Revision b32c97e8bb3647c6e7cbcfcb6d0f48ba43f0d6b4 authored by Ron Burkey on 11 May 2020, 15:50:09 UTC, committed by Ron Burkey on 11 May 2020, 15:50:09 UTC
1 parent b494620
AGC_BLK2_INSTRUCTION_CHECK.agc
### FILE="Main.annotation"
## Copyright: Public domain.
## Filename: AGC_BLK2_INSTRUCTION_CHECK.agc
## Purpose: Part of the source code for Retread 44 (revision 0). It was
## the very first program for the Block II AGC, created as an
## extensive rewrite of the Block I program Sunrise.
## This file is intended to be a faithful transcription, except
## that the code format has been changed to conform to the
## requirements of the yaYUL assembler rather than the
## original YUL assembler.
## Reference: pp. 210-242
## Assembler: yaYUL
## Contact: Ron Burkey <info@sandroid.org>.
## Website: www.ibiblio.org/apollo/index.html
## Mod history: 2016-12-13 MAS Created from Aurora 12 version.
## 2016-12-16 MAS Began.
## 2016-12-17 MAS Completed transcription.
## 2016-12-18 MAS Typos.
## 2016-12-18 MAS Added notes about bugs discovered in the
## tests present here.
## 2016-12-20 MAS Fixed a typo in a program label.
## 2016-12-27 RSB Proofed comment text using octopus/ProoferComments,
## and fixed errors found.
## Page 210
SETLOC ABORT +1
# ADDRESSES OF ERASABLE REGISTERS
ADRS1 ADRES KEEP1
ADRS2 ADRES KEEP2
ADRS3 ADRES KEEP3
ADRS4 ADRES SELF1
ADRS5 ADRES SELF2
ADRS6 ADRES S+MAX
ADRS7 ADRES SELF3
# CONSTANTS USED THROUGHOUT THE INSTRUCTIONS CHECK
SBIT1 OCTAL 00001
SBIT2 OCTAL 00002
SBIT3 OCTAL 00004
SBIT4 OCTAL 00010
SBIT5 OCTAL 00020
SBIT6 OCTAL 00040
SBIT7 OCTAL 00100
SBIT8 OCTAL 00200
SBIT9 OCTAL 00400
SBIT10 OCTAL 01000
SBIT11 OCTAL 02000
SBIT12 OCTAL 04000
SBIT13 OCTAL 10000
SBIT14 OCTAL 20000
SBIT15 OCTAL 40000
## The 00000 below is lightly circled.
S+ZERO OCTAL 00000
S+1 OCTAL 00001
S+2 OCTAL 00002
S+3 OCTAL 00003
S+4 OCTAL 00004
S+5 OCTAL 00005
S+6 OCTAL 00006
S+7 OCTAL 00007
S6BITS OCTAL 00077
S7BITS OCTAL 00177
S13BITS OCTAL 17777
SODD OCTAL 25252 # SEVEN ONE BITS
S+MAX OCTAL 37777
S-MAX OCTAL 40000
ALARMCON OCTAL 40400
SINOUT1 OCTAL 52500
SEVENS OCTAL 52525 # EIGHT ONE BITS
SINOUT2 OCTAL 52552
CYRCON OCTAL 57761
SINOUT3 OCTAL 77725
S-15 OCTAL 77760
S-14 OCTAL 77761
S-7 OCTAL 77770
## Page 211
S-6 OCTAL 77771
S-5 OCTAL 77772
S-4 OCTAL 77773
S-3 OCTAL 77774
S-2 OCTAL 77775
S-1 OCTAL 77776
S-ZERO OCTAL 77777
# NEXT TWO CONSTANTS ARE USED IN THE DEVIDE SUBROUTINE
DV1CON OCTAL 14000
DV2CON OCTAL 37776
# NEXT TWO CONSTANTS ARE ADDRESSESS USED BY EXTRACODE INDEX INSTRUCTIONS
ADRS+1 ADRES S+1
ADRSDV1 ADRES DV1CON
CS A
-0CHK CCS A
TCF ERRORS
TCF ERRORS
TCF ERRORS
TC Q
CS A
-1CHK CCS A
TCF ERRORS
TCF ERRORS
CCS A
TCF ERRORS
TC Q
ERRORS XCH Q
TS SFAIL # SAVE Q FOR FAILURE LOCATION
INCR ERCOUNT # KEEP TRACK OF NUMBER OF MALFUNCTIONS
INHINT # TURN ON PROGRAM ALARM LIGHT
CS ALARMCON
MASK DSPTAB +11D
AD ALARMCON
TS DSPTAB +11D
RELINT
# IF C(SMODE) IS +NON-ZERO START CHECKING AGAIN AT TCCHK
# IF C(SMODE) IS + PUT +0 IN SMODE AND IDLE
## "CA SMODE" below is marked with a line drawn to a +, and "BZMF STOPCHK" is underlined and marked with a -
CA SMODE
EXTEND
BZMF STOPCHK
TC SMODECHK
STOPCHK CA S+ZERO
TS SMODE
## Page 212
TC CHECKNJ
## The following CCS has its branch conditions written in: +>0, +0, -<0, and -0.
SMODECHK CCS SMODE
TC +3
TC SMODECHK -1
TC +1
TC CHECKNJ
CAF STRTCHK
TC BANKJUMP # TO START OF CHECKING ROUTINES
STRTCHK CADR TCCHK
SETLOC 26000
CA S+ZERO # INITIALIZE COUNT REGISTER
TS ERCOUNT
TS SCOUNT
# NORMAL USE OF TC AND TCF
TCCHK TC +2
TC CCSCHK
TCF +2
TC ERRORS
TC Q
TC ERRORS
## Written to the right of the next few lines is "Written for comp self test and in Gyro S.F. test", signed G&S.
# NORMAL USE OF CA, CS, AND CCS
CCSCHK CA S-3
TS KEEP1
CCS KEEP1
TC ERRORS
TC ERRORS
TC +2
TC ERRORS
CCS A # C(A) =+2, RESULT OF CCS -NUMBER
TC +4
TC ERRORS
TC ERRORS
TC ERRORS
CCS A # C(A) = +1, RESULT OF CCS + NUMBER
TC +4
TC ERRORS
TC ERRORS
TC ERRORS
CCS A # C(A) = +0, RESULT OF CCS + NUMBER
TC ERRORS
TC +3
TC ERRORS
TC ERRORS
CS A
CCS A # C(A) = -0, RESULT OF CCS +0
TC ERRORS
## Page 213
TC ERRORS
TC ERRORS
CCS A # RESULT OF CCS -0
TC ERRORS
TC +3
TC ERRORS
TC ERRORS
# NORMAL USE OF MASK
MSKCHK CS S-ZERO
MASK S-ZERO # 00000, 77777
TC -0CHK -1
CS S+ZERO
MASK S+ZERO # 77777, 00000
