuration !l
The ADD, SUBTRACT, and MOVE verbs have a
CORRESPONDING option that permits selective
operation on matching data items only.
(Contd. )
8/65
420: 164.144
PROCESS ORIENTED LANGUAGE: OPERATING SYSTEM/360 COBOL
.144 COBOL-61 Electives Implemented in Ope'ratihg System/360 COBOL (see 4: 161. 3)
Note: the Elective features that are available only with Operating System/360 COBOL F
are marked with an asterisk.
1
2
3
6
Semicolon
Relationship characters
Figurative constants
File Description
Label formats
13
Record Description
Table length
Label handling
22
24
Verbs
COMPUTE
ENTER
25
INCLUDE
26
USE
27
28
29
30
31
32
33
35
36
\,
Characters and Words
Formula characters
10
21
Comments
Elective
Key No.
Verb Options
LOCK
MOVE CORRESPONDING*
OPEN REVERSED
ADVANCING
STOP execution
Formulas
Operand size
Tests
Implied Conditionals*
37
38
Compound conditionals
Complex conditionals*
39
Conditional statements I
sequence*
43
46
Environment Division
FILE-CONTROL
I-O-CONTROL
47
Identification Division
Date-compiled
48
49
Special Features
Library
Segmentation
+, -, *,
I,
**,
=.
can be used for punctuation.
=, >, and < are available.
HIGH-VALUE(S), LOW-VALUE(S).
NONSTANDARD labels are permitted.
the "DEPENDING ON" option is provided so
that lengths of tables and arrays can vary.
labels may be omitted, or standard or nonstandard labels can be used.
permits algebraic formulas.
used for linkage to subroutines (not to· enter
a new language).
library routines can be called (no
REPLACING option).
non- standard I/o error and label handling
routines can be used.
rewound tapes can be locked.
items can be moved in groups.
tapes can be read backward.
paper advance can be specified.
coded message is printed.
algebraic formulas can be used.
up to 18 digits.
IF { } IS NOT ZERO form ·is provided.
implied operators with implied subjects
are permitted.
ANDs and ORs can be intermixed.
conditional statements within conditional
statements are permitted.
!NVALID KEY can follow imperative statements.
library descriptions can be copied.
SAME and APPLY clauses can be used.
current date can be printed when program
is compiled.
library routines can be called.
implemented in nonstandard manner.
* COBOL F version only.
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
420: 164.145
IBM SYSTEM/360
.145 COBOL-61 Electives Not Implemented (see 4: 161.3)
Key No.
4
5
Characters and Words
Long literals
Figurative constants
Comments
7
Computer-name
literals may not exceed 120 characters.
HIGH-BOUND(S), LOW-BOUND(S) not
available.
no alternative computer names.
8
9
12
File Description
BLOCK size
FILE CONTAINS
SEQUENCED ON
HASHED
no range in block size permitted.
approximate file size cannot be shown.
no list of keys can be given.
hash totals cannot be created.
14
Record Description
Item length
11
8/65
Elective
15
16
17
BITS option
RANGE IS
RENAMES
18
19
20
SIGN IS
SIZE clause option
Conditional range
23
Verbs
DEFINE
34
Verb Options
Relationships
40
41
42
44
45
Environment Division
SOURCE-COMPUTER
OBJECT-COMPUTER
SPECIAL-NAMES
PRIORITY IS
1..,0 CONTROL
variable item lengths cannot be specified in
a PICTURE.
items cannot be specified in binary.
value ranges of items cannot be shown.
alternate groupings of elementary items
cannot be specified.
no separate signs allowed.
variable length items cannot be specified.
a conditional value cannot be a range.
new verbs cannot be defined.
IS UNEQUAL TO, EQUALS, and EXCEEDS
are not provided.
only "computer name" is allowed.
only "computer name" is allowed.
no special-names paragraph is permitted.
priorities cannot be assigned to files.
library descriptions cannot be used.
420: 165. 100
IBM System/360
Process Oriented Language
BOS COBOL
PROCESS ORIENTED LANGUAGE: BASIC OPERATING SYSTEM/360 COBOL
•1
GENERAL
.11
Identity:.........
Basic Operating System/360
COBOL.
.12
Origin:..........
IBM Corporation.
• 13
Reference:.......
IBM Publication
C24-3433-0.
• 14
Description
A single COBOL language and compiler is offered
by IBM for use with either the tape-resident or discresident Basic Operating System/360 in a minimum
environment of 16,384 bytes of core storage.
According to preliminary announcements, Basic
Operating System/360 (BOS) COBOL will include aU
but six of the facilities prescribed for implementation in Required COBOL-61. Many useful extended
and elective features will also be provided to permit
effective utilization of the hardware characteristics
of the System/360.
BOS COBOL has been designed at a 10K-byte program design level, to enable it to operate with 16K
System/360 Processing Units in conjunction with the
Basic Operating System's control programs. Included in this basic program design are all of the
language facilities of Operating System/360 COBOL
E. Statements related to the random processing of
records stored on direct-access devices will also be
supplied with BaS COBOL. For a complete description of the restrictions, extensions, and electives of
BOS COBOL - relative to Required COBOL-61 and
to Operating System/360 COBOL F-report Section
420:164 should be consulted. Since BOS COBOL and
Operating System/360 COBOL E provide very similar language facilities, whatever is applicable to the
COBOL E language, in Section 420:164 can also be
considered as applicable to BOS COBOL.
The random processing statements are standard
features of Operating System/360 COBOL E, so they
are described in Section 420:164 of this report •
These statements permit use of the direct-access
and indexed sequential file organization methods and
the direct-access method of random record retrieval. If these optionally available COBOL statements
are not selected, direct-access devices can still be
used for input-output operations in the sequential
record access method •
The minimum equipment configuration required to
operate the 10K-level BOS COBOL compiler includes: a System/360 Processing Unit with at least
16,384 bytes of core storage and the Standard Instruction Set with the Decimal Arithmetic feature;
one 2400 Series Magnetic Tape Unit or 2311 Disk
Storage Drive for Basic Operating System/360 program residence; three additional 2400 Series Magnetic Tape Units or one 2311 Disk Storage Drive for
intermediate storage of compiler results; and any
standard-model card reader, printer, and punch
unit.
A list of the input-output devices supported by the
BOS COBOL compiler includes the following: 2501
Card Reader; 1442, 2520, and 2540 Card Read
Punches; 1403, 1404, 1443, and 1445 Printers;
2400 Series Magnetic Tape Units; 2311 Disk Storage
Drive; 2321 Data Cell Drive; 1052 Printer-Keyboard;
and 1015 Inquiry Display Terminal
IBM has released no performance estimates to date
relative to the compilation speed or object program
efficiency of the Basic Operating System/360 COBOL
compiler. However, IBM expects the performance
of BOS COBOL to be similar to that of Operating
System/360 COBOL E, as described in Section
420:164.
• 141 Availability
Language
specifications: . . . . . September, 1965.
Compiler: . . . . . . . . . 4th quarter 1965; direct access statements by 2nd
quarter 1966.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
420:167.100
IBM System/360
Process Oriented Language
Operating System/360 PL/I
PROCESS ORIENTED LANGUAGE: OPERATING SYSTEM/360 PLiI
.1
GENERAL
.11
Identity:
.12
Origin:
.13
Reference: . . . . . . . . IBM Publication C28-6571,
published May, 1965.
.14
Description
. . . . . . . . . Programming Language/I.
PL/I.
(Formerly NPL or New
Programming Language).
. . . • . . . . . . IBM Corporation and the
SHARE NPL committee.
The PL/I language is a new high-level programming
language, developed jointly by IBM and the SHARE
organization, for use in both commercial and
scientific applications. It includes facilities
that are more powerful and extensive than those of
any other presently-announced computer compilation
language. These facilities include:
o Program check-out operations.
"
Handling of self-identifying input and
output files.
" Random-access file facilities, with a choice
of a number of access methods.
" Dynamic storage allocation facilities.
Q
Asynchronous operation facilities, enabling
different parts of a program to run in parallel
with each other.
o Program modification ability, allowing parameters and other changes to be inserted
into a skeleton text immediately before
compilation.
.
In addition, PL/I includes a number of more con-
ventional functional facilities, such as sorting,
report-writing, data conversion, and communications facilities, as well as the conventional proces s ing statem ents .
\
An important part of PL/I's design has been concerned with the development of useful subsets that
can be selected from the full language. These
subsets will be less complex than the full language
and can be learned and used by programmers who
have no immediate use for the complete range of
language facilities. This subset feature should
be particularly useful in training programmers
for business applications, where PL/I's extensive
scientific facilities will be of little or no value and
could easily cause confusion.
This report covers the PL/I language, as defined
by IBM, rather than any subset or specialized
implementation. IBM has cautioned that not all
of the language features will be available in the
initial versions of the compilers, but no other
specifications have been announced regarding
the specific restrictions that will apply to various
PL/I implementations. Paragraph .15 of this
report presents preliminary information and performance figures for the Operating System/360
PL/I compilers.
Because the PL/I language introduces several
new concepts and language features, and because
many of the terms used in describing it may be
unfamiliar to the reader, this report is unusually
lengthy and comprehensive.
The language facilities offered in PL/I are described on the following pages in a manner consistent with the coverage of other Process
Oriented Languages in these reports. Section
4:160 of the Users' Guide can be consulted for
explanations of the various paragraph headings.
The last portion of this report is an Appendix, or
general reference section, that provides a list
of key terms used in describing the PL/I language
(Table VI); a comparison of the principal features
of PL/I, FORTRAN, and COBOL (Table X); and
other useful information.
In its outward appearance, a PL/I source program
will probably resemble a FORTRAN program
more than a COBOL program, although many
coding conventions of both compilers have been
incorporated. However, unlike COBOL, the
PL/I program is not broken into separate elemental
divisions. Instead, file descriptions and data
characteristics can be defined within the body of
the program.
PL/I offers a wide variety of data formats and
many types of operands and expressions. For
example, the operational repertoire provides for
the use of fixed-point and floating-point operands,
coded in either binary or decimal form. Both real
and complex numbers can be represented, although
complex numbers must be written as expressions
consisting of a real part and an imaginary part,
rather than as single items; e. g., 2.1 + 3.71. In
addition, constants can be entered in the form of
character-strings or bit-strings of binary digits.
Mixed-mode operations make feasible the use of
PL/I's variety of data formats and operand types.
For example, fixed-point decimal fields can be
added to floating-point binary fields in order to
obtain an integer value. PL/I automatically provides the necessary mode and radix conversions
during the evaluation of the expressions. PL/I
follows a number of specified rules in establishing
the size and characteristics of the intermediate
results during an evaluation of mixed-mode expressions and statements. Because these evaluation rules assume worst-case situations, unnecessary time-consuming computations may well
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
9/65
IBM SYSTEM/360
420: 167.140
. 14
become known. Some preliminary information on
the Operating System/360 PL/I compilers is presented in the following paragraphs.
Description (Contd.)
result. Table IX contains a sampling of the manner
in which PL/I treats intermediate results when
evaluating expressions.
.15
.The operators provided in PL/I include the standard
arithmetic and comparison operations, as well as
the AND, NOT, and inclusive OR logical operations.
PL/I also provides the concatenation operator (" ) to
assist in handling character-strings and bit-strings
of data. An example of the concatenation (stringingtogether) operation is the following: 3 + 4 II 5
equals 75.
PL/I's operational repertoire includes more than
70 built-in functions. Several of these functions,
such as ADD, SUBTRACT, etc., appear to have
been provided in imitation of COBOL operations.
Most of the functions, however, add new, mainly
scientific facilities to the language. An interesting
feature of some of the built-in functions is the
capability to select the amount of precision desired
in function evaluation.
The PL/I language, as described by IBM, appears
to offer many valuable and convenient programming
facilities. It will not be possible to assess PL/I's
true value until the various PL/I compilers become
available and their operational characteristics
Operating System/360 PL/I Compilers
Two compilers have been announced by IBM for
compilation of PL/I programs in conjunction with
the Operating System/360. The PL/I F Level
compiler operates in a minimum environment of
65K bytes of core storage, and the PL/I H Level
compiler requires at least 262K bytes of core storage. Both compilers will be capable of handling
the full PL/I language, but IBM indicates that the
implementation of some of the language features
will be "deferred" in the initial version of the
compilers. Full upward and downward compatibility exists between these compilers, and a source
program compiled by each version should produce
identical results when executed, except for operations that are time-dependent, such as asynchronous processing and interrupt-controlled features
of PL/L IBM expects to deliver the F level PL/I
compiler in March, 1966, and the H Level PL/t
compiler in September, 1966.
IBM has provided us with estimated· PL/I compilation
times based on three sample configurations. Table I
describes the sample configurations and indicates
the timing factors that must be summed to arrive
TABLE I: OPERATING SYSTEM/360 PL/I COMPILATION TIME FACTORS
Configuration Components
Sample Configurations
Configuration A
System/360 Model 30 with 65K bytes of core
storage, one Multiplexor Channel, one
Selector Channel, one 2520 Card Read
Punch, one 1403 Printer (Model 3), two
2311 Disk Storage Drives, and the Operating System/360's sequential scheduler. *
Configuration B
System/360 Model 50 with 65K bytes of core
storage, two Selector Channels, two 2311
Disk Storage Drives, and the Operating
System/360's sequential scheduler. **
Configuration C
System/360 Model 65 with 262K bytes of core
storage, two Selector Channels, two 2301
Drum Storage units, four 2311 Disk Storage
Drives, and the Operating System/360's
sequential scheduler. **
Time in Seconds
Configuration A Configuration B
Configuration C
Estimated Timing Factors
Commer- Scien- Commer- Scien- Commer- Sciencial
tific cia!
tific
cial
tific
System overhead for each
independent subprogram
(maximum size of 2,048 bytes)
For each input source card
9/65
16.4
0.47
17.1
0.68
12.3
0.09
13.0
3.1
3.2
0.14 0.026
*
Source program input is assumed to be entered through the Card Read Punch, and
object program instructions and listing are assumed to be written in disc storage.
**
Source program input and object program output are assumed to be read from and
written in disc storage.
0.041
420:167.150
PROCESS ORIENTED LANGUAGE: OPERATING SYSTEM/360 PL/I
.15
Operating System/360 PL/I Compilers (Contd.)
•
Built-in Functions: specific functions (which may
or may not be generic) that are provided in the
PL/I compiler. Built-in functions can return
both scalar and array values to the program
statement, whereas programmer-defined functions can return only a single scalar value.
•
Subroutine Procedure: a procedure block designed to be entered and executed by means of
a special CALL statement. Unlike function
procedures, a subroutine procedure is not
restricted to returning an explicitly-specified
value to the program statement.
at a total time for a PL/I compilation that is performed in conjunction with the Operating System/360.
These estimates are based upon the compilation of
two types of source programs: typical commercial
programs and typical programs for scientific applications. The assumption is made that there is an
average of two cards per executable statement in
the commercial source programs and one card per
executable statement in the scientific programs.
In both cases it is assumed that the number of cards
in the source program is not greater than 200 for
System/360 sample configurations A and B, and
not greater than 800 for configuration C.
.2
PROGRAM STRUCTURE
. 21
Divisions
• Do Group: a collection of statements used to
control program looping .
• Statement: the basic element of the program's
procedures. Examples of various kinds of
PL/I statements follow .
DO I = J TO 10;
A = B + C;
IF A = B THEN GO TO S1; ELSE A = C;
ON OVERFLOW GO TO OVFIX;
PROCEDURE (X, Y, X);
SBPRIM: Z = A**2-A*B+B**2;.
PL/I programs are not broken down into different
types of specification divisions. The characteristics of the files and data can be specified at any
point in the program. No program identity is
established until execution time, so there is no
need for an Identification Division, as in COBOL.
Target computer descriptions are not provided in
the language.
.22
. 23
Procedure Entities
All statements of a P L/I program are organized
into program sections called "blocks. n Program
blocks define the scope of data variables and statement names. They also make possible the dynamic
allocation and release of the core storage required
for data variables within each block.
The procedural entities used in the PL/I language
are listed below.
•
A list of the types of statements in the PL/I
language and their functions is shown in Table II .
Block: a collection of statements that define
a program region. Blocks that are nested in
other blocks are called "internal" blocks, in
contrast to the non-nested "external" blocks.
•
Begin Block: a block that can be executed
wherever it stands in the written program.
•
Procedure Block: a block that can be entered
only through programmed jumps.
•
Function Procedure: a procedure block designed
to be entered and executed as part of the execution of a program statement. Function procedures use the arguments specified in the function
reference in order to return a result to a
referring statement.
• Generic Function: a family of function procedures
with a single name. Reference to a generic
function causes the selection of a certain member of the function family. The member selected
and the characteristics of the value returned
are determined by the attributes of the programmersupplied arguments.
.
©
Data Entities
The data entities used in the PL/I language are
listed below.
•
Structure: a collection of scalar variables,
arrays, and other structures, arranged by
level numbers in a manner very similar to
that of COBOL's Data Division. The outermost or major structure must have a level 1
description; the inner or minor structures
must always have a level number numerically
greater than the structure in which they are
contained.
• Array: an n-dimensional, ordered collection
of elements, all of which must have an identical
data description. The elements of an array
can be non-numeric; they can also consist of
structures.
• Cross Section of an Array: a grouping of all
those elements of an array that fit some general
description. Thus, the cross section of a
3 x 3 array A(i, j, k), written as A(*, 1, 1),
would consist of all elements of A where j
equals k equals 1; i.e., A(1,1,1), A(2,1,1),
and A(3, 1, 1).
• Constant: Any data item whose representation
is both its name and its value. Real arithmetic
constants are allowed in either fixed-point or
floating-point mode, and can be coded in binary
or decimal format. A pure binary constant
consists of a string of ones and zeros, and is
referred to as a bit-string constant. other
forms of constants include character-strings,
. sterling currency figures, and imaginary arithmetic constants (i. e., complex numbers whose
real part is zero).
1965 AUERBACH Corporation and AUERBACH Info, Inc.
9/65
420:167.151
IBM SYSTEM/360
TABLE IT: STATEMENTS USED IN THE PL/I LANGUAGE
STATEMENT
EFFECT
STATEMENT
EFFECT
ALLOCATE
Causes storage to be allocated for specifically controlled data.
FREE
Releases storage that was
reserved by the ALLOCATE
statement.
Assignment
(written
as = sign)
Evaluates expressions and
assigns values to scalars,
arrays, and structures.
GET
Brings data from the current
file and automatically
provides any necessary
editing.
BEGIN
Functions as the heading
statement of a begin
block.
GO TO
Causes control to be transferred to a specified
statement.
Invokes a procedure and
causes control to be transferred to a specified entry
point in the procedure.
GROUP
Releases facilities that
were allocated during the
opening of a file, and
causes proper disposition of
the file.
Releases a group of records
from a program, and
(optionally) skips to
another specified group
of records.
IF
Causes control to be transferred conditionally
according to the value of
an expression.
Provides the attributes for
simple names.
IMPLICIT
Causes execution of the
controlling task to be
suspended for a given
number of milliseconds.
Associates a particular
set of attributes with the
initial letter of the data
name.
LAYOUT
Releases the facilities
allocated to a specific
program which was previously FETCHed and is
now no longer needed.
Specifies the horizontal layout of data on input and
output.
Null Statement
(written as a
semicolon)
Causes no program action.
CALL
CLOSE
DECLARE
DELAY
DELETE
DISPLAY
Causes messages to be displayed to the machine
operator.
ON
Specifies what action is to
be taken when an interrupt
occurs.
DO
Provides controlled looping
facilities.
OPEN
Obtains and prepares files
for subsequent use.
END
Terminates groups and
blocks of program
statements.
PAGE
Specifies the pagination of
files.
POSITION
ENTRY
Specifies a secondary entry
point to a subroutine.
Indicates a variable position
within a file.
PROCEDURE
EXIT
Ends the execution of a task
and, consequently, of any
attached tasks.
Functions as the head of a
procedure, provides its
major entry point, and
specifies the parameters
needed within the procedure.
FETCH
Brings a program from the
program library into central memory and makes it
available as a procedure
through use of a CALL
statement.
PUT
Inserts data into the current
file, and automatically
provides any desired data
editing.
READ
Causes data to be transmitted
from an external medium
to internal storage.
FORMAT
Specifies a format list for
use with data transmitted
under format direction.
(Contd.)
9/65
420:167.230
PROCESS ORIENTED LANGUAGE: OPERATING SYSTEM/360 PL/I
TABLE II. STATEMENTS USED IN.THE PL/I LANGUAGE (Contd.)
STATEMENT
EFFECT
STATEMENT
EFFECT
REPOSITION
Moves back the position of
the pointer by one data
field; normally used when
an error has occurred.
SIGNAL
Simulates the occurrence of
an interrupt condition.
SKIP
Causes a specified number of
records or lines to be
skipped.
SORT
Specifies that the records on
a particular file are to be
sorted.
SPACE
Causes a specified number
of records or lines to be
spaced between each
explicitly specified record
or line.
STOP
Causes the immediate
termination of a program.
WRITE
Causes data to be transmitted from internal
storage to an external
storage medium.
RESTORE
RETURN
REVERT
.23
.24
Causes data previously
saved in auxiliary storage
to be returned to internal
storage.
Ends the execution of a procedure, and returns control to the invoking procedure.
Nullifies the specified
treatment of an interrupt
procedure within the same
program block.
SAVE
Places data in auxiliary
storage for possible later
return to main storage.
SEGMENT
Defines positioning within
a segmented file.
Data Entities (Contd.)
• 25
o Variable: any named data item whose value may
or may not change during the execution of the
program. Variables can consist of the same
data types as constants: real arithmetic, bitstring, character-string, imaginary, and.
sterling.
· 251 Qualified names
Names
. 241 Simple names
A simple name is an identifier of a file, statement,
structure, entry, array, etc. Simple names must
start with an alphabetic character and can include
up to 31 characters. Blanks can be included within
a name provided that two blank characters are not
placed next to each other.
. 242 DeSignators
,
\
'-
PL/I uses very few signals or codes to designate
the characteristics of entities in the source
language. Most frequently, specific declarations
of the attributes of each entity are used; these
declarations are supplemented by default assump'"
tions in the compiler. The designators used by
PL/I include the following:
o
/* ........ */ indicates a comment.
" Names starting with I, J, K, L, M, or N
are assumed to refer to fixed-point real
binary values unless otherwise defined.
Structure of Data Names
A qualified name is a compound name, with its
components separated by periods, that is used
to avoid ambiguities in cases where a particular
simple name may be used within two or more
different structures. Any name of a variable,
array, or structure can be qualified by preceding
the name of the item with the names of one or
more of the containing structures. The structure
names must be sequenced from left to right in
order of increasing level numbers. The sequence
of names does not need to include all of the containing structures, but it must include sufficient
qualification to resolve all potential ambiguities.
.252 Subscripted names
A subscripted name is the name of an array
followed by a list of subscripts. The subscripts
are separated by commas and enclosed in parentheses, and the number of subscripts must equal
the number of dimensions of the array. By means
of subscripted names, the elements of any array
can be individually referenced.
· 253 Subscripted qualified names
A subscripted qualified name is a sequence of
simple and subscripted names separated by periods
and used to identify potentially ambiguous elements
of an array.
• 254 Synonyms
\
" Names starting with any other letter are
assumed to refer to floating-point real binary
values unless otherwise defined.
Two or more labels can be assigned to a statement
and used interchangeably. Through the use of the
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
9/65
IBM SYSTEM/360
420: 167.254
.254 Synonyms (Contd.)
•
DEFINE attribute, named data can be assigned
to occupy the same storage area already assigned
to another named data item. Thus the DEFINEd
data can be referred to by either of the two data
names.
.26
Number of names
• The IMPLICIT statement provides that any name
beginning with a specified alphabetic letter or
letter range shall have certain attributes, such
as: IMPLICIT C BINARY COMPLEX;. This
statement provides that all variableswhosenames
start with C should be considered as binarycoded complex numbers .
Any limitation on the number of names available in
a specific PL/I implementation will be a characteristic of the implementation rather than of the
language.
. 27
Region of Meaning of Names
Data items can be assigned hierarchical levels,
permitting the programmer to refer to groups of
the fields with a single name. The level numbers
are written immediately before the name of the
data items concerned, such as: DECLARE 1
CARDIN, 2 NAME, 2 WAGES, 3 NORMAL, 3
OVERTIME;. This statement declares that
the data record CARDIN consists of the variables
NAME and WAGES, and that the variable WAGES
itself consists of the variables NORMAL and
OVERTIME.
.271 Universal names
The only universal names are those identifying the
built-in functions and the file names.
• 272 Local names
Most names used in PL/I are local names. Names
of items are declared (either explicitly or implicitly)
for use within a specific program block. The name
will be reserved for the specific item throughout
the block, and throughout any program blocks that
are nested within the original block. However,
should a further explicit declaration of an identical
name be made in a nested program block, then this
second declaration supplants the original meaning
of the name within the program block where the
second declaration was made and within any
program blocks nested within that block.
The area of the source program within which a
name refers to the same specific entity is called
the "scope" of the name.
The PICTURE attribute of the DECLARE statement defines the internal and external formats
of numeric or character-string data, and specifies the editing of data, such as: C PICTURE
'XAA. AA' . The item named C is here described as a six-position field of characters (X),
letters (A's), and a decimal point.
. 32
Files and Reels
Data files are described in the PL/I language by
ascribing attributes to them in DECLARE statements. The MEDIUM and~SAGE attributes
specify device-dependent information for selected
input-output files. The precise specifications of
these attributes will be defined in individual P L/I
implementations. Other file attributes include:
•
STANDIN or STANDOUT: defines a file as
being the System Standard Input file or the
System Standard Output file .
. 273 Non-local names
Any name that is used in two independent program
blocks is assumed to refer to two different items
unless the name is made non-local by an explicit
"External" declaration. Names that are desired
to be non-local must be declared to be External
in every block from which they may be referenced.
All file names are considered to be External names.
Any other names are considered to be Internal- --(local to the program block in which they are declared and any nested program blocks) unless
specifically declared to be External.
•3
DATA DESCRIPTION FACILITIES
.31
Methods of Direct Data Description
Data is described in the P L/I language by directly
ascribing "attributes" (specific characteristics) to
each data item. These attributes can be specified
in three ways:
• The DECLARE statement specifies attributes
of simple names, such as: DECLARE A FLOAT
(3), B REAL (10) FLOAT. Two floating point
variables named A and B are defined as requiring
3 and 10 digits of precision, respectively. B is
further defined as !t real arithmetic variable.
• BLOCK: defines the file's blocking factor, the
maximum length of the block, and whether the
records are of fixed or variable length.
• DISCARD or KEEP: defines the final disposition
of a file.
• POOL: specifies that the same buffer areas
can be used for two input-output files.
The CONSECUTIVE, REGIONAL, INDEXED,
and COMMUNICATIONS attributes define the
organization of the data records within a file and
the manner in which these records are located .
REGIONAL indicates that the file is located
within speCified direct-access-device cylinders;
INDEXED specifies that an ordered index is
used to locate records within a file; and
COMMUNICATIONS specifies that records are
obtained from the file in a sequence determined
by a queue. The SEQUENTIAL and DIRECT
attributes specify the manner in which records
within a file are accessed. Only REGIONAL or
INDEXED files can be accessed in DIRECT
fashion.
Standard file and reel labels can be checked
through use of the !DENT option of the OPEN
statement.
(Contd.)
9/65
PROCESS ORIENTED LANGUAGE: OPERATING SYSTEM/360 PL/I
. 33
Records and Blocks
.4
OPERATION REPERTOIRE
PL/I allows the records in a file to be of fixed or
variable length, and to be arranged in groups of
physical blocks and segments as designated by the
programmer. (A segment is a group of records
divided by an arbitrarily assigned segment symbol.)
The language places no limits on block sizes,
although individual compiler implementations may
be more restrictive.
.41
Formulae Used for Numeric Computations and
Comparisons
No facilities are included in the PL/I language for
controlling the input-output error recovery
methods, so the standard routines included in the
Operating System/360 will probably be utilized.
. 34
Data Items
The classes of data items are designated explicitly
by DECLARE and IMPLICIT statements within the
program. Should these designations be incomplete,
"default" assumptions or interpretations are made
by the compiler based on the context. (Every
attribute of a variable and every optional language
specification in PL/I has a predefined default
interpretation.) The various classes of data items
permitted in the language are listed below:
Integer: . . . . . . . only as a sub-class of fixedpoint.
Fixed-point: . . . . either in binary or decimal
coding.
Floating-point: ... either in binary or decimal
coding.
Alphabetic: . . . . . either as a constant or in a
character string.
Alphameric: .... either as a constant or in a
character string.
Binary: . . . . . . . . either as a constant or in a
binary string.
Imaginary: . . . . . represents a complex value
whose real part is assumed
to be zero.
. 35
.411 Operator list
+ •••••••••••••• addition.
- . . . . . • • • . • . . • . subtraction.
* .............. multiplication.
/ . • . • . . . . . . • • . . division.
** .............. exponentiation.
1 . . . . • • . . . • . . . . NOT (used for bit-strings).
& .••••••••••••• AND (used for bit-strings).
I .............. OR (used for bit-strings).
< •..•.•..••.... compare for "less than"
condition .
<= •......•.... ; . compare for "less than or
equal to" condition.
= .............. compare for "equal to"
condition.
1= . . . . . • . • . . . • . . compare for "not equal"
condition.
>= .••••••••••••• compare for "greater than
or equal to" condition.
> . • • . . • . . . . . • • . compare for "greater than"
condition.
II .............. concatenate into a single
bit-string.
. . . . . . . . . . . . . . replace by.
.412 Operands allowed
PL/I allows programmers to use operands that
have different arithmetic modes and different
code representations within the same expression.
These operands can consist of literals, arrays,
and structures. Fixed-point and floating-point
numbers, real and complex numbers, binary and
decimal-coded operands can be mixed as desired.
The necessary conversions will be inserted
into the object coding both before the expression
is evaluated and after evaluation, just prior to
storing the result .
Data Values
The PL/llanguage places no limit upon the values
of the various data items. Specific PL/I implementations and subsets may be more restrictive.
.36
420: 167.330
Special Description Facilities
Several facilities are provided in the PL/I language
to permit data to be described in unusual ways.
Among these facilities are the following data
attributes:
•
LIKE: specifies that the name being described
is to have the same data structure as another
named item.
• DEFINED: specifies that a name data item is
to occupy the same storage area as that already
assigned to another data item. The DEFINED
data can then be referred to by either of two
equated names.
All data tables are expliCitly DECLAREd to be
arrays, whether or not the elements of the array
are numeric items. Thus the array manipulation
features of the PL/I language can be utilized with
any type of tabular information.
©
Various combinations of arrays and structures
can be included within an expression. When only
one of the operands of an expression is an array or
a structure and the other part is a single value,
the expression is evaluated for each element of
the array or value. For example, the expression
3 + A, where A is the two-dimensional array
568
3 4 7,
is evaluated by adding 3 to each element of A,
yielding another two-dimensional array:
8 9 11
6 7 10.
When both operands of an expression are arrays or
structures, they must have identical descriptions
in order for the expression to have meaning. In
this case, the corresponding elements of each
operand are used to evaluate the expression. For
example: A + B, where both A and B are structures
of one floating-point number followed by three
fixed-point numbers, will result in another structure of the same description, each element of
which will be the sum of the corresponding elements
of the original structures.
1965 AUERBACH Corporation and AUERBACH Info, Inc.
9/65
IBM SYSTEM/360
420:167.413
.413 Statement structures
.422 Logical operations
PL/I language statements can include expressions
that involve any combination of operators. For
instance, the expression (A - B**3)/(C*DIIE) is a
valid statement although it includes concatenation
as well as conventional arithmetic operators. The
extent to which operators can be mixed within an
expression appears to be governed simply by the
ability of the expression to produce a meaningful
result, rather than by any arbitrary rules of the
language.
In evaluating an expression, the operations are
performed by the compiler in a strict sequence that
depends upon the priority levels of the operators
(see Table TIl below) and their positions in the
expression. When an expression includes more
than one operator at a particular level, the operations are performed from left to right, just as the
expression is written, except in the case of the top
priority level, where the scan proceeds from right
to left.
The bullt-in functions provide for AND and INCLUSIVE OR operations on all the elements of an
array, yielding a bit-string as long as the longest
element in the array. No standard facilities are
included for EXCLUSIVE OR and NOT operations
on arrays.
.423 Scanning
The built-in function SCAN produces a new array
with one dimension less than the original array.
The function value of the SCANned array is determined by a decimal integer and operator parameters. Provision is made for the use of any
given operator during the creation of the new
array from the elements of the original array.
.424 other array operations
Arrays form an integral part of the data structure
used by PL/I, and a large number of array operations are provided. These include built-in functions for locating the current higher bound, the
current lowerbound, and the current extent of a
given dimension of an array.
TABLE III: PRIORITY LEVELS OF OPERATORS
IN THE PL/I LANGUAGE
LEVEL
OPERATORS
Levell
** , prefix +, prefix -
Level 2
*, /
Level 3
infix
Level 4
>=, >,1 =, =, <, <=
TABLE IV: LIST OF BUILT-IN FUNCTIONS
Level 5
1
FUNCTION NAME
Level 6
&
Level 7
I
GENERIC
ARITHMETIC
FUNCTIONS
Level 8
II
ABS (s)
Absolute value of x.
MAX (x, y ..• )
Value of maximum
argument.
MIN (x, y ... )
Value of minimum
argument.
MOD (x, y)
x-FLOOR (xiy)*y (see
"FLOOR" below).
SIGN (x)
1 if x> 0, 0 if x
-1 if x < O.
FIXED (x, y, z)
x converted to a y-digit
fixed-point value with z
positions following the
decimal or binary point.
FLOAT (x, y)
x converted to a floatingpoint value with y digits
precision.
FLOOR (x)
Largest integer not
exceeding x.
+, infix -
.414 Results representation
The results of an operation can legitimately be
-.:'epresented in more than one manner, such as
rounded or truncated. PL/I permits the programmer to assign specific attributes to final results
through the attributes of defined variables. However,
when the PL/I compiler evaluates mixed-mode
expressions, it represents all intermediate results
according to a specific set of rules. These rules,
illustrated in Table IX of this report, will be of
considerable importance to anyone making use
of mixed-mode expressions.
.42
.43
Other Computation
Numerous built-in functions are provided in the
PL/I language. These functions normally allow
their operands to be either constants, variables,
arrays, or structures. Table IV lists the various
built-in functions provided in the PL/I language.
Operations on Arrays
.421 Matrix operations
PL/I's built-in functions provide for the addition,
subtraction, and multiplication of matrices. No
standard facilities are included for inversion,
normalization, eigen roots or eigen vectors,
correlation or probability moments.
FUNCTION VALUE
=
0, and
(Contd.)
9/65
420: 167.430
PROCESS ORIENTED LANGUAGE: OPERATING SYSTEM/360 PL/I
TABLE IV: LIST OF BUILT-IN FUNCTIONS (Contd.)
F UNCTION NAME
FUNCTION VALUE
FUNCTION NAME
FUNCTION VALUE
CEIL (x)
Smallest integer not
exceeded by x.
COSH (x)
Cosh (x).
SINH (x)
Sinh (x).
TRUNC (x)
FLOOR (x), if x ~ 0; otherwise CEIL (x), if x < O.
ATANH (x)
Arctanh (x).
ATAN (x, y)
Arctan (x/y).
BINARY (x, y, z)
x converted to binary base;
y and z specify the
precision of the result.
DECIMAL (x, y, z)
x converted to decimal
base; y and z specify the
precision of the result.
PRECISION (x, y, z)
x converted to precision
specified by y and z.
ADD (w, x, y, z)
wadded to x; y and z
specify the precision of
the result.
STRING GENERIC
FUNCTIONS
BIT (x, y)
x converted to a bit string of
size y.
CHAR (x, y)
x converted to a character
string of size y.
SUBSTR (x, y, z)
Substring of string x,
starting at position y with
length of z.
INDEX (x, y)
Decimal integer specifying
the index of the initial
position of string y as a
substring contained in
string x.
HIGH (x)
Character string of length
x, composed of the highest
characters of the data
character set.
LOW (x)
Character string length x,
composed of the lowest
characters of the data
character set.
REPEA T (x, y)
String x repeated y times.
UNSPEC (x)
Internally coded representation of x.
LENGTH (x)
Decimal integer with length
of x.
MULTIPLY (w, x, y,
z)
w multiplied by x; y and z
specify the precision of
the result.
DIVIDE (w, x, y, z)
w divided by x; y and z
specify the precision of
the result.
COMPLEX (x, y)
Complex number with x
as the real part and y as
the imaginary part.
REAL (x)
Real part of complex
number x.
IMAG (x)
Imaginary part of complex
number x.
CONJG (x)
Conjugate of x.
EXP (x)
Exp (x).
LOG (x)
Log (x).
LOGlO (x)
Log10 (x).
LOG 2 (x)
Log 2 (x).
ARRAY FUNCTIONS
*ATAN (x)
Arctan (x).
SUM (X)
The sum of all the elements
of X.
*TAN (x)
Tan (x).
PROD (X)
*SIN (x)
Sin (x).
The product of all the
elements of X.
*COS (x)
Cos (x).
ALL (X)
TANH (x)
Tanh (x).
A bit string of the maximum
length of any element of
X, with a 1 wherever all
the elements of X are 1.
ERF (x)
(2/,/ir)
./fX)
ANY (X)
SQRT (x)
ERFC (x)
1 - ERF (x).
A hit string of the maximum
length of any element of
X, with a 1 wherever any
of the elements of X are 1.
o
fX EXP (_t 2) dt.
(always positive).
(Contd.)
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
9/65
420: 167.440
IBM SYSTEM/360
TABLE IV: LIST OF BUILT-IN FUNCTIONS (Contd.)
FUNCTION NAME
FUNCTION VALUE
FUNCTION NAME
POLY (X, Y)
X (M:N) and Y (P:Q) are
vectors; the result is:
OTHER BUILT-IN
FUNCTIONS
N-M
2:
N-M
(X(M+J) * II Y (p+I)).
J=O
I=J
LBOUND (X, S)
Current lower bound of the
Sth dimension of X.
'HBOUND (X, S)
Current higher bound of the
8th dimension of X.
DIM (X, S)
Current extent of the Sth
dimension of X.
SCAN (A, I, operator)
,Specialized function; see
text, Paragraph. 423.
CONDITION BUILT-IN
FUNCTIONS
ONPOINT
I/O buffer pointer position
when interrupt occurred.
ONLOC
Procedure name in which
interrupt occurred.
ONFIELD
Contents of field being
processed when I/O
interrupt occurred.
ONCHAR
Character which caused
an I/O conversion error.
ONCODE
Code character identifying
type of error causing
interrupt.
*
• 44
FUNCTION VALUE
DATE
Date, in form YYMMDD,
using year (y), month (M),
and day (D).
TIME
Time, in form HHMMSSTTT,
using hours (H), minutes
(M), seconds (S), and
milliseconds (T).
A LLOCA TION (X)
"I" if storage has been
allocated for major structure X; otherwise "0 . "
POINT (Filename)
Position within current
logical record.
COUNT (Filename)
Number of data items transmitted during the last
read or write operation.
COMPLETE (Task
Identifier)
"I" if task has been completed; otherwise "0."
ROUND (Expression,
Constant)
A fixed-point operand
rounded at a specific
point; or a floating-point
operand with any bias
removed.
STRING (Structure
Name)
Concatenation of all the
structure elements.
A separate function is available for operands expressed in terms of degrees,
rather than radians; e.g., ATAND (X) •
Data Movement and Format
The assignment statement is used to copy data from
one location to another and to evaluate expressions.
It is written by using the equality sign (=) rather
than the formal word ASSIGN. Operands on the
left of the equality sign are set to the value of the
operand on the right.
Single values, arrays, and data structures can be
used as operands in assignment statements, and
more than one result location is possible; for
example, A, B, C, D = E will set each of the
entities A, B, C, and D equal to E. All the operands
of a single assignment statement should be of the
same class; i. e., either all arrays, all single
values, or all structures.
When arrays and structures are used as operands
in assignment statements, it is not necessary
that the descriptions of the various operands be
the same. The data is treated as a stream of
scalar values: the first value is stored in the
first available position, the second value in the
second position, and so on.
A special BY NAME option is available for use
with array and structure assignments. This
option permits the selective assignment of only
the correspondingly named and ordered values
within each array or structure; it is conceptually
equivalent to the CORRESPONDING option in
COBOL.
During the movement and copying of data, changes
of radices and data class occur automatically. The
result fields have the characteristic attributes
that have been declared or implied according to
the standard rules.
The insertion and removal of editing symbols,
floating of dollar signs, etc., are normally handled
as part of the READ or WRITE statements. However,
if the programmer wishes to have these operations
performed during the execution of an assignment
statement, insertions of editing symbols can be
(Contd.)
9/65
420: 167.441
PROCESS ORIENTED LANGUAGE: OPERATING SYSTEM/360 PL/I
.44
Data Movement and Format (Contd.)
.47
made by using the PICTURE attribute for the
result location; deletions will require a separate
concatenation of sub-fields operation.
During the READ or WRITE operations, a Format
List can supply a list of data values. Two format
modes can be specified in a PL/I Format List:
external format mode, which is designed to be
readable through use of character representation;
and internal format mode, which is coded and
individually defined for each PL/I implementation.
In addition, the actual formats of various items can
be specified either by standard notation or by the
PICTURE attribute. Table VII show" the PL/I
standard notation for the various data formats,
and Table VIII lists the PICTURE characters and
defines their use; both tables are included in the
Appendix to this report.
The PL/I language allows conditional tests to be
made for a number of object-time conditions. This
facility permits specific program actions to be
taken in the event of object-time errors. Whenever
the programmer does not provide specific actions for
various types of object-time error conditions, the
PL/I language provides that standard actions be
taken. The programmer can also specify that the
standard action is to be taken in addition to his own
programmed action. Table V shows various objecttime error conditions and the corresponding actions
that will be taken by the system unless these standard
actions have been overruled by the. programmer.
.5
PROCEDURE SEQUENCE CONTROL
.51
Jumps
File Manipulation
By means of the GO TO statement, the PL/I programmer can write jumps to any labelled statement
within the current program block or any internally
nested blocks. Jumps to accessible external program
blocks are normally handled by returning to a prespecified entry point.
Four types of file manipulation are provided in
PL/I; they are distinguished by the means used to
direct the transmission of data. The four methods
are:
Multiple switching is not explicitly included in the
PL/I language, but it can be programmed by using
variables as statement labels, by using subscripted
label variables in a GO TO statement, etc.
Data movement between main and auxiliary storage
is facilitated through use of the SAVE and RESTORE
statements.
.45
•
List-Directed Transmission: the user supplies
a list of the storage areas to be used.
•
Data-Directed Transmission: input data includes
information defining the storage areas to be
used, and output data includes the name of the
data being transmitted.
•
Procedure-Directed Transmission: a CALLed
program procedure within the READ or WRITE
statement directs the transmission and manipulation of each data field.
•
Format-Directed Transmission: a format list
is used to define the form of the data being
transmitted, and a separate data list is used
to define the program storage area which is to
be used for the data.
. 52
Positioning of a file within or between data records
can be accomplished through use of the SKIP,
SPACE, GROUP, and SEGMENT statements. The
POSITION and TAB statements are supplied specifically to facilitate report writing, but they can be
used as necessary by other file operations; these
two statements control field positioning within data
files.
.46
Operating Communication
PL/I provides tbe facility to display messages to
the machine operator through use of DISPLAY
statements. An internal machine log is maintained
automatically, but no language provision is described
to permit program-controlled log entries.
©
Conditional Procedures
The IF and ON statements control conditional program
jumps according to dynamic execution-time conditions.
In the IF statement, an expression is evaluated, and
the subsequent program flow depends on the result
of the evaluation. IF statements can be nested within
other IF statements, as shown in the following example,
where two IF statements are nested within IF statement A.
A: IF X> Y THEN
IF Z =W THEN
IF W < P THEN Y = 1;
ELSE P = Q
The READ and WRITE statements specify the type
of file manipulation to be used. The GET and PUT
statements are used only when accessing individual
fields during the transmission of proceduredirected data.
''''--
Object Program Errors
ELSE X =4;
The ON statement controls conditional jumps that
are based on conditions such as error status,
program check-out conditions, and programmerdefined conditions. Conditional jumps can also
be effected through use of the logical AND, NOT,
and INCLUSIVE OR operators.
.56
Loop Control
Loop control in the P L/I language is handled by the
DO statement. In its general form, the DO state~
ment includes both controlled iteration with incrementation and loop-inhibiting conditions. The
control operands of the DO loop can be any legal
expressions of the language. Condition testing is
performed at the start of each iteration.
The statement "DO COUNTER = 0 TO 100 BY 5
WHILE A = B" will initiate a loop based on the
1965 AUERBACH Corporation and AUERBACH Info, Inc.
9/65
IBM SYSTEM/360
420: 167.560
TABLE V: STANDARD
ERROR-HANDLING TECHNIQUES
CLASSIFICA TION OF
CONTINGENCY
DEFINITION OF
CONTINGENCY
COMPUTATIONS
Error occurring during conversion
of data from one type to another
(CONVERSION) •
Undefined.
Place comment in log;
continue with program.
Results of fixed-point arithmetic
exceed the machine capability
(FIXEDOVERFLOW) .
Most significant
digits are lost.
Place comment in log;
continue with program.
Exponent of a floating-point number
exceeds maximum permitted value
(OVERFLOW) •
Maximum allowable Place comment in log;
terminate program.
positive number
is substituted.
A fixed-point variable is too large
for the assigned size of the data
field (SIZE).
Most significant
digits are lost.
Place comment in log;
terminate program.
Exponent of a floating-point number
is smaller than the permitted
minimum value (UNDERFLOW).
Smailest positive
non-zero value
is substituted.
Place comment in log;
continue with program.
Zero has been used as the divisor
in division operation.
Undefined.
Place comment in log;
terminate program.
Input-output error, control program
error, or conversion error has
prevented successful record access
(ACCESS).
Record not
available to
program.
Place comment in log;
terminate program.
INPUT/OUTPUT
CONDITIONS
• 56
OBJECT~TIME
STANDARD SYSTEM
ACTION
An illegal character has occurred in
the output data (EDIT).
-
Place comment in log;
terminate program.
Reading past a group delimiter on a
file has been attempted (ENDGROUP).
-
Place comment in log;
terminate program.
Reading past a record delimiter on a
file has been attempted (ENDRECORD)
-
Place comment in log;
terminate program.
The data for an output field cannot fit
into the assigned space (FIELDOVERFLOW).
-
Place comment in log;
terminate program.
Label checks on a file have failed
(IDENT).
-
Place comment in log;
terminate program.
The name of a specified file cannot
be recognized (NAME).
-
Place comment in log;
terminate program.
A record that has been requested by
its key cannot be found (SEARCH).
-
Place comment in log;
terminate program.
An error has occurred during the
transmission of a data file
(TRANSMIT) .
-
Place comment in log;
terminate program.
A requested file is unavailable
(UNDEFINED FILE) •
-
Plac·e comment in log;
terminate program.
Loop Control (Contd.)
value of COUNTER, which starts at zero and is
incremented by 5 after each iteration. The operation will be ended either when the value of COUNTER
passes 100 (either by means of the increment within
the DO statement or by a modification of the counter's
value in the body of the DO loop), or when A is not
equal to B. When the DO statement listed above
9/65.
RESULT
is encountered and A is not equal to B, the entire
DO loop is bypassed.
The values of loop-controlling operands can be
made to depend upon the satisfaction of a series
of conditional values within the DO statement.
Thus, if certain conditional values are never
satisfied during the looping operations, iteration
will continue indefinitely.
(Contd.)
420:167.561
PROCESS ORIENTED LANGUAGE: OPERATING SYSTEM/360 PL/I
.56
ment. The program can then be entered by means
of the CALL statement.
Loop Control (Contd. )
DO loops can be nested to any desired extent, and
control jumps can be made from any point within a
basic loop or nested loop to any point in the containing program.
.6
OTHER LANGUAGE FACILITIES
· 61
Program Check-out
.8
ThePL/Ilanguage includes ABNORMAL, USES, and
SETS attributes which describe unusual uses of procedures and variables to assist the compiler in
optimizing the object program. Optimization of
file handling is assisted by the ACTIVITY option
of the OPEN statement, which indicates the relative
activity of the file that is being OPENed. Details
concerning the utilization of these language facilities
in specific PL/I implementations are not available
to date.
P L/I includes a number of facilities designed for
use during program testing. Two statements are
used for this purpose:
•
•
.62
ON: specifies the action to be taken when an
interrupt occurs for the named condition.
ON-conditions allow monitoring of data transfers, and testing of particular instructions
and the ranges of subscripts that control
loops and data structures.
The P L/I language does not directly provide for a
description of the target computer's environment,
although the file descriptions indicate the storage
areas involved in certain types of direct assessing,
There are no source-language provisions for controlling the amount or type of documentation produced by the compiler.
SIGNAL: causes the simulation of external
interrupt conditions. SIGNAL permits
simulation of overflow, underflow, and other
error conditions to facilitate testing of the
program paths used under these abnormal
conditions.
The storage level of a particular independent data
structure or array can be designated in the PL/I
language as SECONDARY, a permanent attribute
that results in the assignment of a less-efficient
storage area to the data item. Transmission of
data items from main to auxiliary storage is
accomplished by means of the SAVE statement;
retrieval can then be effected by the RESTORE
statement.
Program Modification
A separate language,· the PL/I Macro Language,
allows a P L/I program to be constructed from a
skeleton source text immediately before compilation. This Macro Language allows parameterization and modification of source statements depending
upon the value assignments supplied for the macro
variables. The facilities of the Macro Language
also permit semi-automatic creation of repetitive
PL/I programs.
• 63
The allocation of primary core storage is controlled by the assigned attributes of the various
data items. Storage can be allocated permanently
by means of the STATIC attribute, or temporarily
- for the duration of the execution of the program
block containing the data item - by means of
the A UTOMATIC attribute. In addition, the programmer can retain control of the allocation and
freeing of core storage for individual items by
specifying the CONTROLLED data attribute and
then alternately issuing the ALLOCA TE and FREE
statements.
Asynchronous Execution of Different Tasks
In cases where facilities for multiprogramming
and/or multiprocessing are available, PL/I allows
logically independent parts of a program to be
executed asynchronously, and thus, in some cases,
more economically. For example, it might well
be advantageous to execute a sorting operation concurrently with a lengthy computational operation in
order to keep both the central processor and the
peripheral devices productively occupied. Each
of the independent parts of an overall program is
called a "task," and each task can have subordinate tasks "attached" to it. Any operational
task can initiate execution of an attached task, can
check on whether the asynchronously-initiated
attached task has been completed, and can wait
for its completion before proceeding.
· 64
.7
The arrangement of data items within core storage
can be controlled by using the ALIGNED and
PACKED attributes. These attributes determine
whether individual strings of data items will start
at word boundaries or be packed into contiguous
character positions of core storage.
Separate input-output areas are normally used
for each file, and they are assigned in accordance
with the file description statements in the program.
Common buffer areas can be established by means
of the POOL option of the DECLARE statement,
but the programmer must assume responsibility
for the successful use of the shared input-output
areas.
Extension of the Language
No provision for enabling PL/I programmers to
create new language facilities (as distinguished
from special-purpose functions or subroutines) has
been defined to date.
LffiRARY FACILITIES
All programs written in the PL/I language can be
stored in an on-line program library. Any program can be brought into core memory and prepared for execution by means of the FETCH state-
©
TRANSLA TOR CONTROL AND STORAGE
ALLOCATION
.9
APPENDIX
This section of the PL/I language report includes
tables and charts that provide supplementary
information about the various facilities of the
language. Table X provides a brief summary of
the principal features of PL/I as compared with
those of FORTRAN and COBOL.
1965 AUERBACH Corporation and AUERBACH Info, Inc.
9/65
420: 167. 900
IBM SYSTEM/360
TABLE VI: KEY TERMS IN PL/I LANGUAGE DESCRIPTIONS
TERM
DEFINITION
Array
An array is a special form of a data structure, each element
of which has the same characteristics. This definition
includes, but is not restricted to, conventional mathematical
arrays.
Attribute
Attributes are keywords that specify characteristics (such
as DECIMAL, FIXED, RECURSIVE, SEQUENTIAL, etc.)
of the various data elements and procedural parts of the
language.
Concatenation
Concatenation is the operation that strings together characters
so as to create new character strings of alphabetic, decimal,
or bit-form information.
Cross Section (of an
array)
A cross section of an array consists of all the elements in the
array that have specified common subscripts. The full
subscript range is not specified, since some levels can
be written as asterisks.
Default Interpretation
Default interpretations are standardized assumptions made
by individual PL/I implementations in the absence of the
programmer's explicit specifications of data attributes,
error procedures, and keywords of statements. These
assumptions are called default attributes.
hnplied Attributes
Implied attributes are sets of characteristics that the programmer associates with the first letter of the data name.
They are written expliCitly into the program in an
IMPLICIT statement and are then implied throughout
the program. Implied attributes should be differentiated
from default attributes.
Individual hnplementation
Individual PL/I implementations are specific versions of the
PL/I compiler that have characteristics (such as specific
default interpretations) not defined in the language itself.
Macros
Macros in PL/I are not generative instructions that insert
many instruction codes into an object program; PL/I
macros are instructions in the PL/I Macro Language
that are used to modify a source language program
immediately before compilation.
Structure
A structure is a collection of data, each element of which
is individually described. This is a recursive definition,
permitting structures of structures.
Task
A task is an individually-executed part of a program or
collection of programs.
Variable
A variable is any named data item, including those whose
values are constant throughout the program.
",
/
(Contd.)
9/65
420: 167.901
PROCESS ORIENTED LANGUAGE: OPERATING SYSTEM/360 PL/I
TABLE Vll: STANDARD FORMAT ITEMS AND METHODS OF DESCRIPTION USED IN
PL/I DATA FORMAT LISTS
Format Items
Description for Internal
Representation (1)
Description for External
Representation (1)
Fixed-Point Format
IF (precision) - for decimal data
IFB (precision) - for binary data
F (w, d, p)
Floating-Point Format
IE (precision) - for decimal data
IEB (precision) - for binary data
E (w, d, s)
Complex Format (2)
IC (some internal real format
item)
C (one or two real
format items)
Picture Format (3)
IP 'picture-specification'
P 'picturespecification'
Bit-String Format
B (length)
A (w)
Character-String
Format
A (w)
A (w)
General Format
IG - see note (4)
G (w, d, s)
Remote Format (5)
R (statement-label of Format
List to be used)
-
Spacing Format
X (x)
Positioning Format (6)
SPACE; SKIP; GROUP; or TAB with or without an expression
-
or
\
POSITION (format List)
-
Legend
w = the length of the field in characters including signs, decimal or
binary points, and the letters E and B in the representation of
constants.
d = the number of positions after the decimal or binary point.
p = the scale factor, which may be positive or negative.
s = the number of significant digits to appear.
Notes
(1) The entire format specification need not be used. For example,
F (w) represents an integer value.
(2) Complex numbers are represented internally as imaginary values
with assumed zero real parts; complex numbers with non-zero
real parts must be written as expressions.
(3) See Table VllI for details of the picture specifications.
(4) In the IG representation, the internal format is the same as the
external.
(5) The Remote Format Item is used if it is desired to locate format
items remotely from a prespecified FORMAT statement.
(6) These format items act in the same way as the statements of the
same names.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
9/65
IBM SYSTEM/360
420: 167. 902
TABLE VIlI: PICTURE CODES USED IN PL/I EDITING OPERATIONS
CODE
A
FUNCTION
Specifies that the associated field of a character string may contain any
letter or blank.
B
Inserts a blank in the associated field position.
CR
Specifies that the letters "CR" should appear if the associated field value
is negative.
D
Specifies that the sterling pence indicator "d" should be inserted.
DB
Specifies that the letters "DB" should appear if the associated value is
positive or zero.
E
Inserts the letter "E" to indicate an exponent value.
F
Specifies the location of a decimal or binary point in a fixed-point number.
G
Specifies the start of a sterling currency picture.
H
Specifies that the sterling shilling indicator "s" should be inserted.
Specifies that the + overpunch should appear if the associated field is a
positive value.
K
Specifies that the exponent subfield should be assumed to follow the point
in the field associated with K.
R
Specifies that the - overpunch should appear if the associated field is a
negative value.
S
Specifies that the characters + or - should appear depending on the value
of the associated field.
T
Specifies that an overpunch will be inserted to indicate the value of the
associated field.
v
Specifies that a decimal or binary point should be assumed to appear at
this point in the associated field.
x
Specifies that the associated field of a character string may contain any
character.
y
Specifies suppression of all zeros and replacement by blanks.
z
Specifies suppression of leading zeros and replacement by blanks.
1
Specifies that the associated field position contains a binary digit.
2
Specifies that the associated field position contains a binary digit within
a two's-complement coded field.
3
Specifies that the associated field position contains a binary digit within
a one's-complement coded field.
6
Specifies the position of a sterling currency pence character in the IDM
single-character representation.
7
Specifies the position of a sterling currency pence character in the British
Standards Institution representation.
Specifies the position of a sterling shilling character in the British
Standards Institution representation.
Specifies that the associated field position will contain any decimal
digit.
*
Specifies that the asterisk should replace leading zeros.
$
Specifies that the dollar sign should be inserted. When used in strings
of two or more dollar signs, it is the editing character for a floating
dollar sign.
+
Specifies that a + character should be inserted if the field value is positive.
Multiple plus characters indicate a single floating character.
Specifies that a minus sign should be ins erted if the field value is
negative. Multiple minus characters indicate a single floating character.
Specify that these same characters will appear in the associated field
positions. In case of leading zero supression, these characters are
also suppressed.
(Contd.)
9/65
PROCESS ORIENTED LANGUAGE: OPERATING SYSTEM/360 PL/I
420:167.903
TABLE IX: PL/I'S METHODS OF HANDLING INTERMEDlA TE RESULTS OF EXPRESSION
EVALUATION
CASE
PROBLEM
P L/r TREATMENT
Integer Results
Whether to round or truncate.
No rounding; truncate whenever
necessary.
Decimal Results
How to establish the length of a
decimal operand field that is
equivalent to a giv!fIl binary
operand field.
Divide the binary field length by
3 . 32; take the integer equal to
or higher than the quotient as
the length of the field.
Binary Results
How to establish the length of a
binary operand field that is
equivalent to a given decimal
operand field.
Multiply the decimal field length
by 3.32; take the integer equal
to or higher than the product as
the length of the field.
Floating-Point
Results
How to establish the precision
of a floating-point result.
Use the greater preciSion employed
by the floating point operands.
How to establish the precision of
a floating-point operand that is
equivalent to a given fixed-point
operand.
Use the total length of the fixedpoint operand.
How to establish the length of a
fixed-point field after addition
or subtraction.
Retain all positions on the right
and the left of the decimal or
binary point in both operands;
add an additional most significant position in case of
overflows.
How to establish the length of a
fixed-point field after multiplication.
Add the number of positions to
the right and to the left of
the decimal or binary point of
the two operands; the total
number of pOSitions is provided to the right and to the left
of the decimal or binary point
of the product.
How to establish the length of a
fixed-point field after division.
Use the length of the largest
possible number provided in
the specific PL/I implementation.
Fixed-Point
Results
TABLE X: COMPARISON OF FEATURES OF THREE PROGRAMMING LANGUAGES
LANGUAGE FEA TUBES
FORTRAN
COBOL
No
No
Yes
Yes
Yes
Yes
Business Pro!Q:amming
Record handling
File handling
Decimal arithmetic
Source language debugging
Report writing
Source language readily readable
Yes
Yes
Yes
No
Yes
No
No
No
No
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
No
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
*
Yes
No
No
Scientific Programming
Formulae
Floating-point representation
Complex numbers
Arrays
Source language debugging
Source language readily
readable
PeriEheral Devices
Random-access processing
Asynchronous processing
Free control of tape blocking
*
PL/I
General
Tested language
Compilers now available
*
Yes
Yes
*
*
*
Yes
This feature is not normally implemented in "standard" versions.
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
9/65
420: 171.100
IBM System/360
Machine Oriented Language
Operating System/360 Assembler
MACHINE ORIENTED LANGUAGE: OPERATING SYSTEM/360 ASSEMBLER
. 142 Instruction Elements
.1
GENERAL
.11
Identity: . . . . . . . . . . IBM Operating System/360
Assembler.
.12
Origin: . . . . . . . . . . . IBM Corporation.
.13
Reference: . . . . . . . . IBM Publication C28-6514-2.
.14
Description
.141 General Facilities and Design Levels
The structure of the Operating System/360
Assembler is such as to encourage:
•
Use of system library routines and macros, and
user-prepared macro-instruction definitions.
