GE 235 System Manual

Download

Page 1

Page 2

@En235

SYSTEM

MANUAL

COMPUTER DEPARTMENT

Page 3

GENERAL ELECTRIC COMPANY

In the construction of the equipment described, General Electric reserves the right to reasons of improved performance and operational flexibility.

modify the design for

Page 4

TITLE SECTION INTRODUCTION

PROGRAMMING AIDS SYSTEM CONTROL OF INPUT-OUTPUT PERIPHERAL DEVICES CENTRAL PROCESSDR GE-235 PUNCHED CARD EQUIPMENT GE-235 HIGH-SPEED PRINTER GE-2 35 MAGNETIC TAPE SUBSYSTEM GE-235 MASS RANDOM ACCESS DATA STORAGE GE-235 PERFORATED TAPE EQUIPMENT

GE-235 12-POCKET DOCUMENT HANDLER GE-2 35 DATANET- 15 DATA TRANSMISSION SUBSYSTEM GE-2 35 CUSTOM DIGITAL INPUT/OUTPUT EQUIPMENT

G-235 SYSTEM INSTALLATION DATA

.........................................

......................................

.....................................

.............................

................................

............................

.......................

..........................

..................

............................

..............

1-1

11-1

111-1 IV-1

V-1

VI-

VLI-

VLII-1

IX-1

X-1

XI-

XII-

PAGE

1 1

ALPHABETIC LIST OF GE-235 GAP INSTRUCTIONS REPRESENTATION OF GE-235 CHARACTERS INDEX

.............................................

........................

....................

2-14

17-20

Page 5

The GE-235 is the fastest, most versatile memberof the GE-200 Series of information processing

systems (GE-215, 225, and 235). Like the other members of this family, the GE-235

complemented by a full range of powerful programmingtools. Together, the equipment and programming tools provide an integrated system adaptable to a wide variety of business, scientific, and engineering applications.

The basic design philosophy

oftheGE-235 system is the same as that of the GE-215 and 225, which have proved themselves fast, accurate, reliable, and economical in widely divergent fields. This design similarity makes the

GE-235upwardcompatiblewith

the GE-215 and 225 in respect to logic, programming, and coding. As a result, mostprograms and applications originally designed for the other members of the family can immediately be processed on a GE-235 having the same system configuration. Only in the relatively few programs containing timing loops, minor changes may

necessary.

GE-235 INFORMATION PROCESSING SYSTEM

Page 6

The basic system comprises a central processor with a six-microsecond core memory, card or

pe~forated tape input andoutput, electric typewriter input and output, and an operator's control con-

sole. The system configuration can readily be expanded to fit increasing information processing

needs. Equipment available includes:

Magnetic core memory

Card readers Card punch

400 or 1000 cards per minute

100 or 300 cards per minute

4096-, 8192-, or 16,384-word capacity

Perforated tape reader Perforated tape punch Magnetic tape handlers

15,000/41,600 characters per second (at 200/555.5 bits per inch)

Mass random access data storage

34.4 million numeric digits per unit

High-speed printers

per minute,

Document handler

Data communication controllers

Floating point arithmetic capability (through the Auxiliary Arithmetic Unit)

An outstanding feature

system, allowing great flexibility in system configuration. Some typical system configurations for various applications are shown on the following pages.

Programming aids available for the GE-235 include:

on/off line

250 or 1000 characters per second

110 characters per second

read and write 15,000 characters per second (at 200 bits per inch)

(MRADS)

900 alphanumeric lines per minute, on line, or 900 alphanumeric lines

reads and sorts 1200 documents per minute, on or off line

that up to ten input-output devices may be operated concurrently within the

18.8 million alphanumeric characters or up to

GECOM be used with:

COBOL-type statements (specific, simplified English language statements) ALGOL-type statements (algebraic expressions) TABSOL (structured decision tables) GECOM Report Writer (for programming business reports)

FORTRAN

ZOOM

WIZ

GAP

the general compiler, an all purpose, problem-oriented language program that may

a scientific compiler

a macroassembly system

a highly competent, one-pass algebraic compiler

a fast, compact, machine-oriented assembler

I- 2

Page 7

Standard report generators, sort/merge routines, BRIDGE (an ope rating system), GE-235/~~~ (a major network analysis technique),

Thus, the GE-235 provides economical solution of data processing and scientificproblems, and for potential growth in desired areas. The characteristics and capabilities of this new member of the GE-200 Series are fully described in this manual. However, detailedinformationfor operating and programming the system

in separate manuals on these subjects.

all

andother specializedprograms and routines are also available.

the tools for effective management decision-making, for fast and

I-

Page 8

CENTRALPROCESSOR

CARD READER

HIGH-SPEED PRINTER

CARD PUNCH

SYSTEM CONFIGURATION FOR ENGINEERING CALCULATIONS OR REPORT GENERATION

HIGH- SPEED PRINTER

CARD PUNCH

fl"=-,

CENTRAL PROCESSOR

AND AUXILIARY

ARITHMETIC UNIT

MAGNETIC TAPE

UNIT

C..

CARDREADER

---my

(1000

CPM)

SYSTEM CONFIGURATION FOR SCIENTIFIC CALCULATIONS OR DATA RETRIEVAL AND REDUCTION

Page 9

MAGNETIC TAPE UNIT

T+.

+*q!Fi

_C--

CARD PUNCH

---

PERFORATEDTAPE

HIGH-SPEED * PRINTER

DATANET-

--2

-!u7*-

z7-;

-~-'

-T

CENTRAL PROCESSOR

'v:-???

CARD READER

-1

DATA COMMUNICATIONS SYSTEMS

SYSTEM CONFIGURATION FOR BUSINESS OR MANUFACTURING

Page 10

HIGH-SPEED

PRINTER

MAGNETIC TAPE UNIT

CARD PUNCH

CENTRAL PROCESSOR

DOCUMENTHANDLER

SYSTEM CONFIGURATION FOR BANKING

CARD READER

Page 11

PROGRAMMING AIDS

The General Electric Computer Department has developed a large library of programming aids to help the programmer communicate with the GE-235 and simplify the task of producing useful results from the computer. generators and special programs designed to enhance the use of the GE-235 information processing system.

This section describes some of the available programming tools: compilers,

GECOM,

The General Compiler (GECOM) System introduces a fresh, versatile approach to computer communication. This exclusive General Electric product makes available in one package both proved and newly developed programming techniques. GECOM accepts many languages, so problem statements may be written in familiar terminology. The source languages available to the General Compiler are broad and comprehensive.

GECOM will process English language sentences (COBOL-type statements), algebraic expressions (ALGOL-type statements), structured decision tables ation. combination of the language features for any specificprogram run. Because the machine coding

derived directly from the logic of the problem statement, program check-out on the GE-235 may be

done at the logic level.

Because GECOM problems are written in familiar languages, they can be more easily read and understood. approach allows the user to accommodate the more important common coding languages and still

incorporate later changes conveniently. Several distinct advantages over manual programming

methods can be realized.

GECOM automatically produces a documented record of the program it produces. A permanent

record of the program, in its original source language form and with a detailed listing of its transformation to machine instructions,

THE

The user may select only that portion of the system applicable to his needs, using any

GENERAL COMPILER

(TABSOL), and a language for report gener-

In addition, program format provides a high degree of standardization. The selected

available for reference, revision, or augmentation.

Because plans call for implementing GECOM on the General Electric family of general-purpose computers, programming conversion costs are reduced as installationsoutgrow their present com-

puter equipment.

Using familiar language sharply reduces personnel training time and expense. Manual coding eliminated and debugging cut to a minimum. ~hus, a machine program may be produced quickly and efficiently.

COBOL came into being as a result of a conference on Data Systems Languages sponsored by the U. S. Department of Defense. Computer manufacturers and users developed the language called COBOL languages.

computers. The language first available with the General Compiler

(Common Business griented Language) to achieve standardization of data processing

COBO~

reduces progra&ing effort and achieves a more effective utilization of

based primarily on COBOL,

Page 12

which satisfies the needs of the broadest spectrum of data processing applications. COBOL is so close to English syntax that it can easily be read and understood by management, systems, and accounting personnel. As a result, close coordination between management and computer application is both practical and efficient.

COBOL

well suited to creating and processing information contained in data files.

In contrast, ALGOL provides an excellent means for expressing the mathematics and logic associated with scientific applications.

ALGOL was developed by an international group prompted by a growing interest in a standardized notation for numerical methods for computers.

ALGOL

(ALGOrithmic Language) has proved to be far superior to any of its predecessors and has enjoyed the first widespread acceptance and respect accorded a computer language. ALGOL notations are gaining acceptance internationally in numerical methods, text-books and university classes.

TABSOL, the language of decision making, resulted from a need for a language that could solve an unwieldy number of sequential decisions, without involving extensive data file processing or pro-

found mathematics.

CEHERAL~

C~.,",,.

.1...,.

ELECTIIC

1.7.

.0,",1.

A.110..

OENERAL COMPILER SENTENCE FORM

,3,,.,,,,

SAMPLE TABSOL TABLE

Page 13

TABSOL depicts, by means of tables, the relationships of logical decisions that are written in terms

of the conditions to be satisifedandthe subsequent action to be taken. The

vides a readable, understandable table of decisions.

TABSOL structure pro-

TABSOL encompasses both scientific and business applications.

ALGOL-type statements within the framework of the table, thus providing an even more efficient

method for stating the logic of complex information systems.

The GECOM Report Writer ming of business reports. Readily understandable program documentation and ease of preparation of new and revised report

In brief, the Report Writer performs any or all of the following functions:

Prints report headings once at the beginning of the report

Prints report footings once at the end of the report Maintains page control by line count and or skip to a new page at specified line printings Maintains line spacing on the page Prints page headings at the top of each report page

Prints page footings at the bottom of each report page Numbers pages Issues detail or body lines of the report

an extension to the General Compiler that simplifies the program-

are realized by use of this tool.

GECOM accepts COBOL and

Accumulates detail field values conditionally or unconditionally to one or more levels of total

Counts detail lines and/or detail conditions to one or more levels of total

Page 14

Detects control breaks at one or more levels so as to: a. Control the tabulation procedure

b. Issue logical control totals

c. Issue logical control headings

Edits data fields for reporting (for example, comma, decimal point, and dollar-sign insertion and zero suppression)

The COBOL-61 to GECOM Translator converts programs described in COBOL-61 language into language acceptable to GECOM. The source language of the basic compiler is based in part on COBOL-60. The translator enablesGECOM toprovide for additional functions defined in COBOL-61 specifications.

FORTRAN FORTRAN

Using this compiler with the GE-235, a source program written in the language of FORTRAN

scientific compiler, will produce an assembled program ready for use.

FORTRAN II Compiler With Card Input-Output

This compiler will compile a FORTRAN I1 source program on a GE-235 system with a minimum input-output configuration. The full FORTRAN language (magnetic tape operations, for example, are omitted).

ZOOM

Simplicity and flexibility of coding are principal features of the macroassembler called ZOOM (in some respects a compiler). The simplicity of ZOOM coding is illustrated by the fact that the programmer writes algebraic expressions with such ordinary symbols as the plus, minus, and equal signs. Since they are easily read, the expressions are easily and quickly checked for errors.

ZOOM translates these algebraic expressions into near optimum GAP coding. To the programmer who has a working knowledge of GAP, programs and produces near optimum object programs. Input of GAP coding and ZOOM statements; output can be punched cards, magnetic tape, or printer

listings.

Compiler

II,

not implemented in this compiler

ZOOM allows for more condensed and readable symbolic

punched cards with combinations

Engineers and other users of the GE-235 who are not primarily programmers will find the WIZ

System a simple, easy-to-use algebraic compiler. The compiler translates source programs

written in the simple WIZ language, usingordinary mathematical symbols, into GE-235 object pro-

grams ready to run. WIZ produces GE-235 instructions on cards at a rate of 500 to 700 instructions per minute. WIZ makes iteasyfor the user to perform either simple or complex calculations and print the results in edited form.

Page 15

WIZ works with both floating point and fixed point numbers and handles typical algebraic and

trigonometric problems quickly and easily. Modification I permits use of paper tape as well as

punched cards. Use of the optional

time of the object program.

AAU with the WIZ System significantly decreases the run

GAP, THE GENERAL ASSEMBLY PROGRAM

The General Assembly Program allows the programmer to write instructions for the GE-235 computer in symbolic notation rather than in the absolute code of the computer. Mnemonic codes for each instruction are carefully chosen to provide significance to the user. Memory addresses may be

assigned by using decimal notation or by using symbolic notation chosen for maximum convenience to the particular program or programmer. To extend the use of the General Assembly Program the programmer can call on various subroutines (described below) as required by the

program. The General Assembly Program also provides facility for assembling of programs in

either absolute or relocatable form.

A wide range of assembler (pseudo) operations are available as follows: ALF BSS The BSS DDC DEC

EJT

END

EQO

EQU FDC

LOC This operation performs the same function as the ORG operation but the contents are

LST

The ALF

This This is used to enter a single-word decimal constant in the object program.

This operation causes the printer to slew the GAP listing paper to the top of the following

page.

The END Performs the same function as the EQU operation but the operand is assumed to be an

octal number.

Used to over-rule the normal memory assignment performed by the assembly program.

This scale

number.

assumed to be in octal form. This pseudo-operation may be used to start the listing again after it has been suppressed

by the NLS.

used to enter an alphanumeric constant in the program.

used to reserve a block of memory storage.

used to enter a double-word decimal constant in the object program.

(~nd of Program) indicates the end of the program to be assembled.

used to enter a floatingpointdecimal constant in the object program. If no binary

specified, determines the binary scale and yields a normalized floating point

MAL

NAL

This pseudo-operation can be used to specify from on one card.

A/N

This pseudo-instruction causes any the 2's complement form.

constant or group of constants to be assembled in

to 9 words of alphanumeric constants

Page 16

NAM

Permits a program name to be printed at the top of each page of the GAP listing.

NLS OC T ORG PAL

REM

SBR

SEQ

TCD

ZXX

Suppresses listing of the object program during assembly. The OCT converts up to seven octal digits into a binary equivalent. ORG (Origin) This pseudo-operation can be used to specify from

used to indicate the location of the first instruction of the program.

to 9 words of alphanumeric constants

on one card. The last word generated will have the sign set to terminate a print line. The pseudo-operation PLD will cause the assembly program to punch loader cards. When

the PLD pseudo-operation is encountered, all cards from that point to the end of the assembly will be punched in loader format.

The REM programmer's remarks immediately following are not processed by the assembly

but they do appear on the final program listing.

This pseudo-operation

used to call a specified subroutine master tape during assembly.

Checks the sequence number of each card against the sequence number of the previous card.

