Digital PDFs

DEC-15-GFZA-D-PDP-15

September 1971

122 pages

Original

36MB

Document:	FORTRAN IV Operating Environment
Order Number:	DEC-15-GFZA-D-PDP-15
Revision:
Pages:	122
Original Filename:	http://bitsavers.org/pdf/dec/pdp15/XVM/DEC-15-GFZA-D-PDP-15_FORTRAN_IV_OPERATING_ENVIRONMENT.pdf

OCR Text

DEC-15-GFZA-D

PDP-15 FORTRAN IV
OPERATING ENVIRONMENT

For additional copies of this manual, order DEC-! 5-GFZA-D from Digital Equipment
Corporation, Program Library, Maynard Mass. 01754

Price $6.00

DIGITAL EQUIPMENT CORPORATION • MAYNARD, MASSACHUSETTS

1st Edition September 1971

The material in this manual is for informational purposes and is subject to change
without notice.

The following are trademarks of Digital Equipment
Corporation, Maynard, Massachusetts:
DEC

FLIP CHIP
DIGITAL

PDP
FOCAL
COMPUTER LAB

CONTENTS
Page
CHAPTER 1

INTRODUCTION

1.1
1.2
1.2.1
1.2.2
1.2.3
1.2.4
1.2.5
1.3

Operating Procedures
Software Environments
DOS-IS
ADVANCED tvbnitor Software System (ADSS)
PDP-15/30 Background/Foreground tvbnitor System
RSX-15 Real- Time Execution
BOSS-IS
Hardware Environment

CHAPTER 2

INPUT-OUTPUT PROCESSING

2.1
2.1.1
2.1.2
2.1.3
2.2
2.3
2.3.1
2.3.2
2.3.3
2.4
2.4.1
2.4.2
2.4.3
2.4.4
2.5
2.6
2.7
2.7.1
2.7.2

General Information
Device Assignment
Data Structures
Data Transmission
OTS lOPS Communication (FlOPS)
Sequential Input-Output
OTS Binary Coded Input/Output (BCDIO)
OTS Binary Input/Output (BINIO)
OTS Auxiliary Input/Output (AUXIO)
Direct Access I/O
The DEFINE Routine
Formatted Input/Output (RBCDIO)
Unformatted Input/Output (RBINIO)
Initialization and Actual Data Transfer (RANCOM)
Data-Directed Input-Output (DDIO)
Encode/Decode (EDCODE)
User Subroutines
Magnetic Tape Input-Output Routines
Directoried Subroutines

CHAPTER 3

THE SCIENCE LIBRARY

3.1
3.2
3.2.1
3.2.2
3.2.3
3.2.4
3.2.5
3.2.6
3.3
3.3.1
3.3.2
3.4

Intrinsic Functions
External Functions
Square Root (SORT, DSORT)
Exponential (EXP, DEXP)
Natural and Common Logarithms (ALOG, ALOG 10,
DLOG, DLOG10)
Sine and Cosine (SIN, COS, DSIN, DCOS)
Arctangent (ATAN, DATAN, ATAN2, DATAN2)
Hyperbolic Tangent
Sub- Functions
Logarithm, Base 2 (. EE, . DE)
Polynominal Evaluator (. EC, .DC)
The Arithmetic Package

CHAPTER 4

UTILITY ROUTINES

4.1
4.2

OTS Routines
Floating Point Processor Routines

iii

1-1
1-6
1-6
1-6
1-6
1-7
1-7
1-7

2-1
2-2
2-2
2-3
2-3
2-4
2-5
2-6
2-7
2-9
2-9
2-11
2-12
2-13
2-13
2-15
2-15
2-15
2-16

3-2
3-6
3-6
3-6
3-8
3-9
3-10
3-10
3-11
3-11
3-14
3-14

4-1
4-4

CO NTENTS (Cont)
Page
4.3
4.4

FORTRAN - Callable Utility Routines
RSX Library (.LIBRX BIN) Routines

CHAPTER 5

FORTRAN-IV AND MACRO

5.1
5.2
5.3

Invoking MACRO Subprograms from FORTRAN
Invoking FORTRAN Subprograms from MACRO
Common Blocks

4-5
4-5

5-1
5-2
5-3

APPENDICES
APPENDIX A LANGUAGE SUMMARY

A-1

APPENDIX B

ERROR MESSAGES

B.1
B.2
B.3

Compiler Error Messages
OTS Error Messages
OTS Error Messages in FPP Systems

APPENDIX C

PROGRAMMING EXAMPLES

C.1
C.2
C.3

MACRO-FORTRAN Linkages
IFLOW and IDZERO Examples
Input-Output Examples

B-1
B-7
B-9

C-1
C-5
C-6

APPENDIX D SYSTEM LIBRARIES
.LIBR - Page Mode Non-FPP
• LIBRF - Page Mode FPP

D-1
D-4

APPENDIX E PDP-15 FORTRAN FACILITIES

E-1

D.1
D.2

ILLUSTRATIONS
Title

Figure No.
1-1

Page
1-4

Sample DOS-15 Session
TABLES

Table No.
3-1
3-2
3-3
3-4
4-1
4-2
E-1
E-2
E-3

Title

Page

Intrinsic Functions
External Functions
Sub- Functions
Arithmetic Package
FORTRAN-Callable Utility Routines
FORTRAN-Callable RSX Routines
Versions of the Extended Compiler
Versions of the OTS Libraries for the Extended Compiler
Compilers and Libraries for Extended FORTRAN Distributed
with PDP-9/15 Systems

3-3
3-7
3-12
3-17
4-6
4-9
E-1
E-2
E-2

PREFACE

This manual describes the system software facilities which support the PDP-15 FORTRAN IV compilers
together with hardware features which affect the FORTRAN programmer. Included are discussions of
monitor features which are of interest to the FORTRAN programmer, the FORTRAN IV Object Time
1
2
System (OTS), and the Science library . All descriptions presented are based on the most comprehensive version of the FORTRAN compiler. Appendix E presents overall outlines and descriptions and
detailed data specifying the differences between the various compilers for all of the FORTRAN IV
versions offered.
A companion manual "PDP-15 FORTRAN IV LANGUAGE MANUAL", order code DEC-15-GFWA-D,
describes the elements, syntax and use of the FORTRAN IV language as implemented for the PDP-15
computer.

lThe Object Time System is a set of subroutines which are automatically invoked by certain FORTRAN
language elements. A FORTRAN input-output statement, for example is not compiled directly into
executable object code but becomes a call to the appropriate OTS input-output routine.
2The Science library is a set of intrinsic functions, external functions, subfunctions, and subroutines
which the user may invoke explicitly in a FORTRAN statement.

CHAPTER 1
INTRODUCTION

A FORTRAN-IV program may be compiled and run in several different environments. The FORTRAN
programmer need not be concerned with the details of his environment since the FORTRAN Object-Time
System (OTS) will ensure that his statements invoke the appropriate computer instructions. For example,
an arithmetic statement such as A = A*B will appear the same in any FORTRAN-IV program. In the
object program it may be transformed to a subroutine call, an EAE instruction, or a floating point
instruction, depending on the hardware configuration on which the program is produced.
He will need to know procedures for compiling and loading his program and for using the peripheral
devices available to him. In addition, a number of software facilities may be of interest to a
FORTRAN programmer who requires maximum program efficiency or functions not performed by
FORTRAN statements. In this case, he may invoke FORTRAN-callable functions and subroutines from
the FORTRAN library or augment his program by linking to MACRO assembler programs and invoking
the OTS utility routines. *
In this chapter, we describe the basic procedures for using FORTRAN and the major facilities
available to a FORTRAN program. These facilities are described in greater detail in subsequent
chapters, and Appendix C contains a collection of illustrative programming examples. The main
discussion is based on the DOS-IS monitor, and differences for other environments are noted.

1. 1 OPERATING PROCEDURES
The FORTRAN-IV compiler is a two-pass system program which produces relocatable object code.
This code is then linked with user-specified FORTRAN-compiled or MACRO-IS assembled routines
and with required OTS library routines. Program linkage may be accomplished via the linking
loader, LOAD, which loads the resulting program directly into core in absolute format. The user may,
alternatively, use one of the overlay linkage editors - CHAIN (DOS-IS, ADSS, B/F, Basic I/o
Monitor) or TKB (RSX). These construct core images onto auxiliary storage.
*In all MACRO calling sequences given - when an address is required as an argument, it may be
expressed as +400000 to indicate indirection.
1-1

The FORTRAN-IV compiler is called by typing F4 after the monitor has issued a $. When FORTRAN
has been loaded, the version name is typed at the left margin as in:
F4X Vnn
A carriage return is issued and the character> at the left margin indicates that a command string is
expected with the FORTRAN source program on the appropriate input.
The command string has the form:
optionlist .. filename
where the options are delimited by a left arrow and may optionally be separated by commas, and the
string is terminated by a carriage return or ALT MODE. A carriage return specifies that FORTRAN-IV
should be restarted after the current program has been compil ed. ALT MODE returns control to the
monitor.
The option list may be blank or contain any of the following options:

Meaning

Option*

object listing

symbol map

source listing

binary output

output listing on DECtape unit 2

write output on DECtape unit 1

--.
Filename must be a legal FORTRAN symbol. The output listing always has the extension LST.
At the end of pass 1, the compiler types
END PASSl
to accomodate the repositioning of a paper-tape source file in the reader. When compiling from
paper tape, to initiate pass 2, the user types tP (control P). Otherwise, pass 2 is initiated automatically.

*Refer to Appendix E for list of options applicable to each version of FORTRAN

1-2

The following error messages indicate that the command procedures cannot be carried out:
Message

Meaning
Bad command string - retype

lOPS 4

I/o device not ready - type CTRL R when ready

lOPS

See PDP-15/20 User's Guide for lOPS error codes

Other diagnostics which may be printed at compile time are FORTRAN error messages (see Appendix B,
Section B. 1). OTS errors are given at run time for those routines whose calls are generated by the
compiler (see Appendix B, Section B.2).
When the user program has been successfully compiled, it may be relocated and made absolute
(executable) via LOAD, CHAIN, or TKB (the RSX Task Builder).
The Linking Loader is called by typing LOAD or GLOAD (Ioad-and-go) after a monitor-issued $.
The Linking Loader types
LOADER Vnn

>
and awaits a command string specifying programs to be loaded and output options. See the
1
PDP-15/20 User's Guide for detailed instructions. Figure 1-1 shows the printout from a typical
DOS-15 session from source-program preparation to loading.
With CHAIN, the user generates a system of overlays - a resident main program which may include
resident subprograms, a resident blank COMMON storage area, and a set of subroutines which
overlay each other at the user's request. Subroutines are organized into units called LINKS which
may overlay each other. Several LINKS may overlay a larger LINK without overlaying each other.
A LINK is loaded into core when a subroutine within the LINK is called and it remains resident until
overlayed. A LINK's core image is not recorded or "swapped out" when it is overlayed. The same
image is brought into core each time a LINK is loaded. See the PDP-15 CHAIN and EXECUTE
manual for detailed instructions (DEC-15-YWZA-DN2).

1Order code DEC-15-MG2C-D
1-3

DOS-IS V02
ENTER DATE

(t-'M/OD/YY)

6/8/71

ROG IN DEM

$P IP
DOSPIP VIA
>N OK

>fC
DOS-IS V0?

S':EDIT

EDITOR V10A
>OPEN IOTST
IOTST SRC NOT
INPUT

~ILE

~OUNO.

TTY:

.DAT 6

C
11210

20121

\~RITE
(6,100)
FORMAT (IX,$IN:$)
READ (6,) fn ,1..:2
WRITE (6,2 00 )
~ORt-'AT (IX,
'OUT:')

H3=Pl**R2
ltJRITE (6,) R3
STOP
END

EDIT
>CLOSE
EDITOR V10A
>fC
DOS-IS V02
$F4
~4X

VISA
>8" IOTST
END PASS 1

DOS-IS V02
5A TT 6
$LOAD

Figure 1-1

Sample DOS-1S Session

-"
(continued next page)

1-4

BLOAOER VIlA
>p .. 10TST
p
rOTST
p
0010
007
P
.BE
006
P
.EE
002
P
.EF'
004
P
.Ee
001
P BCOIO
028
P
.SS
005
p
fjfil3
STOP
P SPMSG 004
P .F'UB 00.4
P F'IOPS
016
P DBLI NT 058
p
INTEAE 008
P DOUBLE 004
p
RELEAE 016
P
OTSER 009
P
.CB
003
t

77535
75463
75430
75337

75221
75155
71230
7115el
71 135
71042
70554
67652
67246
67112
66707
65576
65366
65346

S 1S

IN:
11 .2 .. 3.0

OUT:
'R3'=
STOP

1404.9282

000000

00S-15 V02
$

Figure 1-1

Sample DOS-15 Session (Cont)

TKB is similar to CHAIN. Its function is to record core images in a file in the format expected by
the RSX INSTALL MCR Function. The task name is used as the file name, and TSK is used as the
extension. TKB uses the same .DAT slots and accepts the same overlay descriptions as CHAIN.
It is called by typing "TKB" following the Monitor's $ request. When loaded, TKB types its nome
and version number and makes the following requests:
LIST OPTIONS
NAME TASK
SPECIFY DEFAULT PRIORITY
DESCRIBE PARTITION
DESCRIBE SYSTEM COMMON BLOCKS
DEFINE RESIDENT CODE
DESCRIBE LINKS AND STRUCTURE
For further information, see RSX-15 Reference Manual (DEC-15-GRQA-D).

1-5

1.2

SOFTWARE ENVIRONMENTS

Each version of FORTRAN-IV has its own version of OTS and the Science Library so that routines may
utilize both hardware and software features. Each of the monitor systems under which FORTRAN
operates is summarized below.

1.2.1

005-15

DOS-15 is a single-user, interactive, disk-resident Operating System. It includes the DOS-15
Monitor, I/o device handlers r and an integrated set of system programs including FORTRAN-IV.
Program editing, loading, and debugging facilities are provided as well as powerful file manipulation
capabilities. The DOS-IS disk file structure supports both direct and sequential access to disk files,
dynamic disk storage allocation, and file protection. The DOS-IS Monitor itself provides the
interface between the user and peripheral devices via Monitor calls and allows the user to load system
or user programs, for example, FORTRAN programs r via simple commands from the user terminal. The
reader is directed to the 005-15 Software System User's Manual, DEC-15-MRDA-D, for more detailed
information.

1.2.2

ADVANCED Monitor Software System (ADSS)

The ADVANCED Monitor Software System is an integrated system of programs which includes the
ADVANCED Monitor, an Input-Output Processor (lOPS), and a set of system programs which prepare,
compile, assemble, debug, and operate user programs. The monitor itself serves as the interface
between FORTRAN and peripheral devices and between the user console and the system. Detailed
information on the components of ADSS may be obtained in the ADVANCED Monitor Software System
Manual, DEC-15-MR2B-D.

1.2.3

PDP-15/30 Background/Foreground Monitor System

The Background/Foreground Monitor (B/F) is an extension of the ADVANCED Monitor which permits
concurrent, time-shared use of the PDP-15/30. This is done through protected, foreground user
programs with a background of batch processing, through program development, or through low-priority
user programs. Details are available in the PDP-15/30/40 Background/Foreground Monitor Software
System manual (DEC-15-MR3A-D).

1-6

1.2.4

RSX -15 Real-Time Execution

RSX-15 is a monitor system designed to handle real-time information in a multiprogramming environment.
RSX-15 controls and supervises all operations within the system including any number of core- and diskresident programs (called tasks). The user can dynamically schedule tasks via simple time-directed
commands issued from the terminal or from within a task. RSX uses the ADVANCED Software Monitor
(1.2.2) and a Real-Time Monitor. System software includes the FORTRAN-IV compiler, the MACRO
Assembler, the TASK BUILDER, and numerous utility programs required to edit, compile, debug, and
run user programs. Details are available in the RSX-15 Real/Time Executive Reference Manual
(DEC-15-GRQA-D) •

1.2.5

BOSS-15

B05S-15 is a batch-processing monitor which is part of D05-15; it, therefore, utilizes the DOS-15
system program and file structures. DOS-15 itself has a facility to batch commands from cards or
paper tape; BOSS-15, however, is a separate entity from· DOS-15 batch. BOSS-15 I s command
language is batch-oriented, noniterative, easy to use, and highly flexible.
Some highlights of BOSS-15 are:
• Procedure driven command language
• Job timing for accounting purpose
• Line editor
• Facility for user-defined commands
B05S-15 provides the user with the abil ity to use any system program (with exception of some programs
that work only in an interactive environment) and the disk-file structure of DOS-1S.

1.3

HARDWARE ENVIRONMENT

Systems with a Floating-Point Processor (FPP) have a special version of the FORTRAN -IV compiler and
OTS which utilizes hardware instructions rather than software calls. For example, RELEAE, the REAL
arithmetic package, is not included in FPP systems since REAL arithmetic expressions may be compiled
into computer instructions.
The FPP F4X System consists of the standard DOS-1S FORTRAN-IV compiler and Object-Time System
(OTS) interfaced (via conditional assembly, and additional routines) to the hardware PDP-1S FPP
(Floating-Point Processor). The interface applies to Single and Double Precision Floating-Point
Arithmetic and Extended Integer Arithmetic (double integers). Single integer arithmetic is still
handled by software.

1-7

Floating-Point (FPP) FORTRAN-N is available in different forms for use in PDP-1S software systems
other than the DOS-1S system. See Appendix E for descriptions of the available types of FORTRAN-IV.
The following points should be noted with respect to the software modifications which accompany the
FPP software systems:
(1) The calling sequence for integer power involution (raising numbers to integer powers) has
been changed. The associated OTS routines will have to be updated throughout any
systems usi ng F4X.
(2) All systems that support a bank mode will require a bank mode version of the F4X compiler
to go along with their respective OTS libraries in order to suppress generation of PDP-1S
instructions (see Appendix D). Note that a bank mode version of the FPP F4X is not
needed because the FPP cannot be added to a PDP-9.
The FPP libraries (given in Appendix D) include the program .FPP which contains a special FPP
error-handling routine, and routines which handle communication between the hardware CPU AC
used by FORTRAN and the FPP accumulator.
All routines described in the science library and OTS utility programs are available in FPP versions
with the exception of RELEAE, DOUBLE and DBLINT which are no longer required.

'- .

1-8

CHAPTER 2
INPUT-OUTPUT PROCESSING

FORTRAN data-transmission statements automatically invoke a number of OTS subroutines which serve
as an interface between the user program and the Monitor. These routines may also be explicitly
referred to in a MACRO program.
The actual transmission of data between memory and a peripheral device is, in general, performed by
the FlOPS package, a set of routines which communicate directly with the Monitor. Other packages,
each associated with a particular type of data-transmission statement, perform three major functions:
a.

Initialization,

Transmission of data to and from the FORTRAN line-buffer in the appropriate structure, and

Termination;

The packages are:

(l) BCDIO, processes formatted sequential READ or WRITE statements;
(2)

BINIO, processes unformatted sequential READ or WRITE statements;

(3) AUXIO, processes auxiliary input-output statements;
(4)

RBCDIO and RBINIO, processes formatted and unformatted direct-access READ and
WRITE statements;

(5)

0010, manages data-directed input-output;

(6)

ENCODE, processes ENCODE and DECODE statements.

Also described in this chapter is a set oflORTRAN-caliable subprograms which support OTS inputoutput functions.

2.1

GENERAL INFORMATION

The three major I/o functions:
a.

To associate logical devices with physical devices,

To associate user data structures with device data structures, and

To perform actual transfer of data

are described in the following paragraphs.
2-1

2.1. 1

Device Assignment

In all systems except RSX, device assignment is managed through the monitor Device Assignment
Table (.DAT) which associates logical device units to physical ones •• DAT has "slot" numbers which
correspond to the logical device numbers. Each slot, at run time, contains the physical device
number and a pointer to the appropriate device handler. Sixteen* entries in • DAT may be used for
user-program device assignment performed via monitor ASSIGN commands at run time. Default
assignments are defined during system generation.

2.1.2

Data Structures

Each peripheral device has an associated data structure which governs the manner in which data ore
stored. There are basically two modes in which data may be stored externally - serially ordirectoried.
For a sequential file, either structure may be used. If it is serial, the physical sequence of records is
identical to the logical sequence. If it is directoried, the logical sequence is established by pointers
which link one record to another although their physical locations need not be in sequence. For a
direct-access file, only directoried devices may be used.
Serial devices used for FORTRAN Input-Output include magnetic tape and DECtape. Records are
transmitted directly from the user buffer to the device and an end-of-file is written after the last
record by a CALL CLOSE or ENDFILE n. A file is accessed simply by virtue of device assignment.
DECtape may also be used in a directoried mode. In this case, a directory containing file information
is maintained. Each entry contains a filename and extension and a pointer to the first block of the
file. Files stored in this way may be referenced in the OTS directoried subroutine calls.
Directoried FORTRAN input-output to a disk, using DOS-15 file structure, is a special case. This
structure is based on a hierarchy of directories with a Master File Directory (MFD) pointing to user
file directories (UFDs). User files are created sequentially but may be accessed either sequentially
or directly. Data blocks (4008 words per block) which comprise a file are chained via a forward
link word (377 ) and backward link word (376 ), Forward links are also stored in a retrieval
8
8
information block (RIB) for direct access. Files stored in this mode are accessed by name. This name
may be assigned by the user via directoried subroutines (e.g., SEEK and ENTER). If this is not done,
default names are used. A default name has the form .TMOmn OTS where mn is the logical device
number.

