Digital PDFs

DEC-08-CMAA-D

December 2000

78 pages

Original

1.9MB

Document:	MACRO-8 Programming Manual
Order Number:	DEC-08-CMAA-D
Revision:
Pages:	78
Original Filename:	https://svn.so-much-stuff.com/svn/trunk/pdp8/src/dec/dec-08-cma/dec-08-cmaa-d.pdf

OCR Text

MACRO-8

PROGRAMMING
MANUAL

PUP—8

DIGITAL

EQUIPMENT

CORPORATION

MAYNARD,

MASSACHUSETTS

PDP—8

PROGRAMMING

MANUAL

MACRO-8

For additional

copies order No. DEC-08-CMAA—D from Program Library, Digital Equipment
Price $2.00

Corporation, Maynard, Mass.
DIGITAL EQUIPMENT CORPORATION

MAYNARD. MASSACHUSETTS

ist Printing October 1965
2nd Printing Revised November I966
3rd Printing October 1967

4th Printing February 1969

1967
i969

instruction times, operating speeds and the like are included in this manual for reference only; they are not to
be taken as specifications.

The following are registered trademarks of Digital

Equipment Corporation, Maynard, Massachusetts:
DEC

PDP

FLIP CHIP

FOCAL
COMPUTER LAB

DIGITAL

PREFACE

The PDP-8 comes to the user complete with an extensive selection of system programs
and routines making the full data processing capability of the new computer immedi-

ately available to each user, eliminating many commonly experienced initial programming delays.
The PDP—8 programming system takes advantage of the many man-years of program

development and field testing by PDP-5 users. Although in many cases PDP-8 programs
originated as PDP-5 programs, all utility and functional program documentation is issued
in a new, recursive format introduced with the PDP—8.

Programs written by users of

either the PDP—5 or the PDP-8 and submitted to the users' library (DECUS

—

Digital

Equipment Corporation Users' Society) are immediately available to PDP-8 users.

Con-

sequently, users of either computer can take immediate advantage of the continuing
program developments for the other.

The MACRO-8 Manual is divided into two parts:

Language.

Basic Information and the MACRO

Part 1 is especially for the new user, while those experienced with assembly

programs will want to start with Part 2.

iii

CONTENTS
PART I

BASIC INFORMATION

Chapter

Page

INTRODUCTION

THE BINARY SYSTEM

THE PDP-8 INSTRUCTION SET
Instructions

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

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

3-]

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

3-]

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

3-I

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

3-3

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

3-6

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

3-8

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

3-9

Augmented Instructions
The Organization of Memory

Complement Arithmetic
Addition

Subtraction

ORGANIZATION

Coding

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

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

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

3-9

4—1

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

4-2

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

4-4

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

Comments

2—1

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

Memory Reference Instructions

I-I

Address Tags

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

Optimum Use
Subroutines

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

4-5

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

4-6

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

4—6

Symbolic Addressing
Storage Techniques

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

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

4—7

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

4 -9

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

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

........

PART 2
THE MACRO LANGUAGE
5

MACRO-8 PROGRAMMING LANGUAGE

Characters

.......

5-2

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

5-5

Integers

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

Symbols

...........

Expressions

5-I

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

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

Elements

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

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

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

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

......

........

5—5

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

5-5

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

5-7

CONTENTS (continued)

Chapter

Page
........................................

5—10

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

5—10

Current Address Indicator

5 (cont)

Origin Setting
Literals

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

Single Character Text Facility

PSEUDO-INSTRUCTIONS

Radix Control

Numbers

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

6-1

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

6-I

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

6-2

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

6-2

Double Precision Integers

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

6—2

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

6—3

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

6—3

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

6—4

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

6-4

Floating Point Constants
Text Facility
End of Program
End of Tape

Alterations to the Symbol Table

MACROS

5-12

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

Current Location Counter
Extended Memory

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

5-11

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

Restrictions

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

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

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

ERROR DIAGNOSTICS
Error Messages

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

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

OPERATING INSTRUCTIONS

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

Symbol Table Modification

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

Appendix
MACRO-8 SYMBOL TABLE
ASCII CHARACTER SET

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

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

6-4

HLUSTRNHONS

Memory Reference Instruction Format
IOT Instruction Format

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

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

Group I Operate Microinstruction Format

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

Group 2 Operate Microinstruction Format

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

Flow Chart of Program to Calculate Sum of Integers

Program Example

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

......

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

TABLES
Table
2-I

First Sixteen Integers in Three Number Systems

3-I

Memory Reference Instructions

3-2

Group I Operating Microinstructions

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

3-3

Group 2 Operating Microinstructions

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

3-4

Effective Address Calculation

3-5

One's and Two's Complement Representations

9-]

Switch Options

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

2—]

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

3-2

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

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

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

VII

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

3—4
3—6
3—7

3-8
9—2

PART 1

BASIC

INFORMATION

CHAPTER

INTRODUCTION

In describing the solution of the Following equation,

y=mx+b
one can

say,

"First, multiply the quantity x by m. To this product, add the quantity b. The result is the

value of y."

The same problem can be solved on an adding machine in the following steps:
l

Clear the keyboard and registers.

Enter the value of m and press the ADD key.

Enter the value of x and press the keys which initiate the multiplication.

Enter the value of b and press the ADD key.

The result, appearing in the totalizing register,

is the value of y.

This sequence of steps can be thought of as a program for the solution of the problem on an adding machine.
in similar fashion, the steps can be written out for a solution to be performed by a digital computer.

in-

stead of pressing buttons and keys, the programmer writes a sequence of instructions to perform operations
on

data stored in the computer.

Such a sequence (for a hypothetical computer) might appear as follows:

clear
add m

multiply x
add b

deposit y
The variables m, x, and b represent quantities stored in the computer; the variable y represents a storage

The operations are carried out in a register called the accumulator, abbreviated AC.

instruction clears the AC.

The next adds the quantity m into the AC.

The third instruction multiplies

the contents of the AC by the quantity x, and the fourth adds the quantity b to the result.
contents of the AC are deposited into a storage

The first

Finally, the

To be useful to a computer, the instructions of a written program must be translated into a sequence of
numeric codes, each of which represents a specific computer operation.

1-]

To do such translating by hand

from instructions to binary numbers would be tedious and lengthy.
to perform

the translating task,

Computer programs have been written

interpreting the written program and producing from it a second program

which can be executed directly by the computer.

These translators are called assemblers because they

assemble from a source program written by the user, a working program instruction by instruction.

This

output from the assembler is called an object, or binary program.
MACRO-8 is an assembler designed to accept input in the form of a sequence of symbolic instructions

representing the operations capable of being executed on the PDP-8.
which may then be placed in the computer and executed.

It produces a binary program tape

The next chapter explains a few basics of the

binary numbering system; succeeding chapters deal in more detail with the MACRO—8 assembly program.

1-2

CHAPTER

THE

BINARY

SYSTEM

Every item of information stored in or processed by a digital computer is encoded as a binary number.

Consequently, the user should become familiar with the binary system and be able to convert numbers
from binary to decimal representation and back,
as

using the octal system as an intermediate.

Eventually,

the programmer gains more experience, he will find himself using the decimal system less and "thinking

in octal"

more.

This is a useful habit to cultivate.

Table 2—1 gives the first 16 integers (and 0) as they are represented in the decimal, binary, and octal
number systems.

Note that in the decimal system there are ten different symbols, or digits, 0—9 which are
In the binary system there are two, 0 and

1; in the octal system, eight,

The number of digits required in a given system is called the radix.

Therefore, the decimal radix

used to represent any number.

0-7.
is

10; the octal radix is 8, and the binary radix is 2. A subscript is used to identify the radix of the sys-

tem in which

the number is represented.

and

1101012

Thus, 4-19610 indicates a decimal number, 25478 an octal number,

binary number.

TABLE 2-1

FIRST SIXTEEN INTEGERS IN THREE NUMBER SYSTEMS

Octal

Binary

1001

1010

13
14

1011
1100

Decimal

Octal

000

001
010

011

Binary

100

1101

101

1110

110

1111

111

10000

1000

All numbering systems using radices involve positional notation; that is, each successive digit position to
the left represents the next higher power of the radix.

For example, the decimal number

expressed algebraically as

4x103+1x102+9x10l +6x 100

2—1

419610 may be

which when calculated becomes

4x 1000+lx100+9x 10+6x l

=4l9610.

Positionally, this appears as

102

TO]

TOO

1000

100

(Radix to its respective powers)

(Units required of each value)

103

(Positioned radix)

Likewise, the octal number 25478 can be expressed as 2 x 83 + 5 x 82 + 4 x 81 + 7 x 80.
In all systems the integral and fractional parts of a number are separated by a radix point.

Depending on

the system in use, this may be a decimal point, octal point, or binary point.
As can be seen from the table, four binary digit positions are required to represent the decimal integers

The octal integers up to 7 require only three binary positions; furthermore, exactly three positions

up to 9.
are

needed.

octal digits.

In other words, three binary digit positions are necessary and sufficient to represent the eight

This fact makes binary-to-octal and octal—to-binary conversions quite simple.

Example: Binary-to-Octal Conversion
The binary integer

1011100110102
can

be converted to its octal equivalent as follows:

Divide the binary digits into groups of three, starting from the right
101 110 011010

Substitute for each group its octal equivalent
5

The result is the octal number

56328
To perform the reverse conversion, substitute the binary equivalent of each octal digit

47518 =iOOiii 101001 =1001111010012
2-2

To convert an octal fraction, group by threes in each direction away from the radix point

1110101

.0100012 =1 HO 101

2243.5578 =010010 100011

010 01(0)

=1c’>5.228

101 101 iii

=010010100011.1011011112

Example: Decimal-to-Binary Conversion
The "remainder method" may be used to convert a decimal number to a binary number.

The decimal

number

may be converted as Follows:

NNM

(binary radix)—92

Odo—doflaflofl

\lmfl-PN-d—'01NO~OJ-—'

N N w—Iat
sis-IAN
CO

N \l O

+— (remainders)

1011100110102

Notice that this method simply involves halving the number to be converted and noting the remainder after

each division.

Example:

Decimal-to-Octal Conversion

The remainder method may also be used for octal conversions.

The decimal number

297010
would be converted as follows:
0
8 T—

O~U1

8 T4—
8

(octal radix)—D8

37]

2970

_- 0

NQJO~UI

’x

4-- (remainders)

56328

ln one sense, the conversion of a number from one representation to another is a way of encoding the

number; the octal integer 778 can be encoded as the binary integer llllll
a

binary code to any symbol, such as a letter of the alphabet.

