Source Code
|
These source-code files are part of a reconstructed copy of Comanche 44, the
initial release (though not the final mission release) of the Apollo 10
Command Module (CM) Apollo Guidance Computer (AGC) software. Subsequent releases
were Comanche 45 and Comanche 45/2 (the flown release on Apollo 10). The reconstruction
is believed to be exactly accurate, but in the absence of an actual Apollo-era
listing of Comanche 44, some explanation of the reconstruction process is needed
to insure confidence in that accuracy.
The starting point was the source code of Comanche 51
(separately previously reconstructed from an Apollo-era listing of Comanche 55,
flown on Apollo 11). Comanche 51 source code was then modified
by undoing changes known to have been made between revisions 44 and 51. The now-reconstructed
Comanche 44 source code was verified by assembling it and checking that the assembled code had
the expected known-correct memory-bank checksums.
Note that page numbers in the reconstructed code match those on the Comanche 55 printout,
although the source code would likely have different page numbers in a contemporary
Comanche 44 listing.
Annotations that were not present in the contemporary source code have been added to the reconstructed
code to justify each change relative to Comanche 51. Here's a guide to the Apollo documentation referenced
in those annotations:
In most cases, only the titles (rather than the full texts) of PCRs are known. In addition to the
explanatory annotations, there is also
a step-by-step commentary about the reconstruction process.
Comments expected to have been present in the original source code are prefixed with a single '#' symbol,
whereas comments added later are prefixed by "##" or "###". Report any errors
noted by creating an
issue report at the Virtual AGC
Project's GitHub repository.
|
048552,000002: ## Copyright: Public domain.
048553,000003: ## Filename: INTERPRETER.agc
048554,000004: ## Purpose: A section of Comanche revision 044.
048555,000005: ## It is part of the reconstructed source code for the
048556,000006: ## original release of the flight software for the Command
048557,000007: ## Module's (CM) Apollo Guidance Computer (AGC) for Apollo 10.
048558,000008: ## The code has been recreated from a copy of Comanche 055. It
048559,000009: ## has been adapted such that the resulting bugger words
048560,000010: ## exactly match those specified for Comanche 44 in NASA drawing
048561,000011: ## 2021153D, which gives relatively high confidence that the
048562,000012: ## reconstruction is correct.
048563,000013: ## Assembler: yaYUL
048564,000014: ## Contact: Ron Burkey <info@sandroid.org>.
048565,000015: ## Website: www.ibiblio.org/apollo/index.html
048566,000016: ## Mod history: 2020-12-03 MAS Created from Comanche 51.
048567,000017: ## 2021-05-30 ABS Modfied COUNT operands to match Comanche 55.
048568,000018:
|
|
Page 1107
|
048570,000020: # SECTION 1 DISPATCHER
048571,000021:
048572,000022: # ENTRY TO THE INTERPRETER. INTPRET SETS LOC TO THE FIRST INSTRUCTION, BANKSET TO THE BBANK OF THE
048573,000023: # OBJECT INTERPRETIVE PROGRAM, AND INTBIT15 TO THE BIT15 CONTENTS OF FBANK. INTERPRETIVE PROGRAMS MAY BE IN
048574,000024: # VIRTUALLY ALL BANKS PRESENT UNDER ANY SUPER-BANK SETTING, WITH THE RESTRICTION THAT PROGRAMS IN HIGH BANKS
048575,000025: # (BIT15 OF FBANK = 1) DO NOT REFER TO LOWBANKS, AND VICE-VERSA. THE INTERPRETER DOES NOT SWITCH SUPERBANKS.
048576,000026: # E-BANK SWITCHING OCCURS WHENEVER GENERAL ERASABLE (100-3777) IS ADDRESSED.
048577,000027:
048578,000028: 6006 BLOCK 03
048579,000029:
048580,000030: 6006 COUNT 03/INTER
048581,000031: 6006 00003 INTPRET RELINT
048582,000032: 6007 00006 EXTEND # SET LOC TO THE WORD FOLLOWING THE TC.
048583,000033: 6010 22164 QXCH LOC
048584,000034: 6011 30006 +2 CA BBANK # INTERPRETIVE BRANCHES FINISH HERE.
048585,000035: 6012 54165 TS BANKSET
048586,000036: 6013 75001 MASK BIT15 # GET 15TH BIT FOR INDEXABLE ADDRESSES.
048587,000037: 6014 54115 TS INTBIT15
048588,000038:
048589,000039: 6015 54023 TS EDOP # MAKE SURE NO INSTRUCTIONS LEFT OVER
048590,000040:
048591,000041: 6016 16037 TCF NEWOPS # PICK UP OP CODE PAIR AND BEGIN.
048592,000042:
048593,000043: 6017 22006 INTRSM LXCH BBANK # RESUME SUSPENDED INTERPRETIVE JOB
048594,000044: 6020 16011 TCF INTPRET +3
048595,000045:
048596,000046: # DLOAD LOADS MPAC, MPAC +1, LEAVING ZERO IN MPAC +2.
048597,000047:
048598,000048: 6021 00006 DLOAD EXTEND
048599,000049: 6022 50116 INDEX ADDRWD
048600,000050: 6023 30001 DCA 0 # LOAD DP C(C(ADDRWD)) INTO MPAC,MPAX +1
048601,000051: 6024 52155 SLOAD2 DXCH MPAC
048602,000052: 6025 35021 CAF ZERO # ZERO MPAC +2
048603,000053:
|
|
Page 1108
|
048605,000055: # AT THE END OF MOST INSTRUCTIONS, CONTROL IS GIVEN TO DANZIG TO DISPATCH THE NEXT OPERATION.
048606,000056:
048607,000057: 6026 54156 TS MPAC +2 # AND DECLARE DP MODE
048608,000058:
048609,000059: 6027 54163 NEWMODE TS MODE # PROLOGUE FOR MODE-CHANGING INSTRUCTIONS.
048610,000060:
048611,000061: 6030 30165 DANZIG CA BANKSET # SET BBANK BEFORE TESTING NEWJOB SO THAT
048612,000062: 6031 54006 TS BBANK # IT MAY BE SAVED DIRECTLY BY CHANJOB.
048613,000063:
048614,000064: 6032 10023 NOIBNKSW CCS EDOP # SEE IF AN ORDER CODE IS LEFT OVER FROM
048615,000065: 6033 16046 TCF OPJUMP # THE LAST PAIR RETRIEVED. IF SO, EXECUTE.
048616,000066: # EDOP IS SET TO ZERO ON ITS RE-EDITING.
048617,000067:
048618,000068: 6034 10067 CCS NEWJOB # SEE IF A JOB OF HIGHER PRIORITY IS
048619,000069: 6035 15170 TCF CHANG2 # PRESENT, AND IF SO, CHANGE JOBS.
048620,000070:
048621,000071: 6036 24164 INCR LOC # ADVANCE THE LOCATION COUNTER.
048622,000072:
048623,000073: # ITRACE (1) REFERS TO "NEWOPS".
048624,000074: 6037 50164 NEWOPS INDEX LOC # ENTRY TO BEGIN BY PICKING OP CODE PAIR.
048625,000075: 6040 30000 CA 0 # MAY BE AN OPCODE PAIR OR A STORE CODE.
048626,000076: 6041 10000 CCS A # TEST SIGN AND GET DABS(A).
048627,000077: 6042 16331 TCF DOSTORE # PROCESS STORE CODE.
048628,000078:
048629,000079: 6043 00177 LOW7 OCT 177
048630,000080:
048631,000081: 6044 54023 TS EDOP # OP CODE PAIR. LEAVE THE OTHER IN EDOP
048632,000082: 6045 76043 MASK LOW7 # WHERE CCS EDOP WILL HONOR IT NEXT.
048633,000083:
048634,000084: 6046 54020 OPJUMP TS CYR # LOWWD ENTERS HERE IF A RIGHT-HAND OP
048635,000085: 6047 10020 CCS CYR # CODE IS TO BE PROCESSED. TEST PREFICES.
048636,000086: 6050 16216 TCF OPJUMP2 # TEST SECOND PREFIX BIT.
048637,000087:
048638,000088: 6051 16712 TCF EXIT # +0 OP CODE IS EXIT.
048639,000089:
|
|
Page 1109
|
048641,000091: # PROCESS ADDRESSES WHICH MAY BE DIRECT, INDEXED, OR REFERENCE THE PUSHDOWN LIST.
048642,000092:
048643,000093: 6052 75017 ADDRESS MASK BIT1 # SEE IF ADDRESS IS INDEXED. CYR CONTAINED
048644,000094: 6053 10000 CCS A # 400XX, SO BIT 1 IS NOW AS IT WAS IN CYR.
048645,000095: 6054 16115 TCF INDEX # FORM INDEXED ADDRESS.
