Ibm 2030 Manual

Field Engineering

Manual

of

Instruction

Processing Unit

System/36D Model

3D

PREFACE

This

manual

contains

information

about

the

IBM

2030

Processing

unit.

A

companion

manual

on

input/output

control

should

be

obtained

for

information

pertaining

to

the

attachment

of

I/O

devices

to

the

IBM

System/360

Model

30.

The

companion

manual

is

the

IBM

2030

I/O

Control

Field

Engineering

Manual

of

Instruction,

Form

225-3362.

Minor

Revision,

August

1965

This

edition,

225-3360,

is

identical

in

content

to

the

Z25-336Q.

The

IBM

Confidential

classification

has

been

removed

and

the

form

number

has

been

changed

to

allow

free

access

to

the

manual.

Address comments concerning

the

content of this publication to

mM

Product Publications, Endicott, New York 13764.

©

1965

by

International

Business

Machines

Corporation

CONTENTS

SECTION

1.

COMPREHENSIVE INTRODUCTION

System

Configurations

...••.••.......

System

Concepts

............•......••

Numbering

Systems

....•.•.....•••••..

Ari

thmetic

Principles

........•..•...

Data

Flow

..•...•••....•....•..•.•.•.

Basic

Programming

.......•.••..•.•.•.

Instruction

Sequencing

and

Branching

The

System/360

and

Interrupts

.••.••.

Storage

Protection

•••.•.•...•...•.•.

SECTION

2.

FUNCTIONAL UNITS

...•.•...

SLT

Circuitry

.•....••....•••....•...

Remembering

Devices

.••.••.•..••.••..

Combined

Components

..•••...••...•••.

Central

Processing

Unit

(CPU)

Clock

•.

Arithmetic

Logical

Unit

(ALU)

.....•.

Registers

......••..•.••....••......•

Core

Storage

..••...•..•...•••.......

Memory

Control

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

.

SECTION

3.

THEORY

OF

OPERATION

•...•.

Concepts

of

Capacitor

Read

Only

Storage

..•••.....•..•..•

Micro

Programming

Introduction

....••

Micro

Program

Examples

•..........•••

Control

Field

Mnemonics

...•...••..•.

Parity

Bits

•..•....•....•••...•..••.

ROS

Addressing

...............••...•.

ROAR

Con

tro

1s

......•...........•..••

ROS

·T

imings

•.....•.••..••.•.•...•...

Physical

Description

....•.•...•...•.

Machine

Check

Handling

....•.•.•.....

Forced

Micro

Program

Entries

..•.....

Overall

Timing

Relationships

.....•..

SECTION

4.

POWER

SUPPLy

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

.

SECTION

5.

SPECIAL FEATURES

..•.•..•.

1401,

1440,

and

1460

Compatibility

Features

...•......•.

Interval

Timer

..........•....•....•.

APPENDIX 1

APPENDIX 2

••••••••••••••••••••••••••

Upper

Indicator

Panel

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

.

Lower

Indicator

Panel

..........••...

Operator

Panel

.....................•

Control

Keys

•......•••......••...•..

Mode

Control

Panel

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

.

Meter

Panel

.••..•.•...••••..........

1-1

1-5

1-21

1-25

1-31

1-35

1-43

1-49

1-69

2-1

2-3

2-4

2-5

2-10

2-22

2-23

2-62

3-1

3-15

3-23

3-44

3-51

3-53

3-60

3-63

3-66

3-7l

3-76

3-80

4-1

5-1

5-48

6-2

6-5

6-7

6-9

6-11

6-15

6-17

ANSWERS

TO

REVIEW QUESTIONS

••......•

6-19

INDEX

•...•.....•...•...•.......••.••

7-1

SYSTEM

CONFIGURATIONS

•

Single

system

concept.

•

Different

models

provide

a

variety

of

processing

speeds

and

storage

sizes.

•

Broad

range

of

input/output

devices.

•

Uses

an

8-bit

coding

structure

to

represent

data.

•

program

compatibility

throughout

system

models.

Behind

the

decision

to

design

the

IBM

System/360

- a

single

system

which

encompasses

all

areas

of

data

processing

lies

the

awareness

that

apparently

unre-

Figure

1-1.

IBM

System/360,

Model

30

SECTION

1.

COMPREHENSIVE

INTRODUCTION

lated

applications

have

more

similari-

ties

than

differences.

For

example,

because

of

teleprocessing

and

other

factors,

scientific

applications

require

high-speed

input/output

similar

to

that

required

by

commercial

applications.

In

addition

to

this

versatility,

the

System/360,

because

of

its

modularity,

adaptability

and

compatibility,

can

handle

the

many

kinds

of

growth

that

normally

occur

in

computer

installations.

Modularity

in

the

System/360

is

achieved

through

the

availability

of

seven

models.

A

typical

Model 30

system

is

shown

in

Figure

1-1.

In

addition

to

a

choice

of

processing

speed

through'

model

selection,

each

model

offers

a

choice

of

storage

capacities.

1-1

AS

problems

and

workloads

grow

or

change,

the

System/360

can

easily

be

expanded

o.r

changed

to

handle

additional

or

different

operations.

