Atlas Manual

Overview

Preface

Contents

1. Introduction

2. Atlas 1

3. Accumulator

4. B-Registers

5. Routines

6. Accumulator (cont)

7. Extracodes

8. I/O

9. Magnetic Tapes

10. Programs

11. Monitoring

12. Tuning

Appendices

Return

This site is best viewed using a browser with either native SVG support or an SVG Plug-in

Access Key Details

ABL: APPENDICES

Appendix A References

CS 271	Mercury Orion Atlas Autocode
CS 294A	Operating Instructions for the Teletype Punch
CS 294A	Punched Tape Codes
CS 298A	Atlas Magnetic Tape
CS 308B	Punched Tape Codes (5- and 7-track)
CS 309A	Extracode Functions
CS 318B	Making a Fortran II Program suitable for use with the Atlas Fortran Compiler.
CS 339	Operating Instructions for the Model 'B' Flexowriter
CS 345	List of Atlas Basic Instructions
CS 349A	Atlas Programming Exercises
CS 360	Telex Data Links
CS 361	Keyboard used with Ferranti Computers
CS 362	Operator's Conventions for Punched Tapes
CS 363	Specifications of Paper for Punched Tapes
CS 364	Specifications of Dimensions for Punched Tapes
CS 365	Creed Teleprinter with Integral Tape Reader and Punch
CS 366	Model 'S' Flexowriter
CS 367	Creed Keyboard Punch for 7-track tapes
CS 368	Creed Verifier for 7-track tapes
CS 377	Algol 60 Report
CS 378A	Reference manual for Atlas Algol (provisional)
CS 379A	A primer of Algol 60 for Atlas
CS 384	Summarised programming information
CS 390	Primer of Fortran Programming
CS 401	The Analysis of Plane Structural Frames
CS 402/5052	Extended Mercury Autocode for Orion and Atlas
CS 405/5055	Traffic Assignment
CS 411	The Atlas 1 Computer System Operators' Manual
CS 428	Atlas user's description of the L.P. input scheme
R 40	An Assembly programme for a Phrase-Structure Language
R 55	The Manchester University Atlas operating system and Users' description.
R 58	The Atlas supervisor
R 67	One-level storage system
R 68	The Atlas scheduling system
R 70	Processing commer~ial data on Atlas
R 74	Central Control Unit of the Atlas Computer
R 76	The Compiler Compiler

Appendix B: Notation

Most symbols are used with two completely different meanings. The interpretation to be given to a symbol depends on its context. A convenient division is whether it is used in describing the arithmetic or the basic language, so the notation is listed under these two headings.

1. ARITHMETIC

Lower-case letters are used for suffices and for the content of a location, the location being in upper-case letters. The result of an operation is denoted by a prime. Thus, s is the content of address S, and am' = am + s means the content of Am after the operation is equal to the content before plus the content of S.

(a) Suffices

x

the argument of the prefixed location

y

the exponent of the prefixed location

:

two consecutive registers, starting with the one suffixed

*

the register following that suffixed

(b) Accumulator

A

the full double-length accumulator, holding ax
79-bit mantissa ax and 8-bit exponent ay

al

the 48-bit floating-point number formed from l, ls and ay

am

the 48-bit floating-point number formed from m and ay

aq

the single-length number which is obtained by standardising, rounding and truncating the contents of the accumulator

L

the less-significant half of Ax, of 39 bits with no sign

Ls

the sign bit associated with L

M

the more-significant half of Ax, of 40 bits, the most-significant bit being the sign digit

AO

accumulator overflow

DO

division overflow

E or EO

exponent overflow

Q

standardised

R

rounded, by forcing a 1 in the least-significant digit of M if L is not clear

TR+

rounded, by adding a 1 to the least-significant digit of M if the most-significant bit of L is a 1

(c) General

B

any B-register (index register)

Ba

the B-register specified by the Ba digits of an instruction

Bc

the B-carry digit

Bm

the B-register specified by the Bm digits of an instruction

Bt

the B-test register

C

the main control register (B127)

C()

the contents of the location specified within the brackets

E

the extracode control register (B126)

F

the function digits of an instruction

G

the logical accumulator (B98 and B99)

I

the interrupt control register (B125)

N

the unmodified address part of an instruction (a 24-bit number with the point one octal place from the least-significant end).

n

the modified address part of an instruction regarded as a 24-bit number with the point one octal place from the least-significant end

S

the address of a store location. A full-word address in accumulator instructions (digits 21-23 ignored), a half-word address in B-register instructions (digits 22 and 23 ignored), or a 6-bit character address (all digits relevant)

V

the V-store

B

register a of the V-store

B

signifies extracodes suitable for fixed-point working

2. BASIC LANGUAGE

In practice no distinction is made between capital and small letters, t hough capitals are used here. However, as an aid to clarity, it is sometimes advantageous to use lower case letters for the separators m, n, v, x and q. Small Greek letters α, β, are used for 21-bit decimal integers, k for the octal number in the 3 least-significant digits of a 24-bit address, and ρ for a general octal nunber of up to 8 octal digits.