Generates an instruction that transfers control to the location specified in the operand field, at execution time; however, does not indicate end of assembly.

The ZXX pseudo-operation is used to set the operation bits of the assembled instruction to any desired configuration. The operand can be decimal or symbolic, and indexing

optional. In use, a Z is placed in column 8 with the two octal digits (XX) desired as an

and

operation code in columns

10.

SAMPLE GAP CODING

Page 17

SUBROUTINES, SERVICE ROUTINES AND BRIDGE II

Subroutines

Subroutines are designed to handle, manipulate, move, or sort information within the computer memory. Some of the important routines accomplish the following:

Conversion of data from one radix to another (octal, binary; BCD)

Word replacement

Internal memory sort

To solve problems in scientific areas, mathematical routines are available to calculate complex

functions and mathematical procedures such as:

Sine-cosine, square root, arctangent, exponential, and logarithm Matrix transposition, inversion

Scalar multiplication

Linear simultaneous equations

Multiple regression

Roots of a polynomial

Least squares polynomial fit

Linear programming

Service Routines

The main functions of service routines are to assist in debugging programs and in simplifying oper-

ating procedures.

These routines have been prepared in symbolic and/or object program form.

Service routines to perform tasks such as the following are available:

Reset memory Dump memory to cards, magnetic tape, paper tape, or printer

Load programs into memory

Trace programs

Compare, correct, and print out contents of tape

Correct cards

Page 18

Scan memory Convert, analyze, and relativize card decks

Reproduce cards or print out contents

Bridge Bridge

by the programmer, Bridge

The functions of Bridge execution. For installations that have a steady work load, use of Bridge

SIMULATORS AND GENERATORS

a tape maintenance and run sequencing program. Upon the use of simple instructions

provides such functions as:

Run collection and sequencing from cards or tape

Tape correction of binary or symbolic programs Dating of magnetic tape, using either date created or current date Blocking of tape records Provisionof run-to-runlinkage

Provision of altering run sequencing Combining of runs with subroutine or relocatable sections

Provision for loading priority programs for use with API

I1 are directed by control cards that establish the run sequence for run

I1 reduces over-all time.

Forward

The Forward Sort and Merge Generator produces tailored card or tape programs to efficiently sort and merge GE-235 data files, The sorts and merges are tailored at generation time according to descriptive parameters written by the user. Extensive options that allow for use of GAP coding enable users to attain complete flexibility in data format and selection and to utilize media other than tape.

IBM-650 Simulator

The IBM-650 Simulator accepts IBM-650 System programs the required routines to simulate the IBM-650 computer commands, and produces the same results and outputs as the IBM-650 computer. An existing IBM-650 program need not be written in GE-235 language in order to run on the GE-235 computer.

Sort/Merge Generator

input/output

anddata as input, selects and executes

Page 19

Potential users of the GE-235 who have the IBM-650 computer can make a smooth transition to the GE system through the use of simulators provided by General Electric.

The simulator program achieves the following objectives without loss of accuracy or flexibility:

Simulates the basic IBM-650 System, with 2000 words of drum memory, one 533 card reader, and punch with alphabetic device.

An extended version provides the capability of core storage, index registers, floating point, and magnetic tapes. The simulator program runs on the GE-235, with

atleast 8192words of memory, card reader and card punch, and

typewriter. Control cards preceding the IBM-650 program deck define the IBM-650 plugboard wiring

and console switch settings. Can be modified to include other IBM-650 configurations or features with a minimum of

programming effort. Documentation

detailed and complete so that features peculiar to

certain applications may be readily incorporated.

LGP-30 Simulator

The LGP-30 Simulator executes the Royal

McBee LGP-30 system instructions in the GE-235 and

produces essentially the same results and outputs as that computer.

SPECIALIZED PROGRAMS

Special needs of computer users are filledby specialized programs such as the text searching system and the

GE-235/cP~ program. Other programs are tailored to needs of a specific industry

or user.

The Text Searching System

The Text Searching System permits retrieval of information from texts.

The System consists of three principal programs. One converts texts (written in a natural or artificial language) into a form suitable for searching. A second program compiles programs to search the texts for requested symbol occurrences. The third program executes the compiled programs to search converted texts and announce the search results.

The

GE-235/Cp~ program adapts a major network analysis technique to the GE-235. Complex projects (such as new product introduction, large constructionprojects, and assembly-line planning) consisting of as many as 2100 activities and

1000 events can be analyzed by the GE-235 in minutes. Alternate schedules with optimum time and cost data, or other major project parameters, are produced as a printed output. The least an 8K memory,

magnetic tape handlers, a card reader, and a printer.

GE-235/~~~ program can be used with GE systems having at

Page 20

BankPac

series of generalized programs, called BankPac, have been tailored to the needs of commercial

banks. General Electric prepares broad programs to do such jobs as updating and maintaining files, issuing reports, functions. The user can readily add desired detailed programs. BankPac program will cover demand deposit accounting, installment loans, savings accounts, transit items, and personal trusts.

making customer statements, and the handling of many other normal banking

Electric Utility Routines

Groups of Electric Utility Routines were tailored to needs of individual utility companies. These programs are designed to compute load flow, optimal loading, load duration, gas flow and pressure, and short circuit conditions.

Page 21

SYSTEM CONTROL OF INPUT-OUTPUT PERIPHERAL DEVICES

In a GE-235 system up to ten input-output (I/o) devices of various types may operate simultane-

ously with the central processor and with each other. This truly active configuration is extremely flexible and efficient; it is capable of a maximum throughput of 55, 000 20-bit words per second, plus card reading, per second, plus card reading, card punching and printing, by the addition of optional dual access 1/0 controller selector channels.

This performance is the result of two significant design concepts:

Each 1/O device controls itself and executes its own 1/0 commands.

All central processor have access to memory on a time-sharing basis.

card punching and printing. The throughput may be doubled to 110, 000 words

I/O

devices (with exception of perforated tape reader/punch and typewriter) and the

INDEPENDENT-CONTROL OF

The

individual operation of each 1/0 device is determined by the controller through which it attached to the system. The controller receives the commands appropriate to it (such commands as to start, these commands without further instruction. central processor

stop, edit data, and rewind tape are typical) from the central processor and executes

free to continue with the succeeding item in the program.

TIME-SHARIIVG OF MEMORY ACCESS BY

Orderly and efficient time-sharing of memory access among the central processor and 1/O

devices of the GE-235

Allowing only one system element to have access to memory at one time. Allowing each element access to memory on a priority schedule when it needs it and

causing

These conditions are satisfied in the GE-235 by the built-in priority control logic. The success of this feature, in fact, accounts for the high efficiency and capacity of the system and the simultaneity of operation of

ensured by:

to relinquish access when it does not.

its

Garious elements.

1/0

DEVICES

Thus, having given a command to a controller, the

1/0

DEVICES

Priority Control and Time-Sharing

The GE-235 priority control feature is shown schematically in the accompanying block diagram. For the numbered channels shown on the diagram, descending priority is from left to right (from 0 to

6).

The channel assigned to a particular 1/0 device depends upon its information transfer

that

is,

rate

its memory access requirements.

For

example, a high-speed printer has a lower priority requirement than a magnetic tape unit,

since a tape controller cannot wait as long for access as a printer controller can without causing

timing errors. If the printer had the higher priority, it could possibly monopolize access to

III-

Page 22

memory at the expense of the tape unit without increasing its effective printing speed. Thus, the

I/O

most efficient operation follows from assigning the highest-priority channel to the

device

with the highest access speed.

GE-235

MAIN

rlrl

a2."2E

gaaa

PRIORITY CONTROL LOGIC

CARD

READER

,CONTROLLER

A recommended assignment of priorities is:

MRADS Controller Magnetic Tape Controller or MRADS Controller

Magnetic Tape Controller

Document Handler Adapter Datanet Controller, Document Handler Adapter or High-Speed Printer Controller

6 High-Speed Printer Controller

CARD

PUNCH

CONTROLLER

CENTRAL

PROCESSOR

Page 23

Examples of Time-Sharing

Three examples of GE-235 equipment configurations are shown below, with the relative amounts

I/O

of time consumed in each case by

activity and internal computation.

Example Read cards at 400 cards per minute

Read magnetic tape at 15,000 characters

per second (500-character record)

Print at 900 lines per minute (edited) Total

Percent of total time left for computing:

Example Read cards at 400 cards per minute

Mass random access data storage (read

and write cycle at 200 milliseconds)

Print at 900 lines per minute (edited) Total Percent of total time left for computing:

Example

High-speed card reading at 850 cards per minute Read magnetic tape at 41,600 characters per

second (500-character record) 4.2

Write magnetic tape at 41,600 characters per

second (500-character record) 4.2

Print at 900 lines per minute (edited)

Percent of Total Time

96.9*

98.6*

Total Percent of total time left for computing:

Since execution of many instructions does not require access to memory, actual computing

time may range upward from the amount shown.

90.2*

Overall Effect of Time-Sharing

The overall effect of the GE-235 time-sharing arrangement is to create the most efficient balance

I/O

between

operations and internal computation, regardless of the type of application.

Page 24

PERIPHERAL SWITCH CONTROL UNIT

The Peripheral Switch Control Unit

many as seven 1/0 device controllers between two GE-200 Series systems. Any

normally connected to one of the numbered priority control channels may be switched by this unit. This switching capability permits optimum utilization of all peripheral equipment for multiple system operation.

The heart of the unit is the switch control console (see the accompanying illustration). The console is console performs two functions: 1) assigns each controller to the desired central processor and 2) assigns the selected priority to each 1/0 device.

Two lighted pushbuttons are associated with each 1/0 device--one on the left of the panel

(SYSTEM Separating each pair of SELECT buttons the priority level of the associated 1/0 device. Additional circuitry actuates one of the ADDRESS SELECT ERROR indicators if more than one controller one system. A vertical column of eight numbered panel indicators marked SELECTED ADDRESS also appears on each side of the panel. on each system.

connected to each

SELECT PERIPHERAL) and one on the right (SYSTEM

an optional feature that makes possible the switching of as

I/O device

1/O device to be switched and to the two central processors. The

SELECT PERIPHERAL).

arow of ADDRESS SELECTION buttons for determining

switched to the same priority level on

These indicators show which priority levels are in use

1/0 DEVICE SWITCH CONTROL UNIT CONSOLE PANEL

Page 25

110 DEVICE SWITCH CONTROL

UNIT

CONSOLE

Page 26

CENTRAL PROCESSOR

Frr-

transistorized, single address, general purpose digital computer.

accepts and p/rocesses information from punched cards, magnetic tape, perforated tape, magnetic ink character recognition equipment, communications systems and other peripheral equipment.

provides output to magnetic tape, punched cards, perforated tape, high speed printer,

mass random access data storage, data communications systems and other output media.

Control Console

provides for complete operator control and communication with the system. displays contents of the significant registers and provides control signals to the operator

for the effective monitoring of system operations.

mass random access data storage, data

IV-

Page 27

Console Typewriter

accepts instructions from the operator.

prints instructions to the operator.

monitors system operations.

prints direct output from the central processor.

MAGNETIC CORE MEMORY

The memory portion of the Central Processor is the immediate-access storage element for the GE-235 system. Both the data to be processed and the controlling instructions are held in memory and are called for by the control unit as required.

The memory is composed of magnetic cores

storing one unit of information, referred to as a bit. The basic unit of memory storage is the

"word"

address.

The size of the basic memory is 4096 words. The memory design allows for an expansion to an

8192-word memory or to a 16,384-word memory without the necessity of expensive retrofits. The 16,384-word memory is divided into two groups of 8192 words each, referred to as the upper

bank and the lower bank.

Minimum instruction word access and execution time is 6 microseconds. to or from memory, including the instruction word time, is accomplished in 12 microseconds; a double word transfer is made in 18 microseconds. are made in one-word parallel form; that is, the word bits are transferred simultaneously.

Internal ferred to memory, and by recomputing and verifying that parity bit when the word is read from memory. The effect of a console

each word consisting of 20 bits plus a parity check bit. Each word has its own unique

checking is accomplished by generating and storing a parity bit when a word is trans-

puity error may be controlled to meet the needs of the application by

STOP/RUN switch operation.

.050 inch in diameter; each core is capable of

data word transfer

The transfer of words to and from memory

WORD FORMATS

The GE-235 can process data in either binary or alphanumeric form. This feature permits both modes of operation to take advantage of the particular characteristics of a given application.

Alphanumeric (BCD) Words

When cards punched in Hollerith code are used as computer input, the information contained in

each

of the 80 columns is automatically converted into a six-bit binary-coded-decimal (BCD)

character. Thus, 3 alphanumeric characters occupy 18 of the 20 bit positions of an alphanumeric

Page 28

data word. Double length word operations permit the automatic handling of six alphanumeric characters with a single instruction. These convenientword sizes eliminate the need for elaborate

partial word facility. Information must be in the BCD format prior to printing or typing.

The word below illustrates

how 3 random characters are represented in a six-bit (BCD) format

within one word:

The table below illustrates the range of alphanumeric characters which can be represented within

any six bit positions within any word of storage:

Alphanumeric Character

Zone

Bits Numeric Bits

-+v-v-Y-w--vYv

0000 0001 0010 0011 0100 0101 0110 0111 1000 1001 1011 1100 1101 1110

GE-235

Alphanumeric Characters

Binary

binary data word consists of

Words

bits plus the sign bit. For example, the decimal number

+49

is represented in binary form as:

Negative binary numbers are expressed in 2's complement form. For example, the decimal

-10

number

is represented in binary form as:

(minus)

IV-

Page 29

A 20-bit word can accommodate a range of decimal values from -524,288 to +524,287, sometimes referred to as 5-1/2 decimal digits. Double length word operations permit the efficient processing of decimal values between associated with representing numeric data in the true binary format are:

Memory storage efficiency.

An effective increase in the rate with which numeric data can be transferred to and from

magnetic tape.

Increased speed of arithmetic and data handling operations.

+274,877,906,943 and -274,877,906,944. The advantages

Binary and BCD information may be intermixed in memory so as

mode of representing each field in each application, Subroutines are provided to convert numeric

data from BCD to binary, or vice-versa.

to provide the most efficient

Decimal Arithmetic

An optional Decimal Arithmetic feature of the GE-235 provides a decimal add and subtract capability without requiring BCD to binary conversion. These arithmetic operations can be

performed on single decimal words of 3 digits or on double length decimal words of 6 digits.

Automatic carry is provided for larger fields. numbers

+368 and +I5896 would appear in memory.