Similarly, one can assign

The table in Appendix B shows the binary

codes assigned to all the characters of the Model ASR-33 Teletypewriter.

A programmer may invent a symbolic name to refer to the location of a given word in the computer mem-

These symbolic names, or tags, are assembled together with the instruction mnemonic into a binary

ory.

number which indicates the memory location of the word, the instruction code, and the address of the
data

The association of binary code and symbol is the basis of a programming language.

A programmer learns

the symbols for the computer's repertoire of operations and the rules for arranging a sequence of symbolic
instructions in a useful format.
paper tape.

An assembly program accepts this source program input and translates it into an equivalent

sequence of binary numbers,
on an

He prepares a symbolic program for input on a medium such as punched

producing a program which is usable directly by the computer by placing it

output medium, which again may be punched tape.

be read into the computer and executed.

2-4

This binary, or object program tape may then

CHAPTER

THE

PDP-8

INSTRUCTION

SET

INSTRUCTIONS

Every PDP—8 operation is specified by a unique combination of 1's and 0's stored in the twelve bits of one
memory register.

Such an instruction word can be one of two types:

memory reference instructions per-

form operations which require access to the information stored in a memory register; augmented instructions
do not refer to memory cells.

The operation code of an instruction is contained in bits 0, l, and 2 of the word.

Since three binary digits

correspond to one octal digit, it is apparent that there can be no more than eight operation codes, corresponding to the octal digits 0-7.

Codes 0—5 are reserved for memory reference instructions.

codes 6 and 7 are for augmented instructions.

Operation

These two types of instructions are defined, and the instruc-

tions described, in the following sections.

The following special symbols are used in the instruction lists below.
Definition

M
C(A)

The contents of register A

C(A) % C(B)

The contents of register A replace the contents of

The address or location of any memory register

The

ith bit of register Y

Y]_4

Bits

l—4, inclusive, of register Y

C(AO-S) => C(Yé—l 1)

The contents of bits 0—5 of register A replace the contents of bits 6-H of register Y.
not affected

The contents of A are

Inclusive OR

AND

CiAl

The l's complement of the contents of register A

(Boolean)

Memory Reference Instructions
Word format of memory reference instructions is shown in Figure 3—1 , and the instructions are explained
in Table 3-l

OPERATION
CODES 0—5

MEMORY
PAGE

r—*—fi
0

L..Y_J

TABLE 3-I
Mnemonic

Octal

Time

Symbol

Code

(pseC)

AND Y

3.0

\
v

INDIRECT
ADDRESSING

Figure 3-i

ADDRESS

Memory Reference Instruction Format
MEMORY REFERENCE INSTRUCTIONS

Operation

Logical AND. The AND operation is performed between
the C(Y) and the C(AC). The result is left in the AC,
C(AC) are lost. The C(Y) are unchanged.
Corresponding bits are compared independently. This instruction, often called extract or mask, can be considered
=>
/\
as a
bit—by—bit multiplication.
and the original

C(YI) C(ACI)

C(ACI)

Example

C(ACI) original

C(YI)

final

C(ACI)

—'OO

I
0
I

TAD Y

3.0

Two's complement add. The C(Y) are added to the C(AC)
in 2's complement arithmetic
The result is left in the AC
.

and the original

C(AC)

are

lost.

The C(Y) are unchanged.

If there is a carry from AC0, the link is complemented.

This feature is useful in multiple precision arithmetic

C(Y) + C(AC) => C(AC)
[52 Y

3.0

Index and skip if zero.

The C(Y) are incremented by one

in 2's complement arithmetic.

If the resultant C(Y)

the next instruction is skipped.

If the resultant C(Y) )5 O,

the program proceeds to the next instruction
are

The C(AC)

unaffected.

C(Y) + I => C(Y)
If result
0, C(PC) + 1 => C(PC)
=

DCA Y

3.0

Deposit and clear AC. The C(AC) are deposited in core
memory location Y and the AC is then cleared.

The previous C(Y) are lost.
C(AC) => C(Y), then 0 => C(AC)

3—2

TABLE 3-]

MEMORY REFERENCE INSTRUCTIONS (continued)

Mnemonic

Octal

Time

Symbol

Code

(psec)

JMS Y

3.0

0 pe r a f.Ion

Jump to subroutine.
ory location Y.

Y + I

The C(PC) are deposited in core mem—

The next instruction is taken from location

C(PC) + I => C(Y)
Y + 1 => C(PC)
JMP Y

I .5

Jump to Y.

The next instruction is taken from core mem—

ory location Y.
Y => C(PC)

Augmented Instructions
There are two classes of augmented instructions, or instructions which do not reference core memory.

They are the input-output transfer (IOT) which has an operation code of 6, and the operate (CPR), which
has an operation code of 7.

Bits 3 through II within these instructions Function as an extension of the

operation code and can be microprogrammed to perform several operations with one instruction. Augmented
instructions are

I-cycle instructions requiring I .5 psec for execution.

Input-Output Transfer Instruction
Microinstructions of the input—output transfer (IOT) instruction effect information transfers between the
arithmetic and control element and an input-output device via the input-output control element.
format of the IOT instruction is shown in Figure 3—2.

The

Bits 3 through 8 are used to select the I/O device;

and bits 9 through II enable generation of I/O pulses during event times 4,

2, and I, respectively.

Operations performed by IOT microinstructions are explained in Chapter 4 of the PDP—8 Users Handbook.

GENERATES

AN IOP 4

AN IOP 1
PULSE AT
EVENT TIME1

PULSE AT
OPERATION

EVENT TIME 3
IF A 1

DEVICE

SELECTION

IF A 1

k—Vh"
GENERATES
AN IOP 2
PULSE AT

EVENT TIME 2
IF A I

Figure 3-2

IOT Instruction Format

3—3

Operate Instruction
The operate instruction consists of two groups of microinstructions.

Group I is principally for clear,

complement, rotate, and increment operations and is designated by the presence of a O in bit 3. Group 2
is used principally for checking the contents of the accumulator and link and continuing to or skipping the
next instruction based

the check.

A l in bit 3 and a 0 in bit I]

designate a Group 2 microinstruction.

Group I operate microinstruction format is shown in Figure 3-3, and the microinstructions are listed in the
table below.

Any logical combination of bits within this group can be combined into one microinstruction.

For example, it is possible to assign l's to bits 5,

6, and H; but it is not logical to assign l's to bit 8 and

9 simultaneously since they specify conflicting operations.

If RAL,

RAR, RTL, or RTR is specified, lAC

may not be specified, and conversely.

ROTATE 4
POSITION IF A 0.

ROTATE
OPERATION
CODE 7

AC AND L

CLA

CMA

RIGHT

2 POSITIONS
IF A t

(—k‘fi

r—N‘w

r—A—x

r—J—w

L_Y_..J

\_W._J

CONTAINS
A 0 TO
SPECIFY

CLL

CML

kw._)

ROTATE

IAC

AC AND L
LEFT

GROUPI

Figure 3-3

TABLE 3—2

Group I Operate Microinstruction Format

GROUP 1 OPERATE MICROINSTRUCTIONS

Mnemonic

Octal

Event

Symbol

Code

Time

CLA

7200

0 pera r‘Ion

Clear AC.
0 => C(AC)

CLL

7100

Clear L

0 => C(L)
CMA

7040

Complement AC. The C(AC) are set to the T's com—
plement of C(AC).
C(AC) => C(AC)

CML

7020

Complement L.
C(D => C(L)

RAR

7010

Rotate AC and L right.
rotated right one place

C(AC.) => C(AC. +1)

a C(L)l
C(ACll)
:>

C(L)

3-4

C(ACO)

The C(AC) and the C(L) are
.

TABLE 3—2

GROUP I OPERATE MICROINSTRUCTIONS (continued)

Mnemonic

Octal

Event

Symbol

Code

Time

RAL

7004

O p erat'0 n

Rotate AC and L left.

The C(AC) and the C(L) are

rotated left one place.

C(ACO) => C(L)

C(L) 3) C(Ac1 1)
7012

RTR

Rotate two places to the right.

Equivalent to two

successive RAR operations.

RTL

7006

Rotate two places to the left.

Equivalent to two

successive RAL operations.

700i

IAC

Index AC.

The C(AC) are incremented by one in

2's complement arithmetic

C(AC)+T => C(AC)
7000

NOP

No operation.

—

Causes a I .5 psec program delay.

Group 2 operate microinstruction format is shown in Figure 3—4, and the microinstructions are listed in
the table below.

Any logical combination of bits within this group can be composed in one microinstruction.

REVERSE

OPERATION
CODE 7

CLA

r———*-——\

r—Afl

SZA

SENSING OF
BITS 5,6,7

HLT

r—H

H—fi

F—‘fl

\_.Y_J

H—’

k_w_l

9T4

L_w_l

CONTAINS A t
TO SPECIFY

SMA

SNL

05R

CONTAINS A 0
TO SPECIFY
GROUP 2

GROUP 2

Figure 3-4

Group 2 Operate Microinstruction Format

If skips are combined in a single instruction, the inclusive OR of the conditions determines the skip.

For

example, if I's are designated in bits 6 and 7 (SZA and SNL), the next instruction is skipped if either
C(AC)

0, or C(L)

I, or both.

If two reverse sense skip instructions are combined (bit 8 is set), the

logical AND of the conditions determines the skip.

For example, if I's are designated in bits 6,

8 (SNA and SZL), the next instruction is skipped if C(AC) f 0
from
to

2E C(L)

7, and

The CLA microinstruction

Group I can be combined with Group 2 commands. This command occurs at event time 2 with respect

the event times listed in the following table

3—5

TABLE 3-3

Mnemonic

Octal

Symbol

Code

CLA

7600

GROUP 2 OPERATE MICROINSTRUCTIONS

Operation

Clear AC.
0 => C(AC)

SPA

Skip on positive AC. If the C(AC) is a positive
number, the next instruction is skipped.
If
C(ACO) 0, then C(PC) + l => C(PC)

75l0

SMA

7500

Skip on minus AC. If the C(AC) is a negative number, the next instruction is skipped.
It C(ACO)
i, then C(PC) + 1 => C(PC)
=

SNA

7450

Skip on non-zero AC.
lf C(AC) 75 0, then C(PC) +i => C(PC)

SZA

7440

Skip on zero AC
lf C(AC)
0, then C(PC) + l => C(PC)
.

SZL

7430

Skip on zero L. if C(L) O, the next instruction is
skipped.
lf C(L)
0, the C(PC) + l => C(PC)
=

SNL

7420

Skip on non-zero L.
If C(L)
i, then C(PC) + 1 => C(PC)
=

SKP

74l0

Skip, unconditional.
C(PC) + 1 => C(PC)

OSR

7404

OR with switch register.