Storage

can

be

added

and

input/output

and

processing

speeds

increased

--

in

small

increments,

as

needed.

This

adaptability

makes

provision

for

the

inclusion,

either

initially

or

subsequently,

of

a

broad

range

of

input/output

devices.

Versatile

performance

characteristics

permit

handling

of

data

in

virtually

any

desired

character

representation.

Instead

of

the

usual

six-bit

character,

the

System/360

employs

a

new

eight-bit

coding

structure

to

represent

data.

An

8-bit

unit

of

data

is

called

a

byte.

This

8-bit

coding

structure

allows

256

possible

combinations

for

letters,

digits

and

symbols,

providing

greate.r

versatility

in

both

binary

and

decimal

operations.

It

also

means

that

System/360

can

accept

a

character

code

of

fewer

bits,

such

as

a

telegraph

code.

More

important

than

the

modularity

and

adaptability

of

the

System/360

is

its

compatibility.

A

program

written

for

one

configuration

will

run

on

any

other,

if

there

is

enough

memory

capaci-

ty

and

input/output

equipment,

and

if

the

program

is

not

geared

to

the

operat-

ing

speed

of

any

particular

unit.

Sub-

ject

to

these

constraints,

a

program

written

for

a

smaller

System/360

will

run

without

modification

on

a

larger

one.

While

this

-upward-

compatibility

is

certainly

an

advantage,

-downward-

compatibility

can

be

even

more

valuable;

for

example,

a

small

user

can

utilize

1-2

programs

written

for

larger

systems.

This

places

a

total

library

of

programs

at

all

users'

disposal.

Compatibility

further

allows

a

System/360to

be

tailored

to

fit

either

centralization

or

decentralization.

That

is,

a

company'sinstallation

can

be

either

a

large

central

processor

or

a

number

of

smaller

processors.

Shifts

between

the

extremes

are

possible

within

the

same

system.

Traditionally

there

have

been

constraints

on

computer

versatility,

so

that

one

processor

has

lent

itself

to

scientific

and

engineering

application,

another

to

commercial

data

processing

applications,

another

to

process

control,

and

still

another

to

communi-

cations.

The

System/360

provides

a

versatile

set

of

instructions

that

permit

operat-

ing

quickly

and

efficiently

regardless

of

the

system

application.

MODELS

AND

SPEEDS

-

Different

processor

models

provide

a

variety

of

processing

speeds

and

core

storage

siz.es.

-The

2030

is

the

processor

for

the

System/360,

Model 30

(Figure

1-2).

-

The

2030

is

available

in

four

core

storage

sizes,

represented

as

System/360,

Models C30, D30,

E30,

and

F30.

Figure

1-2.

IBM

2030

Processing

Unit

To

fit

the

widely

varied

cost

and

volume

needs

of

all

computer

users,

the

IBM

System/360

is

available

in

many

differ-

ent

models.

For

instance,

to

fit

the

needs

of

the

user

who

needs

a

minimal

number

of

answers

per

month,

a Model 30

is

available

at

a

minimal

cost.

For

the

user

who

needs

a

greater

number

of

answers

per

month,

a Model 70

is

availa-

ble

that

will

give

approximately

50

times

as

many

answers

per

month

as

a

Model

30.

The

answers

will

be

the

sallie;

only

the

number

of

answers

in

a

given

period

of

time

will

be

different.

There

are

two

basic

differences

between

models:

core

storage

capacity

and

inter~

nal

processing

speeds.

Figure

1-3

shows

the

core

storage

capacities

for

the

different

System/360

models.

Internal

processing

speed

is

largely

dependent

on

the

speed

of

the

core

storage

unit

and

the

amount

of

data

involved

on

each

core

storage

access.

Core

storage

speed

in

the

System/360

varies

from

2

micro-

seconds

per

single

byte

access

on

the

Model

30's

to

1

microsecond

per

eight-

byte

access

on

tpe

Model

70's

(Figure

1-4)

•

1-3

Storage Capacity

in

Bytes

524,288

262,144

131,072

65,536

32,768

16,384

8,192

System/360 Models

Figure

1-3.

System/360

Storage

Sizes

Processor System/360 Models Bytes/ Memory Speed

Access

2030 C30, D30, E30,

F30

1

2.0

microseconds

2040

D4O,

E40, F40,

G40,

H40

2

2.5

microseconds

2050 F50, G50,

H50

4

2.0

microseconds

2060

G60,

H60,

160

8 .

2.0

microseconds

2062 H62,

162

8 1

.0

microseconds

2070 H70,

170

8 1

.0

mi

croseconds

Figure

1-4.

System/360

Storage

Access

and

Speed

1-4

GENERAL

DATA

FLOW

•

The

processing

unit

controls

the

system.

•

Information

enters

the

system

from

an

input

device.

The

information

is

manipulated

by

the

processing

unit

to

develop

the

required

answers.

The

a,nswers

are

sent

to

an

output

device

to

be

stored.

In

the

System/360,

Model

30,

the

2030

processing

unit

provides

all

system

control.