The examples below illustrate how the decimal

Floating Point Arithmetic

The GE-235 Auxiljary ~ritimetic Unit can be used to advantage in scientific and engineering applications where numerous floating point or double precision arithmetic calculations are required. The logic of the AAU performs double precision fixed point and floating point arithmetic more efficiently than is possible when using mathematical subroutines. However, subroutines may be preferable when floating point arithmetic is done on a limited basis.

Three floating subtraction, multiplication, and division may be done under any of the three modes of operation. All arithmetic is performed in binary mode.

modes of calculation may be performed by the Auxiliary Arithmetic Unit: unnormalized

point, normalized floating point, and fixed point double precision operations. Addition,

Page 30

During floating point calculations, the data to be operated upon by the internal logic of the

comes

from the main memory of the Central Processor.

floating point number occupies two

words of memory storage and assumes the following format:

Word One

AAU

Bits

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Exponent

Word Two

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Se = Sign Sm= Sign of Mantissa

Exponent

(0 (0

= =

plus;

Mantissa

minus)

The binary point is assumed to be placed before the mantissa. This format produces a binary number with a 30-bit mantissa and a binaryexponent range of -256 to equal to a decimal format of a 9-digit mantissa and a decimal exponent range of -76 to

+255. This is approximately

+76. The

use of two words allows one of the sign positions to be applied to the exponent which, in turn,

allows the use of the full range of the exponent.

The word format for fixed point double precision words in memory

Word One

Bits

as follows:

Sign of Word One

Fixed Point Double Word

Bits

S = Sign of Word Two

lp-lg

Word Two

the Sign of

Page 31

Instruction Words

With the exception of certain peripheral equipment operations, a GE-235 instruction is a single-

address word consisting of 20 bits. Except for branching operations, instructions are executed

sequentially;

the reading of the next instruction from memory occurs after the execution of the

current instruction.

The basic format of the instruction word

Bits

DO THIS

Bits

Opera tion

Code

is:

4 5 6 7

WITH DATA LOCATED HERE

4567

X X

Operand Address

Bits 0 through 4 designate the operation to be performed, bits 5 and 6 indicate whether the instruction is to be automatically modified (indexed), and bits

through

indicate the operand

address.

Because the

five bit positions allowed in the instruction word for the operation code can define

only 32 operations, additional bit positions are required to define the more than 300 instructions

in the repertoire of the GE-235.

This is achieved by using bit positions in the operand address

field for instructions that require only a limited portion of that field.

REGISTERS

With the exception of shift instructions, all information transfers between Central Processor

registers, between the registers and memory, and between the registers and the adder occur in

parallel. That is, all 20 bits comprising words in transit or being operated upon arithmetically

are transferred at the same time.

The

It performs this function by holding:

The augend during addition. The sum after addition. The minuend during subtraction.

IV-

Page 32

The result after subtraction. The most significant half of the product after multiplication. The most significant half of the dividend before division. The quotient after division. The most significant half of a double wordafter the execution of all double length instructions. The word to be shifted during various shift instructions. The word transferred to or from memory or another register or to be modified in some way

during the execution of various data transfer instructions.

The word on which tests are performed during the execution of branch instructions.

The

remainder after division and acts with the

a 20-bit register which holds the multiplier during multiplication and the

during the execution of double word operations. When double word instructions are used, the

least significant half of the double word is automatically transferred into the the

Holds the multiplier during multiplication. Holds the least significant half of the product after multiplication. Holds the least significant half of the dividend before division. Holds the remainder after division. Holds the least significant half of a double word after the execution of all double length

instructions.

Shifts in conjunction with the

and A registers in special shift instructions.

The N register is a 6-bit register which is used as a single character (BCD) buffer between the computer and the typewriter and the perforated paper tape reader and punch. Information

transferred directly between the N register and the A register.

The I register is the instruction register. It contains the 20 bits of an instruction word during

execution of a computer command. While instructions are being processed, bits

0 through

indicate the operation to be performed and bits 5 and 6 control the indexing of instructions when

specified by the program. During the execution of instructions involving reading an operand

from memory (and after appropriate indexing has taken place), bits

through

contain the

memory location of the operand.

Page 33

Index

Words

Index words are locations in memory (0000, 0001, 0002, 0003) reserved for use as counters or for the automatic modification of instructions in the

Unlike GE-235 registers, the index words are integral parts of memory and not separate physical

storage devices.

index words.

128

An optional feature adds 31 index groups of

locations each, giving a total of

GE-235 REGISTER RELATIONSHIPS - BLOCK DIAGRAM

Console typewriter paper tape reader

and punch

Page 34

Counter

The P counter

a 15-bit sequence control counter that contains the memory address of the next instruction to be executed. The contents of the P counter are incremented by one after the current instruction has been selected from memory and placed in the

normally indicates the address of the next instruction in sequence. The contents of the P counter

can be set from the

gram.

M Register

The M register memory and the Central Processor or peripheral equipment. When a word enters the from the Central Processor or punched card equipment for recording in memory, a parity bit

a 21-bit register that acts as an input-output buffer between the magnetic core

computed and 21 bits are storedin memory. Peripheral controllers that are connected through the numbered priority control channels generate a parity bit which to the storage of the 21-bit word in memory. When a word register, parity

again computed and the new parity bit is compared with the one already existing

checked in the M register prior

read from memory into the

to ensure accuracy of data transfers.

Adder

The adder of the Central Processor is a high-speed, parallel, binary adder network that executes

the calculations specified by the instruction code in the

Real Time Clock

The Clock, or C register, Central Processor, the C register is automatically incremented by

a 19-bit register (there is no sign bit). While power is applied to the

binary one every sixth of a second. When the count reaches 518,400, the decimal equivalent of 24 hours in sixths of a second, the C register

The real time

automatically reset on the next increment to all zeros and starts counting again.

clockis loadedfrom the A register, and its contents can be read by transferring them to the A register. The clock can be set or read either by program or by manually inserted instructions.

AUXILIARY ARITHMETIC UNIT

The addition of the GE-235Auxiliary Arithmeticunit (AAU) extends the arithmetic capability of the

GE-235 system. This device, with built-in logic, facilitates floating point and double precision arithmetic through its increased capacity and the speed with which it computes. The Auxiliary Arithmetic Unit contains two 40-bit registers, AX and QX, which correspond functionally to the A and

Q registers in the Central Processor.

and divide extremely large numbers

represented in either fixed or floating point format. At the

option of the programmer, the unit functions in three modes: normalized floating point,

Eighty bits permit the unit to add, subtract, multiply

unnormal-

ized floating point, and fixed point computations. Underflow and overflow conditions are checked

by means of programmed interrogation.

Page 35

The Auxiliary Arithmetic Unit is not a peripheraldevice; it is an extension of the basic arithmetic unit of the Central Processor. All peripheral operations may occur concurrently with the use of this device.

When data in the two-word floating point format enters the Auxiliary Arithmetic Unit, it is converted into one 40-bit word and stored in the AX register. (Number designations for bits in the AAU start with number 1 rather than 0 as used in memory.) The word exists in the 40-bit AX register as:

Bits

Exponent Mantissa

Se = Sign of Exponent

10 20 21 22 40

Sign of Mantissa

Mantissa

The QX register is an extension of the AX register. It also consists of an eight bit exponent with sign and a thirty bit mantissa with sign. The value of the exponent of the QX register

the value

of the AX register minus 30.

When data in the fixed point double precision format enters the Auxiliary Arithmetic Unit, the two

words from memory (Y and

are converted into one 40-bit word and stored in the AX

Bits

2 20 21 22 40

S = Sign

Y+l

An instruction for the Auxiliary Arithmetic Unit is contained in one 20-bit word identical to the

format of Central Processor instructions.

ADDRESS MODIFICATION (INDEXING)

The GE-235 achieved

capable of automatic address modification under program control. This is

through the use of the special index words in memory locations 0001, 0002, and 0003.

Memory location 0000 is not used for address modification, but it can operate as a counter.

Thirty-one additional index word groups are available as an optional feature. tains four used either as counters or for automatic address modification. (groups are numbered

index words; the first word may be used as a counter, the other three words may be

The total of 32 index word groups

0 - 31), if included in the system, occupy memory locations 0000 through

Each group con-

0127. Any of these locations may be used for normal storage if not required for address modification in a program. Thus, the GE-235 features

total of

index words for the address modifi-

cation of instructions and a total of 128 index words which may be used as counters and incremented and tested by special index word instructions.

Page 36

Bit positions 5 and 6 of an instruction word designate the index word to be used in modifying the operand address portion of the instruction. Binary configurations select index words as follows: 01

location 0001, 10 = location 0002, and 11 = location 0003. Bits 00 in positions 5 and 6 indicate that no modification is to be performed, If the additional 31 index word groups are available, an instruction is provided to allow the selection of any one of the 32 groups at any point in the program.

an instruction in the I register calls for automatic address modification, an extra word time

(6 microseconds) is required to accomplish the operation.

The sequence of events in an automatic address modification is:

The instruction to be executed is read into the I register.

Bits

through 19 of thedesignatedindexword and bits 7 through 19 of the I register are added

together and the

result replaces bits

through 19 of the I register. Bits 0 through 4 of the

I register are unchanged.

The modified instruction in the I register is executed. The original instruction in memory is not changed.

INSTRUCTION REPERTOIRE

The descriptions on the following pages are those of instructions which may be completely executed by the Central Processor without assistance from any of the other system units. Descriptions associated with peripheral units may be obtainedelsewhere in the manual following the description of each peripheral device. Instructions in this section have been grouped under the following headings:

Data Transfer

Arithmetic Shift Branch

Address Modification

Auxiliary Arithmetic Unit Real Time Clock

Each description gives the mnemonic operation code for the instruction and an indication of whether a memory location

(Y)

or a constant

(K)

required in the operand address portion. Some instruc-

tions do not require an operand, so for these no entry will be found under the Operand Address

column. The letter

under the Index column indicates that the instruction can be indexed. Exe-

cution times are given in microseconds. The execution times include the fetching of the instruction

and data.

In all instructions which involve the bringing of a word from memory, the word in memory remains unchanged. For most instructions involving the transfer of information from registers, the condition of the register after execution

unchanged.

Page 37

FOR THE PURPOSES OF THIS MANUAL, THE DESCRIPTION OF EACH INSTRUCTION 1s NECESSARILY ABBREVIATED. COMPLETE EXPLANATIONS WITH EXAMPLES MAY BE

FOUND

THE GE-235 PROGRAMMING REFERENCE MANUAL.

Data Transfer

Mnemonic Oper. Code Address

LDA Y X 12 LOAD A register with the contents of

DLD

STA Y

DST

Operand

Index Time

Word Microsec. Description

Y X 18 DOUBLE LENGTH LOAD registers A and

contents of Y and

12 STORE A register contents into memory location Y.

X 18 DOUBLE LENGTH STORE contents of registers A

Q into Y and Y+1.

and

Y+1.

Q with the

MOV Y 24+12N MOVE a block of N words starting at memory lo-

cation Y into another area of memory. Register A contains starting address of new area; register

contains number of words to be moved (in 2's complement form). This instruction is an optional feature.

ST0

18 STORE OPERAND ADDRESS field of register A in

the operand address field of Y.

ORY

EXT

OR A INTO Y byplacing a A has a

bit in the corresponding position.

18 EXTRACT into A by placing a

wherever Y has a

bit in Y wherever register

bit in register A

bit in the corresponding position.

LDO

18 LOAD A FROM

18 LOAD

into A.

FROM A by transferring the contents of

Q by transferring the contents of

18 MOVE A TO Q and replace the contents of A with

zeros.

EXCHANGE A AND of registers A and

Q by interchanging the contents

18 LOAD ZERO into all bit positions of register A. 18

LOAD ONE bit into 19th bit position of register A and set the other bit positions to

IV- 12

Page 38

Mnemonic

Operand

Oper. Code Address LMO

Inex

Word Microsec.

Time

LOAD MINUS ONE into register A by setting all bit

positions to

18 COMPLEMENT register A by replacing each 1 bit

with

Description

and each 0 bit with

NEGATE A by replacing the contents of register A with the 2's complement.

12 CHANGE SIGN of register A by replacing

position with

and a 0 bit in sign position with

NOP 18 NO OPERATION; next

Arithmetic

Mnemonic Operand Index Time Oper. Code Address Word Microsec.

ADD

12 ADD the contents of Y and register A algebraically;

Description

sum in A.

DAD

18 DOUBLE LENGTH ADD the contents of Y and Y+l

and registers

ADD ONE (plus one) algebraically to the contents of

and Q algebraically; sum in A and Q.

18 SUBTRACT the contents of Y algebraically from the

contents of register

DOUBLE LENGTH SUBTRACT the contents of Y and Y+l algebraically from the contents of registers A and

result in A and

instruction

executed.

bit in sign

MPY

DVD

SUBTRACT ONE algebraically from the contents of register A.

54-138 MULTIPLY the contents of Y algebraically by the

contents of register gisters

and Q.

Q, placing the product in re-

156-174 DIVIDE the contents of registers A andQ algebraically

by the contents of Y, placing the quotient in A and the

remainder in Q.

IV- 13

Page 39

OPTIONAL DECIMAL ARITHMETIC

Mnemonic Operand Oper. Code

SET

ADD

DAD

ADO

SUB

DSU

Address BINMODE

DECMODE

Index Time

Word Microsec.

30 DOUBLE DECIMAL SUBTRACT the contents of

SET BINARY MODE, permitting arithmetic instructions to operate in the binary mode as detailed above. (When power automatically in the binary mode.)

SET DECIMAL MODE, permitting arithmetic instruc-

tions to operate in the decimal mode as detailed below.

DECIMAL ADD the contents of

algebraically; sum in A.

DOUBLE DECIMAL ADD the contents of Y and and registers A and Q algebraically; sum in A and

ADD ONE DECIMAL (plus one) algebraically to the contents of register A.

DECIMAL SUBTRACT the contents of Y algebraically from the contents of register A.

and A and

Description

turned on, the computer will be set

and register A

Y+l

algebraically from the contents of registers

result in A and

Y+l

SBO

Shift

The shift instructions shift the contents of register A to the right or left serially either alone or with the contents of the N and/or two bit positions or less; six additional microseconds are required for each additional 3-bit shift or fraction thereof.

Mnemonic Operand Oper. Code

SRA

SRD

SLA

Address

Index Word Microsec.

registers. Twelve microseconds are required for a shift of

Tim

12+

SUBTRACT ONE DECIMAL algebraically from the contents of register A.

Description

SHIFT RIGHT A. The contents of register A are shifted right

SHIFT RIGHT DOUBLE. The contents of registers A

and

shifting out of A into

SHIFT LEFT A. The contents of register A are

shifted left

places.

together are shifted right K places, with bits

places.