The next instruction is skipped.

(May be combined with

CLA.)
C(SR) V C(AC) => C(AC)
HLT

Halt.

7402

Stops the program.

If this instruction is com-

bined with others in the CPR 2 group, the computer

stops immediately after completion of the cycle in
process.

THE ORGANIZATION OF MEMORY
A PDP-8 memory Field can be likened to a book.

The 4,096 words of the memory field correspond to the

lines of text, and if we divide the memory into segments of l28 words each, we have an analogy to a

32-page book in which each page contains 128 lines.
The memory field is in fact segmented in this fashion, and the analogy to a book is affirmed by the fact
that each of the l28-word blocks is called a page.
to allow direct reference to all

be provided.

Because the PDP-8 instruction word is not long enough

of the registers in a memory field, a special type of address reference must

This can be illustrated by pursuing our book analogy a little bit farther.

Suppose that one is reading a text, and taking notes in the margins of each page.
some

of the notes will refer to other parts of the page,

to information

One can expect that

other pages.

if one of the notes refers to a line of text on the same page, write only the line number.

For convenience,
If the note re-

fers to a line on some other page, write the page number followed by the number of the line on that page.

Alternatively, of course, it is possible to begin on the first page and number all the lines of the book con-M

secutively, and refer to them by these numbers alone.
In similar fashion,

given word of memory may be referred to by its page address; that is, its address

within a page, or by its absolute address, which designates its position in the whole of memory.

analogy ends

A PDP—8 memory field is organized as follows:

the 4,096 words are arranged sequentially, with absolute

77778. The field is divided into 32 pages numbered from 0-378. Each page

addresses of 0 through

tains l28 registers, with page addresses of O—l77

As Figure 3-l

If bit 4 of the instruction contains a

con—

shows, the address field of a memory

field of a memory reference instruction contains 7 bits, which is iust enough to allow access to

locations.

Here the

2008

l, the address field of the instruction refers to one of the

addresses on the current page, that is, the page in which the instruction is stored.

If bit 4 contains a 0,

the reference is to an address on page 0.

The state of bit 3 of the instruction determines what is done with the contents of the memory register specified by bits 4-l l

if this bit is O, the contents of the cell addressed by the instruction are taken as the

operand, and the operation is completed.
effective address of the instruction.
are

In this case, the address specified by the instruction is the

lf, however, bit 3 contains a i, the contents of the cell addressed

treated, not as the operand, but as the l2-bit absolute address of the register containing the operand.

in other words, the cell addressed contains the effective address of the instruction.
ory reference instruction

can

in this way,

a mem—

indirectly address any register in the memory field, regardless of which page

it is on.

Thus, there are four ways, depending on the states of bits 3 and 4 of a memory reference instruction, in
which the effective address may be obtained

TABLE 3-4
Bit 3

Bit 4

EFFECTIVE ADDRESS CALCULATION
Effective Address

The operand is in page 0 at the address specified by bits 5

through ll
0

The operand is in the current page at the address specified by

bits 5 through ll.
3—7

TABLE 3—4

EFFECTIVE ADDRESS CALCULATION

Bit 4

Bit 3

(continued)

Effective Address

The absolute address of the operand is taken from the contents
of the location in page 0 designated by bits 5 through II.

The absolute address of the operand is taken from the contents
of the location in the current page designated by bits 5 through II.

COMPLEMENT ARITHMETIC
In the PDP-8, as in other machines utilizing complementation techniques,

negative numbers are repre—

sented as the complement of positive numbers, and subtraction is achieved by complement addition.

Representation of negative values in 2's complement arithmetic is slightly different from that in I's com—

plement arithmetic

The I's complement of a number is the complement of the absolute positive value; that is, all I‘s are re—

placed by 0's and all 0's are replaced by I's.

The 2's complement of a number is equal to the I's com-

plement of the positive value plus one.
In I's complement arithmetic a carry from the sign bit (most significant bit) is added to the least significant bit in an end-around carry.

In 2's complement arithmetic a carry from the sign bit complements the

link (a carry would set the link to I if it were properly cleared before the operation), and there is no
end-around carry

A I's complement representation of a negative number is always one less than the 2's complement repre—
sentation of the same number.

Differences between I's and 2's complement representations are indicated

in the following table.

TABLE 3-5

ONE'S AND TWO'S COMPLEMENT REPRESENTATIONS

Number

One's Complement

Two's Complement

000000000] OI

000000000I OI

000000000I 00

000000000100

0000000000I I

0000000000 I 0

00000000001 I
0000000000 I 0

+ I

00000000000I

000000000000

-0

IIIIIIIIIIII

Nonexistent

—I

IIIIIIIIIIIO

IIIIIIIIIIII

3—8

TABLE 3—5

ONE'S AND TWO'S COMPLEMENT REPRESENTATIONS (continued)

Number

One's Complement

Two's Complement

-2

111111111101

111111111110

-3

llllllllllOO

111111111101

-4

lllllllllOll

111111111100

—5

111111111010

111111111011

Note that in 2's complement there is only one representation for the number which has the value zero,

while in l's complement there are two representations.

Note also that complementation does not interfere

with sign notation in either l's or 2's complement arithmetic; bit 0 remains a 0 for positive numbers and a
l for negative numbers.

PDP-8 arithmetic operations (as exemplified by the TAD instruction) are carried out in 2's complement.

This means that the operands involved in the arithmetic must be in correct 2's complement representation.

The 2's complement of any number is formed by taking the l's complement, and adding 1

to it.

The two

operate class instructions, CMA and lAC, may be combined into a single instruction to perform the operation of taking a 2'5 complement.

Because the operation is often required, the mnemonic CIA is given to

the combined instruction.

Addition

The addition of a number contained in a core memory location and the number contained in the accumulator is performed directly by using the TAD Y instruction, assuming that the binary point is in the same

position and that both numbers are properly represented in 2's complement arithmetic. Addition can be

performed without regard for the sign of either the augend or the addend. Overflow is possible, in which
case

the result will have an incorrect sign, although the eleven least significant bits will be correct.

carry from bit 0 will

complement the link bit.
Subtraction

Subtraction is performed by complementing the subtrahend and adding the minuend.

As in addition, if

both numbers are represented by their 2's complement, subtraction can be performed without regard for
the signs of either number.

3-9

CHAPTER

ORGANIZATION

The rules for writing a program can best be introduced by tracing the steps from the statement of a problem
to the completed routine.

Consider, for example, a program to compute the sum of the first n integers,

s=n+(n-I)+(n—2)+(n—3)+ ...+3+2+I
First, a flow chart, that is, a block diagram of the general steps to the solution of the problem, should be
made.

It might look something like Figure 4-]

Ignoring the first box for the moment, consider the

second, third, and fourth boxes. The operations specified in these three boxes perform the main computation of the sum.

Box 2 specifies the actual arithmetic of computing the partial sum.

which counts the number of partial additions that have been made.

Box 3 is a counter,

Box 4 is a decision point; if the count

indicates that the sum is complete, the program goes on to box 5; if it is not complete, the program returns to

the beginning of the main computation.

program is called a

This sort of continuous recycling through a section of

loop.

CLEAR TEMPORARY
REGISTERS
AND INITIALIZE

CALCULATE
sum

PARTML

‘

DECREMENT
uuosx

NO
4

YES

Figure 4-]

HALT

Flow Chart of Program to Calculate Sum of Integers

The first box in this chart specifies the operations which prepare the program for operation.

Since nearly

every program written will be used more than once, with different data each time, it is necessary at the

4—1

start of the routine to clear out old results from

previous use of the program so that one can start fresh

with new data, in the same way that one clears the keyboard and registers of an adding machine before

starting a new calculation.

In programming terms, this preparation is called initializing.

Besides indicating the general sequence of operations, the flow chart also gives an idea of storage require—
ments.

Space will be needed not only for the instructions which perform the computation, but also for the

data used by the routine.

One register will be needed to hold the partial sum accumulated through re-

peated additions; this register will naturally hold the correct total when the calculation is finished.
Another cell is needed for the M, which is the counter specified in box 3 of the flow chart.
one

Finally,

The program, then, must supply two types of information.

First, it must include the executable instructions,

data, and temporary registers which will occupy memory cells when the program is executed. Second, it
must

include a certain amount of information for the MACRO Assembler itself, to establish the locations

and extent of the program in memory.

These assembler directives are called pseudo—instructions;

included in the program along with the executable instructions.

they are

A pseudo—instruction is rather like a

proofreader's mark on a manuscript; the mark provides information to the editor, but does not itself cause
additional text to appear in the printed book.

Similarly, a pseudo-instruction provides information to the

Assembler, but does not itself cause any coding to be inserted into the object program.

CODING
Now to write the program.

First one must decide where the program is to be stored.

This word is called the MI

necessary only to establish the location of the first register of the program.

and can be designated in two ways.
to be stored;

To do this, it is

The most common way is to designate the page in which the program is

the origin is automatically set at the first register (page address 0) in that page.

Thus, to put

the routine in page 2, the pseudo-instruction, PAGE, would be used as follows:
PAGE 2

The program would then begin in location 0 of page 2.
To set the origin to any register other than the first one in a page, the absolute address of the starting

location must be specified.

This is done by using the special character

"*"
.

To place the origin at loca-

tion 210, the coding would be:

*210

In establishing the origin, an index within the Assembler, called the current location counter (CLC), has
been set equal to the absolute address of the origin.

For example, the use of PAGE 2 would set the CLC

4-2

to a

value of 400, which is the absolute address of the first register in page 2.

the CLC to be set to a value of 210.

Likewise, *210 causes

In programming terms the CLC points to the origin; such an index

is called a pointer.

The coding for the program may now be written.

First three registers for the three items of data must be

Usually, data storage is set aside by placing a 0 in each register, thus allocating a cell.

reserved.

The

coding looks like this:
PAGE 2
O
O

Since the origin has been set at location 400, the three data cells occupy locations 400, 40] , and 402

(page addresses 0, l, and 2, respectively).
Next, write the coding For the computation.
PAGE 2
0
0
0

CLA
DCA 400
TAD 40]

DCA 402
TAD 400
TAD 402

DCA 400
STA
TAD 402
SZA
JMP 406
HLT

33
The program now occupies locations 400—416 (0-16 on page 2).

The special character $ indicates to the Assembler that it is a complete program and that nothing else is
to follow.

This program is now complete, and can be assembled into a working routine.