Aα

Parameter α, (0 ≤ α ≤ 3999), of the current routine

B

An operator in an expression, causing bits 12-23 of the previous element to be set to zero i.e. it gives a "Block Address"

C

Introduces a string of characters on the next line which are translated into internal code and placed in successive character positions

Dα

An operator, causing the previous element to be logically shifted down α places

E

The enter directive, causes the compiler to evaluate any used parameters etc., insert library routines, delete the compiler and enter the program

ER

As E but the compiler is not deleted and may be used again

EX

The enter interlude directive, to enter a short program for any reason during the compiling process

F

Introduces one or more floating-point numbers on a line, after some expressions otherwise interpreted. Also, if necessary, increases * to the next full-word address

Gα

Global Parameter α (0 ≤ α ≤ 3999)

H

Subsequent expressions on a line are interpreted as 24-bit words and, if necessary, * is increased to the next half-word address

Jσ

σ is octally justified to the left, i.e. the most significant digit goes into bits 0-2, the next into 3-5 etc.

Kσ

σ is octally justified to the right, to bit 20, i.e. the least-significant digit goes into bits 18-20, the next into 15-17 etc.

Kσ.k

As Kσ, with k going into bits 21-23

Lα.k

Library routine number α, copy k. (.k is omitted if only one copy of the routine is wanted) ( α = 1 to 1999, k = 1 to 1999 )

M

Alternative to &

N

A separator which non-equivalences the element before it with the element after it in an expression

Pα

Preset parameter α (0 ≤ α ≤ 99 normally)

Q

A separator in an expression which divides the element before it by the element after it, placing the result in digits 0-20

Rα

A directive defining the beginning of routine α (0 ≤ α ≤ 3999)

S

Expressions after the directive S are interpreted as 6-bit characters, i.e. only bits 15-20 of the expression are used

Tα or Tα-β

The title is copied to output channel α or channels α to β inclusive

U^α

An operator causing the previous element to be logically shifted up α places

Uα or Uα-β

Unset preset parameter α, or α to β inclusive

V

A separator in an expression. The element befor it is OR-ed with the element after it

W

An operator in an expression, causing bits 0-11 of the previous element to be set to zero, i.e. it gives an address within a block

X

A separator. The element before it is multiplied by the element after it

Yα

α is placed in bits 0-23 (instead of bits 0-20)

Z

A directive indicating the end of a routine

*

The address of the first character position in the location where the item is placed. If used in an expression on the right-hand side of a directive then * is the address of the next character position

|

All subsequent characters up to NL are ignored (| is not a terminator)

£,π

Alternatives to |

[ ]

All characters between square brackets are ignored. Bracket nesting to any level is allowed

,

Alternative to multiple space as a terminator

&

A separator which logically ANDs the element before it with the element after

:

a special separator used in

1. an element α : β. α modulo 2¹² goes to bits 0-11 and β modulo 2²¹ to bits 0-20, added to α
2. a floating-point number N (α : β) : γ where the value of the number is N x 10^α × 8^β and the exponent is forced to γ or standardised if γ is omitted

/

1. in a parameter, / separates the parameter number and routine number
2. an alternative to :

' (prime)

an operator which forms the logical binary complement i.e. it replaces 1's by 0's and 0's by 1's

?

1. in the context A α ? = expression, A α is only set to the given expression if no other definite setting of A occurs before the porogram is entered. Similarly for G α ? = expression
2. in the context P α ? = expression, P α is set equal to the given expression unless P α is already set, in which case the directive is ignored
3. in the context ? expression, causes the compiler to ignore the remainder of the line if the value of the expression is zero.

Appendix C: V-Store Addresses of Peripherals

Each peripheral is allocated one or more words in a part of the V-store associated with its particular type of equipment.

A V-store address is identified by having 6 as its most significant octal digit; furthermore, since only the more significant half-words are used, the least-significant octal digit of the address is always zero.

That part of the V-store associated with the peripherals is the first 256 words of the block beginning with J6004.

To each type of equipment there corresponds 16 consecutive words, so that peripheral p of type q is allocated the V-store address

J6004 + 16q + p

The type number q is defined in the following table:-

q	Equipment	V-store address of equipment 0 of each type
0	Card Readers	J60040000
1	Spare (London only: High Speed data Link)	J60040200
2	TR7 Paper Tape Readers and N.E.P. Tape	J60040400
3	Graphical Outputs	J60040600
4	Anelex Line-printers	J60041000
5	I.B.M. Magnetic Tape	J60041200
6	Fast Paper Tape Punches	J60041400
7	TR5 Paper Tape Readers	J60041600
8	Teletype Punches	J60042000
9	Card Punches (and, Manchester only: X-Ray Diffractometer)	J60042200
10	Spare (Manchester only: A.T.& E. On-line Data Links)	J60042400
11	Teleprinters	J60042600

The addresses above are all for equipment 0 of the type indicated. Card Reader 1, for example, would be addressed by writing J60040010.

Appendix D: Character Codes

The following table lists all the available Atlas Internal Code characters together with their external representations in terms of punchings on 7-track tape, 5-track tape, and punched cards, using the standard Atlas character codes for these media.

The 7-track tape code is the I.C.T./Ferranti Orion/Atlas code. The 5-track code is the standard I.C.T./Ferranti code as used on Pegasus, Mercury, Sirius, Atlas and Orion. The card code is the Atlas Fortran card code.

