This pUblication provides, for reference
purposes, a detailed definition of the
machine functions performed by System/370. The publication applies only to systems
operating in the System/370 mode. The
IBM 370-XA Principles of Operation, SA22-7085, should be consulted regarding
the functions of the architecture which
apply to systems operating in the 370-XA mode, and the IBM 4300 Processors Prin­
ciples of Operation for ECPS:VSE Mode, GA22-7070, should be consulted regarding
the functions of the architecture which
apply only to systems operating in the
VSE mode.
The publication describes each function
at the level of detail needed to prepare
an assembler-language program that
relies on that function. It does not,
however, describe the notation and
conventions that must be employed in
preparing such a program, for which the
user must instead refer to the appropri­
ate assembler-language publication.
The information in this publication is
provided principally for use by
assembler-language programmers, although
anyone concerned with the functional
details of System/370 will find it
useful.
This publication is written as a refer­
ence and should not be considered an
introduction or a textbook. It assumes
the user has a basic knowledge of data­
processing systems. IBM publications
relating to System/370 are listed and
described in the IBM System/370, 30xx, and 4300 Processors Bibliography, GC20-0001. All facilities discussed in this publi­
cation are not necessarily available on
every model. Furthermore, in some
instances the definitions have been
structured to allow for some degree of
extendibility, and therefore certain
capabilities may be described or implied
that are not offered on any model.
Examples of such capabilities are the
number of channel-mask bits in the
control register, the size of the CPU
address, and the number of CPUs sharing
main storage. The allowance for this
type of extendibility should not be
construed as implying any intention by
IBM to provide such capabilities. For
information about the characteristics
and availability of facilities on a
specific model, see the functional char­
acteristics publication for that model. ~ Largely because this publication is ~ arranged for reference, certain words ; PREFACE and phrases appear, of necessity, earli­
er in the publication than the principal
discussions explaining them. The reader
who encounters a problem because of this
arrangement should refer to the index,
which indicates the location of the key
description.
The information presented in this publi­
cation is grouped in 13 chapters and
several appendixes:
Chapter 1, Introduction, highlights some
of the major facilities of System/370. Chapter~, Organization, describes the
major groupings within the system --the
central processing unit (CPU), storage,
and input/output --with some attention
given to the composition and character­
istics of those groupings. Chapter d, Storage, explains the infor­
mation formats, the addressing of stor­
age, and the facilities for storage
protection. It also deals with dynamic
address translation (OAT), which,
coupled with special programming
support, makes the use of a virtual
storage possible in System/370. Dynamic
address translation eliminates the need
to assign a program to a fixed lo~ation in real storage and thus reduces the
addressing constraints on system and
problem programs. Chapter ~, Control, describes the facil­
ities for the switching of system
status, for special externally initiated
operations, for debugging, and for
timing. It deals specifically with CPU states, control modes, the program­
status word (PSW), control registers,
program-event recording, timing facili­
ties, resets, store status, and initial
program loading. Chapter 2, Program Execution, explains
the role of instructions in program
execution, looks in detail at instruc­
tion formats, and describes briefly the
use of the program-status word (PSW), of
branching, and of interruptions. It
contains the principal description of
the dual-address-space (DAS) facility.
It also details the aspects of program
execution on one CPU as observed by
other CPUs and by channels. Chapter ~, Interruptions, details the
mechanism that permits the CPU to change
its state as a result of conditions
external to the system, within the
system, or within the CPU itself. Six
classes of interruptions are identified
and described: machine-check interrup­
tions, program interruptions, super­
visor-call interruptions, external
iii

interruptions, input/output interrup­ tions, and restart interruptions. Chapter 7, General Instructions,
contains detailed descriptions of
logical and binary-integer data formats
and of all unprivileged instructions
except the decimal and floating-point
instructions. Chapter ~, Decimal Instructions,
describes in detail decimal data formats
and the decimal instructions. The deci­
mal instructions are a part of the commercial instruction set. Chapter 1, Floating-Point Instructions,
contains detailed descriptions of
floating-point data formats and the
instructions provided by the floating­
point facility and by the extended­
precision floating-point facility. Chapter 1!, Control Instructions,
contains detailed descriptions of all of
the semiprivileged and privileged
instructions except for the I/O
instructions. Chapter 11, describes the
correcting,
malfunctions. Machine-Check Handling,
mechanism for detecting,
and reporting machine Chapter lZ, Operator Facilities,
describes the basic manual functions and
controls available for operating and
controlling the system. Chapter 11, Input/Output Operations,
explains the programmed control of I/O
devices by CPUs and by channels. It
includes detailed descriptions of the
I/O instructions, channel-command words,
and other I/O-control formats.
The Appendixes include: • Information about number represen­
tation • Instruction-use examples • Lists of the instructions arranged
in several sequences • A summary of the condition-code
settings • A list of the System/370 facilities • A table of the powers of 2 • Tabular information helpful in
dealing with hexadecimal numbers • An EBCDIC chart • tv A discussion of changes
compatibility between
and System/370 affecting System/360 • A discussion of changes affecting
compatibility within System/370 SIZE NOTATION In this publication, the letters K and M
denote the multipliers 2
10
and 2
20
,
respectively. Although the letters are
borrowed from the decimal system and
stand for kilo (10
3
) and mega (10
6
), they do not have the decimal meaning but
instead represent the power of 2 closest
to the corresponding power of 10. Their meaning in this pUblication is as
follows:
Symbol Value
K (kilo) 1,024 =
2
10
M (mega) 1,048,576 =
2
20
The following are some examples of the
use of K and M: 2,048 is expressed as 2K. 4,096 is expressed as 4K.
65,536 is expressed as 64K
(not 65K).
224 is expressed as 16M.
When the words "thousand" and "million"
are used, no special power-of-2 meaning
is assigned to them. BYTES, CHARACTERS, AND CODES Although the System/360 architecture was
originally designed to support the
Extended Binary-Coded-Decimal Inter­
change Code (EBCDIC), the instructions
and data formats of the architecture are
for the most part independent of the
external code which is to be processed
by the machine. For most instructions,
all 256 possible combinations of bit
patterns for a particular byte can be
processed, independent of the character
which the bit pattern is intended to
represent. For instructions which use
the zoned format, and for those few
instructions which are dependent on a
particular external code, the instruc­
tion TRANSLATE may be used to convert
data from one code to another code.
Thus, a machine operating in the System/370 mode can process EBCDIC, ASCII, or any other code which can be
represented in eight or fewer bits per
character.
In this publication, unless otherwise
specified, the value given for a byte is the value obtained by considering the 4 bits of the byte to represent a binary, code. Thus, when a byte is said to