The coding, however, is

rather stark, and not very useful to another programmer, should he wish to discover what the program does.
Of course, he could figure out the function of the routine by going through it step by step with pencil

and paper, but if this program were much
a

longer, this task would be tedious and impractical. Comments,

method of explaining the functions of the program, make the work immediately useful to someone else.

COMMENTS
Comments are included in a PDP-8 program in a simple way.
our

program.

Consider the three data storage locations of

A comment would identify each register's function immediately; they might appear as:

/PARTIAL SUM AND FINAL TOTAL
/|NTEGER N
/INDEX

0 0-

The slash preceding each comment is the character which identifies a comment field.

All information on

that line following the slash is considered to be a comment, and is ignored by the Assembler.

For clarity

and neatness, a tab has been inserted between the instruction and the comment.

Since a slash identifies all the subsequent information on a line as a comment, one full line can be used

for a title by placing the slash in the first space after the left margin.

With a title, the program might

look like this:

/INTE GER SUMMATION ROUTINE
PAGE 2

/PART|AL SUM AND FINAL TOTAL

Now, an immediate identification of what the program is to be used for is included in the listing.
More comments explain the workings of the program:

/INTE GER SUMMATION ROUTINE
PAGE 2
o
0

/PART|AL SUM AND FINAL TOTAL
/INTEGER N
/INDEX

CLA
DCA 400

/ZERO TO SUM

TAD 401

DCA 402
TAD 400

/SET INDEX
/MAIN LOOP

TAD 402

DCA 400
STA

/DECREMENT INDEX

4-4

TAD 402

/IS IT 0?
/NO: KEEP GOING
/YES: HALT

SZA
JMP 406

HLT

$
Now it is a little easier to understand the program, but some limitations remain.

It is tied to one particu-

lar place in the computer's memory; to put it in some other location, recoding the entire routine with new
addresses would be necessary.

Nor are these octaI page addresses much help in keeping track of program

flow; even in a short routine such as this, it is necessary to do some counting to find out what location the
instruction STA is in, for example.

A simple,

meaningful way of identifying storage addresses is needed.

ADDRESS TAGS

As described above, a symbol is assigned to each of the PDP-8 instruction codes.
mer can

Similarly, the program—

assign a symbol to any one of the storage addresses in the computer memory. For instance, the

routine under discussion contains three important registers:

the beginning of data storage, the beginning

of the program sequence, and the beginning of the main loop.

Each one of these locations can be labelled with a symbolic tag, in the following way:

/INTEGER SUMMATION ROUTINE
PAGE 2

DATA, 0

/PART|AL SUM AND FINAL TOTAL
/INTE GER N
/INDEX

0
0

BEGIN, CLA
DCA 400

/ZERO TO SUM

TAD 401

GO,

DCA 402
TAD 400

/SET INDEX
/MAIN LOOP

TAD 402

DCA 400
STA

/DECREMENT INDEX

TAD 402

SZA
JMP 406

HLT

/Is IT 0?
/NO: KEEP GOING
/YES; HALT

$
Three reference points have been established.

Note that the tags are roughly mnemonic, thus offering an

additional measure of program identification and clarification.
been defined.

In programming terms, three symbols have

In this case, each symbol has a value that is equal to the absolute address of the register in

which the associated item (data word or instruction) is stored.
of BEGIN, 403; of GO, 406.

Thus, the value of the symbol DATA is 400;

Symbolic Addressing
Locating sections of a program has been simplified, but full advantage of address tags has not yet been
taken.

The instruction addresses themselves are still tied to the absolute addresses of the program.

ever, since each

How-

tag has a numeric value corresponding to an address, a tag may also be used as a symbol

in the address part of an instruction.

For example, the absolute address 402 is clearly equivalent to the

expression, BEGIN-l (BE GIN=403; 403—l=402). Similarly, the address 4]] is equivalent to GO+3.

Replacing absolute addresses with symbolic expressions, the program will look like this:
/INTE GER SUMMATION ROUTINE
PAGE 2

/PART|AL SUM AND FINAL TOTAL
/INTEGER N
/lNDEX

DATA, 0
0
o

BEGIN, CLA
DCA DATA

/ZERO TO SUM

TAD DATA+I

Go,

DCA DATA+2
TAD DATA

/SET INDEX
/MAIN LOOP

TAD DATA+2

DCA DATA
STA

/DECREMENT INDEX

TAD DATA+2

SZA
JMP Go
HLT

/IS IT 0?
/NO: KEEP GOING
/YES: HALT

This program that is easily readable, includes a running commentary for explaining its functions, and
carries its own symbolic references that further assist in understanding and keeping track of the routine.

Moreover, the routine has been freed from a specific place in memory by the use of symbolic addresses.
To change the location of the program, it is necessary only to change the origin setting by the PAGE

pseudo-instructions

STORAGE TECHNIQUES

The program above will function but it may be improved.

One of its drawbacks is the fact that wherever

the program is stored, the data must be stored right along with it.

When space is at a premium, as it

often is, it would be desirable to put all of the data in a fixed place, and let the working parts of the
program be arranged as necessary.

In the PDP-8, data that is used by several parts of a program is often stored in page 0.

If each of the

three data words in the program is given a name, page 0 addresses can be assigned to them.

4-6

Call the first

word TOTAL, the second INDEX, and the third N

In effect, the data words have been labeled with

address tags, but since the names can be used in symbolic expressions (for instance, in an instruction

address), programmers usually refer to them as variables.

If the variables are assigned to page 0 and the

instruction address in the program changed to match, the coding will look like this:

/| NTE GER SUMMATION ROUTINE
*20

TOTAL,
INDEX,
N,

0
0

/THESE REGISTERS ARE
/Now ON PAGE 0

PAGE 2

BEGIN, CLA
DCA TOTAL

/ZERO TO SUM

TAD N

GO,

DCA INDEX
TAD TOTAL

/SET INDEX
/MA|N LOOP

TAD INDEX
DCA TOTAL

STA

/DECREMENT INDEX

TAD INDEX
SZA

JMP GO

HLT

/IS IT 0?
/NO: KEEP GOING
/YES: HALT

When the program is loaded, the three items of data (having been assigned addresses on page 0) will be
stored in page 0.

A large step toward our goal of a refined, useful routine has been taken; i.e., a program which may be

placed anywhere in memory and is easily understood and interpreted by someone other than the author,
because of liberal comments and symbolic addressing.

OPTIMUM USE
It remains only to relate this sequence of coding to the context in which it is likely to operate.

Obviously

the summation program is useless in core by itself, waiting for an integer summation to be carried out.

Since the example has served its purpose in illustrating how a complete program can be written, all the

coding except the minimum necessary for the actual computation will be removed, leaving this (the

ellipsis always indicates the presence of unspecified coding):

4-7

BEGIN, CLA
DCA TOTAL

/ZERO TO SUM

TAD N

GO,

DCA INDEX
TAD TOTAL

/SET INDEX
/MAIN LOOP

TAD INDEX

DCA TOTAL

/DECREMENT INDEX

STA
TAD INDEX

/IS IT 0?
/NO: KEEP GOING
/YES: HALT

SZA
JMP GO

HLT

This sequence can now be placed within a larger program wherever it may be needed.

example of a long program which uses this routine two times.

Figure 4-2 is an

Whenever the summation routine occurs,

it is preceded by an instruction which takes the value of N from the AC and places it on the proper location in data storage.

One other change is also necessary.

The address tag G0 has been eliminated from

all but the First occurrence in the coding sequence; it is obvious that one symbol cannot be defined with
more

than one value, or there would be confusion as to which of the values was meant at any one time

As a result, it is necessary to eliminate the reference to GO in the JMP instruction at the end of the
sequence

Notice that the instruction, DCA INDEX always remains exactly seven locations ahead of the JMP in-

struction, regardless of where the routine is stored. Evidently, an address expression which could refer
to the
a

location of DCA INDEX relative to the location of the JMP instruction is needed.

In other words,

construction is desired that will perform the same function as the following but without having to define

the symbol HERE:
DCA INDEX
TAD INDEX

HERE,

JMP HERE-2

Remembering that the current location counter always llpoints" to the location of the instruction currently

being assembled, it becomes apparent that what is desired is a symbol which, whenever it is used, film
takes the value of the CLC

In MACRO-8, this symbol is the period,

Whenever this character is

encountered in an address expression, the value of the CLC is substituted for it.

In the example in

Figure 4-2 it is used in the address of the JMP instruction wherever it appears in the routine.
the necessity of having to define a new set of symbols each time the routine is used is avoided.

In this way,

PAGE 10

BEGIN,

GO,

/INITIALIZATION

DCA

TOTAL

TAD

DCA

INDEX

TAD

TOTAL

TAD

INDEX

DCA

TOTAL

/SET INDEX
/MAIN LOOP

/|NCREMENT INDEX

STA
TAD

INDEX
GO

/IS IT 0?
/NO: KEEP GOING
/YES: ALL DONE
/IN|T|ALlZATION

SZA
JMP

DCA

TOTAL

TAD

DCA

INDEX

TAD

TOTAL

TAD

INDEX
TOTAL

DCA

/SET INDEX
/MAIN LOOP

/DECREMENT INDEX

STA
TAD

INDEX

/Is IT 0?
/NO: KEEP GOING
/YES: ALL DONE

SZA
IMP

.-7

Figure 4-2

Program Example

SUBROUTINES
Now the program example is useful, but one more difficulty must be overcome.

An examination of

Figure 2-4 will make thisproblem obvious: If the routine is used very many times, an extremely large
amount of storage space will

be used simply in repetitions of the same coding sequence.

the general aim of writing a concise, efficient program.

This counteracts

If the routine could be written only once,

placed

in a separate section of memory and called into operation each time it was needed, the whole procedure

would be more efficient.

4~9

A program that may be used in this manner is called a subroutine.

It is a sequence of coding with the

fol lowing properties:

It is self-contained, that is, it can be assembled by itself without having to be part

of a

larger program.

It occupies its own section of memory,

It performs only one task.

It is called into operation only by another program, and when it has finished its

task, it returns to that program

logically separate from other coding sequences.

Each subroutine has a definite purpose.

It can be called any number of times, and upon com-

pletion always returns control to the point from which it was last called.
5.
a

When necessary, it uses data supplied by the calling program, and returns results to

place accessible by that program

When a subroutine is written, four requirements which are implied by the last two properties listed above
must be fullfilled:

The data must be accessible by the subroutine.

The results must be accessible by the calling program.

There must be a way of calling the subroutine into operation

There must be a way of returning control to the calling program

The problem of data transmission has been solved by placing the three data storage registers in page 0,

thereby making them accessible to any program or subroutine in memory.
return is not quite

obvious.