The

processing

unit

is

given

instructions

by

a

programmer.

These

instructions

are

interpreted

and

executed

by

the

processing

unit.

Execu-

tion

of

an

instruction

might

involve

adding

two

numbers

together,

or

it

might

involve

causing

a

printer

to

print

a

check.

Regardless

of

the

instruction,

the

interpretation

and

control

lies

in

the

2030

processing

unit.

General

data

flow

is

divided

into

three

operations

(Figure

1-5).

First,

information

comes

into

the

processing

unit

via

some

input

device.

This

infor-

mation

is

then

used

along

with

input

information

from

other

devices

and

con-

stant

information

contained

in

main

storage

to

develop

the

required

result

or

output.

This

output

is

then

sent

to

an

output

device

where

the

resultant

information

is

stored.

The

information

storage

may

be

a

printed

report,

punched

cards,

or

magnetized

spots

on

a

.reel

of

magnetic

tape.

2030 Processing Unit

Main

Storage

Input/

r--..-(

Output

Channel

Control

Unit

Random

Access

Device

Control

Unit

Random

Access

Device

Data

and

Control Information

-4.----~

Arithmeti

c,

Control,

and

Register

Systems

Figure

1-5.

General

Data

Flow

SYSTEM

CONCEPTS

•

Programming

Systems

support

of

the

IBM

System/360

is

called

Operating

System/360.

Control

Unit

• The

Operating

System/360

supports

the

Computing

System/360.

Together

they

make

up

the

IBM

System/360.

The

introduction

of

the

IBM

Systeml360

marks

the

achievement

of

a

truly

all-

purpose

computer

that

can

solve

any

type

of

data-handling

problem

with

greater

speed

and

efficiency

than

ever

before.

This

opens

up

greatly

increased

computer

potential

in

every

area.

In

order

to

realize

this

potential,

it

was

apparent

to

the

designers

that

the

programming

support

needed

to

be

as

powerful

and

as

extensive

as

the

Control Unit

Printer

computer.

In

fact,

programming

support

should

make

the

IBM

System/360

an

even

mO.re

powerful

system.

This

lead

to

the

concept

of

the

Com-

puting

System/360

(hardware)

being

sup-

ported

by

the

Operating

System/360

(programming),

together

making

up

the

System/360.

r--------------------------,

I I

, I

'r------------------, ,

I

IComputing

System/3601

I

I l

____________________

J I

I I

,

Operating

System/360

IL

__________________________

J

IBM

System/360

This

means

that

a

customer

is

getting

a

system

that

is

powerful

and

flexible,

1-5

yet,

due

to

extensive

programming

sup-

port,

he

can

easily

apply

his

problem

to

the

System.

For

one

thing,

he

can

write

problem

solving

programs

without

the

necessity

of

translating

them

into

a

language

understandable

by

the

machine.

Once

written,

the

operation

of

his

pro-

gram

is

controlled

or

supervised

by

Cperating

System/360,

relieving

the

operator

of

many

tasks

and

increasing

the

utilization

of

the

Computing

System/360.

OPERATING

SYSTEM/360

CONCEPTS

•

Control

programs

allowing

monitored

operation

of

a

system

have

been

proven

by

experience

to

produce

optimum

computer

utilization.

•

Operating

System/360

includes

both

control

programs,

and

IBM

and

user-

written

processing

programs.

•

Basic

programming

Support

programs

will

be

provided

for

System/360

systems

with

8R

bytes

of

storage.

As

stated

previously,

IBM's

single

system

approach

with

the

System/360

recognizes

that

computing

systems

and

programming

systems

should

be

integrated

and

not

developed

independently.

Experience

in

the

past

decade

has

proved

that

the

optimum method

of

producing

this

result

is

with

monitored

operation.

Early

monitors

were

designed

to

mini-

mize

human

intervention.

The new

and

sophisticated

control

techniques

includ-

ed

in

programming

systems

with

the

System/36

0

extend

its

capabilities

so

that

the

monitor

and

control

functions

make up

what

is

called

an

operating

system.

The

basic

purpose

of

Operating

System/360

is

to

permit

the

user

to

solve

problems

and

process

information

effectively.

Included

in

Operating

System/360

are

both

processing

and

con-

trol

programs.

Processing

programs

include

all

application-oriented

pro-

grams,

including

both

IBM

and

user-

written.

For

systems

having

16K

bytes

of

main

storage,

basic

control

program

functions

will

be

supplied

with

magnetic

tapes

or

direct

access

devices.

Additional

1-6

capability

can

be

utilized

as

more

main

storage

is

added.

For

systems

having

8K

bytes

of

main

storage,

programming

systems

is

supply-

ing

basic

programming

support

programs,

which

perform

many

of

the

functions

of

operating

systems,

including

control

functions.

Control

Programs

•

Control

programs

perform

functions

such

as

control

of

administrative

operations,

job

flow

control,

Input/Output

control,

and

program

execution

control.

A

basic

key

to

achievement

of

high

oper-

ating

efficiency

in

a

computing

or

data

processing

installation

is

a

good

con-

trol

procedure.