Also in the table is an indication of which characters are available on the Anelex Line-Printer.

Some characters are designated "Unassigned". This indicates that no external printing characters have been assigned to these Internal Code characters. Most compilers and Input Routines treat these as Illegal Characters. However, these characters have had 7-track paper tape puncbings assigned to them. This serves two purposes: (i) it means that, if at some later date characters are assigned to these paper tape punchings, then Internal Code characters are available and assigned to correspond to them, and (ii) it means that, since there is some internal representation for every 7-track paper tape code with odd parity, it is possible to use Internal Code representation for parity-checked "binary" information, rather than use pure "Binary" mode. However, it should be noted that, since the Supervisor treats the shift characters and the New Line characters in a special manner, the internal representation will not be an exact "image" of the external punchings. It should also be noted that some of these Internal Code characters are used by non-standard external codes to represent non-standard characters. Thus in all cases special care should be taken in using these characters.

One character (Outer Set 02) is designated "Spare". This character has no external printing assigned to it nor any 7-track paper tape code. It may however be used by non-standard external codes, and thus care should be taken over its use.

Internal Code character 00 (Inner and Outer set) is designated "Not Assigned". This character is reserved for special purposes by the compilers and the Supervisor. It will never have a character assigned to it, and should never be used by normal programs.

Certain characters are not available as standard on any input medium, and may be treated as "spare" by input routines (for example, L100 treats them thus). They are

Outer Set 03 £ 
Outer Set 35 10 
Outer Set 36 11

They have the meaning as shown when used as output characters destined for the line-printer.

Certain other characters have alternatives given in parentheses. This is because the characters are alternative: in one or more of the relevant external codes. (For example Inner Set 13 is listed as π or £; these are alternatives on 7-track tape, 5-track tape and cards.) However, in the case of the line-printer, only the first character so listed is relevant, and, except for %, the other character appears elsewhere in the table for line-printer purposes.

Note that, in the column for 5-track tape, the 5-track tape code is given with the sprocket hole after two information holes and not three. This is the reverse of the normal convention and is done because the form as printed corresponds to the internal binary representation of the character when "Binary" mode for Input or Output is in use.

The Fault character (77 Inner Set) has no external representation. It is used under certain circumstances on Input by the Supervisor as a translation of any external character which has no Internal Code representation.

Tabulate (02 Inner Set) is treated as a single space by the Supervisor on output equipment where it does not othenvise exist.

The external shift characters (06 and 07) are ignored by the Supervisor for equipments where they have no relevance (e.g. the line-printer).

The fifteen symbols

: ' [ ] < > = _ | ? , 2 α β π

are on the fourth quadrant of the line-printer wheel. If none of these symbols are used in a line, the time to print the line is 1/1000 minute; otherwise it is 1/810 minute

Notation:

FS

Figure Shift

LC

Lower Case

LS

Letter Shift

UC

Upper Case

**

A paper-tape character appearing in both shifts

..

The character concerned is not available on this peripheral.

Internal Code - Inner Set

Character	Internal Code (octal)	7-track code		5-track code		Atlas Fortran card code (holes punched)	Anelex Line-Printer (availability
		Case	Binary Bits	Shift	Binary Bits
(Not Assigned)	00		..		..	..	..
Space	01	**	0010.000	FS	01.110	None	Yes
Tabulate	02	**	0000.100		..	..	..
Backspace	03	**	0010.101		..	..	..
Shift to outer set	04		..		..	..	..
Shift to inner set	05		..		..	..	..
Shift to LC/LS	06	**	0010.110	**	11.011	..	..
Shift to UC/FS	07	**	0000.111	**	00.000	..	..
( Open bracket	10	LC	0111.000	FS	10.100	0,8,4	Yes
) Close bracket	11	LC	0101.001	FS	01.100	10,8,4	Yes
, Comma	12	LC	0101.111	FS	11.110	0,8,3	Yes
π (£) Pi (Pounds)	13	LC	0111.011	LS	01.111	11,8,3	Yes
? Query	14	LC	0101.100	LS	10.111	11,8,5	Yes
& Ampersand	15	LC	0111.101		..	8,5	Yes
* Asterisk	16	LC	0111.110	FS	11.000	11,8,4	Yes
/ Oblique	17	UC	0011.111	FS	11.101	0,1	Yes
0	20	UC	0100.000	FS	00.001	0	Yes
1	21	UC	0110.001	FS	10.000	1	Yes
2	22	UC	0110.010	FS	01.000	2	Yes
3	23	UC	0100.011	FS	11.001	3	Yes
4	24	UC	0110.100	FS	00.100	4	Yes
5	25	UC	0100.101	FS	10.101	5	Yes
6	26	UC	0100.110	FS	01.101	6	Yes
7	27	UC	0110.111	FS	11.100	7	Yes
8	30	UC	0111.000	FS	00.010	8	Yes
9	31	UC	0101.001	FS	10.011	9	Yes
< Less than	32	LC	0100.011		..	0,8,5	Yes
> Greater than	33	LC	0110.100	FS	10.001	10,8,5	Yes
= Equals	34	LC	0100.101	FS	01.010	8,3	Yes
+ Plus	35	UC	0111.101	FS	01.011	10	Yes
- Minus	36	UC	0111.110	FS	11.010	11	Yes
. Point	37	UC	0101.111	**	00.111	10,8,3	Yes