The transfer and control of

First, restore the symbolic tags to the subroutine as well as the title and

other pseudo-instructions to make it a self-contained program giving:

/INTEGER SUMMATION SUBROUTINE
PAGE 2

INTSUM, DCA N
DCA TOTAL

/SAVE INPUT IN C(AC)
/ZERO TO SUM

TAD N

GO,

DCA INDEX
TAD TOTAL

/SET INDEX
/MAIN LOOP

TAD INDEX

DCA TOTAL
STA

/DECREMENT INDEX

TAD INDEX

/IS IT 0?
/NO: CONTINUE
/YES

SZA
JMP GO

At a first glance, it could seem easy to eliminate the problem of transferring control to the subroutine.
A JMP INTSUM in the calling program every time the subroutine was called would provide the path.

But

The subroutine has no way of knowing where to return.

The

what happens at the end of the calculation?

subroutine cannot merely halt as this violates requirement d above.
The solution lies in the use of the JMS instruction.

Remembering from Chapter 2 when the JMS instruction

is executed, the contents of the program counter are saved‘ in the location addressed by the instruction,

and control is transferred to the register immediately following the one addressed.
contains the address of the next instruction following the JMS is available.

Thus, a register which

This can be used to return to

the proper point in the calling program when the subroutine has finished its computation.

Rearrange the

first part of our subroutine as follows:

/INTEGER SUMMATION ROUTINE
PAGE 2

INTSUM,

DCA N
DCA TOTAL

/SAVE PC HERE
/SAVE INPUT NUMBER
/ZERO TO SUM

TAD N

GO,

DCA INDEX

/SET INDEX

Now, whenever the instruction JMS INTSUM is used, the contents of the program counter are stored in the
location tagged with the name INTSUM and the computer continues operation with the instruction imme-

diately following which is the first instruction of the subroutine.
Now that the subroutine has been called, a means to exit from it must be provided.
stores

the program counter which

The JMS instruction

”points" to the instruction following the JMS. After the execution of a

JMS INTSUM, the register INTSUM contains the contents of the program counter at the time the JMS was
executed.
return to

This can be used as an effective address (See Chapter 3 , the "Organization of Memory") to

the point of call.

To do this, the subroutine may be terminated with a JMP instruction which

references the register INTSUM as an indirect address as follows:

4-II

/lNTE GER SUMMATION SUBROUTINE
PAGE 2

DCA N

/SAVE PC HERE
/SAVE INPUT NUMBER

JMP l lNTSUM

/YES:

lNTSUM, 0

The character "I" is used to signify indirect addressing.

EXIT SUBROUTINE

In use here it means,

"Jump to the register whose

address is contained in the register tagged lNTSUM."

The complete subroutine is shown below:

(not including the page 0 definition of the variables).

/lNTE GER SUMMATION SUBROUTINE
PAGE 2

INTSUM, o
DCA N
DCA TOTAL

/SAVE PC HERE
/SAVE INPUT NUMBER
/ZERO TO SUM

TAD N

GO,

DCA INDEx
TAD TOTAL

/SET INDEX
/MA|N LOOP

TAD INDEX

DCA TOTAL
STA

/DECREMENT INDEX

TAD INDEX
SZA
JMP GO
JMP I INTSUM

/IS IT 0?
/NO: CONTINUE
/YES: ALL DONE

The reader is referred to the PDP—8 Library For other examples of subroutine calling techniques.
The MACRO—8 Programming Language may now be discussed in detail.

PART
THE

MACRO

LANGUAGE

CHAPTER

MACRO—8

PROGRAMMING

LANGUAGE

The MACRO-8 Symbolic Assembler accepts source programs written in symbolic

them into binary form.
memory locations

language and translates

MACRO—8 produces an obiect program tape (binary), a symbol table defining

(for use with DDT), an octal/symbolic assembly listing, and useful diagnostic messages.

MACRO-8 is compatible with PAL Ill, but has the following additional Features:
User-Defined Macro's

Groups of computer instructions required for the
the solution of a specific problem can be defined
as a macro

Double Precision Integers

instruction by the user.

Positive or negative double precision integers are

allotted two consecutive core locations.

Floating Point Constants

The format and rules for defining these constants
are

compatible with the format used by the PDP—8

Floating Point Package (See Digital—8-5-S).
Operators

Symbols and integers may be combined with a
number of operators

Literals

Addition

Boolean AND

Subtraction

Boolean lnclusive OR

Symbolic or integer literals (constants) are automatical ly assigned

Text Facility

There are text facilities for singl 3 characters and

blocks of text.

Link Generation

Links are automatically generated for out of page
references

To incorporate these new features, it was necessary to decrease the size of the symbol table and because

of this, programs that were originally coded to be assembled by PAL lll might have too many symbols to
be assembled by MACRO-8.

5-l

CHARACTERS

Programs in the MACRO—8 language are written using characters from the ASCII character set. The

following characters are used:

Letters
ABCD...XYZ

Digits
i234567890

Punctuation Characters
Since a number of characters are invisible (i .e.,

nonprinting), the following notation is used to represent

them in the examples:

._I

space

___>|

tab

)

carriage return

The Following characters are used to specify operations to be performed upon symbols or numbers:

Character

253

space

Combine symbols or numbers

plus

Combine symbols or numbers (add)

minus

Combine symbols or numbers (subtract)

exclamation point

Combine symbols or numbers (OR)

carriage return

Terminate line

tab

Combine symbols or numbers or format source

Ll
+

tape
comma

Assign symbolic address

equals

Define parameters

;

semicolon

Terminate coding line

dollar sign

End of pass

asterisk

Set location counter

point

Has value equal to current location counter

5—2

slash

Indicates start of a comment

ampersand

Combine symbols or numbers (AND)

quote

Generate ASCII constant

( )

parentheses

Define literal on current page

[ ]

brackets

Define page 0 literal

< >

angle brackets

Define a macro

Ignored Characters
End of a

Form-feed

logical page of a source program

(see Symbolic Editor)

leader/trailer

Blank tape

Used for

Rubouts

Used for deleting characters

Code 200

Used for leader/trailer

Line-feed

Follows carriage-return

All other characters are illegal when not in a comment or a TEXT field, and cause the error message lC
to

be printed.

The form—feed is used at the end of a page of program for editing purposes.

of leader and trailer are self-explanatory.

The functions

This may be either blank tape or code 200.

Tabulations are usually used in the body of a program to provide a neat page for printing; they can separate
fields from one another, as between an instruction and an associated comment.

For example,

line could

be written as:

(30,

TAD TOTAL/MAIN

LOOP)

but it is far easier to read if tabs are inserted:

GO, —>l TAD TOTAL —Dl /MAIN

(the characters —§l and
Either

;

are

LOOP)

nonprinting)

(semicolon) or the combination carriage-return/line-feed may be used as line terminators.

The

semicolon is considered identical to carriage-return/line—feed except that it will not terminate a comment.

Example:
TAD A

/THIS IS A COMMENT

The entire expression up to the carriage return is considered a comment.

5-3

TAD

B’)

As was noted previously, the tabulation is used as a Formatting device to provide a neat appearance for
Use of the semicolon allows the programmer to place several lines of coding

the printed program listing.
on a

single line. If, for example, he wishes to write aysequence of instructions to rotate the C(AC) and

C(L) six places to the right, it might look like:

RTR)
RTR)
RTR)
He may place all three instructions as a single line by substituting the control character

for the line terminator

RTR;

”

(carriage return).

(semicolon)

The above sequence of instructions may be rewritten as:

RTR)

This type of format is particularly useful when setting aside a section of data storage for a list.

For ex-

ample, a lZ-word list could be reserved by:
LIST,

0‘)

A neat printout (or program

listing, as it is usually called) makes subsequent editing, debugging, and

interpretation much easier than if the coding were laid out in a haphazard fashion. The coding practices
listed below are in general use, and will result in a readable, orderly listing.
a

program to produce

listings of this form.)

A title comment begins at the left hand margin.

Pseudo-instructions may begin at the left margin; often,

one

tab stop to line up with the executable instructions.

Address tags begin at the left margin.

tabulation

(See Digital-8—21—U for

however, they are indented

They are separated from succeeding fields by

Instruction fields, whether or not they are preceded by a tag field, are indented

one

tab stop

A comment is separated from the preceding field by a tabulation, unless it occupies

the whole

line, in which case it usually begins at the left margin.
5-4

Elements

Any group of letters, digits, and punctuation characters which represents binary values less than 2]
.

IS an

element.

Integers

Any sequence of numbers delimited by punctuation characters forms a number. Example:
l
l2

4372

The radix control pseudo-instructions indicate to the Assembler the radix to be used in number interpre—
tation.

The pseudo-instruction DEClMAL indicates that all numbers are to be interpreted as decimal until

the next occurrence of the pseudo-instruction OCTAL.
The pseudo-instruction OCTAL indicates that all numbers are to be interpreted as octal until the next
occurrence

of the pseudo-instruction DEClMAL.

less otherwise specified

The radix is initially set to octal and remains octal un-

Symbols
A symbol is a string of one or more alphanumeric characters delimited by a punctuation character.
are

Symbols

composed according to the following rules:
l

The characters must be either alphabetic (A-Z) or numeric

(0—9)

The first character must be alphabetic.

Only the first six characters of any symbol are meaningful; the remainder, if any,

are

ignored.

A symbol is terminated by any nonalphanumeric character.

The MACRO Assembler has, in its permanent symbol table, definitions of the symbols for all PDP—8 operation

codes, operate commands, and many lOT commands (see Appendix A for a complete list).
used without prior definition by the user.

JMS

Example:

is a symbol whose value of 4000 is taken from the operation code

definitions

These may be

is a user—created symbol.

When used as a symbolic address tag, its value

is the address of the instruction it tags.

This value is assigned by the

Assembler.

Note that because of rule 3, a symbol such as INTEGER, for instance, would be interpreted as INTEGE
since the seventh letter is ignored.

if symbols of more than six characters are used, the programmer should

be careful to avoid the error of defining two apparently different symbols whose first six characters are,
in fact, identical.

For example, the two symbols, GEORGEl and GEORGEZ, differ only in the seventh

character, so that the Assembler would treat them as being the same symbol, GEORGE
It is not necessary to define a symbol before it is used in an expression.
end of PASS 1, however.

They must be defined before the

Thus, one may refer to a number of registers by their address tags, and then

define the symbols later.

Parameter Assignments

A symbol may be assigned a value by means of a parameter assignment statement which looks like an alge-

The single symbol to the left of the equal sign is assigned the value of the expression on

braic statement.

the right. No space(s) or tab(s) may appear between the single symbol to the left of the equal sign and the

equal sign

Examples:

EXIT
C

JMP I 0

A + B

All symbols to the right of an "=" sign must be already defined.