This

procedure

must

include

many

functions:

administrative

control

of

job

schedules,

workflow,

and

computer

usage

records;

cqntrol

over

data

and

program

libraries;

control

over

computer

operations;

and

control

over

the

flow

of

programs

and

data

within

the

computing

system

during

Job

rUns.

The

control

programs

for

the

IBM

System/360

set

up

a

comprehensive

con-

trol

framework

to

assist

the

user

in

satisfying

the

above

objectives.

The

control

programs

operate

at

various

levels

of

concept.

For

example:

1.

Operations

control

of

installation

and

administration

and

workflOW,

including

instructions

from

and

to

the

computer

operator,

administra-

tive

records,

logs

of

system

opera-

tion,

and

control

over

library

pro-

grams.

2.

Job

flow

control,

including

I/O

transition

between

jobs

and

job

segments,

unit

assignments,

initial

loading

and

initialization

when

the

computer

is

first

turned

on,

control

between

jobs,

and

control

over

the

type

of

operation

mode,

ranging

from

simple

stacked

jobs

through

telepro-

cessing

systems

performing

concur-

rent

ope.rations.

3.

Input/Output

control,

including

physical

and

logical

control

over

I/O

records,

files

and

units:

buffer

control;

teleprocessing

terminal

and

message

handling;

random

access

I/O

control,

labeling

of

files,

and

error

recovery

procedures.

q.

Program

execution

control

that

mana-

ges

the

flow

of

program

instructions

from

one

routine

to

another,

includ-

ing

the

instantaneous

transitions

that

take

place

when

any

interrupt

occurs,

the

decisions

that

control

the

series

to

be

exe-

cuted,

the

return

of

control

to

an

interrupted

program,

storage

alloca-

tion

and

protection,

diagnostic

programs.

program

loading.

and

man-

agement

of

the

interval

timer.

Processing

Programs

•

Processing

programs

function

under

control

of

operating

system

control

programs.

•

Some

IBM

supplied

processing

pro-

grams

are;

the

System/360

Assembler,

FORTRAN,

new

programming

language,

COBOL,

report

program

generator,

utility

programs,

and

sort/merge

programs.

Complementing

the

control

programs

and

functioning

under

them

are

those

pro-

grams

necessary

to

handle

users·

speci-

fic

data

processing

needs.

These

pro-

grams,

known

collectively

as

processing

programs,

include

application

programs

both

IBM

and

user

written,

compilers,

Report

Program

Generators,

sort/merge,

and

utility

programs.

Symbolic

programming

languages

and

the

programs

that

translate

them

(assemblers

and

compilers)

offer

valua-

ble

aids

to

the

program~er

in

solving

data

processing

problems.

The

System/360

Assembler

language

is

a

symbolic

language

that

permits

the

coding

of

source

programs

in

convenient,

report

program

generators,

sort/merge,

specialized

language.

it

can

be

used

in

all

kinds

of

applications,

including

both

commercial

and

scientific.

The

FORTRAN

language

allows

the

pro-

grammer

to

code

a

mathematical

or

scien-

tific

problem

in

te.rms

closely

resem-

bling

those

he

uses

in

stating

the

prob-

lem

mathematically.

The

new

programming

language

has

some

features

that

are

characteristic

of

FORTRAN

and

incorporates

some

of

the

best

features

of

other

languages,

such

as

extensive

editing

capabilities,

to

take

advantage

of

recent

developments

in

computer

technology.

The

COBOL

language

provides

a

conven-

ient

method

of

coding

programs

in

a

form

closely

resembling

the

English

language,

using

the

method

sponsored

by

th~

Con-

ference

on

Data

Systems

.Languages

(CODASYL),

a

cooperative

effort

by a

number

of

computer

manu.facturers

and

users.

The

report

program

generato.r

(RPG)

provides

a

convenient

method

for

produc-

ing

a

wide

variety

of

reports,

using

IBM-provided

coding

forms.

Utility

programs

provide

the

user

with

standard,

efficient

handling

of

routine

operations

involving

data

trans-

fer

between

I/O

devices.

These

include

such

operations

such

as

card

to

printer,

card

to

punch,

card

to

tape,

tape

to

tape,

tape

to

punch,

tape

to

disk,

and

many

others.

The

sort/merge

program

is

designed

to

satisfy

the

sorting

and

merging

require-

ments

of

all

tape-oriented

or

random

storage-oriented

IBM

System/360

instal-

lations.

It

is

a

generalized

program

that

can

produce

many

different

sorting/merging

programs

in

accordance

with

control

information

specified

by

the

user.

1-7

COMPUTING

SYSTEM/360

CONCEPTS

•

System/360

uses

binary

and

BCD.

•

Systeml360

uses

variable

and

fixed

length

fields.

•

System/360

uses

a new

technology

called

solid

logic

technology.

The

Systeml360

is

a

general

purpose

computer

system.

By

this

we

mean

it

is

designated

to

be

used

for

commercial,

scientific,

and

communications

applica-

tions.