*This number is the standard size for DOS-15 but may be changed by system generation and assembly
parameters.
2-2

2.1.3

Data Transmission

Data is transmitted to and from the FORTRAN-IV I/o buffer via the OTS FlOPS package. A single

I/o buffer of 4008 words is used. The size of the buffer which is to be transmitted for a particular
device is set in accordance with information provided in an .INIT to the device used.

2.2

OTS lOPS COMMUNICATION (FlOPS)

The FlOPS package provides the necessary communication between OTS and Input-Output Processor.
Its two main functions are device assignment and the transfer of data to and from the FORTRAN
i nterna I I/o bu Her.
FlOPS maintains a status table with one-word entries for each file that is opened. A table entry is
as shown below.

I/o Flag

For dir.
acc. only
l=DELETE
O=NO

O=SEQU.
1 = DIR. ACC.

O=READ
l=WRlTE

not
used

Buffer size
(from • I NIT)

The routines of the FlOPS package and their functions are given below.

FlOPS Package
External Calls: OTSER
Errors:

OTS ERROR 10 - illegal device number
Routine

.FC
(initialize I/O Device)
Call:
LAC DEVICE (address of slot number)
JMS* .FC

Function
.DAT slot numbers are initialized by .FC. The
first call to .FC for any device generates a
monitor .INIT call which opens the file for I/o
and enters the buffer size and I/o flag in the
device status table. Subsequent calls to .FC
ca II .INIT only if the I/o flag has been changed
or the file has been closed.

To set I/o flag:
DZM* .FH (input)
LAC (1) (output)
DAC* .FH

(continued next page)

2-3

FlOPS Package (Cont)

Function

Routine
.FQ
Call:

LAC (address of • DA T slot number (bits 9-17)
lOPS mode (bits 6-8)

Data are transferred between the I/O buffer and
an I/O device. .FQ checks the monitor I/O flag.
If it is zero, a • READ call is made; if it is one, a
• WRITE coli is mode. A call to •WAIT is made in
either case.

JMS* .FQ
Sets a" words in the device status table to zero •
Called at the beginning of a" FORTRAN main
programs to indicate that all devices are
initial ized.

• FP
Call:
JMS* .FP

Initializes END or ERR exits. The AC is saved
and restored to accomodate direct access. If one
of the two exit addresses is not to be specified, an
address of 0 should be passed •

.ZR
Call:
JMS* .ZR
• DSA END addr
• DSA ERR addr

-'-

JMS* .FF (.FG)(.RF)(.RG)

Direct and sequential access BCD and BINIO
terminate routines reinitialize OTSER.

An integer function - 10ERR (N) is available to the user and may be invoked at an ERR exit to
determine the I/O error which has occured. The value of IOERR will be one of the following:

Error

Value

-1

Parity error

-2

Checksum

-3

Shortline

-5

End-of-file

-6

End-of-medium

OTS error number

2.3

Other errors (up to 77)

SEQUENTIAL INPUT-OUTPUT

Sequential input-output operations access consecutive records of a file, beginning with the first
record and then record-by-record until the end of the file. A file which is accessed sequentially may

2-4

be stored serially (on magnetic tape or DECtape) or in directoried mode (on disk and DECtape). That
is, the physical sequence of records mayor may not conform to the logical sequence.

2.3.1

OTS Binary Coded Input/Output (BCDIO)

The formatted READ and WRITE statements generate calls to routines in the BCDIO package. Input
and output operations are performed on a characte~to-character basis under the control of a FORMAT
statement. All BCDIO routines use FlOPS to perform transfer of data. BCDIO routines may also be
called directly by MACRO programs.
Each formatted record is an lOPS ASCII line with a two-word header pair. The first character after
the header is always a forms-control character. Record length, given in the header, is always in
terms of word-pairs. The last character in the last word-pair is always a carriage return.
BCDIO routines are described below.

BCDIO Package
External Calls:

FlOPS, OTSER, REAL, RELNON or RELEAE

Errors:

OTS 10 - illegal I/O device number
OTS 11 - bad input data (lOPS mode incorrect)
OTS 12 - illegal format
Routine

• FR (. Fyv)

Function
Inputs (outputs) a data item.

Call:
JMS* • FR (. Fyv)
.DSA (address of • DAT slot number)
.DSA (address of first word of FORMAT
statement or array)*
.FE
Call:

Inputs or outputs a data item using format decoder
(.FD).

JMS* .FE
• DSA {address of data item (first word))
.FA
Call:

Inputs or outputs an entire array using format
decoder (.FD).

JMS* .FA
• DSA (address of last word in array descriptor
block)

(continued next page)
*This word is 0 for data-directed I/O

2-5

BCDIO Package (Cont)

Routine
.FD

Function
Decodes format into four parameters:

Call:

.D - decimal places
• W - field width
• SF - sco Ie factor
.S - mode

JMS* .FD

Terminates the current logical record.

.FF
Call:
JMS* .FF

As described in the language manual * , FORMAT statements may be entered or changed at run time,
at which point they are interpreted by BCDIO. In addition to providing the FORTRAN programmer
with greater flexibility, this feature permits the MACRO programmer to use the formatted I/o
capabilities of BCDIO. (See Appendix C for examples.)

2.3.2

OTS Binary Input/Output (BINIO)

The BINIO package processes unformatted READ and WRiTE statements. Data transfer is on a wordto -word basis. A logical record, the amount of data associated with a single READ or WRITE
statement, may consist of several physical records whose size (except for the last) is always the
standard lOPS I/O buffer size. Thus, when a WRITE statement is processed, each physical record
generated contains an ID word (word 3) in addition to the two required header words. This word
contains a record identification number. For the first record, this is zero. The last record is
indicated by setting bit 0 of the ID word to 1. Up to 3777778 physical records may be generated
for a single logical record.
For example, if four physical records are generated, the four ID words would be:
000000
000001
000002
400003
If only one record is generated, its ID word will be 400000 signifying the first and last of a set.

An unformatted READ statement accepts logical records of the form described above until its I/O list
has been satisfied. If this occurs in the middle of a logical record, the remainder of the record is
ignored. That is, the next READ will access the beginning of the next logical record.
*DEC-1S-GFWA-D
2-6

-------

The routines of BINIO are described below.

BINIO
External Calls:

FlOPS, OTSER

Errors:

OTS 10 - illegal I/O device number
OTS 11 - illegal input data (lOP mode)
Routine

Function
Initializes a device for binary input and reads
first record.

.FS
Call:
JMS* .FS
.DSA (address of .DAT slot)
.FX

Initializes a device for binary output; initializes
I ine buffer.

Call:
JMS* .FX
.DSA DEVICE

Transfers a data item to or from the line buffer
(a II modes). Mode of item indicated by bits
1 - 2 of argument are:

.FJ
Call:
JMS* .FJ
.DSA (address of item (first) word)

00 = INTEGER
01 = REAL
10= DOUBLE PRECISION
11 = DOUBLE INTEGER
Transfers an array.

.FB
Call:
JMS* .FB
• DSA (address of last word in array descriptor
block)

Terminates current logical record. For WRITE,
packs the line buffer with zeroes as required and
sets bit 0 of the ID word.

.FG
Call:
JMS* .FG
2.3.3

OTS Auxiliary Input/Output (AUXIO)

The AUXIO package processes the commands BACKSPACE t REWIND, and ENDFILE which have
different meanings for magnetic tape and disk. AUXIO routines issue .MTAPE monitor calls giving
• DAT slot and a code specifying the magnetic tape function desired:

2-7

Code

Magnetic Tape

Rewind to load point

Close file associated with .DAT slot.

Backspace record

Pointers resumed for previous ASCII or binary line.

Write end-of-file

N.A.

Disk

For magnetic tape, these operations require only calls to system macros. In order to simulate magnetic
tape functions on disk, a file active table (.FLTB) must be referenced. This contains four-word
entries for every positive .DAT slot indicating whether the file is active (open for input or output)
or inactive. The routines of AUXIO and their serial and file-oriented functions are given below.

AUXIO
External Calls:

FlOPS, .FLTB

Errors:

OTS 10 - illegal I/o device
OTS 11 - illegal input data (lOPS mode incorrect)

Routine

.FT
(BACKSPACE)
Call:

Magnetic Tape

Disk

Repositions device at a point just prior to the
first physical record associated with the current
logical record.

Resumes pointer to
previous ASCII or
binary line.

Repositions device at load point.

Closes file. Ifno
file is open,
nothing is done.

Closes file. Writes an end-of-file mark on tape.

Closes file, zeroes
words 0-3 of the
associated • FLTB
entry •

JMS* .FT
• DSA (address of
.DAT slot)
.FU
(REWIND)
Call:
JMS* .FU
• DSA (address of
.DAT slot)
.FV
(ENDFILE)
Call :
JMS* .FV
.DSA DEVICE

On a REWIND to disk, the filename is saved; thus, subsequent sequential input-output operations will
open that file. On an ENDFILE, the filename is lost and subsequent operations will open a default file.

2-8

2.4

DIRECT ACCESS I/O

Direct access input-output files are referenced by name; records are retrieved or Clccessed by number.
The OTS routines which perform direct-access transmission of data are similar to their sequential
counterparts. Before they are invoked, however, the user must provide a detailed description of
his file.

2.4.1

The DEFINE Routine

The FORTRAN user establishes a direct-access file by calling the DEFINE routine which was described
in Part I, Chapter 6. The meanings of its arguments are iterated below for the call:
CALL DEFINE (D, S, N, F, V, M, A, L)
The parameters provided to OTS for performing direct-access functions are:
D - .DAT slot
S - record size
number of ASCII characters
or
number of binary words
N - number of records (S3m77S)
F - array reference to file name and extension - if 0, default name
V - associated variable - set to number of the last accessed record plus one
M - mode -0 = lOPS binary
non-O = lOPS ASCII
A - file size adjustment indicator
= no adjustment
non-O = adjust

L - deletion indicator
o = no deletion
non-O = delete temporary file
The DEFINE routine initializes a file for direct-access in one of four ways, depending on the
combination of parameters suppl ied.
a.

Simple Initialization - If F specifies a file which already exists and no adjustment has
been indicated, DEFINE opens the file for direct access. The mode and record length
parameters must conform to the file's characteristics. The associated variable is set to 1.
The number of records N must be less than or equal to the actual number of records.

Named File Creation - If F specifies a file which does not exist on .DAT slot D, a file
is created according to the characteristics given in the calling arguments. If the mode
is ASCII, the data portion is filled with spaces (040S). If the mode is binary, all data
words are set to 0 and the ID word for each record to 400000 •
8

2-9

Default-Named File Creation - If F=O in the DEFINE call, a file is created as above but
given a default name of the form. TMOab OTS (unless a file of that name already exists
on • OAT slot D) where ab specifies. OAT slot. If L=l , a bit is set in the FlOPS status
table signifying that the file is to be deleted after an ENDFILE or CALL CLOSE to the
• OAT slot.

File Size Adjustment - If a file F exists and A is not zero, N is used to adjust the number
of records in the file. This is done by creating a temporary file ( •• TEMP OTS) on .OAT
slot 0 via .DAT slot -1 which is temporarily loaded with the • OAT slot 0 handler address
and UIC. The file is copied into it one record at a time up to the number N. If the file
is to be lengthened, null records are added. The adjusted file is then assigned a name
according to F. V is set to 1 if the file is reduced. If it is lengthened, it is set to the
old length plus one.

The algorithm used for determining the function of DEFINE from its ar~ments is illustrated in the
following flowchart.

YES, OEFAULT NAM E IS
INDICATED

BUILD DEFAULT NAME
FROM .DAT SLOT
(.TMOab OTS)

INITIALIZE FILE FOR
RANDOM ACCESS
V IA .RAND

AOJUST SIZE
OF FILE

2-10

From user-supplied arguments, the DEFINE routine establishes a parameter table (PRMTB) which is
available to direct-access input-output routines.
Each device which has a file open for direct-access will have an active four-word entry composed as
follows:

Information

Bits

Word
0

File active bit (l if active - always set for ASCII files)

2-11

Number of blocks per record

12-17

.DAT slot number

mode - 0 if binary; 1 if ASCII

5-11

Word pairs per record

12-17

Records per block (0 for binary records larger than one physical block)

1-17

Records/fi Ie

3-17

Address of associated variable

• PRMTB will generally have four such entries but this number may be varied with an assembly
parameter.
DEFINE also initializes the file in FlOPS, setting the appropriate bits in the FlOPS status table.

2.4.2

Formatted Input/Output (RBCDIO)

Direct-access operations may be performed on any formatted data file conforming to DOS-15 file
structure and with a fixed record length. A direct-access WRITE will output formatted records which
have the same form as with sequential operations. The distinction is that the direct-access records are
transmitted into a series of records which already exist on the selected file. A single READ or WRITE
will access records on the I/o device only as specified in the associated FORMAT statement. This
means that a long I/O list will not cause a new record to be accessed, regardless of the length of
the list, unless this access is indicated by the FORMAT statement. A carriage return is, as with
sequential I/O, appended to each ASCII line. Any information from a previous WRITE mode to a record
which remains after the carriage return, is inaccessible.

The FlOPS buffer and tables are used as

with sequential I/O. Data transfer, however, is performed using the. RTRAN system MACRO.
The RBCDIO routines described below correspond to the sequential I/O routines of BCDIO. Control
is transferred to BCDIO for data transmission via the global entry points given.

2-11

RBCDIO
External Calls:

FlOPS, BCDIO (.FE, .FA), OTSER, RANCOM

Errors:

None
Purpose

Routine
• RW (.RR)
Call:
JMS* .RW (RF)
.DSA (address of • DAT slot)
.DSA (address FORMAT)
(AC holds integer record number)

• RF

Terminates current logical record. Sets last record
flag, reinitializes .ER in OTSER and, for WRITE,
• RTRAN out last record.

Call:
JMS*

BCD direct-access WRITE (READ) sets the directaccess flag; sets mode switch to ASCll; initial izes
direct-access READ/WRITE (.INRRW in RANCOM);
checks mode of existing record; initializes - .STEOR
and BFLOC in BCDIO for direct-access , line buffer ,
and form at decoder; sets. HI LIM in BCOIO • • RW
loads record number into .RCDNM and sets I/o
flag in FlOPS to write. • RR loads record number
into .RCDNM, sets I/o flag to read •

.RF

Entry points to BCDIO are:

2.4.3

RBCDIO Entry

BCOIO Routines

.RE

.FE

.RA

.FA

Unformatted Input/Output (RBINIO)

Unformatted direct-access I/O differs from formatted in two respects. If a binary record does not
totally fill the record into which it is written, the previous contents are still accessible. If a directaccess WRITE requires more words than exist in each record, successive records are accessed and
written until the I/O list is exhausted. Records are linked by ID words as for sequential files.
The routines of RBINIO are described below. Direct-access entry points to BINIO follow.
RBINIO
External Calls:

FlOPS, RANCOM, BINIO

Errors:

None
Function

Routine
• RS (. RX)
Call:
JMS* • RS (. RX)
.DSA (address of .DAT slot)
(AC holds integer record number)

Binary direct -access WRITE (READ) sets directaccess flag; sets mode switch to binary; initializes
direct READ/WRITE (.INRRW in RANCOM); checks
mode of existing record; initializes. BUFLC, • RDTV,
and. WRTV in BINIO for direct access; initializes
I/o buffer; loads record number into .RCDNM.
• RX sets I/o flag to WRITE; • RS sets it to READ.

(continued next page)

2-12

RBINIO (Cont)

Routine

Function
Terminates current logical record. Increments
associated variable, reinitializes .ER in OTSER; if
WRITE, sets lost record flog and outputs final records.

.RG
Call:
JMS* .FG

2.4.4

Initialization and Actual Data Transfer (RANCOM)

RANCOM contains two major routines which are used by both RBCDIO and RBINIO. These routines
perform initialization and data transfer functions which are identical to those performed for ASCII and
Binary I/O.

RANCOM
External Calls:

FlOPS, OTSER, DEFINE

Errors:

OTS 10 - illegal I/o device
ors 24 - illegal record number
ors 25 - mode discrepancy
ors 11 - illegal input data (lOPS mode incorrect)
OTS 21 - undefined file
OTS 23 - size discrepancy
Function

Routine
.INRRW

Initializes a direct access READ or WRITE

Call:
JMS* .INRRW
(AC holds address of slot number.)
For I/o cleanup:
Set up header pair and. RTRAN out block of data.

.RIO
Call:
JMS* .RIO

2.5

For end -of - record rout i nes:
Output (if WRITE) and set pointers to new record.

Data-Directed Input-Output (0010)

The Data-Directed Input-Output package permits input or output of ASCII data without reference to a
FORMAT statement. On input, 0010 extracts individual data fields by scanning the line buffer for
terminators. It then determines the mode of the variable to which the item is to be transferred and
converts the item to that mode if necessary. Unlike the format decoder, 0010 does not reject an item
which is too large but simply assigns the maximum value which the variable can accomodate. On output, 0010 has a set of default format parameters for each type of variable.

2-13

The same buffer is used for both data-directed and formatted I/o, and the I/o action for both takes
place between device and I/O list variables or vice versa in both cases. Thus, 0010 uses the same I/o
initialization and termination routines as regular formatted I/o (found within BCDIO for sequential
access and within RBCDIO for direct access). 0010 control routines are, however, unique due to the
special features described above.
The routines of 0010 are given below.

0010
External Calls:

BCDIO, .55, OTSER, FlOPS, REAL, DBLINT

Errors:

OTS 42 - bad input data*
Function

Routine
• GA
Call:
JMS* • GA / radix 50

name
name 2

Outputs a data item in the 'NAME' = value form •
Mode is obtained from bits 1-2 of the pointer word;
if the mode is 0 (integer-logical), bit 0 of the name
word indicates which (0 for integer, 1 for logical).

first characters
lost 3 characters

• DSA address item
.GC
Call:
JMS* • GC / radix 50
name 1
name 2
.DSA item
• GB

Outputs an array element in 'NAME (I)' = value
form. Also uses bits 1-2 for mode •• GC should
only be used when. SS has been used to calculate
the subscript address.

Outputs an entire array in 'NAME(J)' = value form •

Call:
JMS* • GB / radix 50
name 1
name 2
.DSA array description block
(word #4 address)
• GD

Inputs an item. Mode is in bits 1-2 of argument •

Call:
JMS* .GD
.DSA item
• GE

Inputs an array. Mode is in bits 1-2 of argument •

Call:
JMS* .GE
• DSA addr. of array discriptor block word 4
*For Teletype input - 'BAD INPUT DATA - RETYPE FROM INPUT WITH ERROR' is typed.

2-14

2.6

ENCODE/DECODE (EDCODE)

Encode and Decode perform memory-to-memory transfers and conversions using the apparatus established
for formatted input-output. That is, data is transferred from memory to the I/O buffer to memory. Since
no peripheral device is involved, the initial ization and termination mechanisms of EDCODE are unique
while the data transfer is the same as for BCDIO.
The routines of EDCODE are given below.

EDCODE
External Calls:

OTSER, BCDIO

Errors:

OTS 40 - illegal number of characters
OTS 41 - array exceeded
Routine

Function
Encode •

• GF
Call:
JMS* .GF
• DSA number of characters
.DSA array
.DSA format

Decode •

• GG
Call:
JMS* .GG
• DSA number of characters
.DSA array
.DSA format

2.7

USER SUBROUTINES

The subroutines given below are FORTRAN-callable subroutines which support input-output operations.

· 2.7.1

Magnetic Tape Input-Output Routines*
Routine

Call

EOF

CALL EOF{d,@n , @n )
2
1
Where:

Function

d = • DAT slot (must be
assigned to tape)

Control is passed to n if
1
EOF was encountered on last
input operation; otherwise to n

n ,n = statement numbers
1 2

*Not supported with RSX. END, ERR exits can be used in place of EOF.
2-15

(continued next page)

Routine

Call

rOCHECK

CALL IOCHECK (d,@n1 ,@n )
2

Same

UNIT

CALL UNIT (d,@n ,@n ,@n ,
3
2
@n ) 1
4

Control is passed to:

Function

n - device not ready
1
n - devi~e ready, no
2
prevIous error
n3 - EOF sense.d
n - parity or lost data
4
error

2.7.2

Directoried Subroutines

The directoried subroutines described below comprise a package named FILE. These routines interact
with the DOS-J5 file-oriented data structure and with DECtape file structure.