The symbol to the left of the "=" sign

and its associated value is stored in the Assembler's symbol table.

The use of the "=" does not increment the current location counter.

Assembler itself rather than a part of the output binary.

The equal sign may be used to redefine a symbol.

Symbol Definition
A symbol may be defined by the user in one of three ways:

By use of a parameter assignment. Example:
DlSMlS

JMP RESTOR

5-6

It is, rather, an instruction to the

As a macro name.

Example:

DEFINE LOAD A
< CLA

TAD A >

3.
a

By use of the comma. When a symbol is terminated by a comma, it is assigned

value equal to the current location counter.

Example:

/SET CLC T0100

*100

TAG,

CLA

B,
A,

JMP A
DCA B

The symbol "TAG” is assigned a value of 0100, the symbol "B" a value of 0102, and the symbol "A" a

value of 0103.

Expressions
All elements, i.e. , symbols and numbers (exclusive of pseudo—instruction symbols, macro names, and
double precision or floating point constants) may be combined with certain operators to form expressions.
These operators are:

This signifies 2's complement addition (modulo

plus

409610).
-

This signifies 2's complement subtraction (modulo

minus

409610).
I

exclamation point

This signifies Boolean inclusive OR (union).

ampersand

This signifies Boolean AND (intersection).

u—I

space

Space is interpreted in context.

It may signify

inclusive OR or act as a field delimiter.

Symbols and integers may be combined with any of the above operators.
ated from left to right;

A symbolic expression is evalu—

grouping of terms is permitted. Example:

A+ B

A-B

A 1 B

0005

7777

_A_§§

Value

0002

0003

0002

Value

0007

0005

0014

0002

0007

0005

Value

0700

0007

0707

0671

0707

0000

The MACRO-8 Assembler makes a distinction between the types of symbols it is processing.
are

These types

i) permanent symbols, 2) user defined symbols, and 3) macro names. The character "space" is inter—

preted written in the context of the expression.

If a space is used to delimit two or more permanent sym—

bols, space signifies inclusive OR. Example:
CLA

is a permanent symbol whose value is 7200.

CMA

is a permanent symbol whose value is 7040.

The expression:

has a value of 7240.

CLA .__,CMA

If the symbol following the space is a user defined symbol, space acts as an address field delimiter.

Example:
*2117

CLA

JMP

._.

”A" is a user defined symbol whose value is 2H7.

The expression JMP

._.

A is evaluated as follows:

The seven address bits of A are taken, i.e.:
A

010 00]

001

ill

001

iii

The remaining five bits of A are tested to see if they are 0's (page 0 reference); if} they are not, the
current page

bit is set.

000

001

iii

The operation code is ORed into the expression:
000

000

AddressA 000 01]

DOT

iii

001

ill

JMP

or, written more

5317

101

iOi

Oil

concisely:

000

In addition to the above outlined tests, the page bits of the address field are compared with the page bits

of the current location counter.

If the page bits of the address field are nonzero and do not equal the

page bits of the current location counter,

out-of-page reference is being attempted. If the reference

is to an address not on the page where the instruction will be

located, the Assembler will set the indirect

bit (bit 3) and an indirect address linkage will be generated on the current memory page.
page reference is already an indirect one, the error diagnostic II

(Illegal Indirect) will be typed on

When the link is generated, the LG (Link Generated) message will be typed on PASS 2.

PASS 2.
case

If the out-of—

In the

of several out-of-page references to the same address, the link will be generated only once, but the

LG message will be printed each time.

Example:

*2117
CLA

*2600
JMP

._.

The space preceding the user defined symbol "A” acts as an address field delimiter.

The Assembler will

recognize that the register tagged "A" is not on the current page (in this case 2600—2777) and will
generate a link to it as follows:

in location 2600 the Assembler will place the word

5777 which is JMP I 2777
in address 2777 (the last location on the current page), the word 2i I7 (the actual
address of ”A") will be placed.

The address field of a memory reference instruction may be any valid expression.

Example:

A=27O
*

200

TAD A—20

would

produce, in location 200, the word
OOI OIO 101 000 or I250 (TAD 250)

Although the Assembler will recognize and generate an indirect address linkage when necessary, the
programmer may indicate an explicit indirect address by using the special symbol I'I"

between the operation code and the address field.

This must be

The Assembler cannot generate a link for an instruction

that is already specified as being an indirect reference.
II

(Illegal Indirect).
5—9

In this case, the Assembler will type the message

Current Address indicator

The single character period (.) has, at all times, a value equal to the value of the current location
It may be used as any integer or symbol

counter.

(except to the left of an equal sign). Example:

*200

JMP .+2

ls equivalent to JMP 202.
*

300

+2400

would produce, in register 0300, the quantity 2700.
*2200
CALL=JMS l

0027

Since the second line, CALL=JMP l

does not increment the current location counter, 0027 would be

placed in register 2200 and CALL would be placed in the symbol table with an associated value of
100 110 000 000 or 4600.

Origin Setting
The origin (current location counter) is reset by use of the special character asterisk (*).

location counter is set to the value of the expression following the “*".
0200.

All symbols to the right of "*" must already have been defined.

The current

The origin is initially set to

Example:

lf D has the value 250

then
*D+lO will set the location counter to 0260.

To simplify page handling, the pseudo-instruction PAGE may be used.

the origin is reset to the first location of the next page.

A page number may be specified by a legal

expression following the page pseudo-instruction. Example:
*

at this

270

point, either

When "PAGE" is encountered,

*400

PAGE

PAGE 2
will reset the origin to 0400.

Literals
Since the symbolic expressions which appear in the address part of an instruction

usually refer to the loca—

tions of registers containing the quantities being operated upon, the programmer must explicitly reserve

the registers holding his constants.

The MACRO—8 language provides a means for using a constant directly.

Suppose, for example, that the programmer has an index which is incremented by two. One way of coding
this operation would be as follows:

cL'A
TAD INDEX
TAD C2
DCA INDEX

C2, 2

Using a literal, this would become

CLA
TAD

INDEX

TAD

(2)

DCA

INDEX

The left parenthesis is a signal to the Assembler that the expression following is to be evaluated and as-

signed a register in the constants table of the current page.
address linkages are stored.

This is the same table in which the indirect

In the above example, the quantity 2 is stored in a register in a list begin-

ning at the top of the memory page (page address T77), and the instruction in which it appears is encoded
with an address referring to that address.

A literal is assigned to storage the first time it is encountered;

subsequent references will be to the same register.
If the programmer wishes to assign literals to page 0 rather than the current page, he may use square

brackets, "l" and ”] ", in place of the parentheses.
may be omitted.

However, in both cases, the right of closing member

The following examples are acceptable:

TAD

AND

(777)
UMP)

Note that in the second example, the instruction AND [JMP has the same effect as AND [5000.

Literals may be nested.

For example:

*200
TAD

(TAD

(30

will generate
0200

i 376

0376

i 377

0377

0030

This type of nesting may be carried to as many levels as desired.
at page address

page 0).

T77 and extending toward page address 0.

Literals are stored on each page starting

(Only 12710 or 1778 literals may be placed on

If a literal is generated for a nonzero page and then the origin is set to another page, the current

page literal buffer is punched out (during PASS 2).

If the origin is then reset to the previously used page,

the same literal will be generated it used again.

Single Character Text Facility
It a single character is preceded by a double quote, the 8—bit value of the ASCII code for the character
is inserted instead of taking the letter as a symbol.

CLA
TAD

("A

will place the constant 0301 in the accumulator.

Example:

CHAPTER

PSEUDO-INSTRUCTIONS

The pseudo-instructions are directions to the Assembler to perform certain tasks or to interpret subsequent

coding in a certain way. By themselves pseudo-instructions do not generate coding or (in general) effect
the current location counter.

The functions of each pseudo-instruction are described in this chapter.

CURRENT LOCATION COUNTER
PA GE

This pseudo-instruction is used to set the current location counter.

PAGE n

This will reset the current location counter to the first address of page n, where n is an

integer, a previously defined symbol, or a symbolic expression. Examples:
PAGE 2 will set the CLC to 0400
PAGE 6 will set the CLC to 1400

PA GE

When used without an argument, PAGE will reset the CLC to the first location on the
next succeeding page.

Thus, if a program is being assembled into page i and the pro-

grammer wishes to begin the next segment on page 2, he need only insert the pseudo-

instruction PAGE,

follows:

JMP

.—7

PAGE

CLA
The current location counter may be explicitly set by use of the asterisk.

EXTENDED MEMORY
On PDP—8’s equipped with more than one memory bank, the pseudo~instruction

FIELD n

may be used where n is an

integer, a previously defined symbol, or a symbolic expression

within the range OSnS7.
This pseudo-instruction causes a word of the form

XXX

000

where

OOOSXXXS iii

to be

punched on the binary tape during PASS 2.

This word is interpreted by the Extended Memory Binary

Loaders (see Digital-8-2A-U; Digital-8—2B-U).

RADIX CONTROL

Normally, all integers used in a program are taken as octal numbers. If, however, the programmer wishes
to have certain

numbers treated as decimal, he may use the pseudo-instructions:

DECIMAL

When this pseudo-instruction occurs, all integers encountered in subsequent coding will
be taken as decimal until the occurrence of the pseudo—instruction

which will reset the radix to its original (octal) base.

OCTAL

NUMBERS

The types of numbers allowed are integers (See Chapter 5), double precision integers, and double precision Floating point numbers.

Double Precision Integers
Double precision integers may be positive or negative (2's complement) according to their sign but may
not be

combined with operators.

not disturbed.

They are always taken as decimal radix although the current radix is

Each double precision integer is allotted two consecutive registers with the sign indicated

by bit 0 of the first word.
The double precision integer mode is entered through the use of pseudo—instruction DUBL and all numbers
encountered will be taken as double precision integers until an alphabetic character is encountered.
Each number is terminated by the carriage return

*400)
679467)
44)
-3)

DUBL

TAG,

CLA)

would produce
0400
040]

0245
7053

0402

0000

0403

0054

() )

the semicolon (;) or by a comment.

Example:

0404

7777

0405

7775

0406

7200

and the symbol ”TAG" would have a value equal to 0406.

Floating Point Constants
Double precision floating point constants may be positive or negative according to their sign but may not
be combined with operators.

Decimal radix is assumed but the current radix is not altered.

point constants are each assigned three registers and are stored in normalized form.

Floating

(See Digital-8-5-S

for a description of floating point arithmetic.)
The double precision floating point mode is entered through use of the pseudo-instruction FLTG.