048646,000096:
048647,000097: 6055 50164 DIRADRES INDEX LOC # LOOK AHEAD TO NEXT WORD TO SEE IF
048648,000098: 6056 40001 OCT40001 CS 1 # ADDRESS IS GIVEN.
048649,000099: 6057 10000 CCS A
048650,000100: 6060 16164 TCF PUSHUP # IF NOT.
048651,000101:
048652,000102: 6061 77773 NEG4 DEC -4 B-14
048653,000103:
048654,000104: 6062 24164 INCR LOC # IF SO, TO SHOW WE PICKED UP A WORD.
048655,000105: 6063 54116 TS ADDRWD
048656,000106:
|
|
Page 1110
|
048658,000108: # FINAL DIGESTION OF DIRECT ADDRESSES OF OP CODES WITH 01 PREFIX IS DONE HERE. IN EACH CASE, THE
048659,000109: # REQUIRED 12-BIT SUB-ADDRESS IS LEFT IN ADDRWD, WITH ANY REQUIRED E OR F BANK SWITCHING DONE. ADDRESSES LESS
048660,000110: # THAN 45D ARE TAKEN TO BE RELATIVE TO THE WORK AREA. THE OP CODE IS NOW IN BITS 1-5 OF CYR WITH BIT 14 = 1.
048661,000111:
048662,000112: 6064 66220 AD -ENDVAC # SEE IF ADDRESS RELATIVE TO WORK AREA.
048663,000113: 6065 10000 CCS A
048664,000114: 6066 67712 AD -ENDERAS # IF NOT, SEE IF IN GENERAL ERASABLE.
048665,000115: 6067 16074 TCF IERASTST
048666,000116:
048667,000117: 6070 30120 NETZERO CA FIXLOC # IF SO, LEAVE THE MODIFIED ADDRESS IN
048668,000118: 6071 26116 ADS ADDRWD # ADDRWD AND DISPATCH.
048669,000119: 6072 50020 ITR15 INDEX CYR # THIS INDEX MAKES THE NEXT INSTRUCTION
048670,000120: 6073 76242 7 INDJUMP -1 # TCF INDJUMP + OP, EDITING CYR.
048671,000121:
048672,000122: 6074 00006 IERASTST EXTEND
048673,000123: 6075 66105 BZMF GEADDR # GO PROCESS GENERAL-ERASABLE ADDRESS.
048674,000124:
048675,000125: 6076 75054 MASK LOW10 # FIXED BANK ADDRESS. RESTORE AND ADD B15.
048676,000126: 6077 65054 AD LOW10 # SWITCH BANKS AND LEAVE SUBADDRESS IN
048677,000127: 6100 56116 XCH ADDRWD # ADDRWD FOR OPERAND RETRIEVAL. (THIS
048678,000128: 6101 60115 AD INTBIT15 # METHOD PRECLUDES USE OF THE LAST
048679,000129: 6102 54004 TS FBANK # LOCATION IN EACH FBANK.)
048680,000130: 6103 50020 ITR12 INDEX CYR
048681,000131: 6104 76242 7 INDJUMP -1
048682,000132:
048683,000133: 6105 74373 GEADDR MASK LOW8
048684,000134: 6106 65051 AD OCT1400
048685,000135: 6107 56116 XCH ADDRWD
048686,000136: 6110 54003 TS EBANK
048687,000137: 6111 50020 ITR10 INDEX CYR
048688,000138: 6112 76242 7 INDJUMP -1
048689,000139:
|
|
Page 1111
|
048691,000141: # THE FOLLOWING ROUTINE PROCESSES INTERPRETIVE INDEXED ADDRESSES. AN INTERPRETER INDEX REGISTER MAY
048692,000142: # CONTAIN THE ADDRESS OF ANY ERASABLE REGISTER (0-42 BEING RELATIVE TO THE VAC AREA) OR ANY INTERPRETIVE PROGRAM
048693,000143: # BANK, OR ANY INTEGER IN THAT RANGE.
048694,000144:
048695,000145: 6113 37702 DODLOAD* CAF DLOAD* # STODL* COMES HERE TO PROCESS LOAD ADR.
048696,000146: 6114 54020 TS CYR # (STOVL* ENTERS HERE).
048697,000147:
048698,000148: 6115 30120 INDEX CA FIXLOC # SET UP INDEX LOCATION.
048699,000149: 6116 54130 TS INDEXLOC
048700,000150: 6117 24164 INCR LOC # (ADDRESS ALWAYS GIVEN).
048701,000151: 6120 50164 INDEX LOC
048702,000152: 6121 40000 CS 0
048703,000153: 6122 10000 CCS A # INDEX 2 IF ADDRESS STORED COMPLEMENTED.
048704,000154: 6123 24130 INCR INDEXLOC
048705,000155: 6124 16125 NOOP
048706,000156:
048707,000157: 6125 54116 TS ADDRWD # 14 BIT ADDRESS TO ADDRWD.
048708,000158: 6126 77711 MASK HIGH4 # IF ADDRESS GREATER THAN 2K, ADD INTBIT15
048709,000159: 6127 00006 EXTEND
048710,000160: 6130 16133 BZF INDEX2
048711,000161: 6131 30115 CA INTBIT15
048712,000162: 6132 26116 ADS ADDRWD
048713,000163:
048714,000164: 6133 50130 INDEX2 INDEX INDEXLOC
048715,000165: 6134 40046 CS X1
048716,000166: 6135 26116 ADS ADDRWD # DO AUGMENT, IGNORING AND CORRECTING OVF.
048717,000167:
048718,000168: 6136 77713 MASK HIGH9 # SEE IF ADDRESS IS IN WORK AREA.
048719,000169: 6137 00006 EXTEND
048720,000170: 6140 16153 BZF INDWORK
048721,000171: 6141 77711 MASK HIGH4 # SEE IF IN FIXED BANK.
048722,000172: 6142 00006 EXTEND
048723,000173: 6143 16155 BZF INDERASE
048724,000174:
048725,000175: 6144 30116 CA ADDRWD # IN FIXED - SWITCH BANKS AND CREATE
048726,000176: 6145 54004 TS FBANK # SUB-ADDRESS.
048727,000177: 6146 75054 MASK LOW10
048728,000178: 6147 65005 AD 2K
048729,000179: 6150 54116 TS ADDRWD
048730,000180: 6151 50020 ITR11 INDEX CYR
048731,000181: 6152 36242 3 INDJUMP -1
048732,000182:
048733,000183: 6153 30120 INDWORK CA FIXLOC # MAKE ADDRWD RELATIVE TO WORK AREA.
048734,000184: 6154 16161 TCF ITR13 -1
048735,000185:
048736,000186: 6155 35051 INDERASE CA OCT1400
048737,000187: 6156 56116 XCH ADDRWD
048738,000188: 6157 54003 TS EBANK
048739,000189: 6160 74373 MASK LOW8
048740,000190: 6161 26116 -1 ADS ADDRWD
|
|
Page 1112
|
048742,000192: 6162 50020 ITR13 INDEX CYR
048743,000193: 6163 36242 3 INDJUMP -1
048744,000194:
|
|
Page 1113
|
048746,000196: # PUSH-UP ROUTINES. WHEN NO OPERAND ADDRESS IS GIVEN, THE APPROPRIATE OPERAND IS TAKEN FROM THE PUSH-DOWN
048747,000197: # LIST. IN MOST CASES THE MODE OF THE RESULT (VECTOR OR SCALAR) OF THE LAST ARITHMETIC OPERATION PERFORMED
048748,000198: # IS THE SAME AS THE TYPE OF OPERAND DESIRED (ALL ADD/SUBTRACT ETC.). EXCEPTIONS TO THIS GENERAL RULE ARE LISTED
048749,000199: # BELOW (NOTE THAT IN EVERY CASE THE MODE REGISTER IS LEFT INTACT):
048750,000200:
048751,000201: # 1. VXSC AND V/SC WANT THE OPPOSITE TYPE OF OPERAND, E.G., IF THE LAST OPERATION YIELDED A VECTOR
048752,000202: # RESULT, VXSC WANTS A SCALAR.
048753,000203:
048754,000204: # 2. THE LOAD CODES SHOULD LOAD THE ACCUMULATOR INDEPENDENT OF THE RESULT OF THE LAST OPERATION. THIS
048755,000205: # INCLUDES VLOAD, DLOAD, TLOAD, PDDL, AND PDVL (NO PUSHUP WITH SLOAD).
048756,000206:
048757,000207: # 3. SOME ARITHMETIC OPERATIONS REQUIRE A STANDARD TYPE OF OPERAND REGARDLESS OF THE PREVIOUS OPERATION.
048758,000208: # THIS INCLUDES SIGN WANTING DP AND TAD REQUIRING TP.
048759,000209:
048760,000210: 6164 34374 PUSHUP CAF OCT23 # IF THE LOW 5 BITS OF CYR ARE LESS THAN
048761,000211: 6165 70020 MASK CYR # 20, THIS OP REQUIRES SPECIAL ATTENTION.