FILE
External Cal Is:

FlOPS, .DA

Errors:

OTS 10 - illegal device number
OTS 13 - file not found (SEEK)
OTS 14 - directory full (ENTER)
Call

Routine
SEEK

Purpose

CALL SEEK (n,A)

Finds and opens a named input file.

Where:
n = device number
A = name of array containing the
9-character 5/7 AS CII file
name and extension
ENTER

CALL ENTER (n,A)

Creates and opens a named output file.

CALL CLOSE (n)

Terminates an input or output file
(reqJired when SEEK or ENTER are
used).

FSTAT

CALL FSTAT (n,A,I)

Searches for named fi Ie.

Where:
I = 0 if the file not found;
= 1 if found and action complete

(continued next page)

2-16

RENAM

Purpose

Call

Routine

CALL RENAM (n,A,B,I)

Searches for named file and renames it.

Where:
A is on array containing existing name
B is an array containg a new
file nome
I = 0 if file not found; 1 if
found and action complete
DLETE

Searches for named file and deletes it.

CALL DLETE (n,A,I)
Where:
A is an array containing existi ng fil e name
I = 0 if file not found; 1 if
found and action complete

2-17

CHAPTER 3
THE SCIENCE LIBRARY

The FORTRAN Science Library is a set of pre-defined subprograms which may be invoked by a
FORTRAN-IV subprogram reference. These include intrinsic functions, external functions, the
arithmetic-pockage functions, and external subroutines. Each of these may also be referenced by a
MACRO program as may the sub-functions and OTS routines which are also part of the FORTRAN
library •
Descriptions of each type of subprogram are given in the following subsections. Information given for
these include errors, accuracy, size, and external calls (to other library subprograms). Each function
description also includes the MACRO calling sequence. Where there are two arguments, it is assumed
that the appropriate accumulator has been loaded (accumulators are described in Section 3.4). For
calling sequences which use the .DSA pseudo-operation to define the symbolic address of arguments,

400000 must be added to the address field for indirect addressing.
FORTRAN library subprograms are called by FORTRAN programs in the manner described in the
Language Manual (DEC-1S-GFWA-D). Subprograms called by MACRO programs must be declared
with a .GLOBL pseudo-operation as in:
Examples:
Floating Point (FPP) System

Standard System

• TITLE
.GLOBL SIN
FST = 713640

•TITLE
.GLOBL SIN, .AH

JMS* SIN
JMP .+2
.DSA A
JMS* .AH
.DSA X

/JUMP beyond argument
/+400000 if indirect
/store in real format at

/X
X

JMS SIN
JMP .+2
.DSA A
FST
.DSA X
.DSA 0
.DSA 0

X .DSA 0
.DSA 0
The number and type of arguments in the MACRO program must agree with those defined for the subprogram.

3 -1

3.1 INTRINSIC FU NCTIONS
Table 3-1 contains a description of each of the intrinsic functions in the FORTRAN library.
An intrinsic function's type and arguments cannot be changed. It is referenced via an Arithmetic
statement, as in:

x = ABS (A)
(Table 3-1 appears on the following page.)

3-2

Table 3-1
Intrinsic Functions
Definition

Function

Symbclic
Nome

Mode
1=1**1

.BB

Calling Sequence

Errors

ARG1 IN FLT .ACC
JMS* .BB
· DSA ADDR of ARG2

15 if base = 0
and expo 0

Accuracy
(Bits)

External Calls

N.A.

INTEGER

N.A.

REAL

None

N.A.

REAL

13 if base < 0
13 if base ~ 0
14 if base < 0
14 if base ~O

26
26
32
32

.EE,.DF, REAL
.EE,.DF, DOUBLE
.DE,.DF, DOUBLE
.DE,.DF, DOUBLE

.BC
. BC
.BL

R**I(or J)
R=R**I
R=R**J

ARG 1 IN FL T. ACC
{JMS* SUBR
• DSA ADDR of ARG2

.BD
.BD
.BM

D**I(or J)
D-D**I
D=D**J

ARG1 IN FLT. ACC
(JMS* SUBR
• DSA ADDR of ARG2

.BE
.BF
.BG
.BH

R=R**R
D=R* *D
D=D**R
D=D**D

CMS* SUBR
• DSA ADDR of ARG2

.BI
.BI
.BJ
.BK

1**J l J** J(or I)
1- 1** J
J=J**J
J=J**I

ARG1 IN AC (and MQ) None
CMS* SUBR
• DSA ADDR of ARG2

N.A.

DBL!NT

ABS
lABS
JABS
DABS

R=ABS(R)
I=IABS (I)
DI=JABS(DI)
DP=DABS(DP)

None

N.A.

.DA,REAL
.DA
.DA, DBL!NT
.DA, DOUBLE

AINT
INT
!DINT
JINT
JDINT

R=AINT(R)
I=INT(R)
I=!DINT(DP)
DI=JINT(R)
DI=JDINT(DP)

None

N.A.

.DA, REAL
.DA, REAL
.DA, REAL, DOUBLE
.DA, DOUBLE, DBL!NT
• DA, DO UBL EfDBL! NT

None

ARG1 IN FLT. ACC
W
I
W

Absolute
Value

IARGI

Truncation

Sign of ARG times largest
integer ~ I ARG I

*15 if base =Oand exp ~ O.
---

JMS* SUBR
JMP .+2
• DSA ADDR of ARG

~MS'
SURR
JMP .+2
.DSA ADDR of ARG

Table 3-1 (Cont)
Intrinsic Functions
Definition

Function
Transfer of
Sign

Sign of ARG2

Sign of ARGI
Positive
Difference

Conversion

ARG1-MIN(ARG1,ARG2)

VMODE -ARG

W
I

.j>.

Remaindering

ARG1-[ARG1/ARG2]ARG2
Where: [A 1/A2] is an integer whose magnitude does
not exceed the mag ni tude
of A1/A2 and whose sign is
the same

)

Symbolic
Name

Mlde

SIGN
ISIGN
DSIGN
JSIGN

R=SIGN(R,R)
I=ISIGN(I,I)
DP=DSIG N(DP, DP)
DI=JSIGN(DI,DI)

DIM
!DIM
JDIM

R=DIM(R,R)
I=!DIM(I,I)
DI=JDIM(DI,DI)

FLOAT
IFIX
SNGL
DBlE
JFIX
ISNGl
!DBlE
JDFIX
FlOATJ
DBlEJ

R=FLOAT(I)
I=IFIX(R)
R=SNGl(D)
D=DBlE(R)
DI=JFIX(R)
or JFIX(DP)
I=ISNGl(DI)
DI= JDBlE(I)
DI=JDFIX(DP)
R=FlOAT J(DI)
DP=DBlEJ(Di)

AMOD
MOD
DMOD
JMOD

R=AMOD(R, R)
I=MOD(I,I)
DP=DMOD(DP, DP)
DI=JMOD(DI,DI)

Calling Sequence

('MJMPS>.+3SU",
('MJMPS>SU",
.+3

N.A.

.DA, REAL
.DA
.DA, DOUBLE
.DA,DBUNT

None

N.A.

.DA, REAL
.DA, INTEGER
.DA,DBUNT

None

N.A.

.DA, REAL
.DA, REAL
.DA, DOUBLE
.DA, REAL
.DA, DOUBLE, DBUNT
.DA,
.DA, DBUNT
.DA
.DA, DOUBLE, DBUNT
.DA, DBUNT
.DA,DBLINT

None

N.A.

.DA, REAL
.DA, INTEGER
.DA, DOUBLE
.DA,DBUNT

• DSA ADDR of ARG 1
.DSA ADDR of ARG2

JMS* SUBR
JMP .+2
.DSA ADDR of ARG

fMS>
SUB'
JMP .+3

External Calls

None

• DSA ADDR of ARG 1
• DSA AD DR of ARG2

.DSA ADDR of ARGI
.DSA ADDR of ARG2

-----~-

)

Accuracy
(Bits)

Errors

)

Table 3-1 (Cont)
Intrinsi c Functions

W
I

Function

Definition

Maximum/
minimum value

VAR = max or min value of
arglist

Symbolic
Name
Integer
min/ max
(IMNMX)
MAXO
MINO
AMAXO
AMINO
Real
min/max
(RMNMX)
AMAXI
AMINI
MAXI
MINI

Calling Sequence

Mode

Errors
None

Accuracy
(Bits)
N.A.

External Calls
INTEGER, REAL

I=MAXO(Il'·· .In)
I=MINO(Il'·· .In)
R=AMAXO(I I, ... In)
R=AMINO(Il'·· .I n)
INTEGER, REAL

JMS*SUBR
JMP .+n+l
.DSA ADDR of ARGI

R=AMAXI (Rl, ••• Rn)
R=AMINI (Rl,··· Rn)
I=MAXl(Rl,·· . Rn)
I=MINl(R 1 ,·· • Rn)

.DSA ADDR of ARGn
Doubleprecision
(DMNMX)
DMAXI
DP:DMAXI (DP 1,·· . DP
DP-DMINl( DP 1'·· .DP n

DOUBLE

Double
integer
(JMNMX)
JMAXO
JMINO

DBUNT

DI=JMAXO( D1 l'·· .Dl n)
DI=JMINO(DI 1 ,·· .Dl n)
I

3.2

EXTERNAL FUNCTIONS

Table 3-2 describes the external functions of the FORTRAN library. An external function is a subprogram which is executed whenever a reference to it appears within a FORTRAN expression and which
returns a single value.
A description of the algorithm applied in implementing each of these functions is given below.

3.2.1

Square Root (SQRT, DSQRT)

A first-guess approximation of the square root of the argument is obtained as follows:
If the exponent (EXP) of the argument is odd:

Po = .5

(EXP-1)
(EXP-1 )
2
+ARG
2

If EXP is even:
EXP
EXP
Po = .5 ( -2-) +ARG ( 2- 1 )
Newton's iterative approximation, below, is then applied four times.

Pi+ 1 = ~ (Pi + A~~ )
I

3.2.2

Exponential (EXP, DEXP)

The following description also applies to the sub-functions. EF and. DF.
The function eX is calculated as 2xl092E (xlog E will have an integer portion (1) and fractional portion
2

(F».
Then:

Where:

n = 6 for EXP and .EF
n = 8 for DEXP and .DF

(continued page 3-7)

3-6

Table 3-2
External Functions
Mode

Calling Sequence

Errors

ARGl/2

SQRT
DSQRT

R=SQRT(R)
OP=DSQRT(DP)

JMS*SUBR
JMP .+2
• DSA ADDR of ARG

5 if ARG < 0
6 if ARG <0

• OA,. ER,REAL
• DA, .ER,DOUBLE

Definition

Square
root

External Calls

Exponential

ARG

EXP
OEXP

R=EXP(R)
DP=DEXP(DP)

Same

13ifARG<0
14 if ARG <0

26
34

.DA, .EF, .ER,REAL
• OA, .OF,. ER,DOUBLE

Natural
logarithm

Log ARG
e

ALOG
OLOG

R=ALOG(R)
OP=DlOG(OP)

Same

26
32

.OA, .EE,.ER,REAL
• OA, • DE,. ER, DOUBLE

Common
logarithm

Log

ALOGlO
OLOGI0

R=ALOG 10(R)
OP=OlOG lO(OP)

Same

None

26
34

• OA,. EB ,REAL
• OA, • DB, OCXJBLE

ARG

Sine

Sin(ARG)

SIN
OSIN

R=SIN(R)
OP=OSIN(OP)

Same

Cosine

cos(ARG)

COS
OCOS

R=COS(R)
OP=OCOS(OP}

Same

None

26
34

• OA,. EB ,REAL
• DA, .OB,OOUBLE

Arc
tangent

tan

ATAN
DATAN

R=ATAN(R)
DP=DATAN(DP)

Same

None

26
34

• DA, • ED I REAL
• DA, • DD, DOUBLE

-1
tan
(ARGl/
ARG2)

ATAN2
DATAN2

R=ATAN2(R, R)
OP=DATAN2
(DP ,DP)

JMS*SUBR
JMP .+3
None
.OSA ADDR of ARG 1
• DSA ADDR of ARG2

26
34

Same

tanh(ARG)

TANH

R=TANH(R)

JMS*TANH
JMP .+2
• DSA ADDR of ARG

.OA, .EF ,REAL

W
I

'..I

Accuracy
(Bits)

Symbolic
Name

Function

Arc
tangent
(X/y)
Hyperbolic
tangent

-1

(ARG)

- ---

None

- - - - -- -- ---

The values of C. are given below.
I

Value of i

Value of C.
I

1.0

0.34657359

0.06005663

0.00693801

0.00060113

0.00004167

0.00000241

0.00000119

0.00000518
~,

3.2.3

Natural and Common Logarithms (ALOG, ALOG10, DLOG, DLOG10)

The exponent of the argument is saved as,the integral portion of the result plus one. The fractional
portion of the argument is considered to be a number between 1 and 2. Z is computed as follows:

Z = x- fi

X+v'2

Then:

'vVhere:
n = 2 (ALOG)
n = 3 (DLOG)
The values of C are given below:
DLOG and DLOG 10

ALOG and ALOG 10

= 2.8853900

C = 2.8853913
1

C = 0.96147063
3

C = 0.96180076
3

C = 0.59897865
5

C = 0.57658434
5
C

= 0.43425975
(continued next page)

3-8

The final computation is:

3.2.4

ALOG and DLOG:

log X = (log X) (log 2)
e
e
2

ALOG10 and DLOG10:

10910 X = (log X) (log 2)
2
10

Sine and Cosine (SIN, COS, DSIN, DCOS)

This description also applies to the sub-functions. EB and. DB.
The argument is multiplied by 2/1£ for conversion to quarter-circles. The two low-order bits of the
integral portion determine the quadrant of the argument and produce a modified value of the fractional
portion (Z) as follows.
Quadrant

Low-Order Bits

Modified Value (Z)
F

00
01

1-F

III

-F

-(1-F)

The value of Z is then applied to the polynomial expression:
"

s,"X=(~ C2"+1Z

i=O

2~1

)

n = 4 for SIN, COS, .EB
n = 6 for DSIN, DCOS, .DB
The values of C are as follows:
DSIN, DCOS, .DB

SIN, COS, .EB
C
C

1
3

= 1.570796318

C = 1.5707932680
1

= -0.645963711

= -0.6459640975

C = 0.079689677928
5

C = 0.06969262601
5

C = -0.00467376557
7

C = -0.004681752998
7

C = 0.00015148419
9

C = 0.00016043839964
9
C
C

11
13

= -0.000003595184353
= 0.000000054465285

(continued next page)

3-9

The argument for COS and DCOS is adjusted by adding 1T/2. The sin subfunction is then used to
compute the cosine according to the following relationship:
COs X =sin ~+X)

3.2.5 Arctangent (ATAN, DATAN, ATAN2, DATAN2)
The following description also applies to the sub-functions. ED and. DD.
For arguments less than or equal to 1, Z = arg and:
n
2i+1
arctangent arg = (i~O C2i+ 1Z
)
n = 7 for ATAN and ATAN2
n = 3 for DATAN and DATAN2
For arguments greater than 1, Z = l/arg and:
1T (~
2i+l
arctangent arg ="'2 - i=O C + Z
)
2i 1
n = 8 for ATAN and ATAN2
n = 3 for DATAN and DATAN2
The values of C are given below.
ATAN and ATAN2

DATAN and DATAN2

C 1 = 0.9992150

C 1 = 0.9999993329

C = -0.3211819
3

C = -0.3332985605

C = 0.1462766
5

C = 0.1994653599
5

C = -0.0389929
7

= -0.1390853351

C = 0.0964200441

Cll = -0.0559098861
C
C

3.2.6

13
15

= 0.0218612288
= -0.0040540580

Hyperbolic Tangent

The hyperbolic tangent function is defined as:
tanh

IX I = (1-

2
2 IXI )
l+e

eX is calculated as j<log2e (x1092e will have an integral portion (I) and a fractional portion (F)).
(continued next page)
3-10

Then:

Where:

n=6

The values of C. are:
I

Value of i

Value of C.
I

3.3

1.0

0.34657359

0.06005663

0.00693801

0.00060113

0.00004167

0.00000241

SUB-FUNCTIONS

Table 3-3 describes the sub-functions which are included in the FORTRAN library. These functions
are referenced by intrinsic and external functions but are not directly accessible to the user via
FORTRAN. The sub-function. EB, for example, performs the computation of sine and is invoked by
the external function SIN. MACRO programs may reference sub-functions directly. Algorithms for
all sub-functions which have counterparts among external functions were given in the previous subsection. This leaves the two general sub-functions Logarithm, base 2 and polynomial evaluator. Their
algorithms are given below.
3.3.1

Logarithm, Base 2 (. EE, • DE)

The exponent of the argument is saved as the integer portion of the result plus one. The fractional
portion of the argument is considered to be a number between 1 and 2. Z is computed as follows:

x-1"2
X+f2

(continued page 3-14)

3-11

Table 3-3
Sub-Functions
Definition

Function

W
I

I'-.>

Symbolic
Name

Mode

Calling Sequence

Errors

Accuracy
(Bits)

External
Calls

Sine
Computation

Sin (ARG)

.EB
.DB

R=. EB(R)
DP=.DB(DP)

JMS*SUBR
At entry floating
accumulator contains ARG;
at return contains result

None

19
28

.EC,REAL
.DC,DOUBLE

Arc tangent
Computation

tan -1 (ARG)

.ED

R=. ED(R)
DP=.DB(DP)

Same

None

26
34

Same

.DD

Logarithm
(base 2)
Computation

lo92 ARG

.EE
.DE

R=.EE(R)
DP=.DE(DP)

Same

13, ARG <0
14,ARG<0

26
32

.ER,REAL
.ER,DOUBLE

ARG

• EF
.DF

R=. EF(R)
DP=DF(DP)

Same

None

26
34

REAL
DOUBLE

JMS*SUBR
CAL PLIST

None

N.A.

REAL
DOUBLE

Exponential
Computation

Polynomial
Evaluation

VAR=

i C Z2i+l
1 =0 2i+l

.EC
.DC

VAR=

i=O 2i+l

R=.EC(R ,R ,···R )
2 1
n
DP=.DC(DP ,DP ,
2
1
••• DP )
n

·
·
·
PLIST-N/ - number of terms

I
I

+1
C / last term
n
C -l/next to last
n

·
·

Z2i+l

C /2nd term
1
C /1st term
g
--

(continued next page)

)

Table 3-3 (Cont)
Sub-Functions
Function

General Get
Argument

Definition

N.A

Symbolic
Name

Mode

.DA

N.A

Calling Sequence

Calling Routine

Errors

None

Accuracy
(B its)

External

N.A

None

SUBR CAL 0
JMS* .DA
JMP .+n+l
(address of ARG 1)
(address of ARG2)

·
·
·
(address of ARGn)
Is Called By
W
I

JMS*SUBR
JMP .+n+l
.DSA ARGl
.DSA ARG2

.....
W

----_.-

Then:
1
n
2i+1
log2 X =2+\~ C2i + 1Z
)
n = 2 (.EE)

n=3(.DE)
The values of Care:

3.3.2

.EE

.DE

C = 2.8853913
1
C = 0.96147063
3
C5 = 0.59897865

C = 2.8853900
1
C = 0.96180076
3
C = 0.57658434
5
C = 0.43425975
7

Polynominal Evaluator (.EC, .DC)

A polynomial is evaluated as:

3.4

THE ARITHMETIC PACKAGE

The arithmetic package contains the OTS arithmetic routines which are invoked by FORTRAN arithmetic expressions. These routines may also be called directly by MACRO progrcms. Versions of
FORTRAN-IV designed for use with the Floating Point Processor (FPP) require only single integer
arithmetic routines. Double (extended) integer arithmetic will be handled by the hardware.
The three major routines of the arithmetic package are INTEAE, RElEAE, and DOUBLE. INTEAE
contains integer arithmetic routines; RELEAE, real and floating arithmetic; and DOUBLE, doubleprecision arithmetic.
A description of these routines is given in Table 3-4. In the "calling sequence" column, reference
is made to three accumulators - the A-register, the floating accumulator, and the held accumulator.
The A-register is the standard PDP-15 hardware accumulator. The floating and held accumulators are
software accumulators which are part of the RELEAE package. The held accumulator is used as temporary storcge by some routines. Both consist of three consecutive PDP-15 words and have the format
shown below. (Negative mantissae are indicated by a change of sign.)