All

numbers encountered after the use of FLTG will be taken as double precision Floating point constants until
the occurrence of an alphabetic character other than E

The general input format of a floating point

number is

iddd -ddddEidd

where the d's are decimal digits.

Any character which is not legally part of the above format (except

rubouts) terminates input of the number. Example:
Produces

*400
F LTG

+509 32E
.

—62 97E
.

l .00E

TAGZ,

-02)

04)

-2)

CLA)

0400

0003

0401

2427

0402

6670

0403

0024

0404

5462

0405

0740
7772

0406
0407

2436

041 0

5574

041 i

7200

and the symbol ”TAGZ" would be assigned a value of 04]]

TEXT FACILITY

There is a text facility for single characters and text strings.
mode (double quote), see Chapter 5.

6-3

For a description of the single character

A string of text may be entered by giving the pseudo—instruction TEXT followed by a space, a delimiting

character, a string of text, and repeating the same delimiting character. Example:
ATEXTA

TEXT

The character codes are stored two to a register in ASCII code that has been trimmed to six bits.
the last character, a 6-bit zero is inserted as a stop code.

Following

The above statement would produce

2405
3024
0000

TEXT

._.

/BOB/

would produce
02] 7
0200

The TEXT pseudo-op could also be used as part of a calling sequence to a subroutine:
JMS MESS
TEXT /

JMS MESS

/NO WDS IN MESSAGE
/ADDRESS OF MESSAGE

NOWDS
ADDMESS

ADDMESS,

TEXT /

Note that while the TEXT pseudo-instruction causes characters to be stored in a trimmed code, the use of

the single-character control code (") causes characters to be stored as a Full 8-bit ASCII code.

END OF PROGRAM

The special symbol "$" indicates the end of a program.
nates

the PASS

When the Assembler encounters the "$" it termi-

END OF TAPE

When several tapes are to be assembled together, each, except the last (which ends in "$"), should have
as

its last symbol the pseudo-instruction PAUSE.

and halt the computer.

CONTINUE

This causes the MACRO-8 Assembler to stop processing

After placing a new tape in the reader, assembly can be continued by depressing

ALTERATIONS TO THE SYMBOL TABLE
There are two pseudo-instructions that may be used to alter the permanent symbol table (during PASS l):
EXPUNGE

EXPUNGE the entire symbol

FIXTAB

FIX the symbol TABIe.

table, except For the pseudo-instructions.

All symbols that are currently in the symbol table are fixed.

Example:
CLSF=6l4l
FIXTAB

would define CLSF as a permanent symbol

EXPUNGE

TAD=l 000
FIXTAB
would place the symbol TAD in the assembler's permanent symbol table.
been expunged

6-5

All other symbols would have

CHAPTER

MACROS

When writing a program, it often happens that certain coding sequences are used several times with just
the arguments changed.
ment.

If so, it is convenient if the entire sequence can be generated by a single state-

To do this, the coding sequence is defined with dummy arguments as a macro.

A single statement

referring to the macro by name, along with a list of real arguments, will generate the correct sequence
in line with the rest of the coding.

The macro name must be defined before it is used.

The macro is defined by means of the pseudo-instruction

DEFINE followed by the macro's name and a list of dummy arguments.

For example:

A macro to move the contents of register A to register B and also leave the result in the

accumulator, would be coded as follows:
DEFINE
<CLA

MOVE

TAD

DUMMY]

DCA

DUMMY2

TAD

DUMMY2>

DUMMY]

DUMMY2

The actual choice of symbols used as dummy arguments is arbitrary; however, they may
not be defined or referenced prior to

the macro definition.

The above definition of the macro MOVE is identical to the following:
DEFINE

<CLA;TAD

MOVE

ARGI;

ARGIH ARGZ
DCA ARGZ;

TAD ARGZ>

The actual definition of the macro is enclosed in angle brackets.

When a macro name is processed by the assembler, the real arguments will replace the dummy arguments.
For example:

Assuming that the macro MOVE has been defined as above,
*400

A,0
B, —6

0400

0000

0401

7772

MOVE

A, B

0402
0403
0404

7200
1 200
3201

0405

1201

A macro need not have any arguments:

NOTE:

coding

to rotate

as a macro

For example, a sequence of

the C(AC) and C(L) six places to the left might be encoded

by means of

DEFINE.__. ROTL 6

<RTL; RTL; RTL>
The entire macro definition is placed in the Macro Table, two characters per word, with a dummy argu-

replacing the symbolic name. Example:

ment value

DEFINE

LOAD._. A

<CLA
TAD A>
is stored, in the Macro Table,

ICLlA)| TAl

0..

roughly as follows:

l7700|>00|

where the vertical lines indicate successive 12-bit words.
are

Comments and line-feeds

not stored.

The macro definition can consist of any valid coding except for TEXT or

type statements.

RESTRICTIONS
l

Macros cannot be nested; i.e. , another macro name or definition cannot appear in
macro

definition and cannot be brought in as an argument at reference time.

type statements cannot appear in a macro definition.

TEXT or

Arguments cannot be:

Macro name

TEXT pseudo-instruction or

‘'

special character

The symbols used as dummy arguments must not have been previously defined or

referenced

A macro may not be redefined.

DEFINE

LOOPH A

<TAD

Example:

DCA

TAD

COUNT

ISZ

JMP

.-2>

The symbol "COUNT" is not a dummy argument but an actual symbol.

A macro is referenced by giving the macro name,

by commas.
macro

space, and then the list of real arguments,

separated

There must be at least as many arguments in the macro reference as in the corresponding

definition.

When a macro is referenced, its definition is found, expanded, and the real arguments

replace the dummy arguments.

The expanded macro is then processed in the normal fashion.

LOOP_ x, Y2
.....

is equivalent to:

TAD

DCA

TAD

COUNT

ISZ

JMP

.—2

The macro table shares the available space (60410 registers) with the
table
Thus the programmer must be aware of the amount of room resymbol
quired by his macros and the fact that each symbol occupies four words of memory. Also, the arguments of a macro call are temporarily stored in this buffer
space while the macro is being expanded.
NOTE:

CHAPTER

ERROR

DIAGNOSTICS

The format of the error messages is:

ERROR CODE

ADDRESS

Where ERROR CODE is a 2-character code which specifies the type of error, and ADDRESS is either the

absolute octal address where the error occurred or the address of the error relative to the last symbolic
tag (if there was one) on this page.

Assembly will continue or may be continued after all errors except SE (Symbol Table Exceeded).
error

occurs,

If an SE

the Assembler will halt and may not be restarted.

ERROR MESSAGES
PE

Current, Non-Zero Page Exceeded
An attempt was made to

override a literal with an instruction or

override an instruction with a literal.

This can be corrected by

decreasing the number of literals on the page

decreasing the number of instructions on the page

Zero Page Exceeded

Same as PE only with reference to page 0

Illegal Redefinition of a Symbol
An attempt was made to give a previously defined symbol a new value not via the "="

The

symbol was not redefined. (This is similar to the Duplicate Tag diagnostic of PAL lll).
lllegal Character
l.

# % '

? @\

were

processed neither in a comment nor a TEXT field. The character

is ignored and the assembly continued.

8-]

A non-valid character was processed.

The computer will halt with the illegal character

displayed in the accumulator. Assembly may be continued by putting the desired character
in the SWlTCH REGISTER and depressing CONTINUE

lllegal Equals
An equal sign was used in the wrong context.

TAD
A

A +
B

Examples:

(The expression to the left of the equal sign is not a single symbol)

lllegal Indirect
An out of page reference was made, and a link could not be generated because the indirect

bit was already set.
*

Example:

200
TAD

PAGE

CMA

A,
LG

CLL

Link Generated
A link was generated for an out of page reference at this address.

*200

Generated Binary

TAD

Example:

0200

i 777

0377

0400

7140

PAGE

A,
SE

CMA,

CLL

Symbol Table Exceeded
The Symbol Table overlaps the Macro Table or vice versa.

Assembly is halted and cannot

be continued.

lllegal Format in 0 Macro Definition
The expression after the DEFINE pseudo-instruction does not comply with the macro definition,

position, or structural rules. Example:
A macro name is referenced before the macro definition.

8—2

Undefined Symbol
A symbol has been processed during PASS 2 that was not defined by the end of PASS 1

Missing Parameter in 0 Macro Call
An argument, called for by the macro definition, is missing.

Example:
MAC

DEFINE
< TAD

CIA
DCA
MAC

A
B >
SUM

Two MACRO—8 internal tables have overlapped.

This situation can usually be corrected by de-

creasing the number of current page literals used prior to this point on the page.

If the error

persists, please contact the Small Computer Systems Programming Group at Digital Equipment

Corporation for assistance

CHAPTER

OPERATING

INSTRUCTIONS

MACRO-8 is a 2—pass assembler with an optional third pass which produces an octal/symbolic assembly

listing. During the first pass, MACRO-8 processes the source tape and places all symbol definitions and
macro

definitions in its symbol table and macro table,

respectively. During the second pass, MACRO—8

processes the source tape and punches the Binary Format Tape.

At the end of PASS 2, MACRO-8 prints

This punched table may be

the Symbol Table (it is also punched if the 33—ASR PUNCH is turned on).
read by DDT (See Digital-8-4—S)

The third pass provides a listing of the generated octal code and the

original source language.
There are two versions of MACRO—8 which differ with respect to their use of input/output equipment:

the low speed version uses the 33-ASR Reader for all input and the 33-ASR Punch for all output; the high

speed version uses the Type 750 Photoelectric Reader for all input, the Type 75 High Speed Punch for

binary output, and the 33—ASR Punch for printable output such as error printouts, symbol table punching
and listing, and third pass assembly listing.

NOTE: In the high speed version of MACRO—8, the Type 75 Punch may be
used as the printable output device by changing the contents of location 0004
from 2600 to 0600.

This is useful for long third pass listings, since the punched
It is advised that

output from the 75 Punch can be subsequently listed off-line.

this change not be made until pass 3, so that pass I and pass 2 error messages
will come out on the 33-ASR.

Load MACRO—8 with the Binary Loader (See Digital-8—2-U)

Put the source tape in the reader.

Set the SWITCH REGISTER to 0200.

Depress LOAD ADDRESS

Set switch options (See Table 9-1)

Depress START.

Turn on the 33-ASR reader (if using the low speed version).

When MACRO-8 stops reading (after processing a PAUSE statement), place the

next tape in the reader and depress CONTINUE

been processed

Repeat this step until all tapes have

When MACRO-8 encounters the terminating character, dollar sign ($), it performs

one

of the following sets at events depending upon what pass has just been completed.