048762,000212: 6166 66171 AD -OCT10 # (NO -0).
048763,000213: 6167 10000 CCS A
048764,000214: 6170 16202 TCF REGUP # FOR ALL CODES GREATER THAN OCT 7.
048765,000215:
048766,000216: 6171 77767 -OCT10 OCT -10
048767,000217:
048768,000218: 6172 66061 AD NEG4 # WE NOW HAVE 7 - OP CODE (MOD4). SEE IF
048769,000219: 6173 10000 CCS A # THE OP CODE (MOD4) IS THREE (REVERSE).
048770,000220: 6174 50000 INDEX A # NO - THE MODE IS DEFINITE. PICK UP THE
048771,000221: 6175 46213 CS NO.WDS
048772,000222: 6176 16204 TCF REGUP +2
048773,000223:
048774,000224: 6177 50163 INDEX MODE # FOR VXSC AND V/SC WE WANT THE REQUIRED
048775,000225: 6200 46211 CS REVCNT # PUSHLOC DECREMENT WITHOUT CHANGING THE
048776,000226: 6201 16204 TCF REGUP +2 # MODE AT THIS TIME.
048777,000227:
048778,000228: 6202 50163 REGUP INDEX MODE # MOST ALL OP CODES PUSHUP HERE.
048779,000229: 6203 46213 CS NO.WDS
048780,000230: 6204 26166 +2 ADS PUSHLOC
048781,000231: 6205 54116 TS ADDRWD
048782,000232: 6206 50020 ITR14 INDEX CYR
048783,000233: 6207 76242 7 INDJUMP -1 # (THE INDEX MAKES THIS A TCF.)
048784,000234:
048785,000235: 6210 00002 OCT 2 # REVERSE PUSHUP DECREMENT. VECTOR TAKES 2
048786,000236: 6211 00006 REVCNT OCT 6 # WORDS, SCALAR TAKES 6.
048787,000237: 6212 00006 OCT 6
048788,000238: 6213 00002 NO.WDS OCT 2 # CONVENTIONAL DECREMENT IS 6 WORDS VECTOR
048789,000239: 6214 00003 OCTAL3 OCT 3 # 2 IN DP, AND 3 IN TP.
048790,000240: 6215 00006 OCT 6
048791,000241:
|
|
Page 1114
|
048793,000243: # TEST THE SECOND PREFIX BIT TO SEE IF THIS IS A MISCELLANEOUS OR A UNARY/SHORT SHIFT OPERATION.
048794,000244:
048795,000245: 6216 10020 OPJUMP2 CCS CYR # TEST SECOND PREFIX BIT.
048796,000246: 6217 16232 TCF OPJUMP3 # TEST THIRD BIT TO SEE IF UNARY OR SHIFT.
048797,000247:
048798,000248: 6220 77722 -ENDVAC DEC -45 B-14
048799,000249:
048800,000250: # THE FOLLOWING ROUTINE PROCESSES ADDRESSES OF SUFFIX CLASS 10. THEY ARE BASICALLY WORK AREA ADDRESSES
048801,000251: # IN THE RANGE 0-52, ERASABLE ECADR CONSTANTS FROM 100-3777, AND FCADRS ABOVE THAT. ALL 15 BITS ARE AVAILABLE
048802,000252: # IN CONTRAST TO SUFFIX 1, IN WHICH ONLY THE LOW ORDER 14 ARE AVAILABLE.
048803,000253:
048804,000254: 6221 24164 15BITADR INCR LOC # (ENTRY HERE FROM STCALL).
048805,000255: 6222 50164 INDEX LOC # PICK UP ADDRESS WORD.
048806,000256: 6223 30000 CA 0
048807,000257: 6224 54117 TS POLISH # WE MAY NEED A SUBADDRESS LATER.
048808,000258:
048809,000259: 6225 35055 CAF LOW7+2K # THESE INSTRUCTIONS ARE IN BANK 1.
048810,000260: 6226 54004 TS FBANK
048811,000261: 6227 70020 MASK CYR
048812,000262: 6230 50000 ITR7 INDEX A
048813,000263: 6231 16303 TCF MISCJUMP
048814,000264:
|
|
Page 1115
|
048816,000266: # COMPLETE THE DISPATCHING OF UNARY AND SHORT SHIFT OPERATIONS.
048817,000267:
048818,000268: 6232 54004 OPJUMP3 TS FBANK # CALL IN BANK 0 (BITS 11-15 OF A ARE 0.)
048819,000269: # ITRACE (6) REFERS TO "OPJUMP3".
048820,000270: 6233 10020 CCS CYR # TEST THIRD PREFIX BIT.
048821,000271: 6234 50000 INDEX A # THE DECREMENTED UNARY CODE IS IN BITS
048822,000272: 6235 12000 TCF UNAJUMP # 1-4 OF A (ZERO, EXIT, HAS BEEN DETECTED)
048823,000273:
048824,000274: 6236 10163 CCS MODE # ITS A SHORT SHIFT CODE. SEE IF PRESENT
048825,000275: 6237 12017 TCF SHORTT # SCALAR OR VECTOR.
048826,000276: 6240 12017 TCF SHORTT
048827,000277: 6241 12121 TCF SHORTV # CALLS THE APPROPRIATE ROUTINE.
048828,000278:
048829,000279: 6242 4364 FBANKMSK EQUALS BANKMASK
048830,000280: 6242 00122 LVBUF ADRES VBUF
048831,000281:
|
|
Page 1116
|
048833,000283: # THE FOLLOWING IS THE JUMP TABLE FOR OP CODES WHICH MAY HAVE INDEXABLE ADDRESSES OR MAY PUSH UP.
048834,000284:
048835,000285: 6243 16454 INDJUMP TCF VLOAD # 00 - LOAD MPAC WITH A VECTOR.
048836,000286: 6244 17040 TCF TAD # 01 - TRIPLE PRECISION ADD TO MPAC.
048837,000287: 6245 17624 TCF SIGN # 02 - COMPLEMENT MPAC (V OR SC) IF X NEG.
048838,000288: 6246 17350 TCF VXSC # 03 - VECTOR TIMES SCALAR.
048839,000289: 6247 16652 TCF CGOTO # 04 - COMPUTED GO TO.
048840,000290: 6250 16437 TCF TLOAD # 05 - LOAD MPAC WITH TRIPLE PRECISION.
048841,000291: 6251 16021 TCF DLOAD # 06 - LOAD MPAC WITH A DP SCALAR.
048842,000292: 6252 17573 TCF V/SC # 07 - VECTOR DIVIDED BY SCALAR.
048843,000293:
048844,000294: 6253 16450 TCF SLOAD # 10 - LOAD MPAC IN SINGLE PRECISION.
048845,000295: 6254 16567 TCF SSP # 11 - SET SINGLE PRECISION INTO X.
048846,000296: 6255 16472 TCF PDDL # 12 - PUSH DOWN MPAC AND RE-LOAD IN DP.
048847,000297: 6256 17303 TCF MXV # 13 - MATRIX POST-MULTIPLIED BY VECTOR.
048848,000298: 6257 16526 TCF PDVL # 14 - PUSH DOWN AND VECTORLOAD.
048849,000299: 6260 16575 TCF CCALL # 15 - COMPUTED CALL.
048850,000300: 6261 17306 TCF VXM # 16 - MATRIX PRE-MULTIPLIED BY VECTOR.
048851,000301: 6262 17565 TCF TSLC # 17 - NORMALIZE MPAC (SCALAR ONLY).
048852,000302:
048853,000303: 6263 17543 TCF DMPR # 20 - DP MULTIPLY AND ROUND.
048854,000304: 6264 17546 TCF DDV # 21 - DP DIVIDE BY.
048855,000305: 6265 17552 TCF BDDV # 22 - DP DIVIDE INTO.
048856,000306: 6266 17570 TCF GSHIFT # 23 - GENERAL SHIFT INSTRUCTION.
048857,000307: 6267 16720 TCF VAD # 24 - VECTOR ADD.
048858,000308: 6270 16716 TCF VSU # 25 - VECTOR SUBTRACT.
048859,000309: 6271 17005 TCF BVSU # 26 - VECTOR SUBTRACT FROM.
048860,000310: 6272 17300 TCF DOT # 27 - VECTOR DOT PRODUCT.
048861,000311:
048862,000312: 6273 17427 TCF VXV # 30 - VECTOR CROSS PRODUCT.
048863,000313: 6274 17374 TCF VPROJ # 31 - VECTOR PROJECTION.
048864,000314: 6275 16754 TCF DSU # 32 - DP SUBTRACT.
048865,000315: 6276 17031 TCF BDSU # 33 - DP SUBTRACT FROM.