In

the

past,

these

applications

were

handled

by

separate

computer

fami-

lies

(Figure

1-6).

t

Growth

7090

709

704

701

SCIENTIFIC

7080

705

III

705

II

702

COMMERCIAL

One

scientific computer fami Iy and its comparable commercial

equivalent.

Figure

1-6.

Commercial

vs.

Scientific

Computers

1-8

The

scientific

computers

were

usually

fixed

word

length

machines

and

used

a

pure

binary

form

of

coding.

On

the

other

hand,

the

commercial

computers

were

usually

variable

word

length

(character

oriented)

machines

and

used

a

binary

coded

representat.ion

of

decimal

information.

The

System/360

uses

binary

as

well

as

BCD

and

has

both

fixed

and

variable

length

fields.

The

System/360

also

uses

a new

tech-

nology

known

as

solid

logic

technology

(SLT).

It

consists

of

printed

circuitry

instead

of

physical

wiring

on

the

back

panel.

It

also

uses

packaged

logic

circuits.

This

new

technology

reduces

manufacturing

costs,

increases

reliabil-

ity

and

reduces

maintenance

time.

In

Figure

1-1,

you

can

see

the

compo-

nents

that

make

up

a

data

processing

system.

2030 Processing Unit

Main

Storage

Input/

r--~

Output

Channel

Control

Unit

Random

Access

Device

Control

Unit

Random

Access

Device

Data

and

Control Information

.....

i-----

..

Arithmetic,

Control,

and

Register

Systems

Control

Unit

Figure

1-7.

Typical

Data

Processing

System

Primary

Storage

• The

smallest

main

storage

addressa-

ble

unit

is

called

the

byte.

•

A

byte

consists

of

8

data

bits

and

1

parity

bit.

Cdd

parity

is

main-

tained.

The

System/360

uses

a 24

bit

binary

address.

A

byte

can

represent

characters,

binary

numbers,

or

many

different

codes.

A

half

word

is

2

bytes.

A

word

is

4

bytes.

A

double

word

is

8

bytes.

•

Control Unit

Printer

Data

can

be

fixed

length

(2,

4,

or

8

bytes)

or

variable

length

(up

to

256

bytes)

•

Fixed

length

data

must

reside

on

the

correct

boundaries

in

main

storage.

A

program

check

occurs

if

the

bound-

ary

restriction

is

violated.

The

primary

storage

is

that

section

of

a

computing

system

that

contains

the

pro-

gram

to

be

executed

as

well

as

the

data

to

be

processed.

All

data

entering

the

system

goes

into

the

primary

storage

before

it

can

be

processed.

After

proc-

essing,

the

data

must

be

placed

back

into

primary

storage

before

it

can

be

sent

to

an

output

device.

1-9

primary

storage

is

sometimes

referred

to

as

main

storage.

The

System/360

also

uses

ferrite

cores

for

its

main

storage.

The

smallest

addreSsable

unit

of

main

storage

in

the

System/360

is

called

the

byte.

The

byte

consists

of

eight

data

bits

and

one

parity

bit.

r~-T-T-T-T-T~-~'

I I I I I I I I I 1

IP 0 1 2 3 4 5 6

71

L~_~_~~~_~~_J

The

Byte

As

can

be

seen

in

the

preceding

exam-

ple,

the

left-most

bit

of

a

byte

is

the

parity

bit.

The

IBM

System/360

main-

tains

odd

parity

for

all

bytes

in

main

storage.

The

pa.rity

bit

is

added

or

removed

to

make

the

total

bit-count

odd

for

any

byte.

This

method

of

coding

provides

convenient

error

checking:

an

even

number

of

bits

indicates

an

error

condition.

One

thing

you

should

get

clear

is

that

the

byte

is

the

smallest

addressa-

bleunit

of

main

storage.

This

means

that

each

and

every

byte

of

main

storage

is

individually

addressable.

To

read

out

the

first

eight

bytes

of

main

stor-

age,

the

Model 30

takes

eight

storage

cycles.

For

each

cycle,

the

Model 30

changes

its

storage

address

by

one,

using

addresses

0-7.

Main

storage

addresses

start

with

00000

for

the

first

byte

and

increase

by

one

for

each

byte

in

the

particular

main

storage

unit.

Valid

storage

addresses

for

a Model

30

would

start

with

00000

and

continue

up

to

65535.

To

allow

for

program

compatibility

as

well

as

for

future

growth,

the

System/360

uses

a 24

bit

binary

address

to

address

main

stor-

age. age.

A

24-bit

binary

number

allows

us

to

go

as

high

as

16777215

for

an

address.

You

can

see

the

future

growth

that

is

possible

here!

A

binary

rather

than

a

binary

coded

decimal

address

is

used

because

it

is

more

effi-

cient

with

large

addresses.

The

24-bit

binary

address

that

would

be

used

to

address

byte

location

0007

is

written

000000000000000000000111.

You

are

probably

a

little

perplexed

about

this

byte

by

now.

You

know

that

a

byte

consists

of

eight

data

bits

and

a

parity

bit.

You

know

that

each

byte

is

individually

addressable

by

a 24

bit

1-10

binary

address!