Character	Internal Code (octal)	7-track code		5-track code		Atlas Fortran card code (holes punched)	Anelex Line-Printer (availability
		Case	Binary Bits	Shift	Binary Bits
' Prime (alternative n (letter n on 5-track tape only)	40	LC	0100.000	FS	10.111	8,4	Yes
A	41	UC	1010.001	LS	10.000	10,1	Yes
B	42	UC	1010.010	LS	01.000	10,2	Yes
C	43	UC	1000.011	LS	11.000	10,3	Yes
D	44	UC	1010.100	LS	00.100	10,4	Yes
E	45	UC	1000.101	LS	10.100	10,5	Yes
F	46	UC	1000.110	LS	01.100	10,6	Yes
G	47	UC	1010.111	LS	11.100	10,7	Yes
H	50	UC	1011.000	LS	00.010	10,8	Yes
I	51	UC	1001.001	LS	10.010	10,9	Yes
J	52	UC	1001.010	LS	01.010	11,1	Yes
K	53	UC	1011.011	LS	11.010	11,2	Yes
L	54	UC	1001.100	LS	00.110	11,3	Yes
M	55	UC	1011.101	LS	10.110	11,4	Yes
N	56	UC	1011.110	LS	01.110	11,5	Yes
O	57	UC	1001.111	LS	11.110	11,6	Yes
P	60	UC	1110.000	LS	00.001	11,7	Yes
Q	61	UC	1100.001	LS	10.001	11,8	Yes
R	62	UC	1100.010	LS	01.001	11,9	Yes
S	63	UC	1110.011	LS	11.001	0,2	Yes
T	64	UC	1100.100	LS	00.101	0,3	Yes
U	65	UC	1110.101	LS	10.101	0,4	Yes
V	66	UC	1110.110	LS	01.101	0,5	Yes
W	67	UC	1100.111	LS	11.101	0,6	Yes
X	70	UC	1101.000	LS	00.011	0,7	Yes
Y	71	UC	1111.001	LS	10.011	0,8	Yes
Z	72	UC	1111.010	LS	01.011	0,9	Yes
(Unassigned)	73	UC	1101.011		..	..	..
(Unassigned)	74	UC	1111.100		..	..	..
(Unassigned)	75	UC	1101.101		..	..	..
(Unassigned)	76	UC	1101.110		..	..	..
Fault	77		..		..	..	..

Internal Code - Outer Set

Character	Internal Code (octal)	7-track code		5-track code		Atlas Fortran card code (holes punched)	Anelex Line-Printer (availability
		Case	Binary Bits	Shift	Binary Bits
(Not Assigned)	00		..		..	..	..
Space	01	**	0010.000	FS	01.110	None	Yes
(Spare)	02		..		..	..	..
£ Pounds	03		..		..	..	Yes
Shift to outer set	04		..		..	..	..
Shift to inner set	05		..		..	..	..
Shift to LC/LS	06	**	0010.110	**	11.011	..	..
Shift to UC/FS	07	**	0000.111	**	00.000	..	..
(Unassigned)	10	**	0001.000		..	..	..
(Unassigned)	11	**	0011.001		..	..	..
(Unassigned)	12	**	0011.010		..	..	..
(Unassigned)	13	**	0001.011		..	..	..
Stop	14	**	0011.100		..	..	..
Punch On	15	**	0001.101		..	..	..
Punch Off	16	**	0001.110	F	..	..	..
: Colon	17	LC	0011.111		..	6,8	Yes
φ (x) Phi (letter x)	20		..	FS	00.011	..	..
[ Open square bracket	21	LC	0110.001		..	11,7,8	Yes
] Close square bracket	22	LC	0110.010		..	11,6,8	Yes
→ Arrow	23		..	FS	00.101	..	..
≥ Greater than or equals	24		..	FS	01.001	..	..
≠ Not equal	25		..	FS	10.010	..	..
_ Underline	26	LC	0100.110		..	10,6,8	Yes
| Vertical bar	27	LC	0110.111		..	10,7,8	Yes
2 (%) Superscript 2 (Percent)	30	LC	0101.010		..	..	Yes
Curly equals (v) (letter v)	31		..	FS	00.110	..	..
α (10) Alpha (Ten)	32	UC	0101.010		..	..	Yes
β (11) Beta (Eleven)	33	UC	0111.011		..	..	Yes
½ Half	34	UC	0101.100		..	..	Yes
10 Ten	35		..		..	..	Yes
11 Eleven	36		..		..	..	Yes
(Unassigned)	37	UC	1000.000		..	..	..