Page 40

Mnemonic Operand Index

Oper. Code SLD K

~hdress Word

Time

Microsec.

12+

Description

SHIFT LEFT DOUBLE. The contents of registers A

and Q together are shifted left K places, with bits shifting out of Q into A.

SCA

SCD

SAN

NAQ

NOR

12+

18+

SHIFT CIRCULAR A. The contents of register A are

shifted SHIFT CIRCULAR DOUBLE. The contents of registers

A and Q together are shifted K places to the right in a circular fashion, with bits shifting out of A into Q and out of Q into A.

SHIFT A AND N RIGHT. The contents of registers A

and N together are shifted K places to the right, with

bits shifting out of A into N. SHIFT N AND A RIGHT. The contents of registers N

and A together are shifted K places to the right, with bits shifting out of N into A.

SHIFT A INTO N AND Q. The contents of register A are shifted K places to the right into both registers N and Q.

SHIFT N, A, AND Q RIGHT. The contents of registers N, A, and Q together are shifted Kplaces to the right.

NORMALIZE. The contents of registerA are normalized by shifting left and eliminating leading zeros up to K places.

places to the right in a circular fashion.

DNO

Branch

Mnemonic Oper. Code

BRU

SPB

Operand Address

18+

Index Time Word Microsec.

DOUBLE LENGTH NORMALIZE. The contents of registers A and Q are normalized by and the elimination of leading zeros in register A up to K places.

BRANCH UNCONDITIONALLY. Control is transferred to the instruction at memory location

STORE P AND BRANCH. The locationof this SPB instruction replaces the contents of index word X and control

location Y.

then transferred to the instruction at memory

double shift left

IV-

Page 41

The following branch instructions test to determine whether a particular internal condition or false. false, the second sequential instruction is executed.

If the condition tested is true, the computer executes the next sequential instruction; if

true

Mnemonic Oper. Code

BPL 12 BRANCH ON PLUS. Register A tested for plus sign. BMI 12 BRANCH ON MINUS. Register A testedfor minus sign. BZE 12 BRANCH ON ZERO. Register A tested for all zeros. BNZ

BOV

BNO

BOD 12 BRANCH ON ODD. Register A tested for odd value. BEV 12 BRANCH ON EVEN. Register A testedfor even value. BPE

BPC

Operand Address

Index Time Word Microsec.

12 BRANCH ON NON-ZERO. Register A testedfor not all

12 BRANCH ON OVERFLOW. Overflow indicator tested

12 BRANCH ON NO OVERFLOW. Overflow indicator

12 BRANCH ON PARITY ERROR. Parity error indicator

12 BRANCH ON PARITY CORRECT. Parity error

Description

zeros.

for ON.

tested for OFF.

tested for ON.

indi-

cator tested for OFF.

CAB Y

(Optional feature)

DCB Y

(Optional feature)

12-24 COMPARE AND BRANCH. Register A compared alge-

braically with contents of Y. If contents of Y are

greater, the next instruction is executed; if contents of Y and A are equal, the second sequential instruction is executed; if contents of Yare less, the third sequential instruction is executed.

12-24 DOUBLE COMPARE AND BRANCH. Registers A and

together are compared algebraically with contents of

Y+1. Branching occurs in an analogous manner

Y and

to the CAB instruction above.

Page 42

Address Modification

Mnemonic Operand Index Time

Oper. Code Address Word Microsec.

18 INCREMENT X. The constant

Description

added absolutely to the contents of index word X, and the result replaces the contents of X.

LDX

STX

SXG

Auxiliary Arithmetic Instructions

AAU instructions for floating point operations assume that the operands to be acted upon are

already in floating point format. The operands are put in floating point format by means of a subroutine furnished for this purpose.

X 18 LOAD X. The contents of

X 18 STORE

12 INDEX GROUP SELECT. If additional index word

BRANCH IF X IS HIGH OR EQUAL. Contents of index word X are tested for an equal to or greater than condition. Branching follows the normal sequence of branch instructions.

BRANCH IF X IS LOW. Contents of index word X are tested the normal sequence of branch instructions.

groups are available, this instruction selects one of the 32 possible groups.

particular group.

for a less than

The contents of index word X are stored

condition. Branching follows

are placed in index word

(00 - 31) specifies the

Before giving an arithmetic instruction to the AAU it is necessary to set the mode of operation by one of the following SET instructions:

Mnemonic Operand Oper. Code Address

FLPOINT

NFLPOINT

FMPOINT

Index Time Word Microsec.

Description

This instruction sets the format for unnormalized

floating point operations. This instruction sets the format for normalized

floating point operations. This instruction sets the format for double word

fixed point operations.

IY-17

Page 43

The mnemonics for the following AAU instructions consist of the normal three alphabetics plus the tag A in GAP coding sheet column 20. For example, LAQ A moves the contents of the QX

in the central processor.

AAU while LAQ without the A tag moves the contents of Q to A

Mnemonic Oper. Code

LQA

The memory address address modification, must be greater than 001

Mnemonic Oper. Code

FLD

FST

Operand Address

Ope rand Index Time Address Word Microsec. Description

Index Time Word Microsec.

6 LOAD AX FROM QX. The contents of register QX

are moved into AX. QX is unchanged.

6 MOVE AX TO QX.

moved into QX. AX

6 LOAD QX FROM AX. The contents of register AX

are moved into QX. AX is unchanged.

6 EXCHANGE AX AND QX. The contents of registers

(Y)

designated in an AAU instruction, as it appears in the I register after

X 18 LOAD AAU. Contents of Y and

X 18 STORE AAU. Contents of register AX are stored

into

Description

The contents of register AX are

cleared to zeros.

and QX are interchanged.

and

Y+1.

Y+l

are loaded into

FAD

24-36 NORMALIZED FLOATING POINT ADD. The floating

point number in the floating point number in register AX. The result is in AX in normalized form.

24-30 UNNORMALIZED FLOATING POINT ADD. Same as

normalized floating point add except the result is placed in registers AX and QX and may or may not be in normalized form.

24 DOUBLE WORD

and

Y+l

of register AX and the result

38 bit fixed point number.

and

Y+l

added algebraically to

FIXED POINT ADD. The contents

are algebraically added to the contents

placed in AX as a

Page 44

Mnemonic

Oper. Code

Operand

Index Time

Address Word

Microsec.

FSU

FMP

NORMALIZED FLOATING POINT SUBTRACT. The

floating point number in Y and Y+l

subtracted algebraically from the floating point number in register AX. The result

placed in AX in normalized form. UNNORMALIZED FLOATING POINT SUBTRACT. Same as normalized floating point subtract except the result

placed in registers AX and QX and may or may not be in normalized form. DOUBLE WORD FIXED POINT SUBTRACT. The contents of

and Y+l are algebraically subtracted from the contents of register AX and the result is placed in AX as a 38 bit fixed point number.

NORMALIZED FLOATING POINT MULTIPLY. The floating point number in Y and Y+l is algebraically multiplied by the floating point number in register

QX. The 60-bit product of the two mantissas is normalized. The most significant half of the normalized product is stored with its exponent in AX; the least significant half is stored in QX (the exponent 30 less than the floating point exponent in AX). UNNORMALIZED FLOATING POINT MULTIPLY. Same as normalized floating point multiply except

the result is placed in registers AX and QX and may or may not be in normalized form. DOUBLE WORD FIXED POINT MULTIPLY. The contents of Y and

Y+l are algebraically multiplied by

the contents of register QX, giving a 76-bit product

and of the product is stored with the least significant half

identical sign bits. The most significant half

sign bits in AX, and

stored with 2 sign bits in

QX.

FDV

NORMALIZED FLOATING POINT DIVIDE. The floating point number in registers AX and QX

algebraically divided by the floating point number in

and Y+1. The normalized quotient

stored in AX,

and the remainder, which may or may not be nor-

malized,

stored in QX. UNNORMALIZED FLOATING POINT DIVIDE. Same as normalized floating point divide except the quotient that is stored in AX may or may not be in normalized form. DOUBLE WORD FIXED POINT DIVIDE. The contents of registers AX and QX are divided by the contents of Y and

Y+1. The quotient is stored in AX

and the remainder in QX.

IV- 19

Page 45

Mnemonic Operand O~er. Code Address

BAR XXX

Index Time Word Microsec.

Description

BRANCH ON AAU INTERROGATED CONDITIONS.

The BAR instruction interrogates the AAU for spe-

cific conditions. The condition tested

a mnemonic placed in the operand address field and

indicated

condition tested next sequential instruction. If false, the second sequential instruction require a

XXX in the format given here. If the

true, the computer executes the

executed. BAR instructions

in GAP coding sheet column 20.

indicated

BAR

BPL

BMI

BZE

BNZ

BOV

BNO

BU F

BNU

BER

12 BRANCH ON AAU PLUS. Register AX

plus sign.

12 BRANCH ON AAU MINUS. Register AX

a minus sign.

12 BRANCH ON AAU ZERO. Register AX

all zeros.

12 BRANCH ON AAU NOT ZERO. Register AX

tested for not all zeros.

12 BRANCH ON AAU OVERFLOW. AAU

overflow indicator ON.

12 BRANCH ON AAU NO OVERFLOW. AAU

for overflow indicator OFF.

12 BRANCH ON AAU UNDERFLOW. AAU

underflow indicator ON.

12 BRANCH ON AAU NO UNDERFLOW. AAU

for underflow indicator OFF.

12 BRANCH ON AAU ERROR. The error indicator is

tested for ON.

tested for a

tested for

tested

BAR

BNE

BOO

BON

BUO

BUN

12 BRANCH ON AAU NO ERROR. The error indicator

tested for OFF.

12 BRANCH ON OVERFLOW HOLD ON. AAU is tested

for overflow indicator ON.

12 BRANCH ON OVERFLOW HOLD NOT ON. AAU

tested for overflow indicator OFF.

12 BRANCH ON UNDERFLOW HOLD ON. AAU

for underflow hold indicator ON.

12 BRANCH ON UNDERFLOW HOLD NOT ON. AAU

tested for underflow hold indicator OFF.

tested

Page 46

Mnemonic

Oper. Code

Operand Index Time Address Word

Microsec.

Description

ROV

RUN

RIN

RESET OVERFLOW HOLD. The overflow hold indi-

cator

RESET UNDERFLOW HOLD. The underflow hold indicator

RESET INDICATORS. The underflow hold and overflow hold indicators are turned off.

turned off.

Real Time Clock Instructions

Mnemonic Operand Oper. Code

LAC

LCA

Address

Index Time

Word Microsec.

Description

The contents of register A are replaced by the contents of register C and the sign of

Bit positions

through 19 of register A are placed in

registerA is set to zero.

CONTROL CONSOLE

The GE-235 Control Console provides the operator with the controls and indicators necessary for direct manual and visual communication with the Central Processor. Basic alarm conditions are clearly displayed to facilitate accurate operator corrective action. The detailed contents of arithmetic, instruction and program counter registers are displayed. Changes or deletions of instructions or data can be made through convenient toggle switches.

Automatic and manual control modes allow complete operator control of programs. In the auto-

matic mode, the computer executes instructions in the normal sequential manner. In the manual

mode, the computer executes instructions in a step-by-step procedure, going from one instruction

to the next under operator control.

The lower third of the console consists of the Control Panel which contains operating switches that

control Central Processor power, operating mode selection, alarm and register resetting, initial

program loading, and starting and stopping of computer operation.

The upper two-thirds of the Control Console

Switches. The Indicator Panel consists of display lights for the contents of the A,

and the P counter and for the location of the current selected index word group. Toggle switches are provided to allow for the manual entry of information into the A register. In addition, various alarm and ready status indicators that are essential for proper control are provided, as well as an indicator showing the controller that is actively accessed at any given time. For Central Processors equipped with the optional AAU, additional AAU switches and indicators appear above the

standard indicator panel.

occupied by the Indicator Panel and Register Control

and I registers

Page 47

No00

000

66d

CYYUTLR

666

bc)6

R15ET

6QQ QQQ QQQ @Q@ QQQ QQQ QQQ

660 666

OPERA,#OM

,No,X

CONTROL

66iS

dbi4

CONSOLE

666

LDDREIl

666

b66

666

Page 48

INPUT/OUTPUT TYPEWRITER

The GE-235 Input/~utput Typewriter can transmit input data to the Central Processor from the keyboard while simultaneously printing the data. Communication with the Central Processor during

both input and output operations programmed single-character (BCD) operations. Output

through the N register. All input and output operations are

at the rate of 15 characters per second.

The typewriter keyboard has

writer characters include:

Functions that can be programmed are:

Space Carriage Return Tabulation Print Red Print Black

Ignore

Index

44datakeys, a space bar and standard functional control keys. Type-

ALPHABETICS A THROUGH NUMERIC DIGITS 0 THROUGH SPECIAL CHARACTERS AS FOLLOWS:

Console

Mnemonic Operand Oper. Code

RCS

TON

TYP

KON

And

Typewriter Instructions

Address Word Microsec.

Index Time

12 READ CONTROL SWITCHES, Each of the 20 console

12 TYPEWRITER ON. The typewriter

12 KEYBOARD ON. The typewriter input keyboard

Description

control switches for register

switch is down (ON), a sponding position of A.

nected to the N register. TYPE. If the typewriter

the N register

typed.

bit

ON, one BCD character in

are examined.

placed in the corre-

logically con-

activated. Striking of a key will load the corresponding BCD character into the N register.

Page 49

Mnemonic Operand

Oper. Code Address Word

Index Time

Microsec.

Description

OFF

BNN

BNR

AUTOMATIC PROGRAM INTERRUPT

Automatic operation of two or more unrelated programs. When the API the peripherals needed to process an additional program are allowed to interrupt the main program.

Information can be sent to, or takenfrom, a peripheral. Then, while the peripheral

this information, control can be returned to the main program until the peripheral becomes "ready"

to send or receive more information. The programmer can prevent the main program from being interrupted through execution of a Priority Break instruction.

The API monitors the card reader, cardpunch, and the controllers connected through the numbered

priority control channels (magnetic tape, printer, document handler, mass random access data

storage, and Datanet-15). An position before that controller can interrupt the program through the API. This switch effectively inhibits interruptions from main program peripherals. The

Program Interrupt

externaloff/on switchon the peripheral controller must be in the ON

input/output devices cannot interrupt the program through the API.

12 OFF. All N register peripherals are logically discon-

nected from the N register, and the power supply for

the perforated paper tape punch

12 BRANCH ON N REGISTER NOT READY. The N

12 BRANCH ON N REGISTER READY.

tested for the ready condition. Branching follows the

normal branch instruction sequence.

(API)

an optional feature designed to control the simultaneous

tested for the not-ready condition. Branch-

set by programmed instruction,

input/output typewriter and other N

turned off.