TC -0CHK -1
CA S+ZERO
MASK S+ZERO # 00000, 00000
TC -0CHK -1
CA S-ZERO
MASK S-ZERO # 77777, 77777
TC -0CHK
# NO EDIT FEATURE OF MASK IS CHECKED
# BITS 9-14 OF WRITE LINES GO TO BITS 1-7 OF EDOP
CA S-ZERO # 77777
TS EDOP # 00177
MASK EDOP # 00177
TS KEEP1 # 00177
MASK EDOP # 00177
CS A # 77600
AD KEEP1 # 77777
TC -0CHK
# CHECK MASK OF AN SC REGISTER
CA S+1
TS L
CA S-ZERO
MASK L
TC -1CHK -1
# NORMAL USE OF XCH, AD, AND TS
CA S+MAX # 37777
TS KEEP1
AD KEEP1 # 01 - 37776
TS KEEP2 # 37776
TC ERRORS
TC -1CHK -1
XCH KEEP1 # SKEEP1 NOW +0
CS A # 40000
AD A # 10 - 00001
TS KEEP3 # 40001, C(A) = -1
TC ERRORS
## Page 214
AD KEEP3 # C(A) = 40000
AD KEEP2 # C(A) = -1
AD KEEP1 # C(A) = -1
TS KEEP4 # -1
CS KEEP4 # +1
TC -1CHK -1
# NORMAL USE OF INCR
# NOT CHECKING COUNTER INTERRUPT
CA S+MAX # 37777
TS KEEP1
INCR KEEP1 # +0
INCR KEEP1 # +1
INCR KEEP1 # +2
AD S-MAX
TC -0CHK # CHECK C(A) HAS NOT CHANGED
CS KEEP1
TS KEEP1 # -2
INCR KEEP1 # -1
CA KEEP1
TC -1CHK
# CHECK INCREMENT OF AN SC REGISTER
CA S-2
TS L
INCR L
CA L
TC -1CHK
# NORMAL USE OF ADS
CA SBIT13
TS KEEP1 # 10000
ADS KEEP1 # 20000
ADS KEEP1 # OV WITH +0
TS KEEP2
TC ERRORS
CS KEEP1
TC -0CHK
# CHECKS ADS OF AN SC REGISTER
CA S13BITS # 17777
TS L
AD S+1 # C(A) = 20000
ADS L
AD S-MAX
TC -0CHK
CS L
AD S+MAX
TC -0CHK
# NORMAL USE OF LXCH
CA S+1
## Page 215
TS KEEP1 # +1
LXCH KEEP1 # +1 IN L
CS A
TS KEEP2 # -1 IN KEEP2
LXCH KEEP2 # L = -1, KEEP2 = +1
CS KEEP2
TC -1CHK
CA L
TC -1CHK
# UNDERFLOW AND OVERFLOW IS LOST IN L REGISTER
CA S+MAX
AD A
TS L # OV WITH 37776
TC ERRORS
CS S+MAX
AD A # UV WITH 40001
LXCH A # C(A) = 37776, C(L) = 40001
TS KEEP1 # 37776
TC +2
TC ERRORS
CA L
TS KEEP2 # 40001
TC +2
TC ERRORS
AD KEEP1 # -0
TC -0CHK
# NORMAL USE OF DXCH
CA S+MAX
TS KEEP2 # 37777, K+1
CS A
TS L # 40000
AD S+1
TS KEEP1 # 40001, K
CS A # 37776
DXCH KEEP1
# A = 40001, L = 37777 ....... KEEP1 = 37776, KEEP2 = 40000
AD L
TC -1CHK -1
CA KEEP1
AD KEEP2
TC -1CHK
# NORMAL USE OF DAS (6 CHECKS)
# IF ADDRESS OF K DOES NOT = ZERO, C(L) = +0 AND C(A) = NET OVERFLOW
# C(A) = +0 IF NO OVERFLOW OR UNDERFLOW
# DAD++ WITH NO OVERFLOW
DAS++ CAF S13BITS
TS KEEP1 # 17777
## Page 216
TS KEEP2 # 17777
TS L # 17777
AD S+1 # 20000
DAS KEEP1
# C(KEEP1) = 37777, C(KEEP2) = 377776
TC -0CHK -1
XCH L
TC -0CHK -1
CS KEEP1
AD KEEP2
TC -1CHK
# DAS++ WITH OVERFLOW
DAS++OV CA S+MAX
TS KEEP1 # 37777
TS KEEP2 # 37777
TS L # 37777
CA S+1 # +1
DAS KEEP1
# C(KEEP1) = +1, C(KEEP2) = 37776, C(A) = +1,
TC -1CHK -1
XCH L
TC -0CHK -1
CS KEEP1
TC -1CHK
CA S-MAX
AD KEEP2
TC -1CHK
# DAS MIXED SIGNS
DAS+--+ CA S+MAX
TS KEEP1 # 37777
CS A
TS KEEP2 # 40000
CS A
AD S-1
TS L # 37776
CS A # 40001
DAS KEEP1
# C(KEEP1) = +1, C(KEEP2) = -1
TC -0CHK -1
XCH L
TC -0CHK -1
CA KEEP1
TC -1CHK -1
CA KEEP2
TC -1CHK
# DAS-- WITH NO UNDERFLOW
DAS-- CS S13BITS
TS KEEP1 # 60000
TS KEEP2 # 60000
TS L # 60000
## Page 217
AD S-1 # 57777
DAS KEEP1
# C(KEEP1) = 40000, C(KEEP2) = 40001
TC -0CHK -1
XCH L
TC -0CHK -1
CS KEEP2
AD KEEP1
TC -1CHK
# DAS-- WITH UNDERFLOW
DAS--UV CA S-MAX
TS KEEP1 # 40000
TS KEEP2 # 40000
TS L # 40000
CA S-1 # -1
DAS KEEP1
# C:KEEP1) = -1, C(KEEP2) = 40001, C(A) = -1
TC -1CHK
XCH L
TC -0CHK -1
CA KEEP1
TC -1CHK
CA S+MAX
AD KEEP2
TC -1CHK -1
# DAS A. DOUBLES THE CONTENTS OF THE A REGISTER AND THE L REGISTER.
CA S-MAX
TS KEEP2 # 40000
TS L # 40000
CS A
TS KEEP1 # 37777
DAS A
# C(A) = OV 37775, C(L) = 40001
TS KEEP3
TC ERRORS
CA L
AD KEEP3
TC -1CHK
# NORMAL USE OF INDEX WITHOUT EXTRACODE.
# INSTRUCTIONS CHECKED WITH INDEX UP TO FIRST SPACE SKIPPED
# ARE CA, TS, XCH, CCS, AD, TC, TCF, TS WITH OVERFLOW, AND MASK
# FIRST INITIALIZE ERASABLE REGISTERS USED FOR INDEX INSTRUCTION
CA S+ZERO
TS NDX+0
CA ADRS6 # ADDRESS OF S+MAX
TS NDX+MAX
CA ADRS1
TS NDXKEEP1
CA ADRS2
## Page 218
TS NDXKEEP2
CA ADRS3
TS NDXKEEP3
CA ADRS4
TS NDXSELF1
CA ADRS5
TS NDXSELF2
NDXCHK NDX NDX+MAX # CA S+MAX
CA 0000 # A = 37777
NDX NDXKEEP1 # TS KEEP1
TS 0000 # TS WITH NO OV, UV
NDX NDX+0 # CS A
CS 0000 # A = 40000
NDX NDXKEEP1 # XCH KEEP1
XCH 0000 # A = +MAX, KEEP1 = - MAX
NDX NDX+0 # CCS A
CCS 0000 # A = 37776
TC +4
TC ERRORS
TC ERRORS
TC ERRORS
NDX NDXKEEP1 # AD KEEP1
AD 0000 # A = -1
SELF1 NDX NDXSELF1 # TC +2
TC 0003
TC ERRORS # DID NOT PERFORM A TC
NDX NDXSELF1 # TC +2
TC 0006
TC CNTINU # CHECK C(Q)