Proper operator intervention is then required.

Pass Just Completed
I

Events

Operator Intervention

Set up For PASS 2

Turn on 33-ASR punch (in high speed
MACRO-8, symbol table is output via
33-ASR) Put source tape in reader;
.

hit CONTINUE to enter PASS 2.

(a) It PASS 3 is desired:
(In the high speed version of MACRO—8,

Terminate current assembly;
punch out page 0 constants,
checksum and trailer code on

the contents of register 0004 could be

binary tape; print and punch
rubout, the alphanumerically
ordered symbol table, an EOT

altered at this point to change output

code, a rubout, and trailer
code; Set up for PASS I.

switch 3 up at step 5.

devices).

Go to step 2 of the operating

instructions, making sure to set AC

(b)

It PASS 3 not desired:

Turn off 33-ASR punch, put next program
to be assembled in the reader. Hit
CONTINUE to enter PASS I.

Terminate assembly listing:

Turn off 33-ASR punch; put next

Setup for PASS I

program to be assembled in the reader;
hit CONTINUE to enter PASS I

TABLE 9-]

SWITCH OPTIONS

Switch Up
None

Result

MACRO-8 will enter the next pass as defined in the preceding table.

example:

For

if the previous assembly was terminated during or at the end of PASS

restarting MACRO-8 with no switches up would cause PASS 2 to be entered.
MACRO-8 initially starts at PASS I

Restore symbol table to only the pemanent symbols and enter PASS I

Enter PASS 2

TABLE 9—1

SWITCH OPTIONS (continued)

Switch Up

Result

Enter PASS l without erasing any previously defined symbols.

Enter PASS 3.

During PASS 3, MACRO-8 outputs an octal/symbolic listing

of the assembled program.

If this pass is terminated before completion, either

switch options 0 or 2 may be used to return to PASS 1 for subsequent assemblies.
MACRO-8 will output as much of the source statement (symbolic) as its internal

storage capacity will allow.

Because of the internal operations during the

processing of macro statements, the symbolic output may be meaningless.
The double precision integer and double precision floating, point processors

are

deleted and may be used for storage of user defined symbols.

creases

the size of the symbol table by

6410 symbols.

The macro processor and the number processors (above) are deleted and may be
used for storage of user defined symbols.

table by

NOTE:

Switches 10 and ll

to handle

This in-

This increases the size of the symbol

l2510 symbols.

are

sensed whenever PASS l is entered.

MACRO-8 would have to be reloaded

subsequent programs that use macros, double precision integers,

floating point numbers.

The Binary Format Tape produced during PASS 2 may be loaded by the Binary Loader.

When the loading

is completed, the accumulator should contain zero which indicates that it has loaded correctly.

The PASS 3 output is of the following format:
AAAA

NNNN

(Symbolic)

CR/LF

Where AAAA is the absolute octal address and NNNN is the generated code.
of phase with the octal

Example:
*200

0200
020]
0376

1377

TAD (1

3776

DCA A
*4000

4000

0377

0001

4000

0000

A, 0

9—3

Literals are somewhat out

SYMBOL TABLE MODIFICATION
Because of the small amount 6f core
macro

(60410 registers) remaining to be used for programmer symbols and the

table, the Following suggestions are offered which may allow a particular installation or individual

to conserve on

table space.

By use of the pseudo-ops EXPUNGE and FIXTAB, unnecessary instruction mnemonics can be removed from
the symbol table thus making more space available for programmer defined symbols and macros.

This also

decreases assembly time as the never used instruction symbols are not involved in the symbol table searches.

The most often used instruction mnemonics should be assembled first, so that they will be in core next to

the special characters and pseudo-instructions.

This is desirable because the symbol search routine starts

searching at the top of the table and works down.
At an installation that does not have a piece of optional equipment available, the corresponding instruction
set can be removed.

A symbolic tape beginning with EXPUNGE, containing all necessary instruction mne-

monics, and ending with FIXTAB and the $ sign could be assembled (only PASS I is needed) by MACRO-8

prior to any other assemblies. Example:
EXPUNGE

AND=0000
TAD=I 000
C LA=7200

FIXTAB

(The pseudo-op PAUSE could also be used with this tape, the
first of a multiple tape assembly.)

APPENDIX

MACRO-8

/MEMORY REFERENCE
AND=OBOO

SYMBOL TABLE

INSTRUCTIONS

KCC:6032
KRS:6934
KRB:GO$6

TAD=IZOO
ISZ:2000
DCA:50@D

/TELETYPE TELEPRINTER/PUNCH
TSF:GOAI
TCF:6042

JMS=AOOO

JMP=SOOO
IOT:GOOO
OPR:7000
/MICROINSTRUCTIONS
NOP:7000

TPC:6044
TLS:6946

/FLOATING POINT INTERPRETIVE COMMANDS
FEXTZOOOO
FADD:IOOO
FSUB:2000
FMPY=SOOO
FDIV:ADDO
FGET:5§OO
FPUT:6000
FNOR:7000
/OSCILLOSCOPE AND PRECISION CRT

CLA:7ZOO
CLL271OO

CMA:7@4D
CML:702@
RAR27OIO
RTR:7@12
RAL:7004
RTL=7006
IAC:7ZOI
SMA=752O

fDISPLAY
DCX:6051
DXL:6953
DCY:6061
DYL=6065
DIX:6@54
DIY:6664
DXS:6057
DYS:6067
DSF:6071
DCF:6072
DLB=6074
/INCREMENTAL PLOTTER
PLSF:SSOI
PLCF:6502

SZA=7440
SPA:75IO
SNA=7450
SNL:7420
SZL:7460
SKP=74lO
OSR:7404
HLT=7492
/COMBINED
CIA=7041
LAS=7604
STA=7240
STL=7120

MICROINSTRUCTIONS

PLPU:6504

PLPR=65II
PLDU:6512
PLDD:6514
PLUD=6522
PLPL:6521
PLPD:6524
/LINE PRINTER
LCF:6652
LPR:6655
LSF:6661
LCB:6662
LLD:6664
/CARD READER AND CONTROL
CRSF=6652
CERS=6654

GLK:7204
/PROGRAM INTERRUPT
ION:6001
IOF:6002
IANALOG TO DIGITAL CONVERTER
ADC:6@OA
/HIGH SPEED PERFORATED TAPE READER
RSF=6011
RRB:60!2
RFC=SOI4
/HIGH SPEED PERFORATED TAPE PUNCH
PSF:6021

PCF:6@22
PPC:6024
PLS=6026
/TELETYPE KEYBOARD/READER
KSF=6051

CRRB:667I
CRSA:6672

Al-l

CRSB:6674
/CARD PUNCH CONTROL
CPSF:6631
CPCF:664|
CPSE=6542
CPLB=6644
/AUTOMATIC MAGNETIC TAPE CONTROL
MSCR:67OI
MCD:6702
MTS=6706
MSUR=6711
MNC:6712
MTC:6716

MSWF:6721
MDUF:6722
MCWF:6722
MEUF:6722

MIWFz6722
MSEF:6751

MDEF:6752
MCED=6732
MEEF:6752
MIEF26752

MTRS=6734
MCC:6741
MRWC:6742
MRCA:6744
MCA=6745
/COMBINED

MMMM:6757
MMMF:6757
/AUTOMATIC

INSTRUCTIONS

MULTIPLY-DIVIDE

RDAzsllZ
MUL=6111
DIV:6121

SZO:6114
SAF:6124
RDM:6122
/MICROTAPE INSTRUCTIONS
MMLS:6751
MMLM=6752
MMLF:6754
MMSF:6761
MMCF:6772

MMSC=6771
MMCC:6762

/CONVERTER

DIGITAL

ADSF:6551
ADCV=6552
ADRB=6534
ADCC=6541
ADSC=6542
ADIC=6544
ISERIAL MA GNETIC DRUM SYSTEM
Dch:sszs
DRcw=ssa5
DRCF:6611
DREF:6612
DRTS:6615
DRSE=6621
DRSC=6622
DRCN:6624
xzxrsnnzn ARITHMETIC ELEMENT
muy:74zs
DVI:7407
NMI:74ll
SHL:7415
A5R:74xs
LSR:7417
MQL:7421

CAM=6151
LAR:6104
LMQ26102

MMRS=6774

MMLC:6766
MMML:6766
IMEMORY PARITY
smp:slal
cmpzsnaz
ITYPE 133/ 139 ANALOG TO

SCA:744I
MQA:7sz1
/MAGNETIC TAPE SYSTEM
I1Fm:s7a7
TSRD:6716
Isz:s715
TSDF=6721
ISSR=6722
TSST:6724
TSRS:6754
rwnrzsvsx
TCTI:6732
/MEMORY EXTENSION
RDF:6214
RIF:6224
RMF:6244
RIB:6234

CDF:6201
CIF=6222

APPENDIX

ASCII

CHARACTER

SET

These characters may be used in symbols.
Character

8-Bit From

6—Bit From

Character

301

302

303

8-Bit From

6-Bit From

323

324

325

304

326

305

327

306

330

307

331

310

332

31 1

1 1

260

312

261

313

262

314

263

315

264

316

265

317

266

320

267

321

270

322

271

These characters are special.

Character

8-Bit From

6-Bit From

Meaning

243

Inclusive OR
Character pseudo-instruction
Illegal outside of (TEXT) or (") or comment

244

End of PASS

245

246
247

46
47

Illegal outside of (TEXT) or (") or comment
Logical AND
Illegal outside of (TEXT) or (") or comment

250

Define literal

251

Terminate literal

252

Set origin

241

242

‘

(
)
*

+
,
-

.
g

/
'

253

2'5 complement addition

254

Define symbol

255

2'5 complement subtraction

256

Has value of CLC

257

Start of comment

272

Illegal outside of (TEXT) or (") or comment

A2- 1

Character

8-Bit From

273
274

Meaning

Terminate expression

74
75

Start macro definition
End macro definition

277

76
77

300

Illegal outside of (TEXT) or (") or comment
Illegal outside of (TEXT) or (") or comment

333

Define page 0 literal

334
335

34
35

End page 0 literal

336
337

36
37

Illegal
Illegal

275

276

"‘/'—@-ov

6-Bit From

Define Parameter

Illegal outside of (TEXT) or (") or comment

Line-feed

212

Used for formatting (ignored)

Return

2I5

Terminate line

Space

240

Address delimiter or inclusive OR

Rubout
Form-feed
Blank
Code 200

377

Ignored
Ignored
Ignored
Ignored

214
000
200

EJflEflIEII

EQ U l P M E N T
CORPORATION
MAYNARD. MASSACHUSETTS

PRINTED IN USA.