•
Use of separately-prepared program segments
that are linked together as one object program
only at program load time. Many cumbersome
overlay control manipulations are thus removed
from the concern of the programmer.
•
Use of programs which are dynamically
variable at execution time, according to
daily requirements and system configuration availability.
•
Use of the executive/monitor facilities of
the Operating System/360 (see Section
420:191 for a description of the Operating
System).
All the System/360 assemblers are two-pass systems that produce a listing, a symbol table, and
relocatable machine coding. Basically, the output
of each assembly is a relocatable program block;
a number of blocks loaded at one time by the Linkage Editor/Loader constitutes a program.
System/360 instructions can have a variety of diff erent lengths and formats. The instructions
basically deal with the contents of registers (there
are 16 general registers which are dual-purpose
arithmetic and index registers), the contents of
core storage, and "immediate data" (i. e., a
literal written in the instruction itself). Fewer bits
are required to address a register (one of 16
possibilities) than a core storage location (one of
over 16 million possibilities), so the different
instructions have different lengths; in fact, there
are five basic instruction formats. These are the
RR (Register to Register), RX (Register to Indexed
Storage Address), RS (Register to Storage), SI
(Storage with Immediate Operand) and SS (Storage
to Storage) types of instructions. They are discussed in the Central Processor section, Page
420:051. 121.
In the Assembler, the instruction types· are 'distinguished by their mnemonic operation codes, which
differ for each type of instruction format. (There
are, for instance, no fewer than 14 different Add
instructions and 22 Load instructions in the Universal Instruction Set.) From these mnemonics, implied lengths are derived for the instructions and,
in some cases, for the operands. There are different alignment rules for floating-point binary, fixedpoint binary, variable-length deCimal, etc., and
assignment of the proper type and alignment for
each operand is also based on the instruction
mnemonic. The details of length and type are held
internally for each operand during the assembly
process.
Three design levels* of the Operating System/360
Assembler have been announced by IBM - the 12K,
44K, and 200K levels. (Several other Assemblers
have been developed by IBM for use with the Basic
Operating System and the Basic Programming
Support package, as described in Sections 420:192
and 420: 193, respectively.) The 12K and 44K versions of the Operating System/360 Assembler are
expected to become available in December 1965,
with the full 200K version not expected until June
1966. The limitations and restrictions of the 12K
and 44K versions - although these are the versions
soon to be in the hands of users - have not been
specified by IBM to date. Only the 200K Assembler
has been documented.
Both instructions and operands are referred to
symbolically, using up to eight alphanumeric
characters for any name. Thus, JOHN AP FIELD2,
FIELDI creates an instructiop. adding FIELDI into
FIELD2, where both are dechnal fields. This instruction is then referred to as JOHN, and the use
of JOHN as an operand in a Branch instruction will
cause a jump to the instruction AP FIELDl,
FIELD2. In the System/360, this familiar concept
of symbolic operand and instruction addressing is
carried a step farther, in that expressions are
allowed. Thus, instead of allowing only a simple
name plus or minus a displacement factor, the
System/360 allows for multiplication, division,
parenthetical expression, and almost any form of
address that can be arithmetically evaluated.
* Core storage requirements of 12K, 44K, and 200K
are being used as "design levels" for the Operating System/360 software by IBM. These three
levels require, respectively:
• a 16K system with 3 tapes or a random
access device;
• a 64K system with a random access device;
• a 256K system with a random access device.
This flexible addressing capability allows the size
of the symbol table to be reduced, while providing
for handling the complexities of the various data
representations of the System/360. Relocatable
and absolute operands may be mixed within an
expression, although this mode of addressing
naturally has its own rules which a programmer
must carefully observe.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
420: 171.143
. 143 Types of Constants
Constants can be expressed in binary, decimal,
hexadecimal,fixed-point, floating-point, address,
or chli.racter modes. The varied alignment requirements are satisfied by the Assembler. A
distinction is drawn between:
• Self-Defining Constants, which are written
as part of the coding, assembled in position,
and have no symbolic name attached to them,
but which are not part of the actual machine
instruction either on the coding sheet or in
the machine .itself;
•
Literals, which are written as part of the instruction, but which, in the machine, are not
p'art of the instruction but are located in the
"literal pool, " whose contents are set by the
assembler instructions; and
•
Defined Constants, which set the original contents of a location but allow later references
to be made using symbolic names.
Each arithmetic mode has its own rules, but in
general the programmer can write in one mode
(e. g., decimal) and, by explicitly or implicity describing the constants, can have the assembler
perform the necessary conversions. Thus, floatingpoint binary constants, which are stored with an
exponent expressed in powers of 16 rather than of
10, can still be written in decimal form with decimal
exponents.
In the definition of constants, as in the use of algebraic expressions in place of symbols as described
above, the programmer can take advantage of a
number of facilities designed to allow easier coding.
Thus, a group of floating-point constants can be
preceded by a common decimal scale factor (a
power of 10). This would allow, for instance, a
table of times to be expressed in terms of microseconds by the programmer but stored in terms of
seconds by the Assembler. Another programmer
aid of this sort is the ability to define a number of
constants together, thus simplifying the setting up
of initialized tables and other groups of constants
which have some common characteristics.
. 144 Programming in Blocks
System/360 anticipates handling more than one job
at a time, so it follows that absolute storage locations must be allocated when the program is loaded
for execution rather than at assembly time. Where
a loader is able to provide this standard of sophistication, it often takes on the additional task of tying
logically independent parts of a program together
to provide for efficient use of the internal store
itself .
An immediate result of this normal development of
a loader for a multiprogramming computer system
is that "computer programs," as precise entities,
vanish from the programmer's desk and take on
final form only as they are prepared for execution
by the loader. The programmer writes program
segments, or program blocks, and defines what
external routines (outside his blocks) are needed.
These can be picked up, used as necessary, and
assigned the appropriate parameters and priorities
8/65
IBM SYSTEM/360
each time they are executed - right in the computer
room and away from the programmer's personal
supervision.
This change of outlook provides the opportunity
for more efficient equipment utilization through
the use of parameters coming in from the day-today input and priorities adjusted to suit the day's
needs. Potentially, it may provide a great deal
more in specialiZed applications.
.145 Pseudo-Instructions
A number of pseudo-instructions are available for
controlling the Assembler and providing data to the
loader. Some are machine-based and require the
programmer to appreciate that he is, in fact,
working at machine level on an unusually complicated
computer. He must, for instance, ensure that a
sufficient number of base registers are made available, and he is responsible for their settings. The
Assembler does decide which registers to use
among the ones the programmer has made available.
and set, but it definitely does not allow the programmer to think of the System/360 core storage locations as being directly addressable.
Other pseudo-instructions handle routine control of
the printouts, allow initial data to be set up, define
symbols that can be referenced by programs
external to the current block of coding, and allow
coding to be copied from the system library.
.146 Macro-Instructions
The macro-instruction facilities provided by the full
Operating System/360 Assembler to regulate supervisory control and data management are both extensive and powerful. Effective utilization of these
facilities and efficient combinations thereof will
not be easily attainable, due in part to their sheer
weight of numbers , but also due to definite deficiencies in the integration of the total software documentation. In addition, the present documentation offers
little in the way of specific aids to the programmer in
the construction of actual programs. Because the
elements of a System/360 program are numerous
(e. g., assembly control, machine and assembly
language statements, input-output control, Operating
System interrelationships, data management, program linkage and job control at execution time),
and because the choice of statements within each
programming area is potentially great, there is a
definite need for particularized suggestions in program formulation.
The facilities offered to permit the insertion of
user-designed macro-instructions are both comprehensive and flexible. A routine generated by
a given macro-instruction can be modified during
program execution, both in its instruction operands
and its sequence of operation, depending upon
dynamically-encountered circumstances.
.147 Program Diagnostics and Dynamic Control
An Operating System/360 control function called
the Test Translator (TESTRAN) is provided to
perform object program diagnostics at program
execution time. Diagnostic controls of the dumps
and traces are structured by the programmer in
macro-instruction language during preparation
(Contd. )
MACHINE ORIENTED LANGUAGE: OPERATING SYSTEM/360 ASSEMBLER
care and considerable skill. Beyond this consideration, the programmer is faced with several other
potentially difficult tasks. The addition of userdefined macro-instructions implies an understanding
of another set of macro-definition and control instructions (up to 17). Linkage of the problem program to the Supervisor function of the Operating
System/360 is accomplished by using up to 31
different macro-instructions according to specifically
requested control functions. The operand parameters and their associated format rules differ
according to the nature of the desired control
function. The programmer must also code the
input-output data management macro-instructions
(up to 32) and, optionally, any of the 23 possible
diagnostic macros. It seems apparent that coding
simplicity has been sacrificed in favor of sophisticated language/machine utilization.
.147 Program Diagnostics and Dynamic Control (Contd.)
of the assembly language program. The diagnostic
output is edited to print in whatever data structure
the programmer may specify, with source and
machhle language labels and addresses printed as
applicable. Several varieties of effective dumps
and traces are provided, and all are dynamically
modifiable by the object program. Limits and
counters can be set to avoid tracing runaway loops,
for example, and parameters can be included to
diagnose only specified classes of output.
The inclusion of these comprehensive diagnostic
facilities in the full Operating System/360 should
greatly facilitate.program debugging. Such facilities seem, in fact, to be an absolute prerequisite for
the E!ystem/360 since the programmer may be dealing
with decimal, binary, alphabetic, hexadecimal,
fixed-point, and floating-point operands, all within
a few instructions. In addition, program control
may frequently pass between the problem program
and multiple supervisory functions of the Operating
System. Debugging aids must necessarily be comprehensive and convenient.
.149 Assembly Times
IBM has provided us with estimated assembly times
for the Operating System/360 Assemblers on three .
sample System/360 configurations. Table I describes
these sample configurations and indicates the several
timing factors that must be summed to arrive at a
total time for an assembly that is performed in conjunction with the Operating System/360.
No method has been provided for referring to the
working data that the Assembler keeps internally,
and amending the coding depending on the status
of this data during the assembly process. The
lack of such a corrective facility may cause problems in handling the base register assignment
and the displacement count (i. e., the elements of
each main storage address). The inability to get
at and amend the register-displacement relationship
may lead to inefficient storage allocation.
An ffiM-supplied timing example shows the estimated
time requirements to perform a 44K assembly
operation on a main program of 800 source statements on sample Configuration B (a 65K Model 50).
In this example, it is assumed that 1,000 lines of
source-program listing information will be produced.
The estimated time to perform this assembly
operation (reading the source statements from
disc storage and writing the compiled program
and source program listing to disc storage) is
51. 0 seconds.
.148 Machine/Language Complexity
A programmer uses an assembly language to allow
him to control the computer at machine level, but
without all the complexity of the machine language.
In addition to this basic function, assemblers nowadays provide for routines to be inserted or referred
to, for handling the needed relationships with the
operating system (if any), and a variety of other
aids to machine language programmers. It follows,
therefore, that the more complex the computer
system, the more complex the assembler needs to
be if it is to be of maximum utility to the programmer.
The ffiM System/360 is a complex computer. It has
an eight-bit character code (and programmers are
not yet used to handling such codes); it requires
that most addresses be in two-part form (base and
displacement); it uses five different classes of
machine language instructions, each with its own
format requirements; it can manipulate five types
of data, with distinct instruction applicability and
positional rules for each type. In addition, the
System/360 operates in an environment that transfers control between one or more problem programs
and the Operating System according to dynamically
varying conditions. All these factors add complexities which must by handled by the Operating System/
360 Assembler, and the facilities provided by IBM
to do so seem, in general, to be both comprehensive
and effective.
To use the power and flexibility of the System/360
Assembler requires that the machine and assembler
instructions (up to 184) be coded with meticulous
420: 171.147
.15
Publication Date: . . . . April, 1964.
.16
Availability
Design level 12K: . . . . December 1965.
Design level 44K: . . . . December 1965.
Design level 200K: ... June 1966.
.2
LANGUAGE FORMAT
.21
Diagram: . . . . . . . . . refer to System/360 Assembler Coding Form, Page
420:172.820.
.22
Legend
Name field: . . . . . . . . assigns a symbolic name to
a statement.
Operation field: . . . . . specifies a machine instruction or assembler
instruction.
Operand field: . . . . . . identifies and describes
data to be acted upon by
the instructions.
Comments field: . . . . . permits lines of descriptive
information to be listed.
Identification-Sequence
field: . . . . . . . . . . . optional field which identifies the program and/or
sequences statements.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
420:171.230
IBM SYSTEM/360
TABLE I: ESTIMATED ASSEMBLY TIMES
Sample Configurations
Configuration Components
Configuration A
System/360 Model 30 with 32K bytes of core
storage, one Multiplexor Channel, one
Selector Channel, one 2540 Card Read
Punch, one 1403 Printer (Model 3), two
2311 Disk Storage Drives, and the Operating System/360's sequential scheduler. '"
Configuration B
System/360 Model 50 with 65K bytes of core
storage. two Selector Channels. two 2311
Disk Storage Drives. and the Operating
System/3GO's priority scheduler. "'*
'Configuration C
System/360 Model 65 with 262K bytes of
core storage. two Selector Channels. two
2301 Drum Storage Units. four 2311 Disk
Storage Drives, and the Operating System/
360's concurrent job scheduler. **
Assembly Time in Seconds
Estimated Timing Factors
12K Assembler.
44KAssembler,
200KAssembler.
Configuration A
Configuration B
Configuration C
For assembly initiation
For each assembly
For each output line
21.1
28.0
0.46
15.0
14.0
0.022
3.57
6.5
0.006
... Input is assumed to be entered through the Card Read Punchj output is assumed to
consist of a printed source-program listing and an object program written on disc
storage.
.... Source-program input and object-program output are assumed to be read from and
written to disc storage. Formatted lines of the source-program listings are also
written to disc storage.
.23
Corrections:.....
• 24
Special Conventions
. no special provisions.
.241 Compound addresses: · one term or an arithmetic
. 242 Multi-addresses: .
· 243 Literals: ••...••
· 244 Special coded
addresses:
combination of terms
(operators +, -, *, and /).
· separated by commas.
· preceded by = and specified as
to type (e. g., hexadecimal,
floating-point, binary, etc.);
only one literal is permitted per machine instruction. Data constants
are preceded by quotation
marks; address constants
by parentheses.
·*
represents current value
of the location counter.
· 245 othersSelf-defining terms: · the value of the term is
inserted directly into
the instruction and, as
such, is independent of
program relocation. The
value is specified as to
type and is enclosed by
quotation marks.
.3
LABELS
. 31
General
.311 Maximum number of
labels: . . . . . . . . . . >5,000.
• 312 Common label formation rule: • • • • . • . . yes.
.313 Reserved labels: .• . none .
. 314 other restrictions: ... none.
.315 Designators: . . . . . . . ampersand (&) in first
position of label indicates
a prototype, model, or
symbolic parameter
statement as used in macroinstruction definition.
.316 Synonyms permitted: •. yes; EQU pseudo •
.32
Universal Labels
.321 Labels for procedures Existence: . . . . . . . . mandatory if referenced by
other program segments.
Note: Universal labels
defined within a segment
and referred to in another
segment are identified
by the ENTRY pseudo;
referencing segments
must identify universal
labels by the EXTRN
pseudo.
Formation rule First character:. · alphabetic.
Others: . . . . .
• alphanumeric (A-Z, 0-9).
Number of
characters: .
· 1 to 8 characters.
.322 Labels for library
routines: . . . . .
· same as procedures ..,
.323 Labels for constants: · same as procedures.
• 324 Labels for files: ..• · same as procedures (in
macros only) .
.325 Labels for records: . same as procedures (in
macros only).
.326 Labels for variables: . same as procedures .
• 33
Local Labels
. 331 Region: . . . . .
· local to a program segment .
(Contd. )
8/65
/
420: 171.332
MACHINE ORIENTED LANGUAGE: OPERATING SYSTEM/360 ASSEMBLER
.332 Labels for procedures Existence: . . . • . . . . mandatory if referenced by
an instruction in the same
segment.
Formation rule First character: ... alphabetic.
others: . . . • . . . . . alphanumeric (A-Z, 0-9);
no special characters and
no embedded blanks.
Number of characters: •..••..•
. 333 Labels for library
routines: . . . . . . . . .
. 334 Labels for constants: .
. 335 Labels for files:: .•..
1 to 8 characters.
same as procedures.
same as procedures.
same as procedures (in
macros only).
.336 Labels for records: .• same as procedures (in
macros only).
. 337 Labels for variables: . same as procedures.
· 54
Translator Control
.541 Method of control Allocation counter: .. pseudo-operation.
Label adjustment: .•. pseudo-operation.
Annotation: . • . . . . . pseudo-operation and
special cards.
· 542 Allocation counter Set to absolute: . . . . START.
Set to label: . . . . . . . ORG.
Step forward: . . . . . . ORG, CNOP.
Step backward: . . . . . ORG.
Reserve area: .•... DS, ORG, LTORG .
.543 Label adjustment Set labels equal: . . . . EQU .
Set absolute value: .. EQU .
Clear label table: ... none.
· 544 AnnotationComment phrase: ... * in first column.
Title phrase: . . . . . . TITLE or START .
.4
DATA
.6
SPECIAL ROUTINES AVAILABLE
.41
Constants
· 61
Special Arithmetic: .. most can be provided by the
System/360 hardware and
optional features .
Format used by coder Machine format
. 62
Special Functions:
Integer Decimal:.
.63
Overlay Control
. 411 Maximum size constants -
. . decimal (up to 16 bytes) in packed or unpacked
format.
Binary: . . . . . . . . . . binary (up to 256 bytes).
Fixed numeric Decimal: . . • . . . . . . decimal (up to 16 bytes)
or binary (up to 8 bytes) .
Hexadecim:al (up to 512
digits): . . . . • . . . . . binary (stored 2 hexadecimal digits per byte) .
Floating numeric Decimal: . . • • . . . . . binary (4 or 8 bytes).
Alphabetic: . . . . . . . . up to 256 characters (bytes).
Alphameric: . . • . . . . up to 256 characters (bytes).
.412 Maximum size literals:same as constants; see
Paragraph .411 above.
.5
PROCEDURES
· 51
Direct Operation Codes
· 511 MnemonicExistence: . . . . . . • .
Number: . . . • . . . . .
Example: . . . • . . . .
. 512 Absolute: . . . . . . . . .
. 52
mandatory.
142 plus 17 extended codes.
A = Add.
not permitted.
Macro-Codes
.53
· 631 Facilities: . . . . . . . . . provided by linkage macros
and Operating System/360
loader routine .
.632 Method of call: . • . . . . program segmentation and
linkage are indicated by
pseudo-instructions.
· 64
Interludes: . . . • . . . . none.
Data Editing
· 641 Radix conversion: . . . . machine instructions are
provided to convert
between binary and
decimal.
Code translation: •..• a "Translate" machine instruction is provided for
codes of up to 8 bits .
. 642 Format control: ....• provided by "Edit" instruction (part of optional
Decimal Arithmetic feature
in Models 30 and 40).
Zero suppression: .. yes.
Size control: . . . . . . yes.
Sign control: . • . . . . yes.
Special characters: .. yes .
· 643 Method of call: . . • . . . machine instructions .
.65
· 521 Number available For macro-language
control: . . . . . . . . 17.
For supervisor
services: . . . . . ' .. 31.
For data management:32.
For diagnostic
services: . . . . . . . 23 .
. 523 New macros: . . . . . . . defined in source program
or inserted into library
in a separate operation.
. none announced to date .
Input-Output Control
.651 File labels: . • . . . . . . provided by Operating
System/360.
.652 Reel labels: . . . . . . . . provided by Operating
System/360.
.653 Blocking: . . . . . . . . . provided by Operating
System/360.
.654 Error control: . . . . . • provided by Operating
System/360.
. 655 Method of call: •....• macro-instructions .
.66
Sorting:
.•. see Paragraph 420:151.12
for Sort/Merge Program
des cription.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
IBM SYSTEM/360
420: 171.670
.67
Diagnostics:
. provided by Test Translator
function of Operating
System/360.
.7
LIBRARY FACILITIES
.71
Identity:........
. System or Program Library.
.72
Kinds of Libraries:
. expandable master.
.73
Storage Form:.
. magnetic tape, disc, or
drum.
.74
Varieties of Contents: . statements in any source
language, operating system control routines, all
software packages,
macros, data descriptions,
independent programs.
.75
Mechanism
,,751 Insertion of new item: . by means of job control
statements at load time .
. 752 Language of new item:. any acceptable to System/
360.
.753 Method of call: . . . . . . COpy pseudo causes a prewritten section of coding
to be inserted.
. 76
Insertion in Program
.761 Open routines exist: ..
.762 Closed routines exist: .
. 763 Open-closed is
oI\tional: . . • . . . . . .
. 764 Closed routines
appear once: . . . . . .
yes.
yes.
yes .
yes.
.8
MACRO AND PSEUDO TABLES
.81
Macros (Partial List)
Code
Description
WAIT: . . . . . . . . . . . indicates the program cannot
proceed until a certain condition is met; the operating
system will initiate a new
program.
GET:
. moves next logical record of
input file into work area.
PUT:
. moves next logical record
of output file into the output
buffer.
OPEN: . . . . . . . . . . . repositions file to be opened
to the beginning and checks
label if present.
CLOSE: . . . . . . . . . . writes contents of remaining
buffers, processes trailer
labels, and disposes of file
as indicated.
.82
Pseudos (Partial List)
Code
ICTL:.
8/65
Description
. specifies the format of the
source program statements.
Code
Description
!SEQ: . . . . . . . . . . . . causes the assembler to
check the sequence of input
cards .
ORG: . . . . . . . . . . . . alters the setting of the location counter for the
current control section.
LTORG: . . . . . . . . . . specifies the location of the
literal pool into which all
literals thus far encountered are to be
assembled.
CNOP:
.. aligns an instruction at a
specific word boundary.
COPY:
. includes previously written
source-language coding in
a program.
END: . . .
. terminates assembly of a
program.
TITLE: .
. identifies the assembly
listing and assembly output
cards.
EJECT: . . . . . . . . . . causes next line of listing
to appear at the top of a
new page.
SPACE:
. inserts one or more blank
lines in the listing.
PRINT:.
. designates how much detail
is to be included in the assembly listing .
EQU: . . . . . . . . . . . . defines a symbol by assigning to it the attributes of
an expression in the operand field .
DC:
. provides constant data in
storage.
DS:
. reserves areas of storage
and assigns names to them .
. defines and generates an
CCW: . . . . • • • .
8-byte Channel Command
Word.
START: . . . . . • . . . . can be used to name the
first control section of a
program and specifies a
tentative starting location
for the program.
CSECT: . . . • . • . . . . identifies the beginning or
the continuation of a control section.
DSECT: . . • . . . . . . . identifies the beginning or
resumption of a dummy
section.
ENTRY: . • . . . . . . . . identifies linkage symbols
that are defined in this
program but may be used
by some other program.
EXTRN: . . . • . . . • . • identifies linkage symbols
that are used by this program but defined in some
other program.
USING: . . • . . . . . . . . indicates that one or more
general registers are
available for use as base
registers.
DROP: . . . . . . . . . . specifies a previously available register that may no
longer be used as a base
register.
420: 172. 100
IBM System/360
Machine Oriented Language
BOS Assembler
MACHINE ORIENTED LANGUAGE: BASIC OPERATING SYSTEM/360 ASSEMBLER
.1
GENERAL
.11
Identity: . . • . . • . . . . IBM Basic Operating System/360 (BOS) Assembler.
. 12
Origin:
.13
Reference: ..••.••. IBM Publications C24-33611 and C24-3364-1.
• 14
Description
•..••••.•• IBM Corporation.
Two Assemblers have been announced by IBM for
use with the System/360 Basic Operating System:
a 4K Disk version designed for use with the 8K
Disk Basic Operating System and a 10K Tape/Disk
version for use with the 16K Tape/Disk Basic
Operating System. The 4K design level Disk
Assembler is described in this report section; the
advantages of using the larger 10K Assembler with
the Basic Operating System have not been documented by IBM to date. Delivery dates for the two
Basic Operating System Assemblers have been
announced as September and'December 1965 for the
4K and 10K versions, respectively.
The Basic Operating System/360 (BOS) Assembler
can assemble programs written in the Basic
Programming Support assembly languages and in
the assembler language of the 7090/7094 Support
Package for the IBM System/360 (see Sections
420: 173 and 420: 151 for descriptions of these additional assembly languages). The BOS Assembler
language is a proper subset of the Operating
System/360 Assembler described in the previous
report section (420:171), and, as such, any program written in the 8K BOS language can be
assembled with the Operating System/360 Assembler, with one notable exception: the XFR
(Generate a Transfer Card) assembly instruction,
which defines the transfer point or entry point of a
phase or overlay, is not recognized as a valid
mnemonic operation code by the Operating System/
360 Assembler.
I
\
Other limitations and restrictions of the Basic
Operating System Assembler, as related to the full
Assembler of the Operating System/360, center
principally on number of operands per statement,
size of reserved storage, and depth of nesting per
symbolic expression. Those differences are illustrated in Table I. It should be noted that the COM
and COpy assembly instructions are not available
in the Basic Operating System Assembler. (In the
Operating System/360 Assembler, COM identifies
and reserves a common area of storage for use by
independent programs linked at load time; COPY
obtains source language coding from a library and
includes it in the program currently being assembled.)
The facilities provided for use of macro-language
statements are quite comprehensive. The user can
define a macro instruction containing up to 49
symbolically-replaceable operands; he can set and
alter the value of operands within a macro; and he
can alter the sequence of instructions within a
macro-definition depending upon dynamic execution
time conditions by means of 5 conditional instructions. Unlike the Operating System/360 Assembler,
the BOS Assembler cannot use the conditional instruction outside the macro-definition. Similarly,
the use of operand sublists and SET symbol subscripts (to define the value of variable symbols)
is not permitted when defining macro-instructions
for use with the BOS Assemblers. Another limit
to the facilities of the BOS Assemblers is that
nesting of inner macro-instructions can be accomplished only to a depth of three, as compared with
the virtually unlimited nesting potential in the
Operating System/360 Assembler.
The input-output control macro-instructions for the
Basic Operating System Assembler bear the
designations of Logical IOCS and Physical IOCS.
Physical IOCS routines reside in core storage as
part of the system's Supervisor; with the Supervisor
they consume a maximum of 4,096 bytes of core
storage. The facilities they provide will control
as many I/O devices as can be connected to the one
Multiplexor Channel and up to two Selector Channels. The logical record-handling routines provided by Logical IOCS do not require the permanent
use of main storage. These routines reside on disc
storage and are called into main storage by macroinstructions only as required by the programmer.
Physical IOCS consists of four routines to control
the actual transfer of records between the inputoutput device and core storage: Start I/O routine,
Interrupt routine, Channel Scheduler, and Device
Error routines. Logical IOCS, on the other hand,
performs the following functions with regard to
logical data records: blocking or unblocking of
records; flip-flopping between I/O storage areas
when two areas have been allotted to permit data
transfer overlap; end-of-file and end-of-volume
control; and checking and writing of standard tape/
disc labels. The core storage required to use the
macro facilities of IOCS in support of programs
that utilize both tape and disc components will be
a minimum of 16K. The facilities contained within
the logical 10CS package are extensive, but many
of them (such as the optional dual input-output
areas) will prove costly in terms of core storage
requirements for systems of less than 16K bytes.
Preceding the source program at assembly time
must appear the declarative file-definition macroinstructions. The input-output file descriptions
contained on these cards are utilized by the IOCS
function to set up the routines necessary for each
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
IBM SYSTEM/360
420: 172.140
TABLE I: LANGUAGE DIFFERENCES BETWEEN THE BASIC OPERATING
SYSTEM AND OPERATING SYSTEM ASSEMBLERS
OPERATING SYSTEM
SPECIFICATION
BASIC OPERATING SYSTEM
LANGUAGE
fEATURE
.SPECIFICATION
CODING CONVENTIONS
Maximum of two continuotion
lines allowed
Continuation Une~
Maximum of one continuation
line allowed
Address Constants
Only One odd,,,,, comtcnt may
be specified in a DC statement or
a literal
One Or more Clddres5 canllanh may
be specified in a DC Itatement or
a literal
Bit-Length Specification
Feature excluded
Feature i, provided in Operating Sptem
Only one operand allowed per
DC statement
One or more operands allowed
per DC statement
Oupl icalion foctor exprehoble
only by (] decimal self-defining
Duplication factor expres!oble
by any absolute expressiCln
Exponent Modifier
Exponent modifier expresmbJe only
by a decimal self-defining term
Exponent modifier exprellable
by ony abalute expressian
length Modifier
Length modifier expressoble only
by a decimal self-defining te,m
length modifier expreslOble by
any absolute exprelsion
Literal>
In case of duplicate lite,al"
mare than one mClY be ,tOlrd
In cme of duplicate literals,
only On". is stored
Scale modifiel exp,e~sable only
Scale mod'!ie,. "xpresloble by
CONSTANTS
DC Operands
Duplication FocI"
Scale Modifier
by a decimal self-defining tel_m_.~.~.--L_o_c_y_ob_'o_'c_,"_._'_'_p,_,,_siO_c_ _ _--j
EXPRESSIONS
1'10 limit on the number of lets
of parentheles (
) allowed
in an "xp,ellion
Parenthe,es
Only one let of porenthesel (
ollowed·in on expression
Terms
Maximum of three '",m, allowed
per exprellion
OS Operand
Only one operand allowed in a
OS statement
One or male opeland, aliawed
in a DS statement
length Modi fier
Maximum length designation of a
storage field is 256 by tel
Maximum length designation of
a storage field is 65,536 byte,
(NOP
Each opel and explelloble only by
a decimal self-defining term
Eoch operand expresloble by
any absolute exples~ion
COM
Stotement excluded
Statement provided in Operating System
COpy
Statement excluded
Statement provided in Operating System
CSECT and
DSECT
In addition to the combined number
of CSECTs, DSECT" EXTRN" and
V-type address constants not being
allowed to exceed 255, the combined
numbe' of CSECT and DSECT stalements must not exceed 32.
The combined number of (SECT
,tatements, DSECT statements,
EXTRN Itotemenll" and V-type
add res. constants mUlt not exceed
)
No limit on the number of term,
allowed per exp'ession
STORAGE DEFINITION (OS STATEMENT)
ASSEMBLER-INSTRUCTION STATEMENTS
EXTRN ond
ENTRY
nTlE
ed
ane relocatable Iymbal ;, "lloweach EXTRN and ENTRY Ilote-
The fir\t TITLE Itatement provide,
the
for p0ge~ of the
listing that
untt! the
next TITLE statement
any);'
encoulltered
255
One or more relocatable symM
boll are allowed in each
EXTRN and ENTRY statement
The first TITLE statement, in
addition to providing the heading for all pages of the Ii,tthat lie between it and the
TITLE
~Iatement
also provido:;os the
any
of the listing
that
it.
USING and
DROP
Maximum of 5 bale
nations allowed
DROP statement
XFR
Statement provided in Basic
Operating System
de'\gUSING or
Maximum of 15 1.,,,.1' register
de~ignations allowed in each
USING or DROP statement
Statement excluded
Reprinted from IBM Publication C24-3361-1, p. 163.
.14
Description (Contd.)
type of processing, such as consecutive, direct
access, or indexed sequential, and to set aside
specified core storage areas for each file. This
method of controlling input-output operations,
although quite functional, offers little that is new in
the area of automatic I/O device control for relatively small systems.
The Basic Operating System Assembler contains
many of the features of the Operating System/360
Assembler, such as the free-form or fixed-form
coding sheet, operands in the form of multiple terms
and expressions, the use of self-defining terms,
program sectioning by control segments, the use of
symbols external to the current program segment,
and the use of explicit and implied operand lengths.
All of the machine instructions in the full System/
360 instruction s.et can be handled by the BOS
Assemblers, as well as 17 extended mnemonic
codes to express conditional branching instructions.
Up to 24 Assembler control instructions are provided, as well as several macros to enable the
(Contd. )
8/65
MACHINE ORIENTED LANGUAGE: BASIC OPERATING SYSTEM/360 ASSEMBLER
. 14
420: 172. 141
Description (Contd.)
.2
LANGUAGE FORMAT
programmer to communicate with the core-resident
Supervisor in its performance of tasks relating to
interrupt handling, I/O requests, and program
retrieval.. Certain Job Control functions are also
available to control the Assembler's various types
of output decks and lists, and the amount of work
area available to the Assembler.
.21
Diagram: . . . . . . . • . refer to System/360 Assembler Coding Form, Page
420:172.820.
.22
Legend
Name field: . . . . . . . . assigns a symbolic name to
a statement.
Operation field: . . . . . specifies a machine instruction or assembler
instruction.
Operand field: •....• identifies and describes data
to be acted upon by the
instruction.
Comments field: ...• permits lines of descriptive
information to be listed.
Identification-Sequence
field: . . . . . . . . . . . optional field which identifies the program and/or
sequences statements.
In order to perform diagnostic routines while exe-
cuting a program under control of the Basic
Operating System/360, the Autotest functions of the
BOS may be used. Autotest is a group of diagnostic
routines that the programmer can call for by a
special set of control cards while writing in the
Assembly language. The Autotest Control Program
resides in core storage along with the Supervisor
and the user's problem program; as a result, the
computer used for testing the object program with
Autotest diagnostic requests must contain at least
16,384 bytes of core storage. Autotest facilities
for users of 8K disc-oriented systems have not been
offered by IBM to date. It must be noted that when
using the Autotest facilities in a 16K target computer environment, under control of the Basic
Operating System/360, the user's problem program
must be considerably smaller than 16K bytes of core
storage in size.
.23
Corrections: .•....• no special provisions.
.24
Special Conventions
•
Either one Multiplexor or one Selector Channel..
.241 Compound addresses: . one term or an an arithmetic
combination of terms
(operators +, -, *, and/).
.242 Multi-addresses: .••. separated by commas.
.243 Literals: . . • . . . . . • preceded by = and specified
as to type (e. g., hexadecimal, floating-point,
binary, etc.); only one
literal is permitted per
machine instruction.
Data constants are preceded by quotation marks,
address constants by
parentheses.
.244 Special coded
addresses: .•.•.•. * represents current value
of the location counter.
.245 Others Self defining terms: . the value of the term is
inserted directly into the
instruction and, as such,
is independent of program relocation. The
value is specified as to
type and is enclosed by
quotation marks.
•
One mM 2311 Disk Storage Drive.
.3
LABELS
•
One disc work area.
.31
General
•
One Card Read-Punch (1442 or 2540).
•
One mM 1403, 1404, or 1443 Printer.
•
The following Basic Operating System functions:
Loader, Supervisor, Job Control, Assembler,
and Macro Library routines, as necessary.
Autotest facilities include the ability to Autopatch;
i. e., to exchange, add, or delete instructions
without reassembling. Various types of registers
and core storage dumps are also provided, plus a
listing of all program phases fetched by the problem
program.
Users of 8K System/360 configurations are provided
with several independent diagnostic routines in lieu
of the Autotest facilities. These routines include
data file and core storage dumps. Core storage
dumps can be called by the BOS Supervisor at
abnormal end-of-job, without any operator intervention. These dumps can also be called for by the
problem program at normal end-of-job, or by the
operator at any time.
To perform a program assembly, the BOS Assembler requires a System/360 with at least the following features and units:
•
8, 192 bytes of core storage.
•
Standard instruction set.;
.15
Publication Date: ..•. April, 1965.
.16
Availabilitv
Design level 4K: . . . . September, 1965.
Design level 10K: . . . . December, 1965.
© , 965
.311 Maximum number of
labels: . . . . . . . . . . ?
.312 Common label formation rule: . . . . . . . . yes.
.313 Reserved labels: . . . . none.
.314 Other restrictions: .•. none .
. 315 Designators: . • . . . . . ampersand in first position
of label indicates a prototype, model, or symbolic
parameter statement as
used in macro-instruction
definition.
.316 Synonyms permitted: . yes; EQU pseudo.
AUERBACH Corporation and AUERBACH Info, Inc.
8/65
IBM SYSTEM/360
420: 172.320
• 32
Universal Labels
• 321 Labels for procedures
Existence: ••••••• mandatory if referenced by
other program segments.
Note: Universal labels
defined within a segment
and referred to in another
segment are identified by
the ENTRY pseudo;
referencing segments
must identify universal
labels by the EXTRN
pseudo.
Formation rule First character: ••• alphabetic.
others: ••••••••• alphanumeric (A-Z, 0-9).
Number of
characters: •••••• 1 to 8 characters.
• 322 Labels for library
routines: •••••••• same as procedures.
.323 Labels for constants: • same as procedures.
• 324 Labels for files: •••• same as procedures (in
macros only).
• 325 Labels for records: •• same as procedures (in
macros only).
• 326 Labels for variables: • same as procedures.
• 33
Local Labels
• 331 Region: .••••••••• local to a program segment.
• 332 Labels for procedures Existence: ••••••• mandatory if referenced by
an instruction in the same
segment.
Formation rule First character: ••• alphabetic.
Others: ••••••••• alphanumeric (A-Z, 0-9);
no special characters and
no embedded blanks.
Number of characters: ••••••••• 1 to 8 characters.
.333 Labels for library
routines: •••••••• same as procedures.
• 334 Labels for constants: • same as procedures.
• 335 Labels for files: .••• same as procedures (in
macros only).
• 336 Labels for records: •• same as procedures (in
macros only).
• 337 Labels for variables: • same as procedures.
•4
DATA
• 41
Constants
• 411 Maximum size constants
Format used by coder Machine format
hlteger Decimal: ••••..•• decimal (up to 16 bytes)
in packed or unpacked
format.
Binary: •••••..•• binary (up to 256 bytes).
Fixed numeric Decimal: ••••••••. decimal (up to 16 bytes) or
binary (up to 8 bytes).
Hexadecimal (up to
512 digits): ••..••. binary (stored 2 hexadecimal
digits per byte).
Floating numeric Decimal: •••••.•• binary (4 or 8 bytes).
8/65
Alphabetic: ••••..•• up to 256 characters (bytes) .
Alphameric: .••.•.• up to 256 characters (bytes) .
.412 Maximum size
literals: •••••••.. same as constants; see
Paragraph .411 above.
.5
PROCEDURES
• 51
Direct Operation Codes
.511 Mnemonic Existence: ..••••.
Number: ••.•••••
Example: .••..•••
.512 Absolute: .••••.••.
• 52
mandatory.
142, plus 17 extended codes.
A = Add.
not permitted.
Macro-Codes
.521 Number available
For macro-language
control: •••.•••• 12.
For supervisor
communication: • . • 8 •
For supervisor
assembly: .••••.. 4 .
For input-output
control (IOCS): .•• 25 .
• 523 New macros: .•.•••• placed in the macro library
by using the catalogue
function of the system
librarian •
.53
Interludes: . . • . . . . . none.
· 54
Translator Control
• 541 Method of control Allocation counter: •• pseudo-operation.
Label adjustment: .• pseudo-operation.
Annotation: •••••• pseudo-operation and special
cards.
• 542 Allocation counter Set to absolute: •••• START.
Set to label: •••••• ORG •
Step forward: ••••• ORG, CNOP •
Step backward: •••• ORG.
Reserve area: .••.. DS, ORG, LTORG •
.543 Label adjustment Set labels equal: ••• EQU •
Set absolute value: •• EQU.
Clear label table: .•• none •
• 544 Annotation Comment phrase: ••• * in first column .
Title phrase: .••••• TITLE or START •
.6
SPECIAL ROUTINES AVAILABLE
.61
Special Arithmetic: .• most can be provided by the
System/360 hardware and
optional features.
.62
Special Functions: ••• none announced to date.
.63
Overlay Control
.631 Facilities: •••••••• provided by linkage macros
plus Basic Operating
System/360 loader routine.
.632 Method of call: ••••• program segmentation and
linkage are indicated by
pseudo-instructions.
(Contd. )
420: 172.640
MACHINE ORIENTED LANGUAGE: BASIC OPERATING SYSTEM/360 ASSEMBLER
.64
Code
Data Editing
PUT: •••••••••••• moves next logical record
of output file into the output buffer.
OPEN: ••••••••••• repositions file to be opened
to the beginning and checks
label if present.
CLOSE: •••••••••• writes contents of remaining buffers, processes
trailer labels, and disposes of file as indicated .
.641 Radix conversion: ••• machine instructions are
provided to convert between binary and decimal.
Code translation: .••• a "Translate" machine
instruction is provided for
codes of up to 8 bits.
.65
Input-Output Control
.651
• 652
• 653
. 654
.655
File labels: .•••••.• provided by BOS.
Reel labels: •••.••. provided by BOS .
Blocking: ••••..••. provided by BOS •
Error control: ••.••.• provided by BOS.
Method of call: .•••• macro-instructions.
• 66
Sorting: •••••.•••• see Paragraph 420: 151. 12
for Sort/Merge Program
description.
• 67
Diagnostics: •••••.• provided by Basic Operating
System's Autotest function;
not included in the Assembler language.
.7
LffiRARY FACILITIES
.71
Identity: ••••.••••• System or Program Library.
• 72
Kinds of Libraries: ••• expandable master.
• 73
Storage Form: .••••• magnetic tape or disc.
• 74
Varieties of Contents:
.75
Mechanism
core image library, macro
library, and relocatable
library, including system
programs, the Librarian
programs, IBM-supplied
programs and user programs.
.751 Insertion of new item: . System librarian routine.
.752 Language of new item:. any acceptable to System/
360.
.753 Method of call: ••.•• macro-instructions.
• 76
Insertion in Program
• 761 Open routines exist: •• yes.
.762 Closed routines exist: . yes.
.763 Open-closed is
optional: ••.•...•. yes.
. 764 Closed routines
appear once: . • • • . • yes.
.8
MACRO AND PSEUDO TABLES
.81
Macros (Partial List)
Code
,I
\,
'-.
Description
WAIT: ••••.•••••• indicates the program cannot proceed until a certain
condition is met; the
operating system will
initiate a new program.
GET: .••••••...•. moves next logical record
of input file into work area.
©
Description
.82
Pseudos
Description
ICTL: •••••.•••.• specifies the format of the
source program statements .
ISEQ: .••...•.••.• causes the assembler to
check the sequence of input cards.
ORG: ••••....•.•• alters the setting of the
location counter for the
current control section.
LTORG: . . . . . . • . • • specifies the location of the
literal pool into which all
literals thus far encountered are to be
assembled •
CNOP: •••.••.•..• aligns an instruction at a
specifiC word boundary •
XFR: ••••.•..•••• generates a card that defines
the transfer point or entry
point of another program
segment or overlay.
END: •••••••.•••• terminates assembly of a
program.
TITLE: •••••.••.• identifies the assembly
listing and assembly output cards.
EJECT: .•••••.••• causes next line of listing
to appear at the top of a
new page.
SPACE: .•.•••••.. inserts one or more blank
lines in the listing.
PRINT: •••••••••• designates how much detail
is to be included in the
assembly listing .
EQU: ••••.••••••• defines a symbol by assigning to it the attributes of
an expression in the
operand field .
DC: •••.••••.•..• provides constant data in
storage.
DS: .••••••••••.• reserves areas of storage
and assigns names to them.
CCW: .••..•••.•.. defines and generates an
8-byte Channel Command
Word.
START: •.••.•••.• can be used to name the
first control section of a
program and specifies a
tentative starting location
for the program.
CSECT: •••••••••• identifies the beginning or
the continuation of a control section.
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
IBM SYSTEM/360
420: 172.820
.82
Description
Pseudos (Contd.)
EXTRN: •••••••••• identifies linkage symbols
~at are used by this program but defi.ned in some
other program.
USING: • • • • • • • • • • • indicates that one or more
general registers are
available for use as base
registers.
DROP: ••••••••••• specifies a previously available register that may no
longer be used as a base.
register.
Description
DSECT: •••••••••• identifies the beginning or
resumption of a dummy
section.
ENTRY: •••••••••• identifies linkage symbols
that are defined in this
program but may be used
by some other program.
IBM
IBM Syltlml3BO AII'lmblel' Cad.IDg PIIl'IIl
t===~-----------------------'----------~~~~NS Ir.=~~--r-~~-f~~+-~-+'~."~o...
o:~~m---~
I
I
I
I
I:
!I
I!
I
I'I ·,
1I
i I
Ii
I: : :
I
i
I
I
TI'i
I I
TTl I I: : I
II i I i"1
II
I
!
I
,I
I,
II
II
I,
,i'l
II
I
II Ii
Figure 1: mM System/a60 Assembler Coding Form
/
8/65
420: 173.100
IBM System/360
Machine Oriented Language
BPS Assembler,s
MACHINE ORIENTED LANGUAGE: BASIC PROGRAMMING SUPPORT ASSEMBLERS
.1
GENERAL
.11
Identity:
Basic Programming Support
(BPS) Assemblers:
Tape Assembler with
Input-Output Macros.
Card Basic Assembler.
. 12
Origin: . . . . . . . . .
IBM Corporation .
.13
References: . . . . . .
IBM Publications:
C24-3355-1,
C28-6503-3,
C24-3364-1,
C24-3343-0.
.14
Description
Two assembler programs are offered by IBM with
the System/360 Basic Programming Support package.
The Basic Programming Support (BPS) Tape Assembler functions under control of a Supervisor
which permits program relocatability, the use of
input-output macro instructions, interrupt handling,
and control of device error routines. The BPS
Tape Assel.Ilbler is constructed on a 4K-byte design level and will function with any model System/
360 with 8,192 bytes of core storage. The BPS
Tape Assembler was released to the field in July,
1965.
The Basic Programming Support Card Basic Assembler is a simplified and somewhat restricted
version of the other System/360 assembly languages.
Its program design level, like that of the BPS Tape
Assembler, is the somewhat-constricting 4K bytes,
and it will function with any model System/360 with
8,192 bytes of core storage. But since the BPS
Card Assembler uses no intermediate work storage,
such as tape or. disc, its facilities are greatly limited in comparison with those of the BPS Tape Assembler. It is significant to note that despite all
the limitations of the BPS Card Basic Assembler,
it was the only Assembler available to early IBM
System/360 users. The BPS Tape Assembler is
now being used in most small System/360 installations.
• 141 Basic Programming Support Tape Assembler
The machine requirements necessary to use the
BPS Tape Assembler include 8K bytes of core storage, three tape units (at least one must be 9-track),
one card reader (2540 or 1442), one printer (1403,
1404, 1443, or 1445), and one I/O channel (either
Multiplexor or Selector). The Tape Assembler
supports the standard card reader-punches and
printers, as well as the IBM 2400 Series Magnetic
Tape Units, the 1052 Printer-Keyboard, the 1015
Inquiry Display Terminal, and the 2671 Paper Tape
Reader.
©
The BPS Tape Assembler contains virtually all of
the language facilities offered with the Basic Operating System's 8K Disk Assembler. The Disk Assembler is described in detail in the previous report section (420:172); this coverage should be consulted for specific information regarding the assembly language facilities available with the BPS Tape
Assembler. The Tape Assembler has the same
subset relationship to the full Operating System/
360 Assembler as that of the previously-mentioned
8K Disk Assembler.
The macro-language library facilities of the BPS
Tape Assembler are both comprehensive and flexible. The user can insert and alter macro-routines
in the library whenever desired, although not during
the actual assembly operation. The input-output
control macro-instructions (lOCS) should facilitate
the opening and closing of files, blocking and unblocking of logical records, checking and writing of tape
labels, and control of I/O device error routines.
Any input-output channel configuration of up to one
Multiplexor and two Selector Channels can be controlled by the IOCS routines. Since these routines,
as well as other macro-facilities, are under control
of the Basic Programming Support Supervisor, they
can be called into core storage as required and
overlaid whenever they are not needed. Nevertheless, the control routines that must reside in core
storage will restrict the size of object programs
that can operate in the control environment of an
8K system to a maximum of between 4K and 5K
bytes of storage for each program phase.
Program diagnostics can be utilized by calling on the
Autotest functions of the BPS Tape Control System.
However, as noted in the description of the BOS Disk
Assembler, Autotest requires a minimum of 16K
bytes of core storage to perform its operations.
Therefore, if 16K bytes of core storage are not
available for program testing, then the only available
diagnostics are standard data file and core storage
dumps.
The System/360 Basic Programming Support Tape
Assembler can assemble programs written in the
BPS Card Basic Assembler language (described
below) and the Basic Operating System's 8K Disk
Assembler language .
. 142 Basic Programming Support Card Basic Assembler
The BPS Card Basic Assembler (hereafter referred
to as the Basic Assembler) is a two-phase program
that is loaded from punched cards'. Magnetic tape,
if available, can be used as intermediate work storage for the Basic Assembler in order to speed up
the assembly process. The Basic Assembler will
operate on any System/360 system that has 8K bytes
of core storage, the standard instruction set, and
one IBM 1442 Model 2 or 2540 Card Read Punch.
An IBM 1443 or 1403 Printer and an IBM 1052 Printer-Keyboard can be used to provide program listings.
1965 AUERBACH Corporation and AUERBACH info, inc.
8/65
IBM SYSTEM/360
420: 173. 142
• 142 Basic Programming Support Card Basic
Assembler (Contd.)
The Basic Assembler language is not a unique assembly language, but is in fact a proper subset of
the larger Operating System/360 Assembler lanquage. However, the facilities of the Basic Assembler language are significantly restricted. Simplicity of programming has been the laudable goal
in limiting the Basic Assembler'S facilities. By
IBM's admission, the language features of the
Basic Assembler "are designed to greatly simplify
the writing of programs for the IBM System/360.
By avoiding unnecessary complexity, the language
features reduce program errors and, consequently,
the time required to produce a program that is
suitable for execution. ,,*
While refreshing in its lack of complexity, the
Basic Assembler possibly goes to an extreme in
its omission of all macro-facilities. The programmer must provide not only his own input-output
routines and the associated error routines, but
also whatever machine interrupt routines might be
required. The primary purpose of the Basic Assembler was to provide the first System/360 users
with an assembly program, regardless of how restricted.
The object program can be produced in absolute or
relocatable form. Symbols that are external to the
present program can nonetheless be utilized. The
Relocatable Loader program can link related object
program segments just prior to execution. The
Absolute Program Loader loads object programs
into the core storage locations specified in each
text card of the object deck.
The principal features of the Basic Assembler include the ability to translate all System/360 machine
instructions and to use up to 15 assembler control
instructions. Data constants can be represented by
alphameric characters and hexadecimal or decimal
digits, with conversions to internal machine lanquage performed by the assembler. Assembly-language symbolic labels can reference relocatable or
absolute data, and self-defined constants can be included in many instructions. Expressions with multiple terms - connected by arithmetic operators are also perIllitted in simple or compound form. A
special reassembly facility permits time-saving
during program reassembly operations by retrans*IBM Publication C28-6503-3: IDM System/360
Basic Programming Support Basic Assembler
Language.
lating only the source cards which are in error .
Object programs can be patched at execution time
by inserting "replace" cards with storage addresses
specified.
In addition to the lack of all macroinstructions,
the principal deficiences of the Basic Assembler
with repsect to the other System/360 assemblers
can be summarized as follows:
•
The use of literals is prohibited.
•
The insertion of constants in binary representation is not possible.
•
Continuation source cards containing operands,
constants, or comments that cannot fit on one
card are not permitted.
•
The capacity of the symbol table is greatly
restricted:
8K system - 275 symbols;
16K system - 1299 symbols;
32K system - 3347 symbols.
•
The maximum size of a character constant is
limited to 16 characters.
•
Scale and exponent modifiers of constants are
not permitted.
•
More than one related program segment cannot
be assembled at one time.
•
Only one general register can be specified for
use by the Assembler in the formation of storage
addresses.
It should be noted that programs written in the Basic
Assembler language are acceptable to the other Basic
Programming Support, Basic Operating System, and
Operating System Assemblers. Similarly, source
programs written in these other System/360 assembly languages are theoretically acceptable to the
Basrc Assembler, provided, of course, that they do
not use any of the language features that are not implemented in the Basic Assembler.
It is interesting to note that the Basic Assembler
will also accept programs written for the IBM System/360 Model 20 Basic Assembler, except for certain Model 20 machine instructions that are not
present in the larger System/360 models. The
mnemonic operation codes of the incompatible Model
20 instructions are as follows: BAS, BASR, CIO,
HPR, SPSW, TIOB, and XIO.
/
8/65
420: 191.1 00
IBM System/360
Operating Environment
Operating System/360
OPERATING ENVIRONMENT: OPERATING SYSTEM/360
.1
GENERAL
· 11
Identity: ..
· 12
Description
. . . . Operating System/360.
The IBM Operating System/360 consists of a comprehensive set of control and processing programs
integrated within a supervisory network to provide
coordinated and continuous operation of the System/
360 computer system. The Operating System/360
is designed in a modular fashion, in that additional
control facilities can be assembled into the system
as desired, limited only by the amount of core and
auxiliary storage available to the installation.
Disc file and/or magnetic drum storage devices
and a minimum of 32,768 bytes of core storage
are prerequisites for utilization of the Operating
System/360.
Operating System/360, the language translators
can utilize any available control services of the
Operating System. The currently-announced
language translators available with the Operating
System/360 include the following:
•
Assembler: A symbolic assembly system with
impressive addressing and sectioning flexibility,
and with extensive facilities for the use of
literals and macro-language. The Assembler is
available at the 12K, 44K, and 200K levels; see
Section 420: 171 for language descriptions and
assembly times.
•
FORTRAN IV: The 200K version closely resembles the IBM 7090/7094 FORTRAN IV
language, which is described in Section 408: 162.
The 200K Operating System/360 version includes
few restrictions and several useful extensions to
the original FORTRAN IV language. A 12K version of the language is also offered with the
Operating System, for use in a 32K-byte minimum core storage environment. This language
is a restricted subset of the 200K FORTRAN IV
offering no complex, logical, or double precision facilities. Both of the Operating System/
360 FORTRAN IV languages are described in
Section 420: 161.
•
COBOL: Two versions of the COBOL language
have been announced for use with the Operating
System/360: a 12K ,estricted version for compilation with a minimum of 32K bytes of core
storage, and a 44K version that includes nearly
all of the required features of COBOL-61 plus
some valuable extensions (a SORT verb, a
Report Writer, the MOVE CORRESPONDING
option, and debugging facilities). The Operating
System/360 COBOL languages are described in
Section 420: 164.
•
Programming Language I (PL/I): Available at
the 44K and 200K Operating System/360 program
design levels, PL/I is designed to offer a single
set of language facilities amenable to both
scientific and commercial computer users.
Some features include computational facilities
similar to FORTRAN, string manipulation, and
dynamic storage allocation. Section 420: 167
presents a detailed report on PL/I.
•
Report Program Generator (RPG): The Operating System/360 RPG, available in a single
version at the 12K program design level, requires a minimum of 32K bytes of core storage.
In addition to the standard report writer features,
the Operating System/360 RPG has the ability to
operate on· input data from disc storage, and
offers the facilities to perform table-lookup
operations and to insert routines written in other
source languages. Please refer to Section
4~0: 151 for further information on the Report
Program Generator.
Two basic design levels for the Operating System/
360's control programs have been announced by
IBM:
•
•
A 12K version designed to provide 32K and
larger systems with simple stacked-job processing, input-output device control, operator
communication, and service functions for
machine and programming errors. If core
storage is increased to 65K or 131K bytes,
additional control functions can be added to the
basic design if data transcriptions are to be
run concurrently with a main program or if
full telecommunications operations are desired.
A 200K basic design level designed to function
in an environment of 262K bytes of core storage,
providing full multiprogramming capabilities
(the processing of two or more main programs
concurrently), multiprocessing capability
(linking the facilities and processing power of
two or more central processors), the facility to
control all elements of a computer processing
job from a remote location, and the techniques
and routines necessary to control time-shared
operation of a central computer system.
According to IBM, the 12K version of the Operating
System/360 will be delivered during the fourth
quarter of 1965. The full Operating System is not
expected to be delivered until the second quarter
of 1966.
· 121 Processing Programs
I
~
The processing programs provided and controlled
by the Operating System/360 include language translators, service programs, and the user's own
problem programs. The language translators are
designed at three basic design levels: 12K, 44K,
and 200K. The smaller versions are in all cases
proper subsets of the larger, complete language
versions. As with all programs controlled by the
©
1965 AUERBACH Corporotion and AUERBACH Info, Inc.
8/65
IBM SYSTEM/360
420: 191.121
module. If the combined load module exceeds the
limits of available core storage, the Linkage
Editor segments the program as necessary and
provides the linkages to the overlay segments
necessary for proper program execution. Optional
output of the Linkage Editor program is a list of all
processed control statements and a module map of
each program control section and overlay that has
been combined into the single program load
module. If program segmenting is neither desired
nor required by storage restrictions, the Linkage
Editor's functions can be bypassed. The Linkage
Editor operates as a processing program and is
available in design levels of 18K and 44K bytes.
• 121 Processing Programs (Contd.)
In addition to the language translators, a variety of
service programs are included among the processing programs available with the IBM Operating
System/360.
The most noteworthy of the service programs
provided is the Linkage Editor, a sort of intermediate assembler. The principal function of the
Linkage Editor is the combining of separately assembled or compiled "object modules" of a program into one "load module" in a format suitable
for loading and execution under control of the
Operating System. Programming of individual
control sections by several different programmers
thus becomes entirely feasible, and errors in one
segment will not necessitate recompilation of the
entire program.
IBM has provided us with estimated Linkage Editor
processing times based on three sample System/
360 configurations. Table- I describes these sample
configurations and indicates the several timing
factors that must be summed to arrive at a total
time for a Linkage Editor job that is performed in
conjunction with the Operating System/360.
Guided by programmer-supplied control card
statements, the Linkage Editor reconciles and
links into one dictionary the information contained
in the External Symbol Dictionary and the Relocation Dictionary of address constants that are
produced with every assembled or compiled object
An IBM-supplied timing example shows the estimated time requirements to perform a Linkage
Editor (18K version) job that consists of combining
TABLE I: LINKAGE EDITOR PROCESSING TIME FACTORS
Sample Configurations
Configuration Components
Configuration A
System/360 Model 30 with 32K bytes of core storage, one Multiplexor
Channel, one Selector Channel, one 2540 Card Read Punch, one 1403
Printer (Model 3), two 2311 Disk Storage Drives, and the Operating
System/360's sequential scheduler. *
Configuration B
System/360 Model 50 with 65K bytes of core storage, two Selector
Channels, two 2311 Disk Storage Drives, and the Operating System/
360's priority scheduler. **
Configuration C
System/360 Model 65 with 262K bytes of core storage, two Selector
Channels, two 2301 Drum Storage units, four 2311 Disk Storage
Drives, and the Operating System/360's concurrent job scheduler. **
Time in Seconds
Estimated Timing Factors
For Linkage Editor initiation
For each Linkage Editor
process
For each library access
For each object program combined
For each byte of object
programs combined
For each 2, 048-byte control
section in the object
programs
For each byte of the combined loadable object
program
44K Linkage Editor,
Configuration B
44K Linkage Editor,
Configuration C
18.3
12.8
2.87
5.4
2.0
3.9
1.4
1.6
0.32
0.75
0.37
0.088
0.0004
0.00009
0.00002
0.46
0.088
0.045
0.00084
0.00018
0.0001
18K Linkage Editor,
Configuration A
*
Control card input is assumed to be entered through the Card Read Punch, .and program input
and output are assumed to be read from and written to disc storage.
**
Input and object-program output are assumed to be read from and written to disc storage.
(Contd. )
8/65
420: 191.122
OPERATING ENVIRONMENT: OPERATING SYSTEM/360
. 121 Processing Programs (Contd.)
six object program modules stored in two on-line
libraries on sample Configuration A. In this example, it is assumed that the size of the object
program load module will be 10,000 bytes. The
estimated time in seconds to perform this Linkage
Editor job, reading the object program modules
from disc storage and writing the combined load
module to disc storage, is 37.6 seconds.
The other service programs provided by the
Operating System/360 are described in Section
420: 151, Problem Oriented Facilities. Routines
included in this group of Operating System/360
facilities are two generalized Sort/Merge programs,
the Test Translator (TESTRAN) diagnostic routines,
data transcription and data set file maintenance
routines, an IBM 7090/7094 Support Package, and
Simulation programs for IBM 1410, 7010, 7070,
7074, 7080, 7090, and 7094 computer systems.
. 122 Control Programs
The remainder of this report section is devoted to
the Control Programs of the Operating System/360.
These programs can be logically grouped into three
major categories: Supervisor, Job Scheduler, and
Master Scheduler. Each category includes many
distinct control functions. All control functions are
governed and monitored by the Supervisor, just as
user problem programs are controlled, yet they
are not considered to be integral parts of the Supervisor itself. The unique mix of control programs
selected for use in a given installation's Operating
System will be determined by the amount of available core storage and the number of direct-access
devices available for residence of both the processing and control programs (including users' problem programs), and data set files and their
catalogues. Consult Tables III through VI of this
report section for an indication of configuration
requirements necessary to support the various
functions of the Operating System/360.