You

know

that

main

storage

size

can

vary

from

approximately

8K

bytes

on

a Model 30

to

over

500K

bytes

on

a Model

70.

You

know

that

the

Model 30

has

access

to

one

byte

per

storage

cycle.

However,

you

are

probably

asking

yourself:

Is

the

byte

a

character?

Is

it

a

binary

number?

Just

what

is

it?

The

answer

to

these

questions

is

simple.

The

eight

data

bits

of

a

byte

can

be

coded

to

represent

characters,

binary

numbers,

or

anything

you

want

it

to

be.

The

instructions

of

the

System/360

are

many

and

varied.

Some

of

the

instructions

treat

bytes

as

charac-

ters.

Some

instructions

treat

bytes

as

part

of

a

binary

number.

So

the

answer

to

the

question,

·What

does

a

byte

represent?-

is:

It

depends

on

the

par-

ticular

instruction

being

executed

at

the

time.

This

question

will

be

answered

more

to

your

satisfaction

after

you

study

the

data

formats

and

some

of

the

instructions.

As

was

previously

stated,

the

System/360

is

a

general

purpose

data

processing

system.

As

such

it

is

designed

to

operate

with

fixed

length

as

well

as

variable

length

data.

The

byte

as

you

have

already

learned

is

a

very

versatile

unit.

It

is

individually

addressable.

By

further

specifying

the

number

of

desired

bytes,

we

can

have

a

variable

length

field

in

main

storage

starting

and

ending

at

any

byte

address.

To

be

truly

general

purpose,

the

System/360

must

also

be

capable

of

oper-

ating

with

fixed

length

data.

Whereas

variable

length

data

has.a

variable

number

of

bytes,

fixed

length

data

always

has

a

fixed

number

of

bytes.

Letts

go

on

and

define

these

fixed

length

fields.

A

half

word

is

two

bytes

in

length

(Figure

1-8a).

The

data

bit

positions

of

a

half

word

are

numbered

0-15

from

left

to

right

(Figure

1-8b).

Notice

that

the

parity

bits

are

not

shown.

They

will

not

be

shown

form

here

on,

since

they

do

not

represent

data.

Remember,

however,

that

every

byte

does

contain

a

parity

bit

for

checking

pur-

poses.

A word

is

4

bytes

long

(Figure

1-8c).

The

data

bit

positions

o.f a word

are

numbered

0-31

from

left

to

right

(Figure

1-

Sd)

•

A

double

word

is

8

bytes

long

(Figure

1-Se) •

The

data

bit

positions

of

a

double

word

are

numbered

0-63

from

left

to

right

(Figure

1-8f).

a.

Byte Byte

Half

Word

b. I

0:

>>>>>>>>:

1<

11

><

13>4>51

c.

d.

e.

Byte : Byte : Byte :

Half

Word

Byte 1

Word

I0 • •

31

I

Word

Byte Byte

Byt~

Byte Byte

Double Word

Byte Byte Byte

f.

0

"""'.~----------------.

631

Double Word

Figure

1-8.

Word

Formats

Remember

that

each

byte

of

a

half

word,

or

double

word

carries

its

own

parity

bit.

Remember

also

that

it

is

the

instruc-

tion

being

executed

that

determines

whether

to

consider

data

as

variable

or

fixed.

The

instruction

also

determines

in

the

case

of

fixed

length

data

whether

it

is

a

half

word,

or

double

word.

Before

leaving

the

definitions

of

fixed

length

data,

you

must

learn

the

restrictions

placed

on

the

use

of

fixed

length

data.

Byte IByte Byte IByte Byte IByte Byte IByte Byte

0000

0001

0002

0003

0004

0005

0006

0007

0008

Half Word

Half

Word

Half

Word

Half

Word

Word Word

Double Word

Figure

1-9.

Boundary

Restrictions

The

rule

is

that

fixed

length

data

must

reside

on

the

correct

boundaries

in

main

storage

(Figure

1-9).

Fixed

length

data

is

addressed

by

the

high

order

(left-most

byte)

of

the

field.

For

half

words

this

address

must

be

divisible

by

two.

For

words

this

address

must

be

divisible

by

four

•

For

double

words

this

address

must

be

divisible

by

eight.

Another

way

of

stating

this

rule

is

to

say

the

24

bit

binary

address:

1.

Of

a

half

word

must

have

one

low-

order

zero

bit.

2.

Of

a word

must

have

two

low-order

zero

bits.

3.

Of

a

double

word

must

have

three

low-

order

zero

bits.

The

boundary

restriction

placed

on

the

use

of

fixed

length

fields

is

a

restriction

placed

on

the

user.

Viola-

tion

of

these

rules

does

not

produce

a

machine

check.

Instead,

violation

of

these

rules

is

considered

a

program

check.

Because

there

are

other

restrictions

placed

on

the

programmer,

you

should

be

able

to

identify

program

checks

by

type.

The

type

of

program

check

caused

by

a

violation

of

fixed

length

boundaries

is

known

as

a

specification

exception.