TheN register is

operating on

Automatic-Program Interrupt Instructions

Mnemonic Operand 0per.Code

SET

Address

PST 12 SET PRIORITY START. The Central Processor enters

PBK 12

1nde; Time

Word Microsec.

the API mode, where it will remain until a SET PBK instruction is executed.

SET PRIORITY BREAK. The Central Processor leaves the API mode and will not return until a SET PST instruction is executed.

Description

Page 50

GE-235

The

GE-235 information processing system features two types of card readers and two types of

card punch units with which to process tabulating cards.

PUNCHED

CARD

EQUIPMENT

CARD

READERS

400

CARD

PER MINUTE UNIT

Punched cards may be read in decimal (Hollerith), 10-row binary, and 12-row binary modes.

1000

CARD

PER

MINUTE UNIT

Card readers can operate simultaneously with central processor computation and other input/output operations.

Cards are read serially, column by column, by photocells, providing reliable and accurate card reading in either continuous or demand type operations.

Intermixed Hollerith and 12-row binary coded cards can be read by detecting a in column one of binary cards.

7-9

punch

Page 51

CARD

PUNCHES

There are two card punches available:

100

card per minute and 300 card per minute.

100

CARD

PER

MINUTE

PUNCH

Cards are punched in decimal (Hollerith), 10-row binary, or 12-row binary modes. Punching may occur simultaneously with central processor computation and operation of

other peripheral units. Cards are punched in parallel, row by row, with all 80 columns of each row being punched

simultaneously. Input and output bins each have a capacity of 800 cards. Extensive error monitoring systems assure high accuracy of data transfer to cards.

Page 52

CARD

PUNCHES

There are two card punches available: 100 card per minute and 300 card per minute.

100

CARD

PER

MINUTE

PUNCH

Cards are punched in decimal (Hollerith), 10-row binary, or 12-row binary modes. Punching may occur simultaneously with central processor computation

and

operation of

other peripheral units.

Cards are punched in parallel, row by row, with all 80 columns of each row being punched

simultaneously. Input and output bins each have a capacity of 800 cards. Extensive error monitoring systems assure high accuracy of data transfer to cards.

Page 53

MEDIA FORMAT

Cards may be punched and read in either Decimal or Binary Modes.

Decimal (Hollerith) Mode

The characters used are the decimal numbers 0 to characters which include punctuation marks, and a space or blank. See the Representation of

GE-235

Characters are automatically converted to Binary Coded Decimal mode of

One

Characters on page

GE-235

word is made up of 3 characters.

of the appendix for the complete list.

OO~O~lOOOOOOO~~OOOO~~~~~oooo

I I

IlI1I1lllllllllllllllllllllI

2222222222222222222222222222

313333333333333333333333333

444444414444444444444444444

5551555555555555555555555555

666666666666666666666666666

777777777711771177777777777

188818888888888888888888888

11031

the

letters of the alphabet, and

12111a151ei71119121nnrnana

Hollerith

Data Card

bits.

special

Octal Octal

Word

Page 54

and

Row Binary Mode

Each hole punched in a column represents 1 bit of the GE-235 word in memory of the Central

Processor Unit.

Row

Binary

GE-235 MEMORY WORD

10-row Binary mode is used primarily for program instruction words.

12-row Binary mode uses the blank space at the top of the card for rows 12 and

manner provides complete compatibility with other computing systems.

and in this

Page 55

CHECKING FEATURES

GE-235 Card Readers

Errors in feeding and reading are detected by the timing and checking features of the respective GE-235 Card Readers. Indicators on the control panels are illuminated at the time of any of the following errors, depending on the model:

CHECKING FEATURE

Card Timing

synchronization to insure that each card column is read at the proper time.

Optical Checks

are checked for proper operation after each card is read.

Card Feed Error

but card fails to read.

Card Read Error

checks fail.

Stacker Full--Input Hopper Full

time a card is read.

End of File

of file.

Card Jam

monitored for possible card jams or changes in

operating speed.

Checked for slippage and proper Yes Yes

Light source and all photocells

Occurs if a card is called for

Occurs if any of the above

Occurs each

A program check is made for end

Transport mechanism

continuously

MODEL

400

CPM

Yes

1000

Yes

CPM

Invalid Characters

Decimal Mode, each column

character validity.

When reading cards in the

checked for

Yes

GE-235 Card Punches

In the GE-235 card punches, error checking circuitry and the punching mechanism monitor for the following errors:

Any column was erroneously left blank. Any column was double punched with numeric characters. If input hopper is empty, card punching stops. If there is a misfeed, card punching stops. If there

a card punch alarm, card punching stops.

Page 56

CARD READER INSTRUCTIONS

Mnemonic Index Operand Time Description Oper. Code Word Address Microsec.

BCN

BCR

RCB

RCD

RCF

KCM

HCR

CARD PUNCH INSTRUCTIONS

Y 12 READ CARDS BINARY. Initiates continuous reading

Y 12 READ CARDS DECIMAL. Initiates continuous reading

Y 12 READ CARD FULL. Reads one card in 12-row

12 BRANCH ON CARD READER NOT READY. Reader

12 BRANCH ON CARD READER READY. Reader is

12 HALT CARD READER. Halts the continuous feeding

is not ready to read cards.

ready to read cards.

of lo-row Binary Mode cards.

in Decimal (Hollerith) Mode.

READ

Decimal Mode and senses for a

Binary Mode.

of cards.

CARD

If found, reading of card

MEED.

Initiates

one

card

7-9

punch in column

changed to 12-row

read

BPN

BPR

WCB

WCD

WCF

12 BRANCH ON CARD PUNCH NOT READY. Punch is

not ready to punch cards.

12 BRANCH ON CARD PUNCH READY. Punch is ready

to punch cards.

Y 12 WRITE CARD BINARY. Card is punched in a 10-row

Binary Mode.

Y 12 WRITE CARD DECIMAL. Card is punched in Decimal

Mode.

Y 12 WRITE CARD FULL. Card is punched in 12-row

Binary Mode.

Page 57

GE-235 HIGH-SPEED PRINTER

GE-235 HIGH SPEED PRINTER

AND

CONTROLLER

An on-line, buffered type printer.

Prints 900 lines per minute. Prints 120 alphanumeric characters per line. Prints in Open Gothic style, spaced 10 characters per inch horizontally, and 6 lines per

inch vertically. Provides automatic editing of format.

VI-

Page 58

Skipping lines (or paper slew) is accomplished at trol by:

Vertical Format tape; or, Vertical Line Count.

inches per second under program con-

Paper width may vary from Paper may be up to Makes up to 5 copies. Total of

FORTRAN and other special characters are available as options.

characters are provided on the printer:

Numbers

letters of the alphabet;

special characters, plus a space or blank:

through

Plus Sign

Period Minus Dollar Sign Asterisk Percent sign Underscore Space or blank Slash

3-1/2

inches.

inches in length per page in fan folded, continuous form type.

Equal Comma

Left bracket

Right bracket

Number sign

"At" sign

Page 59

FORMATS

Data to be printed on the GE-235 High-Speed Printer controller. Three methods for arranging horizontal line format for printing are available; these methods may be used in any combination.

Initial setup on the data to conform exactly to the print line desired;

Rearrangement of the data in memory by programmed insertion or deletion of characters;

2. Automatic editing performed by logic in the printer controller, based on a combination of

3. related data and format words transferred to the printer controller. suppression, deletion of data, insertion of constants and spaces, and dollar amount field editing. Many program steps and the inherent mechanical limitations of wired plugboards

are eliminated because of the powerful automatic editing capability.

Slewing can be accomplished by program or switch control. The programmer can cause slewing of any number of lines up to 63 by specifying the number of lines to be skipped. Paper can be slewed to the top of apage by either program or manually pressing the Slew Top Page switch on the printer control panel. Special vertical formats can be printed by specifying slews to any one of eight chanriels of a perforated tape loop (vertical format tape) contained in the printer mechanism.

Editing of Data with Automatic Format Control

When a line essor memory in a block of words organized the same as the print line data. The format control data consists of:

to be printed under format control, format words are stored in the Central Proc-

transferred from memory to the printer

Editing includes zero

Any printable character control characters

The Printer Controller, in assembling a formatted line, reads in from memory one data word and

one format word. The first format character

acter, it is printed.

Five characters are available for controlling the format of the printed line. These characters,

their octal representations, and their functions are: IGNORE (OCTAL: 35). On Ignore, the next data character is immediately considered. IGNORE/SKIP (OCTAL: 36). On Ignore/Skip, a blank is printed and the next data character is

considered.

DELETE (OCTAL: 37). On Delete, the next data character is ignored, and the next character considered

DELETE/SKIP (OCTAL: 56). On ~elete/~kip, a blank ignored, and the next character considered is the next format character.

the next format character.

a control character, it

considered initially. If it is a printable char-

treated as described below.

printed, the next data character

@En235

VI-

Page 60

ZERO SUPPRESS (OCTAL: next format character is printed if character, spaces will be inserted in the print line until: (a) a non-zero data character is detected, or has been put into effect, the print line data is inspected only for a non-zero data character, and the format control data is inspected only for a period. A spaces in the print line in the same manner as Zero Suppress except that the next data character is not ignored after the

The following is an example of how editing is performed with the printer controller:

(b) a period appears in the format data. It should be noted that once a Zero Suppress

symbol is printed.

On Zero Suppress, the next data character is ignored; and the

57).

is a printable character. After considering this last format

symbol in the format data also initiates

Format Words

Data Words

Line

Depending upon the value contained in the data words, the following amounts (using nines for

example) would be printed on the high speed printer under control of the format control char-

acters above.

When zero suppression on the print line is in effect as a result of a dollar sign or an octal 57 (zero suppress) in the format words,

printed instead) unless the data character preceding the comma in the format word is a non-zero

character.

53 73 35 35 73 35 35 33 35

1 , I

9 9 9 9 9 9 9 9

9 9, 9 9 9

a comma in the format word is not printed (but a blank is

9 9

CHECKING FEATURES

The control panel on the Printer Controller contains indicator lights that show status of the printer and indicate alert conditions. In addition to the Ready and Off-Line indicators, there are the Parity Alert to show when a parity error has occurred and a Paper Alert to show when paper

torn or used up.

Page 61

Printer Control lnstructions

single Print instruction will cause printing of up to 120 characters on one line from a block of 40 words of BCD data in memory, When the PRINT instruction is given, the printer receives information directly from the main memory through the appropriate numbered controller selector

priority channel.

Printer operations are time shared with other processing. The printer can be interrogated for "ready" status (completion of previous PRINT instructions) by the BCS instructions.

The WPL, WFL, SLW, and SLT commands must be preceded by the SEL P instruction, which selects the specific controller selector channel. These instructions cannot be modified.

Printer Control Instructions

Mnemonic Operand Oper. Code

Address

SLT

SLW N

WFL

Index Time

Word Microsec.

36 SLEW PAPER TO TAPE PUNCH. The printer paper

is format tape in channel

36 SLEW PAPER N LINES. The printer paper is spaced

36 WRITE FORMAT LINE. One line of BCD information

spaced until a hole is detected in the vertical

(0 to 63) number of times.

WPL X N is printed under automatic format control.

starting location in memory of a series of format control words. Instruction should always be followed

by a WPL instruction to specify the location (X) of the first word of information to be printed and whether

WPL

the information

N 36 WRITE PRINT LINE. Print one line of BCD

alphanumeric or numeric (N).

mation (I to 120 characters long), starting at memory

location

numeric only (if N

The N indicates that information is

blank, information is alpha-

numeric). Printer paper is automatically spaced one

line.

Printer

Test

and Branch lnstructions

is the

infor-

The following branch instructions test whether a particular printer controller condition

true or false. If the condition tested is true, the computer executes the next sequential instruction. If false, the P stands for printer controller connected to controller selector channel

computer skips the next instruction and executes the second sequential instruction.

thus referred

printer controller P.

VI-

Page 62

Mnemonic O~er. Code

Operand Address

Index Word

Time

Microsec.

Description

BCS

B CS

BCS

BPR

BPN

BOP

BNP

BER

BNE

BAA

BNA

BSA

BNS

24-36

24-36 24-36

24-36

BRANCH ON PRINTER READY. P

is tested for the

ready status.

BRANCH ON PRINTER NOT READY. P

is tested for

the not ready status.

BRANCH OUT OF PAPER. P

is tested to determine if

the printer is out of paper.

BRANCH ON NOT OUT OF PAPER. P

determine if the printer is not out of paper.

BRANCH ON ERROR. P BRANCH ON NO ERROR. P

is tested for parity error.

is tested for a no parity

error condition.

BRANCH ON PRINTER ANY ALERT. P

is tested for

any alert condition.

BRANCH ON PRINTER NO ALERT. P

is tested for no

alert condition.

BRANCH ON PRINTER SLEW ALERT. P

is tested for

a slew alert condition.

BRANCH ON PRINTER NO SLEW ALERT.

on a no slew alert condition.

is tested to

Pis tested

BCS

BOV

BNO

24-36

BRANCH ON PRINTER BUFFER OVERFLOW. P

tested for a buffer overflow condition.

BRANCH ON PRINTER NO BUFFER OVERFLOW. P

tested for no buffer overflow condition.

VI-

Page 63

THE GE-235 MAGNETIC TAPE SUBSYSTEM

CONTROLLER

The

GE-235

magnetic tape subsystem provides a fast method for transmission of extensive files

and other large quantities of information into

WITH

TAPE HANDLERS

andout of the

GE-235

central processor. It consists

of two sections; magnetic tape controllers, and magnetic tape handlers.

TAPE CONTROLLER INDICATOR AND CONTROL PANEL

VII-

Page 64

TAPE CONTROLLER

Each controller can direct the operations of up to 8 tape handlers. Up to 7 controllers (with a total of 56 tape handlers) can be connected to the GE-235

system. Provides logic for selecting the tape handler, and for reading, writing and rewinding the

magnetic tape.

Monitors the flow of data between tape handlers and memory. Initiates and times the starting and stopping of the selected tape handler. Detects the beginning of tape, end-of-tape, and end-of-file record. Ensures reliability for the magnetic tape subsystem through its error-checking circuitry. Two controllers provide simultaneous tape reading and writing in a GE-235 system. Two controllers available: one for low-densitv taoe: and one for high-densitv taoe.

TAPE HANDLERS

Reads from and writes on tape at 75 inches per second. Two character densities are available:

Low Density - 200 characters per inch

1. (1 5,000 characters per second);

High Density

Millions of characters mav be recorded on a single reel of ta~e.

TAPE HANDLER CONTROL AND INDICATOR PANEL

555.5 characters per inch (41,600 characters per second).