The paragraphs that follow indicate the various
functions of each of the principal categories of
Operating System/360 executive control.
Supervisor: The Supervisor is the control center
of the complex of Operating System/360 facilities.
To perform its coordinating functions, the Supervisor must receive control of the central processing
unit py means of interrupts. The interrupts can
result either from specific requests for supervisory services from another part of the operating
system or from a, problem program, or from automatic signals generated by the.hardware to indicate
such conditions as the end of an I/O operation or a
device malfunction. Through use of privileged instructions, the Supervisor - once summoned attempts to continue the jobs being processed by
entering error routines and abnormal condition
routines, as well as by attempting to reassign
input-output units.
I
\
"-
A principal function of the Supervisor is the
scheduling and controlling of all input-output operations (referred to collectively in IBM publications
as "data management"). Broadly stated, data
management control enables all types of "data sets"
(programs and input-output information) to be
systematically organized and stored on auxiliary
storage devices, and to be retrieved, processed,
and re-stored on external storage devices at the
conclusion df the data processing task.
Using seven basic data access techniques that are
a major expansion of the earlier IBM Input-Output
Control Systems (lOCS) , the Supervisor controls
input-output operations according to. file and
storage descriptive information supplied by the
programmer and by Job Control statements (described below) formulated at program execution
time.
The seven data access methods, designated according to the characteristics of the data set to be
accessed, are the Queued Sequential Access Method
(QSAM), the Basic Sequential Access Method
(BSAM), the Basic Partitioned Access Method
(BPAM), the Queued Indexed Sequential Access
Method (QISAM), the Basic Indexed Sequential
Access Method (BISAM), the Basic Direct Access
Method (BDAM) , and the Queued Telecommunication
Access Method (QT AM). These various methods of
automatic input-output control perform the following
functions: reading and writing data; blocking and
unblocking logical records; oyer lapping data transfers with processing; reading, writing, and
checking standard labels on storage volumes and
data sets; detecting error conditions and correcting
them when possible; and providing linkages to userwritten device error-routines and non-standard
label routines.
It should be noted that the practicality, efficiency,
and overall effectiveness of the so-called automatic
I/O routines of the Operating System/360 are contingent upon the skill with which the associated data
sets are organized, and upon the general competence of the System/360 programmer who specifies
buffer and storage sizes and sequences of I/O
operations. Indiscriminate use of facilities at
hand can lead to a great waste of external storage
space and to extremely inefficient program execution.
In addition to the I/O facilities described above,
the Operating System/360 Supervisor performs
various other functions. A partial listing includes:
allocating main storage space required by programs
during their execution; sharing common areas of
main storage among various routines; loading
object programs into main storage; controlling the
concurrent execution of programs and routines;
and providing timing services and other accounting
information required for indiVidual job and total
system logging.
Job Scheduler: Job scheduling within the Operating
System/360 is controlled by statements expressed
in a "job control language," which the programmer
inserts in the input stream at the time of program
execution. The job control statements specify the
programs to be executed, the avallable inputoutput configuration, input-output disposition, and
data set characteristics. Specific input-output
devices can also be specified by means of these
statements.
The Job Scheduler provides for a continuous flow of
jobs through the computer system. It consists of
three principal functions: reader/interpreter,
© 1965 AUERBACH Corporotion and AUERBACH Info, Inc.
8/65
IBM SYSTEM/360
420: 191. 130
. 122 Control Programs (Contd.)
can inform the system of changes in I/O configuration and can request information on the status of the
overall system. The Master Scheduler also
initializes the Job Scheduler at the operator's command. All system messages to the operator are
also controlled by the Master Scheduler.
initiator /terminator, and output writer. After the
job control statements are read and analyzed, the
Job Scheduler prepares program control tables for
each 'program, ensure's that the control information
is queued in priority sequence (if so specified), and
ensures that all the necessary I/O devices are
allocated. The Job Scheduler then passes control
to the Supervisor to begin program execution. The
terminator performs standard finalizing functions
at the conclusion of the job step. Output data sets
can optionally be stored on a direct-access device
for transcription at a later time to a system
printer or punch.
Several versions of the Job Scheduler are available,
offering facilities that range from processing jobs
in a simple stacked mode to true multiprogramming
and time-sharing operations. When more than one
job is scheduled for simultaneous execution (assuming the presence of the necessary hardware
configuration), the Job Scheduler controls the
dynamic allocation of the system's resources, the
switching of central processor control from one
task to another according to established priorities,
and main storage roll-out when necessary. These
functions of the Job Scheduler are referred to as
"task management" facilities. IBM estimates that
a minimum of 30K to 40K bytes of core storage
will be required for Job Scheduler and Supervisor
control functions in support of multiprogramming
operations.
The specific job control capabilities offered to the
operator by the Master Scheduler include cancelling
or holding a scheduled job, changing job priorities,
varying the status of input-output devices, entering
information into the system log, and requesting log
print-outs.
The various processing and control functions of the
Operating System/360 described above require the
use of considerable core storage and auxiliary
direct-access storage devices. Tables III through
VI are designed to show the amount of storage and
peripheral equipment required for specific Operating System functions and operational objectives.
.13
IBM has provided us with Job Scheduler timing
estimates on the three sample System/360 configurations shown in Table 1. Table II indicates the
timing factors that must be summed to arrive at a
total time for a Job Scheduler operation that is
performed in conjunction with the Operating System/ • 14
360.
· 15
Master Scheduler: The last of the three principal
control functions of the Operating System/360 is the .2
Master Scheduler, consisting of the routines necessary to control two-way communication between the · 21
System/360 operator and the Operating System.
By issuing commands to the Master Scheduler via
· 211
such input-output devices as the 1052 PrinterKeyboard or the 2250 Display Unit, the operator
Availability
Basic monitor functions
(2400 Series Tapes,
2302, 2311, 2314,
2671, 1402, 1403,
1443, and 2201): . . . . 4th quarter, 1965.
Time-sharing and data
communication functions (2701 and 2702
with 1030, 1050,
1060, Telegraph
Terminal Controls
and 2250 Modell,
7340, 7320, ,2301,
2302, and 2321): •... 2nd quarter, 1966.
Originator: •....••• IBM Corporation•
Maintainer: . . . . . • . . IBM Corporation.
PROGRAM LOADING
Source of Programs
Programs from on-line
libraries: . . . . • . • . System Program Library
on direct access device.
TABLE II: JOB SCHEDULER TIMING FACTORS
Time in Seconds
Estimated Timing Factors
For each processing job
to be scheduled
For each program within
the lob
For each data set or data
control definition statement
*
Priority Scheduler,
Configuration B**
Priority Scheduler,
Configuration C**
4.6
2.8
0.46
6.7
4.5
0.66
1.4
1.1
0.35
Sequential Scheduler,
Configuration A*
/
See Table I for configuration components; job control cards are assumed to be entered
through the Card Read Punch.
** See Table I for configuration components; job control card data is assumed to be read
from disc storage.
(Contd.)
8/65
420: 191.211
OPERATING ENVIRONMENT: OPERATING SYSTEM/360
TABLE III: FIXED CORE STORAGE REQUIREMENTS IN TYPICAL OPERATING
SYSTEM/360 COMP VTING SITVA TIONS
Control
Program
Storage
in Bytes*
Control Program
Operational Mode
Computer System Input-Output
Device Configuration
Basic I/O Total Bytes
Support
of Core
Storage Storage Rein Bytest
quired
Sequential scheduling of
jobs in a non-multiprogramming environment.
11,882
One Multiplexor Channel
One Selector Channel
Two 2311 Disk Storage Drives
One card read punch
One printer
One operator's console
(Ten input-output requests can
be queued on channels)
1,170
12,992
Sequential scheduling of
two concurrent jobs, with
each program alternatively occupying core
storage; one of the jobs
is a Telecommunications
task.
14,860
One Multiplexor Channel
Two Selector Channels
Four 2311 Disk Storage Drives
Four magnetic tape drives
One card read punch
One printer
One operator's console
One 2701 Data Adapter Unit
One communications line
group with four lines
(Ten input-output requests can
be queued on channels)
2,111
16,971
Priority scheduling of two
cohcurrent jobs in a
multiprogramming environment, with each program occupying a
partition of storage; one
of the jobs is a Telecommunications task; input
reader /interpreter is included.
19,326
One Multiplexor Channel
Two Selector Channels
Four 2311 Disk Storage Drives
Four magnetic tape drives
One card read punch
One printer
One operator's console
One communications line
group with four lines
(Ten input-output requests can
be queued on channels)
2,111
21,437
Single job scheduling with
20 input entries on the
work queue catalogued in
main storage; priority
scheduling, interval and
job step timing are
included.
33,186
One Multiplexor Channel
Two Selector Channels
Four 2311 Disk Storage Drives
One 2302 Disk Storage Drive
Six magnetic tape drives
One card read punch
One printer
One operator's console
(Thirty input/output requests
can be queued on channels)
2,638
35,824
Multiprogramming of three
jobs with a variable number of tasks queued on
disc storage and selected
by priorities; interval
and job timing, and input reader/interpreter,
and an output writer
(considered to be
operating) are included.
48,003
One Multiplexor Channel
Two Selector Channels
Four 2311 Disk Storage Drives
One 2302 Disk Storage Drive
Three magnetic tape drives
One card read punch
One printer
One operator's console
(Twenty input-output requests
can be queued on channels)
2,320
50,323
*
In addition to the control functions tabulated in Table III, the overlay supervisors require
from 400 to 800 bytes, and the Test Translator control requires from 3,600 to 4, 100 bytes
of core storage.
t Logical record handling will require another 2K to 3K bytes of core storage, depending upon
the data management techniques utilized in the problem program.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
420:191.212
IBM SYSTEM/360
TABLE IV: AUXILIARY STORAGE REQUIREMENTS FOR OPERATING SYSTEM/360
SYSTEM PROGRAM RESIDENCE
Number of Tracks Required
2311 Disk
Storage Drive
2302 Disk
Storage
2301 Drum
Storage
Control Program including all functions for job,
task, and data management (except indexed
sequential data access method routines, and
the interpreter for TESTRAN)
228
153
37
Basic indexed sequential access method
(BISAM) and queued indexed sequential
access method (QISAM)
100
67
17
System Programs
Assembler
(12K)
67
45
11
Assembler
(44K)
49
33
8
Assembler (200K)
100
67
16
Macro Library
463
329
88
COBOL (12K)
91
61
15
COBOL (44K)
96
70
18
FORTRAN IV (12K)
45
33
8
FORTRAN IV (200K)
77
52
13
Linkage Editor (18K)
13
9
3
. Linkage Editor (44K)
16
11
3
36
31
8
291
195
47
23
15
4
Repor:t Program Generator (12K)
Sort/Merge (12K)
TESTRAN interpreter
· 212 Independent
.312 Occupation of working
programs: . . . . . . . . in System Program Library
storage: • . . • • . . . . actual load addresses are
or from unit record equipassigned by Job Scheduler
ment, magnetic tape, or
at load time.
direct access device .
. 213 Data: . . . . . . .. . .. as required by users' pro• 32 Input-Output Units
grams.
· 214 Master routines: . . . . System Program Library
.321 Initial assignment: ... actual units are assigned by
(on disc or drum).
the operating system .
. 322 Alteration: . . . . . . . . provided by the operating
· 22 Library Subroutin es: .. Supervisor provides autosystem or by the operator.
matic library call facility. .323 Reassignment: . . . . . . provided by the operating
system or by the operator.
• 23 Loading Sequence: ... sequential loading of programs as provided by
.4
RUNNING SUPERVISION
operator on system's
direct access device or
.41 Simultaneous Working: controlled by operating
on card reader.
system input-·output control programs .
.3
HARDWARE ALLOCATION
.31
Storage
· 311 Sequencing of program
for movement between levels: . . . . . . Linkage Editor provides for
multiphase linking of
"control sections" (independent and relocatable
parts of segments).
.42
Multiprogramming: ... controlled by Job Scheduler
in operating system.
. 43
Multi-sequencing: . . . . controlled by the Supervisor
in 200K Operating System
(operates an attached peripheral computer and
transmits stacked jobs
and results between computers).
(Contd.)
8/6~
OPERATING ENVIRONMENT: OPERATING SYSTEM/360
420: 191.430
TABLE V: CORE STORAGE REQUIREMENTS FOR VARIOUS OPERATIONAL OBJECTIVES
Objective
~,.,
Core Storage
Required, Bytes
Basic Limitation
To run under control of the
Operating System/360
32K
Design level of minimum Operating
System/360
To stack jobs one after another on
the system input device
32K
Design level of required Operating
System control programs
To operate .a simple inquiry system
either alone or in parallel with a
suitable small main program
32K
Design level of required Operating System control programs
To run restricted FORTRAN or
COBOL compilations
32K
Design level of basic compilers
To operate a simple inquiry system
in parallel with a 12K level program,
such as a compilation
32K
Space limitations
To operate a data communication
system either alone or in parallel
with a main program
32K
Design level of required' Operating
System control programs
To run compile-and-execute
operations, with library calls in the
compiled program being handled
automatically
64K
Design level of required Operating
System control programs
To run peripheral programs such as
card-to-tape conversions in parallel
with a single main program
64K
Design level of required Operating
System control programs
To run a comprehensive COBOL
compilation
64K
Design level of full COBOL
compiler
To run a comprehensive FORTRAN
IV compilation
256K
Design level of full FORTRAN IV
compiler
To run a System/360 installation
from a remote location
256K
Design level of required Operating
System control programs
To allow direct connection between
two computer systems
256K
Design level of required Operating
System control programs
To operate in a full multiprogramming environment
256K
Design level of required Operating
System control programs
-
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
420: 191.431
·IBM SYSTEM/360
TABLE VI: MINIMUM CONFIGURATION REQUIREMENTS FOR OPERATING
SYSTEM/360 PROGRAMS
System Programs and
Storage Requirements (1)
Assembler:
12K
Program
Library (4)
Input-Output
Intermediate
Storage (4)
Instruction
Set
4 tapes or 1
disc file
Standard
4 tapes or 1
disc file or
1 drum
3 tapes or 1
disc file or
1 drum
Standard
4 tapes or
1 disc file
Scientific (2)
2 tapes or 1
disc file or
1 drum
Scientific (2)
4 tapes or 1
disc file
Commercial (3)
1 disc file
or Ldrum
2 of card, tape,
printer units
or 1 disc file
same as above, or
1 drum
4 tapes or 1
disc file or
1 drum
Commercial (3)
1 disc file
or 1 drum
same as above, or
1 drum
Scientific (2)
1 disc file
or 1 drum
same as above, or
1 drum
2 tapes or 1
disc file or
1 drum
1 tape or 1
disc file or
1 drum
Report Program
Generator:
12K
1 disc file
2 of card, tape,
printer units or
1 disc file
4 tapes or 1
disc file
Standard
Utility Programs:
12K
1 disc file
2 of card, tape,
printer units or
1 disc file
depends on
application
Standard
1 disc file
or 1 drum
card, tape, disc
file, or drum
3 tapes or 1
disc file or
1 drum
Standard
1 disc file
2 of card, tape,
printer units
or 1 disc file.
same as above
or 1 drum
44K
1 disc file
or 1 drum
200K
1 disc file
or 1 drum
same as above
1 disc file
2 of card, tape,
printer units
or 1 disc file
same as above,
or 1 drum
FORTRAN IV:
12K
200K
1 disc file
or 1 drum
COBOL:
12K
1 disc file
44K
PL/I:
44K
200K
Control Programs
Standard
Scientific (2)
(1)
At least 13K additional bytes are required for the basic Control Programs.
(2)
Floating-Point Arithmetic option is required.
(3)
Decimal Arithmetic option is required.
(4)
A single disc file or drum can provide both program library storage and· intermediate storage.
(Contd.)
8/65
420: 191.440
OPERATING ENVIRONMENT: OPERATING SYSTEM/360
.44
Errors, Checks, and Action
· 632 Change of normal
progress: . . . . . .
Check or
Interlock
Loading input error:
Action
I/O equipment
check
Allocation impossible: hardware check
In-out error single:
In-out error persistent:
· 45
optional
interrupt.
optional
interrupt.
I/O equipment
check
optional
interrupt.
I/o equipment
check
check
check
check
check
check
check
optional
interrupt.
optional
interrupt.
interrupt.
interrupt.
interrupt.
interrupt.
interrupt.
interrupt.
hardware check
interrupt.
Storage overflow:
hardware check
Invalid instructions:
Arithmetic overflow:
Underflow:
Invalid operation:
Improper format:
Invalid address:
Overwriting of
forbidden area:
hardware
hardware
hardware
hardware
hardware
hardware
Restarts
. 451 Establishing restart
points: • . . . . . . . . . checkpoint facilities are
provided to record the
current status of a program on an auxiliary
storage device.
· 452 Restarting process: .• can be initiated by operator
or by problem program.
.5
PROGRAM
DIAGNOSTICS:
. called in from system
library by Supervisor as
requested by problem
program.
•6
OPERATOR CONTROL
. 61
Signals to Operator
· console typewriter or display device.
· 611 Decision required by
operator: . . . . .. . supervisor-call interruption.
. 612 Action required by
operator: . . . . . . . . supervisor-call interruption.
• 613 Reporting progress
of run: • . . . . . . . . . log of errors, log of
machine time.
.7
LOGGING
.71
Operator Signals: . . . . provided by Operating
System.
.72
Operator Decisions:
. provided by Operating
System.
· 73
Run Progress: . • ..
. provided by Operating
System.
.74
Errors:
· provided by Operating
System.
..75
Running Times:
.76
Multiprogramming
Status: . . . . . • . . . . provided by Operating
System as requested by
the operator.
.8
PERFORMANCE
.81
System Requirements
.. provided by Operating
System, normally by use
of the Interval Timer.
.811 Minimum configuration: . • . . . . . . . . . • Simple Stacked-Job Processing - 32K .
Extended Stacked Job Processing (error procedures
and "load-and-go" facility
- 64K.
Simple Inquiry - 32K.
Telecommunications - 32K.
Multiprogramming - 256K.
Multiprocessing - 256K.
All programs require the
Program Library to be on
disc file or drum.
System Generation and
Loading require 3 magnetic
tapes and one direct access
device •
.812 Usable 'extra
facilities: . .
. . all .
.813 Reserved equipment: .. core storage occupied by
the control routines and
operating system programs .
System Program Library
lmit.
· 82
System Overhead
• 62
Operator's Decisions: . console typewriter or display device.
.821 Loading time: . . . . . . depends upon input device .
· 822 Reloading frequency: .. control programs normally
remain in core storage .
. 63
Operator's Signals
.83
· 631 Inquiry: ... . . . ..
Program Space
Available: .. ,
. console typewriter or display device.
emergency-interrupt button
on console.
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
· minimum control program
requires 13K bytes of
main storage; the remainder is available for
program use.
8/65
IBM SYSTEM/360
420: 191.840
. 84
Program Loading
Time: •....••••.. IBM has provided us with
-.
estimated program fetch
times for the three sample System/360 configurations shown in Table 1.
Table VII shows the factors
that must be summed to
arrive at a total fetch
time. It is assumed that
the programs are fetched
from disc storage residence and loaded into core
storage.
• 85
Program Performance: . . . . . . . . timing estimates for the
Linkage Editor, Job
Scheduler, and Program
Fetcil and Loader are
listed in Tables I, II, and
VII within this report
section. Overhead considerations related to the
use of the other supervisory services of the
Operating System/360
have not been made
available to date.
TABLE VII: PROGRAM FETCH TIMING FACTORS
Time in Seconds
Estimated Timing Factors
Configuration A* Configuration B* Configuration C*
:
For program fetch initialization
For each 2048-byte program block
fetched
For each byte of the fetched program
* See Table I for configuration components.
8/65
12.3
0.026
0.000008
8.9
2.06
0.026
0.018
0.000008
0.000001
420: 192.100
IBM System/360
Operating Environment
Basic Operating System/360
OPERATING ENVIRONMENT: BASIC OPERATING SYSTEM/360
.1
GENERAL
. 11
Identity: . . . . • . . . • . Basic Operating System/360.
.12
Description
The IBM Basic Operating System/360, announced
in March 1965, consists of an integrated set of
control and processing programs designed to provide users of limited-storage System/360 processor
models with some of the system-monitor and language facilities that were previously available only to
users of large-scale System/360's through the
Operating System/360 (Section 420:161). The control provisions offered with the Basic Operating
System (BOB) are limited in their scope, yet they
serve a useful function in the automatic handling of
program/machine interrupt conditions and inputoutput device control.
Two basic .program design levels have been
announced by IBM for implementation of the Basic
Operating System/360:
•
A 4K control-program version is designed to
provide users of 8K and larger disc-fileoriented systems with simple stacked-job operation and input-output control. The 4K level
of the Basic Operating System requires a minimum of one IBM 2311 Disk Storage Drive for
system residence of all control and processing
programs. This disc-oriented 4K version of
the Basic Operating System can support any
I/O channel configuration of up to one Multiplexor Channel and two Selector Channels.
Special versions of the Assembler and Report
Program Generator language, as well as several general service programs described below,
will be available for use with the 4K Disk Basic
Operating System.
•
A 6K control-program version for residence
on either disc or tape is offered to users of
16K and larger systems. The use of this
version of the Basic Operating System in its
minimum 16K-byte storage environment offers
simple operator-system communication (via
the IBM 1052 Printer-Keyboard and/or 1015
Display Unit) and increased input-output device
flexibility, in addition to the simple stacked-job
processing offered in the smaller version. A
larger range of improved language facilities
is also available at the 16K Basic Operating
System level, including restricted versions of
FORTRAN, COBOL, and PL/I.
Operating in an environment of 32K bytes of
core storage, the Tape and Disk Resident 16K
Basic Operating System can control limited
multiprogramming operations, although the overhead requirements of the control functions will
render the cost of Tape Resident multiprogramming
prohibitively high. The extent of BOB-controlled
multiprogramming in a 32K environment is limited
to concurrent processing of one problem program
(user-written or an IBM 10K language compiler)
and up to two IBM-supplied data transcription
routines.
Remote-terminal processing can be supported by
the 16K Disk Resident Basic Operating System
when a minimum of 32K bytes of core storage is
available. The control functions of the BOS will
respond to the inquiry, permit the transaction to
be processed, and transmit the reply message
prior to returning control to the problem program
in the "background." The co-existing background
program (which runs whenever no inquiry is being
processed) can be a 10K language translator or a
disc-resident problem or utility program.
According to IBM, the 4K disc-oriented Basic
Operating System control functions will be delivered
during the third quarter of 1965. The 6K Tape and
Disk Resident control functions are expected during
the fourth quarter of 1965. The control routines
necessary to supervise multiprogramming and
remote terminal processing, however, will not be
available until the second quarter of 1966.
. 121 Processing Programs
The processing programs provided and controlled
by the Basic Operating System/360 include language
translators, service programs, and the user's
problem programs. The language translators are
designed at two basic design levels: 4K and 10K
bytes. Like all programs controlled by the Basic
Operating System/360, the language translators
can utilize any of the available supervisory services.
The language translators available with the Basic
Operating System/360 include the following:
•
Assembler: A symbolic assembly system
with good address modification and macrolanguage facilities, available at the 4K and
10K program design levels. These versions
of the System/360 assembly languages are
described in Section 420:172.
•
FORTRAN IV: One FORTRAN IV compiler
is available for use with the Basic Operating
System - a 10K version designed for use
with the Tape or Disk Resident control functions. The language encompasses the features
listed in the proposed ASA Basic FORTRAN.
The 10K FORTRAN language, described in
Section 420:162, is nearly the same as the 12K
design-level language supplied with the Operating
System/360 (see report Section 420:161).
•
COBOL: The COBOL compiler offered with the
BOS is of the 10K design level, requiring a
minimum of 16K bytes of core storage and one
IBM 2311 Disk Storage Drive or three IBM 2400
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
I~M
420: 192.121
.121 Processing Programs (Contd.)
Series tape units for compilations. This version
of the COBOL language is described in Section
420:165.
•
•
Programming Language I (PL/I): Expected to
appear in the first quarter of 1966, the 10K
PL/I compiler can be used on any 16K discfile or three-tape System/360 equipped with
the Floating Point and Decimal Arithmetic optional features. The facilities and features of
mM's latest language, designed for both commercial and scientific users, are described in
Section 420:167.
Report Program Generator (RPG): Two versions
of the RPG are offered for use with the Basic
Operating System/360:a disc-oriented generator
for 8K System/360's and an expanded version
for use with 16K tape or disc systems. In addition
to standard report writing and file maintenance
capabilities, the BOS Report Program Generator
promises several additional facilities, such as
table-lookup control and multiple input-output
device options. Further information about the
RPG is available in Section 420:152.
In addition to the language translators, a variety of
service programs are included among the processing
programs available with the Basic Operating
System/360. Significant among the service programs
are the Linkage Editor, Librarian, and System Generator programs.
The principal function of the Linkage Editor program
within the Basic Operating System's environment is
to edit into the core' image library every program
that is to be run under control of the BOS. The
output of the language translators is produced in
a relocatable form, which must then be proQessed
by the Linkage Editor to form executable, nonrelocatable program phases which reside in the BOS
core image library. Once catalogued in the library
via the Linkage Editor, the object programs can be
called forth for execution by means of Job Control
statements at program load time.
In more complex situations, the Linkage Editor can
also serve the function of linking together into one
core image program the relocatable output (object
modules) of various separate assembly-type operations. All references between program segments
or control sections are resolved by the Linkage
Editor, and all segments receive specific core storage locations integrated within a single executable
program.
The Librarian is a set of service programs designed
to maintain, service, and organize the three separate
libraries of the Basic Operating System/360.
Essential to every BOS environment is the core
image library of programs in load format. Also
recommended for use with the system are a macroroutine library and a relocatable object module
library (to permit combining with other modules
without reassemblies). The Librarian provides
services to add to or delete from the components
of the three libraries.
The System Generation service program assembles
the Supervisor program from cards (unless supplied
8/65
SYSTEM/360
on disc with system delivery) and edits it into its
residence on disc or tape. By using Job Control
cards (described below) the System Generation
progliam can further provide for loading of all
available language and service programs into the
resident core image library.
The other service programs provided by the Basic
Operating System/360 are described in Section
420:152, Problem Oriented Facilities. Routines
included in this group of facilities are two Sort/
Merge programs for 8K and 16K systems, an
Autotest program to aid in debugging operations
(requiring a minimum of 16K bytes of core storage),
and several data transcription utility programs,
including routines to support the mM 2321 Data
Cell Drive.
.122 Control Programs
The remainder of this report section is devoted to
the Control Programs of the Basic Operating
System/360. These programs can be logically
grouped into two major categories: the Supervisor
and the Job Control programs. The many distinct
control functions within each category ,are themselves
coordinated by the Supervisor to provide the problem
programs with automatic services and to ensure
that automatic transitions occur from phase to
phase within a program and from job to job within
the total processing environment.
The two sections that follow indicate the specific
control functions of each of the two principal
categories of Basic Operating System/360 executive
control.
Supervisor: The Supervisor is the control center
of the complex of BOS facilities. Part of the Supervisor always resides in core storage and occupies
a maximum of either 4K or 6K bytes, depending on
the version of Basic Operating System selected.
Certain other routines are kept in the core image
library (on disc or tape) and are called into a
reserved transient area of core storage by the
Supervisor as necessary.
To perform its coordinating functions, the Supervisor
must receive control of the central processor by
means of interrupts. The interrupts can result
from specific requests for supervisory services
from another part of the operating system or from
a problem program. Interrupts can also result
from automatic signals generated by the hardware
to indicate such conditions as the end of an inputoutput operation or a machine or program malfunction. Depending on the type of interrupt code generated, the Supervlsor determines the proper
routine to enter in order to continue processing or
resolve the problem situation.
A prinCipal function of the Supervisor is its Channel
Scheduler control, through which it regulates all
input-output operations. Channel Scheduler control consists of several routines resident in the
Supervisor that perform the functions of Physical
IOCS (Input-Output Control System): scheduling or
queuing of I/O requests on each channel, starting
and stopping of input-output operations, and handling
of I/O interrupts. The Channel Scheduler routines,
working in conjunction with the hardware design
(Contd. )
420: 192.122
OPERA TING ENVIRONMENT: BASIC OPERATING SYSTEM/360
.122 Control Programs (Contd.)
standard labels on tape and disc; and a checkpoint/
restart procedure that writes checkpoint records
recording job status at various intervals and restarts a job at a given checkpoint after an interruption.
of the central processor, provide for the simultaneous execution of instructions with data transfers
(read/write/compute overlap).
The Logical IOCS routines, for blocking and deblocking records, sWitching between multiple I/O
areas, handling end-of-file conditions, and label
checking and writing, are not part of the Supervisor
itself, but they call upon the services of the Supervisor. The Logical IOCS routines are assembled
into the object program from macro-instructions in
the source program.
Job Control: The Job Control program provides the
linkage between processing jobs to ensure that the
Basic Operating System and its supervised problem
programs operate in a continuous mode. The Job
Control program does not reside in core storage but
is called in at the conclusion of a job to provide
transition to the next processing step. It is called
into storage to perform its functions either by the
initial program loading (IPL) procedure after loading
the Supervisor, by the Supervisor itself (after an
abort job or dump operation), or by the problem
program at normal end-of-job time (by means of the
EOJ macro-instruction in the Assembler language).
Table I shows execution times for the BOS 10CS
routines; the indicated times refer to consecutive
records processed on a System/360 Model 30.
In addition to the I/O facilities mentioned above, the
Basic Operating System/360 Supervisor offers
several other control functions. A partial listing
includes: error routines for each type of inputoutput device; routines to provide messages to and
accept replies from the operator of the system via
the mM 1052 Keyboard-Printer; a system loader
routine to retrieve all scheduled programs from the
core image library and load them into core storage
for processing; routines for checking and writing
The Job Control program performs various functions
on the basis of information provided in job control
cards, as read in the input stream at the time of
program execution. These functions are basically:
•
To prepare a program for execution by transmitting its core image library directory to the
Supervisor's system loader, or by first directing
the Linkage Editor program to perform its core
TABLE I: BOS I/O CONTROL ROUTINE TIMES FOR CONSECUTIVE RECORDS PROCESSED
ON A MODEL 30
16KBOS
8KBOS
Disc
Other
Disc
Other
GET/PUT each record
of unblocked file or
first record of block
in blocked file:
2.9 msec
2.0 msec
1.7 msec
1. 2 msec
Additional time for
successive records
after first record in
blocked file Fixed length:
Variable length:
0.4 msec
0.4 msec
0.4 msec
0.4 msec
0.4 msec
unavailable
0.4 msec
Additional time (plus
GET/PUT time) if
separate work area Fixed length:
Variable length:
0.2 msec
0.3 msec
0.2 msec
0.3 msec
0.2 msec
unavailable
0.2 msec
OPEN and CLOSE file
routines No labels:
Labeled tape input:
Disc input:
Additional time for
each disc label:
Labeled tape output:
Dis c output:
Additional time for
each label in file
directory:
Additional time for
each label deleted
from directory:
8/65
650 msec
850 msec
550 msec
720 msec
2500 msec
1575 msec
2msec
unavailable
2500 msec
1335 msec
25 msec
25 msec
50 msec
unavailable
995 msec
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
IBM SYSTEM/360
420: 192. 123
.122 Control Programs (Contd.)
storage allocation function if the program is
not yet in the core image library, but still in
relocatable text form.
· 13
•
To assign symbolic names to input-output
devices, permitting actual device assignment to be changed at program execution
time.
•
To set up a communication region containing
program name, current date, user's program
switches, and current machine configuration,
for use by both the Supervisor and the user's
problem program.
•
To restart a job from a specified checkpoint
by repositioning tape drives, reassigning I/O
devices, and calling in the Supervisor's restart
program.
Availability
Basic Operating System
for 8K and larger
disc systems: ••..• September 1965.
Basic Operating System
for 16K and larger
tape.or disc
systems: . . . . . . . . December 1965.
Multiprogramming and
Remote Terminal
Processing (for 16K
BOB): •.. ; . . . • . . . second quarter, 1966.
.14
Originator: . . . . . . . . mM Corporation.
· 15
Maintainer:
• 16
.2
First Use: . . . • • . . . . April 1965 for card/tape
I/O subroutines, storage
dump, and relocatable
loader .
PROGRAM LOADING
· 21
Source Programs
.3
HARDWARE ALLOCATION
.31
Storage
.311 Sequencing of program
for movement between
levels: . • . . . • . . . . must be incorporated in
user's program; system
loader of Supervisor
will perform overlays
if so directed.
.312 Occupation of working
storage: . . . • • . . . . storage is allocated in a
fixed fashion by the
Linkage Editor prior to
program load; overlay
areas are also set
aside at that time.
.32
.321 Initial assignment: ... programmer names symbolic
device; Job Control
cards aSSign devices
to the symbolic names
at execution time.
.322 Alternation: . . . . . . • . prepared by Job Control
statements; made operational by direct request
of user's program.
.323 Reassignment: . . . . . . effected by Job Control
cards if job is aborted
prematurely.
.4
RUNNING SUPERVISION
.41
Simultaneous Working: controlled by Channel
Scheduler routines of
Supervisor .
.42
Multiprogramming: .• regulated by Supervisor.
.43
Multi-sequencing: .... no provisions.
.44
Errors, Checks, and Action
. . . . . . . IBM Corporation.
.Error
.211 Programs for on-line
libraries: ••••••.• core image library and
relocatable program
library on tape or disc .
. 212 Independent programs: loaded at execution time
by Job Control cards
from punched cards, tape,
or disc storage.
.213 Data: .••••...•... as required by users' programs.
.214 Master routines: ..•• in core image form on resident
tape or disc file.
.22
.23
Library Subroutines: .. macro-routines can be called
from an on-line tape or
disc library by macroinstructions at assembly
time.
Loading Sequence: ... determined by sequence of
Job Control cards in the
input stream at program
execution time, or by
program calls embedded
in the problem program.
Input-Output Units
Loading input error:
Check or
Interlock
check
Allocation impossible: check
In-out error - single: check
In-out error persistent:
check
Storage overflow:
check
check
Invalid instructions:
Arithmetic overflow: check
Invalid operation:
check
Improper format:
check
Invalid address:
check
Reference to forbidden area of
core memory:
check
.45
coded message on
printer.
Linkage Editor
message.
interrupt routine.
interrupt
interrupt
interrupt
interrupt
interrupt
interrupt
interrupt
routine.
routine.
routine.
routine.
routine.
routine.
routine.
iilterrupt routine.
Restarts: ..•..••.. Supervisor checkpoint
routine writes checkpoint
.
program status records
when directed; a Job Control statement directs
the restart routine to begin
at a specific checkpoint.
(Contd. )
8/65
OPERATING ENVIRONMENT: BASIC OPERA:rING SYSTEM/360
420: 192.500
.5
PROGRAM DIAGNOSTICS
.8
PERFORMANCE
.51
Dynamic
.81
System Requirements
.511 Tracing: .•.•••.•.. provided through use of
.
Autotest routines that
must reside in core
storage with problem
program .
. 512 Snapshots: . . • • . . . . . provided through Autotest.
. 52
Post Mortem: . . . • . . a special dump routine is
processed at any abnormal
end-of-job occurrence,
listing contents of core
storage and general registers.
.811 Minimum configuration: 8,192 bytes of core storage;
standard instruction set;
one I/o channel; one card
reader (1442, 2501, 2520,
or 2540); one 2311 Disk .
Storage Unit or four
2400 Series Tape Units .
.6
OPERA TOR CONTROL
.812 Usable extra
facilities: •.••.•.. all (as incorporated in
program).
.813 Reserved equipment: .. approximately the first
3, 000 bytes of core
storage (reserved for
resident Supervisor) .
• 61
Signals to Operator
.82
· 611 Decision required
by operator: .••••• Supervisor-call interruption.
.612 Action required
by operator: .••..• Supervisor-call.
.613 Reporting progress
of run: . . . • . . . . . . Supervisor-call.
. 62
.63
Operator's Decisions: . through System/360 console
or IBM 1052 PrinterKeyboard.
.7
.821 Loading time: . . . . . • dependent upon speed of
input unit used.
.822 Reloading frequency: . Supervisor need be loaded
only once, but can be
reloaded whenever desired
through the initial program loading (IP L)
procedure .
. 83
Program Space
Available: . • • • • . . . all of core storage except
for the maximum 4K or 6K
bytes of storage reserved
for the Supervisor's control routines.
.84
Program Loading
Time: • • . . . . • . . . . depends upon the speed
--.
of the input device used.
.85
Program Performance: no Supervisor performance
times have been made
available by IBM to date;
see Table I for I/O control
routine times.
Operator's Signals
• 631 Inquiry: .•••.•.... through System/360 console
or IBM 1052 PrinterKeyboard.
.632 Change of normal
progress: . . • . . . • . indicated by coded messages
on console or PrinterKeyboard.
LOGGING: . . . . . . . . as incorporated in user's
program.
System Overhead
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
420: 193.100
IBM System/360
Operating Environment
Basic Programming Support
OPERATING ENVIRONMENT: BASIC PROGRAMMING SUPPORT
.1
GENERAL
.11
Identity:.........
.12
Description
IBM System/360 Basic
Programming Support
(BPS).
The Basic Programming Support (BPS) offered by
IBM for use with small configurations of its ·System/
360 consists of a group of non-sophisticated card and
tape oriented programs designed to provide a restricted set of language and service facilities operating in an environment of minimum supervisory control. Users of System/360 computer systems with
8K or 16K bytes of core storage can choose from
two classes of BPS according to their input-output
configurations. One class of software support applies to systems oriented toward the use of punched
card devices; this class includes only independent
program routines that do not function under the guidance of a supervisory control program. For use
with systems containing a minimum of two 2400
Series Magnetic Tape Units, there is another version of the Basic Programming Support: a series of
tape-oriented control programs designed to provide
semi-continuous supervision of all language and
problem programs during their execution.
. 121 Card Basic Programming Support
The Card BPS for the System/360 contains a set of
independently-operated programs in punched card
form designed to accomplish their functions in an
environment of 8K bytes of core storage, using only
the standard instruction set and a bare minimum of
l.nput-outputdevices. The Card BPS is indeed a basic
system, but the facilities it provides' will help the
user to achieve efficient utilization of the smaller
models of System/360 in simple data processing
situations.
used primarily with systems having limited I/o
configurations and serving only the most basic data
processing problems, the RPG (rather than the Assembler) may become the principal programming
tool in many installations using Card BPS.
The other programs included in the Card Basic Programming Support package are referred to as the
Basic Utilities and include three versions of main
storage/general register dump routines, Absolute
and Relo'cating Loader routines, and a program to
generate the Loader routines according to the specific installation's machine configuration. The Absolute Loader occupies the lower area of core storage and loads assembled programs into Assembleror RPG-assigned core storage addresses. By
means of an REP (Replace) card, specified sections
of the program being loaded can be changed just
prior to execution. The Relocatable Loader also
resides in lower core storage and contains the "replace option" as well. Its unique function is to
provide linkages between separately-assembled
programs so that they can operate as logical units
at program execution time. External references
are resolved in such a way as to permit one program
to refer to instructions or data within another program that is linked to the first by the Relocatable
Loader.
The final group of routines included in the Card
BPS Basic Utilities is called the I/o Support Package. Used by the other Basic Utility routines, the
I/O Support subroutines provide the programmer
with all the coding required to use card and tape input-output devices. Error routines and machinecheck interrupt facilities are included, plus the
capability to display coded error messages to the
operator via the console of the processing unit.
. 122 Tape Basic Programming Support
Included among the IBM-supplied programs are a
punched-card Basic Assembler (described in report
Section 420:173) and Report Program Generator
(discussed in report Section 420:153). The Basic
Assembler is a two-pass operation, and it can utilize up to two magnetic tape units to speed its performance. The assembly language itself is very
restricted, allowing no literals nor macro-language
facilities. Its output can be either in absolute loadable program text or in a relocatable format that
facilitates programming by segments, leaving absolute storage assignments to be performed at
execution time by the Relocatable Loader.
The IBM System/360 Basic Programming Support
available to 8K and larger tape-oriented systems consists of a large group of processing programs (language translators and service programs) and control
programs which are capable of operating in a semiautomatic operating system environment. The basic
supervisory program control functions are performed
on-line in conjunction with and in service of the problem program. Other services normally associated
with an operating system are performed off-line,
due to the limited core storage in an 8K design-level
system. The remainder of this report section is
devoted to a description of the facilities of the Tape
Basic Programming Support system.
The Report Program Generator (RPG) for the Card
Basic Programming Support offers standard facilities to generate listings, perform calculations, and
update files from programmer-supplied problemoriented statements. Since the Card BPS will be
The minimum ,machine configuration required to
operate the Tape BPS includes 8,192 bytes of core
storage, the standard instruction set, two 2400 Series
Magnetic Tape Units (at least one of which must be
9-track), one card reader (2540 or 1442), and one
(Contd.)
8/65
,/
420: 193.122
OPERATING ENVIRONMENT: BASIC PROGRAMMING SUPPORT
. 122 Tape Basic Programming Support (Contd)
I/o channel (either Multiplexor or Selector). To
assemble the Supervisor routines of the Tape BPS
requires the use of another tape unit and a line
printer (1403, 1404, 1443, or 1445). Additional
core storage is desirable in almost all Supervisorcontrolled operations, and is essential for use of
several of the language translators. The other I/o
devices supported by the Tape BPS, in addition to
those mentioned above, are the 1052 Printer-Keyboard, the 1015 Inquiry Display Terminal, and the
2671 Paper Tape Reader. The following machine
features are also supported: the Interval Timer,
Simultaneous Read-While-Write, any channel configuration up to one Multiplexor Channel and two
Selector Channels, and additional core storage.
The language translators offered with the Tape
Basic Programming Support are listed below:
• An 8K Assembler with card and tape IOCS
routines; other IOCS routines are being
prepared by IBM to support the 2321, 1231,
1285, 1412, 1418, 1419, and 1428 input-output
devices. Report Section 420: 173 describes the
8K Tape Assembler with IOCS.
• 16K Card and Tape FORTRAN Compilers.
Descriptions of the BPS FORTRAN languages
can be found in report Section 420:163, where
the restrictions of these language subsets
are listed. The FORTRAN compilers require
the use of the Floating-Point Arithmetic option, and the tape version requires the use of
three tape drives for compilation.
• A 16K Programming Language/I (PL/I) compiler designed for use with punched card systems. The full, Operating System/360 version of this new commercial/scientific language is described in Section 420:169.
• An 8K card RPG and an 8K tape RPG, designed to provide the standard report-writing
features while functioning under control of the
Tape BPS's modified operating system. Section 420:153 lists the salient characteristics
of these two Report Program Generators.
• Several card and tape utility programs, also
described in report Section 420:153, BPS
Problem Oriented Facilities. Two generalized Sort/Merge programs are provided, as
well as Autotest routines to aid in program
debugging and two Multiple utility routines to
provide up to three simultaneous data transcription operations. Disk Storage and Data
Cell device utility programs are also supplied
in the Tape BPS utility package.
At present IBM has no plans to implement a COBOL
Compiler for use with the Basic Programming Support.
The various control functions of the Tape BPS provide a loosely integrated operating environment for
the processing programs of the software system.
Restricted by a maximum program design level of
4K bytes, the control programs are of necessity
©
straightforward and simple, with no one program
exercising supervisory control over all the others.
In most cases, with the Tape BPS, data processing
jobs will be performed in independent job steps, with
the system's operator providing the necessary linkage between steps.
The control functions fall into three major categories:
Control Programs and Operating Environment, Input-Output Control (IOCS), and the Tape Assembly
System. A further group of Tape BPS programs related to these functions provides for the building,
maintenance, and service of the BPS System Tape
and its libraries.
Control Programs and Operating Environment:
principal among these programs are the core storage
resident Supervisor, the Job Control Program, the
Initial Program Loader (IPL), and the Supervisor's
Program Loader.
The Supervisor is defined and assembled to satisfy
the control and I/O requirements of each installation,
In fact, if more I/O devices are to be supported than
the small system's standard card readers, printers,
and magnetic "tape units, then several Supervisors
will have to be assembled for various processing
jobs. Otherwise, one Supervisor with input-output
control for all available devices could consume most
of the low-order 8K bytes of core storage.
The Supervisor consists of several basic elements:
•
A communication region for the storage of
information useful both to the Supervisor and
to the problem program (e. g., addresses of
user-supplied routines, program switches,
today's date, etc.)
•
Routines to handle programmed and machinedetected interrupt conditions, including operator-initiated external interrupts and those
originating from the operation of I/O devices.
•
A set of Channel Scheduler routines to control
the movement of data between main storage
and input-output devices initiated by means of
physical IOCS commands from the Supervisor.
•
Error recovery routines for each I/O device.
•
A group of programs to provide intercommunication with the operator.
•
A program retrieval routine (FETCH) to assist
in overlay control.
•
End-of-job routines.
The size and number of these routines included in
the Supervisor will vary with each installation and
are dependent primarily on the amount of available
core storage. A typical Supervisor operating in a
standard card/tape environment should always fit
within 4,096 bytes of core storage. Once the
Supervisor has been loaded into storage, it is the
user's responsibility to ensure that problem programs do not destroy the resident control routines.
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
420: 193.123
IBM SYSTEM/360
.122 Tape Basic Programming Support (Contd)
to build· a tape of problem programs in loadable
form, and to provide the services of the Linkage
Editor - a quasi-assembler designed to reconcile
external symbolic references and linkages between
separately-assembled program segments, and to
produce a single loadable object program.
The Job Control program performs several distinct
functions in the preparation of a program for execution. Job Control need not reside in core storage
during the problem program's execution. It is required only when certain of its functions are specifically requested by Job Control cards. When requested, the Job Control program must be loaded
by the Program Loader just as any problem program, unless it was already resident in core storage. It can provide the following services: reset
program switches to zeros, reassign symbolic
names to input-output devices, set up restarts for
previously check-pointed jobs, insert data into the
communication region, and edit and store tape label
information for later use by the problem program's
label-checking routines. Once its services have
been performed, the Job Control program can be
overlaid by the problem program.
All of the Tape Assembly System functions are controlled by Job Control statements, which request
speCific functions and specify the parameters necessary to control their execution. The Tape Assembly System is regulated by its own unique Supervisor and Job Control Programs to provide for
the continuous execution of the language translators
and Linkage Editor in an environment of I/o and interrupt control.
The Tape BPS Initial Program Loader (IPL) is provided to begin or renew operation of the system.
The IP L routine must be used to reconstruct the
operating system whenever the resident Supervisor
and Program Loader are overlaid, either delib- erately or accidentally.
The System Program Loader loads the output of the
language translators or Linkage Editor into core
storage locations specified in the loadable text. The
Program Loader operates under control of the
. 13
Supervisor, it remains in core storage unless its
core storage space is required for problem program
use. In the case when the Program Loader is overlaid it must be reloaded by the Initial Program
Loader routine. The Program Loader can load object programs from cards or tape. If any relocation
problems or unresolved program linkages exist in
the object text, the Linkage Editor must first process this text before it can be loaded.
Input-Output Control System (IOCS): The IOCS
system provides logical data record handling routines generated by means of macro-language facilities within the Tape BPS language translators.
Declarative macro-instructions describe to the
translator the characteristics of the input-output
files. Then, during the assembly operation, imperative macros cause the insertion into the object
deck of whatever routines will be required to handle
the program's logical record manipulation. These
routines can calIon the services of the Supervisorcontrolled physical IOCS routines to provide the actual transfer of data from main storage to the input-output devices. Various types of logical records
can be handled by the logical IOCS routines of the
Tape BPS, including fixed-length blocked and unblocked records, as well as variable-length records
in blocked or unblocked format.
Tape Assembly System: the Tape Assembly System
consists of a group of programs that function in an
independent mode, distinct from the problem program execution mode. The Tape Assembly System
must reside on a 9-track 2400 Series Magnetic Tape
Unit, which is referred to as the System Tape. The
principal functions of the Tape Assembly System
programs are to assemble and compile source programs into relocatable or absolute object programs,
The BPS System Tape contains two types of libraries:
the core image library of assemblers, compilers,
loaders, etc., and the macro-routine library which
contains all the macros to be used by the language
translators. The macro-routines are arranged on
the tape in a priority sequence according to frequency of use. Addition and deletion of routines to
both these libraries are performed by a Maintenance
routine, which can also be present on the System
Tape. The catalogues and contents of these librarieS
can be printed or punched as desired.
Availability:
Assemblers, Tap~
FORTRAN, and Basic
utilities: . . . • . . . . . all by third quarter, 1965.
Multiple Device
Utilities and RPG's: fourth quarter, 1965.
Card FORTRAN: ... first quarter, 1966.
Card P L/I: ..•.... third quarter, 1966.
.14
Originator: • . . . . . . . IBM Corporation.
. 15
Maintainer: . . . . . . . . IBM Corporation.
.16
First Use: . . . . . . . . April 1965 for loaders and
card/tape I/O subroutines.
.2
PROGRAM LOADING
. 21
Source of Programs
.211 Programs from on-line
libraries: ..•.•••. System Tape holds core
image and macro libraries .
. 212 Independent programs: from card or tape units.
.213 Data: . • . . . . . . . . • . as required by users'programs .
. 214 Master routines: .... on cards or tape; on System
Tape for control of Tape
Assembly System.
.22
Library Subroutines: . language translators' macrofacilities.
. 23
Loading Sequence: ..• sequential loading of programs
as provided on loadable tape
or system's input device.
(Contd.)
8/65
420: 193.300
OPERATING ENVIRONMENT: BASIC PROGRAMMING SUPPORT
.3
HARDWARE ALLOCATION
.31
storage
. 311 Sequencing of program
for movement
between levels: ..•. must be incorporated in
program; Program Loader
will call in overlays as
requested.
. 312 Occupation of working
storage: . . . • . . . . . storage is allocated as assembly time; program
segments can be combined
with storage allocation performed by the Linkage
Editor.
.32
• 52
Post Morten: ..•••.. not provided in control programs; an independent dump
program can be read into
storage when necessary •
.6
OPERATOR CONTROL
.61
Signals to Operator: ... through the 1052 PrinterKeyboard or the console •
.62
Operator's
Decisions: . . . . • • . . through the 1052 PrinterKeyboard or the Interrupt
Key of the console.
.63
Operator's Signals: ... through the 1052 PrinterKeyboard or the Interrupt
Key of the console •
.7
LOGGING: • . . . • . . . as incorporated within the
users' programs. The
hardware Interval Timer
is available for use by the
programmer•
Input-Output Units
. 321 Initial assignment: ... assigned by programmer
through Supervisor-specified symbolic names.
.322 Alternation: '"
.. as incorporated in the
program.
. 323 Reassignment:. .
. . permitted by means of Job
Control statements.
.4
RUNNING SUPERVISION
.8
PERFORMANCE
.41
Simultaneous Working: as provided by Channel
Scheduler routines of the
Supervisor.
.81
System Requirements
.42
Multiprogramming: ... possible only with Multiple
Device Utility programs
for simple data transcriptions.
• 43
Multi-sequencing: •... not provided .
.44
Errors, Checks, and Action
Check
or
Error
Interlock
Loading input error:
check
Allocation impossible:
check
In-out error - single:
check
In-out error - persistent: check
Storage overflow:
check
Invalid instructions:
check
Program conflicts:
check
Arithmetic overflow:
check
Invalid operation:
check
Improper format:
check
Invalid address:
check
.45
.5
Restarts: . . . . • . .
PROGRAM
DIAGNOSTICS:
coded message.
interrupt routine.
interrupt routine.
coded message.
coded message.
coded message.
coded message.
interrupt routine.
coded message.
coded message.
coded message.
.811 Minimum configura8,192 bytes of core storage;
tion: ..•••..•.•. two 2400 Series Magnetic
Tape Units; one card
reader (2540 or 1442); one
I/o channel (Multiplexor or
Selector).
.812 Usable extra facilities: additional core storage;
1403, 1404, 1443, and 1445
printers;
1052 Printer-Keyboard;
1015 Inquiry Display
Terminal; and 2671 Paper
Tape Reader •
. 813 Reserved equipment: .. none.
.82
.821 Loading time: • . . . . . depends on speed of input
device used.
.822 Reloading frequency: .• program loader must be reloaded with every problem
program if it is overlaid in
core storage by the previous
problem program.
. a CHKPT macro-instruction .83
writes program checkpoint
records as directed; the Job
Control program, via its
RSTRT statement, will
initialize the program to be
restarted at the specified
checkpoint.
. 84
. tracing and dump facilities
are provided through the
Autotest program (Section
420:153) in a 16K environment.
System Overhead
.85
Program Space
Available: . • . . . . . • all of core storage except
the area occupied by the
Supervisor; a basic Supervisor will fit within 4, 096
bytes of core storage.
Program Loading
Time: ..••••..•.. depends on speed of input
device used.
Program Performance: no Supervisor performance
times have been made
available by IBM to date.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
420:201.100
IBM System/360
System Performance
SYSTEM PERFORMANCE
The overall performance of the mM System/360 naturally varies widely, depending
upon the user's choice of Processing Unit model and peripheral equipment. Therefore, the
System/360's performance on the AUERBACH Standard EDP Reports benchmark measures of
system performance has been analyzed separately for each model. For performance.curves,
summary worksheets, and analyses of the results, turn to the System Performance sections
of the subreports on the models of interest:
Model 20: . . . • . . . . . • . . . . . Section 422.201
Model 30: . • . . . . • . . . . . . . . Section 423:201
Model 40: . . . • • . . . . . . . . . . Section 424: 201
Model 50: . . . . . . . . . . . . . . . Section 425:201
Model 65:, • . • . . . . . . . . . . . . Section 426:201
Model 67: . • . • . • . . . • . . . . . Section 427:201
Model 75: . . . . . . . . . . . . . . . Section 428:201
8/65
420:211.101
IBM System/360
Physical Characteristics
PHYSICAL CHARACTERISTICS
Width,
inches
Unit
Depth,
inches
Height,
inches
Weight,
pounds
Power,
KVA
BTU
per hr.
60
29
52
1,200
3.3
9,000
60
29
52
1,400
4.1
11,000
32*
60*
63*
93*
60
60
1,500
1,700
3.8
2.8
10,000
6,000
119*
86*
73
4,500
11.3
19,600
119*
130
193
86*
106
88
73
73
73
5,150
2,400
4,575
14.0
6.9
16.2
25,900
12,000
38,700
30
83
73
2,560
12.5
33,000
26
15
27
195
0.2
670
23
20
9
65
0.1
335
1231 Optical Mark Page
Reader, Mod N1
44
24
45
620
1:2
3,700
1403
1403
1403
1404
48
48
57
67
29
29
29
32
53
53
54
54
750
750
825
1,600
1.0
1.4
1.4
2'.1
3,000
4,600
4,600
5,100
112
41
60
2,475
2.7
6,300
112
41
60
? 650
3.8
8,300
112
41
60
2,700
3.8
8,300
112
41
60
2,750
4.6
10,500
2020 Processor,
Mod B1, C1, D1
2020 Processor,
Mod B2, C2, D2
2030 Processor
2040 Processor
2050 Processor,
Mod F & G
2050 Processor,
ModH
2065 Processor
2075 Processor
2365 Processor Storage, Mod 1
1051 Control Unit,
ModNl
1052 Printer-Keyboard,
Mod 1
Printer,
Printer,
Printer,
Printer,
Mod
Mod
Mod
Mod
2
3
N1
2
1412 Magnetic Character
Reader, Mod 1
1418 Optical Character
Reader, Mod 1 & 3
1418 Optical Character
Reader, Mod 2
1428 Alphameric Optical
Reader, Mod 1 & 3
1428 Alphameric Optical
Reader, Mod 2
1419 Magnetic Character
Reader, Mod 1
1442 Card Read Punch,
Mod N1
1443 Printer, Mod N1
2150 Console
2250 Display Unit,
Mod 1
2250 Display Unit,
Mod 2
2260 Display Station
2280 Film Recorder
2281 Film Scanner
2282 Film Recorderl
Scanner
*
112
41
60
2,800
4.6
10,500
112
41
60
2,675
3.3
8,500
43
56
64
24
43
29
49
46
52
575
800
800
0.7
1.1
0.65
1,500
3,200
1,740
58
62
50
590
2.8
7,200
22
13
105
105
28
21
30
30
50
16
70
70
375
25
1,900
1,900
2.4
?
18.5
18.5
6,600
408
54,500
54,500
105
30
70
1,900
18.5
54,500
Processor dimensions do not include console reading boards.
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
420·211 102
IBM SYSTEM/360
"
""
---"-
Unit
2301 Drum Storage
2302 Disk Storage,
Mod 3
2302 Disk Storag'3,
Mod 4
2311 Disk Storage Drive
2314 Direct Access
Storage Facility
2321 Data Cell Drive,
Mod 1
2401 Magnetic Tape
Unit, Mod I, 2, & 3
2402 Magnetic Tape
Unit, Mod 1, 2, & 3
2403 Magnetic Tape &
Control, Mod 1, 2, & 3
2404 Magnetic Tape &
Control,Mod1, 2, &3
2415 Magnetic Tape
Unit, Mod 1
2415 Magnetic Tape
Unit, Mod 2
2415 Magnetic Tape
Unit, Mod 3
2540 Card Read Punch,
Mod 1
2701 Data Adapter Unit
2702 Transmission
Control
2802 Hypertape Control
2803 Tape Control
2804 Tape Control
2816 Switching Unit,
Mod 1 &,2
2820 Drum Storage
Control
2821 Control Unit,
Mod 1, 2, & 4
2821 Control Unit,
Mod 3,5
2822 Paper Tape Reader
Control Unit
2840 Display Control
2841 Storage Control
Unit
2848 Display Control,
Mod 1
2860 Selector Channel,
Mod 1
2860 Selector Channel,
Mod 2
2860 Selector Channel,
Mod 3
2870 Multiplexor
Channel
7320 Drum Storage
7340 Hypertape Drive,
Mod 3
"Width,--- -Depth;inches
inches
" Height,--inches
Weight;- -Power;KVA
pounds
BTU-per hr.
35
29
64
850
1.5
3,800
86
33
69
4,025
9.0
20,000
86
30
33
24
69
38
4,425
390
12.6
0.75
28,000
2,000
189
32
60
3,500
8.2
19,400
69
51
60
1,950
8.7
19,500
30
29
60
800
1.6
3,500
60
29
60
1,600
3.2
7,000
60
29
60
2,000
2.1
5,500
60
29
60
2,000
2.4
6,300
60
30
70
1,800
3.3
10,000
120
30
70
2,300
4.1
12,000
180
30
70
2,800
4.9
15,000
58
30
45
1,050
1.2
3,000
40
26
40
320
0.3
1,200
29
29
60
60
62
62
29
29
60
60
60
60
900
550
1,400
1,600
2.0
0.6
1.0
1.5
1,800
1,360
2,500
4,000
29
42
60
500"
0.9
1,500
29
62
60
750
1.5
4,000
32
46
60
1,000
2.4
7,000
32
93
60
2,000
4.8
14,000
30
29
26
42
40
60
400
550
2.05
1.4
1,700
4,800
31
44
60
750
1.9
5,500
29
61
72
1,000
1.5
3,542
32
68
71
1,150
3.05
8,200
32
68
71
1,450
3.65
10,000
32
68
71
1,750
4.25
11,600
32
68
71
1,450
4.3
11,000
30
29
60
850
1.1
2,800
29
60
48
1,500
4.0
12,000
General Requirements
Operating ranges Temperature: .••••.••••.••••.•... 60- 90°F. for typical systems; 65- 80°F. for
optical and magnetic character readers.
Relative humidity: .••••••.•••••.•.. 20-80% for typical systems; 20-65% for optical
and magnetic character readers.
Nonoperating ranges Temperature: ...•.••.••.•.•••.••• 50-110°F. for most units.
Relative humidity: .••.••••.••.••.•. 8-80% for most units.
Power: .•••••••..•••.••.••••••••. either 208 or 230 volt, 3-phase, 4-wire, 60-cycle;
voltage tolerance is +10%, -8%.
7/65
420:221.101
IBM System/360
Price Data
PRICE DATA
MODELS 30, 40, 50,65, AND 75*
I
"
"
PRICES
IDENTITY OF UNIT
CLASS
No.