1-11

Another

exception

to

valid

program-

ming

is

addressing

a

byte

location

that

is

not

available

on

your

particular

model

of

System/360.

The

largest

size

main

storage

that

is

available

on

the

Model 30

is

65,536

bytes.

Any

address

other

than

00000-65535

results

in

a

program

check.

This

type

of

check

is

known

as

an

addressing

exception.

Central

Processing

Unit

(CPU)

•

The

two

main

sections

of

the

CPU

are

(1)

the

control

section

and

(2)

the

arithmetic

and

logical

unit

(ALU).

The

CPU

uses

variable

field

length

instructions

for

storage

to

storage

operations.

Variable

fields

can

be

up

to

256

bytes

long.

The

CPU

uses

fixed

length

instruc-

tors

for

storage

to

register

or

register

to

register

operations.

Fixed

length

fields

can

be

half-word,

word,

or

double-word

fields.

Register-to-register

or

storage

to

register

operations

use

any

of

16

general

purpose

registers.

Floating

point

operations

use

any

of

4

double-word

floating-point

reg-

isters

In

Figure

1-10,

you

can

see

the

logical

1-12

structure

of

the

CPU

for

the

System/360

and

its

relationship

to

the

main

storage.

There

are

two

main

sections

in

CPU.

They

are

(1)

the

control

section

and

(2)

the

arithmetic

and

logical

unit

(ALU).

From

Figure

1-10,

you

should

be

able

to

see

some

of

the

functions

of

the

control

section.

They

are:

1.

All

references

to

main

storage,

whether

for

instructions

or

for

data,

are

made by

the

control

sec-

tion.

2.

During

the

first

part

of

any

instruction,

the

control

section

addresses

main

storage

and

causes

the

instruction

to

be

fetched

and

sent

to

the

control

section.

The

instruction

is

then

decoded

by

the

control

section

and

executed

during

the

latter

part.

The

ALU

contains

the

circuits

neces-

sary

for

adding

and

comparing

data

fields

as

well

as

the

other

circuits

necessary

for

operating

on

data

fields.

As

can

be

seen

from

Figure

1-10,

the

ALU

can

do:

1.

Variable

field

length

operations.

2.

Fixed-point

operations

involving

fixed-length

fields.

3.

Floating

point

operations.

Storage Address

Instructi

ons

Computer

System

Control Indexed Fixed

Address Point

Operations

16

General

Registers

Figure

1-10.

ProceSSing

Unit

Logic

Flow

Variable

Field

Length

Operations:

In

looking

at

the

ALU,

let

us

first

consid-

er

variable

length

fields

as

used

in

many

commercial

computers

of

the

past.

Two

main

concepts

were

used.

The

storage-to-storage

concept

was

used

by

computers

of

the

IBM

1401

family.

In

it

the

data

fields

were

b~ought

out

of

main

storage,

operated

upon~

and

the

results

went

back

into

main

storage

(Figure

1-11)

•

Other

computers

such

as

those

of

the

IB~

702-705

family

used

a

storage-to-accumulator

concept.

The

accumulator

was a

small

storage

device.

The

storage

medium

could

be

core

Main Storage

Data

Variable Floating Arithmetic

and

gic

Unit

Field

length

Point

lo

Operations Operations

4

Floating Point

Registers

Primary

Storage

1 1

AlU

Storage to Storage Concept

Figure

1-11.

Storage

to

Storage

Concept

1-13

storage.

vacuum

tubes

or

transistorized

registers.

In

the

storage

to

accumulator

concept

one

of

the

data

fields

would

be

in

main

storage

and

the

other

would

be

in

an

accumulator.

Both

fields

would

be

brought

out

to

the

ALU,

operated

upon,

and

the

result

would

go

back

into

the

accumulator

(Figure

1-12).

Main

Storoge

1

ALU

J 1

Accumulator

Storage to Accumulator

Concept

Figure

1-12.

Storage

to

Accumulator

Concept

Fo.r

its

variable

length

operations

the

System/360

uses

the

storage-to-

storage storage

concept

(Figure

1-13).

Main

Storage

1 f

Variable

Field

Length

ALU

Operations

Figure

1-13.

System/360

Storage

to

Storage

Operations

As

you

have

previously

learned.

variable

length

fields

can

start

at

any

byte

location

in

main

storage.

They

are

not

restricted

by

storage

boundaries

as

are

fixed

length

operands.

However.

1-14

there

fixed

length

operands.

(Data

fields

are

sometimes

referred

to

as

operands.)

However.

there

must

be

some

way

of

indicating

to

the

system

the

length

of

the

fields.

In

computers

of

the

past,

this

was

done

several

ways.

The 1401

used

a

special

word mark

bit

over

the

high-order

position

of

the

data.

The

IBM

70S-II

used

zone

bits.

In

the

System/360

variable

length

opera-

tions

use

binary

and

decimal

operands.

In

order

to

be

code

independent,

System/360

specifies

the

length

of

these

fields

by a

length

code

in

the

instruc-

tion.

The

length

code

can

be

either

4

or

8

bits

long,

depending

on

the

instruction.

The

length

code

is

in

binary.