Page 65

Tapes may be used as on-line storage during a computer run. Tapes may be used as off-line storage to retain masses of information as a file for subse-

quent related computer runs.

Magnetic tapes may be erased and used repeatedly.

Each cabinet contains two magnetic tape handlers.

Each handler holds two 10-1/2 inch diameter reels, one for feeding, and one for taking up the tape.

Each reel holds up to 2,400 feet of standard 1/2-inch wide magnetic tape. Data is recorded in groups of words called blocks with short blanks in tape separating

blocks.

FORMATS

The Tape Handler is capable of reading either backwards or forward in three formats: a Binary Coded Decimal, Binary, and Special Binary. The programmer selects the format he desires for his specific application. The tape has awidth sufficient for the recording of seven tracks, or channels, of data. Six bits of data are recorded across the tape; a parity bit is recorded in the seventh track.

Parity, odd or even, depends on the format being used. In the binary format, the lateral one-bits will always be odd (known as odd parity). In BCD, the parity is even. This is simply a safeguard to prevent an inadvertent programming error that might arise if reading a tape prepared in the binary format with a Read Decimal type command.

Data is written on tape in groups of words called blocks of fixed or differing length. Inter-block gaps physically separate magnetic tape blocks. This gap is a allowed for starting and stopping tapes between blocks. An end-of-file block separates groups of information, or marks the end of information on a specific tape. It is an erased section 3-3/4 inches long, followed by the code combination 0001111 and the parity character.

Binary

In the BCD format, in tape reading and writing operations, three 6-bit characters contained in the GE-235 word are transferred to and from tape. Thus, in the BCD format, some character bit configurations are altered so as to make the GE-235 system compatible with other tape systems in use.

This alteration is accomplished automatically.

Coded

~ecimal

Format

3/4-inch section of erased tape

Page 66

This is a significant feature because tapeswrittenby the GE-235 may be read by another system's tape unit, and, tapes prepared by other units can be read by the GE-235 system. The alphanumeric "962" exists in memory as three

BCD

characters, as follows:

Binary Coded Decimal Memory Format

When written on magnetic tape these characters appear as three 6-bit characters accompanied by

a parity bit, as:

PARITY (EVEN) ZONE BITS

{

Tape movement

NUMERIC BITS

Binary Format

In the binary format,

characters as in the remaining recording on tape and ignored when reading from tape.

As an illustration of the appearance of a binary word on tape, the decimal number 262243 appears

in memory and on Magnetic Tape as follows:

tape bits of this fourth position contain zeros which are automatically inserted when

the word in memory

BCD

format, and afourth character containing two bits of the data word. The

written as four magnetic tape positions: three tape

VII-

Page 67

It is advantageous to write numeric information in the binary format because the equivalent of

5-1/2

decimal digits in memory may be recorded in only

tape characters. In BCD format, however, 3 BCD characters require 3 tape positions. If memory data consists of mixed formats of both BCD and binary in memory, the binary writing format

employed.

Special Binary Format

The GE-235 system also operates in a special binary format. This records 3 positions on tape

instead of

without altering the bit structure. This feature effectively expands compatibility to cover both the decimal and binary modes of other tape systems. The feature is also useful for GE-235 applications using mixed binary and decimal data.

CHECKING FEATURES

Each GE-235 tape handler has two sets of magnetic heads; read heads, and write heads. During reading operations, only the read heads are used. During write operations, both sets of heads operate, as follows:

Write Heads

Record the characters on tape as the tape passes;

Read Heads

Read each character immediately after

written and checks

its

validity.

Thus, if errors occur, they are detected during the write operation and can be corrected

under program control. Write Permit Ring

Fits into a groove on the back side of the tape reel to permit writing

on the tape. Removing the ring from the reel prevents writing or accidental erasure of data

already on a tape.

The GE-235 magnetic tape system incorporates a wide variety of error detection circuits to ensure accurate data transfers between memory and tape.

These circuits are:

CONTROLLER INPUT-OUTPUT REGISTER EXHAUST OR OVERFLOW. Checks against an excessive wait for memory access on the part of the Tape Controller during multiple

input/

output operations.

LATERAL PARITY.

This parity check associated with each tape character ensures the

accuracy of each character when read from, or written onto, tape.

LONGITUDINAL PARITY. This is a parity check on each of the seven record tracks, or

channels, recorded as a parity character at the end of each tape record. MODULO THREE OR FOUR.

made to determine that the proper number of characters (three in BCD and special binary

When information is read from, or written onto, tape, a check

formats, and four in binary formats) have been read or recorded.

All

information words are parity checked when transferred between memory and the Tape

Controller.

Page 68

Checks are also made against the conditions which cause an ALERT HALT. This is indicated by the

turning on of the ALERT HALT light on the tape controller display

paneL These conditions are:

A parity error on instruction words 2 and 3 as these words are transferred from memory

the tape controller.

Addressing a tape unit that is rewinding. Using an address which has not been assigned to any specific tape unit. Any detectable malfunction of the tape handler. Specifying a tape handler logical address which has been assigned to

two or more units.

Giving a write command to a tape unit which does not have the write permit ring attached

the supply reel. Each tape unit is equipped with a file protection ring which is attached to the tape supply reel. When this ring is not attached, the controller does not permit writing on that

tape. Addressing a tape unit with a read backward instruction (RBB, RBD, or RBS) when the tape

on the unit is initially positioned at its beginning. Parity and other such errors detected while reading or writing tape do not cause an ALERT

HALT but do give an indication on the tape controller display panel that a specific error has

occurred on the last record. These errors can be detected under program control by means of the test-and-branch instructions for the Magnetic Tape System.

MAGNETIC TAPE INSTRUCTIONS

All instructions executed by the Magnetic Tape Controller consist of 3 memory words, the first of which has as its function the selection of the specified controller. This first word has the operation

code SEL P. Similarly all test instructionspertaining to the status of a controller bear the common

operation code BCS.

The general format of these instructions utilize the following conventions: P

ment (channel number); M

length of record (number of words).

Coding under the "Index

memory location (starting); T = tape handler assignment (number)

word" column is recorded in column 20 on the GAP coding sheet.

controller assign-

General Tape Instructions

Mnemonic Oper. Code Address Word

SE L P

Operand Index Time Description

Microsec. 24-36 SELECT the peripheral controller connected to

se-

lector channel number P.

BCS

XXX

P 24-36 BRANCH ON CONTROLLER SELECTOR. The pe-

ripheral controller connected to the controller channel

selector is tested for the conditionindicated by a mne-

monic placed in the operand address field identified by

xxx.

;

Page 69

Tape Movement Instructions

The RTD, RTB, RTS, RBD, RBB, RBS, RWD, WTD, WTB, WTS, and WEF tape movement instructions are preceded by the SEL P instruction to designate the specific controller channel number.

Mnemonic Operand Index Time Description Oper. Code Address Word

Microsec.

BKW

RBB (blank)

RBD (blank)

RBS

(blank)

RTB

(blank)

RTD

(blank)

RTS

(blank)

T 36 BACKSPACE AND POSITION WRITE HEAD. The tape

on tape handler T is backspaced one record and is positioned for reading or writing. This instruction must be preceded by the SEL P instruction and it cannot be modified.

T 36 READ BACKWARD BINARY.

tape T backward in Binary Mode into memory, starting at location A.

T 36 READ BACKWARD DECIMAL. N words are read from

tape T backward in binary coded decimal mode into memory, starting at location

T 36 READ BACKWARD SPECIAL BINARY. N words are

read from tape T backward in special binary mode, starting at location A.

T 36 READ TAPE DECIMAL- FORWARD. N words are read

T 36 READ TAPE SPECIAL BINARY-FORWARD. Nwords

36 READ TAPE BINARY-FORWARD. N words are read

from tape T forward in binary mode into memory,

starting at location

from tape T forward in binary coded decimal mode into

memory, starting at location A.

are read from tape T forward in special binary mode, starting at location A.

words are read from

RWD

WEF

WTB

(blank)

WTD

(blank)

WTS

(blank)

36 TAPE REWIND. 36 WRITE END-OF-FILE. The end-of-file is written on

tape.

T 36 WRITE TAPE BINARY. N words are written in binary

mode on tape T, starting at memory location M.

T 36 WRITE TAPE DECIMAL. N words are written in

binary coded decimal mode on tape T, starting at memory location M.

36 WRITE TAPE SPECIAL BINARY. Nwords are written

in special binary mode on tape T, starting at memory location

Page 70

Test and Branch Instructions

The following instructions test to determine whether a magnetic tape controller condition is true or false. If the condition tested is true, the computer executes the next sequential instruction. If false, the computer executes the second sequential instruction. Under Index Word and the Description, P stands for tape controller P.

Mnemonic Oper. Code

BCS BTR P 24-36 BRANCH ON TAPE CONTROLLER READY. P is

BCS BTN P 24-36 BRANCH ON TAPE CONTROLLER NOT READY. Pis

BCS BEF P 24-36 BRANCH ON END-OF-FILE. P is tested for

BCS BNF P 24-36 BRANCH ON NO END-OF-FILE. P is testedfor end-

BCS BER P 24-36 BRANCH ON ERROR. P is tested for any error

BCS BNE P 24-36 BRANCH ON NO ERROR. P is tested for no error

BCS BIO P 24-36 BRANCH ON

BCS BIC P 24-36 BRANCH ON 1NPUT/OUTPUT BUFFER CORRECT. P

Operand Address

Index Time Word Microsec.

Description

tested for the ready status.

tested for the not-ready status.

file indicator

of-file indicator OFF.

cator ON.

indicator

input/output buffer error indicator ON.

for

is tested for

ON.

INPUT/OUTPUT ERROR. P is tested

input/output buffer error indicator OFF.

end-of-

indi-

BCS BME P 24-36 BRANCH ON MOD 3 OR 4 ERROR. P is tested for

modulo 3 or

BCS BNM P 24-36 BRANCH ON NO MOD 3 OR 4 ERROR. P is tested for

modulo 3 or 4 error indicator OFF.

BCS BET P 24-36 BRANCH ON END-OF-TAPE. P is tested for end-of-

tape indicators This indicator is turnedxby the detection of a photo reflective spot located on tape just

prior to the trailer.

BCS

BNT P 24-36 BRANCH ON NO END-OF-TAPE. Pistestedfor

of - tape indicator

BPE P 24-36 BRANCH ON TAPE PARITY ERROR. P is tested for

parity error indicator

error indicator ON.

ON.

end-

VII-

Page 71

Mnemonic Operand Index Time Description Oper. Code Address Word Microsec.

BCS BPC P

BCS BRW P

24-36

BRANCH ON TAPE PARITY CORRECT. P is tested for tape parity error OFF.

BRANCH ON TAPE REWINDING. P is testedfor tape rewinding condition.

BCS BNR P

24-36

BRANCH ON NO TAPE REWINDING. P is tested for

no tape rewinding condition.

Page 72

GE-235 MASS RANDOM ACCESS DATA STORAGE

With the Mass Random Access Data Storage subsystem (abbreviated MRADS), information can be

accessed directly, eliminating the need to search through unrelated information as

the sequential processing of punched card and magnetic tape files. Such random access

extremely valuable in applications where large volumes of information must be stored and

retrieved with minimum delay. Each GE MRADS subsystem consists of a MRADS Controller, a rotating disc assembly (or MRADS unit), and MRADS controllers can be connected to the Controller Selector of the GE-235 Computer, with each controller capable of controlling up to four MRADS units.

electronics control cabinet. One or two

required in

MASS RANDOM ACCESS DATA STORAGE UNIT

Speed of rotation of disc

Effective transfer rate: Average latency time Maximum latency time

Average positioning time

Maximum positioning time Average access time (including track verification)

1200 rpm 11,850 and 23,700 26 ms 52 ms

199 ms 305 ms 225 ms

words/sec

FORMATS

Each MRADS unit consists of 16 circular magnetic discs. Information 256 circular tracks which make up each side of a disc. Eight 64-word records are recorded in each of the 128 inner tracks. Sixteen 64-word records are recorded in each of the 128 outer

tracks.

vm-

recorded serially in

?>p

Page 73

Each word recorded on a disc

a replica of the word as

appears in the GE-235 core memory.

Both binary and alphanumeric (BCD) configurations are retained without change.

Between 1 and 16 records (64 words each) can be read or written with one command. Each data

disc of the MRADS unit

heads

four heads for each side of the disc.

With the arm in one position, 64 on outer heads

Positioning Read-Write Disc-Support

served by its own positioning arm. Each arm contains eight read-write

possible to transfer a total of 96 records (32 on inner heads,

on both sides of the disc).

Read-write heads numbered 0-7

heads per side

Heads 0,

inner zone of a disc.

heads per disc.

and 3 serve the

MRADS POSITIONING

ARM

Heads

5, 6

and 7 serve the

outer zone of each disc.

AND HEADS

The diagram below shows one surface of a disc; the division of the surface into inner zones; and the zones into 8 and 16 sections with 128 tracks each.

Outer

zone

record

23,100

words/second

per sector

transfer

rate.

Page 74

CHECKING FEATURES

Customer-tailored error checking features include:

Odd parity check on read-write transfers Memory access check Clocking check Address confirmation check Longitudinal parity word check Read-after-write check

Each word

parity error should occur on reading, program reconstruction

check word.

on the disc file contains parity plus an additional check word per record.

possible through the use of the

a word

MRADS INSTRUCTIONS

The PRF, RRF, RRD, WRD, WRF, and RAW instructions must be preceded by the SEL P instruction, which selects the specific controller

utilize the following conventions:

mass random access file number (0 to 3).

K = MRADS file number (0 to 3), test and branch instructions.

memory location of data.

P. The general format of these instructions

N = number (1 through 16) of 64-word records to be transmitted from core storage to disc

storage or vice versa.

controller assignment (plug number).

P Note: File number (F)

Mnemonic Operand Index Time Description

Oper. Code Address Word PRF

(MRADS Address) Contains actual address (in octal) of the selected

WRF

normally recorded in column 20 of the GAP coding sheet.

F 36 Positions arm of MRADS unit to receive or transmit

F 36 READ FROM MRADS UNIT. Reads N records from

F 36 WRITE ON MRADS UNIT. Writes N records on

F 36 READ AFTER WRITE CHECK. A parity check is

Microsec.

a specific record.

MRADS.

MRADS file F, into memory, starting at location

MRADS file

made on all words of a file.

F from memory, starting at location M.

record(s) transferred to the

VIII-3

Page 75

Mnemonic Operand Index Time Description Oper. Code Address Word Microsec.

RRD N F 36 READ MRADS AND RELEASE ARM. Performs the (blank) M same function as RRF except that the positioning arm

released after reading.