Name
Model
PROCESSING
UNITS AND
MAIN
STORAGE
Monthly
Rental
Monthly
Maintenance
Purchase
$
$
$
Processing Unit and Main Storage:
Main Storage Capacity
C30
D30
E30
F30
8,192
16,384
32,768
65,536
bytes
bytes
bytes
bytes
1,275
1,775
2,675
3,875
90.00
100.00
115.00
135.00
62,000
85,200
127,800
183,900
D40
E40
F40
G40
H40
16,384
32,768
65,536
131,072
_.~~~".144
bytes
bytes
bytes
bytes
bytes
2,700
3,600
4,800
6,400
10,200
105.00
120.00
140.00
170.00
270.00
139,600
182,200
238,300
316,300
500,300
F50
G50
H50
150
65,536 bytes
131,072 bytes
262,144 bytes
524,288 bytes
8,350
9,950
13,750
19,950
260.00
290.00
390.00
570.00
421,700
499,700
683,700
979,500
G65
H65
165
J65
131,072 bytes
262,144 bytes
524,288 bytes
1,048,576 bytes
19,700
22,750
32,050
51,050
785.00
985.00
1,560.00
2,730.00
865,000
990,000
1,402,000
2,242,000
H75
175
. J75
262,144 bytes
524,288 bytes
1,048,576 bytes
31,750
41,050
60,300
1,250.00
1,825.00
3,005.00
1,310,000
1,722,000
2,575,000
25
100
50
20
75
0.75
1. 25
1. 50
0.25
8.00
1,000
4,000
2,000
800
4,125
2
50
150
NC
2.75
1.50
80
2,000
6,000
225
30
40
40
15
50
8.00
1. 75
1. 50
1. 50
0.75
0.75
9,500
1,500
2,000
2,000
750
2,500
50
50
0.75
0.75
2,750
2,500
Processing Unit O)2tions
For Model 30:
Decimal Arithmetic
Direct Control
Floating Point Arithmetic
External Interrupt
1051 Attachment
1051 Home Component Recognition Adapter
Interval Timer
Storage Protection
3237
3274
4427
3895
7915
7916
4760
7520
1401/1440/1460 Compatibility
Features 1401/1440/1460 Basic Compatibility
Column Binary
1402/1403 Attachment
1442/1443 Attachment
Console Inquiry Station
Disk Storage Drives
Magnetic Tapes For Multiplexor Channel
For Selector Channel
4456
1990
4463
4464
4465
4466
4467
4468
*
See page 422:221.101 for Price Data on Model 20.
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
rBM SYSTEM/360
420:221.102
PRICES
IDENTITY OF UNIT
CLASS
PROCESSING
UNITS AND
MAIN
STORAGE
(Cont'd)
No.
Purchase
$
$
$
Programmed Mode Switch
1620 Compatibility 2540 Compatibility Attachment
2501 Compatibility Attachment
2520 Compatibility Attachment .
20
265
NC
NC
NC
0.50
9.50
NC
NC
NC
800
11,125
NC
NC
NC
3237
3274
4427
4457
4478
7520
For Model 40:
Decimal Arithmetic
Direct Control
Floating Point Arithmetic
14.01/1460 Compatibility
1410/7010 Compatibility
Storage Protection
115
150
100
500
650
150
1.75
2.00
2.50
16.00
17.00
1. 50
4,950
6,400
4,300
22,800
29,000
6,000
3274
4478
7117
7130
For Model 50:
Direct Control
1410/7010 Compatibility
7070/7074 Compatibility
Shared Processor storage
225
650
650
300
2.50
37.00
37.00
10.50
8,600
26,650
26,650
12,100
250
700
700
600
3.00
28.00
24.00
19.00
9,300
29,400
30,500
25,000
225
9.00
10,875
6,500
11,000
375.00
575.00
315,000
525,000
75
100
350
150
2.00
3.00
10.00
2.00
3,100
4,050
15,050
6,900
7920
For Model 65:
Direct Control
7070/7074 Compatibility
7080 Compatibility
709/7040/7044/7090/7094/7094 II
Compatibility .
1052 Adapter
7131
Large Capacity Core Storage
(Models 50, 65 and 75 only):
Modell - 1,048,576 bytes
Model 2 - 2,097,152 bytes
2361 Attachment (required):
On Model 50
On Model.65
On Model 75
Shared Storage feature (optional)
6960
6961
1850
For Model 30:
First channel
Second channel
Channel-to- Channel Adapter
215
185
225
16.00
13.00
3.75
8,550
7,900
10,050
6980
6981
1850
For Model 40:
First channel
Second channel
Channel-to-Channel Adapter
350
325
225
17.00
14.00
3.75
15,50C
14,350
10,050
6980
6981
6982
4580
1850
For Model 50:
First channel
Second channel
Third channel
High Speed Channel
Channel-to- Channel Adapter
700
700
700
100
225
26.00
26.00
26.00
4.00
3.75
31,600
31,600
31,600
4,200
10,050
For Models 65 and 75:
Selector Channel:
Model 1 - one channel
Model 2 - two Chan!lels
Model 3 - three channels
Channel-to- channel Adapter
2,100
3,000
3,900
225
55.00
90.00
125.00
3.75
104,000
148,200
192,500
10,050
2361
8080
SELECTOR
CHANNELS
Monthly
Maintenance
5856
7190
8065
8062
8063
3274
7117
7118
7119
BULK CORE
STORAGE
Name
Monthly
Rental
2860
1850
(Contd. )
7/65
420:221.1 03
PRICE DATA
IDENTITY OF UNIT
\.
CLASS
MULTIPLEXOR
CHANNELS
No.
Name
For Model 30:
Multiplexor Subchannels,
Additional; 128 additional
subchannels
High Speed Multiplexor Channel;
4 sub channels
5250
4600
For Model 40:
High Speed Multiplexor Channel;
4 subchannels
4600
For Model 50:
Multiplexor Subchannels,
Additional; 128 additional
sub channels
High Speed Multiplexor Channel;
4 subchannels
High Speed Multiplexor Channel,
Additional; 4 additional
subchannels
5250
4600
4601
For Model 65 and 75:
Multiplexor Channel (requires
8070) Basic channel
First Selector Sub channel
Second Selector Sub channel
Third Selector Sub channel
Fourth Selector Subchannel
2870
6990
6991
6992
6993
2870 Attachment (req'd)
8070
PRICES
Monthly
Rental
Monthly
Maintenance
Purchase
$
$
$
NC
NC
NC
425
20.00
17,450
825
32.00
35,900
NC
NC
1,200
45.00
54,000
300
25.00
13,500
2,200
400
250
250
250
97..00
15.00
10.00
10.00
10.00
110,000
18,500
11,250
11,250
11,250
50
2.00
2,000
660
525
375
875
100.00
64.50
60.00
128.00
35,000
26,250
18,750
42,000
775
600
120.00
94.00
37,200
36,900
260
320
45.50
49.00
15,500
15,750
775
900
650
900
157.00
166.00
130.00
166.00
34,000
41,200
32,700
41,200
NC
Punched Card and Printer
INPUTOUTPUT
Card Read Punch, Model 1
Card Read Punch, Model N1
Card Punch, Model N2
Card Read Punch, Model B1
Card Punch:
Model B2
Model B3
Card Reader:
Model Bl
Model B2
2540
1442
1442
2520
2520
2501
1403
Printer:
Model 2
Model 3
Model 7
Model Nl
1416
Print Train Cartridge
(1 req'd. for 1403 Models 3
and N1)
1404
1443
1445
Printer, Model 2
Printer, Model Nl
Printer (MICR), Model Nl
2821
Control Unit for 2540, 1403, and
1404:
Modell - for one 2540 and one
1403 Model 2 or 7
Model 2 - for one 1403 Model 2
or 7
©
100
NC
3,000
1,550
875
1,425
277.00
64.50
111.00
75,000
44,275
62,000
970
41.00
46,500
600
32.00
28,800
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360
420:221.104
IDENTITY OF UNIT
CLASS
No.
Name
Model 3 - for two 1403 Model
2's or 7's in any combination
Model 4 - for one 2540 and one
1404 Model 2
Model 5 - for one 2540 and two
1403 Model 2's or 7 's in any
combination
INPUTOUTPUT
(Contd.)
7945
3615
1416
PRICES
Monthly
Rental
Monthly
Maintenance
Purchase
$
$
$
1,200
64.00
57,600
1,050
43.50
50,400
1,570
73.00
75,300
Third Printer Control, (for 2821
Model 3 or 5)
1,100 LPM Printer Adapter (for
connection of 1403 Model 3 or N1
to 2821 Modell, 2, 3, or 5)
500
6.50
24,000
75
1. 00
3,000
Print Train Cartridge (1 req'd
for 1403 Models 3 and N1)
100
NC
335
485
785
385
535
835
62.00
70.00
86.00
74.00
82.00
98.00
16,100
23,400
37,900
18,500
25,800
40,300
Magnetic Tape Unit (two drives):
Modell - 30,000 bytes/sec
Model 2 - 60,000 bytes/sec
Model 3 - 90,000 bytes/sec
Model 4 - 60,000 bytes/sec
Model 5 - 120,000 bytes/sec
Model 6 - 180,000 bytes/sec
620
920
1,520
720
1,020
1,620
120.00
136.00
168.00
144.00.
160.00
192.00
29,800
44,200
73,300
34,600
49,000
78,100
Magnetic Tape Unit and Control
(one drive & single-channel
control):
Modell - 30,000 bytes/sec
Model 2 - 60,000 bytes/sec
Model 3 - 90,000 bytes/sec
Model 4 - 60,000 bytes/sec
Model 5 - 120,000 bytes/sec
Model 6 - 180,000 bytes/sec
885
1,035
1,335
1,085
·1,235
1,535
78.00
86.00
102.00
95.00
103.00
111. 00
43,400
50,900
65,700
53,300
60,800
75,600
3,000
Magnetic Tape
2401
2402
2403
2404
2415
7/65
Magnetic Tape Unit (one drive):
Model 1 - 30, 000 bytes/sec
Model 2 - 60,000 bytes/sec
Model 3 - 90,000 bytes/sec
Model 4 - 60,000 bytes/sec
Model 5 - 120,000 bytes/sec
Model 6 - 180,000 bytes/sec
Magnetic Tape Unit & Control
(one drive & dual-channel
control):
Modell - 30,000 bytes/sec
Model 2 - 60,000 bytes/sec
Model 3 - 90,000 bytes/sec
,Magnetic Tape Unit and Control
(single-channel) :
Modell - 2 drives;
15,000 bytes/sec
Model 2 - 4 drives;
15,000 bytes/sec
Model 3 - 6 drives;
15,000 bytes/sec
Model 4 - 2 drives;
15,000/30,000 bytes/sec
Model 5 - 4 drives;
15,000/30,000 bytes/sec
Model 6 - 6 drives;
15,000/30,000 bytes/sec
A
.11111,,'111,,1111.
AUERBACH
/
1,165
1,315
1,615
93.00
101. 00
117.00
57,100
64,600
79,400
750
100.00
36,750
1,200
180.00
58,800
1,650
260.00
80,850
910
115.00
44,500
1,460
205.00
71,500
2,010
295.00
98,500
(Contd. )
420:221.105
PRICE DATA
PRICES
IDENTITY OF UNIT
CLASS
Name
No.
Monthly
Rental
Monthly
Maintenance
$
$
INPUTOUTPUT
(Contd.)
650
800
20.00
25.00
32,600
40,100
Tape Control (dual channel):
Modell (800 bpi)
Model 2 (800 and/or 1600 bpi)
930
1,080
35.00
40.00
46,700
54,200
Switching Unit, Model 1
550
4.00
26,500
3236
Data Conversion feature:
On 2403, 2415, or 2803
On 2404 or 2804
45
70
1. 00
1. 50
2,160
3,425
3471
3472
3471
Dual Density feature (permits
operation at 800 as well as
1600 bpi):
On 2401 Mod 4, 5, 6
On 2402 Mod 4, 5, 6
On 2403 Mod 4, 5, 6
25
50
25
1. 75
3.50
1.75
1,200
2,400
1,200
5121
5122
Mode Compatibility:
On 2401 Mod 1, 2, 3
On 2402 Mod 1, 2, 3
10
20
NC
NC
450
900
5320
5320
5321
5320
9-Track Compatibility:
On 2403 Mod 4, 5, 6
On 2803
On 2804
On 2415
230
230
280
135
28.00
28.00
22.00
10.00
11,050
11,050
13,440
6,480
7125
7127
7126
7125
7127
7126
7128
7125
7127
7-Track Compatibility:
On 2403 Mod 1, 2, 3
On 2403 Mod 4, 5, 6
On 2404 Mod 1, 2, 3
On 2803 Mod 1
On 2803 Mod 2
On 2804 Mod 1
On 2804 Mod 2
On 2415 Mod 1, 2, 3
On 2415 Mod 4, 5, 6
50
175
75
50
175
75
200
50
95
1.25
19.00
2.00
1. 25
19.00
2.00
21. 00
1. 25
3.50
2,400
8,400
2,650
2,400
8,400
3,650
9,600
2,400
4,560
7135
7135
7136
7135
7 and 9- Track
On 2403 Mod
On 2803 Mod
On 2804 Mod
On 2415
375
375
450
155
47.00
47.00
43.00
13.00
18,000
18,000
21,600
7,440
7160
7161
Simultaneous Read-While-Write:
On 2401
On 2402
10
20
NC
NC
450
900
25
0.75
1,200
2816
~ '3228
I
~
~-~-
$
Tape Control (single channel):
Model 1 (800 bpi)
Model 2 (800 and/or 1600 bpi)
2803
2804
?
Purchase
7185
Compatibility:
4, 5, 6
2
2
16-Drive Addressing (on 2403 or
2803, any model
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
420:221.106
IBM SYSTEM/360
- - - " " " "
"""""---
IDENTITY OF UNIT
CLASS
INPUT
OUTPUT
(Contd.)
No.
Name
PRICES
Monthly
Rental
Monthly
Maintenance
Purchase
$
$
$
1,350
2,100
550
25
130.00
38.00
4.00
0.75
525
250
175
200
200
56.00
2.75
1. 50
1. 00
1. 00
27,250
10,000
7,000
8,000
8,000
35
100
1. 00
4.00
1,400
4,250
5,600
7,900
575
15
2,800
2,300
218.00
340.00
51.00
475.00
51. 50
252,000
355,500
26,300
490
136,500
124,000
2,000
2,300
100
5,250
225.00
75.00
2.00
615.00
96,000
112,300
4,860
252,000
Hypertape
7340
2802
2816
7185
Hypertape Drive, Model 3
Hypertape Control (single-channel)
Switching Unit, Model 2
16-Drive Addressing Feature
67,500
103,000
26,500 "
1,200
Disk and Drum Storage
2841
7950
8079
7144
1024
4385
8100
2302
2311
1316
2321
7320
2301
2820
8170
2314
2316
8170
Storage Control
2302 Attachment (for 2841)
2321 Attachment (for 2841)
7320 Attachment (for 2841)
Additional Storage (for up to 8
additional 2302 access
mechanisms)
File Scan feature (for 2841)
Two-Channel Switch (for 2841)
Disk Storage:
Model 3 - 2 access mechanisms
Model 4 - 4 access mechanisms
Disk Storage Drive, Modell
Disk Pack for 2311
Data Cell Drive, Modell
Drum Storage, Modell
Drum Storage
Drum Storage Control (for 2301)
Two-Channel Switch (for 2820)
Direct Access Storage Facility
and Control, Modell
Disk Pack for 2314
Two-Channel Switch for 2314
-
20
140
-
2.50
650
5,950
140
210
21. 00
8.25
6,700
10,000
505
2,000
41. 50
171. 00
26,700
82,260
2,600
2,900
3,075
180.00
214.00
202.00
120,300
133,800
142,100
2,275
224.00
110,500
3,000
3,300
3,475
214.00
247.00
235.00
138,600
152,100
160,600
100
6.50
5,300
30
20
10
8.25
1. 50
1. 00
1,000
600
300
Paper Tape
2671
2822
Paper Tape Reader
Paper Tape Reader Control
Optical and Magnetic Readers
1231
Ij12
1418
1419
1428
7720
Optical Mark Page Reader
Magnetic Character Reader
Optical Character Reader:
,
Model 1 (3 stackers)
Model 2 (1~ stackers)
Model 3 (3 ·stackers; extended
document. Elize' range) , ,
Magnetic Character Header
Alphameric 'Optical ,Reader:
Modell (3 stackers)
Model 2 (13 stackers)
Model 3 (3 stackers; extended
document size range)
System/360 Adapter (required for
1412, 1418, 1419, and 1428)
Displays
2260
4766
4767
Display Station
Alphameric Keyboard
Numeric Keyboard
(Contd. )
7/65
PRICE DATA
420:221.107
IDENTITY OF UNIT
CLASS
INPUTOUTPUT
(Contd.)
Name
No.
2848
Display Control Modell; 240 char per unit,
up to two Display Adapters
Model 2; 480 char per unit,
maximum of one Display
Adapter
Model 3; 960 char per unit,
max. of one Display Adapter
Display Adapter, one required
per two 2260's:
On Model 1 (12 max)
On Model 2 (8 max)
On Model 3 (4 max)
Expansion Unit; permits connection
of additional Display Adapters to
2848 Model 1(6) , Model 2 (4), or
Model 3 (3)
Expansion Unit; permits connection
of additional Display Adapters to
2848 Modell (4) or Model 2 (3),
and one 1053 Adapter
1053 AdapterFor 2848 Modell or 2 (3858 req'd)
For 2848 Model 3
Line Addressing
Non:"destructible Cursor
Non-destructible Adapter (5340
req'd)
Display Unit:
Model 1 (includes control unit)
Model 2
2250 Options:
Absolute Vectors feature
Alphameric Keyboard
Buffer (4,096 bytes)
Buffer (8,192 bytes)
Character Generator
Light Pen
Operator Control Panel (first)
Operator Control Panel (second)
Programmed Function Keyboard
3355
3356
3357
3857
3858
7927
7928
4787
5340
5341
2250
1001
1245
1498
1499
1880
4785
5475
5476
5855
Film Recorder
Film Scanner
Film Recorder/Scanner
Display Control (for 2250 Model 2
and Film Units)
Display Multiplexor (to connect
additional 2250 Model 2's or Film
Units to a 2840)
Film Unit Attachment (req'd to
attach Film Units to a 2840)
2280
2281
2282
2840
3351
4395
PRICES
Monthly
Rental
Monthly
Maintenance
Purchase
$
$
$
360
23.00
16,700
390
420
23.50
24.00
17,900
19,100
40
80
100
45
2.00
4.00
5.00
NC
1,600
3,200
4,000
1,950
55
NC
2,400
40
40
10
10
5
3.25
3.25
1. 25
1. 00
0.50
1,600
1,600
450
430
215
700
350
125.00
110.00
33,600
16,800
225
50
250
400
300
75
45
35
100
6.00
2'.50
5.00
8.00
14.00
8.00
NC
NC
5.00
9,000
2,400
12,000
19,200
14,500
3,600
2,000
1,550
4,800
5,600
8,200
11,500
1,100
1,275.00
1,100.00
1,850.00
50.00
230,000
377,000
540,000
52,800
50
2.50
2,400
275
5.00
12,500
1,200
625
38.50
27.00
57,600
30,000
Audio Response
Audio Response Unit, Model 3
Audio Response Unit, Model 3
(Note: Optional features are required to connect more than 4lineE
to a 7770 or more than 2 lines to
a 7772).
7770
7772
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
!'8M SYSTEM/360·
420: 221.1 O~
IDENTITY OF UNIT
CLASS
INPUTOUTPUT
(Contd.)
No.
PRICES
Name
Monthly
Rental
Monthly
Maintenance
Purchase
$
$
$
Communication
2701
2702
7955
Data Adapter Unit (including 1
adapter for 1 line)
Transmission Control (for up to 15
lines)
31-line Expansion feature (adds Ill.
more lines to 2702 capacity)
... " ~
(Note: Numerous options permit.!
connection of a wide variety of
communication terminals to the
2701 and 2702.)
Transmission Control (no line
adapters included iii price;
numerous options permit connection to various communications
facilities)
Remote Multiplexor:
Modell
Model 2
Digital Time Unit:
Modell
Model 2
Communications Terminal
Communications Terminal
200
I
15.00
9,700
70.00
40,800
lOq
5.00
4,700
1,450 :
95.00
69,600
460
495
28.75
30.25
20,700
22,275
100
115
80
80
3.00
3.00
23.50
23.50
4,500
5,750
3,200
3,200
/.
850
~I
I
/
2703
2712
1032
2740
2741
Console
INPUTOUTPUT
(Contd.)
I/o
1052
7920
Printer-Keyboard
1052 Adapter (for Models 40-65)
65
225
6.50
9.00
2,725
10,875
1051
Control Unit:
Modell - on-line or "home loop"
Model N1 - "home loop" only
1051 Adapter (for Model 30 only)
Console (for Models 65 and 75)
Operator Control Panel (first)
Operator Control Panel (second)
75
60
75
515
35
35
2.25
2.00
8.00
15.00
3,515
3,050
4,125
25,200
1,550
;L,550
7915
2150
5475
5476
-
NOTES:
7/65
1.
Indicated monthly maintenance charges apply for first 36 months.
2.
Not all input-output devices can be used with all System/360 models; see Section 420:031,
System Configuration, for configuration rules.
422:011.100
IBM System/360
Model 20
Introduction
INTRODUCTION
The Model 20 is the smallest currently-announced member of the IDM System/360
computer family. It was originally announced in November, 1964, as a card-oriented computer
system designed primarily as the first step upward from punched-card tabulating equipment.
The recent addition of magnetic tape facilities has expanded the range of practical applications
for the Model 20. A Model 20 system with card reader and printer can be rented for as little
as $1,280 per month, although typical system rentals will be in the $1,700 to $3,500 range.
Initial deliveries are scheduled for the first quarter of 1966, and the current delivery schedule
is 22 months.
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, gatherwrite, 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.
The Model 20 Processor can contain 4,096, 8, 192, or 16,384 bytes of core storage.
Cycle time is 3.6 microseconds per access of one half-byte (four data bits). Read-only storage
with a cycle time of 0.6 microseconds is used for control of processor and input-output functions.
There are 8 general registers (compared with 16 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 in the Model 20. It is performed
upon 4-bit BCD digits packed two to a byte, with a sign in the rightmost four bits of the loworder byte of each data field. Decimal operands may be up to 16 bytes (31 digits and sign) in
length; the length of each field is specified in the instructions that reference it, rather than by
a word-mark in the field itself.
Because of the instructions which are unique to the Model 20, programs written for a
Model 20 system cannot be directly executed on a larger System/360 model. A special routine
will be provided to trap and interpret these Model 20 instructions and permit all Model 20 programs to be run on other System/360 models with equivalent ,peripheral equipment (provided
that the programs are not dependent upon fixed relationships between input-output times and
instruction execution times). The standard System/360 instructions will be executed normally,
with no interference by the interpretive routine. Throughput can range from below Model 20
speeds (in certain I/O-limited cases) to near-normal speeds for the larger system (in processorlimited cases). The interpretive routine will not be required to generate and run programs
written in Model 20 Report Program Generator language on larger System/360 models.'
A control panel, built into the top of the 2020 Processing Unit 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.
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. Model 20 configurations are subject to the following restrictions:
•
•
•
Only one peripheral device of any given type can be connected.
A 2560 Multi-Function Card Machine and a 2520 Card Punch
cannot be used in the same system.
The maximum configuration possible is three card read stations, two
punch stations, one document print station, one magnetic character reader,
one line printer, and one magnetic tape unit (which can include up to six tape
drives).
The following peripheral devices are available for use in Model 20 systems:
1442
2520
2520
2501
2560
Card Punch, Model 5
Card Punch, Models A2 and A3
Card Read Punch, Model Al
Card Reader, Models Al and A2
Multi-Function Card Machine (MFCM)
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
422:011.101
IBM SYSTEM/360-MODEL 20
1403
2203
2415
1419
2073
Printer, Models 2, 7, and N1
Printer, Model A1
Magnetic Tape Unit, Models 1, 2, and 3
Magnetic Character Reader, Modell
Communications Adapter
These peripheral devices are described in individual sections of the general System/360
report. Note that no random access storage devices are currently available for use with the
Model 20.
Reading, punching, printing, and processing can all occur simultaneously in a System/360
Model 20 through time-sharing of the core storage accesses required by each input-output device
and by the central processor. Magnetic tape reading or writing, however, can be overlapped only
with printing on the 1403 Printer. The 1419 Magnetic Character Reader cannot operate simultaneously with any card reading or card punching device.
The 2560 MFCM is a unique punched card input-output unit developed especially for the
System/360 Model 20. Equipped with two 1,200-card feed hoppers, a reading station, a punching
station, an optional printing station, and five 1, 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; 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 cards per minute. (This
speed is achieved only when no punching or printing is being done on the cards.) 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. The
effective speed depends (as in the IBM 1442) upon the position of the last column punched in each
card; when all 80 columns are punched, the punching rate is 91 cards per minute.
The optional Card Print feature provides a printing unit that can print two, four, or
six lines of information on any or all cards passing through the MFCM. Each line can hold up
to 64 printed characters, spaced 10 to the inch. Rated printing speed is 140 character positions
per second.
The 2073 Communications Adapter provides the Model 20 with limited data communications facilities. With this adapter, a Model 20 can function as a Single-line, point:-to-point processor terminal communicating with: another Model 20, a larger System/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 facilities at speeds ranging 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.
Because of their restricted instruction repertoire, different I/O control methods, and
limited core storage, Model 20 systems will not be able to use the extensive array of software
that IBM is providing for larger System/360 models. The software that is being prepared specifically for Model 20 systems includes a Report Program Generator, utility programs for performing the functions of conventional punched-card machines, a one-for-one Basic Assembler,
I/O control' routines to handle all peripheral operations (including the Communications Adapter),
and a 1419 Magnetic Character Reader I/O Program. Section 422:151 contains more information
about the Model 20 software.
Information in the general System/360 report (420:) which pertains to the Model 20
includes:
Data structure, Section 420:021
System Configuration, Section 420:031
Core storage, Section 420:041
Input-Output Devices, appropriate parts of Sections 420:071 through 420:108
Data Codes, Section 420:141.
Physical Characteristics, Section 420:211.
This subreport concentrates upon the characteristics and performance of the Model 20
in particular.
7/65
/
422:031.100
IBM System/360
Model 20
System Configuration
SYSTEM CONFIGURATION
System configuration possibilities for Model 20 and other System/360 models are summarized in report Section 420:031. This report section shows Model 20 systems arranged in accordance
with the specifications for our Standard Configurations, as defined in the Users' Guide, page 4:030.120 •
•1
TYPICAL CARD SYSTEM; CONFIGURATION I
Deviations from standard Configuration: . . . • . . . . card punch is 50% faster.
printer is 40% slower.
Equipment
Rental
Main Storage (8, 192 bytes)
2020 Processing Unit, Model Cl
2501 Card Reader, Model A2,
and Attachment:
Reads 1,000 cards/minute
275
2520 Card Punch, Model A3,
and Attachment:
Punches 300 cards/minute
475
1403 Printer, Model 7,
and Attachment:
Prints 600 lines/minute
850
TOTAL RENTAL:
$2,300
I
\
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
422:031.200
.2
IBM SYSTEM/360-MODEL 20
4-TAPE BUSINESS SYSTEM; CONFIGURATION II
Deviations from Standard Configuration: •••..••• card reader is 20% faster.
printer is 20% faster.
Equipment
Rental
Main Storage (8,192 bytes)
2020 Processing Unit, Model C2
2501 Card Reader, Model AI,
and Attachment:
Reads 600 cards/minute
215
1442 Card Punch, Model 5, and
Attachment:
Punches 91 full cards/minute
285
1403 Printer, Model 7,
and Attachment:
Prints 600 lines/minute
850
2415 Magnetic Tape Unit and
Control, Model 2 (4 drives total):
15,000 bytes/second
1,200
4658 Input/Output Channel
TOTAL RENTAL:
.3
$3,475
TYPICAL MULTIFUNCTION SYSTEM
Equipment
Main Storage (8, 192 bytes) }
2020 Processing Unit,
Model C1
700
2560 Multifunction Card
Machine and Attachment:
Reads 500 cards/minute
(includes 2-line Card Print
feature)
800
2203 Printer and Attachment:
Prints 300 to 750 lineS/minute
565
TOTAL:
Note: Addition of a 2501 Card Reader, Model Al (600 cpm), and Attachment would
increase the monthly rental to $2, 280 •
7/65
$
$2,065
422:051.100
IBM System/360
Model 20
Central Processor
CENTRAL PROCESSOR
.1
GENERAL
• 11
Identity: . . • • . . . . . . IBM 2020 Processing Unit .
.12
Description
The 2020 Processing Unit, under control of the
stored program, performs all calculations and
controls all input-output devices in a System/360
Model 20 system. It 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 and processor status instructions are
unique to the Model 20 (see Program Compatibility
below). Decimal arithmetic (including multiplydivide), editing, and code translation instructions
are standard. Floating-point arithmetic is not
available.
The processor can contain 4, 096, 8,192, or 16,384
bytes of core storage. Access time is 3.6 microseconds to each half-byte (4 data bits). Read-only
storage with a cycle time of 0.6 microseconds is
used for control of processor and input-output
functions. There are 8 general registers (versus
16 in the larger System/360 models), usable as
fixed-point accumulators or as index registers.
The use of decimal arithmetic is being emphasized
in the Model 20. It 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 field is specified in
the instructions that reference it, rather than by a
word mark in the field itself. Execution times for
decimal arithmetic operations are as follows, where
the operand length is 5 decimal digits (3 bytes):
Time, Microseconds
c = a + b: .•••••.•• 658.
c = ab: .•••.•••••. 6,286.
c=a/b: " •.•..••.. 7,175.
Program Compatibility
There are five major differences between the Model
20 and the Processing Units of the higher-numbered
System/360 models:
•
/
Limited Instruction Repertoire. The Model 20
has 36 instructions, compared with 86 to 138
instructions in the larger models. Among the
instructions not implemented in Model 20 are
fixed-point binary multiply, divide, logical
compare, and shifts; floating-point operations;
radix conversion; multiple register operations;
and stepping instructions. Fixed-point binary
operations are limited to 16-bit operands, and
only eight general registers are available.
•
Addressing. The basic form of addressing
in Model 20 is by means of a 14-bit direct
address contained in the instruction itself.
Addresses can be indexed by using a 12-bit
address in the instruction and referencing
a 16-bit general register to be used as an
increment or decrement. Double indexing
is not permitted in Model 20. The first
144 bytes of main storage are protected,
and program reference to this area results
in an error condition.
•
Interrupts. Interrupts in Model 20 can be
caused only by the completion of data transfers to or from input-output devices. There
is no individual interrupt control for each
channel, although it is possible to inhibit
all interrupts.
•
Multiprogramming. No provisions for multiprogrammed operation are made in Model 20.
•
Input-Output Instructions. Model 20 has its
own unique set of I/O instructions. Also,
the "Halt and Proceed", "Set PSW", and
"Branch and Store" instructions are used
only in Model 20.
Because the instructions for input-output operations and processor status are unique to the
Model 20, programs written for a Model 20
system cannot be directly executed by a larger
System/360 model. A special routine will be
provided to trap and interpret the Model 20
input-output inst"ructions and permit all Model
20 programs to be run on other System/360 models
with equivalent peripheral equipment (provided
that the programs are not dependent upon fixed
relationships between input-output times and
instruction execution times). The standard
System/360 instructions will be executed normally,
with no interference by the interpretive routine.
The instructions which are unique to Model 20
will be trapped as invalid instructions, and the
interpretive routine will then accomplish the equivalent functions on the larger system. Throughput,
according' to IBM, can range from 1;lelow Model
20 speeds (in certain I/O-limited cases) to near
normal speeds for the larger system (in processorlimited cases). The interpretive routine will not
be required to generate and run programs written
in Model 20 Report Program Generator language
on larger System/360 models.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
422:051.200
IBM SYSTEM/360-MODEL 20
-------------------------------------------------------------------------------------- -----
.2
PROCESSING FACILITIES
.21
Operations and Operands
Operation
and Variation
.211 Fixed point Add-subtract:
Multiply Short:
Long:
Divide No remainder:
Remainder:
Provision
Radix
Size
automatic
automatic
binary
decimal
halfword (16 bits) .
variable: 1 to 31 digits.
none.
automatic
decimal
variable: 1 to 15 digits.
none.
automatic
decimal
variable: 1 to 30 digits
in dividend, 1 to 15
in divisor.
• 212 Floating point: .••••. not available .
.213 Boolean AND:
automatic
binary
Inclusive OR:
automatic
binary
Exclusive OR:
none.
.214 Comparison Numbers:
automatic
fixed point binary:
fixed point decimal:
Absolute:
automatic
Letters:
automatic
Mixed:
automatic
Collating sequence ASCII code: .••••• specials, numbers, letters.
Extended BCD code: specials, letters, numbers.
(see Data Code Tables, Section
420:141.)
Provision
From
To
.215 Code translation:
automatic**
any 8-bit
code
any 8-bit
code
.216 Radix conversion:
none.
1 byte.
1 byte.
16 bits.
up to 31 digits.
8 bits or up to 256 bytes.
8 bits or up to 256 bytes.
8 bits or up to 256 bytes.
1 to 256 bytes.
** Special code tables must be provided to use the translate instructions.
Provision
. 217 Edit format Alter size:
Suppress zero:
Round off:
Insert point
Insert spaces
Insert fill
character:
Protection:
Float dollar sign:
• 218 Table look-up:
• 219 othersDecimal shift:
.22
generally make larger
automatic
none
automatic
automatic
can edit multiple
fields with one
instruction
2 to 256 bytes.
must use Move
with Offset
instruction
up to 31 digits.
/
automatic
automatic
none •
none .
semi-automatic
Special Cases of Operands
. 221 Negative numbers Binary: .....••••• 2's complement and sign
hit.
Decimal: . . . . . . . . sign in least significant
byte.
.222 ZeroBinary: . . . . • . . . . only positive zero.
7/65
yomment
Decimal: . • . • . . . . . positive or negative zero;
treated as equal in comparisons .
. 223 Operand size determination Binary: ...•••...• fixed size; halfword (16
hits) .
Decimal (and certain
logical operations):. variable size, indicated by
operand length fields in
instruction.
(Contd. )
CENTRAL PROCESSOR
. 23
422:051.230
Instruction Formats
.231 Instruction structure: . 1, 2, or 3 halfwords (16,
32, or 48 bits), depending
on number of main storage addresses necessary.
.232 Instruction layout and
parts: . • . . . . . . . . • see diagrams below.
Instructions can be two, four, or six bytes in length.
A 2-byte instruction causes no reference to main
storage. A 4-byte instruction causes one reference
to main storage, while a 6-byte instruction causes
two storage references. There are four basic instruction formats in Model 20:
Type RR - Register to Register (2 bytes)
Op
1Rli R21
Type RX - Register to Storage (4 bytes)
op
1Rl 100001 B2 1
D2
Type SI - Storage and Immediate Operand (4 bytes)
Op
12
1 Bl 1
Dl
.2373 Indexing rules: •.••. base address is treated as
a 15-bit binary integer
plus sign bit; displacement
is treated as a 12-bit
positive binary integer.
These are added to form a
16-bit binary integer,
ignoring overflows .
. 2374 Index specification: .. base address (B) field
specifies the number of
a register .
. 2375 Number of potential
indexers: • . . . . . . . 8.
.2376 Addresses which can
be indexed: .••.•.. all core storage addresses
can be indexed by base
register contents .
• 2377 Cumulative indexing:. none.
· 2378 Combined index and
step: .•••.•.•••• none.
.238 Indirect addressing: .• none •
. 239 Stepping: . . • • . . . . • . none.
· 24
Type SS - Storage to Storage (6 bytes)
Category of
storage
4-bit base register specification
12-bit displacement
8-bit literal operand
8-bit operand length specification
Op = 8-bit operation code
R = 4-bit operand register specification.
General registers:
Program Status
Word:
B=
D=
I =
L=
.234 Basic address structure: .••••.•.... 2 + 0; variations in instruction length are due to the
fact that either operand
address may be either a
main storage address or
a register address.
.235 Literals Arithmetic (logical): 1 byte.
Comparisons and
tests (logical): •... 1 byte.
Incrementing modifiers: ..•...•.•• none.
.236 Directly addressed operands -
\.
Special Processor Storage
Number
oflocations
Size in
bits
Program
usage
8
16
indexing,
base addresses,
and accumula'tors.
1
32
holds location counter
and various
flags.
.3
SEQUENCE CONTROL FACILITIES
· 31
Instruction Sequencing
.311 Number of sequence
control facilities: ... 2 Program Status Words
(PSW), only one of
which is active at a
time .
. 312 Arrangement: . . . • . . two PSW's - one new and
one old.
Minimum Maximum Volume
Internal storage
size
size
accessible .313 Precedence rule: .••. test for channel end con~
dition is performed in an
established priority
Core storage:
1 byte
256 bytes 16,383
sequence .
bytes*
. 314 Special sub-sequence
General
counters: " . . . . . . the length of variahleregisters:
1 reg8 reg8 regsize operands is held
ister
isters
isters
in decimal arithmetic
ins tructions.
* If base registers are used for relative ad.315 Sequence control step
dressing, a maximum of 4,096 bytes are
size: . . . • . • . . . . . halfword.
accessible via each register so allocated.
If base registers are not used, 14 bits in
.32 Look-Ahead: . . . . . . none.
the instruction word can be used for direct
addressing.
. 33 Interruption
.237 Address indexing.2371 Number of methods: . L
. 2372 Names: ••.•••...• indexing, using the base
register addresses.
©
.331 Possible causes Input-output units: .. data transfer completed .
Processor errors: .. none.
Other: . . . • . . . . . . none.
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
422:051.332
IBM SYSTEM/360-MODEL 20
.332 Control by routine fudividual control: .. acceptance or nonacceptance of I/o
interrupts can be
controlled.
Method: ..•••..•• specific bit in Program
Status Word.
.333 Operator control: . . . . none.
.334 futerruptionconditions: . . . . . . . . . . (1) channel end condition
occurs and is stored.
(2) CPU is in interrupted
state (channel mask
bit = 1).
(3) processing operation
(instruction execution)
has ended.
.335 futerruption process futerruption action: .• present PSW (Program
Status Word) is stored
and replaced by a
standby PSW.
Registers saved: ... none of the 8 general
registers are saved
automatically. Most
of the necessary optional data is saved in
the old PSW.
Destination: . . . . . . . contained in standby PSW.
.336 Control methods Determine cause: .•. device address is given
in PSW.
Enable interruption: . by setting of bit in the P SW.
.34
Multiprogramming: ... none.
.35
Multisequencing: ...• none.
.4
PROCESSOR SPEEDS
.41
fustruction Times in Microseconds
Note:
B = operand length in eight-bit bytes.
(Operand length in decimal digits =
2B-1. )
.411 Fixed point Add-subtract: ..•.• 269 + 68B
Multiply: .•••.•.•. 20 + 164B + 642B2
Divide: ..•.•••••. 122 + 420B + 737B2
.412 Floating point: .••••. not available.
.413 Additional allowance for Single indexing: ..•• ?
Double indexing: .••• not available in Model 20.
Indirect addressing: . not available in Model 20.
Recomplementing: .• 49B
.414 Control:
Compare Fixed-point binary:. 215
Decimal: .••••..• 269 + 68B
Logical: ..•••••. 163 + 24B
Branch: ..••••..• 110
• 415 Counter control Step: ••••••••••. 206
Test: ••••••••••• 215
.416 Edit: .•••••.••.•• 165 + 26B
.417 Convert: .•••••••.• not available.
.418 Shift: ••.•••.••.•. not available.
7/65
.42
Processor Performance in Microseconds
Fixed point
(decimal)
.421 For random addresses c = a + b: '" ••••• 406 + 84B
b = a + b: .••••••• 269 + 68B
Sum N items: ..•••• (269 + 68B)N
c = ab: .•••••.••• 230 + 197B + 642B2
c = alb: •.•••.••• 332 + 454B + 737B2
.422 For arrays of data ci = ai + bj: .••••.• 1091 + 84B
bj = ai + OJ: ••••••• 954 + 68N
Sum N items: .••••. (774 + 68B)N
2
c = c + aibj: .•••••• 1184 + 265B + 642B
.423 Branch based on comparisonNumeric data: ..••. (1473 + 68B)N
Alphabetic data: .••• (1367 + 24B)N
.424 Switching Unchecked: .•••••• 884
Checked: ..•••.••• 1368
List search: .••••. 130 + (804 + 24B)N
.425 Format control, per character Unpack: .•••••••. 31. 4
Compose: ...••.•• 86.6
.426 Table lookup, per comparison For a match: .•.••. 804 + 24B
For least or greatest: 817.7 + 25. 6B
For interpolation
point: .••••••••• 804 + 24B
.427 Bit indicators Set bit in separate
location: .••••••. 111
Set bit in p'lttern: .•• 136
Test bit in separate
location: ..••••.. 242
Test bit in pattern: .• 250
.428 Moving: '" . . • . . . . 137 + 16B
.5
ERRORS, CHECKS, AND ACTION
Error
Check or
Interlock
Action
Overflow:
check
Zero divisor:
check
Illegal data:
check
Unavailable
operation:
check
programming error
stop.
Illegal storage
address:
check
programming error
stop.
process error stop.
Receipt of data:
parity
check
Dispatch of data: send
parity
bit.
Specification
error:
check
programming error
stop or condition
code set.
programming error
stop.
programming error
stop.
programming error
stop.
422: 111.100
IBM System/360
Model 20
Simultaneous Operations
SIMULTANEOUS OPERATIONS
The capability of the System/360 Model 20 to overlap an input-output operation with
processing or with another input-output operation is under the control of the operator through
a "time-sharing" switch on the console. If the switch is set "on", simultaneous operations
proceed as described below. If the switch is set in the "off" position, the Model 20 becomes
a sequential machine, capable of only one operation at a time.
Reading, punching, printing, and processing can all occur simultaneously at the
option of the operator in a System/360 Model 20 through "time-sharing" of the core storage
accesses required by each input-output device and by the central processor. The processor
delays during most input-output operations are fairly small in relation to the total input-output
time. The notable exception is the 2415 Magnetic Tape Unit. The processor is completely
locked out during a magnetic tape read or write operation. Furthermore, only printing on
the buffered 1403 Printers can proceed concurrently with magnetic tape reading or writing.
Table I summarizes the delays imposed upon central processor operations by most of the
Model 20 input-output devices in the "time-sharing" mode.
Input-output operations in Model 20 systems are handled differently from those in
larger models of the System/360. See Section 420:111, Simultaneous Operations (general), for
a description of how input-output operations are handled in the larger models. Note that data
chaining is not implemented for the Model 20.
In the Model 20, there are only three input-output instructions:
•
•
•
Transfer - controls the transfer of data between main storage and a
peripheral device.
Control - directs a peripheral device to perform a specified function,
such as select a stacker pocket, initiate a forms skip, etc.
Test I/O and Branch - causes a specified peripheral device to be
tested for a particular condition such as device busy or end, of forms,
and a branch to a specified location if the condition is found.
The status of the peripheral device addressed by a Transfer instruction is reflected
in the condition code (comparison and result indicators) at the completion of the Transfer instruction's execution. The three possible status conditions are: aVailable, busy, and not
operational. An interrupt is generated at channel end (data transfer completed) for each peripheral device. See Paragraph 422:051. 33 for details of the Model 20's interrupt procedure.
TABLE I: PROCESSOR DELAYS DURING I/O OPERATIONS
Cycle Time,
Function
Device
Card
Reading
2560 MFCM
2520 Read Punch
2501-A1 Reader
2501-A2 Reader
500
500
600
1,000
2560 MFCM
2520 Read Punch
2520-A2 Punch
2520-A3 Punch
1442-5 Punch
160
500
500
300
160
Card
Punching
Printing
63-char set
1403-2 Printer
1403-7 Printer
1403-Nl Printer
2560 MFCM
Tape
cpm
cpm
cpm
cpm
msee
120
120
100
60
*
col/sec
cpro
cpm
cpm
120
120
200
•
col/sec
Processor
Delay, maee
5
5
5
5
5.2
8
8
8
5.2
2203 Printer -
la-char set
39-char Bet
52-char Bet
Magnetic
Speed
2415 Magnetic
Tape Unit
7501pm
4251pm
3501pm
3001pro
6001pm
6001pm
1,100 lpm
140 col/sec
80
140
172
200
100
100
55
15,000 bytes/
*
t
15
30
33
43
2
2
2
27 max.
t
sec
* Varies with' number of columns punched or printed per card.
t Cycle time varies with size of block. The processor is delayed for the entire duration
of the magnetio tape operation.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
422: 151.1 00
IBM System/360
Model 20
Software
SOFTWARE
Because of their restricted instruction repertoire,
different I/O control methods, and limited core
storage, Model 20 systems will not be able to utilize
the extensive array of software that IBM is providing
for the larger System/360 models. The software
that is being prepared especially for Model 20 systems is summarized in the following paragraphs.
No operating system and no COBOL, FORTRAN, or
PL/I compiler has been announced to date for
Model 20 systems.
.1
System/360 Report Program Generators and run
on larger System/360 models, provided that an
adequate configuration of input-output equipment
is available. The only present exceptions are
the Model 20 RPG specifications supporting the
2560 MFCM (card printing and collating operations)
and the Dual Feed Carriage for the 2203 Printer.
.3
DATA SORTING AND MERGING
This is a group of four routines designed to perform most of the functions of conventional punched
card machines. The user writes specifications
describing the job to be performed on special
preprinted forms. This information is punched
into cards, fed into the Model 20, and used to
generate a program that can be executed inimediately to perform the specified functions. The
principal difference between these routines and
the RPG is that the RPG can handle a wider variety
of applications and can perform more functions
during a single run.
IBM has not announced a magnetic tape sort routine
for Model 20 systems to date, although it is probable that one will be provided to use the 2415
Magnetic Tape Units. See Paragraph. 3 for the
facilities provided for sorting, collating, and merging of punched card files.
.2
REPORT PROGRAM GENERATOR
IBM is emphasizing the use of the Report Program
Generator (RPG) for most Model 20 applications.
The System/360 Model 20 RPG is a generalized program designed to generate coding to perform most
routine business data processing functions. Input
to the RPG consists of specifications written by the
user in a format that is relatively easy to learn and
use. Separate preprinted specification sheets are
used to describe the input to be provided, the calculations to be performed, and the output to be
produced.
The four Punched-Card Utility Programs and
their functions are as follows:
Programs created by the Model 20 RPG can perform
any or all of the following functions (provided that
the necessary input-output equipment is available):
matching, merging, gang-punching, calculating,
reproducing, interpreting, printing, summary
punching, and card selection. Data records can be
obtained from up to three different input files. Calculations can include addition, subtraction, multiplication, division, crossfooting, comparison, and
data movements. Printed reports or punched summary cards can contain up to nine control levels.
(
\
"'-.
Generation of an object program requires a 2020
Processing Unit with 4, 096 bytes of core storage
and any.Qllil of the follOWing: a 2560 MFCM, 8.2520
Card Read-Punch, or a 2501 Card Reader and a
printer. Additional core storage can be utilized if
available. Execution of the object program requires
at least a 2020 Processing unit with 4,096 bytes of
core storage, one card input unit, and one card
punch or printer. The object program can utilize
most of the peripheral units and optional features
available for Model 20 systems. IBM expects most
RPG-generated programs to run at or near the peak
speed of the limiting hardware component.
With respect to program compatibility, IBM states
that virtually all programs written in Model 20
RPG can also be generated by the other IBM
©
PUNCHED-CARD UTILITY PROGRAMS
.4
•
Collate - performs merging, matching,
card selection, card insertion, sequence
checking, consecutive-number checking,
and hash total accumulation; requires 4K
bytes and a 2560 MFCM.
•
Gangpunch-Reproduce - performs the gangpunching, reproduCing, and interpreting
functions of conventional punched card machines; requires 4K bytes and either a 2560
MFCM, a 2520 Card Read-Punch, or a card
reader and a card punch.
•
List-Summary Punch - produces listings from
card files and/or punches a summary card for
each group of data cards; can print headings,
perform editing, and accumulate totals; requires 4K bytes, a printer, and either a 2560
MFCM, a 2620 Card Read Punch, or a card
reader and 3. card punch.
•
Merge-Sort ~ sorts a card file into either
ascending or descending sequence based on
keys contained in up to five alphameric or
numeric fields; practical for use on files that
have large alphameric keys or are largely
presequenced; requires 4K bytes and a 2560
MFCM.
BASIC UTILITY PROGRAMS
The following loading and diagnostic routines will be
provided to facilitate Model 20 operations: Absolute
Loader, Relocatable Loader, Basic Trace, Print
Storage, Punch Storage, and Clear Storage.
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360-MODEL 20
422.151.~00
.5
BASIC ASSEMBLER
The Punched Card 10CS program provides routines
for scheduling I/O devices, handling errorcorrection and restart procedures, supervising
central processor interruption, and generating
linkages to the user's object program.
The System/360 Model 20 Basic Assembler is a
symbolic assembly system that translates source
statements written by the programmer into machinelanguage instructions on a strict one-to-one-basis.
It permits full utilization of all Model 20 hardware
features. Input and output are on punched cards.
The object program can be in either absolute or relocatable format. A printed listing includes the
source statements, object program instructions,
and diagnostic messages.
Generation of an object program requires a 2020
Processing Unit with 4,096 bytes of core storage,
a card reading unit, a card punching unit, and
(optionally) a printer. Programs can be assembled
to run on any Model 20 configuration that includes
a card reading unit.
IBM states that the System/360 Model 20 Basic
Assembler language is upward compatible with the
Basic Assembler language for the larger System/360
models except for those features dealing with
specific hardware differences between the Model 20
and the larger models. These differences are primarily in the control of input and output operations.
6
INPUT/OUTPUT CONTROL SYSTEMS
The Input/Output Control Systems (IOCS) are designed to relieve the programmer of much of the
detailed coding involved in programming inputoutput operations for the Model 20 input-output
units and the Communications Adapter. Two programs are provided: one to generate input-output
control routines for conventional peripheral equipment; the second to generate control routines for
data transmission via the Communications Adapter.
Control routines are generated in symbolic coding
and are tailored to the user's 10CS specifications.
The routines can then be combined and assembled
with a source program in Basic Assembler language\
or they can be assembled separately in relocatable
text.
The Communications 10CS program provides
program-controlled point-to-point data transmission. This program supports data transmission between a Model 20 equipped with the
Communications Adapter and anyone of the
following IBM devices: 1009, 1013, 2701 (equipped
with a Synchronous Data Adapter - Type 1), 7701,
7702, 7710, 7711, or another Mode1.20 system.
.7
1419 I/O PROGRAM
This program facilitates the use of a 1419 Magnetic
Character Reader with a Model 20 system for
Federal Reserve and Commercial Bank Transit
Applications where capture of ON- US data is not
required. The 1419 I/O Program provides the
follOWing functic:ms: engaging and disengaging the
1419 MICR Reader, reading MICR documents,
buffer maintenance, presenting data for user processing, testing of field indicators, issuance of
I/O commands, and servicing of interrupts. Exits
are provided to allow the user to insert own-coded
routines for such functions as stacker determination, formatting of printer and punch output,
document arithmetic processing, recognition of
control levels, or card processing.
IBM states that with this program, when reading
6-inch documents, approximately 12 milliseconds
per document are available to the user, of which
about 5 milliseconds should be allocated to stacker
determination. When reading longer or shorter
documents" a proportiqnate amount of processing
time is available'.
The 1419 I/O Program for the Model 20 cannot be
used on larger System/360 models.
/
7/65
422:201.001
IBM System/360
Model 20
System Performance
SYSTEM PERFORMANCE
GENERALIZED FILE PROCESSING (422:201.100)
These problems involve updating a master file from information in a detail file and
producing a printed record of each transaction. This application is one of the most common
commercial data processing jobs and is fully described in Section 4:200.1 of the Users' Guide.
Standard File Problems A, B, and C show the effects of varying record sizes in the master
file. Standard Problem D increases the amount of computation performed upon each transaction. Each problem is estimated for activity factors (ratios of number of detail records to
number of master records) of zero to unity. In all cases a uniform distribution of activity is
assumed.
In Configuration I, both the master and detail input files are read by the card reader.
The output files are assigned to the card punch (updated master file) and printer (report file) .
For all of the Standard File Problems, the 300-cpm card punch is always the controlling factor
on overall processing time for Configuration I.
In Configuration II, the master files are on magnetic tape. The detail file is assigned to the card reader and the report file to the printer. The blocking of master-file records was held to 528 bytes to permit the Generalized File Processing Problem to be performed in the 8, 192-byte core store.
In a Model 20 system, magnetic tape operations cannot be overlapped with internal
computation by the processor or with any other peripheral operation except printing by a
1403 Printer. Thus, in tape-oriented Model 20 systems, there are three timing factors that
determine the speed with which a file-processing job is performed: (1) central processor
time, including the delays due to input-output operations; (2) 1403 printing time; and (3) the
magnetic tape reading and writing and card reading time. Note that the operations in factor
(3) cannot be overlapped and must be performed sequentially.
At low activity ratios, the central processor is the controlling factor for all of
the Standard File Problems. As the activity factor increases, the magnetic tape reading
and writing and card reading time tends to become the controlling factor. In File Problem
D, where the amount of computation is trebled, the central processor is the controlling factor
for all conditions. At high activity ratios with shorter master file records (Problems A and
B), the printer is the controlling factor.
SORTING (422:201. 200)
The standard estimate for sorting 80-character records by straightforward merging
on magnetic tape was developed from the time for Standard File Problem A by the method explained in Paragraph 4:200.213 of the Users' Guide. A two-way merge was used in System
Configuration II, which has only four magnetic tape units.
MATRIX INVERSION AND GENERALIZED MATHEMATICAL PROCESSING
Automatic floating-point hardware is not available for the System/360 Model 20
Processor; therefore, the mathematically-oriented standard problems have not been coded
for this system.
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360-MODEL 20
422:201.011
WORKSHEET DATA TABLE 1 (STANDARD FILE PROBLEM A)
CONFIGURATION
ITEM
I
1
msee/bloek
InputOutput
Times
msee/switch
msec penalty
SO
(File 1)
Char/bloek
Reeords/bloek
K
52S'
6
0.5
(File 1)
REFERENCE
II
67.2
60 (File 1); 200 (File 2)
File 1 = File 2
f - - - - - . - f- - - - - ---.;. 1 - - - - - 100
60
File 3
f-----.- f - - - - - -1-----125
125
File 4
4:200.112
0
0
File 1 = File 2
f-----f----- ---0
0
File 3
~4---- f - - - - - - f - - - - - 0
0
,
67.2
12 (File 1); 19 (File 2)
File 1 = File 2
--r - - - - - 1 - - - - - " - f - - -12
12
File 3
f-------1-----2
2
File 4
-------
----
2
Central
Processor
Times
msec/block
msec/record
msec/detail
msec/work
msec/report
b5+b9---·-
f - - - 23.1
- - - - - ' - - - -23.1
----
"""b7"+'bs--· -
1 - - - -1-1.-S-- - I - - ---rr:-S---
Standard
File
ProblemA
F = 1.0
for C.P.
and
dominant
column
~---- I-~
~ --_.- f -6.6
-19.2
a3K
File 1: Master In
4:200.1132
Printer
C.P.
1------ I--~
79.2
--_.- f -229.8
--- 1---67.2
12.0
I - - -1 - - - I - - - --19.0
~O_ 1--67~ f - - 72.0
6.0
File 3: Details
- - : - : : - - - - - - - - i - - - - f--- -:o- f-750
12
1.0
File 4: Reports
---.-
4:200.114
File 2: Master Out
200.0
66.0
Total
4
Puneh
C.P.
3
msec/block
2.S
2.S
a1
- - - - f--- --1-3.-2-- - - 1 - - - -13.2
a2
1---- - - - - f--------3.5
3.5
b6
750
530.2
Unit of measure (bytes)
2000 (1)
(?)
Std. routines
- - - - -f - - - - - - - - - - 1 - -1500
144
144
Fixed
Standard
File
Problem A
Space
---- ---_.- - - ' - - 714
714
3(Blocks 1 to 23)
-=-=-.-_
..,-:-:6(Blocks 24 to 48) I -
FUe;.---.--:::----
Working
Total
*
-
--aii60-- - I--
-~4----
----sii6o - - -
4:200.1151
I--~--
' - - -80- - - - I--~--6242
8002
The reduced block length was necessary to fit the program into the available core storage.
/
(Contd. )
7/65
422:201.100
SYSTEM PERFORMANCE
.1
GENERALIZED FILE PROCESSING
.11
Standard File Problem A
. 111 Record sizes Master file: ...•••• 108' data characters, packed
as 88 bytes.
Detail file: .••••.• 1 card.
Report file: .•••..• 1 line.
.112 Computation: •.••••• standard .
• 113 Timing basis: . • . . . . using estimating procedureoutlined in Users' Guille
4:200.113 .
• 114 Graph: . . • • . . . . . . . see graph below .
.115 Storage space required Configuration I: ..•. 6,242 bytes.
Configuration il: ...• 8,002 bytes.
1,000.0
7
4
2
100.0
7
I
4
Time in Minutes to
Process 10,000
Master File Records
-
',2
10.0
7
4
-
2
1.0
7
4
2
0.1
0.0
0.1
0.33
1.0
Activity Factor
Average Number of Detail Records Per Master Record
(Roman numerals denote standard System Configurations.)
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360-MODEL 20
422:201.120
• 12
standard File Problem B
.121 Record sizes Master file: .•.••.. 54 data characters, packed
as 44 bytes.
Detail file: .•.•••. 1 card.
Report file: ...••.• 1 line.
.122 Computation: .•••... standard .
'.123 Timing basis: .••••• using estimating procedure
outlined in Users' Guide,
4:200.12.
.124 Graph: .••••••.... see graph below.
1,000.0
7
4
2
100.0
7
4
Time in Minutes to
Prooess 10,000
Master File Records
~~ I
2
10.0
7
4
~
---
2
1.0
7
4
2
0.1
0.0
0.1
0.33
1.0
Aotivity Faotor
Average Number of Detail Reoords Per Master Record
(Roman numerals denote standard System Configurations.)
(Contd. )
7/65
422:201.130
SYSTEM PERFORMANCE
• 13
Standard File Problem C
.132 Computation: ••••••. standard •
.133 Timing basis: •.•••• using estimating procedure
outlined in Users' Guide,
4:200.13 •
. 134 Graph: ••••••••••• see graph below.
• 131 Record sizes Master File: .••••• 216 data characters,
packed as 176 bytes.
Detail file: •••••.• 1 card.
Report file: ••••••. 1 line.
1,000.0
7
4
2
(~ I
100.0
7
4
Time in Minutes to
Process 10,000
Master File Records
2
----
10.0
7
~
-
4
2
1.0
7
4
2
0.1
0.0
0.1
0.33
1.0
Activity Factor
Average Number of Detail Records Per Master Record
(Roman numerals denote standard System Configurations.)
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
422:201.140
.14
IBM SYSTEM/360-MODEL 20
Standard File Problem D
.141 Record sizes Master file: .••...• 108 data characters, packed
as 88 bytes.
Detail file: ..••••• 1 card.
Report file: .•••.•. 1 line.
.142 Computation: •••••.• trebled •
• 143 Timing basis: ..•••• using estimating procedure
outlined in Users' Guide,
4:200.13 •.
.144 Graph: ..•••.••... see graph below.
1,000.0
7
4
2
100.0
,
7
4
Time in Minutes to
Process 10,000
Master File Records
2
---
10.0
7
~
I
-
4
2
1.0
7
4
2
0.1
O. 0
0.1
0.33
1.0
Activity Factor
Average Number of Detail Records Per Master Record
(Roman numerals denote standard System Configurations.)
(Contd. )
7/65
SYSTEM PERFORMANCE
422:201.200
.2
SORTING
.21
Standard Problem Estimates
.212 Key size: . . • • • . . . . 8 characters .
.213 Timing basis: .•..•. using estimating procedure
outlined in Users' Guide,
4:200.213 .
• 214 Graph: ..••••.••.. see graph below .
. 211 Record size: . . . . . . . 80 characters.
1,000
7
Ll
4
2
~
100
V
V
Lt
I
7
10-'
4
1.1
1I
2
V
Time in Minutes to
Put Records into
Required Order
10
V
I
7
/
/
4
2
1
7
4
2
0.1
100
2
4
7
1,000
2
4
7
10,000
2
4
7
100,000
Number of Records
(Roman numerals denote standard System Configurations.)
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
422:221.101
IBM System/360
Model 20
Price Data
PRICE DATA
IDENTITY OF UNIT
CLASS
CENTRAL
PROCESSOR
No.