048866,000316: 6277 16744 TCF DAD # 34 - DP ADD.
048867,000317: 6300 16300 TCF +0 # 35 - AVAILABLE
048868,000318: 6301 17541 TCF DMP1 # 36 - DP MULTIPLY.
048869,000319: 6302 17562 TCF SETPD # 37 - SET PUSH DOWN POINTER (DIRECT ONLY)
048870,000320:
048871,000321: # CODES 10 AND 14 MUST NOT PUSH UP. CODE 04 MAY BE USED FOR VECTOR DECLARE BEFORE PUSHUP IF DESIRED.
048872,000322:
|
|
Page 1117
|
048874,000324: # THE FOLLOWING JUMP TABLE APPLIES TO INDEX, BRANCH, AND MISCELLANEOUS INSTRUCTIONS.
048875,000325:
048876,000326: 6303 12371 MISCJUMP TCF AXT # 00 - ADDRESS TO INDEX TRUE.
048877,000327: 6304 12376 TCF AXC # 01 - ADDRESS TO INDEX COMPLEMENTED.
048878,000328: 6305 12401 TCF LXA # 02 - LOAD INDEX FROM ERASABLE.
048879,000329: 6306 12405 TCF LXC # 03 - LOAD INDEX FROM COMPLEMENT OF ERAS.
048880,000330: 6307 12411 TCF SXA # 04 - STORE INDEX IN ERASABLE.
048881,000331: 6310 12417 TCF XCHX # 05 - EXCHANGE INDEX WITH ERASABLE.
048882,000332: 6311 12433 TCF INCR # 06 - INCREMENT INDEX REGISTER.
048883,000333: 6312 12442 TCF TIX # 07 - TRANSFER ON INDEX.
048884,000334:
048885,000335: 6313 12425 TCF XAD # 10 - INDEX REGISTER ADD FROM ERASABLE.
048886,000336: 6314 12436 TCF XSU # 11 - INDEX SUBTRACT FROM ERASABLE.
048887,000337: 6315 12514 TCF BZE/GOTO # 12 - BRANCH ZERO AND GOTO.
048888,000338: 6316 12521 TCF BPL/BMN # 13 - BRANCH PLUS AND BRANCH MINUS.
048889,000339: 6317 12474 TCF RTB/BHIZ # 14 - RETURN TO BASIC AND BRANCH HI ZERO.
048890,000340: 6320 12534 TCF CALL/ITA # 15 - CALL AND STORE QPRET.
048891,000341: 6321 12543 TCF SW/ # 16 - SWITCH INSTRUCTIONS AND AVAILABLE.
048892,000342: 6322 12504 TCF BOV(B) # 17 - BRANCH ON OVERFLOW TO BASIC OR INT.
048893,000343:
|
|
Page 1118
|
048895,000345: # THE FOLLOWING JUMP TABLE APPIES TO UNARY INSTRUCTIONS.
048896,000346: 6323 COUNT 00/INTER
048897,000347: 00,2000 BANK 0 # 00 - EXIT - DETECTED EARLIER.
048898,000348: 00,2000 13207 UNAJUMP TCF SQRT # 01 - SQUARE ROOT.
048899,000349: 00,2001 13527 TCF SINE # 02 - SIN.
048900,000350: 00,2002 13516 TCF COSINE # 03 - COS.
048901,000351: 00,2003 13607 TCF ARCSIN # 04 - ARC SIN.
048902,000352: 00,2004 13611 TCF ARCCOS # 05 - ARC COS.
048903,000353: 00,2005 13174 TCF DSQ # 06 - DP SQUARE.
048904,000354: 00,2006 12116 TCF ROUND # 07 - ROUND TO DP.
048905,000355:
048906,000356: 00,2007 17637 TCF COMP # 10 - COMPLEMENT VECTOR OR SCALAR.
048907,000357: 00,2010 13232 TCF VDEF # 11 - VECTOR DEFINE.
048908,000358: 00,2011 13023 TCF UNIT # 12 - UNIT VECTOR.
048909,000359: 00,2012 13176 TCF ABVALABS # 13 - LENGTH OF VECTOR OR MAG OF SCALAR.
048910,000360: 00,2013 13245 TCF VSQ # 14 - SQUARE OF LENGTH OF VECTOR.
048911,000361: 00,2014 16323 TCF STADR # 15 - PUSH UP ON STORE CODE.
048912,000362: 00,2015 13274 TCF RVQ # 16 - RETURN VIA QPRET.
048913,000363: 00,2016 13247 TCF PUSH # 17 - PUSH MPAC DOWN.
048914,000364:
|
|
Page 1119
|
048916,000366: # SECTION 2 LOAD AND STORE PACKAGE.
048917,000367:
048918,000368: # A SET OF EIGHT STORE CODES IS PROVIDED AS THE PRIMARY METHOD OF STORING THE MULTI-PURPOSE
048919,000369: # ACCUMULATOR (MPAC). IF IN THE DANZIG SECTION LOC REFERS TO AN ALGEBRAICALLY POSITIVE WORD, IT IS TAKEN AS A
048920,000370: # STORE CODE WITH A CORRESPONDING ERASABLE ADDRESS. MOST OF THESE CODES ARE TWO ADDRESS, SPECIFYING THAT THE WORD
048921,000371: # FOLLOWING THE STORE CODE IS TO BE USED AS AN ADDRESS FROM WHICH TO RE-LOAD MPAC. FOUR OPTIONS ARE AVAILABLE:
048922,000372:
048923,000373: # 1. STORE STORE MPAC. THE E ADDRESS MAY BE INDEXED.
048924,000374: # 2. STODL STORE MPAC AND RE-LOAD IT IN DP WITH THE NEXT ADDRESS (THE LOAD MAY BE INDEXED).
048925,000375: # 3. STOVL STORE MPAC AND RE-LOAD A VECTOR (AS ABOVE).
048926,000376: # 4. STCALL STORE AND DO A CALL (BOTH ADDRESSES MUST BE DIRECT HERE).
048927,000377:
048928,000378: # STODL AND STOVL WILL TAKE FROM THE PUSH-DOWN LIST IF NO LOAD ADDRESS IS GIVEN.
048929,000379:
048930,000380: 6323 BLOCK 3
048931,000381:
048932,000382: 6323 COUNT 03/INTER
048933,000383:
048934,000384: 6323 30165 STADR CA BANKSET # THE STADR CODE (PUSHUP UP ON STORE
048935,000385: 6324 54004 TS FBANK # ADDRESS) ENTERS HERE.
048936,000386: 6325 24164 INCR LOC
048937,000387: 6326 50164 ITR1 INDEX LOC # THE STORECODE WAS STORED COMPLEMENTED TO
048938,000388: 6327 40000 CS 0 # MAKE IT LOOK LIKE AN OPCODE PAIR.
048939,000389: 6330 67716 AD NEGONE # (YUL CANT REMOVE 1 BECAUSE OF EARLY CCS)
048940,000390:
048941,000391: 6331 54116 DOSTORE TS ADDRWD
048942,000392: 6332 74372 MASK LOW11 # ENTRY FROM DISPATCHER. SAVE THE ERASABLE
048943,000393: 6333 56116 XCH ADDRWD # ADDRESS AND JUMP ON THE STORE CODE NO.
048944,000394: 6334 77671 MASK B12T14
048945,000395: 6335 00006 EXTEND
048946,000396: 6336 75013 MP BIT5 # EACH TRANSFER VECTOR ENTRY IS TWO WORDS.
048947,000397: 6337 50000 ITR0 INDEX A
048948,000398: 6340 16341 TCF STORJUMP
048949,000399:
|
|
Page 1120
|
048951,000401: # STORE CODE JUMP TABLE. CALLS THE APPROPRIATE STORING ROUTINE AND EXITS TO DANZIG OR TO ADDRESS WITH
048952,000402: # A SUPPLIED OPERATION CODE.
048953,000403:
048954,000404: # STORE STORE,1 AND STORE,2 RETURN TO DANZIG, THUS RESETTING THE EBANK TO ITS STATE AT INTPRET.
048955,000405:
048956,000406: 6341 06371 STORJUMP TC STORE # STORE.
048957,000407: 6342 16030 TCF DANZIG # PICK UP NEW OP CODE(S).
048958,000408: 6343 06363 TC STORE,1
048959,000409: 6344 16030 TCF DANZIG
048960,000410: 6345 06366 TC STORE,2
048961,000411: 6346 16030 TCF DANZIG
048962,000412:
048963,000413: 6347 06371 TC STORE # STODL.
048964,000414: 6350 16427 TCF DODLOAD
048965,000415:
048966,000416: 6351 06371 TC STORE # STODL WITH INDEXED LOAD ADDRESS.
048967,000417: 6352 16113 TCF DODLOAD*
048968,000418:
048969,000419: 6353 06371 TC STORE # STOVL.