3-14

Held AC Labels

Floating AC Labels

CE01

.AA

Exponent (2's complement)

.AB

CE02

o
CE03

17
Sign of
mantissa

.AC

High-order
mantissa

17
Low order mantissa
17

The format shown above is that used for double-precision numbers. Single-precision numbers must be
converted before and after use in the floating accumulator to the single-precision format:
Low-order
mantissa
0

Exponent
(2's complement)
89

17
High-order
mantissa

Sign of
mantissa

RELEAE routines check for underflow and overflow and set a flag (.OVUDF) in the REAL store routine
. AH as follows:
Flag

Meaning

Action

non-O positive value

overflow - an attempt to store
a REAL constant whose binary
exponent is greater than 3778

± largest representable real
value stored (DOS-15);

negative value

underflow - an attempt to store
a REAL constant whose binary
exponent is less than -400

zero is stored

default value

va lue is stored

zero

The user may test this flag under program control using the logical function IFLOW. Recoverable OTS
messages are also given (see Appendix B, Section B.2).
Division by zero is also checked and a flag .DZERO set to zero (default value is 777777) in the
general floating divide routine (.CI). The result of the division is ± the largest representable value.
An OTS error message is also given for this condition. The user may test. DZERO under program
control using the logical function IDZERO.

3-15

The flags .OVUDF and. DZERO can only be initialized by reloading the program, by a separate
user program, or by IFlOW or IDZERO. These functions are described below.

Routine

IFLOW

Purpose

Checks underflow and overflow

Call

IORlV = IFlOW(I)

External Calls

.DA

Errors

None

The argument 1 indicates the check to be performed and values are returned as follows:

Action

Value

no check

O(.FAlSE) flag unchanged

underflow check

-l(.TRUE) if underflow - flag set to 0;
else 0 (.FAlSE) and flag unchanged

overflow check

-1 (. TRU E) if overflow - flag set to zero;
else 0 (.FAlSE)

Routine

IDZERO

Purpose

Checks for division by zero

Call

IORl V = IDZERO (I)

Externa I Co II s

.DA

Errors

None

If I=O, no check is made, IORlV = O(.FAlSE) and the flag is unchanged. If I/O, a check is made.

If an attempt at division by zero was made, IORlV = -1 (. TRUE) and the flag is reinitialized. Otherwise the flag is unchanged and IORlV = O(.FALSE).

3-16

))

Table 3-4
Arithmetic Package*
Function

Definition

Symbolic
Name

Mode

Integer
Arithmetic

INTEAE

):
W
I

.......

*Multipli
cation
*Division
* Reverse
division
*Subtraction
* Reverse
subtraction

DOUBLE

ARGl
A-Register

None

.......

ARG2
-

ARG1*ARG2

.AD

1=1*1

mu Itipl icand

multiplier

ARGl/ARG2
ARG2/ARGl

.AE
.AF

1=1/1
1=1/1

dividend
divisor

divisor
dividend

ARG1-ARG2
ARG2-ARGl

.AY
.AZ

1=1-1
1=1-1

minuend
subtrahend

subtrahend
minuend

I
I
,

JMS*SUBR
LAC ARG2

REAL

DoublePrecision
Arithmetic
Load
Store
Add
Subtract
Reverse
subtract
Multiply
Divide
Reverse
divide

External
Ca lis

Ca II ing Sequence

ARGl
FL.AC

ARG2
-

N.A
N.A
ARG1+ARG2
ARG1-ARG2
ARG2-ARGl

.AO
.AP
.AQ
.AR
.AU

DP=.AO(DP)
DP=.AP(DP)
DP=DP+DP
DP=DP-DP
DP=DP-DP

value
augend
minuend
subtrahend

address
address
addend
subtrahend
minuend

ARG1*ARG2
ARGl/ARG2
ARG2/ARGl

.AS
.AT
.AV

DP=DP*DP
DP=DP/DP
DP=DP/DP

mu Itipl icand
dividend
divisor

multiplier
divisor
dividend

"'I

> JMS*SUBR
.DSA ARG2

.......

*FPP versions require only Integer Arithmetic (INTEGE).

(continued next page)

Table 3-4 (Cont)
Arithmetic Package
Function
Real Arithmetic (ineludes floating)
load
Store
Add
Subtract
Reverse
subtract
Multiply
Divide
Reverse
divide
RElEAE
W
I

.......

Symbolic
Name

Definition

External
Calls

Calling Sequence

Mode

.....

ARGl
Fl.AC

ARG2

N.A
N.A
ARG1+ARG2
ARGI-ARG2
ARG2-ARGl

.AG
.AH
.AI
.AJ
.AM

R=.AG{R)
R=. AH{R)
R=R+R
R=R-R
R=R-R

value
augend
minuend
subtrahend

address
address
addend
subtrahend
minuend

ARG1*ARG2
ARG1/ARG2
ARG2/ARGI

.AK
.AL
.AN

R=R*R
R=R/R
R=R/R

multiplicand
dividend
divisor

multiplier
divisor
dividend

JMS*SUBR
.DSA ARG2

Floating
Arithmetic

Float
Fix
Negate

.AW
.AX
.BA

R IARG
I RARG
R RARG

R=.AW{I)
I=.AX{R)
R=. BA{R)

Multiply
Add
Normalize
Hold
Sign Control

ARG1*ARG2
ARG1+ARG2
N.A
N.A
(Note 1)

.CA
.CC
.CD
.CF
.CG

R=R*R
R=R+R
R=.CD{R)
R=.CF{R)
R=.CG{R)

Short get
argument

N.A

.CB

R=.CB(R)

A-Register

Fl.AC
-

integer

F. P num
F.P num

Fl.AC

HELD AC

multiplicand
augend
value
value
value

}

JMS*SUBR

mulHpHe,
addend

JMS*SUBR

value

CAlO
JMS* .CB
CAlO

SUBR ENTRY-EXIT
STORAGE FOR ARG ADDR

_ .. _--

----

---

(continued next page)

)

Table 3-4 (Cont)
Arithmetic Package
Function
Floating
Arithmetic
(Cont)
Divide
*Round and
sign

Definition

ARG1/ARG2
N.A

Symbolic
Name

.CI
.CH

Mode

R=R/R
R=.CHR

FL.AC

HELD .AC

{ d;y;""

d;y;dend }

value

ARGl
AC,MQ

W
I

-0

INT

Load
Store
Add
Subtract
Reverse
subtract
Multiply
Divide
Reverse
divide

Float
Fix
Negate
- - - - _ ... _-

External
Calls

Calling Sequence

ARG2
address

JMS*SUBR **
CONSTl
CONST2

......

N.A
N.A
ARG1+ARG2
ARGI-ARG2
ARG2-ARGI

.JG
.JH
.JJ
.JJ
.JM

J=.JG(J)
J=.JH(J)
J=J+J
J=J-J
J=J-J

value
augend
minuend
subtrahend

address
addend
subtrahend
minuend

ARG1*ARG2
ARG1/ARG2
ARG2/ARGI

.JK
.JL
.IN

J=J*J
J=J/J
J=J/J

multiplicand
dividend
divisor

multiplier
divisor
dividend

AC,MQ

FL.AC

Doub. Int.

F.P. Numb... }
F. P. Number JMS*SUBR

R.....JARG
J+-RARG
J .....JARG

.JW
.JX
.JA

R=.JW(J)
J=.JX(R)
J=.JA(J)

JMS*SUBR
.DSA ARG2

.CD ,REAL
REAL

----

*The sign of the result (exclusive OR of the sign bits of .AB and CE02) is stored in .CE. The sign of .AB is saved in CEOS.
**CONSTl and CONST2 are required for both EAE and NON-EAE operations, however, they are used only by the NON-EAE version of .CI.
CONSTl indicates the number of bits to be generated (-34 for single precision, -44 for double precision). CONST2 is the least significant
quotient bit (400 for single precision, 1 for double precision).

...

CHAPTER 4
UTILITY ROUTINES

Two types of subprogram are described in this chapter - OTS routines, automatically invoked by
FORTRAN statements; and external subprograms which may be invoked via a FORTRAN CALL statement.
Both types are accessible to MACRO programs.

4.1

OTS ROUTINES

OTS utility routines perform a number of functions specified by FORTRAN statements. These functions
of FORTRAN, like the input-output functions discussed previously, use OTS as an interface between
the user program and the monitor environment in which it will operate.
Each of these routines is described below.

Routine

• S5

Purpose

Calculates the address of an array element

Calling
Sequence

.GLOBl .55
JM5* .55
.D5A ARRAY
LAC (K.)

.55

•

·
LAC (K )

k
DAC ALOC
External Calls

None

Errors

None

/ addr wd. 4 - array descriptor block
/ subscr i pt i

/ subscript k
/ return with element address in AC

.55 references the array-descriptor block associated with the array whose element is to be located.
An array descriptor block is a four-word table with the contents depicted below.

4-1

Data
mode

0-2

3-4

Word 1

Size (in words)
17

a - for one-dimensional array

Word 2

Size of first dimension

a - for one- and two-dimensional arrays

Word 3

Size of the first two dimensions
Address of first word of array with mode in bits 1-2.

Word 4

Size is determined by multiplying the dimensions of the array by the number of words (N) used for a
data item of the specified mode (M). Thus, an INTEGER array defined by DIMENSION (2,2,2) has
the size 8 in word 1, the size 2 in word 2, and the size 4 in word 3. A REAL array of the seme
dimensions will have 16, 4, and 8 in these locations.
The values of M and N for the various data modes are:

Array Mode

INTEGER, LOGICAL

DOUBLE INTEGER

REAL

DOUBLE PRECISION

The address of an array element A(K ,K ,K ) is calculated by .SS using the following formula:
1 2 3
addr = WD4 + (K1-l) * N + (K -1) * WD2 + (K -1) * WD3
3
2
Routine

.GO

Purpose

Computes index of computed GO TO

Call ing
Sequence

LAC V
JMS* .GO
-N
STMT(l)
STMT(2)
STMT(N)

GOTO

index value in A-register

number of statement address

External Calls

OTSER

Errors

OTS 7 - illegal index « 0)

4-2

- ,

STOP

PAUSE

Routine

.ST

Purpose

Processes STOP statement {returns to monitor}

Calling
Sequence

LAC
JMS* .ST

External Calls

• SP

Errors

None

Routine

.PA

Purpose

Processes PAUSE. Waits for fP and returns control
to user program

Calling
Sequence

LAC
JMS* .PA

External Calls

• SP

Errors

None

Routine

.SP

Purpose

Prints octal number for PAUSE and STOP.
Zero assumed if none supplied.

Calling
Sequence

LAC
loctal integer
JMS* .SP
.DSA {control return for PAU SE}
LAC {first character}

SPMSG

loctal number to be printed

loctal number

·
·
LAC {sixth character}

OTSER

External Calls

None

Errors

None

Routine

.ER

Purpose

To print error messages on Teletype and take
action according to class of error

Calling
Sequence

JMS* .ER
.DSA {error number}

External Calls

None

Errors

None

Recoverable errors are indicated when bit 0 of the error number is a 1. In this case, the AC and link
are restored to their original contents and control is returned to the calling program at the first location following the error.

4-3

Unrecoverable errors are indicated when bit 0 of the error number is O. Control is returned to the
monitor by means 'of an • EXIT function. In the case of an unrecoverable error in a FORMAT statement,
the current 5/7 ASCII word pair of the erroneous FORMAT is also printed. The calling sequence for
• ER for a FORMAT statement di ffers from other calls and is:
JMS* .ER
.DSA 12
LAC chars
LAC chars

error number

current 5 characters

PARTWD
Routine

.PB

Purpose

Part word fetch result in AC or ACMQ

Calling
Sequence

JMS* .PB
• DSA address

External Calls

None

Errors

None

Routine

.PC

Purpose

Stores contents of AC or ACMQ

Calling
Sequence

JMS* .PC
• DSA address

External Calls

None

Errors

None

PARTWD

4.2

FLOATING POINT PROCESSOR ROUTINES

General
Interface
Routine
.FPP

Routine

.AX

Purpose

FPP version of software .AX

Routine

.AW

Purpose

FPP version of software .AW

Routine

.ZA

Purpose

Loads high order mantissa of FPP AC into the
regular AC

Routine

.ZB

Purpose

Initializes FPP error handling

Routine
Purpose

Error handling

4-4

Extended
Integer
(Double
Integer)
Interface
Routines

4.3

Routine

.ZC

Purpose

Converts integer in CPU AC to extended integer in
FPP AC

Routine

.ZD

Purpose

Converts extended integer in FPP AC to single
integer in CPU AC

FORTRAN - CALLABLE UTILITY ROUTINES

These routines are described in Table 4-1.

4.4

RSX LIBRARY (.LIBRX BIN) ROUTINES

A special set of routines is provided for use with the RSX-15 real-time monitor system. This library
includes, in addition to the subprograms described pre,:,iously, the FORTRAN-callable external subroutines given in Table 4-2. The even variable values have the following meaning:
a.

Positive values signal successfu I compl etion.

Zero indicates a request is still pending.

Negative values indicate rejection or unsuccessful completion.
-5

Illegal header word from device (data mode incorrect or data validity bits improperly
set) (DVH)

-6

Unimplemented or illegal function (DVH)

-7

Illegal data mode (DVH)

-10

File still open (DVH)

-11

File not open (DVH)

-12

DECtape error (DVH)

-13

File not found (DVH)

-14

Directory full (DVH)

-15

Medium full (DVH)

-16

Output word-pair-count or input-bufFer-size error (DVH)

-23

Input word-pair-count error (DVH)

-24

LUN has been REASSIGNed while an ATTACH or DETACH request was in an I/O
request queue (DVH)

-101

Out of range Logical Unit Number (10.)

-102 Unassigned Logical Unit Number (IO.)
-103

Non-resident Device Handler (10.)

-104 Control Table argument error (DVH)
-201

Task not in system (RQ., SC,. RN., SY., DA., EA., FX,. UF., CN.)
(continued page 4-15)

4-5

Table 4-1
FORTRAN-Callable Utility Routines

Routine

Clock
Handling only one
call may be
active at
any point
in a user's
program

ENTRY
Name
TIME *

TIME10*

Purpose

Call ing Sequence

Records elapsed
time in minutes
and seconds on
60-cycle
machine

CALL TIME{IMIN,ISEC,IOFF)
'Nhere: IMIN == minutes
ISEC == seconds
IOFF =non-zero
to stop clock

CALL TIME(IM,IS,IOF)

Records elapsed
time in minutes,
seconds, and
tenths of seconds

CALL TIME10{IMIN,ISEC,
ISEC10,IOFF)
'Nhere: IMIN == minutes
ISEC == seconds
ISEC10 = tenths of
seconds
IOFF == non-zero
stops clock

See TIME

Controls the
number of runtime arithmetic
errors output by
OTSER

CALL ERRSET{N)
'Nhere: N = integer giving
number of ti mes
message to be
output before
suppressi on. If
ERRSET is not
given, OTSER
assumes N = 2.

.j:>.

Error
Handling

ERRSET

Examples

·
A
·
·
IOF == 1

External
Calls
.DA
• TIMER

Errors

None

WRITE{4, lOO)IM, IS
[outputs time to execute A]

.DA
• TIMER

None

IfN~O,no

messages output.
--

*Not supported with RSX. Other RSX supplied routines can be used for this purpose.

)

L----- _ _ _ _ _ _ _ _

(continued next page)

)

Table 4-1 (Cont)
FORTRAN-Callable Utility Routines

Routine

Adjustable
Dimensioning

ENTRY
Name

ADJ1

Ca" i ng Sequence

Examples

To adjust onedimensional
array

DIMENSION B(1)
CALL ADJ 1(B ,A)
Where: B = array name
A= beginning
storage location
of B array el ement
(e.g., C(200)
which is the
beginning storage
location of B)
Note: The dimensions of A
must be sufficient to
hold all entries of
array B. A may be
a dummy argument in
a subroutjne

DIMENSION A(300),B(1),C(l)

.....I
'-J

Adjustable
Dimensioning
(Cont)

ADJ2

External
Calls

Purpose

To adjust a twodimensional
array

DIMENSION B(1, 1)
CALL ADJ2(B ,A, NR)
'lYhere: A and B are as for
ADJ1
NR = the number of
rows to appear
in B

Errors

.DA

None

.DA
.AD

None

·
·
CALL ADJ1 (B,A(101»
CALL ADJ1 (C,A{20l)
·
·
·

Band C may be referenced as
if they had been dimensioned
as (100) each

DIMENSION A(300) ,B(1, 1),
C(1,l)

·
·
· ADJ2(B,A(1),10)
CALL
CALL ADJ2(C,A(10l),20)
·
·
·
Band C may be referenced as

if they had been dimensioned
(10,10) and (20,10), respectively
--

- _._ - -

-_.-

(continued next page)

Table 4-1 (Cont)
FORTRAN-Callable Utility Routines
ENTRY
Name

Routine

ADJ3

Adjustable
Dimensioning
(Cont)

----

Calling Sequence

Purpose

DIMENSION B(1, , ,1)
CALL ADJ3(B ,A, NR, NC)
INhere: A,B, and NR are
as for ADJ2
NC = number of
columns to
appear in
array B

To adjust a threedimensional array

Examples
DIMENSION A(300) ,B{1, 1),
C(1 , 1)
CAll ADJ3(B,A{1), 10,5)
CAll ADJ3(C,A(101),10, 10)
Band C may be referenced as
if they had been dimensioned
(2,10,5) and (2.10,10),
respectively

External
Calls
.DA
.AD

Errors

I
I

None

I
I
I

)

Table 4-2
FORTRAN-Callable RSX Routines*
Routine

Purpose

REQUEST

Requests task execution

Calling Sequence

Event Variables Returned

CALL REQST(nHTSKNAM,IP[ ,IEVJ)

+1, -201, -202, -204, -777

Where:
n = no. of characters in task name
TSKNAM = name of task (l to 5 characters)
IP = task priority (1-512)
may be variable or constant
lEV = event variable
SCHEDULE

Schedules task execution

CALL SCHED(nHTSKNAM,lT ,IP[ ,IEVJ)

+ 1, -201, -203, and -777

Where:
IT = name of 5-word integer array describing
schedule
IT(l) = schedule of hour (0-23)
IT(2) = schedule of minute (0-59)
IT(3) = schedule of second (0-59)
IT(4) = reschedule interval (up to one day)
IT(5) = reschedule units (l = ticks,
2 = seconds, 3 = minutes, 4 = hours)

.j>...

-b

RUN

Run task in delta time

CALL RUN(nHTSKNAM,IT ,IP[,IEV])

+ 1, -201, -203, and -777

Where:
IT = name of 4-word integer array
IT(l) = schedule delta time from now
(up to one day)
IT(2) = delta schedule units (l = ticks,
2 = seconds, 3 = minutes, 4 = hours)
IT(3) = reschedule interval (up to one day)
IT(4) = reschedule units
-------------------------------------------

--- - - - -

*Square brackets indicate that the event variable is an optional argument.
(continued next page)

Table 4-2 (Cont)
FORTRAN-Callable RSX Routines*
Routine
SYNC

Execute task at a
specified interval

Event Variables Returned

Calling Sequence

Purpose

+1, -201, -203, and -777

CALL SYNC(nHTSK NAM, IT, IP[, lEV] )

I
I

Where:
IT = name of 5-word integer array
IT(l) = synchronization units (1 = ticks,

... )

IT(2) = schedule interval from synchronization time (up to one day)
IT(4) = reschedule interval (up to one day)
IT(5) = reschedule units (1 = ticks, ••• )
+1, -201, and -777

CANCEL

Cancel task execution
(no effect for an active
task)

CALL CANCEL(nHTSKNAM[ ,lEV])

SUSPEND

Suspend execution of
tosk issuing this call.
Execution not permitted
until a RESUME call

CALL SUSPEND

RESUME

Resume task execution

CALL RESU ME(nHTSK NAM [, lEV)

+1, -202, and -205

MARK

Set an event variable
in delta time

CALL MARK(IT, lEV)

+ 1, -203, and -777

.j>o.
I

....o

Where:
IT = name of 2-word integer array
IT(l) = delta interval (up to one day)
IT(2) = delta units (l = ticks, ••• )

WAIT FOR

Suspend task if
event variable = 0;
resume when non-zero

CALL WAITER(IEV)

*Square brackets indicate that the event variable is an optional argument.
(continued next page)

)

Table 4-2 (Cont)
FORTRAN-Callable RSX Routines*
Routine

Ca II i ng Seque nce

Purpose

Event Variables Returned
-~-

WAIT

Suspend execution of
task unti I occurrence
of next significant event

CALL WAIT

EXIT

Terminate task execution

CALL EXIT

DSKAL

Allocate disk storage

CALL DSKAL(ICTS,NW[,IEVJ)

+1, -6, -15, -101, -104, and -777

Where:
ICTS = control table (integer array
returned at end of operation)
ICTS(1) = amount actually allocated
ICTS(2) = physical disk unit number
ICTS(3) = absolute starting address of
the space allocation relative
to physical disk unit number
NW = desired storage (in words)

""'I "
DSKDAL

Deallocate disk storage

CALL DSKDAL(ICTS[, IEVJ)

+1, -6, -15, -101, -104, and -777

Where:
ICTS = control table (same address as
used in the corresponding DSKAL)
DSKPUT

Put data on disk

--_._-_._ .. _-

CALL DSKPUT(ICTA,IOA, NW,ARRAY[ ,lEV])

+1 and -N

Where:

ICTA = device control table (same as
for corresponding DSKAL)
lOA = disk offset address
NW = number of words (decimal) to transfer
ARRAY = name of array containing data to
be tra nsferred

"'Square brackets indicate that the event variable is an optional argument.