2020
Name
Processing Unit
Model B1 - 4,096 bytes
Model C1 - 8,192 bytes
Model D1 - 16,384 bytes
Model B2 Model C2 Model D2 -
4, 096 bytes*
8,192 bytes*
16,384 bytes*
PR1CES
Monthly
Rental
Monthly
Maintenance
Purchase
$
$
$
500
700
1,200
37.00
42.00
52. 00
24,300
33,600
56,800
575
775
1,275
40.00
45.00
55. 00
27,900
a7,200
60,400
225
200
275
30
55
22.50
22.50
22.50
2.50
5.00
11,250
11,000
11,750
1,500
2,750
2.00
6.00
3.00
4. 75
960
2,500
1,250
3,750
2. 25
* Model B2, C2, or D2 is
required for attachment of
a Communications Adapter,
Input/Output Channel,
Printer Features Control,
Serial I/O Channel, or Universal Character Set Adapter
4442
4447
4448
4460
8082
8090
8092
8095
8099
2501
2520
2520
2560
Attachment
Model A1 Attachment
Model A2 or A3 Attachment
Attachment
20
50
25
75
1580
3480
5575
8637
7081
4658
2073
Card Print Control
Dual Feed Carriage Control
Printer Features Control*
Universal Character SetAdapter~
Serial I/O Channel *
Input/Output Channel*
Communications Adapter*
25
10
55
15
100
150
150
2.00
3. 50
6. 50
4.75
11. 00
1,200
500
2,640
720
5,300
7,500
6,000
INTERNAL
STORAGE
INPUTOUTPUT
PeriI!heral AdaI!ters {on 2020)
1403 Model 2 Attachment
1403 Model 7 Attachment
1403 Model N1 Attachment
1442 Model 5 Attachment
2203 Attachment
1. 50
Core Storage is included in the
2020 ProceSSing Units (above).
1442
Card Punch (#4460 req'd. )
255
38.00
12,750
2501
Card Reader (#8090 req'd. )
Model A1 - 600 cpm
Model A2 - 1,000 cpm
195
255
30.00
42.00
11,700
11,950
Card Read Punch, Model A1
(#8092 req'd. )
700
84.00
33,500
Card Punch (#8095 req'd.)
Model A2 - 500 cpm
Model A3 - 300 cpm
625
450
80.00
60.00
30,000
29,700
2520
2520
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
422:221.102
IBM SYSTEM/360-MODEL 20
PRICES
IDENTITY OF UNIT
CLASS
INPUTOUTPUT
(Contd. )
Name
No.
2560
1575
1576
1577
1403
1403
1376
4740
5381
6411
8640
8641
2203
5558
3475
7815
Monthly
Maintenance
$
$
Purchase
,
$
Multi-Function Card Machine
(#8099 req'd.)
Card Print Feature First 2 Lines (#1580 req'd.)
Second 2 Lines (#1575 req'd.)
Third 2 Lines (#1576 req'd.)
575
90.00
28,750
125
125
125
13.00
13.00
13.00
6,250
6,250
6,250
Printer (600 lpm)
Model 2 (#4442 req'd.)
Model 7 (#4447 req'd.)
775
650
131. 00
130.00
34,000
32,700
Printer (1, 100 lpm)
Model N1 (#4448 req'd.)
900
166.00
41,200
75
18.75
3,075
On 1403 Model 2 or 7 Auxiliary Ribbon Feeding
Feature
Interchangeable Chain Cartridge
Adapter
On 1403 Model 2 only (#5575
req'd.) Numerical Print Feature
On 1403 Model 2 orN1 Selective Tape Listing Feature
Universal Character Set Feature
(#8637 req'd.) For Model N1
For Model 2
Printer (300-750 lpm; #8082
req'd. )
On 2203 Printer 24 Additional Print Positions
Dual Feed Carriage (#3480
req'd. )
6 Additionru Tape Channels
(#3475 req'd.)
1419
Magnetic Character Reader
(#7081 Serial 170 ChanneIReq'd.)
2415
M~etic
3228
7125
Monthly
Rental
Tal2e Unit
15,000 char/sec (#4658 Input/
Output Channel req'd.) Model 1 (two tape drives and
single-channel controller)
Model 2 (four tape drives and
Single-channel controller)
Model 3 (six tape drives and
single-channel controller)
On 2415, any model Data Conversion
Seven-Track Compatibility
-
3,125
225
7.50
9,050
190
10.00
8,100
10
10
1. 75
1. 75
450
450
510
71. 50
23,000
45
4.00
2,475
100
8.50
5,000
10
1. 00
400
2,275
240.00
110,500
750
100.00
32,750
1,200
180.00
58,800
1,650
260.00
80,850
45
50
1. 00
1. 25
2,160
2,400
75
Note: Indicated monthly maintenance charges are those in effect for the first 36 months after
installation.
7/65
IBM SYSTEM/360
MODEL 30
International Business Machines Corp.
I
~.
-~
J
(.
c
AUERBACH INFO, INC.
PRINTED IN U. S. A.
IBM·SYSTEM/360
MODEL 30
International Business Machines Corp.
'--~~
AUERBACH INFO, INC.
PRINTED IN U. S. A.
423:011.100
IBM System/360
Model 30
Introduction
INTRODUCTION
The Model 30 series of the IBM System/360 family is a small-scale, general-purpose
computer system, with typical system rentals ranging from approximately $4,000 (for a card
system) to $6,600 (for a six-tape system) per month. The Model 30 has approximately the same
processing power as the older IBM 7010 system. The available Compatibility Features indicate
that Model 30 is regarded by IBM as a probable replacement computer for IBM 1401, 1440, 1460,
and 1620 systems.
Model 30 is the smallest System/360 processor that can include the full System/360
instruction repertoire. It cannot be connected to some of the faster direct-access peripheral
devices or to the 2361 Large-Capacity Core storage Unit. The available core storage sizes,
which range from 8,192 to 65,536 bytes, impose restrictions on the software which can be used
with Model 30 systems.
Model 30, with the 1401/1440/1460 Compatibility Feature, is the only System/360
processor that can execute programs written in a "foreign" machine code without requiring
additional core storage. Model 30 is presently faster than Model 40 when both are emulating
the same 1400 Series system, because Model 40 uses a combined hardware/software "hybrid"
approach instead of the direct hardware approach used in Model 30.
The Model 30 series of the IBM System/360 is characterized by:
•
A main core storage cycle time of 1. 5 microseconds, with one byte
being accessed per cycle. Storage accessing is not interleaved, so the
effective core cycle time is 1. 5 microseconds per byte. (As originally
announced, in April 1964, Model 30 had a cycle time of 2 microseconds
per byte. The faster 1. 5-microsecond rate was announced in January 1965.)
•
Main core storage capacities ranging from 8,192 to 65,536 bytes.
•
standard inclusion of the Multiplexor Channel, which allows simultaneous
operation of a number of low-speed input-output devices such as printers,
card readers, and communication terminals. This makes the Model 30
system suitable for use as an independent system and/or as a satellite
system to handle the input-output processing of a larger system.
•
standard inclusion of the following features:
I
'",
System Control Panel
standard Instruction Set.
•
Optional availability of the following features:
Channel-to-Channel Adapter
Decimal Arithmetic
Direct Control
Floating-Point Arithmetic
1401/1440/1460 Compatibility or
1620 Compatibility (not both)Interval Timer
Selector Channels (maximum of 2)
Storage Protection
High-Speed Multiplexor Channel.
•
Non-availability of the following features:
Shared Processor Storage
1410/7010 Compatibility Feature
7070/7074 Compatibility Feature
7080 Compatibility Feature
709/7040/7044/7090/7094/ Compatibility Feature
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
423:011.101
IBM SYSTEM/360-MODEL 30
•
Non-availability of the following peripheral devices:
2150
2361
2301
2314
Console
Large-Capacity Core Storage
Drum Storage
Direct Access Storage Facility
This report concentrates upon the characteristics and performance of the Model 30
series in particular. All general characteristics of the System/360 hardware and software
are described in Computer System Report 420: IBM System/360 - General.
The System Configuration section which follows shows the System/360 Model 30
in the following standard System Configurations:
I
Card System
II
4-Tape Business System
m
6-Tape Business System
v
6-Tape Auxiliary Storage 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.
As a matter of general interest, the rentals that would be incurred when either the 2321 Data
Cell Drive, the 2302 Disk Storage Unit, or the 2311 Disk Storage Drives are used to provide
the necessary random access storage capability are listed separately on the diagram for the
Auxiliary Storage System, Configuration V.
Section 423:051 provides detailed central processor timings for the Model 30. See
Section 420:051 for all the other characteristics of the program-compatible System/360 processors, including a discussion of the System/360 instruction repertoire as a programming tool.
Program compatibility with the IBM 1400 Series systems is discussed in Section 420:131.
The input-output channel capabilities of the System/360 Model 30 can be summarized
as follows (see also Sections 420:111 and 423:111):_
Standard Multiplexor Channel
Maximum number: . . . . . . . . . . . . . . . . .
Maximum data rate, bytes/sec Multiplex mode: . . . . . . . . . . . . . . . . . . .
Burst mode: . . . . . . . . . -• . . . • . . . . . . .
Maximum number of subchannels: • . . . . . . .
1.
31,000.
267,000.
244.
High-Speed Multiplexor Channel
Maximum number: . . . . . . . . . . . . . . . . . . 1.
Maximum data rate, bytes/sec Multiplex mode: . . . . . . . . . . . . . . . . . . 200,000.
Burst mode: .... . . . . . . . . . . . . . . . . • 200, 000.
Maximum number of subchannels: • . . . . . . . 4.
Selector Channels
Maximum number: . . . . . . . . . . . . . . . . • . 2. *
Maximum data. rate, bytes/sec: . . . . . . . . . 333,000 per channel
The software that can be used with any System/360 configuration depends upon the core
storage capacity and the number and type of peripheral devices. Several versions of the Assembler, COBOL, FORTRAN IV, and PL/I will be made available. These languages and the other
support routines for the System/360 are described in Sections 420:151 through 420:193.
The overall performance of any System/360 is heavily dependent upon the Processing Unit
model used. A full System Performance analysis of the Model 30 is provided in Section 423:201.
*If the High-Speed Multiplexor Channel is used, only one Selector Channel can be installed.
7/65
423:031.100
IBM System/360
Model 30
System Confi gurotion
SYSTEM CONFIGURATION
System configuration possibilities for Model 30 and other System/360 models are summarized in
report Section 420:031. This report section shows Model 30 systems arranged in accordance with the
specifications for our Standard Configurations, as defined in the Users' Guide, page 4:030.120.
.1
TYPICAL CARD SYSTEM; CONFIGURATION I
Deviations from Standard Configuration: . . . . . . . . • . . . . card punch is 50% faster.
Equipment
Rental
Main Storage (8,192 bytes)
2030 Processing Unit, Model C30
(includes one Multiplexor Channel)
",275
2821 Control Unit, Modell
970
2540 Card Read Punch, Modell:
Reads 1,000 cards per minute
Punches 300 cards per minute
660
1403 Printer, Model 3
(132 print positions):
Prints 1,100 lines per minute
1100 LPM Printer Attachment
1416 Print Train Cartridge
900
Optional Features Included: . . • • . • • . • • • . . . . . . . . . . . Decimal Arithmetic
TOTAL:
©
i
1965 AUERBACH Corporation and AUERBACH Info, Inc.
75
100
25
$4,005
7/65
IBM SYSTEM/360-MODEL 30
423:031.200
.2
4-TAPE BUSINESS SYSTEM; CONFIGURATION II
Deviations from Standard Configuration: • • • • . . . . . . . . . printer is 20% faster.
reader is 20% faster.
Equipment
Rental
Main Storage (8,192 bytes)
2030 Processing Unit, Model C30
. (includes one Multiplexor Channel)
\
$1.~5
2501 Card Reader, Model Bl:
Reads 600 cards per minute
260
1442 Card Punch, Model N2:
Punches 91 full cards per minute
375
2821 Control Unit, Model 2:
1403 Printer, Model 7:
Prints 600 lines per minute
600
650
2415 Magnetic Tape Unit and
Control, Model 2 (includes 4
drives): 15,000 bytes/second)
Optional Features Included: . . . • • . . • • . . • . . . . . . . . . . Decimal Arithmetic
Selector Channel
TOTAL:
1,200
25
215
$4,600
/
(Contd. )
7/65
/
423:031.300
SYSTEM CONFIGURATION
.3
6-TAPE BUSINESS SYSTEM; CONFIGURATION ill
Deviations from Standard Configuration: . . . . . . . . . • . . . . printer is 20% faster.
reader is 20% faster.
Equipment
Rental
Main Storage (16,384 bytes)
$1,775
2030 Processing Unit, Model D30
(includes one Multiplexor Channel)
1051 Control Unit and Adapter
1052 Printer-Keyboard
225
2501 Card Reader, Model B1:
Reads 600 cards per minute
260
1442 Card Punch, Model N2:
Punches 91 full cards per minute
375
2821 Control Unit, Model 2
1403 Printer, Model 7:
Prints 600 lines per minute
600
650
2403 Magnetic Tape Unit and
Control
2402 Magnetic Tape Unitu (2)
2401 Magnetic Tape Unit
(6 drives total)
AU Modell: 30,000 bytes/sec.
Optional Features Included: • • • . . . . • • . • . . . . . . . . . . . Selector Channel
Decimal Arithmetic
TOTAL:
2,460
215
25
$6,585
(
"'-©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
423:031.400
.4
IBM SYSTEM/360-MODEL 30
6-TAPE AUXILIARY STORAGE SYSTEM; CONFIGURATION V
Deviations from Standard Configuration: • . . . . • • . . . . . . up to 380 million bytes more
auxiliary storage.
printer is 20% faster.
reader is 20% faster.
Equipment
Rental
Main Storage (16,384 bytes)
$ 1,775
2030 Processing Unit, Model D30
(includes one Multiplexor Channel)
1051 Control Unit and Adapter
1052 Printer-Keyboard
225
2501 Card Reader, Model B1:
Reads 600 cards per minute
260
1442 Card Punch, Model N2:
Punches 91 full cards per minute
375
2821 Control Unit, Model 2
1403 Printer, Model 7:
Prints 600 lines per minute
2403 Magnetic Tape Unit and
Control
2402 Magnetic Tape Units (2)
2401 Magnetic Tape Unit
(6 drives total)
All Modell: 30,000 bytes/sec.
2841 Storage Control and 8079
Attachment
2321 Data Cell Drive, Modell:
up to 400 million bytes storage;
access time: 175 to 600 msec.
Optional Features Included: . . • • • • . . • • . . . . . • . . . . . . Selector Channel
Decimal Aritlunetic
TOTAL:
Note:
l
600
650
2,460
3,500
215
25
$10,085
The following can be used in place of the 2321 Data Cell Drive, resulting in
the indicated total rentals:
2302 Disk Storage, Model 3, and 7950 Attaclunent - average access time:
165 msec; 112.14 million bytes storage:
$12,960
2311 Disk Storage Drives (3) - average access time: 128 msec; 21. 75
million bytes storage:
$ 8,830
423:051.100
IBM System/360
Model 30
Central Processor
CENTRAL PROCESSOR
.1
GENERAL
. 11
Identity: ••••.•..•• IDM 2030 Processing Unit .
.12
Description
See Section 420:051 for a comprehensive description of the characteristics of all the System/360
Processing Units.
See Section 423:011 for a summary of the distinguishing features of the 2030 Processing Unit as
used in Model 30 systems.
The Instruction Times and Processor Performance
times for Model 30 systems, in all four modes of
arithmetic, are listed below. See Paragraphs
4:050.41 and 4:050.42 of the Users' Guide for the
definitions of these standard measures of central
processor performance.
.4
PROCESSOR SPEEDS
.41
Instruction Times in Microseconds
. 411 Fixed point -
.412
.413
• 414
.415
.416
• 417
I
\.
Binary
Decimal (optional)
Add-subtract: .••..• 30
45 + 4B
Multiply: .•••.•.... 235
34 + 35B + 21B2
34 + 60B + 54B2
Divide: .•.•••••••• 420
where B = operand length
in eight-bit bytes (2 decimal
digits per byte).
Floating point (optional) Long
Short
Add-subtract: ••••• 87
~
Multiply: •••.••••• 800
245
Divide: .••.•••••• 1, 643
300
Additional allowance for Single indexing: •••• 0.0
Double indexing: •••• 4.5
Indirect addressing: • none.
Recomplementing: •. none •
Control:
Compare Fixed point:. • • • • • 30
Decimal: .•.••••• 45 + 4B
Floating point
(long) : . . • • • • • . 73
optional
Floating point
(short): .••••••• 49
Logical: •••••..• 33 + 4B
Branch: .•••••..• 17
Counter control Step: ••••••.•••• 30
Step and test: •••••. 19 to 23 (increment of -1).
38 (increment of any value).
Test: .•••••.•.•. 30
Edit: .••••••••••• 38 + 16B (optional).
ConvertTo binary: •••••.•• 89 + 0.75H + 3H2
To decimal: .•••••. 46 + 18H + 1. 5H2
where H = number of
significant hexadecimal
digits.
.418 Shift: .••••••••..• variable.
.42 Processor Performance in Microseconds
.421 For random addresses Fixed point Floating point
c = a + b: ••••••... 78 (binary) 161 (long)
75 + 7B
107 (short)
(decimal)
b = a + b: .•••.•••• 78 (binary) 161 (long)
107 (short)
45 + 4B
(decimal)
Sum N items: .•.•••• 30N (binary) 87N (long)
(45 + 4B)N
57N (short)
(decimal)
c = ab: .••..••.••• 296 (binary) 874 (long)
77 + 43B + 295 (short)
21B2
(decimal)
c = alb: .••.•.••.. 481 (binary) 1, 717 (long)
77 + ~8B + 350 (short)
54B
(decimal)
where B = operand length
in eight-bit bytes .
.422 For arrays of data Fixed point Floating point
ci = ai + bj= . . . . • . • . 144 (binary) 227 (long)
141 + 7B
173 (short)
(decimal)
bj = ai + bj: .•.••.•. 144 (binary) 227 (long)
173 (short)
111 + 4B
. (decimal)
Sum N items: .•••••• 68N (binary) 125N (long)
(83 + 4B)N 95N (short)
(decimal)
c = c + aibj: .••.•.•• 392 (binary) 1,027 (long)
188 + 47B + 418 (short)
21B2
(decimal)
.423 Branch based on comparison Numeric data: .•••• 131N
Alphabetic data: .••• 124N
.424 SwitchingUnchecked: .•••••. 125
Checked: .•.••.••• 187
List search: " ••.. 108 + 84N
.425 Format control, per character Unpack: ..••••.•• 6.4
Compose: .••••... 16.1; 23.5 with radix
conversions.
.426 Table lookup, per comparisonFor a match: •••.•• 84
For least or greatest: 86.4
For interpolation
point: . • • . . . . . • . 84
.427 Bit indicators Set bit in separate
location: .••••••. 14
Set bit in pattern: ..• 16
Test bit in separate
location: • . . . . . . . 32
Test bit in pattern: .. 33
.428 Moving: .•.••..••• 30 + 3B t where B = number
of bytes moved.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
/
423: 111.100
IBM System/360
Model 30
Simul taneous Operations
SIMUL TANEOUS OPERATIONS
A System/360 Model 30 system can concurrently execute:
•
One machine instruction; and
•
Up to two input-output operations, one on each of the Selector Channels; and
•
Multiple additional input-output operations via the Multiplexor Channel.
Alternatively, up to four operations on a High Speed Multiplexor Channel can be overlapped with
internal processing and one Selector Channel operation.
When the Multiplexor Channel is operated in the burst mode, only one input-output operation
can proceed on that channel. Internal processing cannot be overlapped with burst mode operation, but operations previously initiated on the Selector Channels will proceed concurrently
with the burst mode operation.
The demand on the central processor (i. e., the "interference" or delay imposed upon the central processor program by each individual input-output operation) will vary depending on
whether the peripheral device is connected to one of the Selector Channels or to the Multiplexor
Channel. (See the general discussion of System/360 Simultaneous Operations in Section420:1l1.)
In Table I the processor demands imposed by each of the peripheral units are listed for both
types of channels.
The specific characteristics of the Model 30 Selector, Multiplexor, and High Speed Multiplexor
Channels can be summarized as follows:
Selector Channels
Maximum number: ..••••••...••.•..••.
Maximum data rate per channel: .•.•••..•••
Maximum data rate; Multiplexor Channel
not operating Selector Channell: ..•.•••..•••.••.
Selector Channel 2: .•.••.•..•••.•..
Maximum data rate; Multiplexor Channel
operating Selector Channell: .••••.•••••.•••.
Selector Channel 2: .•••.•••.....••.
Processor demand: ..••.•••.•••••.•••.•
Number of control unit positions: .•••.•....
2.
250 kilobytes/sec.
250 kilobytes/sec.
200 kilobytes/sec.
250 kilobytes/sec.
125 kilobytes/sec.
see Table I.
8 per channel.
Multiplexor Channel
Maximum number: . • . . . . • . . . . • . . • • . . . .
Maximum data rate; Selector Channels
not operating Multiplexed mode: .•••••••••...••.
Burst mode: ••.•••.••••.•.•••.•.
Processor demand: .••.••••••••...•..•.
Number of control unit positions: .••••....•
Maximum number of subchannels: . . . . • . . . • •
1.
31 kilobytes/sec.
267 kilobytes/sec.
see Table 1.
8.
224.
High Speed Multiplexor Channel
Maximum number: •••••••.•...........
Maximum data rate Multiplexed mode: .•••••••.••••••.
Burst mode: •.••••••••.••.•...••
Number of control unit positions: ••••••..•.
Maximum number of subchannels: ..••••.••.
/
(
\
1 (pre-empts one Selector Channel).
200 kilobytes/sec.
200 kilobytes/sec.
4.
4.
"
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360-MODEL 30
423: 111. 101
TABLE I: INPUT-OUTPUT DEMANDS ON THE SYSTEM/360 MODEL 30 PROCESSOR
Device
Random Access
2302 Disk Storage
2311 Disk Drive
2321 Data Cell Drive
7302 Drum
Punched Card
2540 Card Read Punch:
Read, 1000 cpm
Punch, 300 cpm
1442 Model N1 Card Read Punch:
Read, 400 cpm
Punch, 91 cpm
1442 Model N2 Card Punch, 91 cpm
2520 Model B1 Card Read Punch:
Read, 500 cpm
Punch, 500 cpm
2520 Card Punch:
500 cpm (Model B2)
300 cpm (Model B3)
2501 Card Reader:
600 cpm (Model B1)
1,000 cpm (Model B2)
Peak
Data Rate,
KB/sec*
156
156
54.7
135
156
156
54.7
135
Demand on Processor,
per cent, viaMultiplexor
Selector
Channel
Channel
23
23
8. 2
20
100
100
100
100
1.3
0.40
O. 19
0.06
3.30
1. 60
0.12
0.53
0.12
0.12
0.08
0.02
0.02
3.30
0.75
0.75
0.67
?
0.67
0.67
0.10
0.10
4.20
1. 84
?
?
0.67
0.40
O. 10
0.06
1. 84
1. 20
0.80
1.3
0.80
1.3
0.12
0.19
5.00
8.20
70
70
0.53
O. 12
Printers
1403:
132 columns, 600 lpm (Model 2)
120 columns, 600 lpm (Model 7)
132 columns, 1, 100 lpm (Model 3,
N1)
1404, 132 columns, 600Ipm
1443, 120 columns, 240Ipm
1445, 113 columns, 190Ipm
70
70
1.3
1.2
0.19
0.18
1. 80
1. 20
70
70
58
48
2.4
1.3
0.48
0.36
0.36
0.19
0.07
0.05
3.35
1. 80
1. 40
1. 16
Punched Pal2er Tal2e
2671 Paper Tape Reader, 1,000 cps
1
1
0.15
6.24
30
60
90
15
170
30
60
90
15
170
Maggetic Tal1e
2400 Series:
Model 1, 30 KB/sec
Model 2, 60 KB/sec
Model 3, 90 KB/sec
2415, 15 KB/sec
7340 Hypertape, 170 KB/sec
*Kilobytes (thousands of bytes) per second.
7/65
Average
Data Rate,
KB/sec*
4.5
9. 0
13.5
2.2
25
100
100
100
100
100
423:201.001
IBM System/360
Model 30
System Performance
SYSTEM PERFORMANCE
GENERALIZED FILE PROCESSING (423:201.l(iO)
These problems involve updating a master file from information in a detail file and
producing a printed record of each transaction. This application is one of the most common
commercial data processing jobs and is fully described in Section 4:200. 1 of the Users' Guide.
Standard File Problems A, B, and C vary the record sizes in the master file. Standard
Problem D increases the amount of COJDPutation performed upon each transaction. Each problem
is estimated for activity factors (ratios of number of detail records to number of master records)
of zero to unity. In all cases a uniform distribution of activity is assumed.
The Generalized File Processing problem for the System/3S0 was coded in two ways one using master files in packed decimal format and computation in decimal arithmetic, and the
second USing master files primarily in binary format and computation in fixed-point binary. The
decimal computations required more time than those in fixed-point binary; but in the binary case,
items in the detail and report files needed radix conversion. There were no appreciable differences in the total times for the two cases. The graphs for the file problem are based on use of
the fixed-point binary technique. For simplicity, the very similar curves based on the use of
decimal arithmetic are not shown.
In the master file record layout, alignment of data items in core storage was carefully
considered. Double-word boundaries were observed for input-output purposes to improve performance efficiency on the larger models. Instead of the "chain" mode (scatter-gather) of tape
reading and writing, individual records were moved to the work areas using a high-speed,
multiple-register transfer m~thod.
In Configuration I, the master and detail input files are on the card reader. The output
files are on the card punch (updated master file) and printer (report file). For all of the file
processing problems, the 300-cpm card punch is always the controlling factor on overall processing time for Configuration 1.
In Configurations I! and II!, the master files are on magnetic tape. The detail file is
assigned to the card reader and the report file to the printer. The master-file tapes are connected to a Selector Channel, and tape reading or writing can fully overlap prooessing and
other input-output operations. The curves for Configurations II and III indicatethatforProblems
A, B, and D, the master file tapes are controlling near zero activity and the printer is controlling at activities greater than 0.1. For Problem C, where larger master file records are
used, the master file tapes are controlling at lower activities and the printer at higher activities.
Coding space for Configuration I! had to be minimized, and the blocking of master file records
was held to 528 bytes to permit the Generalized File Processing Problem to be performed in
the 8, 192-byte core store.
Because multiprogramming of two or more independent programs is afeatured capability of
the System/360, the time actually used by the central processor (CP) is also plotted. By comparing the curves of total time for the various configurations with the central processor curves,
it can be seen that even in the worst cases (Configuration II! with trebled processing or short
master records) some 65% of the available processing capacity is not in use. A comparison of
the central processor curves for a standard amount of computation and for trebled computation
(i. e., the curves for Problems A and D, respectively) shows the effect of increasing the computatIOnal workload.
SORTING (423:201. 200)
The standard estimate for sorting 80-character records by straightforward merging
on magnetic tape was developed from the time for Standard File Problem A by the method explained in Paragraph 4:200. 213 of the Users' Guide. A two-way merge was used in System
Configuration I! (which has only four magnetic tape units) and a three-way merge in Configuration m. The results are shown in Graph 423:201. 200.
MATRIX INVERSION (423:201. 300)
The standard estimate for inverting a non-symmetric, non-singular matrix was computed by the simple method described in Paragraph 4:200.312 of the Users' Guide. In order to
execute this procedure in floating-point form, the optional Floating-Point Arithmetic feature must
be included on the Model 30. Two lines are shown on the graph, one using the short floatingpoint format (S-digit precision) and the other using the long format (IS-digit precision).
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
423:201.002
IBM SYSTEM/360-MODEL 30
GENERALIZED MATHEMATICAL PROCESSING (423:201.400)
Standard Mathematical Problem A is an application in which there is one stream of
input data, a fixed computation to be performed, and one stream of output results. Two variables
are introduced to demonstrate how the time for a job varies with different proportions of input,
computation, and output. The factor C is used to vary the amount of computation per input
record. The factor R is used to vary the ratio of input records to output records. The procedure
used for the Standard Mathematical Problem is fully described in Section 4:200.4 of the Users'
Guide.
The optional Floating-Point Arithmetic feature must be included on the Model 30 in
order to perform the floating-point computations required in this problem. Double-length operations (16-digit precision) were used since a minimum precision of 8 digits is prescribed. The
input file is on the card reader, and the output file is assigned to the printer.
/
WORKSHEET DATA TABLE 1
CONFIGURA TlON
ITEM
I
1
msec/block
Standard
File
Problem A
InputOutput
Times
(estimated)
Central
(File 1)
msec/block
maee/record
Processor
meec/detail
Tlmee
maee/work
maee/report
Standard
for C.P.
File
Problem A
and
F=1.0
column.
dominant 1/0
- - 9 - a - - · - I-- - - 1 2 - 5 - - - r--~4:200.112
File 1 = File 2
'Filea--· -
0
0
0
- - - 0 - - · - f - - - o- - · - t - - - -O
- - - 0 - - - 1 - - - - - - - t - - - -O0
File = File 2
fru.3---.-
'File4 ---. -
9.6
f-----.9.6
0.8
1 - - - - 9.6
- - - r--
t-~.-4--·- t--~.-4--
-
-----;;:s -
t---14.4 2.2
----u-
~---.- 1 - - - - _ . - 1 - - - - - - \ - - _ . 2.5
2.5
b6
1.9
~9---- 1 - - - - - - f - - - - - - - - - - - 2.7
2.7
2.0
1------.-1-----_.- f - - - - - b7 + bS
a.s
C.P.
1.7
I-~
a1
i---=-- - - . a2K
~---.-
I - a.s-
File 1: Master In
9.6
~---.-=-=-
9.6
File 2: Master Out
1 - - - - - . -1 - - 4.8
5.1
Punch
C.P.
1.7
Printer
--.- ~."B""
O.S
t--- -
37.2
200
----5.1
C.P.
1.7
Printer
~4-
1.6
f---roo- I- o:B""""""
I---- -
4:200.1132
1 - - - --9-.6- ! - - -
-_.- 1---:4s.~ 1 - - -
-7:"2
Total
j---
---
4:200.114
I--- - t--~
1 - - -1 - - 115:2- t-- 57.6
8 6 . 4 I-- 750:0""""
750.0
198.3
172.8- t-"J.50Q.0
394.9
1500.0
Unit of measure (bytes)
~outines
6,000
2,000**
_
f-~8--- I-- ----:t2s - - - - - 1 2 8 ~--- 1 - - - - - - - - - _ . - --~-
Standard
File
Problem A
3 (Blocks 1 to 23)
--::-':::--.---'::- f~ke24to~
Space
648
4 ; 0 9 2 - - - I--
I-~--.-
f--
~--- 1 - - -80- - - f-Working
5,692
Total
Fixed/Floating point
2,512 - -
--4~-
----we; --- - - 1 0 0 -
Floating point
15,592
Floating point
Unit name
input
f--'---- output
Size of record
~--.output
80 bytes
---_.1------.t-80 bytes
80 bytes
~-2L_
60
100
t-----------55
100
msec/block
msec penalty
(estimated)
output
T2
~---.!L_
output
T
m~ec/report
T7
.
38S
4:200.1151
3,062 - - - r---~-
8,190*
Floating point
maec/record
T5
--'='-._- - - - ~5loops
___ ~
*
-
1.6
2.2
1.7
File a: Details
Standard
Mathernatical
Problem A
t - - 100 -
- - . - 1 - - - - - - t--~---- 1 - - -O.S
1.1
File 4: Reports
5
51.2
1------.-
-=-:---:-_. -
4
12
6
67.2
60/200
- - 6 - 0 - - · - c- - - - 1 0 0 - - -
3
m ••c/block
1,056
528*
0.5
File1-File2
r-ru,;-a-·--' -
~4---
maee penalty
2
K
File 4
maee/switch
80
(File 1)
Char/block
Records/block
REFERENCE
m
II
2540 Card Read Punch 2501 CardReader, ModB1 2501CardReader, ModB
--.--:-:cr-::-::-_-- ':-:-=1403 Printer, Mod 3
1403 Printer, Mod 7
1403 Printer, Mod 7
r-:-.-- --_-
80 bytee
9.6
c------9.6
9.6
I------~-
9.6
7.9
7.9
---r--""42.Q- 1 - - - - 42.0
'-------4.8
-
t-----4.8
80 bytes
80 byte~100
r---liiO-
4:200.413
9.6
---_
.9.6
7.9
\-~-
----_.4.8
,/
Coding space has been mm1mized and the length of tape blocks halved m order to fit the program's reqlllrements lOto the
available storage .
•• Estimated coding space required for non-standard tape-handling and error routines.
(Contd. )
7/65
SYSTEM PERFORMANCE
423:201.1 00
·1
GENERALIZED FILE PROCESSING
· 11
Standard File Problem A
.113 Timing basis: ..••.. using estimating procedure
outlined in Users' Guide,
4:200.113 .
. 114 Graph: . . . . . . . . . . . see graph below.
· 111 Record sizes Master file:· . . . . . . . 108 data characters, packed
as 88 8-bit bytes.
Detail file: . . . . . . . 1 card.
Report file: . . . • . • . 1 line •
. 112 Computation: . . . • . . . standard, using fixed-point
binary or decimal
arithmetic.
.115 Storage space required
Configuration I: ••••
Configuration II: .•••
Configuration III: ..•
5,692 bytes.
8,190 bytes.
15,592 bytes.
1,000
7
4
2
100.0
7
1
4
Time in Minutes to
Process 10,000
Master File Records
2
~
10.0
..,
7
~
./
/'
II
4
___ Cl'
/
2
fJll../
/
V,-
---- -
C~"
/
1.0
7
L
.'
4
I'
2
If
0.1
0.0
0.1
0.33
1.0
Activity Factor
Average Number of Detail Records Per Master Record
(Roman numerals denote standard System Configurations; curve marked "CP"
shows central processor time, for Configuration III only.)
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360-MODEL 30
423:201.120
• 12
Standard File Problem B
. 121 Record sizesMaster file: .•.•.. 54 data characters, packed
as 44 8-bit bytes.
Detail file: . . . . . . • 1 card.
Report file: .•.••.. 1 line.
.122 Computation: . . . . . . . standard, uSing fixed-point
binary or decimal
arithmetic •
. 123 Timing basis: . • . . . . using estimating procedure
outlined in Users' Guide,
4:200.12 .
• 124 Graph: . • . . . . . • . . . see graph below.
1,000
7
4
2
100.0
7
4
Time in Minutes to
Process 10,000
Master File Records
c~ I
2
10.0
7
,./
4
/
II
2
7
1.0
2
/
/
----.
Cl' - - - - -
c~
~"
,,
7
4
-
~
I
1/
/
0.1
0.0
0.1
0.33
1.0
Activity Factor
"Average Number of Detail Records Per Master Re-cord
(Roman numerals denote standard System Configurations; curve marked "CP"
shows central processor time, for Configuration III only.)
/
(Contd. )
7/65
SYSTEM PERFORMANCE
. 13
423:201.130
Standard File Problem C
. 131 Record sizesMaster file: . . . . . . . 216 data characters, packed
as 176 8-bit bytes.
Detail file: . . • • . . . 1 card.
Report file: . . . . • . . 1 line.
. 132 Computation: . • . . . . . standard, using fixed-point
binary or decimal
arithmetic .
• 133 Timing basis: . . . . . • using estimating procedure
outlined in Users' Guide,
4:200.13.
.134 Graph: . . . . . • • . . . • see graph below.
1,000
7
4
2
c~ I
100.0
7
4
Time in Minutes to
Process 10,000
Master File Records
2
10.0
~
IT
7
./
./
4
.---- CP-....
/
f-ill
[,...-,'"
2
/
1.0
------
c~/
7
~
4
/
1/
2
0.1
0.0
0.1
0.33
1.0
Activity Factor
Average Number of Detail Records Per Master Record
(Roman numerals denote standard System Configurations; curve marked "CP"
shows central processor time, for Configuration ill only.)
©
1965 AUERBACH Corporotion and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360-MODEL 30
423:201.140
.14
Standard File Problem D
• 141 Record sizes Master file: . • . . . • . 108 data characters, packed
as 88 8-bit bytes.
Detail file: . . • . . . . 1 card.
Report file: ..•••.• 1 line.
.142 Computation: . . . . . . • trebled, using fixed-point
binary or decimal
arithmetic.
.143 Timing basis: .••..• using estimating procedure
outlined in Users' Guide,
4:200.14.
.
.144 Graph: . . . • • • . . . . . see graph below.
1,000
7
4
2
100.0
..
7
I
4
Time in Minutes to
Process 10,000
Master File Records
2
~
10.0
.-
7
./'
2
L
IT
4
./
~
V
~.
,""-
.---C'" -
---.
----
~/
",C
1.0
,
7
.'
I
4
2
I
I
0.1
0.0
0.1
1.0
0.33
Activity Factor
Average Number of Detail Records Per Master Record
(Roman numerals denote standard System Configurations; curve marked "CP"
shows central processor time, for Configuration III only.)
(Contd. )
7/65
SYSTEM PERFORMANCE
423:201.200
.2
SORTING
.21
Standard Problem Estimates
.213 Timing basis: ..•.•. using estimating procedure
outlined in Users' Guide,
4:200.213 (2-way tape
merge in Configuration il;
3-way tape merge in
Configuration ill) •
.214 Graph: . . . . . . . . . . . see graph below.
.211 Record size: . • . . . . • 80 characters.
. 212 Key size: . . . . . . . . . 8 characters.
1,000
7
4
/
2
100.0
/
,
7
/
1/
4
~
Time in Minutes to 2
Put Records Into
Required Order
10.0
,
7
Y
V
/
1/
V
"
V
I
~
V
i/
/
4
/
/
2
/
1.0
)'
~
/
~
7
4
2
0.1
100
2
4
7
1,000
2
4
2
7
10,000
4
7
100,000
Number of Records
(Roman numerals denote standard System Configurations.)
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
423:201.300
~
3
.31
IBM SYSTEM/360-MODEL 30
MATRIX INVERSION
.312 Timing basis: . . . . . . using estimating procedure
outlined in Users' Guide,
4:200.312.
.
Standard Problem Estimates
.311 Basic parameters: . . . general, non-symmetric
matrices, using the optional Floating-Point Arithmetic feature; precision
is approximately 6 decimal
digits in the SHORT format
or 16 digits in the LONG
format.
.313 Graph: . • . . • . . . . . . see graph below.
100
L
7
I
I
1~
4
II
II
2
VI
10
7
II
4
J
V
2
Time in Minutes
for Complete
Inversion
IJ
IJ
1/
~~~
1) hO
0
1
I@
I
7
LI
J
4
I I
I IJ
2
I
/
0.1
I
7
I
11
I
4
I
I I
/
2
V
0.01
1
2
4
7
10
2
4
7
2
100
4
7
1,000
Size of Matrix
/
7/65
SYSTEM PERFORMANCE
(
423:201.400
.4
GENERALIZED MATHEMATICAL PROCESSING
.41
Standard Mathematical Problem A Estimates
.413 Timing basis: . . . . . . using estimating procedure
outlined in Users' Guide,
4:200.413.
.
. 411 Record sizes: .••..• 10 signed numbers; average
size 5 digits, maximum
size 8 digits.
.412 Computation: . . . . . . . 5 fifth-order polynomials,
5 divisions, and 1 square
root; computation is in
"long" floating-point mode
(16-digit precision).
.414 Graph: . . . . . . . . . . . see graph below .
Note: Use of the optional
Floating-Point Arithmetic
feature is assumed, although its cost is not included in the rentals for
Configurations I, II, and III.
1,000
,
,I
7
I
4
~
~"
V\:i
2
~
R=1.0, 0.1, 0.01
./
100
\)
.", \.:Jt
7
_ _ C1?;iiii'
~
C1?~
~
2
..........
?-
\)'
'\-"
\)'
~
1-
~"').
....... '?-.......Jj "'''').,7\).
4
\
.I
~
~.",
\).
V
Time in
10
Milliseconds
per Input Record 7
4
2
1
7
4
2
0.1
2
0.1
4
7
2
1.0
4
7
2
10.0
4
7
100.0
C, Number of Computations per Input Record
(R = number of output records per input record; the times shown here apply to
Configurations II and III; curves marked "CP" show central processor time.)
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360
MODEL 40
International Business Machines Corp.
i
~.
c·
(
AUERBACH INFO, INC ..
PRINTED IN U. S. A.
'
IBM SYSTEM/360
MODEL 40
International Business Machines Corp.
./
AUERBACH INFO, INC.
PRINTED IN U. S. A.
424:011.100
IBM System/360
Model 40
Introduction
INTRODUCTION
\
'.
The Model 40 series of the IBM System/360 family is a small-to-medium scale,
general-purpose computer system, with typical system rentals ranging from approximately
$7,000 (for a 4-tape system) to $11,000 (for a 65K, 6-tape system) .per month. It has approximately the same processing power as the older IBM 7080 system. The available Compatibility
Features indicate that Model 40 is regarded by IBM as a probable replacement computer for
the IBM 1401, 1410, 1460, and 7010 systems.
The Model 40 series of the IBM System/360 is characterized by:
•
A main core storage cycle time of 2.5 microseconds, with two
bytes being accessed per cycle. Storage accessing is not interleaved, so effective core cycle time is 1. 25 microseconds per byte.
•
Main core storage capacities ranging from 16,384 to 262,144 bytes.
•
Standard inclusion of the Multiplexor Channel, which allows simultaneous operation of a number of low-speed input-output devices
such as printers, card readers, and communication terminals.
This makes the Model 40 system suitable for use as an independent
system and/or as a satellite system to handle the input-output proceSSing of a larger system.
•
Standard inclusion of the following features:
System Control Panel
Standard Instruction Set
Interval Timer.
•
Optional availability of the following features:
Channel-to- Channel Adapter
Decimal Arithmetic
Direct Control
Floating-Point Arithmetic
1401/1460 Compatibility or
1410/7010 Compatibility(not both)
Selector Channels (maximum of 2)
Storage Protection
High-Speed Multiplexor Channel.
•
Shared Processor Storage
1440 Compatibility Feature
1620 Compatibility Feature
7070/7074 Compatibility Feature
7080 Compatibility Feature
709/7040/7044/7090/7094 Compatibility Feature.
\.
•
I
\,-
(
Non-availability of the follOwing features:
Non-availability of the following peripheral devices:
2150
2361
2301
2415
1231
1412
1418
1428
Console
Large-Capacity Core Storage
Drum Storage
Magnetic Tape Unit and Control
Optical Mark Page Reader
Magnetic Character Reader
Optical Character Reader
Alphameric Optical Reader.
This report concentrates upon the characteristics and performance of the Model 40
series in particular. All general characteristics of the System/360 hardware and software
are described in Computer System Report 420: IBM System/360 - General.
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
424:011.1 01
IBM SYSTEM/360-MODEL 40
The System Configuration section which follows shows the System/360 Model 40 in the
following standard System Configurations:
II
4-Tape Business System
III
6-Tape Business System
V
6-Tape Auxiliary Storage System
VI
6-Tape Business/Scientific 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.
As a matter of general interest, the rentals that would be incurred when either the 2321 Data
Cell Drive, the 2302 Disk Storage Unit, or the 2311 Disk Storage Drives are used to provide
the necessary random access storage capability are listed separately on the diagram for the
Auxiliary Storage System, Configuration V.
Section 424:051 provides detailed central processor timings for the Model 40. See
Section 420:051 for all the other characteristics of the program-compatible System/360 processors, including a discussion of the System/360 instruction repertoire as a programming tool.
Program compatibility with the IBM 1401 and 1410/7010 systems is discussed in Sections 420:131
and 420: 132, respectively.
The input-output channel capabilities of the System/360 Model 40 can be summarized
as follows (see also Sections 420:111 and 424:111):
Standard Multiplexor Channel
Maximum number: . . . . . . . . . . . . • . • . . .
Maximum data rate, bytes/sec Multiplex mode: . • • . . . • . . • . . . . . . • . •
Burst mode: •.•....••.•••..• ; . . . .
Maximum number of subchannels: . . • . . • . .
1.
31,000.
200,000.
128.
High-Speed Multiplexor Channel
Maximum number: . . . . . . . . . . . . . . . . • •
Maximum data rate, bytes/sec Multiplex mode: . . . . • . . . . . . • . . . . • .
Burst mode: • . . . . . . . . . . . . . . . . . . . .
Maximum number of subchannels: ••••••••
1.
200,000.
200,000.
4.
Selector Channels
Maximum number: • . . . • • . . . . . • . . • . . . 2. *
Maximum data rate, bytes/sec: .•••.•... 400,000.
The software that can be used with any System/360 configuration depends upon the core
storage capacity and the number and type of peripheral devices. Several versions of the Assembler,
COBOL, FORTRAN IV, and PL/Iwill be made available. These languages and the other support
routines for the System/360 are described in Sections 420:151 through 420:193.
The overall performance of any System/360 is heavily dependent upon the Processing
Unit model used. A full System Performance analysis of the Model 40 is provided in Section
424:201.
*If the High-Speed Multiplexor Channel is used, only one Selector Channel can be installed.
7/65
424:031.100
/
\
'~
IBM System/360
Model 40
System Configuration
SYSTEM CONFIGURATION
System configuration possibilities for Model 40 and other System/360 models are summarized
in report Section 420:031. This report section shows Model 40 systems arranged in accordance with
the specifications for our Standard Configurations, as defined in the Users' Guide, page 4:030.120 .
• J.
4-TAPE BUSINESS SYSTEM; CONFIGURATION II
Deviation from Standard
Configuration: ••••••••••••••••••••••••••.•.. core storage is 100% larger.
printer is 20% faster.
reader is 20% faster.
magnetic tape is 100% faster.
Equipment
Main Storage
(16,384 bytes)
$2,700
2040 Processing Unit, Model D40
(includes one Multiplexor Channel)
2501 Card Reader, Model B1:
Reads 600 cards per minute
260
1442 Card Punch, Model N2:
Punches 91 full' cards per minute
375
2821 Control Unit, Model 2
1403 Printer, Model 7:
Prints 600 lines per minute
600
650
2403 Magnetic Tape Unit and
Control
2402 Magnetic Tape Unit
2401 Magnetic Tape Unit
(4 drives total)
All Modell: 30,000 bytes/sec.
Optional Features Included: .•.••••••.....••..•.•. Decimal Arithmetic
Selector Channel
TOTAL:
\
1,840
115
350
$6,890
'-,
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360-MODEL 40
424:031.200
.2
6-TAPE BUSINESS SYSTEM; CONFIGURATIONm
Deviations from Standard Configuration: ••••••••••••• printer is 20% faster.
reader is 20% faster.
Equipment
Main Storage
(16,384 bytes)
2040 Processing Unit, Model D40
(includes one Multiplexor
Channel)
1052 Printer-Keyboard and Adapter
290
2501 Card Reader, Model B1:
Reads 600 cards per minute
260
1442 Card Punch, Model N2:
Punches 91 full cards per minute
375
2821 Control Unit, Model 2
1403 Printer, Model 7:
Prints 600 lines per minute
600
650
2403 Magnetic Tape Unit and
2402 Magnetic Tape Units (2)
2401 Magnetic Tape Unit
(6 drives total)
All Modell: 30,000
bytes/sec.
Optional Features Included: ••••••••••••••••••.••• Selector Channel
Decimal Arithmetic
TOTAL:
contrOl~
2 460
,
350
115
$7,800
(Contd. )
7/65
SYSTEM CONFIGURATION
.3
424:031.300
6-TAPE AUXILIARY STORAGE SYSTEM; CONFIGURATION V
Deviations from Standard Configuration: ..••••••••••• up to 380 million bytes more auxiliary
storage.
printer is 20% faster.
reader is 20% faster.
Equipment
Main Storage
(16,384 bytes)
$2,700
2040 Processing Unit, Model D40
(includes one Multiplexor
Channel)
1052 Printer-Keyboard and Adapter
290
2501 Card Reader, Model B1:
Reads 600 cards per minute
260
1442 Card Punch, Model N2:
Punches 91 full cards per minute
375
2821 Control Unit, Model 2
1403 Printer, Model 7:
Prints 600 lines per minute
600
650
2403 Magnetic Tape Unit and contrOll
2402 Magnetic Tape Units (2)
2401 Magnetic Tape Unit
2 460
(6 drives total)
,
All Modell: 30,000 bytes/sec.
2841 Storage Control and 8079
Attachment
2321 Data Cell Drive, Modell:
up to 400 million bytes storage;
access time: 175 to 600 msec.
Optional Features Included: ••••••.•••••..•.•••..• Selector Channel
Decimal Arithmetic
TOTAL:
NOTE:
3,500 ---350
115
$11,300
The following can be used in place of the 2321 Data Cell Drive,
resulting in the indicated total rentals:
2302 Disk Storage, Model 3, and 7950 Attachment - average
access time: 165 msec; 112.14 million bytes storage:
$14,175
2311 Disk Storage Drives (3) - average access time:
128 msec; 21. 75 million bytes storage: v..
$10,050
'.,-
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
424:031.400
.4
IBM SYSTEM/360-MODEL 40
6-TAPE BUSINESS/SCIENTIFIC SYSTEM; CONFIGURATION VI
Deviations from Standard Configuration: .••••••• ' ••••• printer is 20% faster.
reader is 20% faster.
direct-access device (2311) is
included to permit use of
Operating System/360 32K
FORTRAN IV.
Equipment
Rental
Main Storage (65,536 bytes)
2040 Processing Unit, Model F40
(includes one Multiplexor
Channel)
1052 Printer-Keyboard and Adapter
2841 storage Control and
2311 Disk Storage Drive
1,100
2501 Card Reader, Model B1:
Reads 600 cards per minute
260
1442 Card Punch, Model N2:
Punches 91 full cards per minute
375
2821 Control Unit, Model 2
1403 Printer, Model 7:
Prints 600 lines per minute
600
650
2403 Magnetic Tape
2402 Magnetic Tape
2401 Magnetic Tape
(6 drives total)
All Modell: 30,000
Unit and Control
Units (2)
Unit
2,460
bytes/sec.
Optional Features Included: •••••••••••••••••..••• Decimal Arithmetic
Floating Point Arithmetic
Selector Channel
TOTAL:
7/65
290
115
100
350
$11,100
424:051.100
IBM System/360
Model 40
Central Processor
\
\
CENTRAL PROCESSOR
.1
GENERAL
.11
Identity: •.•••••••. IBM 2040 Processing Unit.
.12
.4
.41
.411
.412
.413
.414
.415
.416
.417
.418
.42
Processor Performance in Microseconds
.421 For random addresses Fixed point Floating point
c
= a + b: .••••••• 36 (binary)
62 (long)
See Section 420:051 for a comprehensive description
45 + 6. 3B
43 (short)
of the characteristics of all the System/360 Proc(decimal)
essing Units.
b = a + b: .••••••. 36 (binary) 62 (long)
See Section 424:011 for a summary of the distin43 (short)
29 + 3. 8B
guishing features of the 2040 Processing Unit as
(decimal)
used in Model 40 systems.
Sum N items: .••••• 12N (binary) 28N (long)
19N (short
29 + 3. 8B
The Instruction Times and Processor Performance
(decimal)
times for Model 40 systems, in all four modes of
c = ab: .••••••••• 113 (binary) 294 (long)
arithmetic, are listed below. See Paragraphs
51. 2 + 30.8B 105 (short)
4:050.41 and 4:050.42 of the Users' Guide for the
+ 3.8B2
definitions of these standard measures of central
(decimal)
processor performance.
c = alb: ..••••••. 216 (binary) 511 (long)
PROCESSOR SPEEDS
46 + 67. 2B 157 (short)
+ 11.3B2
Instruction Times in Microseconds
(decimal)
where B = operand length
Fixed point Decimal (optional)
Binary
in eight-bit bytes.
.422 For arrays of data 29.35 + 3.75B
Add-subtract: ••••• 11.88
Fixed point Floating point
25.58 + 25.78B +
Multiply: .•••••.•. 84.72
3.75B2
ci = ai + bj: .•.•••. 63 (binary) 89 (long)
72 + 6.3B
70 (short)
Divide: ..••••.••. 176.88 to 20.63 + 62.2B +
(decimal)
196.88
11. 25B2
bj = ai + bj: ••••••• 63 (binary) 89 (long)
where B = operand length
56 + 3. 8B
70 (short)
in eight-bit bytes (2 decimal
(decimal)
digits per byte).
Sum N items: •••••• 28N (binary) 44N (long)
Floating point (optional) (46 + 3. 8B)N 35N (short)
Long
Short
(decimal)
Add-subtract: .•••• 27.66
18.66
80.63
Multiply: •.•.••••• 259. 38
c = c + aibf .••.••• 148 (binary) 349 (long)
132.5
Divide: .•••••.•.. 476.88
108 + 34. 6B 151 (short)
+ 3.8B2
Additional allowance for (decimal)
Single indexing: •••• 0.0
Double indexing: .••. 1. 25
.423 Branch based on comparison Indirect addressing: . none.
Numeric data: ..••• 61. 33N
Recomplementing: .• none.
Alphabetic data: ••.• 61.33N
Control:
.424 SwitchingCompare Unchecked: .•••••• 54.39
Checked: .•••••••• 91. 91
Fixed point: .••.•• 11.88
\
Decimal: .•••.••• 22.5 + 2.5B
List search: .••••• 54.40 + 37. 52N
.425 Format control, per character Floating point
t·
I
(long): .•••••.. 24.33
.op lona
Unpack: . . . . • • • . . 4.3
Floating point
Compose: •••••.•• 8.1; 11.8 with radix
(short): ..••.•. 16.38
conversions.
Logical: .••••• ' •. 15.68 + 2.81B
.426 Table look-up, per comparison Branch: ..••••.•• 9.38
For a match: .••••. 37.52
Counter control For least or greatest: 38.71
Step: ••••••..... 11. 88
For interpolation
Step and test: .•••.• 10.63 (increment of -1).
point: ••••.....• 37. 52
16.26 (increment of any
.427 Bit indicators value).
Set bit in separate
Test: .••••••.••• 11. 88
location: . . . . . . . • 9.38
Edit: •••••••••••• 21. 58 + 5. 08B, approxSet bit in pattern: ••• 9.38
imately (optional).
Test bit in separate
Convertlocation: ..••..•. 18.76
To binary: •••••••• 31.88 to 87. 81
Test bit in pattern: •. 18.76
To decimal: .•••••• 28.15 to 98.75
.428 Moving: .•.•••.••. 16.25 + 2. 5B, where B =
Shift: .••••••••••• variable.
number of bytes moved.
Description
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
/
424:111.100
\.
IBM System/360
Model 40
Simultaneous Operations
SIMULTANEOUS OPERATIONS
A System/360 Model 40 system can concurrently execute:
•
One machine instruction; and
•
Up to two input-output operations, one on each of the Selector Channels; and
•
Multiple additional input-output operations via the Multiplexor Channel.
Alternatively, up to four operations on a High Speed Multiplexor Channel can be overlapped with
internal processing and one Selector Channel operation.
When the Multiplexor Channel is operated in the burst mode, only one input-output operation
can proceed on that channel. Internal processing cannot be overlapped with burst mode operation, but operations previously initiated on the Selector Channels will proceed concurrently
with the burst mode operation.
The demand on the central processor (i. e., the "interference" or delay imposed upon the central processor program by each individual input-output operation) will vary depending on
whether the peripheral device is connected to one of the Selector Channels or to the Multiplexor
Channel. (See the general discussion of System/360 Simultaneous Operations in Section420:111.)
In Table I the processor demands imposed by each of the peripheral units are listed for both
types of channels. It is noteworthy that at press time, 15 months after announcement of the
System/360, mM still had not officially specified many of the pertinent figures regarding
channel data rate capacities and processor demands. Estimated figures have been inserted,
and marked accordingly, wherever the specifications are suffiCiently firm to allow reasonable
estimates to be made.
The specific characteristics of the Model 40 Selector, Multiplexor, and High Speed Multiplexor
Channels can be summarized as follows:
Selector Channels
Maximum number: . • . • . . . . . • • . . . . • . . . •
Maximum data rate per channel: . . . . . . . . . . .
Maximum data rate for all Selector
Channels Multiplexor channel not operating: ....•.
Multiplexor channel operating: . . . . . . . .
Processor demand: •.•••.••...•••.•...•
Number of control unit positions: .•.••..•.•
2.
400 kilobytes/sec.
600 kilobytes/sec.
90 kilobytes/sec.
see Table I.
8 per channel.
Multiplexor Channel
Maximum number: . . . . . . . . • • . . . . . . . . . .
Maximum data rate, Selector Channels
not operating Multiplexed mode: . . . . . . . . . • • . . . . .
Burst mode: . . . . • • • • • . . . . • . . . . . .
Processor demand: . . . . . . . . . . . . . . . . . . . .
Number of control unit positions: . . . • • . . . . .
Maximum number of subchannels: ......•.••
\,
'
I
\.
..
1.
31 kilobytes/sec.
200 kilobytes/sec.
see Table I.
8.
128.
High Speed Multiplexor Channel
Maximum number: . . . . . • • . . . . . . . • . . . . .
Maximum data rate Multiplexed mode: . • • • . . . . . . . . . . . .
Burst mode: . • • • . . • . • • . . . . . . . . . .
Number of control unit positions: ..••...•..
Maximum number of subchannels: ....•.•..•
©
1 (pre-empts one Selector Channel).
not specified by IBM to date.
not specified by IBM to date.
4.
4.
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
424: 111.10 1
IBMSYSTEM/360-MODEL 40
TABLE I:
INPUT-OUTPUT DEMANDS ON THE SYSTEM/360 MODEL 40 PROCESSOR
Device
Random Access
2302 Disk Storage
2311 Disk Drive
2321 Data Cell Drive
7320 Drum
2314 Direct Access Storage
Punched Card
2540 Card Read Punch:
Read, 1000 cpm
Punch, 300 cpm
1442 Model N1 Card Read Punch:
Read, 400 cpm
Punch, 91 cpm
1442 Model N2 Card Punch, 91 cpm
2520 Model B1 Card Read Punch:
Read, 500 cpm
Punch, 500 cpm
2520 Card Punch:
500 cpm (Model B2)
300 cpm (Model B3)
2501 Card Reader:
600 cpm (Model B1)
1, 000 cpm (Model B2)
Printers
1403:
132 columns, 600 lpm (Model 2)
120 columns, 600 lpm (Model 7)
132 columns, 1, 100 lpm (Models
3, N1)
1404, 132 columns, 600 lpm
1445, 113 columns, 190 lpm
Punched PaI1er Tape
2671 Paper Tape Reader, 1,000 cps
Magnetic TaI1e
2400 Series:
Modell, 30 KB/sec
Model 2, 60 KB/sec
Model 3, 90 KB/sec
7340 Hypertape:
170 KB/sec
340 KB/sec
Peak
Data Rate,
KB/sec*
156
156
54.7
135
312
156
156
54.7
135
312
Selector
Channel
20
20
7.1
18
41
Multiplexor·
Channel**
100
100
100
100
100
1.3
0.40
0.17
0.05
1. 65
0.80
0.53
0.12
0.12
0.53
O. 12
0.12
0.07
0.02
0.02
1. 65
0.37
0.37
0.67
?
0.67
0.67
0.09
0.09
2.10
0.92
?
?
0.67
0.40
0.09
0.05
0.92
0.60
0.80
1.3
0.80
1.3
0.10
0.17
2.50
4.20
70
70
1.3
1.2
0.17
0.16
0.90
0.60
70
70
48
2.4
1.3
0.36
0.31
0.17
0.06
1. 67
0.90
0.58
1
1
0.13
3.12
30
60
90
30
60
90
3.9
7.8
11.7
100
100
100
170
340
170
340
21. 8
43.6
100
70
70
*Kilobytes (thousands of bytes) per second.
"'*Estimated.
7/65
Average
Data Rate,
KB/sec*
Demand on Processor,
per cent, via-
-
424:201.001
i
i
\
IBM System/360
Model 40
System Performance
SYSTEM PERFORMANCE
GENERALIZED FILE PROCESSING (424:201.100)
These problems involve updating a master file from information in a detail file and
producing a printed record of each transacti0n. This application is one of the most common commercial data processing jobs and is fully described in Section 4:200.1 of the Users' Guide.
Standard File Problems A, B, and C vary the record sizes in the master file. Standard Problem
D increases the amount of computation performed upon each transaction. Each problem is estImated for activity factors (ratios of number of detail records to number of master records) of
zero to unity. In all cases a uniform distribution of activity is assumed.
The Generalized File Processing problem for the System/360 was coded in two ways one using master files in packed decimal format and computation in decimal arithmetic, and the
second using master files primarily in binary format and computation in fixed-point binary. The
decimal computations required more time than those in fixed-point binary; but in the binary case,
items in the detail and report files needed radix conversion. There were no appreciable differences in the total times for the two cases. The graphs for the file problem are based on use of
the fixed-point binary technique. For simplicity, the very similar curves based on the use of
decimal arithmetic are not shown.