SELF2 NDX NDXSELF2 # TCF +2
TCF 003
TC ERRORS
TC Q # SHOULD GO TO SELF1 +2
CNTINU TC -1CHK
CA S+MAX
AD S+1
NDX NDXKEEP1 # TS KEEP1 WITH OV
TS 0000 # A = +1, KEEP1 = +0
TC ERRORS
AD KEEP1 # A = +1
NDX NDX+MAX # MASK S+MAX
MASK 0000 # A = +1
TC -1CHK -1
# INDEX USED WITH ADS, NDX, AND INCR
CA S13BITS
TS KEEP1
NDX NDX+0 # INDEX +0
NDX NDXKEEP1 # ADS KEEP1
ADS 0000 # C(A) AND C(KEEP1) = 377776
## Page 219
NDX NDXKEEP1 # INCR KEEP1
INCR 0000 # C(KEEP1) = 37777
CS A # 40001
AD KEEP1 # A = +1
TC -1CHK -1
# INDEX USED WITH LXCH, DAS, AND DXCH
CA S-MAX
TS KEEP2 # KEEP2 HOLDS 40000, KEEP1 HOLDS 37777
CA S+1
TS KEEP3 # +1
NDX NDXKEEP3 # LXCH KEEP3
LXCH 0000 # C(L) = +1
CA S-2
NDX NDXKEEP1 # DAS KEEP1
DAS 0000
# BEFORE DAS, K = 37777 K+1 = 40000
# A = -2 L = +1
# AFTER DAS, K = 37775 K+1 = 40001
# A = +0 L = +0
NDX NDXKEEP1 # DXCH KEEP1
DXCH 0000
AD L
TC -1CHK
CS KEEP1
TC -0CHK
CA KEEP2
TC -0CHK -1
# INDEX INSTRUCTION USED WITH OVERFLOW
CA ADRS7 # ADDRESS OF SELF3
AD SBIT14
TS KEEP7
SELF3 NDX KEEP7
2 0002
TC ERRORS
# CHECK INDEX OF AN SC REGISTER
CA ADRS1 # 01371, ADDRESS OF KEEP1
NDX A
TS 0000 # PUT 01371 IN KEEP1
CS A
AD KEEP1
TC -0CHK
## Page 220
# START CHECKING EXTRACODE INSTRUCTIONS
# NORMAL USE OF DCA, DCS, AND SU
STRTXTRA EXTEND
DCA SBIT1
# C(A) = +1, C(L) = +2
TS KEEP2 # +1
XCH L
TS KEEP1 # +2
EXTEND
SU KEEP2 # C(KEEP2) = +1
TC -1CHK -1
EXTEND
DCS KEEP1
# C(A) = -2, C(L) = -1
EXTEND
SU L
TC -1CHK
# COMPLEMENTING OF THE DOUBLE PRECISION ACCUMULATOR
CA S+MAX
TS L # 37777
AD A # OV37776
EXTEND
DCS A
TS KEEP1
TC ERRORS
CA KEEP1
EXTEND
SU L
TC -1CHK -1
## The following test is, in the words of AGC developer Hugh Blair-Smith, "wrong wrong wrong".
## The code appears to be assuming that the result of DCA L is A <-- L and L <-- Q. However,
## in reality, the L <-- Q happens bfore L is copied into A, so the net result is A <-- Q
## and L <-- Q.
## Hugh's theory is that Retread made it up to revision 44 mostly before hardware was
## available to run it on. And the self-check code, which was tied to the hardware, was likely
## not exercised as part of its development. Moreover, it's possible that the digital simulator
## itself had an incorrect implementation of DCA based on a misunderstanding of the hareware
## design.
## There's a couple of ways to fix this test, but the most "correct" would probably be:
## DCA L # A = L = -2
## CS A # A = 2
## AD L # A = -2 + 2 = -0
## TC -0CHK
# CHECKS DCA OF AN SC REGISTER
CA S-2 # -2
TS Q
CA S+1
TS L # +1
CA SBIT8
EXTEND
DCA L
AD L # -1
TC -1CHK
# NORMAL USE OF QXCH
CA QXCHCON1
TS KEEP1 # STORE ADDRESS OF AUGCHK IN KEEP1
TC +2 # Q NOW HOLDS ADDRESS OF QNMBR
QNMBR TC ERRORS
EXTEND
QXCH KEEP1 # Q NOW HOLDS ADDRESS OF AUGCHK
TC Q # SHOULD GO TO QXCHCON2 +1, NOT QNMBR
TC ERRORS
QXCHCON1 ADRES QXCHCON2 +1
## Page 221
QXCHCON2 ADRES QNMBR
CS KEEP1 # CHECK THAT KEEP1 HOLDS B(Q)