WRD N F 36 WRITE MRADS AND RELEASE ARM. Performs

(blank) M same function as WRF except that the positioning arm

is released after writing.

Branch and Test Instructions

The following branch instructions test to determine whether a particular MRADS condition true or false. If the condition tested is true, the computer executes the next sequential instruction. If false, the computer skips the next instruction

Mnemonic Operand Index Time Description

Oper. Code Address Word Microsec.

BCS BRR P 24-36 BRANCH ON MRADS READY. BCS BRN P 24-36 BRANCH ON MRADS NOT READY. BCS FKR P 24-36 BRANCH ON FILE K READY. BCS FKN P 24-36 BRANCH ON FILE K NOT READY. BCS BIO P 24-36 BRANCH ON BCS BIC P 24-36 BRANCH ONINPUT/OUTPUT BUFFER CORRECT.

BCS RPE P 24-36 BCS RPC P 24-36 BCS BER P 24-36 BRANCH ON ERROR. BCS BNE P 24-36 BRANCH ON NO ERROR.

BRANCH ON PARITY ERROR. BRANCH ON PARITY CORRECT.

and executes the second sequential instruction.

INPUT/OUTPUT BUFFER ERROR.

BCS F AE P 24-36 BRANCH ON FILE ERROR. BCS FAC BCS FKE P 24-36 BCS FKC P 24-36 BRANCH ON NO ERROR, FILE K.

24-36 BRANCH ON FILE CORRECT.

BRANCH ON ERROR, FILE K.

VIII- 4

Page 76

GE-235 PERFORATED TAPE EQUIPMENT

Transactions are commonly recorded on perforated tape as by-products of equipment such as

tape-punching typewriters, billing machines, remote terminal links and calculators. The GE-235 system provides a perforated tape reader, punch and spooler in one cabinet; or alternatively, a reader and punch, a reader and spooler, a reader only or a punch only.

Reader

The Reader operates at either of two speeds

Reads Permits optional selection of spool feed at 250 characters per second, or strip feed at 1000

characters per second.

Operates simultaneously with other

Punch Mechanism

The Punch is capable of punching up to 110 characters per second (10 characters per inch). Punches 5 or 6, 7 and 8 channel tapes.

Operates simultaneously with computation and other Input/Output operations.

Mechanism

250 Characters per second (10 characters per inch at 25 inches per second).

1000 Character6 per second (10 characters per inch at 100 inches per second).

6, 7 and 8 channel tapes.

Input/Output operations.

IX-

Page 77

Spooler Mechanism

The Spooler has both takeup and supply spools for perforated tape, which automatically maintain proper tension during reading.

FORMATS

Tapes may be read or punched in standard

hole spacing and with the following tape widths:

channel 11/16 inch

6 or

The sprocket hole of the tape serves as a timing source and must be present with every tape character.

and 6 Channel Code Characters

All code combinations are recognized. The presence of a hole in the sprocket hole channel, only, indicates

4 S 3 2 1 +Channel number

channel

channel 1 inch

tape leader or tape strip character and is sensed

Channel Sprocket

7/8

inch

Punch position

6, 7 or 8 channel codes using Chad sprocket

a valid character.

6 Channel

6548321

Example of 5-channel Perforated Tape

Example of 6-channel Perforated Tape

Page 78

and 8 Channel Code Characters

The characters of the 7 and 8 channel codes are entered into the N register in a like fashion.

The presence of a hole in the 5th Position is

not transferred to the N register as it is used for a

Parity Check.

Sprocket hole punched only, indicating a tape leader or tape strip character, is ignored.

All holes punched in Positions

thru 7 are sensed as a Delete code and are not transferred to

the N register.

The transference of all positions of the

Channel code can be accomplished by a special optional

feature.

Channel

7654S321

Example of 7 -channel Perforated Tape

Example of

a-channel Perforated Tape

CHECKING FEATURES

Parity checking is performed when reading or punching 7 or 8 channel tape. Position 5 represents

the parity bit.

Page 79

PERFORATED TAPE INSTRUCTIONS

Mnemonic Oper Code

BNN

BNR

HPT

PON

RON

RPT

WPT

Operand Time Address Microsec.

BRANCH ON N REGISTER NOT READY.

the N register is not ready for input or output.

BRANCH ON N REGISTER READY. Branch if the N register is ready for input or output.

HALT PERFORATED TAPE. Halts the reading of the tape.

PUNCH ON. Punch is logically connected to the N register and other N register peripherals are disabled.

READER ON. Reader is logically connected to the N register and other N register peripherals are

disabled. READ PERFORATED TAPE. Initiates the continuous

reading of the tape.

WRITE PERFORATED TAPE. Initiates the continuous writing (punching) of the tape.

Branch if

Page 80

GE-235 12-POCKET DOCUMENT HANDLER

Reads and sorts up to 1200 magnetically encoded source documents per minute.

On-line, sends information to the central processor for further processing while sorting

documents; off-line, acts as a sorter only.

Sorts documents in a wide variety of formats and ranging in size from 2-1/2 by 5-1/4

inches to 3-3/4 by 2500 documents can be placed in the document handler feed hopper at one time; 1500

documents in Handles defaced and mutilated documents also.

Reads E13b font characters which are as easily recognizable to the human eye as to the

document handler.

each pocket.

FORMATS

BCD

six

The BCD character memory is a cue character,

the BCD character in memory

only.

bits occupy the six least significant bit positions of each memory location. If the

a numeric digit, all remaining bit positions are zeros. If the BCD character in

8-3/4

inches.

one bit is placed in the sign and the most significant bit positions.

an invalid character, a one bit

placed in the sign position

Page 81

MAGNETIC INK CHARACTER RECOGNITION SYSTEM

EXAMPLE OF MICR CHARACTERS IN BANKING APPLICATION

Decimal Digits

Cue characters are normally used to separate fields

of decimal digits. cation the cue characters separate fields of decimal digits in the following manner:

For example, in a banking appli-

The cue characters in this illustration separate such fieldsas the Federal Reserve routing district, a numberassignedto the bank by the American Bankers sociation, the account, and the dollar amount.

As-

Page 82

MAGNETIC

INK

CHARACTER RECOGNITION SYSTEM

EXAMPLE OF MICR CHARACTERS IN BANKING APPLICATION

Cue characters are normally used to separate fields of decimal digits. For example, in a banking application the cue characters separate fields of decimal

digits in the following manner:

11'0

22a011'

2 2

L1110

LL1:

3111 3111

The cue characters fields as the Federal Reserve routing district, a berassignedtothe bank by the American Bankers sociation, the account, and the dollar amount.

311' La02 ?qlllOOOOOO

Decimal Dwts

Cue Characters

this illustration separate such

L00011'

num-

As-

VIEW OF CONTROL PANEL OF GE-235 DOCUMENT HANDLER

Page 83

CHECKING FEATURES

Total Cue Check - Counts the total number of all cue symbols on the document to insure that all

fields are read.

Long Character misinterpreted by the document handler due to the presence of magnetic ink particles just before or after the printed character.

Missing Digit Detection are legitimate and accepts or rejects the document accordingly.

Transposition

Jam Sensing equipment.

and Multiple Read

Decides whether or not spaces encountered in the document's coding

Check Digit

Checks for stoppage of documents to prevent damage to both the documents and

Verifies the correctness of predetermined fields on the document.

Checks for unusually "longn characters or multiple reads

DOCUMENT HANDLER INSTRUCTIONS

The following instructions must be preceded by an SEL P instruction. In the instructions shown

below or reject, P

Mnemonic Oper. Code

memory location, N = document handler unit,

controller assignment number.

Operand Address

Index Time

Word Microsec.

N 36 READ SINGLE DOCUMENT into memory. The first

character goes into location multiple of 64).

pocket number 0 through

Description

must be a

special,

N 36 POCKET SELECTED to receive document.

N 36 HALT continuous feeding.

READ DOCUMENT into feeding next document. for computation.)

63 and CONTINUE

(50

milliseconds are available

X-3

Page 84

The following branch instructions test whether or not a particular document handler condition is true or false. If the conditiontestedis true, the computer executes the next sequential instruction. If false, the second sequential instruction is executed. The letter K in the mnemonic specifies document handler

or 2.

Mnemonic Operand Oper. Code

BCS SKR BCS BCS PDK P 24-36

BCS NPK P 24-36 BRANCH TOO LATE FOR POCKET DECISION,

BCS FSK P 24-36 BRANCH ON SORTER K READY AND FEED

BCS NFK P 24-36 BRANCH ON SORTER K NOT READY OR FEED

BC S VC K BCS IC K P 24-36 BCS

BCS SKE P 24-36 BRANCH ON SORTER K ERROR (error indicator is

Address Word Microsec. Description

SKN

Index Time

P 24-36 BRANCH ON SORTER K READY. P 24-36 BRANCH ON SORTER K NOT READY.

BRANCH IN TIME FOR POCKET DECISION, SORTER K

SORTER K.

COM-

MAND GIVEN.

COMMAND NOT GIVEN.

P 24-36 BRANCH ON VALID CHARACTER, SORTER K.

BRANCH ON INVALID CHARACTER, SORTER K.

P 24-36 BRANCH ON SORTER K CORRECT (error indicator

on).

BCS

DQK 24-36 BRANCH ON DOCUMENT TCD CORRECT, SORTER

NQK 24-36 BRANCH ON DOCUMENT TCD NOT CORRECT,

SORTER K.

X-4

Page 85

GE-235

DATANET-15 DATA TRANSMISSION SUBSYSTEM

The General Electric DATANET-15 enables the GE-235 Computer to automatically receive and

process information originated at locations remote from the computer center and also

matically send information (replies, results, etc.) to the remote locations.

to auto-

DATANET-15

The DATANET-15 serves as the primary control and connecting link between the GE-235 Computer and the transmission line and remote data-originating and receiving equipment.

Remote stations may be connected to the DATANET-1 5 through a variety of transmission facilities, including leased or public telephone and telegraph lines and privately owned, two-wire cables. The DATANET-15 can also be used to connect the GE-235 Computer to public message networks, such as AT&T's Teletypewriter Exchange Service (TWX) or Western Union's TELEX service.

DATA

TRANSMISSION CONTROLLER

XI-

Page 86

EQUIPMENT

The basic DATANET-15 controller includes the equipment arrangements for co~ecting two channels.

The specific code level and speed of the DATANET-15 is determined by the type of remote station equipment used.

Code level may be any one of the following:

Five channel Six channel Seven channel with or without odd parity

Eight channel with or without odd parity

Speed of transmission is controlled by timing units.

Any single speed between 60 and 2400 bits

per second may be used. Standard timing units are as follows:

75 bits per second

100 bits per second

1050 bits per second

Options available include:

4 additional channqls

13 additional chan els

Paper Tape Statqn Adapter

a a

Interface Adapters

for 3-to-6 channel systems.

for 3-to-15 channel systems.

for connection to

one for each teletype grade channel facility.

FUNCTIONS

The DATANET-15 will perform all of the following functions:

Receive

Recognize request for access by a remote station.

Signal remote station that access is granted (under program control).

Provide a remote station "Request for Access" queing arrangement.

Mode

Paper Tape Reader and Punch.

Initiate data flow (on instructions from computer program) from the remote station to the GE-235 Computer memory,

Convert serial data to parallel form. Strip off special bits necessary for character transmission to GE-235 Computer.

XI-

Page 87

Supply own timing source for clocking bits of a received character.

Receive inputs from interface adapters or digital subsets. Assemble data in GE-235 Computer memory for processing. Notify main computer program that data is ready for processing. Generate word parity for character to be transferred to GE-235.

Signal GE-235 at the end of each finite character message block. (This

accomplished by

recognition of a pre-selected end of message character code.)

Transmit

Mode

Recognize request from main computer program to transmit data to a designated remote station.

Select the addressed remote station. Initiate data flow from the GE-235 memory to remote station. Accept parallel-bit code from GE-235 central processor and convert to serial form. Add special bits necessary for character transmission. Check parity on characters received from GE-235 Computer. Generate parity bit for seven and eight-channel code characters to be transmitted. Supply own timing source for transferring character bits serially into the transmission

system.

Supply transmissionpCltput to interface adapters or digital subsets.

CHECKING FEATURES

Odd parity check when using seven and eight-channel codes. Automatic "Operator Error Code" detection.

Two-to-thirty second delay alarm prevents noise pulses from giving false indication of remote station's request to transmit. Alarm will normally be set for a 15 second delay.

Automatic alert halt upon detection of command word parity error.

XI-3

Page 88

DATANET-15 INSTRUCTIONS

Mnemonic Operand Index Time Oper. Code

Address Word Microsec. Description

SE L P 24-36 SELECT. Selects the DATANET-15 connected to Con-

troller Selector plug P.

RRM C 36 RECEIVE REMOTE MESSAGE. Sets receive mode

(blank) Y 36 without affecting any other controller functions.

In-

cludes starting address in memory (Y) and maximum

number of characters to be received (C).

WRT C+S 36 WRITE REMOTE MESSAGE. Positions DATANET-15

(blank) Y 36 to correct station address, switches it to the transmit

mode, and initiates transmission of data from GE-235 Computer memory.

SCN (blank)

RRT C (blank) Y

0 0

12 12

Places DATANET-15 in receive mode and scanning.

36 READ REMOTE TAPE. Turns DATANET-15 to not 36 ready status, turns Paper Tape Reader on, and initiates

the reading of paper tape. Includes the starting memory address and the maximum message length.

WRT C

(blank) Y

36 WRITE REMOTE TAPE. Turns the DATANET-15 36 to the transmit mode, positions the scan counter on

00,

station number

gives the starting memory address

from which data is to be removed, and controls the

-----

maximum message length.

The following illustration shows the DATANET-15 command words as they appear in the GE-235 Com?uter memory. The Select command words which precede the DATANET-15 commands are

not shown.

12 13 14 15 16 17 18 19

S S S S

0 0 0

Bits SetTransmitMode Set Receive Mode Start Scanning

Command Word

0 1 2 3 4 5 6 7 8 9 10

10 0 0 0 C C C C C C C C C C C OlOOOCCCCCCC C CC C 0 0 0 0 0110000000 0

Read Paper Tape

PunchPaperTape

Bits

01010CCCCCCC C CC C 10 0 0 1 CCCCC C C C C C C

0 0 0

0 0 0 0

Command Word 2

0 1 2 3 4 5 6 7 8 9 10 000000YYYY Y Y Y Y Y Y Y Y Y

S = Station Address;

Beginning memory location for

input or output message.