In the master file record layout, alignment of data items in core storage was carefully
considered. Double-word boundaries were observed for input-output purposes to improve performance efficiency on the larger models. Instead of the "chain" mode (scatter-gather) of tape
reading and writing, individual records were moved to the work areas using a high-speed,
multiple transfer method.
In Configurations II, III, and VI, the master files are on magnetic tape. The detail
file is assigned to the card reader and the report file to the printer. The master file tapes are
connected to a Selector Channel, and tape reading or writing can fully overlap processing and
other input-output operations.
For all three configurations on all four of the standard File Problems, the master-file
tape times are the controlling factor at very low activities, and the printer is the controlling
factor at activities higher than about O. 1.
Because multiprogramming of two or more independent programs is a featured capability of
the System/360, the time actually used by the central processor (CP) is also plotted. By comparing the curves of total time for the various configurations with the central processor curves,
it can be seen that even in the worst case some 80% of the available processing capacity is not in
use. A comparison of the central processor curves for a standard amount of computation and for
trebled computation (i. e., the curves for Problems A and D, respectively) shows the effect of
increaSing the computational workload.
SORTING (424:201. 200)
The standard estimate for sorting 80-character records by straight-forward merging
on magnetic tape was developed from the time for Standard File Problem A by the method explained in Paragraph 4:200.213 of the Users' Guide. A two-way merge was used in System Configuration II (which has only four magnetic tape units) and a three-way merge in Configurations m
and VI. The results are shown in Graph 424:201.200.
MATRIX INVERSION (424:201. 3,00)
The standard estimate for inverting a non-symmetric, non-Singular matrix was computed by the simple method described in Paragraph 4:200.312 of the Users' Guide. In order to
execute this procedure in floating-pOint form, the optional Floating-Point Arithmetic feature must
be included on the Model 40. Two lines are shown on the graph, one using the short floating-point
format (6-digit precision) and the other using the long format (16-digit precision).
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360-MODEL 40
424:201.002
GENERALIZED MATHEMATICAL PROCESSING (424:201.400)
Standard Mathematical Problem A is an application in which there is one stream of input data, a fixed computation to be performed, and one stream of output results. Two variables
are introduced to demonstrate how the time for a job varies with different proportions of input,
computation, and output. The factor C is used to vary the amount of computation per input
record. The factor R is used to vary the ratio of input records to output records. The procedure
uSEjd for the Standard Mathematical Problem is fully described in Section 4:200.2 of the Users'
Guide.
The optional Floating-Point Arithmetic feature must be included on the Model 40 in
order to perform the floating-point computations required in this problem. Double-length operations (16-digit precision) were used since a minimum precision of 8 digits is prescribed. The
input file is on the card reader, and the output file is assigned to the printer.
Graph 424:201.400 shows two curves. The curve marked "R = 1.0" is for one output
record for every input record. The other curve is for one output record for every tenth (R = O. 1)
and every hundredth (R = 0.01) input record. (There is no effective difference between these two
cases). For all configurations, the central processor becomes the controlling factor when more
than 6 times the standard amount of computation is performed.
WORKSHEET DATA TABLE 1
CONFIGURATION
ITEM
1
InputOutput
Times
II
m
VI
1)
1,056
1,056
1,056
(File 1)
12
(File
Char/block
Reoords/block
K
msec/block
~=File2_
~--mSBe/switoh
---~--~-
Centrs!
Processor
Times
msee/block
_
_
---~_ _ _8._0_ _
~--- ---~-~File 4
2
~---
, _ _ 0.:2!,- _ _ _ _0_.5_1_ _
-~~--~2_ _
J!.6_ _ _. _ r- ___O~ _ _
4:200.1132
--~_
~3
_
_
I___
O~
_
~9 _ _. _
0.93
mBBe/report
b7 +b8
1.81
Printer
C P
Msec/block
for C.P.
and
aL- _ _ _
r--~
~--.- f- 6.:..LaSK
dominant I/O
..El!!!J. Msster..J!L
column.
~:..lI1.aste".tM
Total
Unit of Measure (bytes)
.l.L-
r--9~
144.0
293.5
I--'-- , - - , - , 6 (Blocks 24 to 48)
~.--.--,-
43.9
,--1_.3_ _
1---
1500.0
293.5
1500.0
. 293.5
6,000
1500.0
6,000
,---~-
I--
----wo - 15,616
15,616
Floating Point
Floating Point
--~-
15,616
Floating Point
~---.- ~ardReadeX:~~ ~Ol cardReade~ 2~ad~
output
1403 Printer, Mod 7
Input
1403 Printer, Mod 7
80 bytes
80 bytes , - f-
80 bytes
80bytes-
1403 Printer, Mod 7
- 8080-bytes
-bytes
Size of record
routput
--'--- -
msec/block
-~- I--~- 1--..l&L- ~--output T2
100
100
100
msec penalty
(estimated)
input
Ta
r-=-----output T4
~~--~
~ loops
_ _ T__ _._
msec/report
4:200.1151
1-~8--- 1--4,648-- - - 4 - ; 6 4 8 -
Fixed/Floating point
Unit name
4:200.114
~3_
----r -4,092
- - - - 1 - 4,092
Total
ProblemA
...-U.. _
- - _1_.3_ ; - - - - --L.A _ - - - - ~O_ r - - 96.0 - - - - - - - -- - 144.0 r1500.0
144.0
1500.0
1500.0
r::--:---1---,-- 100
Working
Mathematical
Printer
r - - - - - f - - - - - - - 1-------- - - - - 128
128
128
~--- r--- 6 4 8 - - - - - . - --~3 (Blocks 1 to 23)
648
Standard
File
Problem A
Space
standard
C.P.
- - - ~1_ ' - - -
6,000
Std. routines
Printer
C P
43.9
...L.L
r
--1:81-
1.81
-_.- ~9_ - - - ~- f - - -
439
File 4: Reports
5
-~-
mBee/record
msee/detall
mBBe/work
~:Details_
4
12.0
12.0
12.0
_ _ _ O~_ _ _ _0_.7_8_ _
3
standard
File
Problem A
F = 1.0
-~-
0
0
_ _ 1_.3_ _ _ _ _1_.3_
0
~=F1le2.
12
-~-
4:200.112
125
125
125
_ _ _0_ _ _ _ _ _ 0_ _ _ _ _ _
0 _
_ _ _0_ _. _ _ _ _ 0_ _ _ _ _0_ _
File 4
.E!k..l = File 2_
.!!k.3 _ _ _
File 4
mseo penalty
(estimated)
12
_ __
51._2_ _
_ _ _100_ _ _
HEFERENCE
6
T7
4:200.413
- - - - - r---~- -8.0
- - -1 - -8.0
8.0
8.0
8.0
4.02
I--~- I--~4.02
- -13.13
--t--~- f--~2.52
2.52
(Contd. )
7/65
424:201.100
SYSTEM PERFORMANCE
i
.1
GENERALIZED FILE PROCESSING
• 11
Standard File Problem A
• 111 Record sizes Master file: . . • . . . 108 data characters, packed
as 88 8-bit bytes.
Detail file: . . . . . . • 1 card.
Report file: ...••.. 1 line.
. 112 Computation: . . . . . . . standard, using fixed-point
binary or decimal
arithmetic •
.113 Timing basis: . . . . . . using estimating procedure
outlined in Users' Guide,
4:200.113 .
. 114 Graph: ..••..••.•• see graph below .
• 115 Storage space required Configuration II: ••• 15,616 bytes.
Configuration III: ..• 15,616 bytes.
Configuration VI: .•. 15,616 bytes.
1,000.0
7
4
2
100.0
7
4
Time in Minutes to
Process 10,000
Master File Records
2
~
11, 111,
10.0
7
~
/
4
2
--1
~
/'
.._____01>
/O~
",.,--
~--
--
1.0
,
7
/
4
2
.r
I
I
.'
0.1
0.0
0.1
0.33
1.0
Activity Factor
Average Number of Detail Records Per Master Record
(
\
(Roman numerals denote standard System Configurations; curve marked "CP"
shows central processor time.)
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
424:201.120
. 12
IBM SYSTEM/360-MODEL 40
Standard File Problem B
. 121 Record sizes Master file: . • . . . . 54 data characters; packed
as 44 8-bit bytes.
Detail file: . . . . . . . 1 card.
Report file: . . . . . . . 1 line.
.122 Computation: . . . . . . • standard, using fixed-point
binary or decimal
arithmetic •
. 123 Timing basis: . . • . . . using estimating procedure
outlined in Users' Guide,
4:200.12.
.124 Graph: . . . • . . . . . . . see graph below.
1,000.0
7
4
2
100.0
7
4
Time in Minutes to
Process 10, 000
Master File Records
2
~
11, 111,
10.0
7
/
4
V
2
1.0
7
L
-
4
_~Cl'
L
L
""
c~/
- --
V'
L
If
1
L
,I
2
0.1
L
.-
,/
V
0.0
0.1
0.33
1.0
Activity Factor
Average Number of Detail Records Per Master Record
(Roman numerals denote standard System Configurations; curve marked "CP"
shows central processor time. )
(Contd.)
7/65
SYSTEM PERFORMANCE
. 13
424:201.130
Standard File Problem C
.131 Record sizes Master file: . . . . . . 216 data characters, packed
as 176 8-bit bytes.
Detail file: . . . . . . . 1 card.
Report file: . . . . . . . 1 line.
.132 Computation: . . . . . . . standard, using fixed-point
binary or decimal
arithmetic .
. 133 Timing basis: . . . . . . using estimating procedure
outlined in Users' Guide,
4:200.13.
.134 Graph: . . . . . . • . . . . see graph below.
1. 000.0
1
4
2
100.0
1
4
Time in Minutes to
Process 10,000
Master File Records
2
lll.
~
11.
10.0
1
-
./'
./
4
/'
~C'P-
~--
2
'
, /
1.0
""Cl?
,
1
4
2
-
/
.'
II
0.1
0.1
0.0
0.33
1.0
Activity Factor
Average Number of Detail Records per Master Record
(Roman numerals denote standard System Configurations; curve marked "CP"
shows central processor time.)
(
'.
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360-MODEL 40
424:201.140
.14
Standard File Problem D
. 141 Record sizesMaster file: . . . . . . 108 data characters, packed
as 88 8-bit bytes.
Detail file: . . . . . . . 1 card.
Report file: .••.••. 1 line.
.142 Computation: .••.... trebled, using fixed-point
binary or decimal
arithmetic •
.143 Timing basis: . . . • . . using estimating procedure
outlined in Users' Guide,
4:200.14.
. 144 Graph: . . . • . . . • . . . see graph below.
1. 000. 0
7
4
2
100.0
7
4
Time in Minutes to
Process 10,000
Master File Records
2
~
\\, 1\.\,
10.0
7
./
/'
4
2
~
V
/'
~ ....
/C'?
~
--
.--C1?-
---
1.0
,
7
,/
4
2
I
V
0.1
0.0
0.1
0.33
1.0
Activity Factor
Average Number of Detail Records per Master Record
(Roman numerals denote standard System Configurations; curve marked "CP"
shows central processor time.)
(Contd.)
7/65
SYSTEM PERFORMANCE
424:201.200
.2
SORTING
.21
Standard Problem Estimates
.213 Timing basis: . . . . . . using estimating procedure
outlined in Users' Guide,
4:200.213 (2-way tape
merge in Configuration II;
3-way tape merge in
Configurations III and VI) .
. 214 Graph: . . . . . . . . . . . see graph below.
.211 Record size: . • . • . . . 80 characters.
.212 Key size: . . . • . . . . . 8 characters.
1,000
7
4
2
II
100
J
7
1/
/
4
V
VV
2
Time in Minutes to
Put Records Into
Required Order
V
~~
~'
10
7
17
/
4
~
V.I
2
0
1
\I'
V
/
7
4
2
0.1
\
100
2
4
7
2
4
7
1,000
2
4
10,000
7
100,000
Number of Records
(Roman numerals denote standard System Configuration.)
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
424:201.300
IBM SYSTEM/360-MODEL 40
.3
MATRIX INVERSION
.31
Standard Problem Estimates
.312 Timing basis: .••••. using estimating procedure
outlined in Users' Guide,
4:200.312 .
• 311 Basic parameters: .•. general, non-symmetric
matrices, using the optional Floating- Point
Arithmetic feature; precision is approximately 6
decimal digits in the
SHORT format or 16 digits
in the LONG format.
.313 Graph: ..••..•.•.• see graph below.
10.0
7
II 1
L
1
4
J
~
2
~
1.0
V
IJ
7
I
II II
4
II II
j J
2
Time in Minutes
for Complete
0.1
Inversion
I
7
IJ
I
I
;'4
4
J/
2
0.01
L
7
I
I
I
4
2
0.001
1
2
4
7
10
2
4
7
100
2
4
7
1,000
Size of Matrix
(Contd.)
7/65
SYSTEM PERFORMANCE
424:201.400
.4
GENERALIZED MATHEMATICAL PROCESSING
. 41
Standard Mathematical Problem A Estimates
.413 Timing basis: . . . . . . using estimating procedure
outlined in Users' Guide,
4:200.413 .
.414 Graph: . . . . . . . . . • . see graph below .
• 411 Record sizes: . . . . • . 10 signed numbers; average
size 5 digits, maximum
size 8 digits .
. 412 Computation: . . . . . . . 5 fifth-order polynomials,
5 divisions, and 1 square
root; computation is in
"long" floating-point mode
(16-digit precision).
Note: Use of the optional
Floating- Point Arithmetic
feature is assumed, although its cost is not included in the indicated
rentals for Configurations
II and III.
1,000
"
7
I
~
4
I:)
~
2
~
V
~
1:)'
I:)~'
01:)'
~
100.0
R
;~
= 0.01, 0.1, 1.0
IL'
7
;;;!l.,: 1. 0
2
"
,,-. / .,...
4
l.-
- , 01-
= .2l,:..;':...-..-B.
I.........
... ""
~
"-c~ /
c-c.
__
~_
4,092 _ _ _ _4,~ _ _
4.648 _ _ _ _ _
__
~_
~~lto2:3~_~
~~24to_~_.
~
_____
orking
____
ToULl
~
4,648
100
15616
_
100
15,616
Floating point
input
~
4,~
100
Fixed/Floating point
UnIt of Measure
~~ru:~Ucal
25.6
--10,'il
~'----
c.P.
Problem A
1,056
~. FU,:~
~.----
macc/report
St."",,,.
Ffl.
VIm
(unblocked)
-----wO--I----24--:D* --'9:4
~~Flle~ _ _ _ ~ _ _ _ ~ _ _ _ ~_
msecpenalty
(estimated)
Central
(blocked)
1,056
FUe 2
FUe4 . - - - - --""125- - - --,,--- Input-
VIIB
VIlA
15,616
Floating point
20,280
Floatlngpoint
Size of record
St ,---- - :~ ::: --
msec/block
~~ _ _ - - ; _ _ _
msecpenalty
(estimated)
Input
:~:::
-- -
:::~:
15,616
FloaUng point
2501CardReader, ModB12540 CardReadPunch 2501CardReader,ModBl
,--. -140'3" Print.sr, Mod 7 1403 Printer, Mod 3""""" ""1403 Printor~
-
2400 Serles Tape, Mod 2
~es Tape. :M:0d2
--To:~:
.-
*- __ + ___ *_
Prohl~A r------E~~~--t-~~~--+-~~~--1-~~~--+--~~.--~
meee/record
~5100p'
msecJreport
*
T3
~~
-=-
TS
T;-=
'l'7
3.2
-
---a:z -
== 2.82
1.82
1.05
3.2
-- ~ -
._~
1.62
'.82
1.05
.==
3.2
-a:z -
4:200.413
0.04
---""ii":Ci4-
1.82
1.82
1.05
,1.011
2.82.==
2.~~
FUes ~ and 4 blocked 12 reeords/block.
./
(Contd. )
7/65
SYSTEM PERFORMANCE
425:201.1 00
.1
GENERALIZED FILE PROCESSING
• 11
Standard File Problem A
• 111 Record sizes Master file: ••.... 108 data characters, packed
as 88 8-bit bytes.
Detail file: .•..••• 1 card.
Report file: . . . . . . . 1 line.
. 112 Computation: . . . . . . . standard, using fixed-point
binary or decimal arithmetic .
.113 Timing basis: . . . . . . using estimating procedure
outlined in Users I Guide,
4:200.113.
.114 Graph: , ••.•....•• see graph below.
ON-LINE CARD READER AND PRINTER
4
2
111,
L;"
7
~
./ ./
4
2
."
1.00"'"
./
Time in Minutes to
Process 10,000
Master File Records
V/
Im,vy V
~.
I
t
4
...",
I
C~
I"'"
2
I
....
----
CP-
---
/"
V
/
0.1
7
-
________
//
1.0 - ~~;./
1-.::7
-
~~
~
10.0
I
4
2
0.01
0.0
0.1
0.33
1.0
Activity Factor
Average Number of Detail Records Per Master Record
(Roman numerals denote standard System Configurations;
curve marked "CP" shows central processor time.)
(
\
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
425:201.116
• 11
IBM SYSTEM/360-MODEL 50
Standard File Problem A (Contd. )
. 115 Storage space required
Configuration m: ...
Configuration IV: ••.
Configuration VI: •..
Configuration VIlA: ..
Configuration VIm .
(blocked): •.•••.•
Configuration VIIB
(unblocked): •••..
.116 Graph: .•••..•...•
15,616 bytes.
15,616 bytes.
15,616 bytes.
15,616 bytes .
20,280 bytes.
15,616 bytes.
see graph below.
OFF-LINE CARD READER AND PRINTER
100.0
7
4
2
10.0
7
4
~D)
Time in Minutes to
2
Process 10, 000
Master File Records
mi13LOC
~
1.0
7
V
4
./
~2
--
(13LOCKE~_
------~
VUB
-
_____ C P -
~C~'
~--
0.1
~
7
1/
-'
4
2
0.01
0.0
0.1
0.33
1.0
Activity Factor
Average Number of Detail Records Per Master Record
(Roman numerals denote standard System Configurations;
curve marked "CP" shows central processor time.)
(Contd. )
7/65
SYSTEM PERFORMANCE
425:201.120
• 12 Standard File Problem B
.121 Record sizesMaster file: . . . . . . 54 data characters, packed
as 44 8-bit bytes.
Detail file: ..•..•• 1 card.
Report file: • . . . . . . 1 line.
. 122 Computation: . . . • • . . standard, using fixed-point
binary or decimal arithmetic.
.123 Timing basis: •..••. using estimating procedure
outlined in Users' Guide,
4:200.12.
.124 Graph: ••••...... see graph below.
ON-LINE CARD READER AND PRINTER
100.0
7
4
2
______
~
1.'J
ln., \fl,
10.0
7
~
L
4
Time in Minutes to
Process 10,000
Master File Records
:",...0"
./'
//
m'IVI
2
'iJ
1.0
7
4
-'
.-
//
/
____ CF-
....
-~IA
-
-- -
.. '
-
--
-~
//
C-q.
l/
2
,I
0.1
7
I
I
J
4
2
0.01
0.1
0.0
0.33
1.0
Activity Factor
Average Number of Detail Records Per Master Record
(Roman numerals denote standard System Configurations;
curve marked "CP" shows central processor time.)
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
7/65
IBM SYSTEM/360-MODEL 50
425:201.125
.12
Standard File Problem B (Contd.)
.125 Graph: •••••.•..•• see graph below.
OFF-LINE CARD READER AND PRINTER
100.0
7
/
4
2
10.0
7
4
Time in Minutes to
Process 10,000
Master File Records
2
~
"\11113 ~
1.0
7
."
/'"
4
2
~~
L-- ...
,---
-. -- -
--
"\11113 {13 LOCKED) - -
....
~,-
.---CP-
-
-
--
"..C""
0.1
~
7
./
4
/'
/
2
0'.01
0.0
0.1
0.33
1.0
Activity Factor
Average Number of Detail Records Per Master Record
(Roman numerals denote standard System Configurations;
curve marked "CP" shows central processor time.)
,/
(Contd.)
7/65
SYSTEM PERFORMANCE
• 13
425:201.130
Standard File Problem C
. 131 Record sizes Master file: . . . . . . 216 data characters, packed
as 176 8-bit bytes.
Detail file: . . . . . . . 1 card.
Report file: .••...• 1 line.
.132 Computation: . . . • . . . standard, using fixed-point
binary or decimal
arithmetic .
.133 Timing basis: . . • . . . using estimating procedure
outlined in Users' Guide,
4:200.13 •
• 134 Graph: . . . . . . . . . . . see graph below.
ON-LINE CARD READER AND PRINTER
100.0
7
4
2
~
111, "\11,
10.0
7
./
//
4
!-III
Time in Minutes to
2
Process 10,000
Master File Records
1.0
7
.-
--- -
....,
./
/C:_
VIIIB
4
Time in Minutes to
Process 10,000
Master File Records
2
---- ----::.=
-------------\TUB
1.0
VIIIB (HYPERTAPE)
0.1
7
4
.;C~"
2
.'
0.01
.
----- -------------
~-
~
....-
__ Cl'
-
-
/'
I
7
I
1/
4
2
0.001
0.0
0.1
0.33
1.0
Activity Factor
Average Number of Detail Records Per Master Record
(
"'.
(Roman numerals denote standard System Configurations)
LEGEND
_ _ _ _ _ _ _ _ _ _ _ Elapsed time; unblocked Files 3 & 4
- - Elapsed time; blocked· Files 3 & 4
_
C P - Central Processor time (all configurations)
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
426:201.140
.14
IBM SYSTEM/360-MODEL 65
Standard File Problem D
.141 Record sizesMaster file: . . . . . . . 108 data characters, packed
as 88 8-bit bytes.
Detail file: . . . . . . . 1 card.
Report file: . . . . . . . 1 line.
.142 Computation: . . . . . . . trebled, using fixed-point
binary or decimal arithmetic.
. 143 Timing basis: . . . . . . using estimating procedure
outlined in Users I Guide,
4:200.14.
.144 Graph: . . . . . . . . . . . see graph below.
10.0
7
4
2
1.0
7
~
4
Time in Minutes to
Process 10,000
Master File Records
~
./
\T11l"S
--.;.... -L .__/
VIIm
\TUB
~-
--
-
~
,
r---------~
2
VIIIB (HYPERTAPE)
0,1
7
~CV
...
,...-
-
--
,.".-
4
~C~",
2
~
V
0.01
I
7
I
I
4
II
2
0.001
0.0
0.1
0.33
1.0
Activity Factor
Average Number of Detail Records Per Master Record
(Roman numerals denote standard System Configurations)
LEGEND
_ _ _ _ _ _ _ _ _ _ _ Elapsed time; unblOCKed Files 3 & 4
- - Elapsed time; blocked Files 3 & 4
C P - Central Processor time (all configurations)
(Contd.)
8/65
SYSTEM PERFORMANCE
426:20J .200
.2
SORTING
. 21
Standard Problem Estimates
. 212 Key size: . . . • • . . . . 8 characters .
.213 Timing basis: .••.•• using estimating procedure
outlined in Users r Guide,
4:200.213 .
. 214 Graph: . . . • . . . . . . . see graph below .
• 211 Record size: . . . . • . . 80 characters.
100.0
7
4
~
2
/
10.0
7
II~
I
,/ ,/
',1
~~£
4
4'4.$?
2
~
Time in Minutes to
Put Records Into
Required Order 1. 0
//
~
,
IJ
",-W
~~
7
~
)1III ~~
~
L
L
7
/
L
L
/
I
11'/
4
/ V
1//
V
II
L
VV
2
/
0.1
,/
,/
7
"
4
2
0.01
100
2
4
7
2
1,000
4
7
2
10,000
4
7
100,000
Number of Records
(Roman numerals denote standard System Configurations)
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
IBM SYSTEM/360-MODEL 65
426:201.220
. 22
Operating System/360 Sort/Merge Times
.223 Timing basis: .••.•• IBM Form C28-6543;
input-output blocking
factors as indicated in
legend below •
.224 Graph: . . . . . . . . . . . see graph below.
.221 Record size: . . • . . . . 80 characters.
. 222 Key size: . . . • . . ... 8 characters.
100.0
7
4
2
10.0
7
,J
4
v/
~.
£
2
II
1~~
V ~/
Time in Minutes to
Put Records Into
Required Order 1. 0
~
7
~
~
~ I'-"
4
$
~
.--
2
0.1
7
4
2
0.01
100
2
4
7
1,000
2
4
7
10,000
2
4
7
100,000
Number of Records
(Roman numerals denote standard System Configurations)
Curve
A
B
C
D
E
Configuration
Standard Configuration VIlB
Standard Configuration VIIIB
with Hypertape
Model 65 with one 1302 Disk
Storage Unit
Model 65 with three 2311
Disk Storage Drives
Model 65 with one 2301 Drum
I/o Blocking
12 records/block
12 records/block
40 records/block
30 records/block
120 records/block
(Contd.)
8/65
426:201.300
SYSTEM PERFORMANCE
\
.3
MATRIX INVERSION
. 31
Standard Problem Estimates
.312 Timing basis: . . . . . . using estimating procedure
outlined in Users' Guide,
4:200.312 •
.313 Graph: . . . . . . . . . . . see graph below •
. 311 Basic parameters: ... general, non-symmetric
matrices, using floating
point to a precision of
approximately 6 decimal
digits in the SHORT format and 16 digits in the
LONG format.
1.0
,,
7
/I
4
II
II
r
2
0.1
7
JII
If,
4
N
'l
2
l~;
if
Time in Minutes
for Complete
Inversion
0.01
"f
7
"
'I
JI
4
II
II
2
~
""
JL
0.001
7
I
\
"
11
4
/I
Y
2
(
r
0.0001
1
2
4
7
10
2
4
7
100
2
4
7
1,000
Size of Matrix
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
IBM SYSTEM/360-MODEL 65
426:201.400
.4
GENERALIZED MATHEMATICAL PROCESSING
. 41
Standard Mathematical Problem A Estimates
.412 Computation: •.•.•.. 5 fifth-order polynomials,
5 divisions, and 1 square
root; computation is in
"long" floating-point mode
(16-digit precision) .
.413 Timing basis: . . . . . . using estimating procedure
outlined in Users I Guide,
4:200.413 •
• 414 Graph: • . . . . . . . • . . see graph below.
. 411 Record sizes: . . . . . . 10 signed numbers; average
size 5 digits, maximum
size 8 digits.
1,000.0
7
4
2
100.0
7
4
~
2
A
VIIB
Time in
10.0
Milliseconds
per Input Record 7
VIIIB
~
'I
V
If
4
~~
"
c"9-~ /~
2
CP- f-
1.0
~'"
CP-
7
fi~;§
. . nV-"'~~
~.\.'
~.
~\.
4
2
0.1
0.1
2
4
7
2
1.0
4
7
10.0
2
C, Number of Computations per Input Record
(Roman numerals denote Standard Configurations;
R = number of output records per input record;
Curve marked "CP" shows Central Processor time.)
8/65
4
7
100.0
IBM SYSTEM / 360
MODEL 67
International Business Machines Corp.
(
AUERBACH INFO, INC.
PRINTED IN U. S. A.
IBM SYSTEM /360
MODEL 67
International Business.Machines 'Corp.
AUERBACH INFO, INC.
PRINTED IN U. S. A.
427 :011. 100
IBM System/360
Model 67
Introduction
INTRODUCTION
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 virtually instantaneous 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 A.pril 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 a:vailable only through special proposals and individual negotiations. In August 1965, Model 67 became a member of the standard IBM product line. Typical
monthly rentals will range from $45,000 to $60, 000 for a single-processor Model 67 system,
and from $100,000 to $200,000 for a multi-processor system. Initial customer deliveries of
Model 67 systems are scheduled for the second quarter of 1966.
The supporting software for time-shared operation of Model 67 systems has not yet
been documented, and this documentation is not scheduled to be available before Spring, 1966just before the first Model 67 systems are due to be delivered. Most of the software programs
themselves are scheduled for release during the Summer of 1967.
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 that speeds translations between virtual and physical addresses.
•
A main core storage cycle time of O. 75 microsecond, with eight
bytes being accessed per cycle; storage accessing is two-way interleaved.
.
•
Main core storage capacities ranging from 262, 144 to 2,097,152 bytes,
in independent modules of 262, 144 bytes.
•
Up to four 2067 Processing Units and up to four 2846 Channel
Controllers per system.
.
•
Up to seven Multiplexor or Selector Channels per Channel Controller; each channel can be controlled by any Processing Unit and
can transfer data to or from any core storage module.
•
Ability to connect virtually all of the standard System/360 peripheral
eqUipment and a variety of non-IBM terminal devices.
•
Ability to "partition" the system, by setting manual switches on the
2167 Configuration Console, to make certain components "unavailable"
for use by certain other components.
•
Availability of Compatibility Features that permit emulation of IBM
7070/7074, 7080, or 709/7040/7044/7090/7094 systems.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
B/65
427:011.1 01
IBM SYSTEM/360-MODEL 67
The standard complement of System/360 software will not be usable for time-sharing
operations in Model 67 systems, so specialized software support will be provided. The TimeSharing Operating System is being designed to include all the generalized software support necessary for most time-sharing scientific installations. A key component is the Time-Sharing
Monitor, which will control the execution of all jobs entering a Model 67 system and the hardware
environment in which they operate ... Programming languages available to Model 67 users include
FORTRAN IV, PL/I, a symbolic Assembler, and a Terminal Command Language to facilitate
communication between the system and users at remote terminals. No time-sharing COBOL
compiler has been announced, but a modified form of the Operating System/360 COBOL compiler will be made available for use as a batch-processing compiler for "background" operations.
All t.'1e other Model 67 compilers will produce re-entrant coding and will permit conversationalmode editing and syntax checking during source-program input.
Potential Model 67 users have also been informed that work is proceeding to provide a
Desk Calculator Language for "real-time" (interpretive-mode) computation and a GeneralPurpose System Simulator (GPSS) compiler; no general announcement of either of these programs
has been made by IBM to date.
.
The performance of Model 67 systems is difficult to predict accurately. Central
processor execution times for our standard measures of processor performance are·listed in
Section 427:051 of this subreport. These times are somewhat slower 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 ex~cution in Model 67. Other factors that
will tend to degrade the performance of a Model 67 system, and estimates of their quantitative
effects, are shown in Section 427:201, System Performance. It can safely be concluded that 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.
8/65
427:031.001
IBM System/360
Model 67
System Configuration
SYSTEM CONFIGURATION
A Model 67 system can include from one to four 2067 Processing Units. Systems with a single
Processing Unit and without switching capabilities use a 2067-1 Processing Unit, up to four 262Kbyte 2365-2 Processor Storage Units, and at least one 2860 Selector Channel or 2870 Multiplexor
Channel; a maximum of one Multiplexor Channel and six Selector Channels can be connected.
Model 67 systems with one to four Processing Units and with switching capabilities use 2067-2
Processing Units, up to eight 262K-byte 2365-12 Processor Storage Units, at least one 2846
Channel Controller, and at least one 2860 Selector Channel or 2870 Multiplexor Channel. A
maximum of four 2846 Channel Controllers can be used, and each 2846 can control up to seven
channels in any combination of Selector and Multiplexor Channels. Systems with more than one
Processing Unit require a 2167 Configuration Console.
The system configurations shown on the next two pages are considered typical of the Model 67
systems being offered by mM. Below each configuration diagram is a list of the components and
their prices. (Because of the highly specialized desigu objectives of the Model 67, the two representative configurations shown here do not match the specifications for any of our Standard System Configurations, as defined in Section 4:030 of the Users r Guide.)
Two special pieces of equipment are included in most Model 67 systems: the 2167-2 Configuration Console and the 2846-2 Channel Controller. These units were developed specifically for
the Model 67, and they are briefly described below.
Model 2167 Configuration Console
The 2167 is a switching unit that enables a Model 67 system to be manually reconfigured by the
operator. Each input-output controller, each core storage unit, and each processor can be
interconnected as required, uSing the switches on the console. If ciesired, two or more independent computer systems can be established by "partitioning" the hardware; this is done by
manually setting the switches so as to make certain components of the system "unavailable" for
use by certain other components.
Model 2846 Channel Controller
The mM 2846 Channel Controller permits interconnection of multiple I/O channels to multiple
processors and multiple core storage units. A 2846 permits concurrent operations on all channels attached to it. Up to seven I/o channels, four processors, and eight core storage units
can be interconnected by one controller. Channel Controllers provide operational control signals to start and terminate I/O operations, perform the necessary addressing, establish priorities, handle interrupts and perform diagnostic and maintenance functions.
To control interrupt conditions, the 2846 accepts mask bits from each processor to indicate
whether that processor will accept interrupt signals from each of the attached channels. When
a channel signals an interrupt, the Channel Controller directs the interrupt to the processor(s)
presently enabled for interruption.
Each 2846 Channel Controller has its own unique interface with each of the connected core storage units. The maximum total data rate of each controller is 6,400,000 bytes per second.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
427:031.100
.1
IBM SYSTEM/360-MODEL 67
TYPICAL SINGLE-PROCESSOR MODEL 67 SYSTEM
2860-2
SELECTOR
CHANNEL
2311
DISK
Quantity
Type
1
1
1
2067
1
1
1
2365
Oeserl tlon
Model or S/F
Unit Rental
$17,075
225
50
1
7920
B070
Processing Unit
2
7123
8035
Processor Storage
Seven Bit Storage Protection
2067 Attachment
9,250
1052 Adapter
2870 Attachment
II/C
N/C
1
2860
2
Selector Channel
3,000
1
1
2870
1
6990
Multiplexor Channel
first Selector Sub-channel
2,200
400
1
2820
1
Storage Control
2,300
1
2301
1
Drum Storage
2,000
1
1
2841
1
6118
Storage Control
Record OverHow
525
10
2
2311
1
Disk Storage Drive
575
2
1316
1
Disk Pack
1
2803
1
Tape Cortrol
650
920
15
2
2402
2
Magnetic Tape Units
1
1052
7
Printer .. Keyboard
1
1
2821
1
8637
Control Unit
Universal Character Set Adapter
970
15
b5
1
2540
1
Card Read Punch
660
1
1
1
1403
2
4740
8641
Printer
Interchangeable Chain Cartridge Adapter
Universal Character Set. Feature
775
75
10
1
1
1
I>
2702
1
4615
8055
3233
IBM Tennlnal
Type' (19684)
2741 Break
Data Set Line Adapters
6
I>
2741
1
4708
eo_,
Transmission Cortrol
:;:=catlons Tennlnal
850
35
10
20
80
2.50
Total Syste~ Rental: $44, 785
(Contd.)
8/65
SYSTEM CONFIGURATION
.2
427:031.200
TYPICAL MULTI-PROCESSOR MODEL 67 SYSTEM
Quantit
2
2
2
2
2
2
Type
2067
2
3274
7920
5495
1102
5494
3
3
1
3
6
1
2365
2
1
Descrjr:ation
Model or S/F
Unit Rental
$ 17,275
Direct Control
250
1052 Adapter
225
Partitioning Sensing
85
Additional 2846 Attachment
140
Partitioning Logic & Extended Direct
100
Control
Processing Unit
12
7123
3846
8036
8088
5518
2067 Switching Feature
2846 Switching Feature
Power Sequencing
1052
7
Printer - Keyboard
2167
2
5496
5497
Configuralian Control
Magnetic Tape Units
5
3615
8037
8038
8100
6148
Control Unit
1l00LPM Printer Adapter
Universal Character Set Adapter
Universal Character Set Adapter
Two Channel Switch
Remote Switch Attachment
920
1,570
75
15
15
200
2540
1
Card Read Punch
660
65
4
4
1403
N1
8640
Printer
Universal Character Set Feature
900
10
4
1416
1
Interchangeable Train Cartridge
100
1
12
1
1
1
2848
1
3355
3857
3858
7927
Display Control
Display Adapter
Expansion Unit
Expansion Unit
1053 Adapter
360
40
.45
55
40
1
1
2814
3
6148
Switching Unit
Remote Switch Attachment
24
24
2260
1
4766
Display Station
Alpha'merie Keyboard
1
1053
1
Printer
1
2702
1
1
4615
12
1
1
1
3233
BOSS
8110
6148
Transmission Control
IBM Terminal Control TYPe I
1t9684)
Data Set line Adapter
2741 Break
Two Processor Switch
Remote $.witch Attachment
1
4708
Communication Terminal
Interrupt
1,250
25
25
Partitioning Sensing
Partitioning Sensing
Channel Controller
1,950
Selector Channel
3,900
Address Prefixing
2
2
2
2870
1
1
1
1
1
1
1
2250
1
1
2814
2
2
2
2820
2
2301
1
2314
1
8170
6148
2
2
2821
. 650
100
Nle
2
1
24
2402
2
4
2
S
2
3
1095
40
2,200
MuttiplelCor Channel
Selector Subchannel
400
115
Address Prefixing
1
1245
1498
1880
4785
5485
5486
Second Operator Control Panel
1
6148
~::~1w~~~ Attachment
N/e
1
8170
6148
Storage Control
2,300
,
3
Unit Rental
9,425
2860
3
Tape Control
Two Channel Switch
Remote Switch Attachment
N/e
N/e
100
100
N/e
2846
3
1
8100
6148
Description
2803
Processor Storage
Expansion Feature
2
,
Model or S/F
1
1
1
7-61t Storage Protection
2
6
1
6990
1095
Quantit" T"pe
Display Unit
Alphameric Keyboard
4096 Position Buffer
Charac:lerGtrlerator
light Pen
First Operator Control Panel
700
50
250
300
75
60
50
125
~~:aSr:~l~::~~~hment
Io",m 5......
2,000
~!reet Access Storage Facility
5,250
~~o Channel Switch
emote Switch Attachment
hi'" Pa.'
100
NlC
2
2
12
12
2741
Nle
200
N/e
30
20
50
850
35
20
10
75
N/C
80
2.50
140
NlC
20
Total System Rental: $128,585
©
1965 AUERBACH Corporation .and AUERBACH Info, Inc.
8/65
~
427:051.100
~
IBM System/360
Model 67
Central Processor
CENTRAL PROCESSOR
.1
GENERAL
• 11
Identity: •••••••••• IBM 2067 Processing Unit.
.12
Description
The Model 2067 Processing Unit is essentially a
2065 Processing Unit, as used in Model 65 systems,
with modifications to improve its effectiveness in a
time-shared operational environment. Details of
these modifications and performance data on the
2067 are included in this section. (See Section
420: 051 for a comprehensive description of the characteristics of all standard System/3BO processing
units; these characteristics also apply to the Model
2067 except where differences are specifically noted
in this section.)
The modifications to the normal System/3BO Processing Units which are incorporated in the 20B7 are:
a new method of address interpretation, which
allows virtual addresses to be used; a new set of
registers, which facilitates the control of timeshared operations; an extended version of the storage protection scheme; new instructions to control
these added facilities; and a High-Resolution Interval
Timer. A detailed description of each of these
modifications follows.
.
.121 Address Interpretation
In the standard System/3BO processor,each address
contains 24 bits and refers to one specifiC byte in the
core storage units connected to the processor. Approximately IB million bytes are therefore directly
addressable, but only those addresses which correspond to byte positions which are actually installed
can be used during program execution.
In Model 67, the same 24 bits are used to specify an
address; however, the 12 most significant bits now
specify the virtual address of a 4, 09B-byte "page."
The same maximum number of bytes can be addressed, but in Model 67 any of the addresses can be
used, regardless of the amount of core storage actually included in the installation. In fact, there is
no direct connection between the addresses used in
the Model 2067 during program execution and the
physical position of the "pages" within core storage.
This addressing characteristic of Model 67 permits
dynamic relocation of programs by the Time-Sharing
Monitor (Section 427 :191) without requiring any modification of the programs themselves.
The only exceptions to this separation of actual and
program addresses apply to core storage addresses
o through 4095. These addresses are reserved for
specific processors to allow various diagnostic and
other system requirements to be met. All references to these addresses must be accompanied by a
prefix identifying the processor concerned.
The technique used in Model 67 to perform the
translations from the virtual addresses used in the
program to the physical addresses of the data required at any particular instant includes the use of
an "associative memory." The information stored
in an associative memory is addressable by its contents as well as by its location. The Model 67's
associative memory contains the virtual and phySical
addresses of eight of the 4, 096-byte pages which are
being used by the currently-operating program.
When a virtual address in any of these eight pages
is encountered during execution of the stored program, the corresponding physical address can be
quickly read out of the associative memory. Since
associative memories are, in general, much faster
than table look-up operations, this is a very fast
way to perform the necessary translations between
virtual and physical addresses; the translation process takes only 150 nanoseconds in Model 67.
When the associative memory does not currently
contain the address of the required page, microprogrammed operations are used to obtain the appropriate physical address through the use of two
tables held in main core storage. This naturally
involves two core storage accesses, which delays
computation by another 1. 9 microseconds per address. A special register is used to hold the present physical address of the instruction counter, so
the additional 1. 9 microseconds will rarely be required for instruction accessing.
The associative memory's contents are updated each
time an unsuccessful attempt has been made to find
a page address in it. One of the presently-stored
addresses is dropped out and replaced by the address
of the new page which has just been requested. The
selection as to which of the eight addresses in the
associative memory shall be overwritten in this
manner is made automatically, based on a randomizing technique. Figure 1 is a flow chart summarizing the overall technique for address interpretation
in Model 67.
From the system programmer's point of view, the
addressing structure of the Model 67 is completely
different from the addressing structure of the other
System/360 models. In Model 67, the programmer,
like the hardware, uses "pages" of 4,096 bytes. But
unlike the hardware, which regards each page as a
completely separate entity, the programmer groups
his pages into "segments." A segment consists of
up to 256 pages. Each page is loaded as a single unit
during program execution.
.122 Additional Registers
Additional registers included in the Model 2067
Processing Unit are:
•
Eight 26- or 32-bit associative registers that are
used for address interpretation. The virtual and
(Contd.)
8/65
/'
427:051.122
CENTRAL PROCESSOR
I
~OGICA~
ADDRESS
I I
SEGMENT
PAGE
I
BYTE
Storage Key. An additional Test and Set instruction
is also available to temporarily segregate part of
the core storage; this allows updating operations to
take place safely without risking intervention by a .
different processor or a different program during
the update operation.
II :r
I
'""oc.m, '
HIGH ORDER
I
LOW ORDER
REGISTERS (8)
.125 High-Resolution Interval Timer
PHYSICA~
An addressable automatic interval timer has been
added to the Model 67 system. The timer is incremented every 13 microseconds, and incrementing
requires no processor or core storage cycles. An
interrupt can be programmed to occur whenever the
timer count reaches zero.
ADDRESS
I
I
I
I
IF AN EQUAL
I
I
I
I
I
SEGMENT
PAGE
ON COM_
PHYSICAl.
ADDRESS
~,~O
:
I
I
.2
I - - :~sg::'~~';\'
}
COfIRESPONDING
I
.+ - -
Model 67 uses virtual addressing, rather than absolute addressing, during program execution. The
additional registers used for this purpose are listed
in Paragraph. 122. These registers are used to
determine the physical addresses within core memory of the operands to be used by the program!. as
described in Paragraph .121 and Figure 1 of thIS
report section. In all other respects, the processing facilities of Model 67 are the same as those of
other System/360 Processing Units, as described
in Section 420:051.
SEGMENT AND PAGE
I
i . . . . . / ' F NO EQUAc ON COMPARE
IF NO
EQUAL ON
COMPARE
SEGMENT TAB~E
PAGE
E~
--
/
.L't'1
. \'L.I
PHYSICAL
ADDRESS
-,
~
SEGMENT TABLE ORIGIN
TAB~E
TAB~E
PAGE TABL.E ORIGIN
-- I
.3
Figure 1. Address Interpretation in Model 67
· 122 Additional Registers (Contd.)
physical page addresses occupy 12 bits each, and
B bits are used to hold frequency-of-use data.
Access time is 150 nanoseconds.
CD
Sixteen 32-bit Multiple Program Control
Registers.
•
One 24-bit Instruction Page Address Register •.
• 123 Storage Protection
So far as the central processor hardware is concerned, interruption, multiprogramming, and
multi-sequencing are handled in the same way in
Model 67 systems as in the standard System/360
processors.
.4
In the Model 2067, the standard four-bit storage
protection keys used in the other System/360 models
have been extended to seven bits for each 2, 04B-byte
block of core storage. The standard four-bit key
provides protection against overwriting by permitting writing only by programs that supply the key
value- (between 1 and 15) presently being used to
protect each block. An attempt by any other program to write data in the block results in an interrupt. In the 2067, a f~th bit determines whether or
not reading from each block shall be permitted, and
two test bits record whether or not any reading or
any writiIig has occurred in the block since the bits
were reset. The same four-bit key value is used
for protection against both overwriting and (if the
. fifth bit is set "on") reading.
• 124 New Instructions
Additional or modified instructions provided to control the new facilities are: Load Multiple Control,
Store Multiple Control, Set Storage Key, and Test
SEQUENCE CONTROL FACILITIES
The Mask Register and I/o addressing methods
have been modified so that each processor in a
Model 67 system can address and control each of
the 2B possible input-output channels. In addition,
a special single bit can inhibit all input-output
channel interrupts during changeovers from one
program to another. Using these facilities, it is
possible to control which processor within the system will respond to interrupts from each channel.
REGISTER
©
PROCESSING FACILITIES
PROCESSOR SPEEDS
The processor speeds of the Model 67 are significantly affected by.the necessity to locate each operand by using the virtual address in the instruction
to find its physical address in core storage. When
the address of a particular page is in the associative
memory, the time spent in transforming the virtual
address to the physical address is 0.150 microsecond. However, when references to both the segment table and the page table in core storage are
necessary before the physical address of an operand
can be found, an additional delay of 1. 9 microseconds is involved. These delays are incurred
each time an operand address is used.
In the following performance times, it has been assumed that 95 percent of the operands sp'ecified in a
program will have their page addresses held in the
associative memory, so the average delay caused
by the necessity to change the virtual addresses to
physical addresses will be 0.245 microsecond per
operand. No account has been taken here of the
delays caused when more than one processor and/or
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
)':J
427:051.400
.4
IBM SYSTEM/360-MODEL 67
PROCESSOR SPEEDS (Contd.)
Fixed point
.41
Instruction Times in Microseconds
. 411 Fixed point -
• 418
Binary Decimal
Add-subtract:
1. 64
3.88 + O. 5B
4.25 + 4. 2B + B2
Multiply: •••••• 5.04
6.44 + 4. 6B + 2.2 B2
Divide: ..••.•.. 8.70
where B = operand length in
eight-bit bytes (2 decimal
digits per byte).
Floating point Long
Sho-rt
Add-subtract:
2.70
2.67
Multiply: •.••..• , 7. 84
4.64
Divide: . • • . . . . . 14.34
7.54
Additional allowance for Single indexing: • •• O. 0
Double indexing: •• 0.39
Indirect addressing: none.
Recomplementing:
none.
Control:
Compare Fixed point:
1.64
Decimal: •••••• 3.71+0.43B
Floating point
(long): ••••••• 2.24
Floating point
(short): •••••• 2.27
LOgical: •••••• 3.34+0.4B
Branch: •••••••• 1.39
Counter control Step: . • . . . . . . . • 1.84
1.29 (increment of -1)
Step and test:
1.84 (increment of any
value).
Test: •••••••••• 1.64
Edit: ••••••••••• 3.44+0.8B
Convert
To binary: •••••• 7.88
To decimal: ••••• 8.53 (positive); 8.93
(negative) •
Shift:
variable •
.42
Processor Performance in Microseconds
.412
.413
.414
.415
.416
.417
.422
.423
.424
.425
.426
.427
.428
.421 For random addresses Floating point
Fixed point
.5
c = a + b: ••••••.• 4.2 (binary)
5.5 (long)
6.7 + 1. OB
5.4 (short)
(decimal)
b = a + b: •••.•••• 4.2 (binary)
5.5 (long)
5.4 (short)
6.7 + O.5B
(deCimal)
Sum N items: .••••. 1. 7N (binary)
2.6N (long)
(3.6 + O. 5B)N 2.6N (short)
(decimal)
8/65
= ab:
•••.••.• 7.7 (binary)
10.4 (long)
8.9 + 5. 2B +
6.8 (short)
. 1. OB2
(decimal)
16.9 (long)
c = alb: •••••••• 12.0 (binary)
12.1 + 5. 6B + 10.0 (short)
2.2B2
(decimal)
where B = operand length
in eight-bit bytes (2
decimal digits per byte) .
For arrays of data .;..
FiXed point
Floating point
8.2 (long)
ci = ai + bj: .•.•••• 6.9 (binary)
8.1 (short)
9.4 + LOB
(decimal)
8.2 (long)
b j = ai + b j : •••••• 6.9 (binary)
8.1 (short)
9.4 + 0.5B
(decimal)
Sum N items: ••..• 3.4N (binary) 4.4N (long)
(5.4 + 0.5B)N 4. 4N (short)
(decimal)
13.5 (long)
c = c + aibj: •••••• 9.3 (binary)
13.4 + 5.2B + 10.0 (short)
1.0B 2
(decimal)
Branch based on comparison Numeric data: • • .• 5.8N
Alphabetic data: •• 5.8N
Switching Unchecked: • • • • •• 4.8
Checked: ••••••• 8.0
List search: ••••• 6.2 + 3.6N
Format control, per character Unpack: • • . • • . •• 0.5
Compose: •••••• 1. 5
Table lookup, per comparison For a match: •••• 3.8
For least or
greatest: •••••• 4.8
For interpolation
point:. • • • • • • •• 3.8
Bit indicators Set bit in separate
location: • • . • • •. 1.6
Set bit in pattern: • 2. 7
Test bit in separate
location: •••••• 2.7
Test hit in pattern: • 3.4
Moving: ••••••••• 2.6 + 0.20B, whereB =
number of bytes moved.
c
more than one storage unit are used in the. same
Model 67 configuration. Details of these delays,
which may be substantial, are included iIi the Simultaneous Operations section, on page 427:111.100.
Floating point
ERRORS, CHECKS, AND ACTIONS
Model 67 has the same hardware error-checking
facilities as the standard System/360 models; see
Paragraph 420:051.5.
/"
.
427: 111.100
IBM System/360
Model 67
Simultaneous Operations
SIMUL TANEOUS OPERATIONS
A System/360 Model 67 system can concurrently execute:
• One maclti,ne instruction per processor;
• One input-output operation per Selector Channel; and
• Multiple additional input-output operations via any connected Multiplexor
Channel.
The interference or delay imposed upon central processor programs will vary depending upon a
number of factors. Specific possible causes of interference in Model 67 systems are individually
discussed below. In some cases it is possible to overlap one type of interference with another,
thus reducing the net total interference.
Multiprocessor Systems
In a multiprocessor system, where two or more processors have access to the same core storage
areas, some conflict will probably arise as a result of simultaneous access requests from two or
more processors, only one of which can be immediately satisfied. Simulation studies performed
by IBM indicate that where two processors are involved, these storage access conflicts can cause
each of the two processors to be delayed about 13 per cent of the time, thereby proportionately
reducing their efficiency.
.
Multiple Core Storage Modules
Where a configuration includes more than one core storage module ,cable-length delays will occur in
the physically larger installation. SimUlation indicates that the memory access time in multiplemodule configurations is increased by 50 nanoseconds for each core storage module which physically intervenes between the processor and the accessed core module.
Data-Flow Interference
Data flowing into and out of core storage modules has priority over all central processor operations in obtaining access to the storage modules. The resulting delay in processor operations
appears to be approximately 0.10 percent per 100,000 bytes per second of data flow. "Data
flow" includes all input-output data plus all data being moved from one level of storage to another
for systems reasons.
Channel Characteristics
The characteristics of the 2846 Channel Controller, 2860 Selector Channel, and 2870 Multiplexor
Channel used in Model 67 systems can be summarized as follows:
2846 Channel Controller
Maximum number: .•••••••••••.•••••.•••.••
Maximum data rate per controller: .••.•••••••••••
Maximum number of channels connected: •.•.•.•••••
4.
6,400 kilobytes/sec.
7.
2860 Selector Channel
\
Number of channels per unit ;"
2860 Modell: . • • . • . . . . • . . • • . • • • . • . • . • . . .
2860 Model 2: .•.••••.•••..••.....•.•••••
2860 Model 3: .•...•.•.•.••••..•.•.•.••..
Number of channels per system: . . . . . • • . . . . . . . . . •
Maximum data rate per channel: .•.••....•.•..•.•
Maximum data rate, all channels: ..••.•.•..•...•.
Number of control unit positions: .....•.•.•••••••
1.
2.
3.
up to 7 channels can be
connected to each Channel
Controller.
1,300 kilobytes/sec.
6,400 kilobytes/sec per
Channel Controller.
8 per channel.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
IBM SYSTEM/360-MODEL 67
427:111.101
2870 Multiplexor Channel
Maximum number: •••••••••••.•••.••••••••••
Maximum total data rate: ••.••.••••.•••••••••.
Basic Channel Maximum data rate:
With no Selector Sub channel operating: •••••••••
With 1 Selector Sub channel operating: ••••••••••
With 2 Selector Sub channels operating: •••••••••
With 3 Selector Sub channels operating: , ••••••••
With 4 Selector Sub channels operating: •••••••••
Number of control unit positions: ••••••••••.••••
Number of subchannels: ••••••••••••••••••••
Selector Sub channels Maximum number: •••••.••••••••••••••••.•
Maximum data rate per Selector Sub channel: ••••••
Maximum data rate, all Selector Subchannels: •••••
Number of control unit positions per Selector
Subchannel: •••••••••••••••••••.•••••••
Maximum number of peripheral devices per
Selector Subchannel: •••.•.•••.•••••••••••
8/65
7 per Channel Controller.
450 kilobytes/sec.
110 kilobytes/sec.
95 kilobytes/sec.
80 kilobytes/sec.
65 kilobytes/sec.
50 kilobytes/sec.
8.
192.
4.
100 kilobytes/sec.
400 kilobytes/sec.
8.
16.
427: 161.1 00
(
IBM System/360
'.
Model 67
Programming Languages
PROGRAMMING LANGUAGES
.1
GENERAL
The first set of computer languages to be made
available for users of the Model 67 will include:
• A mnemonic assembler with macro capabilities.
" A .FORTRAN IV compiler.
• A terminal Command Language.
At later dates a PL/I compiler, a non-conversational COBOL compiler, and a Sort/Merge routine
will be added. Table I shows the schedule<;l release dates for the documentation of these languages and for the programs themselves. In
addition to these languages, potential Model 67
users have been informed that a General-Purpose
Systems Simulator (GPSS) compiler and a Desk
Calculator language will be made available.
Preliminary descriptions of these languages
and the associated translators are provided below.
In discussing the software for time-shared computer systems, an understanding of several
specialized concepts is important:
• Re-entrant Coding: A routine is considered
to be re-entrant when it can be used by more
than one program at the same time. This
means that the routine cannot modify the contents of any of its own locations, and that any
required temporary storage must be supplied
along with each program using the re-entrant
routine. Re-entrant coding has the virtue of
being economical in its demands upon core
storage since only one copy of any re-entrant
routine is needed in core storage regardless
of how many different programs are simultaneously utilizing the routine. An additional
virtue of re-entrant coding is that since it is
never modified, it does not need to be re-
written in aUXiliary storage when it is displaced from core storage by another routine
(i. e. , when "page swapping" occurs).
• Interpretive Mode: An interpretive translator performs the functions directed by each
program statement immediately upon receipt
of the statement, without waiting for the next
statement to arrive.
•. Conversational Mode: This implies a "dialogue" between the user and the computer, in
which the translator examines the input supplied by the user and formulates questions
which are directed back to the user. The
user's responses to the questions are then
transmitted back to the computer. An example
of conversational-mode operation would be
checking for errors in the formation of statements, and correcting them, during the input of
a FORTRAN source program.
All Model 67 language translators will be written
in re-entrant coding, and all will use only direct
access storage; i. e., no "scratch" tapes will be
used. Translators will be available that operate
in the conversational mode, the interpretive mode,
and the conventional batched mode. IIi several
cases, the user will have a choice of modes, as
described below.
.121 Time-Sharing Assembler
The Time-Sharing Assembler will use the same
symbolic input language as the standard Assembler which operates under control of the Operating
System/360 (Section 420: 171). Like the other
Model 67 translators, it will be treated by the TimeSharing Monitor as an ordinary problem program.
During the entry of an assembly-language source
program from a terminal, it will be possible to
perform editing and diagnostic checking of the
TABLE I: MODEL 67 PROGRAMMING LANGUAGE AVAILABILITY
Documentation
Available
Program
Available
Conversational Assembler
1st qtr. 1966
2nd qtr. 1967
Conversational FORTRAN IV
1st qtr. 1966
2nd qtr. 1967
Terminal Command Language
1st qtr. 1966
2nd qtr. 1967
PL/I
3rd qtr. 1966
3rd qtr. 1967
Sort/Merge (n::m-conversational)
3rd qtr. 1966
3rd qtr. 1967
COBOL (non-conversational)
4th qtr. 1966
4th qtr. 1967
Language
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
'7
IBM SYSTEM/360-MODEL 67
427:161.121
.121 Time-Sharing Assembler (Contd.)
source program in a conversational mode. Facilities will be available for symbolic updating of assembly-language programs held on file and for obtaining cross-reference listings.
The Time-Sharing Assembler itself will be written
in re-entrant coding. The object-program coding
it produces will not necessarily be re-entrant. No
new facilities are currently planned to assist the
programmer to write re-entrant programs in the
assembly language.
The time-sharing GPSS compiler will use re-entrant coding; the object programs it produces will
also be re-entrant.
.126 Terminal Command Language
The Terminal Command Language will enable users
at remote terminals to:
• Control the construction and execution of programs.
• Control the construction, maintenance, and use
of files .
. 122 FORTRAN IV
The time-sharing FORTRAN IV compiler will use
the .same input language as the Operating System/
360 FORTRAN IV Compiler (Section 420: 161), with
some relaxation of the rules involved.in program
construction. Like the other Model 67 compilers,
it will be treated by the Time-Sharing Monitor as
an ordinary problem program. The compiler itself and all the programs it compiles will be reentrant.
During the entry of a FORTRAN IV program from a
terminal, it will be possible to make corrections
in a conversational mode of operation. No interpretive FORTRAN processor is being provided.
. 123 Programming Language/I
The time-sharing Programming Language/I (PL/I)
compiler will use the same input language as the
Operating System/360 PL/I compilers (Section
420: 167). Like the other Model 67 compilers, it
will be treated by the Time-Sharing Monitor as an
ordinary problem program. The compiler itself
will be written in re-entrant coding, as will all of
the programs it compiles.
During the entry of PL/I program from a terminal,
it will be possible to make syntax and grammatical
corrections in a conversational mode of operation.
. 124 Desk Calculator Language
The Desk Calculator Language (which may be a
subset of the PL/I compiler) can be used for "realtime" mathematical computations, including the use
of functions such as log, exponent, root, sin, cos,
etc. Its operational mode will be interpretive; 1. e. ,
the translator will execute each input statement as
soon as the user finishes typing it.
.
. 125 General-Purpose Systems Simulator (GPSS) Lan~
A GPSS compiler will be provided for both batch
(conventional) and conversational modes of operation. The source language will be GPSsm. An
assembly program will provide free-form format
and symbolic addressing of blocks. Thenumber of
parameters associated with each transaction can be
dynamically varied by the user. Chains of transactions can be created and manipulated. System
attribute functions can be specified as ordinate values to change dynamically with the system. System
variables can be specified in virtually all block
fields. Debugging facilities permit sampling and
display of the state of the system during the simulation run.
8/65
• Direct program checkout operations.
• Perform immediate calculations.
The language has not been fully specified to date;
therefore, the following function list is illustrative,
rather than definitive:
Control Functions:
• LOG-ON - Display standard information such as
time and date, and request the user to identify
himself.
• LOG-OFF - Display elapsed time, used time,
and time of log-off; and cause an orderly return
of all facilities still associated with the terminal
to public availability.
• RUN - Initiate a user program at a specified
point.
• ATTENTION - Halt execution of a program.
• CHECKPOINT - Save the current user status in
a manner permitting reconstruction by the restore function.
• RESTART - Reconstruct a specified user status
that was previously saved.
•
LOAD - Load a program and prepare it for execution.
Debugging Functions
• DUMP - Dump specified regions of a program
(including files) in a specified format.
• DISPLAY - Exhibit specified registers, indicators, locations in standard formats .
• SET - Alter the contents or conditions of specified registers or indicators.
\
• TRACEBACK - Provide a symbolic traceback
through the hierarchy of subroutine calls.
.127 COBOL
The COBOL language for Model 67 is similar to the
IBM Operating System/360 COBOL language, which
is described in Report Section 420: 164. The COBOL compiler is non-conversational; i. e., it will
operate only in the batch-processing mode, and is
designed for use as "background" program.