AD QXCHCON2
TC -0CHK
# CHECKS QXCH OF AN SC REGISTER
CA S+ZERO
TC +2
TCF +5
EXTEND
QXCH A
LXCH Q
TC A
CA L
TC -0CHK -1
# NORMAL USE OF AUG
AUGCHK CA S+ZERO
TS KEEP1 # +0
CS A
TS KEEP2 # -0
EXTEND
AUG KEEP1 # +1
EXTEND
AUG KEEP1 # +2
TC -0CHK # CHECK C(A) HAS NOT CHANGED
EXTEND
AUG KEEP2 # -1
EXTEND
AUG KEEP2 # -2
EXTEND
AUG KEEP2 # -3
TC -0CHK -1
CA KEEP2
AD KEEP1
TC -1CHK
# CHECKS AUG OF AN SC REGISTER
CA S-ZERO
EXTEND
AUG A
TC -1CHK
# NORMAL USE OF DIM
DIMCHK CA S+ZERO
TS KEEP1 # +0
EXTEND
DIM KEEP1
CA KEEP1
TC -0CHK -1
CS A
TS KEEP1 # -0
## Page 222
## There is a physical page break here, but the next page does not have a header or its own number.
EXTEND
DIM KEEP1
CA KEEP1
TC -0CHK
CA S+2
TS KEEP1 # +2
EXTEND
DIM KEEP1 # +1
AD S-1
TC -1CHK -1 # CHECK C(A) HAS NOT CHANGED
CA KEEP1
TC -1CHK -1
EXTEND
DIM KEEP1
TC -0CHK -1
CA KEEP1
TC -0CHK
CS S+2
TS KEEP2 # -2
EXTEND
DIM KEEP2 # -1
CA KEEP2
TC -1CHK
EXTEND
DIM KEEP2 # -0
TC -0CHK -1
CA KEEP2
TC -0CHK
# CHECKS DIM OF AN SC REGISTER
CA S-2
EXTEND
DIM A
TC -1CHK
# NORMAL USE OF MSU
# MSU SAME (S+MAX AND S+MAX), RESULT +0
CA S+MAX
TS KEEP1
EXTEND
MSU KEEP1
TC -0CHK -1
# MSU SAME (+0 AND +0), RESULT +0
TS KEEP2
EXTEND
MSU KEEP2
TC -0CHK -1
# MSU SAME (-0 AND -0), RESULT +0
CA S-ZERO
TS KEEP3
EXTEND
## Page 223
MSU KEEP3
TC -0CHK -1
# MSU +0 AND 77777, RESULT = +1
EXTEND
MSU KEEP3
TC -1CHK -1
# MSU 77777 AND +0, RESULT = -1
CS A
EXTEND
MSU KEEP2
TC -1CHK
# MSU +6 AND +7, RESULT = -1
CA S+7
TS KEEP4
CA S+6
TS KEEP5
EXTEND
MSU KEEP4
TC -1CHK
# MSU +7 AND +6, RESULT = +1
CA S+7
EXTEND
MSU KEEP5
TC -1CHK -1
# MSU 77770 AND 77771, RESULT = -1
CA S-6
TS KEEP6
CA S-7
TS KEEP7
EXTEND
MSU KEEP6
TC -1CHK
# MSU 77771 AND 77770, RESULT = +1
CA S-6
EXTEND
MSU KEEP7
TC -1CHK -1
# CHECKS MSU OF AN SC REGISTER ( -0 AND -0 = +0)
CA S-ZERO
TS L
EXTEND
MSU L
TC -0CHK -1
# NORMAL USE OF BZF
BZFCHK TC +2
TC BZMFCHK # CORRECT ADDRESS IN Q
CAF S+5
EXTEND
BZF ERRORS
## Page 224
CS A
EXTEND
BZF ERRORS
CAF S+ZERO
EXTEND
BZF +2
TC ERRORS
CS A
EXTEND
BZF +2
TC ERRORS
TC Q # SHOULD GO TO BZFCHK +1
TC ERRORS
BZMFCHK TC +2
TC MP1++ # CORRECT ADDRESS IN Q
CAF SBIT9
EXTEND
BZMF ERRORS
CS A
EXTEND
BZMF +2
TC ERRORS
CA S+ZERO
EXTEND
BZMF +2
TC ERRORS
CS A
EXTEND
BZMF +2
TC ERRORS
TC Q # SHOULD GO TO BZMFCHK +1
TC ERRORS
# NORMAL USE OF MP
# 37777 X 2
MP1++ CA S+MAX # 37777
EXTEND
MP S+2 # C(A) = +1, C(L. = 37776
AD L
TS KEEP1 # 37777
MP1+- EXTEND
MP S-2 # C(A) = -1, C(L) = 40001
AD L
TS KEEP2 # 40000
MP1-+ EXTEND
MP S+2 # C(A) = -1, C(L) = 40001
AD L
TS KEEP3 # 40000
MP1-- EXTEND
## Page 225
MP S-2 # C(A) = +1, C(L) = 37776
AD L # 37777
AD KEEP3 # 77777
AD KEEP2 # 40000
AD KEEP1 # 77777
TC -0CHK
# 37777 X 37777
MP2++ CA S+MAX # 37777
EXTEND # CHECKS RSC PULSE
MP A # C(A) = 37776, C(L) = +1
AD L
TS KEEP1 # 37777
MP2+- EXTEND
MP S-MAX # C(A) = 40001, C(L) = -1
AD L
TS KEEP2 # 40000
MP2-+ EXTEND
MP S+MAX # C(A) = 40001, C(L) = -1