Character	Internal Code (octal)	7-track code		5-track code		Atlas Fortran card code (holes punched)	Anelex Line-Printer (availability
		Case	Binary Bits	Shift	Binary Bits
(Unassigned)	40	LC	1000.000	..	..	..	..
a	41	LC	1010.001		..	..	..
b	42	LC	1010.010		..	..	..
c	43	LC	1000.011		..	..	..
d	44	LC	1010.100		..	..	..
e	45	LC	1000.101		..	..	..
f	46	LC	1000.110		..	..	..
g	47	LC	1010.111		..	..	..
h	50	LC	1011.000		..	..	..
i	51	LC	1001.001		..	..	..
j	52	LC	1001.010		..	..	..
k	53	LC	1011.011		..	..	..
l	54	LC	1001.100		..	..	..
m	55	LC	1011.101		..	..	..
n	56	LC	1011.110		..	..	..
o	57	LC	1001.111		..	..	..
p	60	LC	1110.000		..	..	..
q	61	LC	1100.001		..	..	..
r	62	LC	1100.010		..	..	..
s	63	LC	1110.011		..	..	..
t	64	LC	1100.100		..	..	..
u	65	LC	1110.101		..	..	..
v	66	LC	1110.110		..	..	..
w	67	LC	1100.111		..	..	..
x	70	LC	1101.000		..	..	..
y	71	LC	1111.001		..	..	..
z	72	LC	1111.010		..	..	..
(Unassigned)	73	LC	1101.011		..	..	..
(Unassigned)	74	LC	1111.100		..	..	..
(Unassigned)	75	LC	1101.101		..	..	..
(Unassigned)	76	LC	1101.110		..	..	..
Erase	77	**	1111.111	**	11.111	..	..

Appedix E: Summary of Extracodes

Allocation of Function Numbers

There are 512 function numbers available for extracodes, 1000-1777. Of these, 1000-1477 are singly-modified instructions i.e. B-type, and 1500-1777 are doubly-modified i.e. A-type.

The extracodes are divided into sections as shown below:

Function	Subjects
1000-1077	Peripheral routines
1100-1177	Organisational routines.
1200-1277	Test instructions and character data processing
1300-1377	B-register operations.
1400-1477	Complex arithmetic, vector arithmetic and other B-type accumulator routines.
1500-1577	Double-length arithmetic and accumulator operations using the address as an operand.
1600-1677	Logical accumulator operations, trogonometric routines and half-word packing.
1700-1777	Logarithmic, exponential, square root etc., and miscellaneous arithmetic operations.

Were possible, the last two octal function digits correspond to those of similar basic operations.

The extracode function is listed at the left of the page and followed by a reference and a description. The number of basic instructions obeyed is given at the right of the page. This number includes the extracode instruction and its entry in the jump table; where necessary a range or formula is given.

The extracodes are listed in numerical order, and are also classified by type; some extracodes are therefore given twice.

The timings in this Appendix may differ from those in the main text. The figures here are updated in the light of Supervisor changes.

Extracode	Reference	Description	Instructions Obeyed
E.1 Magnetic tape
Block Transfers
1001	9.3.1	Search for section n on tape Ba
1002	9.3.1	Read next K + 1 sections from tape Ba to store blocks P, P + 1, ..., P + K.
1003	9.3.1	Read previous K + 1 sections from Ba to P + K, ..., P + 1, P.
1004	9.3.1	Write P, P + 1,..., P + K to next K +1 sections on Ba.
1005	9.3.1	Move tape Ba forwards K + 1 sections
1006	9.3.1	Move tape Ba backwards K + 1 sections
Organisational Instructions
1007	9.5.1	Move next reel of file Ba
1010	9.5.1	Mount
1011	9.5.1	Mount free
1012	9.5.1	Move on logical channel K
1013	9.5.1	Move free on logical channel K
1014	9.5.2	Write title
1015	9.5.2	Read title or number
1016	9.5.2	Unload
1017	9.5.2	Free tape
1020	9.5.2	Release tape (pass to another program)
1021	9.5.2	Release mechanisms
1022	9.5.2	Re-allocate
1023	9.5.2	How long?
1024	9.5.2	Where am I?
Variable Length Organisation
1030	9.4.2	Start reading forwards
1031	9.4.2	Start reading backwards
1032	9.4.2	Start writing forwards
1033	9.4.2	Start writing backwards
1034	9.4.2	Start reading forwards from fixed blocks
1035	9.4.2	Start reading backwards from fixed blocks
1036	9.4.3	ba' = selected magnetic tape
1037	9.4.3	s' = mode of magnetic tape Ba
Variable Length Transfers
1040	9.4.3	Transfer
1041	9.4.3	Skip
1042	9.4.3	Mark
1043	9.4.3	Stop
1044	9.4.3	Word search
1046	9.7	Read next block on Orion tape
1047	9.7	Read previous block on Orion tape
E.2 Input
1050	8.4	Select input n
1051	8.4	Find selected input
1052	8.14	Find peripheral equipment number
1053	8.14	Test whether binary or internal code
1054	8.14	Read next character to Ba. Jump to n at end of record
1055	8.14	ba' = number of blocks read
1056	8.14	Read ba half-words to S
1057	8.14	Read next record to S
E.3 Output
1060	8.4	Select output n
1061	8.4	Find selected output
1062	8.15	Find peripheral equipment type
1063	8.15	Delete output n
1064	8.15	Write character n
1065	8.15	End this record
1066	8.15	Write ba half-words from S
1067	8.15	Write a record from S
1070	8.15	Rename output n as input ba
1071	8.15	Break output n
1072	8.15	Define output n