./
427:191.100
IBM System/360
Model 67
Operating Environment
OPERATING ENVIRONMENT: TIME.SHARING MONITOR
.1
GENERAL
. 11
Identity:.........
. 12
II
Time-Sharing Monitor.
Description
The principal design objective of the System/360
Model 67 is to furnish continuous computing service to many users simultaneously, while providing
virtually instantaneous responses to each of the
users. To achieve this objective, Model 67 incorporates specialized hardware facilities' that permit
dynamic storage allocation, multiprocessor configurations, extended storage protection, and multipath access to most system components. Equally
important in achieving Model 67's design objective
is the provision of an operating system capable of
serving as an interface between the equipment and
its users and of effectively employing all of the
hardware facilities.
The Time-Sharing Operating System for Model 67
includes many of the facilities and features of the
Operating System/360 (Section 420:191) with the
necessary extensions to enable it to control timesharing operations in multiprocessor, multiprogrammed systems. The programming languages
included in the Time-Sharing Operating System are
described in Section 427:161. Also included in the
Time-Sharing Operating System are a sort/merge
routine and an open-ended library of mathematical
and utility routines. The remainder of this report
section describes the Time-Sharing Monitor, the
key routine that controls the execution of all jobs
entering a Model 67 system and the hardware environment in which they operate.
Error-handling, checkpoint, and restart functions
are performed by separate, re-entrant routines
that are called into core storage by the Monitor as
required.
To provide the required rapid response to every
user, the Time-Sharing Monitor employs the "timeslice" technique. The time-slice is the length of
time each individual user terminal is allowed access
to a processor before the next terminal in line is
serviced. Each installation will be able to select
the time-slice length that will provide the best overall balance between rapid response and efficient '
processor utilization. To minimize the inefficiencies that arise from_too:"'frequent changeovers between programs, IBM currently advocates the use
of relatively long time-slices of around 250 milliseconds; but such long time-slices could easily lead
to excessive delays in response as the number of
simultaneous users increases.
The Monitor's principal task is to respond to interrupts. It must determine the type and reason for
each interrupt signal and then initiate execution of
the appropriate routine to service the interrupt condition. The Monitor recognizes and services four
basic categories of interrupts:
• Processor Control, involving the allocation
of central processor time.
•
It will be possible for Model 67 users to replace
many parts of the IBM-supplied Monitor with their
own private routines to accomplish the corresponding functions, and many will probably do so. The
subsequent description of the Time-Sharing Monitor
applies to the general-purpose version which IBM
plans to make generally available.
Prevent duplicate copies of the same re-entrant
subroutines from occupying core storage at the
same time .
Storage Control, involving the allocation and
protection of both primary (core) and secondary
(disc, drum, etc.) storage. (The Monitor currently contains no provision for using Model
2361 Large Capacity Core Storage as either
primary or secondary storage, although 2361
units can be included in Model 67 systems.)
The principal functions of the Time-Sharing Monitor
are to:
• I/O Control, involving the assignment of I/O
devices to specific programs and the establishment of data paths between the I/o devices and
core storage.
•
Maintain continuous indications of the status and
usage of every system facility (storage units,
I/O devices, processors, and files).
•
•
Supervise the handling of all hardware-originated
interrupts.
The Time-Sharing Monitor uses a two-part Task
Status Index (TSl) to hold the necessary status information about each task curreritly in the system.
•
Record the data required for job scheduling and
cost allocation.
System Control, which involves the logical control of the above three functions plus operator
communication, partitioning, and recovery.
•
Allocate the system's resources in such a way
as to minimize response time to users.
Part of each TSI holds certain user-oriented infor- '
mation that must be kept in core storage at all
times; the other part holds information that can be
relegated to secondary storage when necessary.
•
Provide for "fail-soft" operation by enabling,
whenever possible, the remainder of the
system to keep operating when one component
malfunctions.
The Monitor's storage allocation algorithms deal
with 4, 096-byte blocks or "pages" of core or secondary storage. The Monitor keeps a record of
the current allocation of each block of core storage.
© 1965 AUERBACH Corporation ane;! AUERBACH Info, Inc.
8/65
427:191.120
. 12
IBM SYSTEM/360-MODEL 67
.44
Description (Contd.)
In allocating core storage to programs or data
called in from secondary storage, the Monitor first
tries to assign free blocks. If not enough blocks
are free, the Monitor assigns space that was previously allocated to "inactiv!3" programs (i. e. ,
programs awaiting a user response). If neither of
these sources can provide the required core storage space, the Monitor attempts to take it away
from active programs that have lower priorities
than the program requesting the space. If this is
not possible either, the task is deferred until sufficient core space becomes available. (If the contents of a page about to be overwritten have changed
since it was last called in from secondary storage,
as indicated by one of the seven bits in the storage
protection key, then the page must be rewritten
into secondary storage before it can be overwritten
in core storage by the new page. )
The Monitor's job scheduling algorithm uses a
"pointer" that indicates the Task Status Index of
the task which should be executed next if possible.
When the time-slice allocated to each task ends.
the pointer moves on to the next task. Whenever
a processor needs work, a dispatcher routine
checks the status of the task currently indicated by
the pointer. If the task is in a ready state, it is
dispatched to the appropriate processor for execution. If the task is not ready (e. g. , because it is
awaiting completion of an I/O or "page-turning"
operation), the dispatcher searches through the
Task Status Index chain for the next ready task,
which is then assigned to a processor for execution
until an interrupt occurs. Then control is returned
to the Monitor and the dispatching process.is reinitiated, with the task now indicated by the pointer
again receiving priority.
More than one chain of Task Status Index entries can
be maintained if desired, and each chain can use a
different time-slice value. This permits various
types of terminal equipment and various classes of
users to be serviced in a manner appropriate to
their differing needs.
Errors, Checks, and
Action
Check or
Interlock
Action
Allocation
impossible
In-out error single
In-out error persistent
software
check
interrupt
Invalid instructions
Arithmetic
overflow
Underflow
interrupt
program is
deferred.
retry I/O operation;
I/O unit is closed
down and diagnostic routine
informed.
program is terminated.
varies with program.
varies with program.
program is terminated.
format is queried
with user.
log entry is made
and program
terminated.
Error
interrupt
interrupt
interrupt
interrupt
Invalid operation
Improper
software
format
check
Reference to
interrupt
forbidden area
.45
Restarts
Restart points must be established by the user's
own coding. When an error is detected, processing is halted in all processors within the system.
A diagnostic program is called in and an attempt
is made to retry the instruction or to roll back to
a previous restart point within the same program.
If a machine error occurs either during this process or when the restarted program again reaches
the point where the original error occurred, the
program is aborted and the operator is informed.
The operator has the option of taking the processor
concerned off-line. If a successful reconstruction
of the original error occurs, or if the restarted
program proceeds normally, the necessary log
entries are made and the other processors reactivated.
.6
OPERATOR CONTROL
Detailed documentation is scheduled to be published
in the first quarter of 1966, and the Time-Sharing
Monitor will be available for use in Model 67 systems in the second quarter of 1967.
.61
Signals to Operator:.
handled by Monitor.
.62
Operator's
Decisions: .•.••.
via operator's console.
.2
PROGRAM LOADING
.63
Operator's Signals:..
via operator's console .
. 21
Source of Programs: .. from on-line library or
any terminal.
.7
LOGGING:. . . . . . ..
handled by Monitor •
.22
Library Subroutines: .. only from on-line library.
.8
PERFORMANCE
.23
Loading Sequence: .•• handled by Monitor; see
preceding Description.
.81
System Requirements
.3
HARDWARE ALLOCATION: . . . . . . .
handled by Monitor.
.4
RUNNING SUPERVISION
.41
Simultaneous Working:. handled by Monitor.
.42
Multiprogramming: .•. handled by Monitor.
. 43
Multi-sequencing:.... handled by Monitor. using
program description to
define independent tasks.
8/65
.811 Minimum configuration:. . . • . . . . ..
812. Usable extra
facilities: • . • . . ••
813. Reserved equipment: .••••.•• "
. 82
System Overhead:. ..
Model 2067 with at least
262K bytes of core storage and a direct access'
device •
any Model 67 configuration.
4,096 bytes per processor .
see System Performance
section, page 427:201. 100.
427:201.100
IBM System/360
Model 67
System Performance
SYSTEM PERFORMANCE
A Model 67 includes between one and four modified Model 65 Processing Units, so
its maximum potential processing power is the sum of the processing power of the individual
processors which make up the system. The performance of Model 65 systems has been calculated and is shown in the System Performance section of the Model 65 subreport, beginning on
page 426: 201.001. This report section concentrates upon methods of calculating the performance
of specific Model 67 systems based on: (1) the available performance figures for Model 65, and
(2) the various performance degradation factors that result from the Model 67 system's configuration, the software, the number of program changeovers* per second, and the input-output
loading. All these factors have a direct effect upon the processing power of Model 67 systems,
and all are summarized in Table 1.
TABLE I: FACTORS AFFECTING THE PROCESSING POWER OF MODEL 67 SYSTEMS
FACTOR
RESULTING SYSTEM DEGRADATION
Use of Associative Memory Addressing
All addresses must be converted from
virtual addresses to physical addresses
before they can be used; this effectively
lengthens the memory cycle.
12%
Use of Multi-Processor Systems
Where two or more processors in a multiprocessor system simultaneously attempt to
address the same core storage module, conflicts occur. Time spent in settling the
conflicts causes delays in the system.
13% for a two-processor system.
Use of Multiple Core Storage Modules
Where more than one core storage module is
included in a system, so that the module to
be accessed is not physically adjacent to the
processor, delays occur due to the increased
cable lengths.
Approx. 2% per module for the third and
subsequent core storage modules; this
delay can be overlapped with the system
degradation due to use of multi-processors.
Interference Due to Input-Output Loading
Input-output operations have priority over
processor operations in obtaining access to
core storage; this causes delays in processor
operations.
Approx. 0.1% per 100,000 bytes/second
transferred into or out of core storage
modules.
Use of the Time-Shared Monitor
Monitoring the operation of the system and
constantly reorganizing the core storage
areas will require a significant amount of
processor time. No detailed estimate of
the system degradation due to this factor is
available, but 10% would not be unusual.
Estimated at 10%.
Changeovers from Program to Program
Each time one user's program halts and
another user's program is initiated or reinitiated, a certain amount of data storage,
flag resetting, etc. is required.
Estimated at 0.1 millisecond per program
changeover.
* i. e., The number of times one user program halts and another user program starts; this
is equal to the number of "time-slices" used.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
427:201.101
IBM SYSTEM/360-MODEL 67
TABLE II: SYSTEM DEGRADATION OF A SPECIFIC TWO-PROCESSOR MODEL 67 CONFIGURATION
No. of Program
Changeovers
per Second
1
2
4
10
20
30%
30%
30%
30%
30%
40
100
200
400
1,000
31%
31%
32%
34%
40%
*
Processing Power Available
in Each Time-Slice
(In Equivalent Model 65
Milliseconds)
System Degradation
From All Causes
Except I/O Loading*
1,400
700
350
140
70
34
14
7
3.3
1.2
I/O loading causes an additional system degradation of approximately
0.1% per 100,000 bytes per second transferred into or out of the core
storage modules.
TABLE III: ESTIMATED CYCLE-TIMES, IN MILLISECO~S, FOR A
SPECIFIC TWO-PROCESSOR MODEL 67 SYSTEM
AVERAGE PROCESSING POWER USED PER TIME-SLICE
(in terms of equivalent Model 65 processing power)
1
msec
rim
j:;P'i
2
msec
4
msec
10
20
meec
msec
40
msec
100
msec
3.0
7.5
15
30
1
0.9
1.6
2
1.8
3.3
4
3.6
6.5
6.0
200
msec
msec
73
150
290
400
15
30
60
150
290
580
30
60
120
290
580
1,200
E-
To Satellite
System
(at right)
1,100
2404 Magnetic Tape Unit
and Control (2)
2402 Magnetic Tape Units (6)
2401 Magnetic Tape Units (2)
(16 drives total)
All Model 3: 90,000 bytes/sec;
112.5 inches/sec*
13,920
Simultaneous Read-While-Write
(on 2402 and 2401)
140
TOTAL ON-LINE EQUIPMENT:
$55,120
TOTAL SATELLITE EQUIPMENT:
$ 7,010
TOTAL RENTAL:
$62,130
* Using alternative tape drives and their appropriate controllers, total system rentals are as follows:
With Model 5 drives (75.0 inches/second, 120,000 bytes/second): • . . • • . . . . • . . ••
With Model 6 drives (112.;5 inches/second, 180,000 bytes/second):. . . • . . . . • . . . •.
With 7340 Hypertape drives (112.5 inches/second, 340,000 bytes/second):. • . • . . . .•
$58,330
$62,410
$66,930
(Contd.)
8/65
SYSTEM CONFIGURATION
428:031.301
SATELLITE EQUIPMENT
Deviations from Standard Configuration: . . . • . . . . punch is 50% faster.
Equipment
Rental
Main Storage (8,192 bytes)
2030 Processing Unit, Model C30
(includes one Multiplexor Channel)
$1,275
1051 Control Unit and Adapter
1052 Printer-Keyboard
225
2821 Control Unit, Modell
970
2540 Card Read Punch, Modell:
Reads 1, 000 cards per minute
Punches 300 cards per minute
660
1403 Printer, Model 3:
Pr-ints 1,100 lines pet: minute
1100 LPM Attachment
900
75
i
\.
2403 Magnetic Tape
2402 Magnetic Tape
2401 Magnetic Tape
(4 drives total)
All Model 2: 60,000
75 inches/sec
Unit and Control
Unit
Unit
bytes/sec;
1850 Channel-to-Channel Adapter
225
Selector Channel
215
Decimal Arithmetic
TOTAL SATELLITE EQUIPMENT:
To Main System
(at left)
2,440
25
$7,010
!
\
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
428:051.100
IBM System/360
Model 75
Centrol Processor
CENTRAL PROCESSOR
.1
GENERAL
.11
Identity: . . . . . . . . . . mM 2075 Processing Unit.
. 12
Description
See Section 420:051 for a comprehensive description
of the characteristics of all the System/360 Processing Units.
See Section 426: 011 for a summary of the distinguishing features of the 2075 Processing Unit as
used in Model 75 systems.
The Instruction Times and Processor Performance
times for Model 7 5 systems are listed below. Timin.g
data for the two-way-interleaved, 262K-byte
Model H is shown separately from the times- for the
larger, four-way-interleaved Models I and J. Interleaving of core storage can yield improved performance through overlapped accessing of the individual
memory banks, but the comparative times shown
here indicate that the advantage of four-way interleaving over two-way interleaving in Model 75 is
small.
Timing data is listed for all four arithmetic modes
available in System/360. See Paragraphs 4: 050. 41
and 4: 050.42 of the Users' Guide for the definitions
of these standardized measures of central processor
performance.
.4
(
.415 Counter control Step: .•.•••.- •••• 0.80
Step and test: . . . . . 1. 06 (increment of -1).
1. 24 (increment of any
value).
Test: .•••...•.•• 1.13
.416 Edit: . . . . . . . . . . . . 3.47 + 0.53B
.417 ConvertTo binary: .......?
To decimal: ...• _..?
.418 Shift: ...••••.•••• variable.
. 41
Instruction Times in Microseconds (Models I and J)
.411 Fixed point Binary
Decimal
Add-subtract: . • . . . 0.70
Multiply: .•...••. 2.80
3.56+0.39B
3.52 + 2. 7B +
1. 14B2
3.68 + 3.0B +
Divide: . . . . . . . . . . . 6.70
1.9B2
where B = operand length in
eight-bit bytes (2 decimal
digits per byte).
.412 Floating point -
PROCESSOR SPEEDS
.41 Instruction Times in Microseconds (Model H)
. 411 Fixed point Decimal
Binary
Add-subtract: . . . . . 0.80
3.69+0.39B
Multiply: . . . . . . . . 2.80
3.57 + 2.7B +
1. 17B2
Divide: . • . . . . . . . . 6. 70
3.87 + 2.96B +
1. 89B2
where B = operand length
in eight-bit bytes (2
decimal digits per byte).
.412 Floating point Long
Short
0.92
Add-subtract: . • . . . 0.92
Multiply: . . . . . • . . 4. 10
2.10
Divide: . . . . . . . . . 7. 10
3.90
.413 Additional allowance for Single indexing: . . . • O. 0
Double indexing: . . . ?
Indirect addressing: . none.
Recomplementing: •• none.
.414 Control:
Compare Fixed point: .•..• 0.80
Decimal: . . . . . . . 3.50 + O. 39B
Floating point
(long): . . . . . . . . 0.90
Floating point
(short): . • . • . . . 0.90
Logical: . . . . . . . . 0.85
Branch: .•••••... 1. 02
Add-subtract: .••.• 0.89
Multiply: •.•.•••. 4.10
Divide: . . . . . . • . • . 7.10
0.89
2.10
3.90
.413 Additional allowance for Single indexing: . • • . O. 0
Double indexing: . . . ?
Indirect addressing: . none.
Recomplementing: .. none.
.414 Control:
Compare Fixed point: . • . . . O. 7
Decimal: .•.•••. 3.5 + O. 39B
Floating point
(long): .• , •.••. 0.85
Floating point
(short): ...••••. 0.85
Logical: ••••.••. 0.70
Branch: .••••.•• 1. 06
.415 Counter control Step: •.••••.•••. 0.70
Step and test: •••.• 1. 04 (increment of -1).
1. 17 (increment of any
value).
Test: ••••••.•... 1. 04
.416 Edit: .•..• : •.••.• 3.2 + O. 5B
.417 Convert To binary: .•.••.• 3. 8
To decimal: •••..• ?
.418 Shift: •••••••••••• variable.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
IBM SYSTEM/360-MODEL 75
428:051.420
.42
Processor Performance in Microseconds (Model H)
.42
Processor Performance in Microseconds
(Models I and !D
.421 For random addresses .421 For random addresses Fixed point
Floating point
Fixed point
c = a + b: . • . • • • • • 2. 54 (binary) 2. 66 (long)
6.51 + 0.57B 2.66 (short)
(decimal)
b = a + b: . . . . . . . . 2.54 (binary) 2.66 (long)
3.69 + 0.39B 2.66 (short)
(decimal)
Sum N items: .•••• 0.8N (binary) 0.92N (long)
(3.7 +O.4B)N 0.92N (short)
(decimal)
c = ab: . • • • • • • • • • 5. 34 (binary) 5. 84 (long)
6.4 + 2. 9B
3.84 (short)
+ 1. 2B2
(decimal)
8.84 (long)
c = alb: ••••••.•• ? (binary)
6.4 + 2.9B
5. 64 (short)
+ 1. 9B2
(decimal)
where B = operand length
in eight-bit bytes.
c = a + b: •••••••• 2.3 (binary)
2.4 (long)
6.1 + 0.4B
2.4 (short)
(decimal)
b = a + b: .••••••• 2.3 (binary) 2.4 (long)
3.6 + 0.4B
2.4 (short)
(decimal)
Sum N items: ••••• O. 7N (binary) 0.9N (long)
(3.6 + O. 4B)N 0.9N (short)
(decimal)
c = ab: . • • • • • • • • • 5. 1 (binary)
5. 6 (long)
6.1 + 2. 9B
3.6 (short)
+ 1. 14B2
(decimal)
c = alb: •.••••••• 9.0 (binary)
8.6 (long)
6.1 + 3. 2B
5.4 (short)
+ 1. 89B2
(decimal)
where B = operand length
in eight-bit bytes •
. 422 For arrays of data Fixed point
.423
.424
.425
.426
.427
.428
Floating point
ci = ai + b j : ••..••. 4.54 (binary) 4.70 (long)
8.61 + O.57B 4.70 (short)
(decimal)
b j = ai + b j : . • • • . . . 4.54 (binary) 4.70 (long)
8.61+0.57B 4.70 (short)
(decimal)
Sum N items: .•••. 2. ON (binary) 2.14N (long)
(4.9 +O.4B) N 2.14N (short)
(decimal)
c = c + aibf .••.•. 6.04 (binary) 7.57 (long)
6.8+2.7B
5.57 (short)
+ 1.17B2
(decimal)
Branch based on comparison Numeric data: ••..• 4.46N
Alphabetic data: .•.. 4.46N
Switching Unchecked: •••••.• 4.06
Checked: ..•••••• 7 . 86
List search: ..•••• 3.9 + 3.1N
Format control, per character Unpack: .••••••.• 0.4
Compose: .••••••. 1. 5
Table lookup, per comparison For a match: .••.•. 2.8
For least or
greatest: ••.•••. 3.2
For interpolation
.
point: .••••••.•. 2.8
Bit indicators Set bit in separate
location: ..•••... 0.9
Set bit in pattern: . . . 1. 5
Test bit in separate
location: . . . . • . • . 1. 8
Test bit in pattern: .• 2.5
Moving: ..••.•.•.• 2.8 + 0.18B, where B =
number of bytes moved.
8/65
Floating point
.422 For arrays of data Fixed point
c. =
+ b.: ••••••• 4.1 (binary)
4.2 (long)
J
8.0+0.4B
4.2 (short)
(decimal)
b j = ai + bj: ••••••• 3.1 (binary)
4.2 (long)
5.5+0.4B
4.2 (short)
(decimal)
Sum N items: ••••• 1. 9N (binary) 2.1N (long)
(5.5+0.4B)N 2.1N (short)
(decimal)
c = c + aibj : .•..•• 5.8 (binary) 7.3 (long)
6.6+2.7B
5.3 (short)
+ 1. 14B2
(decimal)
Branch based on comparison Numeric data: •..•. 4.2N
Alphabetic data: •..• 4. 2N
Switching Unchecked: ••.•••• 3.8
Checked: •••••••• 7. 3
List Search: •••••• 3.7 + 2. 8N
Format control, per character Unpack: ••••••••• 0.4
Compose: •••••••• 1. 5
Table lookup, per comparison For a match: •••••• 2.6
For least or greatest: 3.2
For interpolation
point: •••••••.•• 2.6
Bit indicators Set bit in separate
location: . • • • • • • • O. 8
Set bit in pattern: ••• 1.4
Test bit in separate
location: •••••••• 1. 6
Test bit in pattern: •• 2.3
Moving: •••••••••• 2.6 + 0.16B, where B =
number of bytes moved.
1
.423
.424
.425
.426
.427
.428
~
Floating point
428: 111.100
IBM System/360
Model 75
Simultaneous Operations
SIMUL TANEOUS OPERATIONS
A System/360 Model 75 system can concurrently execute:
•
One machine instruction; and
•
Up to six input-output operations, one on each of the installed
Selector Channels; and
•
Multiple additional input-output operations via the basic channel of the
2870 Multiplexor Channel; and
•
Up to four input-output operations, one on each Selector Subchannel
included in the 2870 Multiplexor Channel.
The demand on the central processor (i. e., the "interference" or delay imposed on the central
processor program by each individual input-output operation) will vary depending on the type
of input-output channel. (See the general discussion of System/360 Simultaneous Operations .
in Section 420:111.) In Table I, the processor demands imposed by each of the peripheral units
are listed for each type of channel to which it can be connected.
The specific characteristics of the 2860 Selector Channel and the 2870 Multiplexor Channel
can be summarized as follows:
'
2860 Selector Channel
Number of channels per unit 2860 Modell: • • • • • • • • • • • • • • • • • • • • • • . • . . 1.
2860 Model 2: .•••••••••• '••••.•••••••••• 2.
2860 Model 3: ••••••' ••••••••••••••••••.• 3.
Number of channels per system: • • • • . • • • • • • • • • • up to two 2860 Selector Channels
in any combination of models
(maximum of 6 channels).
Maximum data rate per channel: ••••••.•••••••• 1,300 kilobytes/sec.
Maximum data rate, all channels:
Processor demand: ••••••••••••••••.••.••• see Table I.
Number of control unit positions: ••••••••••••.• 8 per channel.
.............?
2870 Multiplexor Channel
Maximum number: .••.•••••••••••••••.•••• 1 (optional).
Maximum total data rate: ••••.•••••••.•••••• 450 kilobytes/sec.
Basic Channel Maximum data rate:
With no Selector Subchannel operating: •••.•••• 110 kilobytes/sec.
With 1 Seleetor Subchannel operating: •.••••••• 95 kilobytes/sec.
With 2 Selector Subchannels operating: •••••..• 80 kilobytes/sec.
With 3 Selector Subchannels operating: .•.••••• 65 kilobytes/sec.
With 4 Selector Subchannels operating: •••••••• 50 kilobytes/sec.
Processor demand: ••.•••••••...••••••••• see Table I.
Number of control unit positions: •..•••••••.•• 8.
Number of subchannels: .•••..•.•.•.•.••••• 192.
Selector Subchannels Maximum number: ••.•••••.• " ••••....••• 4.
Maximum data rate per Selector Subchannel: .••• 100 kilobytes/sec.
Maximum data rate, all Selector Subchannels: .••• 400 kilobytes/sec.
Processor demand: •••••••••••••••••••••• see Table I.
Number of control unit positions per Selector
Subchannel: • • • . • • • • • • • • • • • • • . • • • • • • • . • 8.
Maximum number of peripheral devices per
Selector Subchannel: •••••••••••.•••••••• 16.
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
IBM SYSTEM/360-MODEL 75
428: 111. 101
TABLE I: INPUT-OUTPUT DEMANDS ON THE SYSTEM/360 MODEL 75 PROCESSOR
Device
Random Access
2302 Disk Storage
2311 Disk Drive
2321 Data Cell Drive
7320 Drum
2301 Drum
2314 Direct Access storage
Punched Card
2540 Card Read Punch:
Read, 1,000 cpm
Punch, 300 cpm
1442. Model N1 Card Read Punch:
Read, 400 cpm
Punch, 91 cpm
1442 Model N2 Card Punch, 9:\. cpm
2520 Model B1 Card Read Punch:
Read, 500 cpm
Punch, 500 cpm
2520 Card Punch:
500 cpm (Model B2)
300 cpm (Model B3)
2501 Card Reader:
600 cpm (Model B1)
1,000 cpm (Model B2)
Printers
1403 Printer:
132 columns, 600 lpm (Model 2)
120 columns, 600 lpm (Model 7)
132 columns, 1, 100 lpm (Models 3,
N1)
1443; 120 columns, 240 lpm
Average
Peak
Data Rate, Data Rate,
KB/sec* KB/sec*
156
156
54.7
135
1,200
312
156
156
54.7
135
1,200
312
Demand on Processor,_per cent, via2870 Multiplexor Channel
2860
Selector
Basic
Selector
Channel (1) Channel (1) Subchannel (1)
1.5
1.5
0.52
1.3
11
2.9
not usable
not usable
4.0
not usable
not usable
not usable
1.3
0.40
0.012
<0.01
0.1
<0.1
?
?
0.53
0.12
0.12
0.53
0.12
0.12
<0.01
<0.01
<0.01
<0.1
<0.01
<0.01
?
?
?
0.67
?
0.67
0.67
<0.01
<0.01
<0.1
<0.1
?
?
?
?
0.67
0.40
<0.01
<0.01
<0.1
<0.1
?
?
0.80
1.3
0.80
1.3
<0.01
0.012
<0.1
0.1
?
?
70
70
1.3
1.2
0.012
0.011
0.1
0.1
?
?
7_0
?
2~4
0.48
0.023
<0.01
0.2
<0.1
?
?
70
70
Magnetic Tape
2400 Series:
Modell, 30 KB/ sec
Model 2, 60 KB/sec
Model 3, 90 KB/sec
Model 4, 60 KB/sec
Model 5, 120 KB/sec
Model 6, 180 KB/sec
30
60
90
60
120
180
30
60
90
60
120
180
0.28
0.56
0.85
0.56
1.12
1.70
2.25
4.50
6.75
4.50
not usable
not usable
?
?
?
?
not usable
not usable
7340 Hypertape, 340 KB/sec
340
340
3.2
not usable
.not usable
(1) Interleaved accesses to core storage under certain conditions may reduce these
demands by up to 30%.
* Kilobytes (thousands of bytes) per second.
8/65
not usable
not usable
?
not usable
not usable
not usable
428:201.001
\
IBM System/360
Model 75
System Performance
SYSTEM PERFORMANCE
i
(
GENERALIZED FILE PROCESSING (42S:201.100)
These problems involve updating a master file from information in a detail file and
producing a printed record of each transaction. This application is one of the most common
commercial data processing jobs and is fully described in Section 4:200.1 of the Users' Guide.
Standard File Problems A, B, and C vary the record sizes in the master file. Standard Problem D increases the amount of computation performed upon each transaction. Each problem is
estimated for activity factors (ratios of number of detail records to number of master records)
of zero to unity. In all cases a uniform distribution of activity is assumed.
The Generalized File Processing problem for the System/360 was coded in two ways one using master files in packed decimal format and computation in decimal arithmetic, and
the second using master files primarily in binary format and computation in fixed-point binary.
The decimal computations required more time than those in fixed-point binary; but in the binary
case, items in the detail and report files needed radix conversion. There were no appreciable
differences in the total times for the two cases. The graphs for the file problem are based on
use of the fixed-point binary technique. For simplicity, the very similar curves based on the
use of decimal arithmetic are not shown.
In the master file record layout, alignment of data items in core storage was carefully
considered. Double-word boundaries were observed for input-output purposes to improve performance efficiency on the larger models. Instead of the "chain" mode (scatter-gather) of tape
reading and writing, individual records were moved to the work areas using a high-speed,
multiple-register transfer method.
All files - master, detail, and report - are assigned to magnetic tape. The detail
and report file tapes are assumed to be transcribed off-line from punched cards and to the printer.
Five cases are considered for the Generalized File Processing Problem. Configurations VIlB and
VillB are shown with the report and detail file tapes both blocked (dashed curves) and unblocked
(solid curves). Configuration VIlIB is also shown using 7340 Hypertape Drives and blocked detail
and report files.
Because multiprogramming of two or more independent programs is a featured capability of the System/360, the time actually used by the central processor (CP) is also plotted.
By comparing the curves of total time for the various configurations with the central processor
curves, it can be seen that even in the worst case (Configuration VIIIB using blocked detail and
record files, when the computational load has been trebled), some 75% of the available processing
capacity is not in use. However, if 7340 Hypertape Drives were used instead of the 2400 Series
Magnetic Tape Units, then the processor could be utilized to about 70% of its capacity. A comparison of the central processor curves for a standard amount of computation and for trebled
computation (i. e., the curves for Problems A and D, respectively) shows the effect of increasing
the computational workload.
SORTING (42S:201. 200)
/'
The standard estimate for sorting SO-character records by straightforward merging
on magnetic tape was developed from the time for Standard File Problem A by the method explained in Paragraph 4:200.213 of the Users' Guide. A three-way merge was used in all system
configurations for the Model 75. The results are shown in Graph 42S:201. 200.
Graph 42S: 201.220 shows the times required to sort SO-character records by means
of the Operating System/360 Sort/Merge routine, as described in Section 420:151. In addition
to the tape sorting times for Standard Configurations VIIB and VIIIB, the times required for
internal sorting on two different types of direct-access devices are shown.
MA TRIX INVERSION (42S: 201. 300)
The standard estimate for inverting a non-symmetric, non-singular matrix was computed by the simple method described in Paragraph 4:200.312 of the Users' Guide. Two lines
are shown on the graph, one using the short floating-point format (6-digit precision), and the
other using the long format (1S-digit precision).
GENERALIZED MATHEMATICAL PROCESSING (42S:201.400)
The Standard Mathematical Problem A is an application in which there is one stream
of input data, a fixed computation to be performed, and one stream of output results. Two
variables are introduced to demonstrate how the time for a job varies with different proportions
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
428:201.002
IBM
SYSTEM/360-~ODEL
7S
of input, computation, and output. The factor C is used to vary the amount of computation per
input record. The factor R is used to vary the ratio of input records to output records. The
procedure used for the Standard Mathematical Problem is fully described in Section 4:200.2 of
the Users' Guide.
Computations were performed in double-length floating-point arithmetic (16-digit
precision) since a minimum precision of 8 digits is prescribed. In Configurations VIIB and
VlIIB, the input and output files are on magnetic tape.
Graph 428:201.400 shows two curves. The curve marked R = 1. 0 is for the case in
which one output record is written for each input record. The other curve is for the case in
which one output record is written for every tenth (R = O. 1) and every hundredth (R = 0.01) input
record. (There is roo effective difference between these two cases.) In Configuration VIIB, one
magnetic tape unit is the controlling factor for amounts of computation up to 24 times the
standard (1. e •• C = 1. 0). In Configuration VmB, magnetic tape controls for up to 16 times
the standard amount of computation.
WORKSHEET DATA TABLE 1
CONFIGURATION
ITEM
VIlB (blocked)
1
Char/block
Output
Times
1-----~_3_ _
,_._
File 4
msee/swltch.
2
Standard
File
Problem A
msee/block
25.6
32.0"
10.0
22.9"
1 056
12
17.0
6.6
6.8
8.1'
- - - -0°- - - - °
--
°
°
0.10
r---- Ms,-
0.03~_
~~---
msee/work
0.162
0.162
C.P.
C.P.
I/O
..!!L,_ _ _ _ 0.03 __ _0____
_0.4o __
1-.0.40__
~-------
-----
3_ 10
a3K
Master ~ ~O
~2: MasterO_u!.. ....Q-lL
~:Details_ _
..J!:.~
0_12
File 4: Reports
~_1:
Total
(bytes)
Unit of measure
_
-'-':--- - - _ : ' 6 (Blocks 24 to 48)
t---i'.!L
3_10
-,-
:-::::-----Working
I/O
C.P.
I/O
C.P.
~O~
-,-
~0.L.
----
I-0....QL1-0.4o __
-,--- _~-40 __ -,--- - - - r-~.4~_
3_10
3.10
o. 10 -,-- -E:.!L - ,
rJ!=!L - - - 0.1o __ ---
25.6._ f-~.-
3.10
---
0_ 10
4:200.114
O_O~_
57_ 6
3_ 93
145.6
, 6,000_ _ _ _ _ _ _6,000 ____
128
128
-'--6~-
22_ 9
3.93
I-~~-128
4,092
- - 1---9",312
--
-----wo---- --Wo---- I----Wo---20,280
3_ 93
81.6
3.93
I-~_O_ _ _
8_1
_ _ 6,000_ _ _ _
128
128
- - - - - --648---- - -648
- - - -,-----648
-~O~-- -~---- 1--"""4,O9~-
Total
I/O
-,_.-
~lL_ ------ - ~.!L
_Q.08
_.- ~L.. --- I-_
r-!!'.QL
----------32_ 0
0_12
0_12
120.0
0_12
8_ 1
22.9
81.6
0.12
~;-----'- - 9 , 3 1 2 - - - 4 , 6 4 8
Space
0.162
15,616
--4;092 - - , - - 4 , 0 " 9 2 - -
- - 1 0 0 - - - - '100---16,616
20,280
Fixed/Floating point
20,280
VIlB
VIllB
F'toatlng point
Floating point
Unit name
input
Series Tape, Mod 2
Series Tape, Mod 3
1-------- - - -2400
-- - - - - - - - - -1---- -2400
- --------,output
2400 Series Tape, Mod 2
2400 Series Tape, Mod 3
Size of record
r---'-----
input
output
80 bytes
80 bytes
r-----------------1------------------120 bytes
120 bytes
I - - - - - - - - - ! . . L - - - - - .- - -- I--------~-------9.7
6.4
~~!L_
output T4
~~c/rec~ _ _ _ _T.&....-_
...l!!!!!'_c/510~ _ _ ~_
msce/report
T7
*
4:200_ 1151
---"'4;"648-- 1-- 9312- - ' -
CONFIGURATION
ITEM
Input Tl
output T2
4:200.1132
- ' - - 0 . ' i 6 2 -1 - - - - - - 0.162
C.P.
I/O
O_!~
25.6
3.93
-,-----3 (Blocks 1 to 23)
Standard
File
Problem A
~4
- - - - -f - - - -
b 7 + b8
maee/block
lor C.P.
and
dominant
1/0 column_
mace penalty
(estimated)
4:200.112
---0-'- ---0--
O~_
msee/detail
msce/block
REFERENCE
03~
~,----
msee/record
Fixed
Standard
Mathematical
Problem A
25.6
1 056
12.0
- - - ----_.- - - - - - , - - - - - - - 1 - - - - - - - - f - - - -0.12*
File 4
0.01
0.12'
0.01
0.12'
_ _ _0.030._
0.030
a1
0 • 03 °. _
f-:- ,__ O. _
1-__
f-'--~O------_ ._ _
0.033 __
0.033
0.033
1--.__
'2
-=------1--------'- - - - - - ---~-_ _ _ 0.052 _ _ I-- __
1-___
O. 05?_
0.052
r----o.~~.--------,--,_ _
__
0.044
0.044
0.044
0.044
~~outine_s_ _
5
°
File 1 = File 2
0.10
0.10
0.10
0.10
F i l e - 3 - - - ' - - - - - 0 . " 0 8 * - 1--------- - - - 0 . 0 8 * - 1 - ' - - - - 0.01
0.01
F=l.O
4
1 056
12.
17.0
Ftle-4----- - - - - 0 - - -
msee/report
3
1 056
12
VIlIB (blocked)
using Hypertape
VIlIB (unblocked)
File 1 = File 2
- - - 1 - - - -°0-- , - - - - -°- - - - ° - 1 - - °' - - - r---- o°- - ' File 3
msee penalty
Central
Processor
Times
VIllB (blocked)
----_.- ------ - - - - - - r - - - - - - - - 1-------,:--- _ _
16~_
---I - - - E - f---~--~~- 1 - - - - 9.4
File 1 =::. File 2
msee/block
Standard
File
Problem A
Input-
1 056
12
File 1
(File 1)
Records/block K
VIlB (unblocked)
--------~------- 0.01
_______
________
0.01
I--- _ _ _ _ _ _ ~O_ _ _ _ _ _ _ _
I-------~-------- 1-- _ _ _ _ _ _
--------~--------0_ 37
~_1
4:200.413
_ _ _ _ _ _. _
~4
0.37
Files 3 and 4 blocked 12 records/block
(Contd.)
8/65
428:201.100
SYSTEM PERFORMANCE
.1
GENERALIZED FILE PROCESSING
• 11
Standard File Problem A
• 111 Record sizesMaster file: •••••. 108 data characters,
packed as 88 8-bit bytes.
Detail file: ••••••• 1 card.
Report file: • • • • • . • 1 line •
• 112 Computation: •••••.• standard, using fixed-point
binary or decimal arithmetic.
.113 Timing basis: •••••• using estimating procedure
outlined in Users' Guide,
4:200.113 •
• 114 Graph: •.•••••.•• see graph below •
• 115 Storage space required Configuration VnB
(blocked): ••.•.•• 20,280 bytes.
Configuration VllB
(unblocked): .••••• 15,616 bytes.
Configuration VIllB
(blocked): ••••••• 20,280 bytes.
Configuration VIllB
(unblocked): .••.•• 15,616 bytes.
Configuration VIIIB
(blocked) - Hypertape: ••••••••••• 20,280 bytes.
10.0
4
2
~
1.0
1
.JII'
4
(
"'-
..
Time in Minutes to
Process 10,000
Master File Records
.-
VUB
-...""
J'
lL_ 1 - - -
.-.:---
\flUB
-
./
L __ -------------VIIIB
...... --
2
-.--
VIIIB (HYPER TAPE)
O. 1
1
4
_ Cll -
---
....
Cl'-
-
.,,/
2
~/
ifu
0.01
,
1
I
I
4
I
I
2
0.001
0.0
0.1
o .33
1.0
Activity Factor
Average Number of Detail Records Per Master Record
(Roman numerals denote standard System Configurations.)
-
-
©
LEGEND
Elapsed time; unblocked Files 3 & 4
- - - - - - _ Elapsed time; blocked Files 3 & 4
CP - Central Processor time (all configurations)
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
428 :201.120
• 12
IBM SYSTEM/360-MODEL 75
Standard File Problem B
. 122 Computation: ••••..• standard, using fixed-point
binary or decimal
aritlunetic •
.123 Timing basis: •.•..• using estimating procedure
outlined in Users' Guide,
4:200.12.
.124 Graph: ..•...••..• see graph below.
. 121 Record sizesMaster file: . . . • . . 54 data characters, packed
as 44 8-bit bytes.
Detail file: ••••.•• 1 card.
Report file: ••.••• 11ine.
10.0
7
4
2
7
4
2
"'"
./
~
..""
.",
V_ 10----7
L___
;
0.1
---
-- - ---
--
~.-
Cl'-
VV~'
0.01
4
~~
.-
VIIIB
"
I----~
pf..)
l'f..R'tt>.
'11.\.11'> t1:l!..
10--- ~---
2
--
//
./,- i""
7
\T11,B_-
.-
7
4
------
~1?!
~
./
Time in Minutes to
Process 10,000
Master File Records
-
~
1.0
,.
I
If
2
0.001
0.0
0.1
0.33
1.0
Activity Factor
Average Number of Detail Records Per Master Record
(Roman numerals denote standard System Configurations.)
LEGEND
_ _ _ _ _ _ _ _ _ _ _ Elapsed time; unblocked Files 3 & 4
_' _ _ _ _ _ _ _ Elapsed time; blocked Files 3 & 4
CP Central Processor time (all configurations)
(Contd.)
8/65
SYSTEM PERFORMANCE
• 13
428:201.130
.132 Computation:
•..••. standard, using fixed-point
binary or decimal
arithmetic.
. 133 Timing basis: •••••• using estimating procedure
outlined in Users' Guide
4:200.13
.134 Graph: •.••..••.• see graph below.
Standard File Problem C
• 131 Record sizesMaster file: .••••• 216 data characters, packed
as 176 8-bit bytes.
Detail file: •••.•. 1 card.
Report file: •••••• 1 line.
4
-
2
4
VIIIB----------i
----------
2
~~~~~~~~~4-~~_VIITB(HYPERTAPE) __________~
-
Time in Minutes to
Process 10,000
Master File Records
--r-.-----
~--------
........ --
.----
4
2
0.01
,
7
4
1
.'
I
2
0.001 -+----~----------~--------------------~----------~
0.0
0.1
0.33
1.0
Activity Factor
Average Number of Detail Records Per Master ,Record
(Roman numerals denote standard System Configurations. )
'LEGEND
- - - - - - - - - - _ Elapsed time; unblocked Files 3 & 4
_ _ _ _ _ _ _ _ Elapsed time; blocked Files 3 & 4
- _
CP Central Processor time (all configurations)
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
428:201.140
• 14
IBM SYSTEM/360-MODEL 75
Standard File Problem D
• 141 Record sizes Master file: •••••• 108 data characters, packed
as 88 8-bit bytes.
Detail file: ••••••• 1 card.
Report file: •••••• 11ine.
.142 Computation: •....•• trebled, using fixed-point
binary or decimal
arithmetic •
.143 Timing basis: ..•..• using estimating procedure
outlined in Users' Guide,
4:200.14.
.144 Graph: •••••.•.•• see graph below.
10.0
7
4
2
1.0
----
7
1--",.4-VUll-.
4
./
~
~
./
Time in Minutes to
Process 10,000
Master File Records
--
__
VIIIB
---~-------...."""..
2
-
,.,--
VIIIB (HYPERTAPE)
0.1
7
4
_cl'
---.
--
2
0.01
I
7
4
.r
I
-'
2
0.001 -+----+-----------r-------~----------------------~
0.0
0.1
0.33
1.0
Activity Factor
Average Number of Detail Records Per Master Record
(Roman numerals denote standard §ystem Configurations. )
LEGEND
-
-
- - -
-
- CP -
Elapsed time; unblocked Files 3 & 4
Elapsed time; blocked Files 3 & 4
Central Processor time (all configurations)
(Contd.)
8/65
SYSTEM PERFORMANCE
428:201.200
.2
SORTING
.21
Standard Problem Estimates
.212 Key size: . . . . . . . . . 8 characters .
• 213 Timing basis: • . . . . . using estimating procedure
outlined in Users' Guide,
4:200.213
.214 Graph: . . . . . . . . . . . see graph below .
• 211 Record size: •••.••• 80 characters.
100.0
7
4
2
II
~ ~Ii
V~
10.0
7
/
4
~/ /
~~JI
2
Time in Minutes
to Put Records
Into Required
Order
~~
1;1
1. 0
,
7
i\
L L
~~
4
/ V
//
2
/
0.1
,,
7
,
/
"
II
/7 f.-~
~
~
.L'I'
~.q,
;J1
,
)'1'
1I
~
V
4
2
0.01
100
2
4
7
1,000
2
4
7
10,000
2
4
7
100,000
Number of Records
(Roman numerals denote standard System Configurations.)
©
1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
428:~Ql.220
• 22
IBM SYSTEM/360-MODEL 75
Operating System/360 Sort/Merge Times
.223 Timing basis: ..•..• IBM Form C28-6543; inputoutput blocking factor s as
indicated in legend below .
. 224 Graph: .••..•.... See graph below •
• 221 Record Size: ••••••• 80 characters.
• 222. Key size: • . . • . • • •. 8 characters.
100.0
7
4
2
10.0
7
Ii
1lV.
~~
4
A
2
~ VII"
~~
Time in Minutes
to Put Records 1.0
Into Required
Order
7
~
4
~
....
~ ~ ~ 1/II~
Ioool!:l
~
9..--
.-
"""'"
'""
2
0.1
7
4
2
0.01
100
2
4
7
1,000
2
4
7
10,000
2
4
7
100,000
Number of Records
Curve
Configuration
I/O Blocking
A
Standard Configuration VTIB
90 records/block
B
Standard Configuration VTIIB
90 records/block
C
Standard Configuration VTIIB
with Hypertape
"90 records/block
/
D
E
Model 75 with one 1302 Disk
Storage Unit
40 records/block
Model 75 with two 2301 Drums
40 records/block
(Contd.)
8/65
SYSTEM PERFORMANCE
428:201.300
.3
MATRIX INVERSION
• 31
Standard Problem Estimates
.312 Timing basis: . • . . . . using estimating procedure
outlined in Users' Guide,
4:200.312 .
.313 Graph: • • . . . . . . . • . see graph below .
• 311 Basic parameters: ••. general, non-symmetric
matrices, using floating
point to a precision of
approximately 6 decimal
digits in the SHORT format and 16 digits in the
LONG format.
1.0
7
/1
4
/I
II
2
IlL
0.1
7
IJ
4
VII
~~N
2
o~§
II ~
Time in Minutes
for Complete
Inversion
0.01
7
I
I
III
4
rl
h
2
f
0.001
I
7
(
I
/I
\
II
4
II
II
r
2
0.0001
1
2
4
7
10
2
4
7
100
2
4
7
1,000
Size of Matrix
~~
..
(
"'-.
© 1965 AUERBACH Corporation and AUERBACH Info, Inc.
8/65
428:201.400
IBM SYSTEM/360-MODEL 75
.4
GENERALIZED MATHEMATICAL PROCESSING
.41
standard Mathematical Problem A Estimates
.412 Computation:
• 411 Record sizes: •.•••• 10 signed numbers; average
size 5 digits, maximum
size 8 digits.
....... 5 fifth-order polynomials,
5 divisions, and 1 square
roet; computation is in
"long" floating-point
mode (16-digit precision) •
.413 Timing basis: ••..•.• using estimating procedure
outlined in Users' Guide,
4:200.413 .
• 414 Graph: •••••..•.. see graph below.
1,000.0
7
\.
4
2
100.0
7
4
)",
2
./
~.
VllB
Time in
10.0
Milliseconds
per Input Record 7
7
,
VIIIB
/
v~
4
I
/'11
2
..J~ ~
1.0
CP
7
1t"'.1!. c~c~
l ~ \)'\:
~ 1.
cP ~1t"'()' ,
.
4
2
0.1
2
0.1
4
7
2
1.0
4
2
7
10.0
C, Number of Computations per Input Record
(Roman numerals denote Standard Configurations;
R = number of output records per input record;
Curve marked "CP" shows Central Processor time.)
8/65
4
7
100.0
(
\
IBM SYSTEM/360
MODEL 44
International Business Machines Corp.
AUERBACH INFO, INC.
PRINTED IN U. S. A.
IBM SYSTEM/360
MODEL 44
International Business Machines Corp.
AUERBACH INFO, INC.
PRINTED IN U. S. A.
1.
IBM SYSTEM/360
MODEL 44
INTRODUCTION
/&EDP
I
"--
435:011. 100
ST"'ARD
...
AUER8AC~
R[PORTS
~
INTRODUCTION
The Model 44 is a special-purpose scientific data processing system that was officially
added to the IBM 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 $5,215 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 are expected
to begin 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 instruction repertoire is a fully compatible subset of the full System/360 repertoire.
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.
The Model 44 Processor can contain 32, 76S, 65,536, 131,072, or 262, 144 bytes (SK
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 IBM 2315 Disk Cartridge that can store up to 1, 171, 200 bytes of data.
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 4S, 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 and described in Section 420:111. The standard Multiplexor
Channel provides up to 64 subchannels, and each of the High-Speed Multiplexor Channels provides
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
IBM SYSTEM/36(}-MODEL 44
435:0 II. 10 I
up to 4 subchannels, setting the theoretical limit of simultaneous input-output data transfer
operations at 72. No Selector Channels are currently 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 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 IBM 7340 Hypertape Drive, but can use both the 800
and 1,600 bpi models of the 2400 Series Magnetic Tape Units. Display devices (Models 2250 and
2260) and data communications devices (2701 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 chaIlI).els.
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 not compatible with the extensive array of standard System/360 software. Designated the Model 44 Programming System, the software centers around a Disk
Resident Monitor program that controls the sequential execution of batched jobs in a nonmultiprogramming environment. Input-output device control routines are also included within
the Monitor. A disk-resident FORTRAN IV compiler with extensive capabilities, an assembler,
and an array of disk-resident utility routines are also provided. The FORTRAN and Assembly
languages are directly compatible with the same languages offered in IBM's Basic Programming
Support (BPS) package of the standard System/360 software. (Three independent card/tapeoriented programs that do not utilize the system disk can be used instead of the Model 44 Programming Support, providing basic FORTRAN language, Assembly language, and data transcription
facilities. )
5/66
A
AUERBACH
®
435:031. 001
A
AUERBACH
STANDAilD
IBM SYSTEM/360
MODEL 44
SYSTEM CONFIGURATION
EDP
REPORTS
~
SYSTEM CONFIGURATION
A Model 44 system consists of a Processing Unit
and from 8,192 to 65,536 four-byte words of
internal core storage. A Single Disk Storage
Drive (Section 435:041) is provided as standard
equipment, and is housed in the Processing Unit's
qabinet. A second Single Disk Storage Drive can
be installed as an optional feature. A Model 44
system also includes a console printer-keyboard as
standard equipment.
•
Peripheral devices are connected to a Model 44 system via a Multiplexor Channel and/or one or two
High-Speed Multiplexor Channels. A minimum of
one channel of either type must be included in every
Model 44 system. A Direct Data Channel can be
included instead of the second High-Speed Multiplexor Channel to permit the connection of up to
eight "foreign" I/O devices, i. e., devices not
manufactured by IBM. Table I summarizes the
input-output channel possibilities in a Model 44
system. Section 435:111, Simultaneous Operations,
provides detailed information on the characteristics
and performance capabilities of the three varieties
of Model 44 I/o channels.
All of the peripheral devices currently available
with the System/360 Model 44 are listed in Table II,
together with the rules governing their connection
to the system. Mix possibilities and total number
of devices per channel and per overall system are
also explained.
One 2315 Disk Cartridge.
•
One Multiplexor or High-Speed Multiplexor
Channel.
•
One console printer-keyboard for output
(standard software requires the use of one
printer or magnetic tape unit for output).
As noted in Table II, Peripheral Devices, a Model
44 system can be connected to an IBM 1800 Data
Acquisition and Control System via control units
connected to a subchannel of either a Multiplexor or
High-Speed Multiplexor Channel. The IBM 1800 is
a specialized computer system designed specifically
for process control and high-speed data acquisition
systems.
Representative standard configurations of the
System/360 Model 44 are presented on the following
pages, with itemized monthly rental prices for
each configuration. The configurations follow
the standards and specifications explained in the
Users' Guide, Section 4:031.
One 2044 Processing Unit, Model E, F,G,
or H (including one Single Disk Storage Drive).
•
1442 Model N1 Card Read Punch;
2501 Model B1 or B2 Card Reader;
2520 Model B1 Card Read Punch;
2540 Model 1 Card Read Punch;
2401/2402 Model 1, 2, 3, 4, 5, or 6
Magnetic Tape Unit with 9-track
capability;
2403/2404 Model 1, 2, 3, 4, 5, or 6
Magnetic Tape Unit with 9-track
capability.
Multiple Mode144 Processing Units can at present communicate with each other only through
shared input-output devices. However, a Model
44 processor can be connected directly to any other
model of the System/360 that has the Channel-toChannel adapter feature. The adapter uses one
control-unit position and one sub channel of a
Multiplexor or High-Speed Multiplexor channel.
The minimum equipment requirements for a valid
Model 44 system include:
•
One input unit from among the folloWing:
TABLE I: SYSTEM/360 MODEL 44 I/O CHANNELS
Channel Type
Maximum
Number
(1)
Multiplexor
High-Speed Multiplexor
Direct Data (2)
Notes:
(1)
(2)
(3)
(4)
(5)
1
2
1
Subchannels
per Channel
(5)
32 or 64 (4)
1 to 4 (3)
0
Control Unit
Positions per
Channel
!~
8
2 to 8
-
None of the I/O channels is standard equipment. Up to
three I/O channels can be used with a Model 44 system.
To use the Direct Data Channel, the first High-Speed
Multiplexor Channel must also be installed, and the
second High-Speed Multiplexor Channel is prohibited.
A High-Speed Multiplexor Channel provides a single
subchannel with two control unit positions as a standard
feature. Three additional sub channels per High-Speed
Multiplexor channel can be prOVided, each with two control unit positions.
The Model E (32, 768 bytes of core storage) Processing
Unit provides control of 32 subchannels; Models F, G,
and H provide control of 64 subchannels.
Up to 8 Multiplexor subchannels and all of the HighSpeed Multiplexor subchannels can operate as shared
subchannels, i. e .• can control as many as 16 I/o units
(only one of which can transfer data at a given time).
© 1966 AUERBA~H Corporation and AUERBACH Info, Inc.
5/66
435:031. 002
IBM SYSTEM/360
TABLE II: SYSTEM/360 MODEL 44 PERIPHERAL DEVICES
Multiplexor
Subchannels Used
per Controller
Maximum
Controllers
per System
Maximum Peripheral
Devices per Controller
Peripheral Device
Controller
I/O Channel
Requirements
Random Access
2311 Disk Drive
Single Disk Storage Drive
2841
special adapter
MPX, HSMPX
MPX, HSMPX
IS
1
22
2
8
1
Console
Console, Printer-Keyboard
special adapter
MPX, HSMPX
1
1
1
2403, 2404,
2803, or 2804
MPX, HSMPX
IS per 2403 or 2803
2S per 2404 or 2804
22 (2403/2803)
11 (2404/2804)
8
self-contained
MPX, HSMPX
1
22
1
self-contained
MPX, HSMPX
1
22
1
self-contained
MPX, HSMPX
1
22
1
2821+
MPX, HSMPX
1
22
1
\
Mal?;netic TaEe
2401 and 2402 Magnetic
Tape Units
Punched Card
2501 Card Reader
Models B1, B2
2520 Card Read Punch
(
Model B1
2520 Card Punch
Models B2, B3
2540 Card Read Punch
Modell
1442 Card Read Punch
Model N1
1442 Card Punch
Model N2
Punched Paper Tape
2671 Paper Tape Reader
Modell
Printers
l403Printer
Models 2, 3, 7, N1
1443 Printer
Model N1
self-contained
MPX, HSMPX
1
22
1
self-contained
MPX, HSMPX
1
22
1
2822
MPX, HSMPX
1
22
1
2821+
MPX, HSMPX
1 per 1403
22
3
self-contained
MPX, HSMPX
1
22
1
self-contained
MPX, HSMPX
IS
22
1
2840·
MPX, HSMPX
IS
22
8
2840·
2840·
2840·
MPX, HSMPX
MPX, HSMPX
MPX, HSMPX
IS
1S
IS
22
22
22
4
4
4
2848
2848
2848
MPX, HSMPX
MPX, HSMPX
MPX, HSMPX
25 or 2S
17 or IS
9 or IS
22
22
22
24
16
8
self-contained
MPX, HSMPX
4 max.
22
up to 4 lines
self-contained
MPX, HSMPX
31 max.
22
up to 31 lines
special adapter
MPX, HSMPX
1
22
1
5
22
1
DisEla~
Units
2250 Display Unit
Modell
2250 Display Unit
Model 2
2280 Film Recorder
2281 Film Recorder
2282 Film Recorder/
Scanner
2260 Display Station:
Modell
Model 2
Model 3
Data Communications
2701 Data Adapter
Unit
2702 Transmission Control
Data Acguisition!Process
Control
1801 or 1802 ProcessorController
1827 Data Control Unit
self-contained
+ A single 2821 Control Unit can also control a combination of one 2540 Card Read Punch and one
1403 Model 2 or 3 Printer, using 3 subchannels; or, a combination of one 2540 Card Read Punch
and two 1403 Printers (Models 2 and/or.7), using 4 subchannels.
• A single 2840 Display Control Unit can also control any mixture of up to 4 Film ·Units. Combinations of 2250 Model 2 Display Units and 2280/2281/2282 Film Units can also be handled by a
2840 Control Unit in the following mixes: 1, 2, or 3 Film Units can be used with 5, 4, or 1
Display Units, respectively.
/"
Legend
MPX = Multiplexor Channel.
HSMPX = High-Speed Multiplexor Cbannel.
IS = One shared Multiplexor or High-Speed Multiplexor subchannel.
,/
5/66
A
(Contd.)
AUERBACH
'"
MODEL 44: SYSTEM CONFIGURATION
.1
435:031. 100
6-TAPE AUXILIARY STORAGE SYSTEM; CONFIGURATION V
Deviations from Standard Configuration: . . • . . . . • . 20,000 more bytes of core storage.
printer is 20% faster.
reader is 20% faster.
Equipment
Rental
Main Storage (32,768 bytes)
2044 Processing Unit, Model E 44
$ 3,465
Single Disk Storage Drive
(1,171,200 bytes)
Console Printer-Keyboard
~.
.
2501 Card Reader, Model B1:
Reads 600 cards per minute
260
1442 Card Punch, Model N2:
Punches 91 full cards per minute
375
2821 Control Unit, Model 2
1403 Printer, Model 7:
Prints 600 lines per minute
600
650
2403 Magnetic Tape Unit and Control
2402 Magnetic Tape Units (2)
2401 Magnetic Tape Unit
(6 drives total)
All Modell: 30,000 bytes/sec .
2841 Storage Control and three 2311
Disk Storage Drives (21. 75 million
bytes)
Optional Features Included: . . . • . . . . . . . . . . . . . . External Interrupt
Multiplexor Channel
High-Speed Multiplexor Channel
with two subchannels
TOTAL:
(
2,460
2,250
30
350
775
$11,215
NOTE: Inclusion of the High-Speed General Registers feature ($700) and the Floating-Point
Arithmetic Feature ($275) increases the monthly rental price to $12,190.
(
\.
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
IBM SYSTEM/360
435:031. 200
.2
6-TAPE BUSINESS/SCIENTIFIC SYSTEM; CONFIGURATION VI
Deviations from Standard Configuration: • . . . • . . . . printer is 20% faster.
reader is 20% faster.
Equipment
Rental
Main Storage (65,536 bytes)
2044 Processing Unit, Model F 44
Single Disk Storage Drive
(1,171,200 bytes)
$ 4,665
/
Console Printer-Keyboard
2501 Card Reader, Model B1:
Reads 600 cards per minute
260
1442 Card Punch, Model N2:
Punches 91 full cards per minute
375
2821 Control Unit, Model 2
1403 Printer, Model 7:
Prints 600 lines per minute
600
650
2403 Magnetic Tape
2402 Magnetic Tape
2401 Magnetic Tape
(6 drives total)
All Modell: 30,000
Unit and Control
Units (2)
Unit
2,460
bytes/sec.
Optional Features Included: . . . . . . . . . . . . . . . . • . External Interrupt
Multiplexor Channel
High-Speed Multiplexor Channel
with one subchannel
Floating-Point Arithmetic
TOTAL:
30
350
650
275
$10,315
NOTE: The total rental using the High-Speed General Registers feature is $11,015.
/
"
(Contd. )
5/66
A
AUERBACH
'"
435:031. 300
MODEL 44: SYSTEM CONFIGURATION
.3
10-:-TAPE GENERAL SYSTEM, INTEGRATED; CONFIGURATION VIlA
Deviations from Standard Configuration: . . . . . . . . . printer is 20% faster.
reader is 20% faster.