048970,000420: 6354 16432 TCF DOVLOAD
048971,000421:
048972,000422: 6355 06371 TC STORE # STOVL WITH INDEXED LOAD ADDRESS.
048973,000423: 6356 16435 TCF DOVLOAD*
048974,000424: 6357 06371 TC STORE # STOTC.
048975,000425: 6360 35030 CAF CALLCODE
048976,000426: 6361 54020 TS CYR
048977,000427: 6362 16221 TCF 15BITADR # GET A 15 BIT ADDRESS.
048978,000428:
|
|
Page 1121
|
048980,000430: # STORE CODE ADDRESS PROCESSOR.
048981,000431:
048982,000432: 6363 50120 STORE,1 INDEX FIXLOC
048983,000433: 6364 40046 CS X1
048984,000434: 6365 16370 TCF PRESTORE
048985,000435:
048986,000436: 6366 50120 STORE,2 INDEX FIXLOC
048987,000437: 6367 40047 CS X2
048988,000438: 6370 26116 PRESTORE ADS ADDRWD # RESULTANT ADDRESS IS IN ERASABLE.
048989,000439:
048990,000440: 6371 40116 STORE CS ADDRWD
048991,000441: 6372 65034 AD DEC45
048992,000442: 6373 10000 CCS A # DOES THE ADDRESS POINT TO THE WORK AREA?
048993,000443: 6374 30120 CA FIXLOC # YES.
048994,000444: 6375 16402 TCF AHEAD5
048995,000445: 6376 35051 CA OCT1400 # NO. SET EBANK & MAKE UP SUBADDRESS.
048996,000446: 6377 56116 XCH ADDRWD
048997,000447: 6400 54003 TS EBANK
048998,000448: 6401 74373 MASK LOW8
048999,000449: 6402 26116 AHEAD5 ADS ADDRWD
049000,000450:
|
|
Page 1122
|
049002,000452: # STORING ROUTINES. STORE DP, TP, OR VECTOR AS INDICATED BY MODE.
049003,000453:
049004,000454: 6403 00006 STARTSTO EXTEND # MPAC,+1 MUST BE STORED IN ANY EVENT.
049005,000455: # ITRACE (5) REFERS TO "STARTSTO".
049006,000456: 6404 30155 DCA MPAC
049007,000457: 6405 50116 INDEX ADDRWD
049008,000458: 6406 52001 DXCH 0
049009,000459:
049010,000460: 6407 10163 CCS MODE
049011,000461: 6410 16423 TCF TSTORE
049012,000462: 6411 00002 TC Q
049013,000463:
049014,000464: 6412 00006 VSTORE EXTEND
049015,000465: 6413 30160 DCA MPAC +3
049016,000466: 6414 50116 INDEX ADDRWD
049017,000467: 6415 52003 DXCH 2
049018,000468:
049019,000469: 6416 00006 EXTEND
049020,000470: 6417 30162 DCA MPAC +5
049021,000471: 6420 50116 INDEX ADDRWD
049022,000472: 6421 52005 DXCH 4
049023,000473: 6422 00002 TC Q
049024,000474:
049025,000475: 6423 30156 TSTORE CA MPAC +2
049026,000476: 6424 50116 INDEX ADDRWD
049027,000477: 6425 54002 TS 2
049028,000478: 6426 00002 TC Q
049029,000479:
|
|
Page 1123
|
049031,000481: # ROUTINES TO BEGIN PROCESSING OF THE SECOND ADDRESS ASSOCIATED WITH ALL STORE-TYPE CODES EXCEPT STORE
049032,000482: # ITSELF.
049033,000483:
049034,000484: 6427 37701 DODLOAD CAF DLOADCOD
049035,000485: 6430 54020 TS CYR
049036,000486: 6431 16055 TCF DIRADRES # GO GET A DIRECT ADDRESS.
049037,000487:
049038,000488: 6432 35001 DOVLOAD CAF VLOADCOD
049039,000489: 6433 54020 TS CYR
049040,000490: 6434 16055 TCF DIRADRES
049041,000491:
049042,000492: 6435 36056 DOVLOAD* CAF VLOAD*
049043,000493: 6436 16114 TCF DODLOAD* +1 # PROLOGUE TO INDEX ROUTINE.
049044,000494:
|
|
Page 1124
|
049046,000496: # THE FOLLOWING LOAD INSTRUCTIONS ARE PROVIDED FOR LOADING THE MULTI-PURPOSE ACCUMULATOR MPAC.
049047,000497:
049048,000498: 6437 50116 TLOAD INDEX ADDRWD
049049,000499: 6440 30002 CA 2 # LOAD A TRIPLE PRECISION ARGUMENT INTO
049050,000500: 6441 54156 TS MPAC +2 # THE FIRST THREE MPAC REGISTERS, WITH THE
049051,000501: 6442 00006 EXTEND # CONTENTS OF THE OTHER FOUR IRRELEVANT.
049052,000502: 6443 50116 INDEX ADDRWD
049053,000503: 6444 30001 DCA 0
049054,000504: 6445 52155 DXCH MPAC
049055,000505: 6446 35017 TMODE CAF ONE
049056,000506: 6447 16027 TCF NEWMODE # DECLARE TRIPLE PRECISION MODE.
049057,000507:
049058,000508: 6450 22007 SLOAD ZL # LOAD A SINGLE PRECISION NUMBER INTO
049059,000509: 6451 50116 INDEX ADDRWD # MPAC, SETTING MPAC+1,2 TO ZERO. THE
049060,000510: 6452 30000 CA 0 # CONTENTS OF THE REMAINING MPAC REGISTERS
049061,000511: 6453 16024 TCF SLOAD2 # ARE IRRELEVANT.
049062,000512:
049063,000513: 6454 00006 VLOAD EXTEND # LOAD A DOUBLE PRECISION VECTOR INTO
049064,000514: 6455 50116 INDEX ADDRWD # MPAC,+1, MPAC+3,4, AND MPAC+5,6. THE
049065,000515: 6456 30001 DCA 0 # CONTENTS OF MPAC +2 ARE IRRELEVANT.
049066,000516: 6457 52155 DXCH MPAC
049067,000517:
049068,000518: 6460 00006 ENDVLOAD EXTEND # PDVL COMES HERE TO FINISH UP FOR DP, TP.
049069,000519: 6461 50116 INDEX ADDRWD
049070,000520: 6462 30003 DCA 2
049071,000521: 6463 52160 DXCH MPAC +3
049072,000522:
049073,000523: 6464 00006 +4 EXTEND # TPDVL FINISHES HERE.
049074,000524: 6465 50116 INDEX ADDRWD
049075,000525: 6466 30005 DCA 4
049076,000526: 6467 52162 DXCH MPAC +5
049077,000527:
049078,000528: 6470 45017 VMODE CS ONE # DECLARE VECTOR MODE.
049079,000529: 6471 16027 TCF NEWMODE
049080,000530:
|
|
Page 1125
|
049082,000532: # THE FOLLOWING INSTRUCTIONS ARE PROVIDED FOR STORING OPERANDS IN THE PUSHDOWN LIST:
049083,000533: # 1. PUSH PUSHDOWN AND NO LOAD.
049084,000534: # 2. PDDL PUSHDOWN AND DOUBLE PRECISION LOAD.
049085,000535: # 3. PDVL PUSHDOWN AND VECTOR LOAD.
049086,000536:
049087,000537: 6472 00006 PDDL EXTEND
049088,000538: 6473 50116 INDEX ADDRWD # LOAD MPAC,+1, PUSHING THE FORMER
049089,000539: 6474 30001 DCA 0 # CONTENTS DOWN.
049090,000540: 6475 52155 DXCH MPAC
049091,000541: 6476 50166 INDEX PUSHLOC
049092,000542: 6477 52001 DXCH 0
049093,000543:
049094,000544: 6500 50163 INDEX MODE # ADVANCE THE PUSHDOWN POINTER APPRO-
049095,000545: 6501 36213 CAF NO.WDS # PRIATELY.
049096,000546: 6502 26166 ADS PUSHLOC
049097,000547:
049098,000548: 6503 10163 CCS MODE
049099,000549: 6504 16521 TCF ENDTPUSH
049100,000550: 6505 16517 TCF ENDDPUSH
049101,000551:
049102,000552: 6506 54163 TS MODE # NOW DP.
049103,000553: 6507 54156 ENDVPUSH TS MPAC +2
049104,000554: 6510 52160 DXCH MPAC +3 # PUSH DOWN THE REST OF THE VECTOR HERE.
049105,000555: 6511 50166 INDEX PUSHLOC
049106,000556: 6512 51775 DXCH 0 -4
049107,000557:
049108,000558: 6513 52162 DXCH MPAC +5
049109,000559: 6514 50166 INDEX PUSHLOC
049110,000560: 6515 51777 DXCH 0 -2
049111,000561:
049112,000562: 6516 16030 TCF DANZIG
049113,000563:
049114,000564: 6517 54156 ENDDPUSH TS MPAC +2 # SET MPAC +2 TO ZERO AND EXIT ON DP.