AD L
TS KEEP3 # 40000
MP2-- EXTEND
MP S-MAX # C(A) = 37776, C(L) = +1
AD L # 37777
AD KEEP3 # 77777
AD KEEP2 # 40000
AD KEEP1 # 77777
TC -0CHK
# C(A) = NON-ZERO, C(K) = ZERO
# RESULT IS ALWAYS POSITIVE ZERO
MP3++ CA S+MAX # 37777
EXTEND
MP S+ZERO
AD L
TC -0CHK -1
MP3+- CA S+1
EXTEND
MP S-ZERO
AD L
TC -0CHK -1
MP3-+ CA S-1
EXTEND
MP S+ZERO
AD L
TC -0CHK -1
MP3-- CA S-ZERO
EXTEND
MP S-ZERO
AD L
TC -0CHK -1
# C(A) = ZERO, C(K) = NON-ZERO,
## Page 226
# RESULT IS + ZERO FOR A POSITIVE SIGN AND NEGATIVE
# ZERO FOR Z NEGATIVE SIGN
MP4++ CA S+ZERO
EXTEND
MP S+MAX
AD L
TC -0CHK -1
MP4+- EXTEND
MP S-1
AD L
TC -0CHK
MP4-+ CS A
EXTEND
MP S+5
AD L
TC -0CHK
MP4-- CS A
EXTEND
MP S-ZERO
AD L
TC -0CHK -1
# MULTIPLY ZERO X ZERO
# RESULT IS ALWAYS PLUS ZERO
MP5++ CA S+ZERO
TS KEEP1
EXTEND
MP KEEP1
AD L
TC -0CHK -1
MP5+- EXTEND
MP S-ZERO
AD L
TC -0CHK -1
MP5-+ CA S-ZERO
EXTEND
MP S+ZERO
AD L
TC -0CHK -1
MP5-- CA S-ZERO
EXTEND
MP S-ZERO
TC -0CHK -1
CA CONTINU
TC BANKJUMP
CONTINU CADR DV1++ -1 # CONTINUE WITH INSTRUCTION CHECK
SETLOC 30000
TC CHECKNJ # CHECK FOR NEW JOB
## Page 227
# NORMAL USE OF DV ... REMAINDER HAS SIGN OF DIVIDEND
# 1/4 DIVIDED BY 3/8
# C(A) = 25252 WITH A + QUOTIENT AND 52525 WITH A - QUOTIENT.
# C(L) = REMAINDER = /100000/ WITH SIGN OF DIVIDEND.
DV1++ CA DV1CON # 14000
TS KEEP7 # 14000, +3/8
TS Q
CS A
TS KEEP6 # 63000, -3/8
CA S+ZERO
TS L
CA SBIT13 # 10000
EXTEND
DV Q # CHECKS RSC PULSE
TS KEEP1 # 25252
CA S+ZERO
DV1+- LXCH A
EXTEND
DV KEEP6
AD KEEP1
TC -0CHK
CA S-ZERO
LXCH A
DV1-+ CS A
EXTEND
DV KEEP7
TS KEEP1 # 52525
CA S-ZERO
DV1-- LXCH A
EXTEND
DV KEEP6
AD KEEP1
TC -0CHK
CA L
AD SBIT13
TC -0CHK
# 1/2 TO 15TH DIVIDED BY 1/2 TO 14TH
# C(A) SHOULD BE 1/2 AND CONTENTS OF L SHOULD BE ZERO
DV2++ CA S+1
TS KEEP7 # 00001, DIVISOR
CS A
TS KEEP6 # 77776, DIVISOR
CA S+ZERO
CA SBIT14 # 20000
TS L
CA S+ZERO
EXTEND
DV KEEP7 # C(A) = 1/2, C(L) = +0
TS KEEP1
DV2+- LXCH A
## Page 228
EXTEND
DV KEEP6
TS KEEP2 # -1/2
AD KEEP1
TC -0CHK
CA KEEP2
LXCH A
DV2-+ CS A
EXTEND
DV KEEP7
TS KEEP2
DV2-- LXCH A
EXTEND
DV KEEP6
TS KEEP1
AD KEEP2
TC -0CHK
CS KEEP1 # MAKE SURE QUOTIENT IS 1/2
AD SBIT14
TC -0CHK
CA L
TC -0CHK
# SAME AS PREVIOUS DIVISION EXCEPT A AND L WILL HAVE OPPOSITE SIGNS
# BEFORE DIVISION. SINCE A WILL ALWAYS BE ZERO, THE SIGN OF THE QUOTIENT
# WILL DEPEND ON THE SIGN OF L AND THE SIGN OF THE DIVISOR.
DV3++ CA SBIT14 # 20000
TS L
CA S-ZERO
EXTEND
DV KEEP7
TS KEEP1 # 20000
DV3+- LXCH A
CS A # A = -0
EXTEND
DV KEEP6
AD KEEP1
TC -0CHK
CS SBIT14 # -1/2
DV3-+ LXCH A
EXTEND
DV KEEP7
TS KEEP1
DV3-- LXCH A
CS A # A = +0
EXTEND
DV KEEP6
AD KEEP1
TC -0CHK
CS L
TC -0CHK -1
## Page 229
# C(A) = 17777 AND C(L) = 37777. THIS IS DIVIDED BY 20000. THE RESULT
# SHOULD BE +-/37777/ AND THE REMAINDER +-/17777/
DV4++ CA S+MAX
TS L
CA SBIT14
TS KEEP7 # 20000
CS A
TS KEEP6 # 57777
CA S13BITS # 17777
EXTEND
DV KEEP7
TS KEEP1
DV4+- LXCH A
EXTEND