Extracode	Reference	Description	Instructions Obeyed
E.4 Subroutine Entry
1100	7.7	Enter subroutine at s, ba' = c+1	6
1101	7.7	Enter subroutine at S, ba' = c+1	5
1102	7.7	Enter subroutine at bm, ba' = c+1	6
1362	7.7	Enter subroutine at n, b90' = c+1	3
E.5 Branching
1103	12.3.2	Establish Ba branches
1104	12.3.2	Start branch Ba at S
1105	12.3.2	Kill Ba. If Ba=64 kill current branch
1106	12.3.2	Halt current branch if Ba is active
1107	12.3.2	Jump to n if Ba is active
E.6 Monitor
1112	11.3	Set Monitor jump to n
1113	11.4.1	Do restart
1117	11.1.3	End program
E.7 Miscellaneous Transfers
1120	7.8	ba' = clock
1121	7.8	ba' = date
1122	12.4	ba' = local instruction counter
1123	12.4	set instruction counter = n	2048
1136	12.4	Read instruction counter
1124	7.8	v6' = n	3
1125	7.8	ba' = v6 & n	4
1126	12.9	v7' = n (hoot)	2
1127	12.9	ba' = v7' & n (read handswitches)	3
E.8 Traps
1131	7.5.2	See E.12 Character Data Processing
1132	11.2	Set trap/normal mode
1133	11.2	Ba' = trap address
1134	11.2	Trap
1135	12.1	See E.10 Store
1136	12.1	See E.7 Miscellaneous Transfers
E.9 Compiler and Supervisor
1140	12.9	Read parameter Ba to s
1141	12.9	Define Compiler
1142	12.9	End compiling
1143	12.9	Reserve Supervisor Tape
1147	12.9	Call complier n
1150	12.9	Assign ba blocks, labels P to (P + ba -1) to overflow K
1151	12.9	Set up blocks P onwards from overflow K
1155	12.1	See E.10 Store
1156	12.9	Enter extracode control at n if the "In Supervisor switch" is set
1157	12.9	Enter extracode control at n if the "Process switch" is set
E.10 Store
1135	12.1	Jump to n when block ≥ba defined
1155	12.1	JFind smallest block label defined
1160	12.1	Read block P
1161	12.1	Release block P
1162	12.1	Duplicate read
1163	12.1	Duplicate write
1164	12.1	Rename
1165	12.1	Write block P
1166	12.1	Read to absolute page
1167	12.1	Lose block P
1170	12.1	Clear blocks
1171	12.1	Store allocation = n blocks
1172	12.1	ba' = number of pages available
1173	12.1	ba' = number of blocks available
1174	12.1	Reserve band n
1175	12.1	Read K+1 blocks
1176	12.1	Write K+! blocks
1177	12.1	Lose band n

Arithmetic and Logical Extracodes

Accumulator operations are rounded floating-point unless marked X, when they are suitable for fixed-point working

Extracode	Reference	Description	Instructions Obeyed
E.11 Tests
1200	7.6.1	ba' = n if AO set; clear AO	9
1201	7.6.1	ba' = n if AO clear; clear AO	7
1204	7.5.3	See E.19 Logical Operations	7
1216	7.6.2	ba' = n if bm > 0	4-5
1217	7.6.2	ba' = n if bm ≤ 0	4-5
1223	7.6.2	ba' = n if Bc=1	4
1226	7.6.2	ba' = n if bt > 0	4-6
1227	7.6.2	ba' = n if bt ≤ 0	3-5
1234	7.6.1	c' = c' + 2 if am approximately = s	11-12
1235	7.6.1	c' = c' + 2 if am not approximately = s	11-12
Approximate equality is defined by |(am-s)/am| < C(ba), with am standardised If am = 0, am is not approximately = s
1236	7.6.1	ba'=n if ax > 0	4 or 6
1237	7.6.1	ba'=n if ax ≤ 0	3 or 5
1250-53	7.6.1	See E.12 Character Data Processing	3-5
1255	7.6.1	ba'=n if m is neither zero nor all ones	9
1265	7.5.3	See E.19 Logical Operations
1727	7.6.1	c' = c + 1, c +2 or c + 3 as am >, = or < s	7
1736	7.6.1	c' = c + 2, c +2 if |am| ≥ s	7 or 9
1737	7.6.1	c' = c + 2 if |am| < s	7
E.12 Character data processing
1131	7.5.2	Table search	8+6n
In 1250 and 1251, S is taken as a character address
1250	7.5.2	ba' (digits 18-23) = s, ba' (digits 0-17) = 0	7-10
1251	7.5.2	s' = ba (digits 18-23)	11-18
1252	7.5.2	Unpack n characters starting from character address C(ba), to half-words from C(ba*)	14 to 17 + int.pt. (6.5n)
1253	7.5.2	Pack n characters starting from character address C(ba*), to character address C(ba)	19 to 28 +3.75n