0-15

12 13 14 15 16 17 18 19

Character Count;

(must be

2's

1-2048

complement)

XI-4

Page 89

Test and Branch Instructions

Mnemonic Operand

BCS XXX BCS RC

BCS RCN BCS RTD BCS RNT BCS RAH BCS RN

BCS REC

BCS RNC

BCS RDP BCS RND BCS RCP BC S RNP

Index Time Word Microsec.

Description P 24-36 Branch on controller selector. P 24-36 Branch if DATANET-15 is ready. P 24-36 Branch if DATANET-15 is not ready.

24-36 Branch if N-Second delay occurred. P 24-36 Branch if N-Second delay did not occur. P 24-36 Branch on alert halt. P 24-36 Branch on no alert halt. P 24-36 Branch if error code is detected. P 24-36 Branch if error code is not detected. P 24-36 Branch on data parity error. P 24-36 Branch on no data parity error, P 24-36 Branch on command word parity error.

24-36 Branch on no command word parity error.

BCS RSP

BCS RSN

BCS RAE BCS

RNE BCS REM BC S RNM BC S REX BCS RNX BCS RPH BC S RP T

P 24-36 Branch

,~~--

access.

P 24-36 Branch if scanner

scanner is positioned on station requesting

not positioned on station re-

questing access.

24-36 Branch on any error. P 24-36 Branch on no error. P 24-36 Branch if end of message code

detected. P 24-36 Branch if no end of message code is detected. P 24-36 Branch on end of transmission. P 24-36 Branch on no end of transmission. P 24-36 Branch on paper tape halt. P 24-36 Branch on no paper tape halt.

XI- 5

Page 90

Mnemonic Oper. Code Address

Operand Index Time

Word Microsec. Description

BCS BCS RNO P 24-36 Branch on no character counter overflow. BCS RAI P 24-36 Branch if DATANET-15 attempted Automatic Program

BCS RNI P 24-36 Branch if DATANET-15

ROV P 24-36 Branch on character counter overflow.

Interrupt.

didnot attempt Automatic Pro-

gram Interrupt.

Page 91

TOM

Enable the GE-235 to receive information from specialperipherals such as sensors, readers, analog devices, etc.

Operate, through the controller selector, simultaneously with the central processor and standard GE-235 peripherals.

Are custom-tailored to fit special problems of customer's present needs and equipment. Perform parity checks on all input and output data transfers. Require only minor circuit changes to adapt the system to new applications in the interface

area.

'PU

MEN

DIGITAL INPUT/OUTPUT ADAPTOR (DI/OA)

Provides a single, one-directional channel for transferring data between the GE-235 and an external device.

Transfers data at the rate of 50,000 twenty-bit words per second. Consists of swing-out modules located within the central processor. Requires no programming except for the use of the SEL P instruction.

DIGITAL INPUT/OUTPUT BUFFER (DI/OB)

Moves data bFemtwo external buffers and the GE-235. Transfers data at the rate of 18,000 twenty-bit words per second. Housed in a cabinet 40-1/8 inches wide, 32-1/8 inches deep, 76 inches high; has its own power

supply.

DIGITAL INPUT/OUTPUT CONTROLLER (DI/OC)

Moves data between the GE-235 and up to 64 external devices.

Transfers data at the rate of 18,000 twenty-bit words per second. Housed in a cabinet 40-1/8 inches wide, 32-1/8inches deep, 76 inches high; has its own power

supply.

XII-

Page 92

DIGITAL INPUT/OUTPUT DISTRIBUTOR (DI/OD)

Moves data between four external buffers and the GE-235.

Transfers data at the rate of 50,000 twenty-bit words per second. Housed in a cabin

40-1,'8 inches wide, 32-1/8 inches deep, 76 inches high; has its own power

supply.

For programming the DI/OB and DI/OC, the SEL P instruction is always followed by twc command words. The first command word gives, in octal, the memory address of information coming from a peripheral. The second command word gives, in octal, the memory address of the information goin? a peripheral.

For Multiple Cycle:

SEL OCT 1 OXXXXX OCT

04XXXXX

For Multiple Cycle with

Automatic Priority Interrupt:

SEL

OCT 1 1 XXXXX OCT 05XXXXX

For Single Cycle:

SEL P OCT

l2XXXXX

OCT 06XXXXX

For Sin~le Cycle with

Automatic

SEL OCT

F'riority Interrupt:

3XXXXX

OCT 07XXXXX

The following branch conditions may be tested for through the use of the OCT command.

Octal 2514P20 DI/OB (DI/OC) Not Ready

516P20 DI/OB (DI/OC) Ready

2514P21 2516P21

No parity error on DI/OB (DI/OC)

Parity error on DI/OB (DI/OC)

2514P22 CWI Not Busy

2516P22 CW1 Busy

2514P23 CW2 Not Busy 2516P23 CW2 Busy

Page 93

Page 94

TYPE REFERENCE

TYPE

REFERENCE

Address Modification Arithmetic ARITH

Automatic Priority

Interrupt API Auxiliary Arithmetic Unit Card Reader or Punch CARD Console CONSL Controller Selector SE L DATANET-15 DN15 Data Transfer TRANS Document Handler DOC

ADMOD

AAU

High Speed Printer PRINT Internal Test and Branch ITAB Magnetic Tape MAG

Mass Random Access Data

Storage MRADS

Optional Instruction OPTNL Perforated Tape Punch or Reader PAPTP

Pseudo Instruction PSUDO

Real Time Clock CLOCK

Shift SHIFT Typewriter TYPE

Page 95

MNEMONIC

OCTAL CODE USE

DESCRIPTION

TIME

Microsec.

BAR BNO BAR BNU

BAR BNZ BAR BON BAR BOO

BAR

B*'BPL

BAR BUF BAR BUN BAR BUO BAR BZE BCN BCR BCS BAA BCS BEF BCS BER

(MU) (AAU)

(MU) (MU) (AAU)

(M U) (AAU) (MU) (MU) (AAU) (MU)

(CARD)

(CARD) (PRINT)

(MAG) (MAG)

Branch on AAU No Overflow Branch on AAU No Underflow

Branch on AAU Non- Zero Branch on Overflow Hold not On Branch on Overflow Hold On

Branch on AAU Overflow Branch on AAU Plus

Branch on AAU Underflow Branch on Underflow Hold not Branch on Underflow Hold On Branch on AAU Zero

Branch on Card Reader Not Ready Branch on Card Reader Ready Branch on Printer any Alert Branch on Mag Branch on Error

Tap'? End-of- File

BCS BET

BCS BIC

BCS BIO

BCS BME BCS BNA BCS BNE

BCS BNF

(MR ADS (PRINT)

(MAG)

(MRADS)

(MAG) (MRADS)

(MAG) (PRINT)

(MAG) (MRADS)

(PRINT)

(MAG)

)

Branch on Mag Branch on

Branch on 1/0 Buffer Error

Branch on Mod Branch on Printer No Alert Branch on No Error

Branch on Mag Tape No End-of-File

Tap-- End-of-Tape

1/0

Buffer Correct

or 4 Error

Page 96

MNEMONIC

BCS BNM

OCTAL CODE

(MAG)

Branch on No Mod 3 or 4 Error

BCS BNO

BCS BNP

BNR

BCB'BNS

BCS BNT BCS BOP BCS BOV BCS BPC BCS BPE

BCS BPN

BCS BPR BCS BRN BCS BRR

BCS BRW BCS BSA BCS BTN

(PRINT) Branch on Printer No Buffer Overflow (PRINT) Branch on Printer Not Out of Paper

(MAG) Branch on Mag Tape Not Rewinding (PRINT) (MAG) Branch on Mag Tape No End-of-Tape (PRINT) Branch on Printer Out of Paper (PRINT) Branch on Printer Buffer Overflow

(MAG) Branch on Mag Tape Parity Correct (MAG) Branch on Mag Tape Parity Error (PRINT) Branch on Printer Not Ready (PRINT) Branch on Printer Ready (MRADS) (MRADS) Branch on MRADS Ready

(MAG)

(PRINT) Branch on Printer Slew Alert

Branch on Printer No Slew Alert

Branch on MRADS Not Ready

Branch on Mag Tape Rewinding

24 24

24 24 24

24 24

24 24 24

BCS BTR

BCS

DQ1

BCS

DQ2

BCS FAC BCS FAE

FS1

BCS

BCS FS2

BCS FOC BCS FOE

(MRADS) (MRADS)

(DOC)

(MRADS) (MRADS)

Branch on Sorter 1 Document TCD Correct

Branch on Sorter Correct

Branch on File Correct

Branch on File Error

Branch on Sorter Command Given

Branch on Sorter

Command Given Branch on File Branch on File

Document TCD

Ready and Feed

0 Correct

0 Error

Page 97

MNEMONIC USE

CODE

OCTAL

DESCRIPTION

TIME

Microsec.

BCS FON BCS FOR

BCS

F1C

BCS FIE

BC~

FIN

BCS FIR

F2C

BCS BCS F2E

BCS

F2N

BCS F2R BCS F3C BCS F3E BCS F3N

BCS

IC1

BCS IC2

(MRADS) (MRADS)

(MRADS) (MRADS) (MR ADS (MRADS) (MRADS)

(MRADS)

(MRADS) (MRADS)

(MRADS) (MRADS) (MRADS) (MR ADS

(DOC)

Branch on File 0 Not Ready Branch on File

Branch on File 1 Correct Branch on File 1 Error Branch on File 1 Not Ready

)

Branch on File 1 Ready Branch on File 2 Correct

Branch on File Branch on File 2 Not Ready Branch on File 2 Ready Branch on File Branch on File Branch on File

Branch on File

)

Branch on Sorter Character

Branch on Sorter 2 Invalid Character

0 Ready

Error

3 Correct

Error

Not Ready

Ready

Invalid

NF1

BCS

BCS NF2

BCS NP1 BCS NP2 BCS NQ1

NQ2

BCS

PD1

BCS

PD2

BCS

(DOC)

Branch on Sorter 1 Not Ready or Feed Command Not Given

Branch on Sorter 2 Not Ready or Feed Command Not Given

Branch too late for Pocket, Sorter Branch too late for Pocket, Sorter 2

Branch on Sorter 1 Document TCD not Correct

Branch on Sorter 2

Doc7~ment TCD

not Correct Branch in Time for Sorter 1 Pocket

Decision Branch in Time for Sorter 2 Pocket

Decision

Page 98

MNEMONIC

OCTAL

CODE

USE

DESCRIPTION

TIME

Microsec.

BCS RAE

BCS RAH

BCS RAI

BCS

RCY

BCS RCR BCS RDP

BCS REC BCS REM

BCS REX

BCS RNA

BCS RNC

BCS RND

Branch on DATANET-15 Any Error Branch on DATANET-15 Alert Halt Branch on DATANET-15 API

Attempted Branch on DATANET-15 Not Ready Branch on DATANET-15 Command

Word Parity Branch on Branch on DATANET-15 Data Parity

Error Branch on DATANET-15 Error Code Branch on DATANET-15 End of

Message Branch on DATANET-15 End of

Transmission Branch on DATANET-15 No Alert

Halt Branch on DATANET-15 No Error

Code Branch on DATANET-15 No Data

Parity Error

DATANET- 15 Ready

BCS RNE BCS RNI

BCS RNM

BCS RNO

BCS RNP

BCS RNT

BCS RNX

BCS ROV

Branch on DATANET-15 No Error Branch on DATANET-15 API Not

Attempted Branch on DATANET-15 No End of

Message Branch on DATANET-15 No Counter

Overflow Branch on DATANET-15 No Command

Word Parity Branch on DATANET-15 No 15 Second

Delay Occured Branch on DATANET-15 No End of

Transmission Branch on DATANET-15 Counter

Overflow

Page 99

MNEMONIC

OCTAL CODE

USE

DESCRIPTION Microsec.

TIME

BCS RPC BCS RPE BCS RPH

BC& RPT

BCS RSN

BCS RSP

BCS RTD

SIC

BCS BCS S1E BCS SIN BCS SIR BCS S2C

BCS S2E BCS S2N

(MRADS) (MRADS) (DN15)

Branch on MRADS Parity Correct Branch on MRADS Parity Error

Branch on DATANET-15 Paper Tape Halted

Branch on DATANET-15 Paper Tape Not Halted

Branch on DATANET-15 Scanner Not 24 Positioned

Branch on DATANET-15 Scanner 24 Positioned

Branch on DATANET-15 Second Delay Occured

Branch on Sorter Branch on Sorter 1 Error Branch on Sorter Branch on Sorter

Branch on Sorter 2 Correct Branch on Sorter 2 Error Branch on Sorter 2 Not Ready

Correct 24 - 36

Not Ready

Ready

24 24 24

24 24 24 24

24 24

36 36 36

36 36 36 36 36

S2R

BCS BCS VC1 BCS VC2 BCS O+F BCS

O+T

BCS ItF BCS 1tT

BCS 2tF

2tT

BCS BCS 3tF

3tT

BCS BCS 4tF BCS 4tT

Branch on Sorter 2 Ready

Branch on Sorter Branch on Sorter 2 Valid Character Controller Condition Controller Condition Controller Condition

Controller Condition Controller Condition 2 False

Controller Condition

Controller Condition 3 False

Controller Condition Controller Condition 4 False Controller Condition 4 True

Valid Character

0 False 0 True 24 - 36

False

True

3 True

24 24 24

24 24 24 24 24 24

24 24

36 36 36

36 36 36 36

36 36

Page 100

MNEMONIC

OCTAL CODE

USE

DESCRIPTION

TIME

Microsec.

BCS 5+F BCS 5+T BCS 6+F

@EL) @EL)

BCS 6+T

7+F

BCS BCS 7+T

BCS 8+F

BCS 8+T BCS 9+F

BCS 9+T BEV

BKW BMI BNN

@EL) (SE L)

(SEL)

(SE L) @EL)

(ITAB) (MAG) (ITAB) (CONSL)

Controller Condition 5 False Controller Condition 5 True Controller Condition 6 False Controller Condition 6 True Controller Condition 7 False

Controller Condition 7 True Controller Condition 8 False

Controller Condition 8 True Controller Condition 9 False Controller Condition 9 True

Branch On Even Backspace and Position Write Head Branch on Minus Branch on N Register Not Ready

(PAPTP) BNO BNR

(ITAB)

(CONSL)

Branch On No Overflow

Branch On

BNZ BOD

BOV BPC BPE BPL BPN BPR BRU BSS

(PAPTP)

(ITAB) (ITAB)

(ITAB) (ITAB) (ITAB) (ITAB) (CARD) (CARD) (ITAB)

(PSUDO)

Branch On Non-Zero Branch On Odd

Branch On Overflow

Branch on Parity Correct

Branch on Parity Error Branch On Plus Branch On Card Punch Not Ready Branch On Card Punch Ready Branch Unconditionally Block Started by Symbol