049115,000565: 6520 16030 TCF DANZIG
049116,000566:
049117,000567: 6521 54163 ENDTPUSH TS MODE
049118,000568: 6522 56156 XCH MPAC +2 # ON TRIPLE, SET MPAC +2 TO ZERO, PUSHING
049119,000569: 6523 50166 +2 INDEX PUSHLOC # DOWN THE OLD CONTENTS
049120,000570: 6524 53777 TS 0 -1
049121,000571: 6525 16030 TCF DANZIG
049122,000572:
|
|
Page 1126
|
049124,000574: # PDVL - PUSHDOWN AND VECTOR LOAD.
049125,000575:
049126,000576: 6526 00006 PDVL EXTEND # RELOAD MPAC AND PUSH DOWN ITS CONTENTS.
049127,000577: 6527 50116 INDEX ADDRWD
049128,000578: 6530 30001 DCA 0
049129,000579: 6531 52155 DXCH MPAC
049130,000580: 6532 50166 INDEX PUSHLOC
049131,000581: 6533 52001 DXCH 0
049132,000582:
049133,000583: 6534 50163 INDEX MODE # ADVANCE THE PUSHDOWN POINTER.
049134,000584: 6535 36213 CAF NO.WDS
049135,000585: 6536 26166 ADS PUSHLOC
049136,000586:
049137,000587: 6537 10163 CCS MODE # TEST PAST MODE.
049138,000588: 6540 16557 TCF TPDVL
049139,000589: 6541 16460 TCF ENDVLOAD # JUST LOAD LAST FOUR REGISTERS ON DP.
049140,000590:
049141,000591: 6542 00006 VPDVL EXTEND # PUSHDOWN AND RE-LOAD LAST TWO COMPONENTS
049142,000592: 6543 50116 INDEX ADDRWD
049143,000593: 6544 30003 DCA 2
049144,000594: 6545 52160 DXCH MPAC +3
049145,000595: 6546 50166 INDEX PUSHLOC
049146,000596: 6547 51775 DXCH 0 -4
049147,000597:
049148,000598: 6550 00006 EXTEND
049149,000599: 6551 50116 INDEX ADDRWD
049150,000600: 6552 30005 DCA 4
049151,000601: 6553 52162 DXCH MPAC +5
049152,000602: 6554 50166 INDEX PUSHLOC
049153,000603: 6555 51777 DXCH 0 -2
049154,000604:
049155,000605: 6556 16030 TCF DANZIG
049156,000606:
049157,000607: 6557 00006 TPDVL EXTEND # ON TP, WE MUST LOAD THE Y COMPONENT
049158,000608: 6560 50116 INDEX ADDRWD # BEFORE STORING MPAC +2 IN CASE THIS IS A
049159,000609: 6561 30003 DCA 2 # PUSHUP.
049160,000610: 6562 52160 DXCH MPAC +3
049161,000611:
049162,000612: 6563 30156 CA MPAC +2
049163,000613: 6564 50166 INDEX PUSHLOC # IN DP.
049164,000614: 6565 53777 TS 0 -1
049165,000615: 6566 16464 TCF ENDVLOAD +4
049166,000616:
049167,000617: # SSP (STORE SINGLE PRECISION) IS EXECUTED HERE.
049168,000618:
049169,000619: 6567 24164 SSP INCR LOC # PICK UP THE WORD FOLLOWING THE GIVEN
049170,000620: 6570 50164 INDEX LOC # ADDRESS AND STORE IT AT X.
049171,000621: 6571 30000 CA 0
049172,000622: 6572 50116 STORE1 INDEX ADDRWD # SOME INDEX AND MISCELLANEOUS OPS END
049173,000623: 6573 54000 TS 0 # HERE.
|
|
Page 1127
|
049175,000625: 6574 16030 TCF DANZIG
049176,000626:
|
|
Page 1128
|
049178,000628: # SEQUENCE CHANGING AND SUBROUTINE CALLING OPTIONS.
049179,000629:
049180,000630: # THE FOLLOWING OPERATIONS ARE AVAILABLE FOR SEQUENCING CHANGING, BRANCHING, AND CALLING SUBROUTINES:
049181,000631: # 1. GOTO GO TO.
049182,000632: # 2. CALL CALL SUBROUTINE SETTING QPRET.
049183,000633: # 3. CGOTO COMPUTED GO TO.
049184,000634: # 4. CCALL COMPUTED CALL.
049185,000635: # 7. BPL BRANCH IF MPAC POSITIVE OR ZERO.
049186,000636: # 8. BZE BRANCH IF MPAC ZERO.
049187,000637: # 9. BMN BRANCH IF MPAC NEGATIVE NON-ZERO.
049188,000638:
049189,000639: 6575 24164 CCALL INCR LOC # MAINTAIN LOC FOR QPRET COMPUTATION.
049190,000640: 6576 50164 INDEX LOC
049191,000641: 6577 30000 CAF 0 # GET BASE ADDRESS OF CADR LIST.
049192,000642: 6600 50116 INDEX ADDRWD
049193,000643: 6601 60000 AD 0 # ADD INCREMENT.
049194,000644: 6602 54004 TS FBANK # SELECT DESIRED CADR.
049195,000645: 6603 75054 MASK LOW10
049196,000646: 6604 50000 INDEX A
049197,000647: 6605 32000 CAF 10000
049198,000648: 6606 54117 TS POLISH
049199,000649:
049200,000650: 6607 30165 CALL CA BANKSET # FOR ANY OF THE CALL OPTIONS, MAKE UP THE
049201,000651: 6610 74364 MASK BANKMASK # ADDRESS OF THE NEXT OP-CODE PAIR/STORE
049202,000652: 6611 64364 AD BANKMASK # CODE AND LEAVE IT IN QPRET. NOTE THAT
049203,000653: 6612 60164 AD LOC # BANKMASK = -(2000 - 1).
049204,000654: 6613 50120 INDEX FIXLOC
049205,000655: 6614 54052 TS QPRET
049206,000656:
049207,000657: 6615 30117 GOTO CA POLISH # BASIC BRANCHING SEQUENCE.
049208,000658: 6616 77711 +1 MASK HIGH4
049209,000659: 6617 00006 EXTEND
049210,000660: 6620 16631 BZF GOTOERS # SEE IF ADDRESS POINTS TO FIXED OR ERAS.
049211,000661: 6621 30165 +4 CA BANKSET # SET EBANK PART OF BBANK. NEXT, SET UP
049212,000662: 6622 54006 TS BBANK # FBANK. THE COMBINATION IS PICKED UP &
049213,000663: 6623 30117 CA POLISH # PUT INTO BANKSET AT INTPRET +2.
049214,000664: 6624 54004 TS FBANK
049215,000665: 6625 75054 MASK LOW10
049216,000666: 6626 65005 AD 2K
049217,000667: 6627 54164 TS LOC
049218,000668: 6630 16011 TCF INTPRET +3
049219,000669:
049220,000670: 6631 E3,1400 EBANK= 1400 # SO YUL DOESN'T CUSS THE "CA 1400" BELOW.
049221,000671:
049222,000672: 6631 30117 GOTOERS CA POLISH # THE GIVEN ADDRESS IS IN ERASABLE - SEE
049223,000673: 6632 66220 AD -ENDVAC # IF RELATIVE TO THE WORK AREA.
049224,000674: 6633 10000 CCS A
049225,000675: 6634 30117 CA POLISH # GENERAL ERASABLE.
049226,000676: 6635 16644 TCF GOTOGE
049227,000677:
|
|
Page 1129
|
049229,000679: 6636 30120 CA FIXLOC # WORK AREA.
049230,000680: 6637 60117 AD POLISH
049231,000681: 6640 50000 INDEX A # USE THE GIVEN ADDRESS AS THE ADDRESS OF
049232,000682: 6641 30000 CA 0 # THE BRANCH ADDRESS.
049233,000683: 6642 54117 TS POLISH
049234,000684: 6643 16616 TCF GOTO +1 # ALLOWS ARBITRARY INDIRECTNESS LEVELS.
049235,000685:
049236,000686: 6644 54003 GOTOGE TS EBANK
049237,000687: 6645 74373 MASK LOW8
049238,000688: 6646 50000 INDEX A # USE THE GIVEN ADDRESS AS THE ADDRESS OF
049239,000689: 6647 31400 CA 1400 # THE BRANCH ADDRESS.
049240,000690: 6650 54117 TS POLISH
049241,000691: 6651 16616 TCF GOTO +1
049242,000692:
049243,000693: 6652 50164 CGOTO INDEX LOC # COMPUTED GO TO. PICK UP ADDRESS OF CADR
049244,000694: 6653 30001 CA 1 # LIST.