Extracode	Reference	Description	Instructions Obeyed
E.13 B-register operations
1300	7.5.1	ba' = integral part of s, am' = fractional part of s	10
1301	7.5.1	ba' = integral part of am, am' = fractional part of am	9
1302	7.5.1	ba' = ba.n, rounded away from zero	24-27
1303	7.5.1	ba' = -ba.n rounded away from zero	23-26
1304	7.5.1	ba' = integral part of (ba/n), b97'=remainder	26-31
In 1302-1304, ba and n are 21-bit integers in digits 0-20
1312	7.5.1	ba' = ba.n	23-26
1313	7.5.1	ba' = -ba.n	22-25
1314	7.5.1	ba' = integral part of (ba/n), b97'=remainder	27-32
In 1312-1314, ba and n are 24-bit integers
1340	7.5.1	ba' = ba.2-2n; unrounded arithmetic shift right	12-24
1341	7.5.1	ba' = ba.22n; unrounded arithmetic shift right	11-23
1342	7.5.1	ba' = ba; circularly shifted right n places	10-19
1343	7.5.1	ba' = ba circularly shifted left n places	9-18
1344	7.5.1	ba' = ba logically shifted right n places	12-23
1345	7.5.1	ba' = ba logically shifted left n places	11-22
1347	7.5.1	s' = s or ba	5
1353	7.5.1	ba' = position of most-significant 1 in bits 16-23 of n (as B123)	7
1356	7.5.1	bt' = ba neqv s	5
1357	7.5.1	bt' = ba neqv n	4
1362	7.5.1	See E.4 Subroutine Entry
1364	7.5.1	ba' = (ba & n) or (bm & n); b119' = (ba neqv & n	6
1371	7.5.1	b121' = ½Ba, b119' = N + bm	2
1376	7.5.1	bt' = ba & s	5
1377	7.5.1	bt' = ba & n	4

The complex accumulator, Ca, is a pair of consecutive registers, the first register having address ba. If Ba = 0, Ca is locations 0,1. s: is a number pair. Ca may coincide with S: but otherwise overlap with it. A is spoiled.

Extracode	Reference	Description	Instructions Obeyed
E.14 Complex arithmetic
1400	7.4.6	ca' = log s:	132-149
1402	7.4.6	ca' = exp s:	109-128
1403	7.4.6	ca' = conj s:	5
1407	7.4.2	See E.16 Miscellaneous B-type accumulator operations	15-30
1410	7.4.6	ca' = sqrt s:	90-93
1411	7.4.6	am' = arg s: radians	45-62
1412	7.4.6	ca' = mod s:	43
1413	7.4.6	ca' = s cos s*, s sin s*	70-80
1414	7.4.6	ca' = 1/s:	15
1415	7.4.2	See E.16 Miscellaneous B-type accumulator operations
1420	7.4.6	ca' = ca + s:	8
1421	7.4.6	ca' = ca - s:	8
1424	7.4.6	ca' = s:	6
1425	7.4.6	ca' = - s:	6
1456	7.4.6	s:' = ca	5
1462	7.4.6	ca' = ca.s:	18
E.15 Vector Operations
The vectors are of order n. s1 is stored in consecutive locations from ba, and s2 from ba*. A is spoiled.
1430	7.4.7	s1' = s1 + s2	9+4n
1431	7.4.7	s1' = s1 - s2	9+4n
1432	7.4.7	s1' = am. s2	9+4n
1433	7.4.7	s1' = s1 + am.s2	10+5n
1434	7.4.7	s1' = s2 (forwards or backwards)	11 or 13 +3n
1436	7.4.7	am' = Σ s1i' . s2i , i= 0 to n-1	10+5n
1437	7.4.7	a' = Σ s1i' . s2i , i= 0 to n-1	10+136n
E.16 Miscellaneous B-type accumulator operations
1407	7.4.2	Remainder and adjusted integral quotient	15-30
1415	7.4.2	Generate pseudo random number	4
1441	7.4.5	See E.18 Arithmetic using address as operand	6
1452	7.4.3	m'=m.sx.8ay+sy-bay , ay' =bay	21-28 (X)
1456-62	7.4.6	See E.14 Complex arithmetic
1466	7.4.2	a' = C(N+bm+ba) × C(N+bm) + a	21
1467	7.4.2	am' = Σ si xi from i=0 to r where x=am, Si = S+i, r=ba	6+3n
1473	7.4.3	m' = (ax/sx).8ay-sy-bay ,ay'=bay	26-33 (X)
1474	7.4.3	C(ba') = quotient (am/s), am' = remainder	22-33 (X)
1475	7.4.3	C(ba') = quotient (a/s), am' = remainder	21-32 (X)
1476	7.4.3	C(ba') = quotient (integral part of am/s), am' = remainder	30-41
E.17 Double-length Operations
The double-length number is s: = s + s* where sy-13 ≥ s*y. s* and al are asssumed to be positive numbers.
1500	7.4.4	a' = a + s:	13
1501	7.4.4	a' = a - s:	13
1502	7.4.4	a' = a + s:	17
1504	7.4.4	a' = s:	4
1505	7.4.4	a' = - s:	3
1520-35	7.4.5	See E.18. Arithmetic using address as operand	10
1542	7.4.4	a' = a.s:	18
1543	7.4.4	a' = -a.s:	22
1556	7.4.4	a:' = a	5
1562	7.4.5	See E.18. Arithmetic using address as operand
1565	7.4.4	a' = - a	6
1566	7.4.4	a' = |a|	4 or 7
1567	7.4.4	a' = |s:|	5
1574-75	7.4.5	See E.18 Arithmetic using address as Operand
1576	7.4.4	a' = a / s:	21
E.18 Arithmetic using address as Operand
The address is taken as a 21-bit integer with one octal fractional part. Fixed-point operations imply an exponent of 12
1441	7.4.5	sx' = ba, sy' = 12	6
1520	7.4.5	am' = am + n	10
1521	7.4.5	am' = am - n	11
1524	7.4.5	am' = n, l' = 0	8
1525	7.4.5	am' = - n, l' = 0	8
1534	7.4.5	am' = n, l' = 0	9 (X)
1535	7.4.5	am' = - n, l' = 0	10 (X)
1562	7.4.5	am' = am . n	8
1574	7.4.5	am' = am / n	17
1575	7.4.5	am' = aq / n	16
E.19 Logical accumulator operations
The logical accumulator G is B98 and B99
1204	7.5.3	ba' = number of 6-bit characters from most-significant end identical in g and s	9-25
1265	7.5.3	g' = 26g + n, ba' = 6-bits shifted out of g	11
1601	7.5.3	g' = s	3
1604	7.5.3	g' = g + s	7
1605	7.5.3	g' = g + s with end around carry	12
1606	7.5.3	g' = g neqv s	4
1607	7.5.3	g' = g & s	3
1611	7.5.3	g' = z where z is the logical complement of g	3
1613	7.5.3	s' = g	3
1615	7.5.3	am' = g	4
1624-26	7.4.8	See E.20 Half-word Packing	6
1630	7.5.3	g' = g & z where z is the logical complement of g	5
1635	7.5.3	g' = am	4
1646	7.5.3	g' = g or s	3
1652	7.5.3	bt' = g - s	6 or 8
E.20 Half-word Packing
s has an 8-bit exponent and a 16-bit mantissa
1624	7.4.8	am' = s	6
1626	7.4.8	s' = am, with s rounded	7