N = the contents of the disk status
reg i ster on error

(continued next page)

Table 4-2 (Cont)
FORTRAN-Callable RSX Routines*
Routine

Calling Sequence

Purpose

Event Variables Returned

DSKGET

Get data from disk

CALL DSKGET{ICTA,IOA, NW,ARRAY[ ,lEV])

+land-N

ATTACH

Attach I/o Handler task

CALL ATTACH{LUN[ ,lEV])

+1, -6, -24, -1OJ, -103, and -777

Where:
LU N = logical unit number
DETACH

Detach I/O Handler task

CALL DETACH{LUN[,IEV]

+1, -6, -101, -103, and -777

SEEK

Seek open file for input

CALL SEEK{LU N,nHFLNAM,nHEXT[,IEV])

+ 1, -6, - 10, - 12, - 13, - 101, - 102 ,
-103, and -777

Where:
LU N = logical unit number
n = number of characters in file name or
extension
FLNAM = 1-5 character file name
EXT = 1-3 character extension

~
I

I-..>

Open fil e for output

CALL ENTER{LUN,nHFLNAM ,nHEXT[ ,lEV])

Closes file

CALL CLOSE{LU N,nHFLNAM,nHEXH ,lEV])

+ I, -6, - 11, - 12, - 13, - 14, - 10 I,
-102, -103

HINF

Provides information
about the physical
device and the I/O
Handler associated
with a particular
Logical Unit Number
(LUN)

CALL HINF{LUN,IEV)

Single word containing the following
Handler information:

ENTER

+1. -6, -11, -12, -14, -101, -102,
-103, and -777

Bit 0 - unused
Bit 2 - input - set to 1 if data can be
input
Bit 2 - output - set to 1 if data can be
output
Bit 3 - file-oriented - set to 1 if fileoriented (SEEK and ENTER have
been used)
--

~~--

*Square brackets indicate that the event variable is an optional argument.

)

(continued next page)

)

Table 4-2 (Cont)
FORTRAN-Callable RSX Routines*
Routine

Calling Sequence

Purpose

Event Variables Returned
Bits 4-11 - unit number
Bits 12-17 - device code (1 to 63
decimal devices). Codes
below are fixed for standard devices
1 - TTY (console, LT15, LT19)
2 - DK - RF 15 fixed-head DECdisk
3 - DP - RP02 disk pack
4 - DT - TC02D DECtape
5 - MT - TC59 MAGtape
6 - PR - PC15 paper-tape reader
7 - CD - CR03B card reader
10 - PP - PC15 paper-tape punch
11 - LP - LP15 line printer
12 - VP - VP15 storage scope
13 - VT - VT15 display
Users should assign codes to their own
devices starting at 63 and working back

HINF(Cont)

.j>..
I
~

Disable task

CALL DISABL(nHTSKNAM[ ,lEV])

+1, -201, -210

ENABLE

Enable task

CALL ENABLE(nHTSKNAM[ ,1M)

+1, -201, -210

FIX

Fix task in core

CALL FIX(nHTSKNAM[,IEVJ)

+1, -201, -207

UNFIX

Unfix task in core

CALL U NFIX(nHTSKNAM[ ,IEVJ)

+1, -201, -207

DECLAR

Declares a significant event

CALL DECLAR

TIME

Obtain time from
Executive

CALL TIME(ITIME)

DISABLE

Where:
ITIME == 3-word integer array
ITIME(1) == hours (0-23)
ITIME(2) == minutes (0-59)
ITIME(3) == seconds (0-59)
- ----

*Square brackets indicate that the event variable is an optical argument.

(continued next page)

Table 4-2 (Cont)
FORTRAN-Callable RSX Routine*
Calling Sequence

Purpose

Routine

Obtain time and date
from Executive

DATE

Event Variables Returned

CALL DATE(IDATE)
Where:
IDATE:= 6-word integer array
IDATE(1) = month (1-12)
IDATE(2) = day (1-31)
IDATE(3) := year (0-99)
IDATE(4) = hours (0-23)
IDA TE(5) = mi nutes (0-59)
IDATE(6) = seconds (0-59)

-- -

-----------

------

--- - - -

----

- -- - -

""Square brackets indicate that the event variable is an optical argument.

..."""
I

"""

)

-202

Task is active (RQ., FX.) or not active (RS.)

-203

CAL not Task issued (SC., RN., SY., MT.)

-204

Task is DISABLED (RQ., SC., RN., SY., FX.)

-205

Task not suspended (RS.)

-207

Task already FIXed (FX.) or not FIXed (UP.)

-210

Partition occupied (FX.)

-301

Line number rejected (CI., DI.)

-302

Line is CONNECTed (CI.) or DI CONNECTed (DI.)

-777

Pool is empty

DVH - Device Handler
10.

'QUEUE I/o' Directive

RQ.

'REQUEST' Directive

SC.

'SCHEDULE'Directive

RN.

'RUN'Directive

SY.

'SYNC' Directive

CN. -

'CANCEL'Directive

RS.

- 'RESUME' Directive

CI.

- 'CONNECT'Directive

DI.

- 'DISCONNECT' Directive

FX.

UFo
DA.

- 'UNFIX'Directive
- 'DISABLE' Directive

EA.

- 'ENABLE' Directive

MT.

'FIX IN CORE' Directive

'MARK' Directive

OTS routines which have been modified for RSX are:
FlOPS

modified to use the RSX I/o CAL'S •• FP, which initializes the I/O status table
has been converted to a dummy subroutine.
If a Negative Event Variable occurs as a result of a FlOPS issued I/O request, an
error message (OTS 20) is issued and the task is EXITed.

SPMSG -

rewritten to include the task name. The message is output to LU N 4 in the following format:
STOP - 000000 - TSKNAM

STOP

uses RSXEXIT CAL

PAUSE

SUSPENDs the issuing task. To continue, the RESUME MCR function is used.

OTSER

passes its name and an octal OTS error message number to SPMSG.

Additional routine used by RSX for bank/page mode determination is • BP.

4-15

Two additional OTS routines are given below:

• ASCII
to
• SIXBT
Conversion

Routine

.FTSB

Purpose

To convert two words from • ASCII to • SIXBT

Calling Sequence:

SUBA

0
JMS* .DAA /
JMP ARGEND

FROM

get call args

PTR to ASCII word-pair

two 6-bit words

ARGEND JMS* .FTSB
.DSA FROM
.DSA TO

·
·
·

BLOCK 2

.DAA is a routine which performs the argument I ist transfer function formerly performed by • DA. The
calling sequence has not been changed, but the transfer stops with the end of the shortest argument.

4-16

CHAPTER 5

FORTRAN-IV AND MACRO

In previous chapters, MACRO calling sequences have been given for OTS and Science Library Subprograms. This general form is used in a MACRO program to call any FORTRAN external subroutine
or function. A FORTRAN program may also invoke MACRO subprograms. The method for each type
of linkage is given below.

5.1

INVOKING MACRO SUBPROGRAMS FROM FORTRAN

A FORTRAN program may invoke any MACRO program whose name is declared in a MACRO .GLOBL
statement. The MACRO subprogram must also include the same number of open registers as there are
arguments. These will serve as transfer vectors for arguments supplied in the FORTRAN CAll statement
or function reference. A FORTRAN-IV program and the MACRO subprogram it invokes are shown
below. More extensive examples are given in Appendix C.

MACRO

FORTRAN

TEST MACRO SU BR

READ A NUMBER(A)

READ(1 , lOO)A

100

FORMAT(E12.4)

C
C

NEGATE THE NU MBER
AND PUT IT IN B

MIN

CALL MIN(A,B)
C

WRITE OUT NUMBER(B)
WRITE(2,100)B
STOP
END

MINl
MIN2

• TITLE MIN
.GLOBL MIN, .DA
0
JMS* .DA

/ entry/exit
/ general get

JMP .+2+1

/
/
/

.DSA 0
.DSA 0
lAC* MIN1
DAC* MIN2
ISZ MINl
ISZ MIN2
lAC* MINl
TAD (400000)
DAC* MIN2
JMP* MIN
.END
5-1

argument
(OTS)
jump around
argument
registers

/ ARGl
/

ARG2
/ first word of A
/ store at B
/ poi nt to second word
/ of A and B
/ second word of A
/ sign bit = 1
/ store in second
/ word of B
/ exit

The FORTRAN statement CALL MIN(A,B) is expanded by the compiler to:
00013 JMS* MIN
00014 JMP$ 00014
00015 .DSA A
00016 .DSA B
$00014 = 00017

I to MACRO subprog

When the FORTRAN-IV program is loaded, the addresses (plus relocation factor) of A and B are stored
in registers 15 and 16, respectively. When the MACRO program invokes .DA, these addresses are
stored in MINI and MIN2 and the values themselves are accessed by indirect reference.
Arguments are, as described above, transmitted by • DA using a single word. Bits 3-17 contain the

IS-bit address of the first word. Bits 0-2 serve as flag. FORTRAN uses bit 0 to indicate that the word
specifying the argument contains the address of a word containing the address of the first word of the
argument. The MACRO argument word always contains the address of the first word of the argument.
For array name arguments (unsubscripted), the address of the fourth word of the array descriptor block
is given.

.55 must be invoked to locate the element.

For external functions, the MACRO subprogram must return with a value in the AC (LOGICAL,
INTEGER), AC-MQ (DOUBLE INTEGER) or in the floating accumulator (REAL or DOUBLE PRECISION).

5.2

INVOKING FORTRAN SUBPROGRAMS FROM MACRO

The MACRO calling conventions for FORTRAN subprograms are:

the name of the subprogram must be

declared as global; there must be a jump around the argument address; and the rumber and mode of
arguments in the call must agree with those of the subprogram. This form is shown below.
TITLE
.GLOBL
JMS*
JMP
.DSA
.DSA

ARGI
ARG2

.DSA

ARGN

SUBR
SUBR

.+N+l

I jump around arguments ignored by • DA
I address of first argument - bit 0 set to 1
I indicates indirect reference

When the subprogram is compiled, a call is generated to .DA which performs the transmission of
arguments from MACRO. The beginning of a subroutine might be expanded as follows.

5-2

-...

C
000000
000001
000002
000003
000004
$ 000002 = 000005

TITLE SUBR
SUBROUTINE SUBR(A,B)
CALO
JMS* .DA
JMP $000002
.DSA A
.DSA B

If a value is to be returned by the subroutine, it is most convenient to have this be one of the calling
arguments. An external function is called in the same manner as a subroutine but returns a value in
the AC (single integers), AC-MQ (double integers), or floating accumulator (real and double-precision).
To store the AC, the MACRO program uses a DAC instruction. Values from the floating accumulator
may be stored via the OTS routines .AH (real) and .AP (double-precision). For FPP systems, values
are returned in a hardware accumulator and stored with an FST instruction.
A number of examples of MACRO-FORTRAN linkage are given in Appendix C.

5.3

COMMON BLOCKS

FORTRAN COMMON blocks (and block-data subprograms) may be linked to MACRO programs. When
the MACRO program is loaded, global symbols are first sought in the user and system libraries. Any
remaining are matched, where possible, to COMMON block names. For example:

FORTRAN
INTEGER A,B,C
COMMON/NAME/C
COMMON A,B

MACRO
.GLOBL NAME, .XX
DZM* .XX
ISZ
.XX
DZM* .XX
DZM* NAME

/ .XX is name given to blank COMMON
/ by the F4 Compiler
/ CLEAR A - NOTE INDIRECT REFERENCE
/ BUMP COUNTER
/ CLEAR B
/ CLEAR C

Note that if the values are REAL (two words) or DOUBLE PRECISION (three words), the MACRO program
must account for the number of words when accessing specific variables. This cannot be done if programs
are loaded via CHAIN and EXECUTE,.

5-3

APPENDIX A
LANGUAGE SUMMARY

Statement
Arithmetic

Model

Effect

var = value
array (i) = value

value is assigned to

Text
Reference
2.1

~ or array (i)

ASSIGN

ASSIGN n TO label

Statement.!!. is assigned
the symbol name label

2.2

BLOCK DATA

Identifies subprogram
which enters data into
COMMON block at run time

4.4

CALL

CALL subr(a ,a ,· •• a )
1 2
n
CALL subr

Control is transferred to the
subroutine; a ,a , ••• a are
- 1 -2
-n
substituted for dummy variables

5.2.2

COMMON

COMMON/ b /vlist /b /
1 2
1
vlist 2 /·· •

vlist items are allocated to b
blocks where they are shared
by other programs

4.2.2

CONTINUE

Dummy statement used to
prevent illegal termination
of DO loops

3.2.3

DATA

DATA vlist /clist}/,vlist /
2
1
clist /, . .. vlist clist
n
n
2
DECODE(c, v, f, ERR=n) list

clist is assigned to its corresponding vlist

4.3

Converts character data stored
in the array (v) into binary and
assigns them to variables in list

6.3.4

DIMENSION

DIMENSION a (11) ,a 2 (12)'···
1
a (I )
n n

Storage is allocated for array
(a) to the dimensions specified
by the subscript I ist (I)

4.2.1

DO n ~=m1,m2,m3
DO n ~=m1,m2
DO n ,=m ,m ,-m
3
1 2

Statements following the DO
are executed repeatedly for
values!!! 1 through !!l2 in increments or aecrements of .!!!3

3.2

DECODE

A-I

Effect

Model

Statement

Text
Reference

ENCODE

ENCODE(c, v, f, ERR=n)1 ist

Converts binary data represented by variables in list
into characters accordi ng
to FORMAT specification
(f) or data-directed I/O
rules and stores them in the
array (v)

6.3.4

EQUIVALENCE

EQUIVALENCEO 1) ,(12)'···
(I )
n

Elements of each I ist (I)
are assigned to the sa~
storage location

4.2.3

EXTERNAL

EXTERNAL a ,a , ••• a
n
1 2

Defi nes subprograms
named ~ for use as arguments of other subprograms

4.1.3

FORMAT

n FORMAT(sl ,s2'·· .sn)

FORMA T statement .!l established as field-specification
reference

6.1

FUNCTION

m FUNCTION f(a ,a , ••• a )
n
1 2

Defines FU NCTION named

5.1.2

.!. with dummy arguments ~
and optional mode specification m

GOTO

GO TO n

Control is unconditionally
transferred to statement n

3.1.1

GO TO(n ,n , ••• n ),i
k
1 2

Control is transferred
to the Ith statement in
the list of n's

3.1.2

GO TO label
GO TO label,{n ,n , ••• n )
k
1 2

Control is transferred to the
location specified by label; the
list of ~ may specify legally
ASSIGNable statement numbers

3.1.3

IF{expr)n ,n ,n
1 2 3

Control is transferred to
statement number or ASSIGNed
label !}1' !}2' or !}3 if evaluated
expr is :( 0, = 0, or >0 respec-

3.3.1

tivey

IF (expr)s

Statement ~ is executed if expr
is • TRUE. (non-zero), ignored
if .FALSE. (zero)

3.3.2

IMPLICIT

IMPLICIT m (11),m (1 ), •••
2 2
1
m (I )
n n

Declares mode (m) for variables
beginning with alphabetic characters in list (I)

4.1.2

PAUSE

PAUSE
PAUSE n

Interrupts program execution;
if present, integer.!l is printed
on the console to distinguish
one PAUSE from another

3.4.1

A-2

Statement

Effect

Text
Reference

PRJ NT(d, f)1 ist

The values of variables in
list are converted to ASCII
according to FORMAT
reference (f) and transferred
to external-device (d)

6.3.2

PRJNT(d)list

The values of variables in
J.lg are written in binary on
external device ~)

6.3.2

PRINT(d,)list

The variable names in list
are written on external
device (d), each followed
by its value in the form
IAI = value

6.3.2

PRINT(d,f)

FORMAT reference (f) is
written on external device

6.3.2

Model

(~)

READ(d, f)list

The values represented by
variables in list are read
from externaldevice (d)
and converted accordi;:;-g
to FORMAT reference (f)

6.3.2

READ(d)1 ist

The binary values represented by variables in list
are read from external
device (d)

6.3.2

READ(d,)list

The values represented by
variables in list are read
from external device (d)

6.3.2

READ(d,f)

Values are read into FORMAT
reference <D

6.3.2

READ(d)

A binary record is read from
external device (d) and
ignored

6.3.2

STOP

STOP
STOP n

Signifies the logical end of
a program and returns control
to the MONITOR after.!l is
printed; if present, .!l distinguishes one STOP from
another

3.4.2

SUBROUTINE

SUBROUTINE nome
(a ,a ,· .. an)
1 2
SUBROUTINE name

Defines an external subroutine
named ~i a IS are dummy
arguments representing values
supplied by the calling program
or returned by the subroutine

5.2.1

READ

A-3

Statement
TYPE

WRITE

Text
Reference

Model

Effect

TYPE(d, f)1 ist

The values of variables
in list are converted to
ASCII according to FORMAT
reference (f) and transferred
to externaldevice (d)

6.3.2

TYPE(d)list

The values of variables in
list are written in binary on
external device (d)

6.3.2

TYPE(d ,)list

The variable names in list are
written on external device (d),
each followed by its value f;:;
the form 'A' = value

6.3.2

TYPE(d,f)

FORMA T reference (f) is
written on external device (~)

6.3.2

WRITE(d, f)list

The values of variables in
list are converted to ASCII
according to F ORMAT reference (f) and transferred to
extern~1 device (d)

6.3.2

WRITE(d)1 ist

The values of variables in
list are written in binary on
external device (~

6.3.2

WRITE(d,)list

The variable names in list qre
written on external device (d),
each followed by its value in
the form 'A' = va lue

6.3.2

WRITE(d,f)

FORMAT reference (f) is
written on external device @

6.3.2

A-4

APPENDIX B
ERROR MESSAGES

S.l

COMPILER ERROR MESSAGES

In the F4X version of FORTRAN, compiler error messages are printed in the form:
>mnA<
where:
mn is the error number
A is the alphabetic mnemonic
characterizing the error class.
In F4I and F4A versions, only the alphabetic character is printed, in the form:
>A<
All error messages and the version(s} of FORTRAN to which they are applicable are given below.