Equipment
Rental
Main Storage (65,536 bytes)
2044 Processing Unit, Model F 44
$ 4,665
Single Disk Storage Drive
(1,171,200 bytes)
Console Printer-Keyboard
2501 Card Reader, Model Bl:
Reads 600 cards per minute
260
1442 Card Punoh, Model N2:
Punches 91 full cards per minute
375
2821 Control Unit, Model 2
1403 Printer, Model 7:
Prints 600 lines per minute
600
650
2403 Magnetic Tape Unit and
Control (2)
2402 Magnetic Tape Units (4)
(10 drives total)
5,750
All Model 2: 60,000 bytes/sec.
Optional Features Included: . . • . . . . . . • . . . . . . . . External Interrupt
Multiplexor Channel
Higlr-Speed Multiplexor
Channel with two subchannels
Floating-Point Arithmetic
TOTAL:
30
350
775
275
$13,730
NOTE: The total rental using Model 3 tape drives (90,000 bytes/second) and the High-Speed General
Registers feature is $17,430.
(
"-
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
435:031. 400
.4
IBM SYSTEM/360
4-TAPE SCIENTIFIC SYSTEM; CONFIGURATION XI
Deviations from Standard Configuration: • . • . . . . • . printer is 140% faster.
reader is 20% faster.
punch is 119% slower.
magnetic tapes are 100% faster.
Equipment
Rental
Main Storage
(65,536 bytes)
2044 Processing Unit, Model F 44
$ 4,665
Single Disk Storage Drive
(1, 171, 200 bytes)
Console Printer-Keyboard
2501 Card Reader, Model B1:
Reads 600 cards per minute
260
1442 Card Punch, Model N2:
Punches 91 full cards per minute
375
1443 Printer, Model N1:
Prints 240 Lines per minute
875
2403 Magnetic Tape Unit and
Control
2402 Magnetic Tape Unit
2401 Magnetic Tape Unit
(4 drives total)
All Modell: 30,000 bytes/sec.
Optional Features Included: . . . • . . • . . . • . . . . • . . External Interrupt
Multiplexor Channel
High-Speed Multiplexor
Channel with one sub channel
Floating Point-Arithmetic
TOTAL:
NOTE: The total rental using the High-Speed General Registers feature is $10,020.
5/66
A
AUERBACH
1,840
30
350
650
275
$ 9,320
./
435:041. 100
A
AUERBACH
STANDARD
EDP
IBM SYSTEM/360
MODEL 44
DISK DRIVE
REPORTS
INTERNAL STORAGE: SINGLE DISK STORAGE DRIVE
,
(
.1
GENERAL
. 11
Identity: ..
IBM System/360 Model 44
Single Disk Storage Drive.
IBM 2315 Disk Cartridge.
. 12
Basic Use: .
System disk and auxiliary
working storage.
.13
Description
The dual read/write heads (one head per surface)
move in unison directly to the addressed track
under control of a single Control Seek command.
The two tracks that can be read and recorded in
each of the 200 positions of the access arm are
considered a logical cylinder of data. The specialized Read Data and Write Data instructions can
store and fetch either a full track or a specified
number of sectors. Error detection is provided
by regenerating a check code when reading data
and comparing this code to the check code that was
generated and written with the data during the
write operation.
The Model 44 Processing Unit houses, as standard
equipment, a Single Disk Storage Drive with a
removable IBM 2315 Disk Cartridge. The on-line
storage capacity is 1, 171,200 bytes, and this
storage is used for residence of programming systems, for auxiliary processor storage, and for
standard input-output operations. A second Single
Disk Storage Drive of equal storage capacity and
identical functional characteristics can be added to
the Model 44 Processing Unit as an optional feature. A complete description of the Model 44
Processing Unit is provided in Section 435:051.
The Single Disk Storage Drive is connected to one
of the input-output channels by means of a Multiplexor or High-Speed Multiplexor Channel adapter
that is installed as standard equipment within the
Model 44 processor. The Disk Storage Drive takes
up one control unit position and one sub channel of
either a Multiplexor or a High-Speed Multiplexor
channel. If the optional second Single Disk Storage
Drive is installed in the processing unit, it connects
to the same sub channel as the standard Single Disk
Storage Drive. When both drives are installed, a
seek operation on one unit can be overlapped with a
seek, read, or write operation on the other unit
because the channel is released immediately after
. the seek operation is initiated.
The data transfer rate of the Single Disk Storage
Drive is 90,000 bytes per second; the unit always
operates in burst mode over a Multiplexor or
High-Speed Multiplexor Channel. The average
access time to randomly-addressed data is 70
milliseconds.
A single recording disk is permanently contained
in each IBM 2315 Disk Cartridge. Cartridges are
exchangeable by the operator through a side access
slot in the cabinet of the processor.
No programming compatibility exists between the
Single Disk Storage Drive and the other System/360
mass storage units that are controlled through the
IBM 2841 Storage Control Unit.
Information is written on or read from the magnetic
disk by a pair of read/write heads, one servicing
each disk surface. Each surface is divided into
eight sectors, and the effective storage capacity of
each sector is fixed at 366 data bytes. Therefore,
the storage capacity of each disk track (and the
maximum data load that can be transferred by a
single instruction) is 2,928 bytes. Since 400
tracks are available on the single disk, a total
storage capacity of 1, 171,200 bytes of data is
provided.
Of the 1, 171, 200 bytes of information that can be
stored in a single IBM 2315 Disk Cartridge, approximately 500, 000 bytes will be required for residence
of the Model 44 Programming Support package. See
Section 435:151 for descriptions of the control programs, language processors, and utility routines
that comprise the software support for the Model
44 system.
. 14
Availability: . . . . .
not specified by IBM .
. 15
First Delivery: ...
3rd quarter 1966.
(
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
/--
435:051. 100
4
STANDARD
EDP
AUERBACH
IBM SYSTEMi.360
MODEL 44
CENTRAL PROCESSOR
REPORTS
CENTRAL PROCESSOR
GENERAL
.1
.11
Identity:
.12
Description
IBM 2044 Processing Unit.
The 2044 Processing Unit provides arithmetic,
logic, and control functions, and internal core and
disk storage for the IBM System/360 Model 44
system. The 2044 Processing Unit uses highspeed internal circuitry to implement four-byte
(one-word) data flow and processor control. The
basic System/360 processor design has been retained in the Model 44 processor, but modifications
have been made to optimize performance of scientific and data acquisition tasks.
The Model 44 processor uses the same basic data
and instruction formats as the System/3GO Processor Models 30, 40, 50, 65, and 75. The instruction repertoire is an upward-compatible subset of
the full System/360 repertoire. The Model 44 instruction set can optionally include floating-point
arithmetic, but has no decimal arithmetic capabilities.
The 2044 Processing Unit is available in four
models, E through H, providing core storage capacities of 32,768, 65,536, 131, 072, or 262,144
8-bit bytes (SK, 16K, 32K, or 65K 32-bit binary
words). The 32K-byte Model E processor can be expanded to a Model F, G, or H processor through field
modification. Four bytes in parallel are read from
or written into core storage during each one-microsecond cycle. A parity bit is associated with each
8-bit byte, and parity is checked during each data
fetch from core storage.
Like the System/360 Models 30 through 75, the
Model 44 processor contains sixteen 32-bit general
registers that can be used as fixed-point accumulators or as index registers. The general registers
of the Model 44 processor are normally implemented in a non-addressable extension of core
storage with a cycle time of one microsecond per
word. The optional High-Speed General Registers
(Feature 4583) can be supplied instead of the standard general registers, providing read-write
times of 250 nanoseconds per 4-byte word. The
High-Speed General Registers are implemented in
"solid logic technology" (SLT) circuitry. Byproviding general registers that are four times faster
than the standard Model 44 general registers, this
feature substantially reduces the generation time of
all addresses and the basic execution time of all
fixed-point instructions in particular. See Paragraph .4, Processor Speeds, for a performance
comparison between a Model 44 processor equipped
with the standard general registers and one equipped with the optional High-Speed General Registers.
I
(
\"
If the optional Floating Point Arithmetic feature is
installed, four 64-bit floating-point registers are
supplied. Short 32-bit floating-point operands
utilize the high-order halves of these registers,
implemented in high-speed circuitry. Long 64-bit
floating-point operands use the high-speed halves
of the floating-point registers for the high -order
pOSitions of the operands, and an extensIon of
processor storage as the registers for the loworder bits of the operands.
In the Model 44 processor, the fractional parts
(mantissas) of long floating-point operands can be
of varying lengths. By means of a console switch
on the processor, the user can ad.iust long-precision
instructions for execution with 32, 40, 4,'!, or the
full 56 bits of precision in the floming--point fraction. For maximum speed of exeeution of long
floating-point instructions, the switch is set to
truncate the fraction to 32 bits of pl'ecil:lion, providing floating-point precision that is still greater
than that provided by the 24-bit fraetional precision
of short floating-point operations. The execution
times of long floating-point instruetiolls increase
as greater fractional preci.sion is selected. The
full 56-bit precision must be used if long floatingpoint operations are expected to produce identical
answers when performed on a System/360 Model
44 processor and the other System/360 processors
that are capable of floating-point arithmetic.
_121 Model 44 Instruction Set
The standard Model 44 instruction set contains about
three-fourths of the standard System/360 instructions
listed in Section 420:121; all decimal arithmetic
and most variable field-length instructions are excluded, and floating-point al·ithmetic instructions
(44 in number) are provided only as an extra-cost
option. The basic arithmetic mode is fixed-point
binary, using 32-bit operands and two's-eomplement
notation. Sinee the Medel 44 processor is optimized
toward scientific processing, the common data base
is the 32-,bit word and the 16-bit halfword, composed
of four and two bytes, respectively. However, a
few instructions are provided to manipulate bytes
Singly, such as the Insert Character, Store Character, Test Under Mask, and Compare Logical instructions. Binary to decimal and decimal to
binary radix conversion instructions are not provided (though their presence is especially desirable
(J
in computers that cannot perform decimal arithmetic),
nor are any code translation instructions supplied.
All Storage-to-Storage (Type SS) instructions ar.e
excluded, including all forms of the Move instruction.
Another notable exclusion from the Model 44 instruction set is the Load Multiple and Store Multiple instructions that are designed to load and store designated groups of the 16 System/360 general registers
with one instruction. Effective control of multiprogramming operations is facilitated in other
System/360 models through use of these Load and
Store Multiple instructions. However, the Model
44 processor has not been designed for multiprogramming; it is advertized as a high-speed sequential
batch processor.
Instructions can be two or four bytes in length. A
2-byte instruction causes no reference to main
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
435:051. 121
IBM SYSTEM/360-MODEL 44
. 121 Model 44 Instruction Set (Contd.)
second. It has a replaceable printing head and a
typewriter-style keyboard. The Model 44 console
device takes up one control-unit position and one
subchannel of either a Multiplexor Channel or HighSpeed Multiplexor Channel.
storage and a 4 -byte instruction causes one reference to main storage. There are four basic instruction formats:
•
Type RR - Register to Register (2 bytes)
•
Type RX - Register to Indexed Storage (4 bytes)
The functions and operations of the console printerkeyboard are the same as those described for the
IBM 1052 Printer-Keyboard, Modell, in Report
Section 420:061.
The System/360 Model 44 provides a system control
panel on the processor for use by the operator and
the IBM customer engineer. This system control
panel visually and functionally resembles the
System/360 control panel described in Section
420:061.
Op
•
Type RS - Register to Storage (4 bytes)
.125 Channel Capabilities: .. see Simultaneous Operations,
Section 435:111.
Op
. 126 Interrupt System
•
The basic Interrupt System of the System/360
Model 44 Processing Unit is essentially the same
as that used with the System/360 processor models
30 through 75. See Paragraph 420:051. 123 for a
detailed description of the five classes of System/
360 interrupts.
Type SI - Storage and Immediate Operand
(4 bytes)
Op
I
B = 4-bit base register specification.
D = 12-bit displacement.
I = 8-bit literal operand.
Op= 8-bit operation code.
R = 4-bit operand register specification.
X = 4-bit index register specification.
. 127 Optional Features
The Store and Fetch Protection feature provides
both read and write protection for the contents of
specified 2, 048-byte blocks of core storage. The
protection is achieved by identifying blocks of
storage with a 4-bit "storage key" and comparing
this key with a 4-bit "protection key" associated
with the data to be fetched or stored. If the two
keys match, or if the storage key is zero, the data can
be read from or written to core storage. The
write-protect portion of this feature is identical to
that available with System/360 Models 30, 40, 50,
65, and 75. Two additional instructions are provided with this feature: Set Storage Key (SSK) and
Insert Storage Key (ISK).
.122 Compatibility
A machine-language program written for a System/
360 Model 44 and not implicitly dependent upon internal machine timing can be executed, and will
produce identical results, on System/360 Models 30
through 75 that are similarly equipped. Machinelanguage compatibility in the reverse direction is
generally not possible since the Universal System/
360 instruction set is considerably larger than that
provided with the Model 44. Model 44's Assembly
and FORTRAN languages are unilaterally compatible
with the Assembly and FORTRAN languages provided for use with the System/360 Basic Programm
ing Support software.
By making the core storage read-protect option
available for the Model 44, IBM has provided an
essential hardware feature for proceSSing in realtime and time-sharing environments. The Model
44 is emphasized by IBM as being particularly
suitable for real-time processing, but the software
support for the Model 44 system apparently has no
special provisions to facilitate processing in timesharing environments.
.123 IBM 2315 Disk Cartridge
A single-disk storage drive is supplied as standard
equipment with the 2044 Processing Unit for programming system r,esidence. It is built into the
processor cabinet, and the single-disk 2315 Cartridge can be interchanged with other cartridges by
the operator. The 2315 Cartridge can store up to
1,171,200 8-bit bytes. A second single-disk drive
of equal capacity to the first can be added as an
optional feature. Report Section 435:041 describes
the single-disk drive in more detail.
The High-Resolution Interval Timer feature occupies three bytes of core storage and one byte of
a special register. This register decrements
every 13 microseconds in its rightmost position.
This feature is an extension of the standard Model
44 interval timer.
The Priority Interrupt feature permits quick responses to changes in external conditions, improving the real-time processing efficiency of the
System/360 Model 44. Thirty-two levels of
priority interrupts are provided, and each main
level has 256 sublevels. Thus a total of 8,192 external lines can directly and independently initiate
processor interrupts. Two additional instructions
are provided with this feature: Change Priority
Mask (CHPM) and Load PSW Special (LPSX).
.124 Console Printer-Keyboard
A console printer-keyboard is also supplied as
standard equipment with the System/360 Model 44.
This device provides for communication between
the operator and the system. Facilities are provided for interrupting the processing unit and for
signalling the end of the operator's data transmission. The keyboard unit (for input data) and
the printer (for computer output) are electrically
and mechanically independent of each other. The
printer-keyboard prints at 14.8 characters per
5/66
/
The External Interrupt feature provides six lines
for sending a single-priority external interrupt
A
(Contd. )
AUERBACH
'"
,/
/
CENTRAL PROCESSOR
435:051. 127
instructions are provided to move the device control word to and from storage: Read Direct Word
(RDDW) and Write Direct Word (WRDW).
.127 Optional Features (Contd.)
signal to the Model 44 processor. These interruptions are independent of data channel operations.
The Direct Word feature provides-for the transfer
of 32 bits (a full word) of information between an
external device and the processor's core storage on
a one-word-per-instruction basis. Two additional
.21
.13
Availability: . . . . . . . not specified by IBM.
.14
First Delivery:
.2
PROCESSING FACILITIES
. . . . 3rd qual·ter 1966.
Operations and Operands
Operation
and Variation
Provision
Radix
Size
automatic
binary
full or halfword.
automatic
automatic
binary
binary
halfword (32-bit product).
full word (64-bit product).
none.
automatic
binary
full word (64-bit dividend).
automatic
binary
Multiply:
automatic
binary
Divide:
automatic
binary
.213 Boolean AND:
Inclusive OR:
Exclusive OR:
automatic
automatic
automatic
binary
binary
binary
.211 Fixed point (binary only)
Add-subtract:
Multiply Short:
Long:
Divide No remainder:
Remainder:
.212 Floating point (optional)
Add -subtract: *
*
24
56
24
. 56
24
56
and
and
and
and
and
and
7
7
7
7
7
7
bits
bits
bits
bits
bits
bits
(short).
(long). +
(short).
(long). +
(short).
(long). +
1 word.
1 word.
1 word.
Both normalized and unnormalized.
+ Variable-precision long-length floating-point operations are possible. A console switch
sets the fraction length to 32, 40, 48, or 56 bits.
Provision
.214 ComparisonNumbers:
.215
.216
. 217
. 218
. 219
/
. 22
automatic
fixed-point binary
floating-point binary
Absolute:
Letters:
Mixed:
automatic
automatic
automatic
Collating sequence ASCII code: . . . . . .
Extended BCD code:
specials, numbers, letters.}
specials, letters, numbers.
Code translation: . . . .
Radix conversion: ...
Edit format: . . . . . . .
Table look-up: . . . . .
OthersBinary shift:
32
32
32
32
32
or 16 bits.
or 64 bits.
bits or up to 256 bytes.
bits or up to 256 bytes.
bits or up to 256 bytes.
see Data Code Tables,
Section 420:141.
none.
none .
none .
none •
automatic
Special Cases of Operands (binary only)
.221 Negative numbers: .•. 2 's complement and sign bit.
.222 Zero: ••.••...... only positive zero.
.223 Operand size determination Fixed size: ••••... halfword (16 bits), full
word (32 bits), or double
word (64 bits), implied
by instruction used.
Variable size: . . . . . 32-, 40-, 48-, or 56-bit
floating-point fraction,
set by a console switch.
binary
• 23
32 or 64 bits .
Instruction Formats
.231 Instruction structure: . 16 bits, or 32 bits for
instructions that reference storage .
. 232 Instruction layout
and parts: . . . . . . . see Paragraph 435:051.121.
. 234 Basic address
structure:. . . . . • .• • 2 + 0; variation in instruction length is due to the
fact that one operand may
be designated by either a
register address or a
main storage address.
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
IBM SYSTEM/360-MODEL 44
435:051. 235
.235 LiteralsArithmetic: ...••. none.
Comparisons and
tests (logical): ... 1 byte .
. Incrementing
modifiers: . . . . . . none; increment is either
-lor contained in a
register.
. 236 Directly addressed operands -
.2394 End value: .•••... implied as zero for Branch
on Count; for Branch on
Index, the value is in a
storage location specified
by the instruction.
.2395 Combined step and
test: • . . . • . . . . . yes.
. 24
Internal storage Minimum Maximum Volume
type
size
size
accessible
Core storage:
General
registers:
1.byte
1 register
256 bytes
16 registers
16,777,216
bytes*
16 one-word
registers
'* If base registers are used for relative addressing,
a maximum of 4, 096 bytes .is accessible via each
register so allocated.
.237 Address indexing .2371 Number of methods: . 2.
.2372 Names: . . • . . . . . . (1) indexing using the base
register addresses.
(2) indexing using the X
field (in instruction
format RX only);
permits double indexing if used with
method (1) .
. 2373 Indexing rule: . . . . . base address and index field
are treated as 24-bit positive binary integers; displacement is treated as
a 12-bit positive binary
integer. All these are
added to form a 24-bit
binary integer, ignoring
overflows.
.2374 Index specification: .. base address (B) field and
index (X) field both specify
the number of a register.
.2375 Number of potential
indexers: . . . . . . . 16.
.2376 Addresses which can be indexed Type of address
.2377
. 2378
. 238
. 239
.2391
.2392
.2393
5/66
Application
Storage reference: .. ail can be indexed by base
register contents.
Storage address in
RX instruction
format: ...••.•• can have double indexing (by
base register and index
register).
Cumulative indexing:. via double indexing .
Combine index and
step: . . . . . . . . . '. none .
Indirect addressing: .. none.
SteppingSpecification of increment: . . • . . . . . . . always minus one for
Branch on Count.
Increment sign: .. . . minus for Branch on Count;
minus or plus for Branch
on Index.
Size of increment: .• always one for Branch on
Count; 32 bits for Branch
on Index.
Special Process Storage
.241 Category of
storage
Number
of
locations
Size
in
bits
Program
usage
General
registers:
16
32
indexing, base
addresses, and
accumulators.
Floatingpoint
registers
(optional) :
4
64
floating-point
operations.
Program
Status
Word:
1
64
holds location
counter and
various flags.
Channel
Control
Word:
1
64
holds I/O control
information.
.3
SEQUENCE CONTROL FEATURES
· 31
Instruction Sequencing
.311 Number of sequence
control facilities: ..• 11 Program Status Words
(PSW), only one of which
is active at a time.
· 312 Arrangement: . . . . . . one PSW for initial program
loading, two for each of
the 5 types of interruption:
I/O, program, . supervisorcall, external,and machine
check.
.313 Precedence rule: . . . . priority of interrupts:
(1) machine check
(2) program or supervisorcall
(3) external
(4) I/O.
· 314 Special sub-sequence
counters: . . . . . . . . none.
.315 Sequence control step
size: . . . . . . . . . . . halfword.
.32
Look-Ahead: . . . . . . . none.
.33
Interruption (see also Paragraph 420:051. 123)
· 331 Possible causes Input-output units: .. unit available.
unit ceased transmission .
unit malfunction before
transmission starts.
Input-output controllers: . . . . . . . controller available •
controller ceased transmission (perhaps because of
error noted during transmission).
controller malfunction before transmission starts.
(Contd.)
fA
AUERBACH
'"
CENTRAL PROCESSOR
435:051. 332
Processor errors: ... illegal operation code.
operation code and data incompatible.
overflow, underflow, or
divide error.
all-zero floating-point
result.
operand incorrectly aligned.
violation of storage protection.
Other: . . • . . . . . . . . supervisory routine violation.
system malfunction.
external action from console
or another system.
. 332 Control by routine Individual control: . . . . . . . . . . acceptance or non-acceptance of I/O interrupts can
be controlled by channel.
length checks are controlled
by specific instruction.
Method: . . . • . . . . . specific bits in Program
Status Word. (These bits
are normally controlled by
the system, not by the
user's program.)
• 333 Operator control: . . . . operator may only initiate a
request for an external
interrupt.
.334 Interruption con(1) interruption condition
ditions: . . . . . .
signalled.
(2) interruption condition
attains the necessary
priority on a channel
to be forwarded to the
computer interface
(I/O interrupts only).
(3) this interrupt not
masked out by program or system
masks .
. 41 Instruction Times in Microseconds
(4) processor in a mode of
operation in which this
type of interrupt is
allowed to occur.
.335 Interruption process Interruption action: present PSW (Program
Status Word) is stored
and replaced by a standby
PSW.
Registers saved: ... none of the 16 general registers are saved automatically. Each must be stored
one at a time. Most of the
necessary operational data
is saved in the old PSW .
Destination: . . . . . . . contained in standby PSW;
one of 5 locations corresponding to the 5 levels of
interrupt conditions.
.336 Control methods Determine cause: ... analysis of flags by appropriate standard routines.
Enable interruption: . . . . . . . . . . by setting of bits in the PSW
or an I/O control word.
.34
Multiprogramming: .. no special provisions.
.35
Multisequencing: .... none.
.4
PROCESSOR SPEEDS
All System/360 Model 44 instructions use one or
more of the 16 general processor registers during
their execution. The standard Model 44 general
registers are in a non-addressable extension of
the I-microsecond main core storage. Optionally,
the High-Speed General Registers feature can be
installed, replacing the standard general registers
with 16 high-speed general registers with a cycle
time of 250 nanoseconds per word. Most Model 44
Processing Unit performance timings will vary
depending on whether the general registers reside
in core storage or in high-speed circuits.
.411 Fixed point (binary only) Add-subtract: ••••••••••
Multiply: ••••••••••••••
Divide: •......•••••.•.•
.412 Floating point -
Standard Registers
4.75
18.0*
32.75
Long +
Short
Add-subtract: •••••••••••
8.32
5:31
Multiply: •••••••••••••••
63.13
15.50
Divide: •....•...•.••.... 127.0
24.75
.413 Additional allowance for Single indexing: ••••••••• O. 0
Double indexing: •••••••• 1. 0 (standard registers)
or 0.75 (high-speed registers).
Indirect addressing: ••••• none.
Recomplementing: ••••••• none .
. 414 Control:
Compare Standard Registers
Fixed point: . . . . . . . . 4.0
Floating point (long): .. 7.35
Floating point (short): . 4.74
Logical: . . . . . . . . . . . 4.0
Branch: . . . . . . . . . . . . 2.75
*
High-Speed Registers
2.25
15.75 *
29.0
Long +
Short
-;r,-57
~6
62.38
126.25
14.75
24.00
High-Speed Registers
2.25
6.60
3.99
2.25
2.0
Average figure.
+ Based on full 56-bit preciSion for floating-point fractions.
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
435:051. 415
.415 Counter control Step: . . . • • . . . • . .
Step and test: . . . . .
Test . . . . . . . . . . .
.416 Edit: . . . • • . . . . . . .
.417 Convert: . . . • . . • . .
.418 Shift: . • . . . • . . . . .
IBM SYSTEM/360- MODEL 44
Standard Registers
High-Speed Registers
4.75
3.75
4.0
none .
none •
widely variable .
2.25
2.75
2.25
. 42 Processor Performance in Microseconds
.421 For random addresses High-Speed Registers
Standard Registers
Fixed Point
Floating
Floating
Floating
Fixed Point
Binary
Point (Long)
Point (Short)
Point (Long)
Binary
7.0
16.57
c = a + b: . . • . . .
13. U
11.56
18.82
7.0
16.57
b = a + b: . . . . . .
13.0
11.56
18.82
Sum N items: ...
4. 75(N)
2. 25(N)
7. 57(N)
5. 31(N)
8. 32(N)
20.50
71.38
c = ab: . . . . . . . . . . 26.25
21. 75
73.63
31. 00
33.75
135.25
c = alb: . . . . . . . . . 41.00
137.50
.422 For arrays of data 14.50
33.96
ci = ai + bj: . . . . . . . 26.25
38.21
.25.43
25.43
14.50
33.96
bj = ai + bf . . . . . . . 26.25
38.21
17. 89(N)
Sum N items: . . . . . 13. 25(N)
14. 37(N)
7.25(N)
20. 39(N)
27.50
89.59
c = c + ai bj: . . . . . . 39.0
35.18
92.84
.423 Branch based on comparison High-Speed Registers
Standard Registers
24. 75(N)
Numberic data: . . . . 43. 50(N)
24. 75(N)
Alphabetic data: . . . . 43. 50(N)
.424 -Switching 7.25
Unchecked: . . . . . . . 11.25
14.75
Checked: . . . . . . . . 21. 75
11. 0 + 11. 25(N)
List Search: . . . . . . 20.0 + 18. 5(N)
.425 Format control Unpack: . . . . . . . . . none.
Compose: . . . . . . . . none.
.426 Table look-up, per comparison High-Speed Registers
Standard Registers
10.25
For a match: . . . . . . 17.50
For least or
16.0
greatest: . . . . . . . 26.0
For interpolation
10.25
point: . . . . . . . . . 17. 50
.427 Bit indicators Set bit in separate
2.75
location:. . . . . . .. 3.50
2.25
Set bit in pattern:. .. 4.75
Test bit in separate
3.75
location: . . . . . . . . 5.25
3.5
Test bit in pattern: .. 5.0
4.75(B)
.428 Moving: . . . . . . . . . . 8. 25(B)
where B = number of bytes rnoved.
ERRORS, CHECKS, AND ACTION
.5
Action
.Check or Interlock
Overflow:
check
forced transfer to program
interrupt routine.
forced transfer to program
Underflow (floating-point):
check
interrupt routine.
forced transfer to program
Zero divisor:
check
interrupt routine.
forced transfer to program
Illegal data:
check
interrupt routine.
Forbidden operation:
check
forced transfer to supervisor
interrupt routine.
Unavailable operation:
check
forced transfer to program
interrupt routine.
Illegal storage address:
check
forced transfer to program
interrupt routine.
Receipt of data:
parity check
forced transfer to machine
error or input/output
interrupt routine.
Dispatch of data:
send parity bit.
5/66
A
AUERBACH
®
Floating
Point (Short)
9.31
9.31
4. 56(N)
19.50
28.75
21.18
21.18
11. 87(N)
31. 93
,/
435: 111. 100
!
~
~
AUERBACH
•
STAMDARD
EDP
IBM SYSTEM/360
MODEL 44
SIMULTANEOUS OPERATIONS
REPORTS
SIMULTANEOUS OPERATIONS
A System/360 Model 44 system can concurrently execute:
o
One machine instruction; and
o Up to four input-output operations on each of the two (maximum) HighSpeed Multiplexor channels; and
/I)
Multiple additional input-output operations (up to 64) via the Multiplexor Channel.
When the Multiplexor Channel or High-Speed Multiplexor Channel is operated in the
burst mode, only one input-output operation can proceed on that channel, but "interference"
with central processor operations is less than in the multiplex mode. Multiplex-mode operations on either type of channel permit concurrent input-output data transfers on each subchannel, provided that the aggregate data rate does not exceed the capacity of the channel.
The optional Direct Data Channel can be installed in place of the second HighSpeed Multiplexor Channel. The Direct Data Channel provides a 32-bit-wide data path between
core storage and up to eight "foreign" input-output devices, i. e., external devices from a
manufacturer other than IBM. In its normal mode of operation, the Direct Data Channel permits a single data transfer operation to be overlapped with internal processing. The maximum
data transfer rate in this mode is 500,000 32-bit words per second. When operating in the
"Priority-In" mode, the Direct Data Channel locks out all other input-output and internal processing operations and p'ermits data transfer to occur at up to 1,000,000 words per second.
Use of the Direct Data Channel requires installation of the first High-Speed Selector Channel.
The specific characteristics of the Model 44 Multiplexor,
and Direct Data Channels can be summarized as follows:
High~Speed
Multiplexor,
Multiplexor Channel
Maximum number:
... ' .•...
Maximum data rate, Channels
not operating Multiplexed mode: . . . . . . . . . . . . . .
Burst mode: . . . . . . . . . . . . . . . . . .
Processor demand: . . . . . . . . • . . . . . . .
l.
50 kilobytes/sec.
200 kilobytes/sec.
1 JJ.sec/byte in Burst mode, up to
5 JJ.s,ec/byte in Multiplexed mode;
see Table 1.
Number of control unit positions: . . . . . . . 8.
MiLximum number of subchannels: . . . . . . 64.
High-Speed Multiplexor Channel
Maximum number: . . . . . . . . . . . . . . . . .
Maximum data rate Multiplexed mode (each): . • . . . . . . . . . .
Burst mode Each, with two operating: . . . . . . . . .
One operating: . . . . • ' . . . . . . . . . . . .
Processor demand: . • . . . . . . . . . . . . . .
Number of control unit positions: . . . . . . .
Maximum number of subchannels: . . • . . . .
2.
200 kilobytes/ sec.
400 kilobytes/sec.
500 kilobytes/ sec.
0.5 JJ.sec/byte; see Table 1.
8/channel.
4/channel.
Direct Data Channel
Maximum number: . . • . . . . . . . . . . . . . .
Maximum data rate Normal mode: . . . . . . . . . . . . . . • . . . .
Priority-In mode: . . . . . . . . . . . . . . . .
Processor demandNormal mode: . . . . . . . . . . . . . • . . . . .
Priority-In mode: . . . . . . . . . . . . . . . .
Number of control unit positions: . . . . . . .
Maximum number of subchannels: . . . . . .
l.
500,000 words/sec.
1,000,000 words/sec.
25% of data transfer time.
100% of data transfer time.
8.
none.
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
IBM SYSTEM/360- MODEL 44
435:111. 101
TABLE I: INPUT-OUTPUT DEMANDS ON THE SYSTEM/360 MODEL 44 PROCESSOR
Demand on Processor,
per cent, via Device
Peak
Data Rate,
KB/sec*
Average
Data Rate,
KB/sec*
156
90
156
90
High-Speed
Multiplexor
Channel
Multiplexor
Channel
7.8
4.5
15.6
9.0
Random Access
2311 Disk Drive**
Single Disk Storage Drive**
Punched Card
2540 Card Read Punch:
Read, 1,000 cpm
Punch, 300 cpm
1442 Model N1 Card Read Punch:
Read, 400 cpm
Punch, 91 cpm
1442 Model N2 Card Punch, 91 cpm
2520 Model Bl Card Read Punch:
Read, 500 cpm
Punch, 500 cpm
2520 Card Punch:
500 cpm (Model B2)
300 cpm (Model B3)
2501 Card Reader:
600 cpm (Model B1)
1, 000 cpm (Model B2)
Printers
1403:
132 columns, 600 lpm (Model 2)
120 columns, 600 lpm (Model 7)
132 columns, 1,100 lpm
(Models 3, N1)
1443, 120 columns, 240 lpm
Punched PaEer TaEe
2671 Paper Tape Reader, 1,000 cps
Magnetic Tape
2400 Series:
Modell, 300 KB/sec**
Models 2 and 4, 60 KB/sec**
Model 3, 90 KB/sec**
Model 5, 120 KB/sec**
Model 6, 180 KB/sec**
1.3
0.40
0.65
0.02
6.5 max.
0.20 max.
0.53
0.12
0.12
0.53
0.12
0.12
0.03
0.01
0.01
0.26 max.
0.06 max.
0.06 max.
0.67
?
0.67
0.67
0.03
0.03
0.33 max.
0.33 max.
?
?
0.67
0.40
0.03
0.02
0.33 max.
0.20 max.
0.80
1.3
0.80
1.3
0.04
0.65
0.40 max.
6.5 max.
70
70
1.3
1.2
0.65
0.06
6.5 max
0.60 max.
70
70
2.4
0.48
0.12
0.02
1. 2 max.
0.24 max.
1
1
0.05
0.50 max.
30
60
90
120
180
30
60
90
120
180
70
70
* Kilobytes (thousands of bytes) per second.
** Operates in burst mode exclusively.
5/66
A
AUERBACH
'"
1.5
3.0
4.5
6.0
9.0
3.0
6.0
9.0
12
18
435:151.100
A•
AUERBACH
STANDARD
ED:!?
IBM SYSTEM/360
MODEL 44
SOFTWARE
REPORTS
SOFTWARE
Because of its restricted instruction repertoire
and its built-in Single Disk Storage Drive for use
by the programming systems, the Model 44 does
not use the extensive array of standard System/360
software that IBM is providing for Models 30, 40,
50, 65, and 75. However, two levels of software
are currently offered for use with the Model 44:
Programming System and Basic Programming
Support. The Model 44 Programming System
package is especially designed for use with the
Model 44 processor, whereas the three Basic
Programming Support programs (Assembler,
FORTRAN, and utilities) are virtually identical to
comparable programs in the standard System/360
Basic Programming Support (BPS) package.
. 11
The Disk Resident Monitor of the Model 44 Programming System is a sequential batch processing
monitor program, designed to supervise the
execution of one program at a time and to control
the automatic transition from program to program
at end of job time. The Monitor includes full
interrupt handling facilities and error condition
routines. The interrupt control routines provide
exit points for optional insertion of programmersupplied subroutines. Utility routines, such as
program and machine diagnostic routines, are also
provided by the Monitor. All Monitor facilities are
accessible to the problem prqgrams.
Another essential function pe~formed by the Disk
Resident Monitor is control of input-output devices
and supervision of their data transfer operations.
The devices presently supported by the Monitor
are listed in Table 1. No automatic control routines
are currently provided for data communications
and display devices. The Disk Resident Monitor
and its I/O control routines will be available in
January 1967.
MODEL 44 PROGRAMMING SYSTEM
\
'-
IBM is providing specialized software for use with
the Model 44 in order to use the processor's
restricted instruction repertoire as efficiently as
possible, and to take advantage of the built-in
Single Disk Storage Drive. All Model 44 Programming System software is disk-resident, and
the total package occupies approximately 500,000
bytes of the 1,171,200 bytes provided by an IBM
2315 Disk Cartridge used with the Single Disk
Storage Drive. The Model 44 Programming
System can function with the minimum amount of
processor core storage (32,768 bytes), and can
utilize any additional core storage to the system
maximum of 262,144 bytes.
Additional Model 44 configuration requirements for
use of Programming System software include a
2315 Disk Cartridge for software systems reSidence,
a Multiplexor or High-Speed Multiplexor Channel,
a card reader and card punch, a line printer, and
any associated control units. The optional Floating Point Arithmetic feature is also required to
use the Programming System FORTRAN N compiler and to execute all programs generated by that
compiler.
Disk Resident Monitor
. 12
Disk Resident FORTRAN IV
The Model 44's Disk Resident FORTRAN IV language provides all of the facilities proposed for
inclusion in the American Standards Association's
Basic FORTRAN language. The Disk Resident
FORTRAN IV language is directly compatible with
the 10K-byte level FORTRAN supplied with the
standard System/360 Basic Programming Support
package (and described in Section 420:163).
Both single- and double-precision floating-point
arithmetic operations can be specified in the Disk
Resident FORTRAN IV language. The IBM System/
360 Scientific Subroutine Package (SSP) is also
provided, giving the Model 44 FORTRAN programmer access to a library of 133 FORTRAN-language
subroutines to perform statistical functions, matrix
.~
TABLE I: I/O DEVICES SUPPORTED BY DISK RESIDENT MONITOR
Unit Supported
(
',,-
Console Printer Keyboard
1403 Printer, Models 2, 3, 7, N1
1442 Model N1 Card Read Punch
1442 Model N2 Card Punch
1443 Model N1 Printer
2501 Card Reader, Models B1, B2
2520 Model Bl Card Read Punch
2520 Card Punch, Models B2, B3
2540 Card Read Punch
2311 Disk Storage Drive
2400 Seri-es Magnetic Tape Drives,
Models 1 through 6
Standard Single Disk Storage Drive
Optional Single Disk Storage Drive
Function
Operator communication
Output, compiler listings
Input, output
Output
Output, compiler listings
Input
Input, output
Output
Input, output
Input, output
Input, output, compiler listings
Software residence, input, output
Input, output
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
IBM SYSTEM/36Q-MODEL 44
435: 151. 120
. 12
Disk Resident FORTRAN IV (Contd.)
core storage, but can effectively utilize additional
core storage up to only 65, 536 bytes.
manipulations, and other mathematical functions.
The Floating Point Arithmetic feature must be
installed to compile and/or execute Disk Resident
FORTRAN IV programs.
other minimum configuration requirements for
use of the Basic Programming Support software
include a Multiplexor or High-Speed Multiplexor
Channel, a card reader and card punch, a line
printer, and all associated control units. Use of
the magnetic tape-oriented FORTRAN compiler
requires in addition a minimum of three tape
drives and the Floating Point Arithmetic feature.
The built-in Single Disk Storage Drive is not
utilized by this software for program residence.
The Disk Resident FORTRAN IV compiler will be
available in January 1967.
. 13
Disk Resident Assembler
.14
The Disk Resident Assembler supplied for use with
the System/360 Model 44 provides symbolic language to handle all available Model 44 instructions,
including floating-point instructions and instructions
designed for the control of special features in the
processor, such as Read/Write Storage Protect
and Priority Interrupt. Standard pseudo-operations
for control of the assembly operation and the
facility to use literal data are also provided. The
Model 44 Assembler is directly compatible with
and provides facilities identical to the Basic Programming Support Tape Assembly language that is
available in the standard System/360 software and
described in Section 420:173. The Disk Resident
Assembler will be available in January 1967.
Disk Resident Utilities
The utility programs supplied for use with the
Model 44's disk-resident Programming System
include a program loader and file-to-file data
transcription routines for all the input-output units
listed in Table I. Like all other elements of the
Model 44 Programming System software, the Disk
Resident Utilities package will be delivered in
January 1967.
.2
.21
Like the Model 44 Disk Resident FORTRAN language, the Tape FORTRAN language provides all
the facilities of the American Standards Association's Basic FORTRAN language. The language
also provides all of the facilities available with
the Disk Resident language described in Paragraph
. 12 of this section.
The Tape FORTRAN compiler is a "stand-alone"
program; it requires the use of three magnetic
tape units for compilation and a fourth tape unit
for the compile-and-go capability. (Magnetic tape
units that record data at the 1, 600-bpi density
cannot be used with the Tape FORTRAN compiler.)
The compiler also requires use of the Floating
Point Arithmetic processor feature. The Tape
FORTRAN compiler will be available in October
1966.
.22
Card Assembler
The Model 44 Basic Programming Support Card
Assembler is a stand-alone program that provides
all the symbolic language facilities of the standard
System/360 card assembler that is described in
Section 420:173. The Model 44 Basic Programming
Support software does not provide any I/O device
control routines, error routines, or interrupt
handling routines; all such routines must be handcoded by the problem programmer. Pseudoinstructions are provided, but specialized macro
routines to handle optional features in the processor, such as Direct Data Channel, Direct Word,
and Priority Interrupt, are not available. The
Card Assembler will be delivered with the first
Model 44 systems, beginning in .october 1966.
MODEL 44 BASIC PROGRAMMING SUPPORT
Early users of the System/360 Model 44 will be
provided with three independent software support
programs, called the Model 44 Basic Programming
Support, to be used until the specially-designed
Model 44 Programming System package is made
available. First deliveries of the Model 44 system
are expected in October 1966, but its Programming
System software system will not be available until
January 1967, as noted above.
The interim Basic Programming Support software
is limited in scope and automatic programming
facilities. The only input-output devices supported
by this software include: 1442 and 2540 Card Read
Punch, 2501 Card Reader, 2520 Card Punch, 1403
and 1443 Printer, and Models 1, 2, and 3 of the
2400 Series Magnetic Tape Units. Basic Programming Support software functions with the
Model 44 processor's minimum 32,768 bytes of
Tape FORTRAN
.23
utility Programs
Model 44 Basic Programming Support software
also includes a package of utility routines that
provide file-to-file data transcription capabilities
for the devices listed in Paragraph. 2.
/'
5/66
A
AUERBACH
@
435: 20 1. 001
A
AUERBACH
SUMOAAD
EDP
IBM SYSTEM/360
MODEL 44
SYSTEM PERFORMANCE
RfPIIR'S
SYSTEM PERfORMANCE
(
GENERALIZED FILE PROCESSING (435:201.100)
The System/360 Model 44 is designed primarily as a scientific-oriented system, yet it can
perform standard business data processing functions. For this reason, the Model 44's
performance is measured on our four standard file processing problems. The lack of
radix conversion instructions and good output editing instructions degrades the overall
performance of the Model 44 when processing business-oriented problems of this type.
The problems involve updating a master file from information in a detail file and producing a printed record of each transaction. This application is one of the most common
commercial data processing jobs and is fully described in Section 4:200.1 of the Users'
Guide. Standard File Problems A, B, and C show the effects of varying record sizes in
the master file. Standard Problem D increases the amount of computation performed upon
each transaction. Each problem is estimated for activity factors (ratios of number of
detail records to number of master records) of zero to unity. In all cases a uniform distribution of activity is assumed.
The master files are arranged primarily in binary format to accommodate the 4-byte
binary word organization of the Model 44 processor. All computation is performed in
fixed-point binary mode. Processing of the detail records and report files requires the
use of time-consuming binary-to-decimal and decimal-to-binary conversion subroutines.
In the absence of definitive execution times for IBM's radix conversion subroutines, these
times have been estimated by our editorial staff.
For all four configurations (V, VI, VIlA, and XI), the master files are on magnetic tape,
arranged in various block sizes. The detail file is assigned to the card reader and the report file to the printer. The master-file tape units are connected to a High-Speed Multiplexor Channel, and tape reading and writing can fully overlap processing and the other
input-output operations. The card reader and printer are each connected to a subchannel
of a standard Multiplexor Channel.
For all configurations on all four of the Standard File Problems, the master-file tape times
are the controlling factor at very low activities, and the printer is the controlling factor
thereafter.
SORTING (435:201. 200)
Although no sort program is supplied as part of the Model 44' s standard software, estimated times to perform standard 3-way tape sort operations are graphed to permit comparisons with other systems. The estimating procedure is described in Paragraph
4:200.213 of the Users' Guide.
MATRIX INVERSION (435:201. 300)
The standard estimate for inverting a non-symmetric, non-singular matrix was computed
by the simple method in Paragraph 4:200.312 of the Users' Guide. In order to execute this
procedure in floating-point form, the optional Floating-Point Arithmetic feature must be included on the Model 44. Two lines are shown on the graph, one using the short floatingpoint format (7-digit precision) and the other using the long format (17-digit precision). The
Model 44's standard general registers are assumed to be present in the processor.
GENERALIZED MATHEMATICAL PROCESSING (435:201. 400)
Standard Mathematical Problem A is an application in which there is One stream of input data,
a fixed computation to be performed, and one stream of output results. Two variables are
introduced to demonstrate how the time varies with different proportions of input, computation, and output. The factor C is used to vary the amount of computation per input record.
The factor R is used to vary the ratio of input records to output records. The procedure
used for the Standard Mathematical Problem is fully described in Section 4:200.2 of the
Users' Guide.
The optional Floating-Point Arithmetic feature must be included on the Model 44 in order to
perform the floating-point computations required in this problem. Double-length operations
were used, since the problem specifications demand a minimum of 8-digit precision and
single-length floating-point operations provide only 7-digit precision. However, by setting
an operator-controlled switch on the Model 44's console, the length of double-length floatingpoint arithmetic operations can be reduced from 17-digit precision to 9-digit precision,
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
IBM SYSTEM/3~0 - MODEL 44
435:201. 002
with consequent reduction in the floating-point instruction execution times (see Model 44
Central Processor, page 435:051). Therefore, 9-digit precision, "long" floating-point
computations have been performed throughout this problem.
Graph 435:201. 400 shows three curves. The curve marked "R = 1. 0" shows one output record for every input record. The other two curves show one output record for every tenth
(R = 0.1) and every hundredth (R = 0.01) input record. There is very little difference
between these latter two curves. For configurations VI and VIlA (and Configuration XI at
the O. 1 and 0.01 output ratios), the central processor becomes the controlling factor when
more than 40 times the standard amount of computation is performed. The printer in Configuration XI is the controlling factor virtually throughout the problem when there is one output record for every input record.
,
WORKSHEET DATA TABLE 1
!
CONFIGURATION
ITEM
REFERENCE
V
1
Char/block
Records/block
1,056
1,056
12
12
12
51. 2
125
File 4
O_ _. _
~=~ f----- _ _
..!.ik2. _ _ _ f . . - .__O_ _. _
°
Central
Processor
Times
msee/record
msee/detail
msee/work
msee/report
Standard
File
Problem A
F = 1.0
[or C.P.
1-- ___
O.4~
b7 + ba
~---
~.--a3K
dominant I/O
~asterIn
column.
~asterOut
Printer
Printer
C.P.
0
~~-
---~-
____
O.~
_ _ _0_.4~
0.60
0.60
0.44
0.44
------0.97
-----0:97 -
-----_.- - - - ' : Printer
C.P.
Printer
C.P.
0.44
- - 'l:~ ---- -' 6 4 . 1 - - - - - Ei4.i""'- - 0.53
0.53
0.53
~- - - ---;;:-53-- - -0.-5-3-- - - --0:53- - ~.
-- 4.'8-- -4 . 8 0.44
0.44
0.44
~- 1 - - - - 1.1t--
~-
~-
~-
64.1
4:200.114
~-
1500.0 -
7.2
1500.0
7 : 2 - 1500.0- 7 : 2 - 1500.-0-78.76
1500.0
,_ I - - G,OOO (?)
~--Standard
File
Problem A
Space
78.76
1,566
5,5_62_ _ _
5,562
~---
4,624
4,624
100
100
17 980
17 980
Fixed floating point
Unit name
- -~.--output
Size of record
~.
~ Tl
output
T2
~2L- ---output
T4
T5
Tn
T7
output
msec penalty
~ord
~ops
msec!report
_
1500.0
7.2
""3600:0 -
78.76
3600.0
I--~)- I-~O~
~-
128
1,5_6_6_ _ _
WorkinQ'
msec/block
6,000 (?)
__
12_8_ _ _
~ckslt023)
~kS24t048)
To"!
5/66
___0 _
0
Unit of measure (bytes)
~routines
Standard
Mathematical
Problem A
-------
78.76
Total.
_ _0_ _ _
0.15 . - ----O:~ 4:200.1132
- . - - - . - _ _ _ _ 1_.9_ _ _ - - 1 - . 9 - - ' - - ' - - - 1 9 _
1.9
--3-.3-3.3
3.3
3.3
~.
File 4: .Reports
5
0.60
0.44
I--- - 0.44
~--- - - I-----o~
I--0.097
~------~
-----0.15-
~etails
4
0
___0._53_ _
0.60
C.P.
and
1-------'-- _ _ _ °
0_ _ _. _ _
~--- '---~-
~.--~.---
- - -10051. 2
---125
1 - - - _ _0_ _ _ _ _ _0 _ 4:200.112
1--
3
mseclblock
125
125
..Elk.!..o~ ,...--~File 4
msec/block
12
I----~- 1------_.- I--~1- ___10_o_ _. - f---~~--- I----~-
File 4
2
1,056
1,056
msee/switch
msee ,penalty
XI
(File 1)
~~~
Times
VIIA
(File 1)
msec/block
Standard
File
Problem A
InputOutput
K
VI
.- f-----~
-
I-- _ _1_2_8
_
I--~-
f----~- I--~100
100
17 980
17,980
Floating point
:waung pu,"
4:200.1151
~-
t----~
2501 Card Reader, Mod Bl 2501 Card Reader, Mod. Bl 2501 Card Rea~
1443 Printer. Mod Nt
1403 Printer, Mod 7
1403 Printer. Mod 7
80 bytes
80 bytes
80 bytes
80 bytes
80 bytes
80 bytes
100
100
100
100
100
250
~
0.40
0.40
0.40
0.40
0.40
0.40
2.22
2.22
2.22
2.52
2.52
2.52
1. ~8
1. 58
1. 58
4:200.413
(
-
(Contd. )
A
AUERBACH
'"
/
435: 20 1. 100
SYSTEM PERFORMANCE
.1
GENERALIZED FILE PROCESSING
. 11
Standard File Problem A
.113 Timing basis: . . . . . . using estimating procedure
outlined in Users' Guide,
4:200.113.
.
.114 Graph: . . . . . . . . . . . see graph below .
.115 Storage space required Configuration V: ... 17, 980 bytes.
Configuration VI: ... 17,980 bytes.
Configuration VIlA: . 17, 980 bytes .
Configuration XI: ... 17,980 bytes.
. 111 Record sizes Master file: . . . . . . 108 data characters, packed
as 88 8-bit bytes.
Detail file: . . . . . . . 1 card.
Report file: . . . . . . . 1 line.
. 112 Computation: . . . . . . . standard, using fixedpoint binary arithmetic.
100.0
7
2
/
10.0
7
..,
#'
W
\
"-
2
1.0
~
~
~
/
J
4
Time in Minutes to
Process 10,000
Master File Records
V--
~
I
/'
~7o/
.-
7
1.0
_C1'
~
4
I'VIIA
"",,"'"
I'
2
~
/O'?'
/'
I
"
0.0
0.1
0.1
7
I
,,,
4
2
0.01
0.33
1.0
I
\
Activity Factor
Average Number of Detail Records Per Master Record
(Roman numerals denote standard System Configurations; curve marked
CP shows central processor time.)
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
IBM SYSTEM/360 -
435: 20 1. 120
.12
Standard File Problem B
MODEL 44
.122 Computation: . . . . . . . standard, using fixed-point
binary arithmetic .
. 121 Record sizes -
.123 Timing basis: . . . . . . using estimating procedure
Master file: . . . . . . 54 data characters,
packed as 44 8-bit bytes.
Detail file: . . . . . . . 1 card.
Report file:. . . . . . . 1 line.
outlined in Users' Guide,
4:200.12.
.124 Graph: . . . . . . . . . . . see graph below.
100.0
7
4
2
"",--
/
10.0
~
------ -----
~~
L
7
I
~
~
4
Time in Minutes to
Process 10.000
Master File Records
2
1.0
7
4
I
/"
rrt;;~r/
---
.,
...c"P
IIVIIA
~
./
-""
G~/
2
/~
j
0.1
~.
7
I
I
4
II
,/
2
0.01
0.0
O. 1
0.33
1.0
Activity Factor
Average Number of Detail Records Per Master Record
(Roman numerals denote standard System Configurations; curve marked
CP shows central processor time.)
5/66
(Contd.)
A
AUERBACH
OJ
435: 20 1. 130
SYSTEM· PER FOR MANCE
I
'"
.13
.132 Computation: . . . . . . . standard, using fixed-
Standard File Problem C
point binary 3y.i.thmetic .
basis: . . . . . . using estimating oro(~edure
outlined in Usen," Guide,
4:200.13.
. 134 Graph: . . • • • . . • . . . see graph below.
. 131 Record sizes -
. 133
Master file: . . . . . . 216 data characters,
packed as 176 8-bit
bytes.
Detail file: . . . . . . . 1 card.
Report file: . . . . . . . 1 line.
Timi~g
100.0
7
4
2
10.0
4
Time in Minutes to
Process 10,000
Master File Records
,
,,
7
/
2
~
/
AI
AfIY
L
Y
./
7
III
~VIIA
C~'
.,;'
2
l
~
",- ---
l'
-
-
=-",,-
1.0
/"
JI.
0.1
7
..-
-----
~
1/
1.0
4
~
.--..
L
4
2
0.01
0.0
o. 1
0.33
1.0
Activity Factor
Average Number of Detail Records Per Master Record
(Roman numerals denote standard System Configurations; curve marked
CP shows central processor time.)
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
435: 20 I. 140
. 14
IBM SYSTEM/360 -
Standard File Problem D
. 141 Record sizes Master file: . . . . . . 108 data characters, packed
as 88 8-bit bytes.
Detail file: . . . . . . . 1 card.
Report file: . . . . . . . 1 line.
MODEL 44
.142 Computation: . . . . . . . trebled, using fixed-point
binary arithmetic .
.143 Timing basis: . . . . . . using estimating procedure
outlined in Users' Guide,
4:200.14.
. 144 Graph: . . . . . . . . . . . see graph below.
100.0
7
4
2
/
10.0
II
Time in Minutes to
Process 10,000
Master File Records
2
"..
./
J
~o//
1.0
7
,JI'-
4
C~~
7
._____Ci?-
--
~---.... 1.0
/
2
0.1
~
i/VIlA
------
~
-
,
7
4
/'
~
..--
/
1
,
I'
4
2
0.01
0.0
0.1
0.33
1.0
Activity Factor
Average Number of Detail Records Per Master Record
(Roman numerals denote standard System Configurations; curve marked
CP shows central processor time.)
5/66
fA
AUERBACH
(!)
(Contd. )
435: 20 1. 200
SYSTEM PERFORMANCE
I
\.
.2
SORTING
.21
Standard Problem Estimates
. 213 Timing basis: . . . . . . using estimating procedure
outlined in Users' Guide,
4:200.213; 3-way tape
merge .
. 214 Graph: . . . . . . . . . . . see graph below .
.211 Record size: . . . . . . . 80 characters.
. 212 Key size: . . . . . . . . . 8 characters.
1,000
7
4
2
100
7
4
'I
2
Time in Minutes to
put Records into
Required Order
10
/
,
, ,,
7
"
4
I
~
2
~'f'
/
V
I
I
,'I
./
/
~'\.' ~
~,'1
/
1
,
,
7
"/
4
/'
2
1,/
4.11
V
/
0.1
100
2
4
7
1,000
2
4
7
10,000
2
4
7
100,000
Number of Records
(Roman numerals denote standard System Configurations.)
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
435: 20 I. 300
IBM SYSTEM/360 -
. 3· -MATRIX INVERSION
.31
MODEL 44
decimal digits in the
SHORT format or 17
digits in the long
format.
.312 Timing basis: . . . . . . using estimating procedure
outlined in Users' Guide,
4:200.312.
.313 Graph: . . . . . . . . . . . see graph below.
Standard Problem Estimates
.311 Basic parameters: ... general, non-symmetric
matrices, using the optional Floating-Point
Arithmetic feature; precision is approximately 7
1.0
7
I
4
,
l
~
2
VIJ I
0.1
I
7
I
II
I
If
4
I J
I II
d!,F-.'/
2
,-
§~
'-lJ
;g
Time in Minutes
for Complete
Inversion
O. 01
I
7
I
I
I
I
II
I
4
I I
III
2
'1
0.001
•
7
4
2
0.0001
1
2
4
2
7
4
10
7
2
100
4
7
1,000
Size of Matrix
5/66
A
(Contd.) .
AUERBACH
'"
435: 20 1. 400
SYSTEM PERFORMANCE
\
;"
I
\
.4
GENERALIZED MATHEMATICAL PROCESSING
. 41
Standard Mathematical Problem A Estimates
"long" floating-point mode,
set to 9-digit precision .
.413 Timing basis: . . . . . . using estimating procedure
outlined in Users' Guide,
4:200.413.
Note: use of the optional
Floating-Point Arithmetic Feature is assumed.
.414 Graph:. . . . . . . . . . . see graph below.
.411 Record sizes: . . . . . . 10 signed numbers; average size 5 digits, maximum size 8 digits .
. 412 Computation: . . . . . . . 5 fifth-order polynomials;
5 divisions and 1 square
root; computation is in
2-+----~--~+-~~~----~--+_+_+1_rHH----_r--+_~~rrrH
4
~----_+--~_+_+~++++----_+--~_+_+~r++r----_+--~_+_r~rrH
XI: R
2
= 1. 0
,
_+----_r--+_~~r++++_----r-_+_;_+~++++----~--+_;~~~~.
~~1~
l' ... 1:\'
VI, VIlA: R
Time in
100
Milliseconds
per Input Record 7
~
1. 0, 0.1, 0.01; XI: R
=
~ ~~,
0.1, 0.01:
-~~~~~~~~~·-3~·~l-~-~~~~~~~~~~tt~~~~~~~~I:\·
IV "9"+--Ht-t-H
1',-f~+-H
.~
"
2
2
0.1
4
7
2
1.0
4
2
7
4
10.0
7
100.0
C, Number of Computations per Input Record
(R = number of output records per input record;
curve marked "CP" shows central processor time;
Roman numerals denote standard System Configurations.)
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
/
'.
./
./
435:221. 101
A
saMDARD
EDP
AUERBACH
IBM SYSTEM/360
MODEL 44
PRICE DATA
REPDRTS
~
PRICE DATA: MODEL 44
PRICES
IDENTITY OF UNIT
CLASS
Name
No.
Monthly
Rental
$
PROCESSING
UNIT
2020
Unit
- 32,768 bytes
- 65,536 bytes
- 131,072 bytes
- 262,144 bytes
3,465
4,665
6,265
10,065
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)
Purchase
$
$
200.00
220.00
250.00
350.00
123,600
179,700
257,700
441,700
-
-
30
275
700
350
650
1. 00
11.00
28.00
17.75
29.25
1,260
11,550
29,400
14,680
27,300
4560
4561
4562
4599
Additional High-Speed Multiplexor
Subchannels:
First
Second
Third
High-Speed Multiplexor Channel (second)
125
125
125
360
6.00
6.00
6.00
18.00
5,250
5,250
5,250
15,100
4565
4566
4567
3275
3288
3621
4555
9509
5625
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
125
125
125
800
275
NC
100
NC
400
6.00
6.00
6.00
30.00
3.00
NC
2.00
NC
10.00
5,250
5,250
5,250
33,600
11,550
NC
42,000
NC
16,800
Store and Fetch Protection, for:
Model E
Model F
Model G
Model H
225
275
325
400
4.00
4.75
5.50
7.00
9,450
11,550
13,650
16,800
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
230
40.00
9,775
90
3895
4427
4583
5248
4598
/
Processing
Model E
Model F
Model G
Model H
Monthly
Maintenance
I
\
7531
7531,7532
7531-7533
7531-7534
I
i
\
"
2251
2252
7500
7501
6415
2315
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
-
-
For prices of the System/360 peripheral devices, please refer to the general System/360 Price Data section,
which begins on page 420:221. 101.
© 1966 AUERBACH Corporation and AUERBACH Info, Inc.
5/66
""
..
,
. ..'~"', .,:
......
,
....
;.
./
..
~
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-c041 52.342996, 2008/05/07-21:37:19 Create Date : 2018:05:10 16:23:34-08:00 Modify Date : 2018:05:11 14:00:59-07:00 Metadata Date : 2018:05:11 14:00:59-07:00 Producer : Adobe Acrobat 9.0 Paper Capture Plug-in Format : application/pdf Document ID : uuid:b124abc3-9a1f-0946-b8b0-d578d38d9e51 Instance ID : uuid:74ea684b-cec5-b84b-8a7b-2bec26790260 Page Layout : SinglePage Page Mode : UseOutlines Page Count : 978EXIF Metadata provided by EXIF.tools