Extracode	Reference	Description	Instructions Obeyed
E.21 Functions and Miscellaneous routines
1700	7.4.1	am' = log s	47
1701	7.4.1	am' = log aq	46
1702	7.4.1	am' = exp s	44 or 53
1703	7.4.1	am' = exp aq	43 or 52
1704	7.4.2	am' = integral part s	5
1705	7.4.2	am' = integral part of am	4
1706	7.4.2	am' = sign s	5-6
1707	7.4.2	am' = sign am	4-5
1710	7.4.1	am' = sqrt s	36
1711	7.4.1	am' = sqrt aq	35
1712	7.4.1	am' = sqrt (aq2 + s2)	M41
1713	7.4.2	am' =aq s	87-96
1714	7.4.2	am' = 1 / s	5
1715	7.4.2	am' = 1 / am	4
1720	7.4.1	am' = arc sin s (-π/2 ≤ s ≤ π/2)	76-91
1721	7.4.1	am' = arc sin aq	75-90
1722	7.4.1	am' = arc cos s (0 ≤ s ≤ π)	76-91
1723	7.4.1	am' = arc cos aq	75-90
1724	7.4.1	am' = arc tan s (-π/2 < s < π/2)	38-53
1725	7.4.1	am' = arc tan aq	37-52
1726	7.4.1	am' = arc tan (aq/s) (-π ≤ s ≤ π)	44-61
1727	7.6.1	See E.11 Test Instructions
1730	7.4.1	am' = sin s	32-37
1731	7.4.1	am' = sin aq	31-36
1732	7.4.1	am' = cos s	32-37
1733	7.4.1	am' = cos aq	31-36
1734	7.4.1	am' = tan s	66-76
1735	7.4.1	am' = tan aq	65-75
1736-37	7.6.1	See E.11 Test Instructions
1752	7.4.3	m' = ax.812, ay' = ay - 12	10 or 13 (X)
1753	7.4.3	ax' = m.812, ay' = ay + 12	5 (X)
1754	7.4.3	Round am by adding; standardise	6
1755	7.4.3	ax' = ax.8ay - ny ; ay' = ny	18-26 (X)
1756	7.4.2	s' = am, am' = s	8
1757	7.4.2	am' = s / am	5
1760	7.4.2	am' = am 2	3
1762	7.4.3	m' = ax.812	9 or 12 (X)
1763	7.4.3	ax' = m.8-12	5 (X)
1764	7.4.3	ax' = ax.8n	15-23 (X)
1765	7.4.3	ax' = ax.8-n	14-22 (X)
1766	7.4.3	am' = |s|	4 (X)
1767	7.4.3	am' = |am|	3 (X)
1771	7.5.1	b121' = ½Ba, b119' = N + ba + bm	2
1772	7.4.3	m' = (m.sx) 8 12; ay' = ay + sy - 12	11 or 14 (X)
1773	7.4.3	m' = (ax/sx) 8 ay-sy -12; ay' = 12	28 to 36 (X)
1774	7.4.2	am' = am / s	11
1775	7.4.2	am' = aq / s	10
1776	7.4.2	Remainder	14

Appendix F: Summary of Basic Instructions by Function

See Atlas Instruction card