049245,000695: 6654 50116 INDEX ADDRWD # ADD MODIFIER.
049246,000696: 6655 60000 AD 0
049247,000697: 6656 54004 TS FBANK # SELECT GOTO ADDRESS.
049248,000698: 6657 75054 MASK LOW10
049249,000699: 6660 50000 INDEX A
049250,000700: 6661 32000 CA 10000
049251,000701: 6662 54117 TS POLISH
049252,000702: 6663 16616 TCF GOTO +1 # WITH ADDRESS IN A.
049253,000703:
049254,000704: 6664 30165 SWBRANCH CA BANKSET # SWITCH INSTRUCTIONS WHICH ELECT TO
049255,000705: 6665 54004 TS FBANK # BRANCH COME HERE TO DO SO.
049256,000706: 6666 50164 INDEX LOC
049257,000707: 6667 30001 CA 1
049258,000708: 6670 54117 TS POLISH
049259,000709: 6671 16616 TCF GOTO +1
049260,000710:
|
|
Page 1130
|
049262,000712: # TRIPLE PRECISION BRANCHING ROUTINE. IF CALLING TC IS AT L, RETURN IS AS FOLLOWS:
049263,000713: # L+1 IF MPAC IS GREATER THAN ZERO.
049264,000714: # L+2 IF MPAC IS EQUAL TO +0 OR -0.
049265,000715: # L+3 IF MPAC IS LESS THAN ZERO.
049266,000716:
049267,000717: 6672 10154 BRANCH CCS MPAC
049268,000718: 6673 00002 TC Q
049269,000719: 6674 16676 TCF +2 # ON ZERO.
049270,000720: 6675 16710 TCF NEG
049271,000721:
049272,000722: 6676 10155 CCS MPAC +1
049273,000723: 6677 00002 TC Q
049274,000724: 6700 16702 TCF +2
049275,000725: 6701 16710 TCF NEG
049276,000726:
049277,000727: 6702 10156 CCS MPAC +2
049278,000728: 6703 00002 TC Q
049279,000729: 6704 16706 TCF +2
049280,000730: 6705 16710 TCF NEG
049281,000731:
049282,000732: 6706 50002 Q+1 INDEX Q
049283,000733: 6707 00001 TC 1
049284,000734:
049285,000735: 6710 50002 NEG INDEX Q # IF FIRST NON-ZERO REGISTER WAS NEGATIVE.
049286,000736: 6711 00002 TC 2
049287,000737:
049288,000738: 6712 6710 Q+2 = NEG
049289,000739:
049290,000740: # ITRACE (3) REFERS TO "EXIT".
049291,000741:
049292,000742: 6712 30165 EXIT CA BANKSET # RESTORE USER'S BANK SETTING, AND LEAVE
049293,000743: 6713 54006 TS BBANK # INTERPRETIVE MODE.
049294,000744: 6714 50164 INDEX LOC
049295,000745: 6715 00001 TC 1
049296,000746:
|
|
Page 1131
|
049298,000748: # SECTION 3 - ADD/SUBTRACT PACKAGE.
049299,000749:
049300,000750: # THE FOLLOWING OPERATIONS ARE PROVIDED FOR ADDING TO AND SUBTRACTING FROM THE MULTI-PURPOSE ACCUMULATOR
049301,000751: # MPAC:
049302,000752: # 1. DAD DOUBLE PRECISION ADD.
049303,000753: # 2. DSU DOUBLE PRECISION SUBTRACT.
049304,000754: # 3. BDSU DOUBLE PRECISION SUBTRACT FROM.
049305,000755: # 4. TAD TRIPLE PRECISION ADD.
049306,000756: # 5. VAD VECTOR ADD.
049307,000757: # 6. VSU VECTOR SUBTRACT.
049308,000758: # 7. BVSU VECTOR SUBTRACT FROM.
049309,000759: # THE INTERPRETIVE OVERFLOW INDICATOR OVFIND IS SET NON-ZERO IF OVERFLOW OCCURS IN ANY OF THE ABOVE.
049310,000760:
049311,000761: 6716 35001 VSU CAF BIT15 # CHANGES 0 TO DCS.
049312,000762: 6717 16721 TCF +2
049313,000763:
049314,000764: 6720 34371 VAD CAF PRIO30 # CHANGES 0 TO DCA.
049315,000765: 6721 26116 ADS ADDRWD
049316,000766: 6722 00006 EXTEND
049317,000767: 6723 50116 INDEX ADDRWD
049318,000768: 6724 00003 READ HISCALAR # DCA 2 OR DCS 2
049319,000769: 6725 20160 DAS MPAC +3
049320,000770: 6726 00006 EXTEND # CHECK OVERFLOW.
049321,000771: 6727 16731 BZF +2
049322,000772: 6730 06763 TC OVERFLWY
049323,000773:
049324,000774: 6731 00006 EXTEND
049325,000775: 6732 50116 INDEX ADDRWD
049326,000776: 6733 00005 READ CHAN5 # DCA 4 OR DCS 4
049327,000777: 6734 20162 DAS MPAC +5
049328,000778: 6735 00006 EXTEND
049329,000779: 6736 16740 BZF +2
049330,000780: 6737 06760 TC OVERFLWZ
049331,000781:
049332,000782: 6740 00006 EXTEND
049333,000783: 6741 50116 INDEX ADDRWD
049334,000784: 6742 00001 READ LCHAN # DCA 0 OR DCS 0
049335,000785: 6743 16747 TCF ENDVXV
049336,000786:
049337,000787: 6744 00006 DAD EXTEND
049338,000788: 6745 50116 INDEX ADDRWD
049339,000789: 6746 30001 DCA 0
049340,000790: 6747 20155 ENDVXV DAS MPAC # VXV FINISHES HERE.
049341,000791: 6750 00006 EXTEND
049342,000792: 6751 16030 BZF DANZIG
049343,000793:
|
|
Page 1132
|
049345,000795: 6752 06766 SETOVF TC OVERFLOW
049346,000796: 6753 16030 TCF DANZIG
049347,000797:
|
|
Page 1133
|
049349,000799: 6754 00006 DSU EXTEND
049350,000800: 6755 50116 INDEX ADDRWD
049351,000801: 6756 40001 DCS 0
049352,000802: 6757 16747 TCF ENDVXV
049353,000803:
049354,000804: 6760 54001 OVERFLWZ TS L # ENTRY FOR THIRD COMPONENT.
049355,000805: 6761 35022 CAF FIVE
049356,000806: 6762 16765 TCF +3
049357,000807:
049358,000808: 6763 54001 OVERFLWY TS L # ENTRY FOR SECOND COMPONENT.
049359,000809: 6764 36214 CAF THREE
049360,000810: 6765 56001 XCH L
049361,000811:
049362,000812: 6766 50000 OVERFLOW INDEX A # ENTRY FOR 1ST COMP OR DP (L=0).
049363,000813: 6767 45000 CS LIMITS # PICK UP POSMAX OR NEGMAX.
049364,000814: 6770 54130 TS BUF
049365,000815: 6771 00006 EXTEND
049366,000816: 6772 24000 AUG A # FORCE OVERFLOW.
049367,000817: 6773 50001 INDEX L
049368,000818: 6774 26155 ADS MPAC +1
049369,000819: 6775 54007 TS 7
049370,000820: 6776 35021 CAF ZERO
049371,000821: 6777 60130 AD BUF
049372,000822: 7000 50001 INDEX L
049373,000823: 7001 26154 ADS MPAC
049374,000824: 7002 54007 TS 7
049375,000825: 7003 00002 TC Q # NO OVERFLOW EXIT.
049376,000826: 7004 17121 TCF SETOVF2 # SET OVFIND AND EXIT.
049377,000827:
049378,000828: 7005 00006 BVSU EXTEND
049379,000829: 7006 50116 INDEX ADDRWD
049380,000830: 7007 30003 DCA 2
049381,000831: 7010 52160 DXCH MPAC +3
049382,000832: 7011 00006 EXTEND
049383,000833: 7012 40001 DCOM
049384,000834: 7013 20160 DAS MPAC +3
049385,000835: 7014 00006 EXTEND
049386,000836: 7015 17017 BZF +2
049387,000837: 7016 06763 TC OVERFLWY
049388,000838:
049389,000839: 7017 00006 EXTEND
049390,000840: 7020 50116 INDEX ADDRWD
049391,000841: 7021 30005 DCA 4
049392,000842: 7022 52162 DXCH MPAC +5
049393,000843: 7023 00006 EXTEND
049394,000844: 7024 40001 DCOM
049395,000845: 7025 20162 DAS MPAC +5
049396,000846: 7026 00006 EXTEND
049397,000847: 7027 17031 BZF +2
049398,000848: 7030 06760 TC OVERFLWZ
049399,000849:
|
|
Page 1134
|
049401,000851: 7031 00006 BDSU EXTEND
049402,000852: 7032 50116 INDEX ADDRWD
049403,000853: 7033 30001 DCA 0
049404,000854: 7034 52155 DXCH MPAC
049405,000855: 7035 00006 EXTEND
049406,000856: 7036 40001 DCOM
049407,000857: 7037 16747 TCF ENDVXV
049408,000858:
|
|
Page 1135
|
049410,000860: # TRIPLE PRECISION ADD ROUTINE.