DV KEEP6
TS KEEP2
AD KEEP1
TC -0CHK
CA KEEP2
LXCH A
DV4-+ CS A
EXTEND
DV KEEP7
AD KEEP1
TC -0CHK
CA KEEP2
DV4-- LXCH A
EXTEND
DV KEEP6
TS KEEP3
AD KEEP2
TC -0CHK
CS KEEP3
AD S+MAX
TC -0CHK
CA L
AD S13BITS
TC -0CHK
# C(A) = +-/17777/ AND C(L) = +-/37777/ WITH OPPOSITE SIGN BEFORE DEVISION
# THE QUOTIENT SHOULD BE +-/37774/ WITH THE SIGN DEPENDING ON THE SIGN OF
# A AND THE SIGN OF THE DEVISOR. THE C(L) = +-/1/ DEPENDING ON THE SIGN
# OF A.
DV5++ CS S+MAX
TS L # 40000
CA S13BITS # 17777
EXTEND
DV KEEP7
TS KEEP1 # 37774
XCH L
TC -1CHK -1
## Page 230
DV5+- CA S-MAX
TS L # 40000
CA S13BITS # 17777
EXTEND
DV KEEP6 # C(A) = -37774, C(L) = +1
AD KEEP1
TC -0CHK
XCH L
TC -1CHK -1
DV5-+ CA S+MAX
TS L # 37777
CS S13BITS # 60000
EXTEND
DV KEEP7 # C(A) = -37774, C(L) = -1
TS KEEP2
AD KEEP1
TC -0CHK
XCH L
TC -1CHK
DV5-- CA S+MAX
TS L # 37777
CS S13BITS # 60000
EXTEND
DV KEEP6 # C(A) = 37774, C(L) = -1
AD KEEP2
TC -0CHK
XCH L
TC -1CHK
CA KEEP2 # -37774
AD S-2 # -37776
AD S+MAX # +1
TC -1CHK -1 # CHECK THAT QUOTIENT IS +-/37774/
# DIVIDE SAME (37776). THE RESULT SHOULD BE MAXIMUM AND THE REMAINDER
# SHOULD BE THE SAME VALUE AS THE DIVISOR WITH THE SAME SIGN AS THE
# DIVIDEND
DV6++ CA S+ZERO
TS L
CS DV2CON # 37776
TS KEEP6 # 40001
CS A
TS KEEP7 # 37776
EXTEND
DV KEEP7
CS A
AD L
TC -1CHK
CA S+ZERO
DV6+- LXCH A
EXTEND
## Page 231
DV KEEP6
AD L
TC -1CHK
CA S-ZERO
DV6-+ LXCH A
CS A
EXTEND
DV KEEP7
CS A
AD L
TC -1CHK -1
CA S-ZERO
DV6-- LXCH A
EXTEND
DV KEEP6
AD L
TC -1CHK -1
CS L
AD KEEP6
TC -0CHK
# DIVIDE SAME (ZERO). THE RESULT SHOULD BE MAXIMUM AND THE REMAINDER
# SHOULD BE THE SAME VALUE AS THE DIVISOR WITH THE SAME SIGN AS THE
# DIVIDEND.
DV7++ CS S+ZERO
TS KEEP6 # -0
CS A
TS KEEP7 # +0
TS L
EXTEND
DV KEEP7
AD S-MAX
TC -0CHK
DV7+- LXCH A # C(A) = C(L) = +0
EXTEND
DV KEEP6
AD S+MAX
TC -0CHK
CS A
DV7-+ LXCH A
CS A # C(A) = C(L) = -0
EXTEND
DV KEEP7
AD S+MAX
TC -0CHK
CS A
DV7-- LXCH A # C(A) = C(L) = -0
EXTEND
DV KEEP6
AD S-MAX
TC -0CHK
## Page 232
CS L
TC -0CHK -1
# DEVIDE SAME (ZERO). THE CONTENTS OF THE A REGISTER AND L REGISTER WILL
# HAVE OPPOSITE SIGNS BEFORE DEVISION. THE SIGN OF THE QUOTIENT WILL
# DEPEND ON THE SIGN OF THE L REGISTER BEFORE DEVISION AND THE SIGN OF
# THE DEVISOR. THE SIGN OF THE REMAINDER IS THE SAME SIGN AS THE SIGN OF
# THE L REGISTER BEFORE DEVISION. C(L) REMAINS SAME
DV8++ CA S+ZERO
TS KEEP7 # +0
TS L
CS A
TS KEEP6 # -0
EXTEND # A = -0, L = +0
DV KEEP7 # A = L = +0
TS KEEP1
CA L # C(A) = C(L) = +0
TC -0CHK -1
DV8+- CS A
EXTEND # A = -0, L = +0
DV KEEP6 # A = -0, L = +0
AD KEEP1
TC -0CHK
CS A
XCH L # PUT -0 IN L
TC -0CHK -1 # CHECK C(L)
DV8-+ EXTEND # A = +0, L = -0
DV KEEP7 # A = L = -0
TS KEEP2
AD KEEP1
TC -0CHK
CS A
XCH L # PUT -0 IN L
TC -0CHK # CHECK C(L)
DV8-- EXTEND # A = +0, L = -0
DV KEEP6 # A = +0, L = -0
AD KEEP2
TC -0CHK
CA S+MAX # CHECK QUOTIENT IS CORRECT
AD KEEP2
TC -0CHK
XCH L
TC -0CHK # CHECK C(L)