"!umber

Letter

Meaning
Common, equivalence, data errors:

No open parenthesis after variable name in DIMENSION
statement

No slash after common block name

Common block name previously defined

Variable appears twice in COMMON

EQUIVALENCE list does nat begin with open parenthesis

Only one variable in EQUIVALENCE class

EQUIVALENCE distorts COMMON

EQUIVALENCE extends COMMON down

Inconsistent EQUIVALENCing

EQUIVALENCE extends COMMON down

Illegal delimiter in EQUIVALENCE list
{continued on next page}

B-1

Number

Letter

Meaning
Common, equivalence, data errors: (cont)

Non-COMMON variables in BLOCK DATA

III egal repeat factor in DATA statement

DATA statement stores in COMMON in non-BLOCK DATA
statement or in non-COMMON in BLOCK DATA statement
DO errors:

Statement with unparenthesized =sign and comma not a DO
statement

DO variable not followed by =sign

DO variable not integer

Initial value of DO variable not followed by comma

Improper delimiter in DO statement

I11egal terminating statement for DO loop

"-,

External symbol and entry-point errors:
01

Variable in EXTERNAL statement not simple non-COMMON
variable

ENTRY name non-unique

ENTRY statement in main program

No = sign following argument list in arithmetic statement
function

No argument list in FUNCTION subprogram

Subroutine list in CALL statement already defined as variable

Function or array name used in expression without open
parenthesis

Format errors:
01

Bad delimiter after FORMAT number in I/o statement

Missing field width, illegal character or unwanted repeat
factor

Field width is 0

Period expected, not found

Period found, not expected

Decimal length missing (no "d" in "Fw .d")

Unparenthesized comma

(continued on next page)

B-2

Number

Letter

Meaning
Format errors: (cont)

Minus without number

No P after negative number

No number before P

No number or 0 before H

No number or 0 before X

Too many left parentheses
Hollerith errors:

Number preceding H not between 1 and 5

Carriage return inside Hollerith field

Number preceding H not an integer

JIiore than five characters inside quotes

Carriage return inside quotes
Various illegal errors:

Unidentifiable statement

Misspelled statement

Statement out of order

Executable statement in BLOCK DATA subroutine

Illegal character in I/o statement, following unit number

Illegal delimiter in ASSIGN statement

Illegal type in IMPLICIT statement

Logical IF as target of logical IF

RETURN statement in main program

Semi colon in COMMO N statement outside of BLOCK DATA

Illegal delimiter in IMPLICIT statement

Misspell ed REAL or READ statement

Misspelled END or ENDFlLE statement

Misspelled ENDFILE statement

Statement function out of order or undimensioned array

Typed FUNCTION statement out of order

III egal character in context

Illegal logical or relational operator
(continued on next page)

B-3

Number

Letter

Meaning
Various illegal errors: (cont)

Illegal lener in IMPLICIT statement

Illegal letter range in IMPLICIT statement

Illegal delimiter in letter section of IMPUCIT statement

III egal character in context

Illegal comma in GOTO statement

Illegal variable used in multiple RETURN statement
Pushdown list errors:

DO nesting too deep

Illegal DO nesting

Subscript/function nesting too deep

Backwards DO loop (also caused by some illegal I/o lists).
Appears after END statement.
Overflow errors:

EQUIVALENCE class list full

Program size exceeds 8K

Array length larger than 8K

Element position in array larger than 8K (EQUIVALENCE,
DATA)

Integer negative or larger than 131071

Exponent of floating point number larger than 76

Overflow accumulating constant - too many digits

Statement number errors:
01

Multiply defined statement number or compiler error

Statement erroneously labeled

Undefined statement number

FORMA T statement without statement number

Statement number expected, not found

Statement number more than five digits

Illegal statement number

(continued on next page)

B-4

Number

Letter

Meaning
Partword errors:

Expected co Ion, found none

Q.2

Expected close bracket, found none

Last bit number larger than 35

First bit number larger than last bit number

First and last bit numbers not simple integer constants
Subscripting errors:

Illegal subscript delimiter in specification statements

IVore than three subscripts specified

Illegal delimiter in subroutine argument list

Non-integer subscript

Non-scalar subscript

Integer scalar expected, not found

Two operators in a row

Close parenthesis following an operator

Non-integer subscript

Non-scal ar subscri pt

Two arguments ina row

Digit or letter encountered after argument conversion

Number of subscripts stated not equal to number declared
Table overflow errors :

Arithmetic statement, computed GOTO list, or DATA statement list too large

Too many dummy variables in arithmetic statement function

Symbol and constant tables overlap
Variable errors:

Two modes specified for some variable name

Variable expected, not found

Constant expected, not found

Array defined twice

Error : variable is EXTERNAL or argument (EQUIVALENCE,
DATA)

IVore than one dimension indicated for scalar variable

(continued on next page)

B-5

Number

Letter

Meaning
Variable errors: (cant)

First character after READ or WRITE not open parenthesis in
I/O statement

III egal constant in DATA statement

Variables outnumber constants in DATA statement

Constants outnumber variables in DATA statement

Illegal dummy variable (previously used as non-dummy variable)

Logical operator has non-integer, non-logical arguments

Illegal mixed mode expression

Logical operator has non-integer, non-logical arguments

Signed variable left of equal sign

Illegal combination for exponentiation

• NOT. operator has non-integer, non-logical argument

Function in specification statement

28
29

Two exponents in one constant

Illegal redefinition of a scalar as a function

No number after E or D in a constant

Non-integer record number in random access I/O

Illegal delimiter in I/O statement

Illegal syntax in READ, WRITE, ENCODE, or DECODE
statement

END and ERR exists out of order in I/O statement

Constant and variable modes don't match in DATA statement

ENCODE or DECODE not followed by open parenthesis

Illegal delimiter in ENCODE/DECODE statement

Array expected as first argument of ENCODE/DECODE
statement

Illegal delimiter in ENCODE/DECODE statement
Expression errors:

Carriage return expected, not found

Binary WRITE statement with no I/o list

Illegal element in I/O list

Illegal statement number list in computed or assigned GOTO

Illegal delimiter in computed GOTO

Illegal computed GOTO statement
(continued on next page)
B-6

---

)

Number

Letter

Meaning
Expression errors: (cont)

B.2

Illegal delimiter in DATA statement

No close parenthesis in IF statement

Illegal delimiter in arithmetic IF statement

Expression on left CJf equals sign in arithmetic statement

Too many right parentheses

III egal open parenthesi s (i n speci fi cation statements)

III egal open parenthesis

19
20

Too many right parentheses

Illegal alphabetic in numeric constant

Symbol contains more than six characters

• TRUE., • FALSE. , or • NOT. preceded by an argument

Unparenthesized comma in arithmetic expression

Unary minus in I/o list

Illegal delimiter in I/o list

27
28

Unterminated implied - DO loop in I/O list

Illegal equals sign in I/o list

Illegal partword operator

III egal arithmetic expression

OTS ERROR MESSAGES

Following is a list of OTS error messages.

Error Number

05 (R)
06 (R)

(R) indicates a recoverable error; (T) a terminal error.

Error Description

Possible Source

Negative REAL square root argument

SQRT

Negative DOUBLE PRECISION square root
argument

DSQRT

07 (R)

Illegal index in computed GO TO

.GO

10 (T)

Illegal I/O device number

.FR, .FW, .FS, .FX,
DEFINE, RAN COM

11 (T)

Bad input data - lOPS mode incorrect

.FR, .FA, .FE, .FF, .FS,
RANCOM, RBINIO,
RBCDIO
(continued on next page)

8-7

Error Number
12

(T)

Error Description

Possible Source

Bad FORMAT

.FA, .FE, .FF

13 (T)

Negative or zero REAL logarithmic argument
(terminal)

.BC, .BE, ALOG

(R)

Negative or zero DOUBLE PRECISION logarithmic argument

.BD, .BF, .BG, .BH,
DLOG, DLOG10

(R)

Zero raised to a zero or negative power (zero
result is passed)

.BB, .BC, • BO, . BE, . BF,
.BG, .BH

Fatal I/O error (RSX only)

FlOPS

Undefined file

RANCOM

Illegal record size

DEFINE

Size discrepancy

RANCOM

Illegal record number

DEFINE, RAN COM

Mode discrepancy

RANCOM

Too many open fil es

DEFINE

Single integer overflow*

RELEAE, .FPP

Extended (double) integer overflow****

DBUNT, JFIX, JDFIX,
lSNGL

(R)
**33 (R)

Single fit. overflow

RELEAE

**34 (R)

Single fit. underflow

(T)
,... 21 (T)
22 (T)
direct
23 (T)
access <
errors
24 (T)
25 (T)
,-26 (T)
30 (R)
**31 (R)
20

**32

(R)
**36 (R)
***37 (R)
40 (T)
**35

Double fit. overflow t

Double fit. underflow

RELEAE
t

Fit. divide check

RELEAE

Integer divide check

INTEAE

Illegal number of characters specified [legal:
o<c<625l

ENCODE

(R)

Array exceeded

ENCODE

42 (T)

Bad input data

DD10

**50

(T)

51 (T)

FPP memory protect/non-existent memory
(READ to WRITE Illegal I/O Direction Change
to Disk) without intervening CLOSE or REWIND

BCOlO, BINIO

*Only detected when fixing a floating point number.
**Also prints out PC with FPP system
***If extended integer divide check, prints out PC with FPP system.
****With software F4 system only detected when fixing a floating point number.
tNot detected by software system (only by FPP system).

B-8

B.3

OTS ERROR MESSAGES IN FPP SYSTEMS

In software systems, arithmetic errors resulting in the OTS error messages summarized above are detected in the arithmetic package (RELEAE and INTEAE). In the hardware FPP systems, these errors
are detected by the hardware (with the exception of single integer divide check) and serviced by a
trap routine in the FPP routine. FPP.
Where applicable, on such error conditions, the result is patched for both software and hardware systems as summarized in the following table.

PATCHED VALUE***

Error

FPP Hardware System

Software System

Single Floating Overflow
(.OTS 32)

± largest single floating value

same

Double Floating Overflow
(.OTS33)

± largest single floating value

not detected

Single Floating Underflow
(.OTS34)

zero

same

Double F looting Underflow
(.OTS 35)

zero

not detected

Floating Divide Check
(.OTS36)

± largest single floating value

same

Integer Overflow
(.OTS 30)

limited detection*

same

Double Integer Overflow
(.OTS 31)

none**

limited detection*

Integer Divide Check
(.OTS 37)

none

same

*When fixing a floating point number, integer and extended integer overflow is detected. In these
instances, plus or minus the largest integer for the data mode is patched as result.
**With the FPP system all extended integer overflow conditions are detected, but the results are
meaningless.
***Where "none" is specified, the result is meaningless unless otherwise indicated.
17
Further, when converting an extended integer, the magnitude of which is >2 _1, to a single
integer, no error is indicated and the high order digits are lost.

B-9

---.

APPENDIX C
PROGRAMMING EXAMPLES

C.1

MACRO-FORTRAN Linkages

Example 1.

A New Dimension Adjustment Routine

The present versions of the OTS routines ADJ1, ADJ2, and ADJ3 do not alter the size of the array
being adjusted. If only the array name of an adjusted array is given in a READ or WRITE argument list,
FORTRAN uses this size information; therefore, undesired results can occur. A new routine (ADJ) can
be loaded with a user program which completely handles all cases of dimension adjustment, although it
occupies 72 octal locations. (ADJ3 occupies 41 octal locations.) Consider the following programs:
C

PROGRAt-' )
DI~ENSION

A(4,3,2)

C
C

~AKE ARRAY A ACT LIKE IT
WAS DIMENSIONED A (2,3,4)
CALL ADJ(A,A(I,1,1),2,3 1 4)

PROGRAM 2
DI~ENSION

C
C

A(3,2)

ADJUST ARRAY A TO BE A C2,3)
CALL ADJ (AIA<l,1 ),2,3,0)
THE LAST ARGUMENT MUST BE 0

PROGRAM 3
DIIVoENSION A(2)

ADJUST ARRAY A TO BE ACt)
CAL L AD J CA I A C1 ) I ) .. 0 , 0 )
THE LAST 2 ARGUMENTS MUST BE ZEKO
THE NO. OF SU8SCRIPTS IS NOT ADJUSTABLE

C
C

(continued on next page)

C-l

.TITLE ADJ
I

ISUBROUTINE TO PERFORM DIMENSION ADJUSTMENT
I

IMACRO-lS CALLING SEQUENCE
.GLOBL ADJ
I
JMS* ADJ
I
JMP .+6
I
.DSA ARRAY
IADDRESS OF WD4
I
.DSA 8
INEW WD4
I
.DSA Kl
IADDRE:SS OF NEW MAXIMUM 1ST SUBSCRIPT
I
.DSA K2
IADDRESS OF NEW MAXIMUM 2ND SUI:lSCRIPT
I
.DSA K3
IADDRESS OF NEW MAXIMUM 3RD SUBSCklPT
I

ADJ

.GLORL ADJ,.DA,.AD
'"
JMS* .DA IGET ARGUMENTS
JMP .+5+1 IN OF ARGUMENTS = 5

ARRAY

(I)

KI
K2

'"
0

'"
LAC (LAC* B
IINITIALIZE SUBSCRIPT POINTER
DAC C
LAC B
ISET NEW STARTING ADDRESS
DAC* ARRAY
LAW -3
DAC CTRI IMAXIMUM OF 3 SUBSCRIPTS
TAD ARRAY
DAC ARRAY IPOINT TO FIRST WORD
DAC ARRAYP,
10F ARRAY DESCRIPTOR BLOCK
LAC. ARRAY
IARRAY TYPE IN BITS 3-4
AND (60000
IZERa OUT ARRAY SIZE
DAC* ARRAY
ISAVE CLEAN ARRAY TYPE
RTL
RTL
RTL
TAD (I
IADD 1 FOR' OF WORDS
~ND (3
lAND TREAT DOUBLE INTEGER
SNA
lAS 2 WORD PER ARRAY ELEMENT
LAC (2
ISZ C
IPOINT TO NEXT SUBSCRIPT
JMS* .AD IMULTIPLY INTEGERS
LAC* Kl
IPROGkAM MODIFIED
SNA
lIS SUBSCRIPT PRESENT
JMP D
IRAN OUT OF SUBSCRIPTS
DAC SIZF.# IUPDATE SIZE
I5Z CTR
IARE WE FINISHED?

LOOP
C

SKP

JMP E

IYES

ISZ ARRAYP
DAe * ARRAYP
JMP LOOP IOFFSET
DZM* ARRAYP
ISZ ARRAYP

ISTORE INTO ARRAY
IDESCRIPTOR BLOCK
~JORDS

(2~3)

IZEIW THE REST

IOF THE OFFSET WORDS

(continued on next page)

C-2

Example 2.

IARE ';lE FINISHED
ISZ CTR
JMP LOOP INO
LAC SIZE IFINISHED
IPACK SIZE
AND (17777
IARRAY DESCklPTOk BLOCK
XOR* ARRAY
DAC* ARRAY
JMP* ADJ IRETURN
.END

A Function to Read the AC Switches

It is very often desirable to use the AC switches to alter the sequence of instructions executed in a
FORTRAN program. The following program can be used as a function in an arithmetic IF statement to
conditionally branch •

• TITLE ITOG
I

ISUBROUT I NE TO READ AC SW ITCHES
I

IMACRO-15 CALLING SEQUENCE
I
.GLOBL ITOG
I
JMS* ITOG
I
JMP .+2
IJUMP OVER
I
.DSA (MASK

AkGU~ENT

Example 3.

IADDRESS OF t'-iASK
IRETUkN WITH ~ASKED ACS IN AC

.GLOSL ITOG ... DA
II NTEGEk FUNCTI ON
JMS* .DA IGET ARGUt.t'ENTS
JMP • +1 +1 I 1 AHGU~ENT
IMASK ADDRESS
ILOAD AC FROM SWITCHES
LAS
AND* MASK 1i"1ASK AC
J~1P* ITOG IRETURN WITH MASKED AC SW ITCHES
.END

ITOG

MASK

A Routine to Read an Array in Octal

A MACRO subroutine which reads octal information (REDAR) is as follows:

• T ITl.E REDAR
I

ISUBROllTINE TO READ Af< f<AY IN OCTAL
I

IMACRO-15 CALLING SEQUENCE
.GLOBL kEDAR
JMS* REDAR
J~P .+5
.DSA SLOT IADDRESS OF SLOT #
IADDRESS OF FORMAT STATEMENT ADDRESS
.DSA FORMAT
IADDRESS # OF DIGITS
.DSA DIGITS
IADDRESS OF ARRAY DESCRIPTOR
.DSA ARRAY
/BLOCK ,·IORD 4

I
I
I
I
I
I
I
I
I

(continued on next page)

C-3

.GLOBL REDARI.DAI.FHI.FEI.FF

REDAR
SLOT
FORMAT
DIG ITS
ARRAY

A
8

JMS* .DA IGET ARGUMENTS
JMP .+4+1 I#ARGUMENTS = 4

"o
€I

LAC SLOT
DAC A
LAC* FORI'1AT
DAC 8
JMS* .FR IFORt-~ATTED WR ITE
XX
IADDRESS DAT SLOT #
xx
IADDKESS OF FORMAT STATEMENT
LAlv -3

TAD ARRAY
DAC SLCT IADDkESS OF ARRAY DESCklPTOk BLOCK WORD
LAC* SLOT IPICK UP PACkED SIZE OF AkRAY
AND (17777
ICLEAN OFF MODE #
SNA
JMP E
INO ELEMENTS IN ARRAY
Ct-'~A

DAC SLOT
ISZ SLOT ICOUNTER FOR # WORDS IN ARRAY
LAC* DIGITS
I#DIGIT5 IN EACH WORD
AND (7
ICLEAN ARGUMENT
SZA
SAD (7
JMP E
10 OR 7 DIGITS ILLEGAL
CMA
TAD (1
DAC C
IINITIALIZE LAW INSTRUCTION
LAC* ARRAY
DAC ARRAY IPOINTER TO FIRST WORD OF ARRAY
XX
ILAW -DIGITS
DAC DIGITS
CLA
IINITIALIZE DIGIT PACk
DAC TEMP' 1ST ORE DIGIT PACK
JMS* .FE IREAD DIGIT
.DSA FORMAT
IDIGIT READ INTO FORMAT
LAC TE~P ILOAD DIGIT PACK
CLL

CTL
RAL
TAD FORMAT
ISZ DIGITS
JMP D
DAC* ARRAY
ISZ ARRAY
ISZ SLOT
JMP C
JMS* .FF
JtvlP* REDAl"<
.END

IMULTIPLY BY B
IADD DI GIT
ICOUNT DIGITS
IGO BACK FOR MORE
ISTORE VALUE IN ARRAY ELE MENT
IPOINT TO NEXT ARRAY WORD
ICOUNT ARRAY WORDS
IREAD ANOTHER WORD
/END OF READ
IEXIT

C-4

Example 4.

A FORTRAN Program Using the Foregoing Programs

This FORTRAN program uses the preceding three MACRO programs to read in an array from the
Teletype in octal and type it in decimal. The Teletype should be assigned to .DAT slot 4. Note
how the arguments ore specified. Notice that EQUIVALENCE performs the array element calculation at compile time.

C FORTRAN PROGRAM TO READ AN ARBITRARY INTEGER ARRAY IN OCTAL
C AND WRITE IT IN DECIMAL
DIMENSION J(2000)
C USE EQUIVALENCE TO GET J(I) WITHOUT USING .55
EQUIVALENCE (J(J ) .. K)
C I CONTAINS ADDRESS OF FORMAT
C STATEMENT + 1 TO ~OVE OVER JMP INSTRUCTION
ASSIGN 1 TO 1
1=1+1

FOR MA T ( 6 1 1 .. 1 X.. 6 I 1 .. 1 X.. 6 1 1 .. 1 X.. 6 I 1 • 1 X.. 6 I 1 .. 1 X. 6 I 1 .. 1 X.. 6 I 1 .. 1 X..
1611 )

C TO SIMULATE FORMAT(06 .. IX .. 06 .. IX .. 06 .. 1X,06 .. IX .. 06.1X.06 .. IX ..
C 06 .. 1X.06)
C WRITE SOMETHING TO SHOW INFORMATION NEEDED
2
WRITE(4 .. 3)
3
FORMAT(/19H READ Kl K2 K3(314»
C READ IN DIMENSION INFORMATION
READ(4 .. 4) Kl .. K2 .. K3
4
fORMAT<3I4)
C ADJUST ARRAY J TO THE PROPER SIZE
CALL ADJ<J .. K.KI.K2.K3)
C READ IN ARRAY IN OCTAL
5
CALL REDAR(4 .. I .. 6 .. J)
C WRITE OUT ARRAY
WRITE(4.6) J
6
FORMAT(SI7)
C WAIT FOR tP
PAUSE
C IF A0S17-0 READ IN IDENTICAL ARRAY TYPE
IF (ITOG<l»
2 .. 5 .. 2
END

C.2

IFLOW AND IDZERO EXAMPLES

The following is a programming example of both the IFLOW and IDZERO functions.
C

C
C
2

~AIN PROGRAM TO SHOW USE OF IFLOW AND IDZERO
A=10.**70
8=10.**10
C=A*B
CALL SUBROUTINE TO CHECK FOR UNDERFLOW, OVERFLOW
AND DIVISION BY ZERO.
CALL CHECK <I)
PAUSE 1
C=(10.**(-70»*10 .. **(-20)
CALL CHECK (I)

(continued on next page)

C-5

PAUSE 2
3

C=A/0.

CALL CHECK
PAUSE 3
STOP

(})

END

C
C
C
C
C

10
11

SUBROUTINE TO CHECK FOR UNDERFLOW, OVERFLOW OR
DIVISION BY ZERO IN FLOATING POINT ARITHMETIC.
PASSING A NON-ZERO POSITIVE ARGUMENT WILL CHECK
FOR ALL. A ZERO ARGUMENT RESULTS IN NO
CHECKING.
SUBROUTINE CHECK (N)
LOGICAL IFLOW,IDZERO
If (IFLOW(N» WRITE (1,10)
IF (IFLOW(-N» WRITE (1,11)
IF (IDZERO(N»
WRITE (1,12)
FORMAT C/9H OVERFLOW)
FORMAT Cl10H IJNDERFLOt-l)
FORMAT (/13H DIV. BY ZERO)
RETURN
END

The result of running those programs is (with . DAT slot 1 assigned to the ny):

OVERFLOW
PAUSE 000001
tP

UNDERFLOW
PAUSE 000002
tP

DIV. BY ZERO
PAUSE 0V'1V'1003
,p

STOP 000000

C.3

INPUT -OUTPUT EXAMPLES

The following is a program composed mainly of I/O statements with no connected purpose. The program is presented to illustrate the possible combinations of the different types of I/O (sequential access,
direct access, data-directed, ENCODE/DECODE).

C-6

001

002
003
004

005

c
c
c
c
c

PROGRAM EXAMP~E TO SHOW OBJECT CODE OUTPUT FOR
VARIOUS TYPES OF 1/0

STATE~ENTS

IMP~ICIT REAL eN)
DIMENSION R~1(2), R~2C3), ARR(20), NM1(2), NM~(2)
DATA NMl/5HNAME1, 4HASRC/,NM2/5HNAME2, 4HASRCI

006

007
006

00603 472031 542542
00605 406472 24150~
00613 4720;'1 542~.d4
0061~

AI1I6472

0Q'9

01~

100
JMP

00000
00001
00002
00003
00004

24150~

FORMAT
!~00~0

.OSA 242226
,DSA 526216
.DSA 3~5405
.OSA ~31530

00005

.DSA 31121e.