As

a

result

the

maximum

length

can

be

either

16

or

256

bytes.

The

values

of

the

code

is

one

less

than

the

total

number

of

bytes.

Length

code

of

0000 = 1

Byte

Length

code

of

1111 = 16

Bytes

Length

code

of

11111111 = 256

Bytes

Fixed-Length

Operations:

When

operat-

ing

on

fixed-length

fields

(such

as

half

words.

words,

or

double

words).

the

System/360

uses

the

storage-to-accumulator

concept.

These

fixed-length

operations

use

binary

operands.

For

use

as

accumulators,

the

system/360

has

16

registers

available

to

the

programmer.

As

these

registers

can

be

used

for

purposes

other

than

accumu-

lating.

they

are

called

general

reg-

iste.rs

(Figure

1-14).

Control

Section

Fixed

Point

Operatior

Sixteen

General

Registers

Main

Storage

Figure

1-14.

16

General

Registers

ALU

These

registers

are

numbered

0-15

and

are

addressed

in

an

instruction

by

a

4-bit

binary

address

field.

Being

a word

in

length,

a

general

register

can

easily

contain

a

half

word

data

field.

As

can

be

seen

in

Figure

1-15,

the

bits

of

a

general

register

are

numbered

left

to

right

starting

with

the

number

O.

Also

we

can

see

that

a

half

word

operand

is

placed

in

the

low-order

half

(bits

16-31)

of

a

General

Register.

BitO

---15

16-

-31

I Half Word I

.

Operand

.

GENERAL

REGISTER

Figure

1-15.

General

Register

None

of

the

General

Registers

0-15

can

contain

a

double

word.

For

those

operations

that

use

a

double

word

oper-

and,

such

as

fixed

length

divide,

a

pair

of

adjacent

registers

are

used.

In

these

cases,

an

even-odd

pair

of

reg-

isters

(such

as

0-1

or

6-7)

are

used,

and

the

even

register

is

addressed.

In

this

case

bits

0-63

of

the

double

word

would

be

in

the

registers

as

shown

in

Figure

1-16.

0

31

0

31

10

DOUbl~

Word

63 1

Reg

12

Reg

13

Figure

1-16.

Using

Two

General

Registers

Fixed-length

operands

in

main

storage

must

be

on

integral

boundaries

or

a

program

check

will

occur

indicating

a

specification

exception.

The

general

registers

are

also

used

for

purposes

other

than

accumulating.

For

example,

a

general

purpose

register

can

be

used

as

an

index

register.

Indexing

is

a

form

of

indirect

address-

ing.

An

increment

contained

in

an

index

register

is

added

to

the

data

address

in

the

instruction

to

form

an

effective

main

storage

address.

Neither

the

index

register

nor

the

instruction

in

storage

is

changed

by

indexing_

1-15

Register-to-Register:

With

sixteen

general

registers,

sometimes

both

fixed

length

binary

operands

will

be

in

gener-

al

registers.

In

these

cases,

another

data

flow

concept

is

used

(register-to-register

operation,

Figure

1-17) •

Addresses

Contral

Section Instruction

Main

Storage

Fixed

length

I I

AlU

General

Registers

0-15

Figure

1-17.

System/360

Register

to

Register

Operations

Floating-Point

Operation:

Floating

point

is

the

term

given

to

arithmetic

operations

involving

a

fraction

and

an

exponent.

For

instance:

217,000

can

be

expressed

as:

.217

x

10·

296,000

can

be

expressed

as:

.296

x

10.

Fixed

point

arithmetic

would

add

the

numbers

thusly:

1-16

217,000

+

296,000

513,000

I

Floating

point

arithmetic

would

do

it

like

this:

.217

x

10·

+

.296

x

10.

.513

X

10

6

Floating-point

arithemetic

is

most

useful

for

expressing

very

large

numbers

and

operating

on them

with

much

precision.

To do

floating

point

arith-

metic

the

System/360

has

four

floating

point

re~isters

(Figure

1-18).

Addresses Main

I Storage

Control Instructions I

Section • IFloating

AlU

Point

1 1

Four

floating

Point

Registers

Figure

1-18.

Floating

Point

Registers

The

four

floating

point

registers

are

numbered

0,

2,

4,

6.

These

are

not

the

same

as

general

registers

0,

2,

4,

6.

The

floating

point

registers

are

separ-

ate

registers

used

only

as

accumulators

during

floating

point

operations.

The

floating

point

registers

are

double

word

registers

and

are

addressed

by a

four

bit

binary

address

in

floating

point

instructions.

Bits

0 63

r------------------,

I

F~P.

Reg 6 I

L-

_________________

J

F.P.

Reg

Address

= 0110

Logical

vs

Hardware

Structure

of

System/360:

The

structure

of

the

System/360

which

you

have

been

learning

is

its

logical

structure:

By

this

we

mean

that

this

is

the

way

the

System/360

appears

to

the

programmer.

The manner

in

which

this

logical

structure

is

implemented

will

vary

between

the

dif-

ferent

models

of

the

System/360.

IBM 2030 Manual

Popular Desktop manuals by other brands