# INPUT-OUTPUT INSTRUCTIONS
# NORMAL USE OF READ AND WRITE
IN-OUT1 CA S-1
EXTEND
WRITE L # 77776
CS A # 00001
## Page 233
EXTEND
READ L # 77776
TC -1CHK
CA S-MAX
AD S-MAX # C(A) = 10 - 00001
EXTEND
WRITE Q
CS A # 01 - 37776
EXTEND
READ Q # 10 - 00001
TS KEEP1
TC ERRORS
CA KEEP1
AD S+MAX
TC -1CHK -1
# NORMAL USE OF RAND, RAND = READ AND MASK
RANDCHK CA S+ZERO
TS L
EXTEND
RAND L # 00000, 00000
TC -0CHK -1
CA S-ZERO
EXTEND
RAND L # 77777, 00000
TC -0CHK -1
CS A
TS L
CS A
EXTEND
RAND L # 00000, 77777
TC -0CHK -1
CA S-ZERO
EXTEND
RAND L # 77777, 77777
TC -0CHK
## The following test is slightly wrong. Its proper operation relies on the Q register
## storing an overflow condition. However, it is run without interrupts inhibited, so
## it's possible for an interrupt to occur and destroy Q's overflow (via QRUPT). The
## test designers were likely either unaware of this possibility, or thought that
## overlfow in Q would inhibit interrupts like it would in A. Or, perhaps, the
## digital simulator could have been wrong as discussed above for "DCA L".
## A correct version of the test would have an INHINT immediately prior to "XCH Q",
## and a RELINT sometime after "RAND Q".
RANDOV CA S+MAX
AD S+2 # 01 - 00001
XCH Q
CA S-ZERO # 77777
EXTEND
RAND Q # 01 - 00001
TS KEEP1
TC ERRORS
TC -1CHK -1
CS KEEP1
TC -1CHK
# NORMAL USE OF WAND, WAND = WRITE AND MASK
WANDCHK CA S+ZERO
## Page 234
TS L
EXTEND
WAND L # 00000, 00000
AD L
TC -0CHK -1
CA S-ZERO
EXTEND
WAND L # 77777, 00000
AD L
TC -0CHK -1
CS A
TS L
CS A
EXTEND
WAND L # 00000, 77777
AD L
TC -0CHK -1
CA S-ZERO
TS L
EXTEND
WAND L # 77777, 77777
AD L
TC -0CHK
## The following test is slightly wrong. (See the discussion for RANDOV above.)
WANDUF CA S+MAX
AD S+2 # 01 - 00001
XCH Q
CA S-ZERO # 77777
EXTEND
WAND Q
TS KEEP2
TC ERRORS
CA Q
TS KEEP1
TC ERRORS
CS KEEP1
TC -1CHK
CS KEEP2
TC -1CHK
# NORMAL USE OF ROR, READ AND SUPERIMPOSE
RORCHK CA S+ZERO
TS L
EXTEND
ROR L # 00000, 00000
TC -0CHK -1
CA L
TC -0CHK -1
CA S-ZERO
EXTEND
ROR L # 77777, 00000
## Page 235
TC -0CHK
CA L
TC -0CHK -1
CS A
TS L
CS A
EXTEND
ROR L # 00000, 77777
TC -0CHK
CA L
TC -0CHK
CA S-ZERO
EXTEND
ROR L # 77777, 77777
TC -0CHK
## The following test is slightly wrong. (See the discussion for RANDOV above.)
ROROV CA S-MAX
AD S-2 # 10 - 37776
XCH Q
CA S+MAX
AD S+1 # 01 - 00000
EXTEND
ROR Q # 11 - 37776
TS KEEP1
TC +2
TC ERRORS
CA KEEP1
TC -1CHK
# NORMAL USE OF WOR, WOR = WRITE AND SUPERIMPOSE
WORCHK CA S+ZERO
TS L
EXTEND
WOR L # 00000, 00000
TC -0CHK -1
CA L
TC -0CHK -1
CA S-ZERO
EXTEND
WOR L # 77777, 00000
TC -0CHK
CA L
TC -0CHK
CS A
TS L
CS A
EXTEND
WOR L # 00000, 77777
TC -0CHK
CA L
TC -0CHK
## Page 236
CA S-MAX
EXTEND
WOR L # 77777, 77777
TC -0CHK
CA L
TC -0CHK
## The following test is slightly wrong. (See the discussion for RANDOV above.)
WOROV CA S-MAX
AD S-2 # 10 - 37776
XCH Q
CA S+MAX
AD S+1 # 01 - 00000
EXTEND
WOR Q # 11 - 37776
TS KEEP2 # SHOULD NOT SKIP
TCF +2
TC ERRORS
XCH Q
TS KEEP3
TC +2
TC ERRORS
CA KEEP3 # CHECK C(Q)
TC -1CHK
CA KEEP2
TC -1CHK # CHECK C(A)
# NORMAL USE OF RXOR
RXORCHK CA S+ZERO
TS L
EXTEND
RXOR L # 00000, 00000
TC -0CHK -1
CA S-ZERO
EXTEND
RXOR L # 77777, 00000
TC -0CHK
CA L
TC -0CHK -1
CS A
TS L
CS A
EXTEND
RXOR L # 00000, 77777
TC -0CHK
CA S-ZERO
EXTEND
RXOR L # 77777, 77777
TC -0CHK -1
CA L
TC -0CHK
CS A
## Page 237
TS Q
## The following test is slightly wrong. (See the discussion for RANDOV above.)
RXORUV CA S+MAX
AD S+2 # 01 - 00001
EXTEND
RXOR Q # 10 - 37776, C(Q) = -0
TS KEEP1
TC ERRORS
CA KEEP1
TC -1CHK
TC CHECKNJ # CHECK FOR NEW JOB
TC XTRANDX
# NEXT THREE CONSTANTS ARE ADDRESSESS USED BY EXTRACODE INSTRUCTIONS
ADRSBZMF ADRES NDXBZMF
ADRSDCA ADRES NDXDCA
ADRSQXCH ADRES NDXAUG
# NORMAL USE OF INDEX WITH EXTRACODE INSTRUCTIONS
# INDEX INSTRUCTION USED WITH INDEX AND BZF
XTRANDX CA S+ZERO
EXTEND
NDX S+ZERO
NDX ADRSBZMF
BZF 00000 # BZF +2
TC ERRORS
# INDEX INSTRUCTION USED WITH BZMF
NDXBZMF EXTEND