0000~

.OSA 530544
.OSA 271445
.OSA 12450~

00007
0001~

$000e0' = 0'0011

011
00~11

00012
00013

200
FORMAT (lX,15,G10.3,2CE12.2))
JMP S~~011
,DSA 2 4 1433
.DSA e26222

.DSA 32531e
00015 ,DSA 73~540
~0016
.OSA 271465
00017 .DSA A3112~
0002~
.DSA A25426
00021 .OSA 227144
00022 .DSA '-45224
00023 .DSA ~~~1~0
500011 = 0V'!024
00~14

012
00024
00025

00026
00027

0003t?1
00031
00032

00033
00034

CA~L

JMS* DEfINE
JMP etP.!036

.DSA (0Cl0002
.DSA (0~0144
.DSA (NH'I1'l05
.OSA (QI~000~
.DSA JVB

.OSA
.OSA

(00fiJCI!00
cr00Q1Ql~

013

.OSA Ce.000rl!CiI
CAL.L

00036

JMS*

DEF INE.

00037

JMP
.OSA
.DSA

~005l"

00035

01/le4~

00041
00042

00043
00044
00045
00046

00047
014
0005~

(~00004
(~0113~

.DSA (0Q10U2
.OSA (e.?lI1IQlI'1'"
.DSA JVA
.OSA ((~~(1IQl0~
.DSA (000000
.DSA (C?i00'-'0f!'
CA~L

SEEK

(S,NH1)

JMS* SEEK

C-7

00051
00052

JMP ~0~!54
.DSA ce00~05

~0053

.05A

015
016
017

018
019
00054
00055

00056
00057

C
C

00061
00Ql52

00063

ee064
00065
00066
0"'067
U070
Ci'l0071

pINARY
A) DIRECT ACCESS

JNS'" P !TER

JMP 0'HHiCJ
.05A (~00r015

.DSA

020
0006e

+NMl
CAL.L ENTER C6,N M2)

lV!~0"'0

10V'l0N! +NM2
READ (2jfJVB)

LAC JV8
JMSt .RS
.DSA (0(110"'02
JMS* .RJ
.DSA INT
.05A 777776

TAD

TAD
TAO

INT, RL.2(3), RLl
-"

(e~00~3
(Vqil0CJ1~3

-..,

RL2

DAe $0e.!1!71
00"'72 JMS*
$00071 II 00073
.DSA $P.o0073
00073
0!1l07~

00075
0~076

."J

JMSt .RB
.05A 1~00~0 +RL.1
JMS* .RG

WRITE (2 I 3) INT, RL,2(3), RLl

021

00077
eIiIJ1 et"
NH~l

00102
00103
001~A

00105

00106
00107

L.AC (000003
JMS* .FX
.OSA (~e0(1102
JMS* .RJ
.OSA INT
.OSA 777775
TAO (000003

- ,

TAD (0111011'03
TAD RL2
OAe $001110
JMS* .RJ

00110
00111
$00110 • ~Qll1~
• DSA $r.0112
00112

022
023
024

0121113
00114
0011~

025

00116
e!0117
r0120

0111121
00122
00123
12'0124
00125

00126
00127

El) SEQUENTIAL ACCESS
C
C
JMS* .Ra
.OSA 1\~0000 +RL1
JM5* .RG
REAO (ll INT, RL.2(3), RI..1
JMS* .FS
.D5A (~~Ql~01
JMS* • ~ J
.1)5A tNT
.05A 771776-

TAO

(~' 00~~3

TAO (~~1!I"'03
TAO RL2
OAe $IilIH26
JMS* .FJ

--.."

$00126 • 00130
• DSA $'101 30
00130

C-8

00131
00132
1210133
025

00134
00135
00136
00137

00140
00141
00142

1210143
ft'I014.4
1210145
500144
~0146

12127
028
029

030
031
121"147
121015li1
00151
032
£10152
00153

1210154
00155

00156
00157
0016121

1/10151
1210162
00163
1210164
12101155

$00164
00166

JMS • • fB
.DSA 10Q100~ +RL1
JMS* .FG
wRITE (3) INT, RL,2(3), RL,1
JMS* .FX
.liSA (~~0ti103

JMS • • FJ
.DS4 INT
.DSA 777776
TAO (N'I~H?l03
TAO CQH"I0Q!03
TAO RL2
04C Hi0144
JMS. • F J
;: 00146

,DSA $~014e
C
C II) ASCII
A) DIREq ACCESS
C
C

1) FORMATTED

JMS • • F6

.DSA

1{,'0000

L.AC

JVA

JMS • • fG

.DSA (~~0004
.DSA .111'0
JMS • • ~E
.DSA INT
.OSA 77777~
TAO ce~0!(l1il3
TAO (000(111213
TAO RL2
OAC selHIS4
JMS • • RE

• "0166

• DSA $0~166

JMS • • RA

00171

JM$* .RF

00172
001n
~017A

00175
e0176

fi'l0177
00200
~1lI201

r0202
N!l203

0021114
1/1020~

S002V04
~0206

034
035
036

00207
00210

READ (4#JVA,1P10) INT, RL.2(J), RL.1

JMS • • RR

00167
0017('

033

+RL,1

.DSA H~0000 +RL.l

LAC

wRITE (4'5,200) INT, RL2(3), RL.1

C000f'05

JMS. .RW
.DSA (000"'''A
.DSA .200

Jfo1S • • RE
.OSA INT
.DSA 777776
TAD (000e03
TAD (~00f11fr33
TAO RL..2
OAC !(I!~204
JMS. .RE
I:

0~:e06
$~02~tl

.DSA
C

2) DATA .. DIRECTED
C
C
J t1 $ '* .RA

.DSA

1~1?00e

+RL.1

C-9

1'.10211

Jr-iS* .RF

~37

00212
01'.1213
0021.4
002\:5
~I'J216

00217
00220
~0221

RE,4D (4'7,)

LAC (h!l0Q\07
JMS* .RR
.OSA (0000rt4

TAD

(0(;'011.103

~0223

RL2
OAC Het22A
JMS • • GD

00227
00230
1'.10231
038

00232

(r ~ 011103

TAD
:II

e~226

.DSA $~0226
JMS. .GE
.OSA 1 ('I0 0Q!0 +RL1
JMS* .RF
loI~nE

LAC

021236
00237
~0240

.05,4 NHII0Q''''

0023A
'-'10235

(4NS,)

INT,

RI..2 (3) ,

RL.l

( ~) 00010

JMS • • Rill
.OSA (e ~ 0004
.OSA M'I ~ 0C!!0
JMS • • GA
.OSA e35204

~0233

Rl.,l

.DSA e\00"'!?1[1

TAD

00226

RL.~(3),

JI1S. .GO
.DS,4 I~T
.05,4 777776

00222
00224
00225
$00224

INT,

.--

.OSA PT
JMS· .55
Q.\0243
.OSA R~. 2
00244 LAC ( Q'. ~17JC'-03
1111'.1245 OAC
$k' 1il245
00246 JMS* .GC
00247 .DSA ~71177
Ql02!5~
.OSA M 0 000
$0et245 III 00251
00251 .OSA $"0251
C1!0241
e.0242

2139
04~

C
C

042

C
C

00252

JMS. .(;8

041
~0253

00254
00255
0f1l256

043
00257
0026\?!

00261
00262
00253

00264
00265
00266

00267
00270
00271

SEQUENTIAL. ACCESS
1) FORM,4TTED
'~"';"""

,DSA e71176
,DSA 0~000 0
.DSA 1001l1e.0 +RLl
JMS* .RF
REAO (5,100) INT, Rl2(3), RL.l
JMS. .fR

.OSA
.DSA

( 0 00~05

• 1'11£11

JMS • .FE

• OSA INT
.DSA 777776

TAO
TAO
TAD

~L2

OAC

$00270

(~~ti'l0f!J3

(0 t11 0et03

JMS· .FE
S00270 11 0121272

{1!02n

.DSA $r0272

00273

JMS. .FA

00274

.OSA

1~00"'0

+RLl

C-10

. ~

00275
0044
00276

JHs* .Ff

00277

.DsA (1Z00006
.DSA .21210

00300

JMS* .fiol

WRITE

003",1

JMS • • FE

00302
1110303
00304
00305
00306

.DSA I lilT
.USA 77777e
TAD C~00003

003~7

00310
$003~7

00311
00312
00313
00314
045
021315
00316
00317
00320
00321
00322
00323
00324
00325
00326
00327
00330
$00327
00331
00332
0033~

00334
04tl

00335
00336
~0337

00340
00341
00342
00343
00344
0121345
0121346
~0347

~035'"

TAD

CfS, 200)

INT, RL.2(3), ~\"1

(0001:"03
RL2

oAC $0i!l3~7
JMS· .fE
• 00311
.DSA 5"0311
JMS. .FA
.DSA 10\11000 +R\"1

JMS* .FF

JHS* .Gf

ENCODE (U,ARR,100) INT, RL2(3), RLl

.DSA (0~0012
.DSA 10f.i1l.l0k1, +ARR
.OSA .11210
,JMS. .fE
.OSA INT
.OSA 777776
TAD

TAO

C0\210Q103
(000U3

TAD RL2
OAC $00327
JMS* .FE
"'0331
.DSA $1210331
JMS* • F .A
.DSA le~000 +RL,l
JM5* .FF
DECODE C10,ARR,1III0) INT, RL2(3), FILl
JMS* .GG
.DSA (017.!0012
.DSA li2'00Q'0 +ARR

•

.DS4 .1f1l0

JMS. .FE

.OSA INT
.O$A '177776

TAD

TAO

C~00003
C~\ ~0N"3

TAO RL2
OAe $1(10347
JMS. .FE

•

0~351
500347
00351 .OSA U;?J3!51
12147
C
2) DATA-OIRECTEO
C
12148

04g

00352

JMS* .FA
.05A 10V'I0rlll +f?L.l
JMS* .FF
READ C5, ) INT,RI.,2(3), RLl
JM5* .fR
.OSA (000"'05
.DSA QlVil?O!e'e.
JMS* .GO

00353
0121354
050
00355
00356

00357
1210361(0

C-ll

00361
0~362

1110363
L'l03~4

00365
Ql0366

00367
$0.,366
~037~

0(1.1371
00372
Ql0373

051
00374
00375
003H
00377
0040~)

00401
0041112
00403
00404
004~~
004~6

00407
0041E1
00411
S0(114~6

00412

00413
00414
00415
00416
00417
0!S2
00420

00421
00422
00423
00424
00425
00426

00427
0043e.
00431

.D5A INT
.OSA 777775
TAD (e,iiJ0f"03
TAD (V'flI0003
TAD R1.2
OAt $'~ ""36f5
JMS* .GO
II

~~37e.

.DSA H0370
JMS* .c;E
,05A le~M0 +RLl
JMS* .fF

WRITE C6,) tNT, RL2(3), RL.l

JMS* .Fw
,OSA (0"'0006
.DSA Qle.e00PJ
JMS* .GA
.OSA 035204
.D5A Ql0~000
,DSA IfIlT
JMS· .55
.OSA RL.2
LAC C000ti'~3
OAe $~04f!16
JMS. ,Ge
,05A 0'1177
,DSA 0e000~

• Ql0412
• 05A

3i~;flI412

JMS. ,GB
.DSA e.71176
• DSA 000011'0
,DSA 1000~0 +RLI
JMS· .FF
DECODE (15,ARR,) INT, RL2(3), RLl
JMS • • GG
.OSA ((1100017
.OSA lV,0011'0 +ARR
.OSA 0ftl0000
JMS* .GA
,OS. ~3S204
.DSA 11\(!I00~0
.OSA I/IIT
JMS. ,55
I.AC

~0435

.05A 071171
,DSA 0(1100021

S004~3

00437
00440

00441
00442
00443
00444

053
00445

00446
00447
00450

.OSA PL.2
(0V1P1C'!03

00432
00433
00434
00436

OAe U~433
JMS· .Ge

,'"

004;51
•,OSA

$00437

JMS • • GB
.OSA Gi!71176
.DSA 1'1 0 "Jr. 00
.DSA 1~0~~0 +RLl
JMS· • F f
ENCODE (2~,ARR,) INT, RI.2(3), Rlol
JMS* .GF
.DSA C000~31
,DSA 1 t?lCIII1lCl' +ARR
.DSA 0r.00e.0

C-12

00451
00452
00453
00454
00455
00456
00457
004tH'l
$00457
00461
054
1210462
00463
00464
055
00465
0121465
056
00467
00470
057
00471
1210472
058

00473
00474
059
00475
00476
060

00477
00500
061
00501

JMS* .GO
.OSA INT
.OSA 777775
TAP (v.'e!0Q\03
TAD (000003

TAD

RL2

DAe $\'10457
JMS* .GO
II

~ft!461

.DSA $00461
C
JMS* .GE
,DSA 1~0000 +RL.1
JMS* .Ff
E~DFILE

ENDFII..E 3

JMS* .FV
,OSA C0~~~03
ENDFIL.E 4
JI"IS* .FV
.OSA (000!H'!4
ENDFII.E 5
JMS* .fV
.OSA (000005
ENOFILE 6
JMS* .fV
.DSA (000006
END

CLA

011211503

JMP* .5T
JMS* .FP

005"'.4

J~lP

00502
00505
1210511
"'0512
0121513
00~14

00515
0121523

00!524
00!525
00526

00!527
00577
00600
00601

1210602
012161213
00607

0061(11
00611
00612
00613
1210617
006211'

1210621
00622

JMS* .rv
,OSA C01,H3001
ENOFIL.E 2
JMS* .FV
,DSA (f)~0Vi02

00~00

.6LK 1/!000Q14

.I)SA 0200~4
.DSA 0000",0
,DSA ~0tl1!IJCil~

.DSA 100011!0 +RL.l

,BLK 'H:J00~6
.OSA V\2~0~6
.OSA ~-'IiI~Qle0
.OSA 1i1000V,e.
,DSA 11i1Q101i10 +R1.2

.eLK "'QI~050

.DSA 02~050
.OSA ol/HlIQIf210
.DSA ~000fl10
.DSA H'0000 +A~R

.eL.K "H"00~4

,DSA 02 iii",,,, 4
.OSA Pl0iiHilQlQ!
.DSA 00001110
.DSA 1121001210 +NMl

.BLK PJ000P1,d

.0511 020i1l04
.D5A 000000

.DSA ~~HII0"H~
.DSA 10i'!01i1fil +NM2

C-13

00623
0121~24
0062~

00626

00627
e063~

0121531

Pl0632
1210633
00634
00153!S

.OSA DEFINE
.6LK 012!~'H~' 1
.BLoK ~f!l1!I0f!11
.OSA SEEK
.OSA ENTER
.DSA .~S

.BLK 1lIQ1~001
.OSA .RJ

00644

.RB
.OSA .RG
.OSA .RX
.DSA • F'S
.OSA .F"J
.DSA .FB
.DSA .Fe;
.OSA .FX
.DSA .RR
.DSA .RE

00645

.05.A • RA

00646

.DSA .~F
.DSA • Rill
.OSA .GO
• OSA .GE
.05A .GA
.DSA .SS
.OSA .Ge
.OSA .G8
.DSA .FR
.OSA .f E
.DSA .FA
.DSA .FF
.DSA .n~
.OSA .GF
.DSA .GG
.DSA .FY
.OSA .ST
.DSA .FP
.OSA 0000Vo2
.DSA 011J1i'144
.OSA 00~005
.DSA 0e0l'100

~063(:i

0121637
00640
1210641
1210642
0121643

001547
00651':'1
00651
~0652

1210653
00654

00655
0065el
00657
006tH'
1210661
00662
00663
00664
006155

00666
00667
Q!0670
~0671

0IiHS72

00673
00574

Pl0675
00676
~0677

007l'10

00701
00702
00703
00704
0070~

RLI
RL2

.05A

.DSA 0e1130
.DSA PI~0012
.OSA (1I000M
.DSA Ql0011l~3
.DSA 0IiH'I0~1
.DSA MH'!0G!7
.DSA Vle0U0
.DSA 00~017

.DSA ~00031

.1~0

w DEFINE

JVB
JVA
w SEEK

• os" ~("~'H'IA

ARR
NMl
NM2
.200

005~5

00515

00527
0061?3
0~613
000~~
~0011
~0fi23

00tJ2A
00625
006215

- - "'"

C-14

• ENTER
.RS
* INT
* .RJ
.RS
* .RG

• .R)(
• .FS
•* .FJ
• FE)
* .FG
•
• .F)(
• .RR
* .RE
.RA
* .RF
* .Rw

•

.GO
• .GE

•* .GA
.SS

00627

00630
00631
00632
00e33
006341
00635
00636
00637
0fi.H5~~'

006.11

00642
00643
00644
00645
00646
00647
~06!50

00651
00t;i!'52

• ,Ge
• .GB
• .FR
• ,FE
• .FA

00653
006!54
00655

.FW
• .GF
* .GG
• .FV
• .ST
• .FP

00662

•... .fF

IH'I656
0vH5~7

00660
00661
00663
006604
00665
11J0666
00667

C-15

APPENDIX D
SYSTEM LIBRARIES
D.1

.LIBR - Page Mode Non-FPP

LIBRARV FIlE LISTING FOR .l.IBR
PROGRAM
NA"IE
RBCPIO
R8It-:IO

SOURCE
EXTf.~.SI(lN

~06
(~?5

RANCOI-1
DEFINE

('11

001(1

r112

EDCODE
EOF
UNIT
JABS
JDFIX
JFIX
Fl.OATJ
JOSLE
ISNGl.
JSIGN
JDIM
JMOO
JMN"'X

~t2
0~'0

ERR SET

IOERR
FILE

0~19

t:Wl
Cle.1
0~1
~Hq

et01
I7W2
"f' 3
0~1
~~1

(!lIP
" ~:0
~W2

epa

TIME
TIME. 10

r.H~9

AOJl
ADJ2
AOJ3

"'V1(l
.WV';

~f8

tH:'~

ASS

~lVj 2

lABS
DARS
AINT

N'0

M'l
~1 r. 2

INT

(,[,q

1DINT

005

AMOD
MOD
OMOC'

~~ ~\

PROGRAI-1
SIZE

255
30
56
15
13
13
13
10
30
23
21
23
1~3

25
40
376
45
72
17
36
41

16
14
16
15
13
lJ
27
24

FLOAT
IFIX

e.V:2

~C'4

31
31

DIM

~Hl't(\

!H'l

JD1I-1

\N'~

SNC~L

0~4

DSLE

~¥' 1

IMN~lX

~5P

~M'IIMX

~ep

ACTION

136
113
504
1130
2037

li'tl.
N:4
QlV12

SIGN
DSIGN
ISIGN

PAGE 1

20
22
15
27

11
1 ;l7
120

D-1

.~ ,

LIBRARY FILE lISTING FOR .LlBR
PROGRAM

SOURCE

PROGRAM

NAME

EXTEl-iSION

SIZE

OMNMX

(HIP

1V16

.88

V,'1;'4

6'2'
132
132

.BC

~Q19

.60

~e9

.BE

.BF
.BG
.BH

.BI

~05

33
34

008

'2'~5
0~3

34
120
73
13
20

~~6

SQRT

0Q'S

SI'"

~V.3

ATAN
ATA.N2
EXP

~03
0~2
0~7
0~' 2

COS

AL.OG
AL.OG10

"'~2
0~2

20
20

1~2

~~5

67
71

.EF

Qlb4

.Et
DSQRT
DSIN
DtOS

0el

116
44
71

.ED
.Ef

OAHN
OATAN2

OEXP
DL.OG
OL.OG10

r.e,2
M'7

0el
(1101

IHl1
QlI1'7
0~1

ee.3

13
21
13
46
13
21
21
16
30

IFLOW

I? ~11
('Ie!
M :l
QH'l

.00

0~5

146

~Ql4

120

\ll03
Qle1
eo~, 1

IOZERO
ISENSI'I

.DB
.DE

.OF
.DC

137
47
56

.DA

pre

eCOIO
BINIO

033
015

AUXIO

ru'

363
133

.55

e05
11'03
€e3

GOTD
STOP

~04

.E8

ACTION

44
13

2104

TANH

PAGE tl

3724

,.-...,

50
26

D-2

LIBRARV FILE LISTING FOR .LIBR
PROGRAM
NAME

PAUSE
SPMSG
.FLTB
FlOPS
PARTIIID
DBLlNT
INTEAE
DOUBLE
RELfAE
OrSER
.CB

SOURCE

E)(TEt-SION

PAGE :5

PRQGRAM
SIZE

ee5

0e,4
0~4

73
266

e17

735

~3P

140

e.lF

377

e.71=>

0r-4

203

1l'P

1077

C'l~; 9

210

~e,4

ACTION

CI.OSE

0-3

D.2 .LIBRF - Page Iv'ode FPP
\.IBRARY FII.E LISTING FOR .LIBRF
PROGR4M
NAME

SOU~CE

PROGRAM

EXTENSION

SIZE

RBCOIO
RBINIO
RANCOM
DEFINE

~e· 5

136

r.05
rli'9

5;114

0010

F12

2012

ED CODE
fOF

0P12

2~5

0~~

UNIT

0~1

30
66
14
12
12

JABS
JDF"IX
JFIX
FL.OATJ
JD8LE
ISNGL
JSIGN
JDIM
JMOv
JMNM)/
fRRSET
IOEPR
FIL.E
TIME
TIME10

AOJ!
AOJ2
ADJ3
AaS
UBS

DABS
AINT
INT
lOINT
AMOD
MOO

OHOD
FL.OAT
IFl)(
SIGN
OSIGN
ISIGN
DIM
101M
SNGI.
DaL.E
IMNMX
RMNMX

011

n)l

F01
Fe1
F'~1

F01

Fe2
F~i3

Fe· I
F~l

F1P
00QJ
0~2
~e:8

009
!!!08
~~0
~~0

QI!l:l1!