NDX ADRSDCA
BZMF 0000 # BZMF+2
TC ERRORS
# INDEX INSTRUCTION USED WITH DCA
NDXDCA EXTEND
INDEX ADRS+1 # DCA S+1
DCA 0000 # C(A) = +1, C(L) = +2
CS A
AD L
TC -1CHK -1
# INDEX INSTRUCTION USED WITH DCS
EXTEND
INDEX ADRS+1 # DCS S+1
DCS 0000 # C(A) = -1, C(L) = -2
CS A
AD L
TC -1CHK
# INDEX INSTRUCTION USED WITH MP AND SU
CA S+MAX # 37777
EXTEND
NDX ADRS+1
MP 0001 # C(A) = 1, C(L) = 37776
## Page 238
TC -1CHK -1
CA S+MAX # 37777
EXTEND
NDX S+1
SU 0000
TC -1CHK -1
# INDEX INSTRUCTION USED WITH DV
NDXDV CA DV1CON # PUT 14000 (3/8) IN KEEP3
TS KEEP3
CA S+ZERO
TS L
CA SBIT13 # 10000
EXTEND
NDX ADRS3
DV 0000 # C(A) = 25252, C(L) = 10000
TS KEEP1
CA S-ZERO
XCH L
CS A
EXTEND
NDX ADRS3
DV 0000
AD KEEP1
TC -0CHK
# INDEX USED WITH MSU (C(A) = +0, C(K) = -0) (RESULT = -1)
NDXMSU CA S+ZERO
TS KEEP1
CS A
EXTEND
NDX ADRS1 # MSU KEEP1
MSU 0000 # C(A) = -1
TC -1CHK
# INDEX USED WITH QXCH
NDXQXCH CA ADRSQXCH
TS KEEP1
TC +2
TC ERRORS
EXTEND
NDX ADRS1 # QXCH KEEP1
QXCH 0000
TC Q
TC ERRORS
# INDEX USED WITH AUG
NDXAUG CS S+ZERO
TS KEEP1 # 00000
EXTEND
NDX ADRS1 # AUG KEEP1
AUG 0000
CA KEEP1
TC -1CHK
## Page 239
# INDEX USED WITH DIM
NDXDIM CA S+2
TS KEEP1
EXTEND
NDX ADRS1 # DIM KEEP1
DIM 0000
CS KEEP1
TC -1CHK
# NORMAL USE OF INDEX WITH IN-OUT INSTRUCTIONS
# INDEX USED WITH WRITE AND READ
NDXINOUT CA S-1
EXTEND
NDX S+1 # WRITE L
WRITE 0000
CS A
EXTEND
NDX S+1 # READ L
READ 0000
TC -1CHK
# INDEX USED WITH RAND
CA S+1 # 00001
TS L
CA S-ZERO # 77777
EXTEND
NDX S+1 # RAND L
RAND 00000
TC -1CHK -1
# INDEX USED WITH WAND
CA S-MAX # 40000
AD S+1 # 40001
TS L
CA S+MAX # 37777
EXTEND
NDX S+1 # WAND L
WAND 0000 # C(A) = C(L) = +1
TC -1CHK -1
CS L
TC -1CHK
# INDEX USED WITH ROR
CA S+1
TS Q
CA S+ZERO
EXTEND
NDX S+2 # ROR Q
ROR 0000
TC -1CHK -1
# INDEX USED WITH WOR
CA S-7
TS L # 77770
## Page 240
CA S+MAX
AD S-1 # 37776
EXTEND
NDX S+1 # WOR L
WOR 0000 # C(A) = C(L) = -1
TC -1CHK
CA L
TC -1CHK
# INDEX USED WITH RXOR
CA S-1
TS L # 77776
CA S+MAX # 37777
EXTEND
NDX S+1 # RXOR L
RXOR 0000 # C(A) = 40001
AD S+MAX
TC -1CHK -1
# CHECKS EXTRACODE INDEX OF AN SC REGISTER
CA S+2
TS KEEP1
CA ADRS1 # ADDRESS OF KEEP1
TS L
CA S+1
EXTEND
NDX L
SU 0000
TC -1CHK
# CHECK OF SPECIAL AND CENTRAL REGISTERS
# CHANGE OF SIGN BY ADDING SAME NUMBER (ADDER)
ADDCHK CA SBIT14 # 20000
AD A # 01 -00000
TS A
TC ERRORS
AD A # 10 - 00000
TS A
TC ERRORS
AD A # 00001
TS A
TC +2
TC ERRORS
TC -1CHK -1
# NORMAL OPERATION OF CYCLE RIGHT REGISTER
CYRCHK CA CYRCON # 57761
TS KEEP5 # COUNTDOWN REGISTER
CA S-MAX # 40000
TS CYR
CYRLOOP CCS CYR
TC CYRCNTDN
## Page 241
TC ERRORS
TC ENDCYR
TC ERRORS
CYRCNTDN INCR KEEP5
TC CYRLOOP
ENDCYR CA KEEP5 # 57777
AD CYR # C(CYR) = 20000
CCS A # -0 = END OF CYCLE RIGHT CHECK
TC ERRORS
TC ERRORS
TC ERRORS
# NORMAL OPERATION OF CYCLE LEFT REGISTER
CYLCHK CA S-15 # 77760, -15
TS KEEP5 # COUNT REGISTER
CA S-MAX # 40000
TS CYL
CYLLOOP CCS CYL
TC CYLCNTDN
TC ERRORS
TC ENDCYL
TC ERRORS
CYLCNTDN INCR KEEP5
TC CYLLOOP
ENDCYL CA CYL # C(CYL) SHOULD = +1
TC -1CHK -1
CA KEEP5
TC -1CHK
# NORMAL OPERATION OF SHIFT RIGHT REGISTER
SRCHK CA S-14 # 77761, -14
TS KEEP5 # COUNT REGISTER
CA S-MAX # 40000
TS SR
SRLOOP CCS SR
TC ERRORS
TC ERRORS
TC SRCNTDN
CA KEEP5 # HAS SHIFTED 14 TIMES
TC -1CHK
TC EDOPCHK # NEXT SUBROUTINE
SRCNTDN INCR KEEP5 # INCREMENT COUNT REGISTER
TC SRLOOP
# NORMAL OPERATION OF EDOP REGISTER. BITS 8 - 14 OF G REGISTER GO TO
# BITS 1 - 7 OF EDOP.
EDOPCHK CA S-15 # 77760, -15
TS KEEP5 # COUNT REGISTER
CA S7BITS # 00177
TS CYL
## Page 242
EDOPLOOP CA CYL
TS EDOP
TS CYR # SHIFT LEFT 7 TIMES
CA CYR
CA CYR
CA CYR
CA CYR
CA CYR
CA CYR
CA CYR
MASK S7BITS
CS A
TS KEEP1 # COMPLEMEMT OF C(EDOP)
CA S-ZERO
MASK EDOP
AD KEEP1
TC -0CHK
INCR KEEP5 # INCREMEMT COUNT REGISTER
CCS EDOP
TC EDOPLOOP
TC ENDEDOP
TC ERRORS
TC ERRORS
ENDEDOP CA KEEP5 # SHOULD HAVE PERFORMED EDOPLOOP 14 TIMES
TC -1CHK
INCR SCOUNT # INCREMENT UPON SUCCESSFUL COMLETION
TC SMODECHK
Computing file changes ...