F(iJ2
'!.!rl?-

113Q1

100
25
40
376
45
72
17
3(5

Fe-A
IH1'2

FeA
e0f6

Filii

17
15
16
10

F04

~e~~

Ff'4
FI11
e5P
F8P

23
11
12
24
24

Frtl2

-~,

41
13
14
14
12
12
23
2.

FfIl5

.-~

1121
10
13
16
17
17

F~l

F03
"'00

ACTION

113

FC2

Fe2

PAGE 1

1l'i1

115

D-4

I.IBRARY FII..E I..ISTING FOR .I.I6RF
SOURCE

PROGRAM

EXTENSIOt-;

SIZE

DMNt-,X
,68

r8P

,BC

n 'q

Fe.7

1:114
60
126
126
30
31
31
31
113
73
12
16
t2
35

PROGRAM
NAME

.60

.BE
.SF

.BG
.8101

;.\~) .4

Hl9
F06

Fes

F~8

n"5

.BI

r:e3

SQRT

Fe,R

SIN

Fe3

ATAN
ATAN2
EXP

F'P2
F~'2

AL.OG

F~2

ALOG10

Ht2

16
16

UNH

.EB
,ED

Fe·4
F~' 4
Ff?l5

.E~

Ffi.I~

cos

F03

, E~'

F04

,Ee

n: 1

DSQRT

F~7

OSH~

FU
Fe' 1

I)ATAN

F~1 1
F07

Dens

OAT/lN2

OEXP

~6
6~

72
111
40
70
12
17
12
42
1:2

FV'1
F03

DLOG1~

F~;

IDZERO

iAe 1

ISENSW
IFLOW

NIl

VIe 1

,DO
.OB
.OE.
,OF

.DC
.DA

17
17
16

F c.~ 5

137

Fe:4

115

F~'3

\1,14

FO
F0\

1:~0
43

PP5

AUXIO
ISS

F33
e15
eli/!
0Ul

GOTO
STOP

f?I~'3

3634
363
133
60
26

'103

BCOID
eINIO

ACTION

OLOG

PAGE 2

D-5

L.IBRARY fILE LISTING FOR .L.IBRF

pROGRAM
SIZE

PROG~AM

SCI"iRCE

NAME

EXTHSION

PAUSE

",U5

SP~\SG

l&'4
,;, e, d

13
266
715
145
131
41i'17
2U
22

.FL.TB
flOPS
PARTWC
INTFAE

.FPP
OTSFR

.ca

"' 17
F3P
(.li7P

F12

o~; 9

N' 4

PAGE ~

ACTION

D-6

APPENDIX E
PDP-15 FORTRAN FACILITIES

The extended FORTRAN language described in this manual and in the companion manual (Operating
Environmental Manual DEC-15-GFZA-D) is available only on the systems described below. The
FORTRAN existing on otherPDP-15 systems is described in a manual entitled "PDP-15 FORTRAN IV
Programmer's Reference Manual" (DEC-15-KFZB-D).
The following tables describe the existing versions of the extended compiler, the extended Object
Time System Libraries, and the compiler-library pairs available for different systems. All versions of
the compiler are written in PDP-9 code, however, 'PDP-9 mode' versions produce only PDP-9 code as
output while 'PDP-15 mode' versions may produce PDP-15 instructions where suitable. Page and Bank
Node libraries differ not only in the use of the PDP-15 versus PDP-9 code, but also in the values of
address masking constants used in a few of the routines. Note that the Floating Point Processor (FPP)
is supported only on the PDP-15, thus there is no PDP-9 mode version.
The library names used in the following tables are given for designational purposes within this appendix
only and do not necessarily reflect the names under which the libraries are distributed.

Table E-l
Versions of the Extended Campi I er

Main
Version

F4X

Features

All

Version

e
4X

F4RX

All except
direct-access I/O
All except
direct-access I/O

Approx.
Size (8)

Non-FPP, PDP-15 mode D05-15
Non-FPP, PDP-9 mode DOS-15
FPP, PDP-15 mode D05-15

15406
15363
15661

< F4B9
\...FPF4B

Non-FPP, PDP-15 mode, ADSS (V5B)
Non-FPP, PDP-9 mode ADSS (V5B)
FPP, PDP-15 mode ADSS (V5B)

15251
15226
15522

{F4RX
FPF4RX

Non-FPP, PDP-15 mode RSX
FPP, PDP-15 mode RSX

F4X9
FPF4X

F4B

System

r F4B

E-l

Table E-2
Versions of the OTS Libraries for the Extended Compiler
System

Contents

Libraries

Subsystem

005-15 (BOSS-15)

Contains all routines, assembled for DOS-15
operation.

r
r

Non- FPP, Page
Non- FPP, Bank
FPP, Page
FPP, Bank

ADSS

Contains all routines except direct-access
(DEFINE, RANCOM, RBINIO I RBCDIO)
assembled for ADSS operation.
Contains 011 routines except direct-access
(DEANE, RANCOM, RBINIO I RBCDIO)
ond magtape subroutines (UNIT, EO F) ,assembled for RSX operation and includes
added routines applicable to RSX only.

RSX

BXP

.LBXB
.LBXPF
.LBXBF

BRP

.LBRB
.LBRPF
.LBRBF

{LIB~

.LIBFX

Non- FPP, Page
Non- FPP, Bank
FPP, Page
FPP I Bank
Non- FPP I Pagel
Bank
FPP I Page/Bank

Table E-3
Compilers and Libraries for Extended FORTRAN
Distributed with PDP-9/15 Systems
FPP

Non-FPP

System

Page

Bank

Page

Bank

DOS-15

Compiler

F4X

F4X or F4X9

FPF4X

(BOSS-15

Library

.LBXP

.LBXB

.LBXPF

.LBXBF

ADSS V5B

Compiler

F4B

F4B or F4B9

FPF4B

Library

.LBRP

.LBRB

.LBRPF

.LBRBF

Compiler

F4RX

FPF4RX

Library

.LIBRX

.LIBFX

RSX

E-2

INDEX
A

C (cont)

A-register, 3-13

Command string format, 1-2

Accumulators, 3-13

Command string options, 1-2

Address calculation for array elements, 4-1

Command (BACKSPACE, ENDFILE, REWIND),
2-7, 2-8

Adjustment of array dimension, C-1
ADVANCED Nonitor Software
System (ADSS), 1-6

COMMON blocks, 5-3

ALT MODE, 1-2

Compiler, 1-1

Arctangent (ATAN, DATAN, ATAN2,
DATAN2), 3-9

Control P(tP), 1-2

Arithmetic package functions, 3-1, 3-13
Arrays
data mode values, 4-2
dimension adjustment, C-1
element address, 4-1
size, 4-2
unsubscripted, 5-2

storage area, 1-3

Conversion, .ASCII to .SIXBT, 4-16
Cosine - see Sine and cosine
D

• DAT see Device assignment
Data-directed Input/Output (DDIO), 2-13, 2-14
Data storage, external, 2-2

.ASCII to .SIXBT conversion, 4-16

Data structures of peripheral devi ces, 2-2

ASSIG N command, 2-2

Data transfer

AUXIL (OTS Auxiliary Input/Output, 2-7

EDCODE (memory to memory) 2-15
FlOPS, 2-3
RANCOM, 2-13

B
Background/Foreground Nonitor System, 1-6

Data transmission, 2-1, 2-3

BACKSPACE command, 2-7, 2-8
Backward links, 2-2

DDIO data-directed input/output routines, 2-13,
2-14

Batch processing monitor (BO$$-15), 1-7

DECODE routine, 2-15

BCDIO (OTS Binary coded I/O), 2-5, -6
global entry points, 2-12
routines, 2-5
BINIO (OTS binary input/output), 2-6, 2-7

DECtape, 2-2
DEFINE routine, 2-9, 2-11
parameter table, 2-11
Device assignment, 2-2

BOS5-15 batch processing monitor, 1-7
Buffer size, OTS FlOPS package, 2-3
C

FlOPS, 2-3
Device data structure 2-2
Di rect access to formatted fil e, 2-11
READ,2-11
WRITE, 2-11

Carri age return, 1-2
CHAIN (overlay linker) 1-1, 1-3
CHAIN and EXECUTE loading, 5-3
Comma ( ,) usage, 1-2

Direct access input/output, 2-9
Directoried storage, 2-2
Directoried subroutines, 2-16

Command error messages, 1-3

1-1

INDEX (Cont)
Division by zero in RELEAE routine, 3-14

FlOPS (OTS lOPS communication, 2-1, 2-3

Dollar sign ($) usage, 1-2

routines, 2-3
status tab I e, 2-3

DOS-15
FORTRAN directoried I/O, 2-2
operating system, 1-6
sample session, 1-4, 1-5
DOUBLE function, 3-16
Double integers, 1-7

Floating accumulator, 3-13
Floating-point processor (FPP), 1-7, 1-8
routines, 4-4
Fo rmat for single (double) precision numbers, 3-14
FORMA T statements, 2-5, 2-6

Double precision floating-point arithmetic, 1-7

errors, 4-4
READ,2-5
record length, 2-5
WRITE, 2-5

Double precision number format, 3-14
DOUBLE PRECISION values, 5-3
E

EDCODE routines, 2-15
ENCODE routine, 2-15
END FILE command, 2-7, 2-8

Formatted input/output (RBCDIO) ~ 2-11
FORTRAN callable utility routines, 4-5 through
4-8
FORTRAN sequences called by MACRO, 5-2
Forward links, 2-2

Error messages
command, 1-3
FORTRAN Appendix B
OTS Append ix B

FPP see Floating-point processor
FPP F4X system, 1-7
Functions, 3-16, 3-17, 3-18

Errors, unrecoverable, 4-4
G

Examples
IFLOW and IDZERO, C-5
input/output, C-6
programming, C-1

Global entry points BCDIO, 2-12
.GLOBl pseudo operation, 3-1
H

• EXIT function, 4-4

Exponential (EXP, DEXP), 3-5
Extended integer arithmetic, 1-7
External functions, 3-1, 3-5, 3-6

Hardware, 1-7, 1-8
Header pair, 2-5
Held accumulator, 3-13

External storage, 2-2
External subroutines, 3-1

F
File access on serial devices, 2-2
FILE package, 2-16
Filename, 1-2

ID word (BINIO), 2-6
IDZERO, logical function, 3-14
IFlOW and IDZERO, programming
examples, C-5
Initialization and actual data transfer
(RANCOM), 2-13

1-2

INDEX (Cant)
Input/output

direct access, 2-9
examples, C-6
formatted (RBCDIO), 2-11
sequential, 2-4
unformatted (RBINIO), 2-12, 2-13

MACRO-15, 1-1
MACRO-FORTRAN linkages, C-1
MACRO sequences called by FORTRAN, 5-1
fv\agneti c tape, 2-2

Input/output processing
data directed I/O (DDIO), 2-13
direct access, 2-9
ENCODE/DECODE (EDCODE), 2-15
general, 2-1
OTS lOPS communication (FlOPS), 2-3
sequential, 2-4
user subroutines, 2-15
Input/output routines, fv\agtape, 2-15
INSTALL MCR (RSX function), 1-5

input/output routines, 2-15
fv\agtape tape functions simulated on disk, 2-8
Master File Directory (MFD), 2-2
Memory to memory transfers, 2-15
MFD see Master File Directory
Nbdes, array, 4-2
Nbnitor control, 1-2
Multiprogramming environment, 1-7

INT function, 3-18

INTEAE function, 3-16
INTEGER array size, 4-2
Intrinsi c functions, 3-1, 3-2

Natural and common logarithms
(ALOG, ALOG10, DLOG, DLOG10), 3-7
Number formats, single/double precision, 3-14

IOERR(N} integer function (FlOPS), 2-4

o
L
Language summary, Appendix A
Left arrow (-) usage, 1-2
Li brari es, System, D-1
.LIBR, D-1
.LIBRF, D-4
Linkage
MACRO-FORTRAN, C-1
program, 1-1
Linking loader, 1-1, 1-3
LINKS, 1-3
Links, backward/forward, 2-2
Loading FORTRAN IV, 1-2
Logarithm, Base 2 (. EE, • DE) subfunction, 3-10

Operati ng procedures, 1-1
OTS arithmetic routines, 3-13
OTS Auxiliary input/output (AUXIO), 2-7
OTS binary coded input/output (BCDIO), 2-5,
2-6
OTS binary input/output (BINIO), 2-6, 2-7
OTS error messages, Appendix B
OTS lOPS communication (FlOPS)
Buffer size, 2-3
routines, 2-3
status table, 2-3
OTS routines, 4-1 through 4-4, 4-15, 4-16
di rect access, 2-9
floating point processor, 4-4, 4-5
FORTRAN callable utility, 4-5 through 4-8
RSX library, 4-9 through 4-14

Logarithms, natural and common
(ALOG, ALOGlO, DLOG, DLOG10), 3-7

Output listing, 1-2

Logical function IDZERO, 3-14

Overlay linkage editors, 1-1

Logical record size unformatted statements, 2-6

Overlaying of LINKS, 1-3

Overflow, 3-14

1-3

INDEX (Cont)

R (cont)

Paper-tape source file, 1-2

RSX-15 real-time execution, 1-7

PDP-15/30 Background/Foreground tv'onitor
System, 1-6

RSX library (.LIBRX BIN) routines, 4-5,
4-9 through 4-15

Polynomial evaluator (. EC, .DC)
subfunction, 3-13

5
Sample DOS-15 session, 1-4, 1-5

Program linkage, 1-1

Science library, 3-1

examples, C-1
Pseudo-operation, .GLOBL, 3-1

Sequential file storage, 2-2
Sequential I/o, 2-4

R
RANCOM (initialization and actual data
transfer), 2-13
RBCDIO, formatted input/output, 2-11
RBI NI 0, unformatted i nput/o utput, 2-12, 2-13
READ statement
formatted, 2-5
formatted direct access, 2-11
unformatted, 2-6

Serial file storage, 2-2
Sine and cosine (SIN, COS, DSIN, DCOS), 3-8
Single integer arithmetic, 1-7
Single precision number format, 3-14
Single precision floating point arithmetic, 1-7
Software environments
ADVANCED tv'onitor (ADSS), 1-6
BOSS-15, 1-7
DOS-15, 1-6
PDP-15/30 B/F tv'onitor, 1-6
RSX-15, 1-7

REAL array size, 4-2
REAL values, 5-3
Real-time execution, see RSX-15

Square root (SQRT, DSQRT), 3-5

Record identification number, 2-6

Statements
READ, 2-5, 2-6
WRITE, 2-5, 2-6

Record length, formatted records, 2-5
RELEAE, REAL arithmetic package, 1-7, 3-14,
3-17, 3-18

Storage, external, 2-2
directoried mode, 2-2
sequential files, 2-2
serial mode, 2-2

Relocation of program, 1-3
Restart FORTRAN IV, 1-2
Retrieval information block (RIB), 2-2

Subfunctions in FORTRAN library ,

REWIND command, 2-7, 2-8

logarithm, base '2 (.EE, .DE), 3-10
polynomial evaluator (.EC, .DC), 3-13

RIB see Retrieval information block
Right angle bracket (» usage, 1-2

Subprograms, science library, 3-1

Routines, MACRO-15, 1-1

System generation, 2-2

Routines, OTS, 1-1,4-1 through 4-4, 4-14,
4-16

System Ii brari es, D-l

floating point processor, 4-4, 4-5
FORTRAN callable utility, 4-5 through 4-8
RSX library (.LlBRX BIN), 4-5,4-9
through 4-15

TKB (task builder), 1-1, 1-5
fil ename, 1-5
Time sharing, 1-6

1-4

INDEX (Cont)
U
UFO see User File Directory

Utility routines, 4-1 through 4-16

Underflow, 3-14

FORTRAN callable utility, 4-5, 4-9 through
4-14
FPP,4-4
OTS, 4-1
RSX library (.LIBRX BIN), 4-5, 4-15

Unformatted input/output (RBINIO), 2-12, 2-13
Unformatted statements, 2-6

READ, 2-6
WRITE, 2-6
Unsubscri pted array nome arguments, 5-2
User file directory (UFO), 2-2

W
Word pairs, 2-5
WRITE statement

User subroutines, input/output

formatted, 2-5
'formatted direct access, 2-11
unformatted, 2-6

directoried subroutines, 2-16
magtape I/o, 2-15
operations, 2-15

1-5

HOW TO OBTAIN SOFTWARE INFORMATION
Announcements for new and revised software, as well as programming notes,
software problems, and documentation corrections are published by Software
Information Service in the following newsletters.
Digital Software News for the PDP-8 & PDP-12
Digital Software News for the PDP-II
Digital Software News for the PDP-9/15 Family
These newsletters contain information applicable to software available from
DigitalIs Program library, Articles in Digital Software News update the
cumulative Software Performance Summary which is contained in each basic
kit of system software for new computers. To assure that the monthly Digital
Software News is sent to the appropriate software contact at your installation,
please check with the Software Specialist or Sales Engineer at your nearest
Digital office.
Questions or problems concerning DigitalIs Software should be reported to
the Software Specialist. In cases where no Software Specialist is available,
please send a Software Performance Report form with details of the problem to:
Software Information Service
Digital Equipment Corporation
146 Main Street, Bldg. 3-5
Maynard, Massachusetts 01754
These forms which are provided in the software kit should be fully filled out
and accompanied by teletype output as well as listings or tapes of the user
program to facilitate a complete investigation. An answer will be sent to the
individual and appropriate topics of general interest will be printed in the
newsletter.
Orders for new and revised software and manuals, additional Software Performance Report forms, and software price lists should be directed to the
nearest Digital Field office or representative. U.S.A. customers may order
directly from the Program library in Maynard. When ordering, include the
code number and a brief description of the software requested.
Digital Equipment Computer Users Society (DECUS) maintains a user library
and publishes a catalog of programs as well as the DECUSCOPE magazine
for its members and non-members who request it. For further information
please write to:
DECUS
Digital Equipment Corporation
146 Main Street, Bldg. 3-5
Maynard, Massachusetts 01754

- - - - - - -

- - -

- - - - - - - - - -

- - - -

Fold Here - - - - - - - - - - - - - - - - - - - -

- - Do Not Tear - Fold Here and Staple -

- - - - -

- -

FIRST CLASS
PERMIT NO. 33
MAYNARD. MASS.
BUSINESS REPLY MAIL
NO POSTAGE STAMP NECESSARY IF MAILED IN THE UNITED STATI-S

Postage will be paid by:

mamaoma
Digital Equipment Corporation
Software Information Services
146 Main Street, Bldg. 3-5
Maynard, Massachusetts 01754

READER'S

COMMENTS

Digital Equipment Corporation maintains a continuous effort to improve the quality and usefulness
of its publications. To do this effectively we need user feedback -- your critical evaluation of
this manual.
Please comment on this manual's completeness, accuracy. organization, usability and readability.

Did you find errors in this manual?

If so, specify by page.

How can this manual be improved?

Other comments?

Please state your position . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Date: _ _ _ _ _ __
Name:

----------------------- Organization: ------------------

Street: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Department: _ _ _ _ _ _ _ _ _ _ _ _ __
City: _ _ _ _ _ _ _ _ _ _ _ State: _ _ _ _ _ _ _ _ _ _ Zip or Country________

- - - - - - -

- - -

- - - -

Fold Here - - - - - - - - - - - - - - - - - - - -

- - - - - - - - - -

- - Do Not Tear - Fold Here and Staple -

- - -

- -

FIRST CLASS
PERMIT NO. 33
MA YNARD, MASS.
BUSINESS REPLY MAil
NO POSTAGE STAMP NECESSARY IF MAILED IN THE UNITED STATl'S

Postage will be paid by:

mamaama
Digital Equipment Corporation
Software Information Services
146 Main Street, Bldg. 3-5
Maynard, Massachusetts 01754

•