049411,000861:
049412,000862: 7040 00006 TAD EXTEND
049413,000863: 7041 50116 INDEX ADDRWD
049414,000864: 7042 30002 DCA 1 # ADD MINOR PARTS FIRST.
049415,000865: 7043 20156 DAS MPAC +1
049416,000866: 7044 50116 INDEX ADDRWD
049417,000867: 7045 60000 AD 0
049418,000868: 7046 60154 AD MPAC
049419,000869: 7047 54154 TS MPAC
049420,000870: 7050 16030 TCF DANZIG
049421,000871:
049422,000872: 7051 16752 TCF SETOVF # SET OVFIND IF SUCH OCCURS.
049423,000873:
|
|
Page 1136
|
049425,000875: # ARITHMETIC SUBROUTINES REQUIRED IN FIXED-FIXED.
049426,000876: # 1. DMPSUB DOUBLE PRECISION MULTIPLY. MULTIPLY THE CONTENTS OF MPAC,+1 BY THE DP WORD WHOSE ADDRESS
049427,000877: # IS IN ADDRWD AND LEAVE A TRIPLE PRECISION RESULT IN MPAC.
049428,000878: # 2. ROUNDSUB ROUND THE TRIPLE PRECISON CONTENTS OF MPAC TO DOUBLE PRECISION.
049429,000879: # 3. DOTSUB TAKE THE DOT PRODUCT OF THE VECTOR IN MPAC AND THE VECTOR WHOSE ADDRESS IS IN ADDRWD
049430,000880: # AND LEAVE THE TRIPLE PRECISION RESULT IN MPAC.
049431,000881: # 4. POLY USING THE CONTENTS OF MPAC AS A DP ARGUMENT, EVALUATE THE POLYNOMIAL WHOSE DEGREE AND
049432,000882: # COEFFICIENTS IMMEDIATELY FOLLOW THE TC POLY INSTRUCTION (SEE ROUTINE FOR DETAILS.)
049433,000883:
049434,000884: 7052 50002 DMP INDEX Q # BASIC SUBROUTINE FOR USE BY PINBALL, ETC
049435,000885: 7053 30000 CAF 0 # ADRES OF ARGUMENT FOLLOWS TC DMP .
049436,000886: 7054 24002 INCR Q
049437,000887: 7055 54116 -1 TS ADDRWD # (PROLOGUE FOR SETTING ADDRWD.)
049438,000888:
049439,000889: 7056 50116 DMPSUB INDEX ADDRWD # GET MINOR PART OF OPERAND AT C(ADDRWD).
049440,000890: 7057 30001 CA 1
049441,000891: 7060 54156 TS MPAC +2 # THIS WORKS FOR SQUARING MPAC AS WELL.
049442,000892: 7061 35021 CAF ZERO # SET MPAC +1 TO ZERO SO WE CAN ACCUMULATE
049443,000893: 7062 56155 XCH MPAC +1 # THE PARTIAL PRODUCTS WITH DAS
049444,000894: 7063 54135 TS MPTEMP # INSTRUCTIONS.
049445,000895: 7064 00006 EXTEND
049446,000896: 7065 70156 MP MPAC +2 # MINOR OF MPAC X MINOR OF C(ADDRWD).
049447,000897:
049448,000898: 7066 56156 XCH MPAC +2 # DISCARD MINOR PART OF ABOVE RESULT AND
049449,000899: 7067 00006 EXTEND # FORM MAJOR OF MPAC X MINOR OF C(ADDRWD).
049450,000900: 7070 70154 MP MPAC
049451,000901: 7071 20156 DAS MPAC +1 # GUARANTEED NO OVERFLOW.
049452,000902:
049453,000903: 7072 50116 INDEX ADDRWD # GET MAJOR PART OF ARGUMENT AT C(ADDRWD).
049454,000904: 7073 30000 CA 0
049455,000905: 7074 56135 XCH MPTEMP # SAVE AND BRING OUT MINOR OF MPAC.
049456,000906: 7075 00006 DMPSUB2 EXTEND
049457,000907: 7076 70135 MP MPTEMP # MAJOR OF C(ADDRWD) X MINOR OF MPAC.
049458,000908: 7077 20156 DAS MPAC +1 # ACCUMULATE, SETTING A TO NET OVERFLOW.
049459,000909:
049460,000910: 7100 56154 XCH MPAC # SETTING MPAC TO 0 OR +-1.
049461,000911: 7101 00006 EXTEND
049462,000912: 7102 70135 MP MPTEMP # MAJOR OF MPAC X MAJOR OF C(ADDRWD).
049463,000913: 7103 20155 DAS MPAC # GUARANTEED NO OVERFLOW.
049464,000914: 7104 00002 TC Q # 49 MCT = .573 MS. INCLUDING RETURN.
049465,000915:
|
|
Page 1137
|
049467,000917: # ROUND MPAC TO DOUBLE PRECISION, SETTING OVFIND ON THE RARE EVENT OF OVERFLOW.
049468,000918:
049469,000919: 7105 35021 ROUNDSUB CAF ZERO # SET MPAC +2 = 0 FOR SCALARS AND CHANGE
049470,000920: 7106 54163 +1 TS MODE # MODE TO DP.
049471,000921:
049472,000922: 7107 56156 VROUND XCH MPAC +2 # BUT WE NEEDNT TAKE THE TIME FOR VECTORS.
049473,000923: 7110 60000 DOUBLE
049474,000924: 7111 54001 TS L
049475,000925: 7112 00002 TC Q
049476,000926:
049477,000927: 7113 60155 AD MPAC +1 # ADD ROUNDING BIT IF MPAC +2 WAS GREATER
049478,000928: 7114 54155 TS MPAC +1 # THAN .5 IN MAGNITUDE.
049479,000929: 7115 00002 TC Q
049480,000930:
049481,000931: 7116 60154 AD MPAC # PROPAGATE INTERFLOW.
049482,000932: 7117 54154 TS MPAC
049483,000933: 7120 00002 TC Q
049484,000934:
049485,000935: 7121 54121 SETOVF2 TS OVFIND # (RARE).
049486,000936: 7122 00002 TC Q
049487,000937:
|
|
Page 1138
|
049489,000939: # THE DOT PRODUCT SUBROUTINE USUALLY FORMS THE DOT PRODUCT OF THE VECTOR IN MPAC WITH A STANDARD SIX
049490,000940: # REGISTER VECTOR WHOSE ADDRESS IS IN ADDRWD. IN THIS CASE C(DOTINC) ARE SET TO 2. VXM, HOWEVER, SETS C(DOTINC) TO
049491,000941: # 6 SO THAT DOTSUB DOTS MPAC WITH A COLUMN VECTOR OF THE MATRIX IN QUESTION IN THIS CASE.
049492,000942:
049493,000943: 7123 35016 PREDOT CAF TWO # PROLOGUE TO SET DOTINC TO 2.
049494,000944: 7124 54136 TS DOTINC
049495,000945:
049496,000946: 7125 00006 DOTSUB EXTEND
049497,000947: 7126 22137 QXCH DOTRET # SAVE RETURN.
049498,000948: 7127 07056 TC DMPSUB # DOT X COMPONENTS.
049499,000949: 7130 52160 DXCH MPAC +3 # POSITION Y COMPONENT OF MPAC FOR
049500,000950: 7131 52155 DXCH MPAC # MULTIPLICATION WHILE SAVING RESULT IN
049501,000951: 7132 52131 DXCH BUF # THREE WORD BUFFER, BUF.
049502,000952: 7133 30156 CA MPAC +2
049503,000953: 7134 54132 TS BUF +2
049504,000954:
049505,000955: 7135 30136 CA DOTINC # ADVANCE ADDRWD TO Y COMPONENT OF
049506,000956: 7136 26116 ADS ADDRWD # OTHER ARGUMENT.
049507,000957: 7137 07056 TC DMPSUB
049508,000958: 7140 52156 DXCH MPAC +1 # ACCUMULATE PARTIAL PRODUCTS.
049509,000959: 7141 20132 DAS BUF +1
049510,000960: 7142 60154 AD MPAC
049511,000961: 7143 60130 AD BUF
049512,000962: 7144 54130 TS BUF
049513,000963: 7145 17147 TCF +2
049514,000964: 7146 54121 TS OVFIND # IF OVERFLOW OCCURS.
049515,000965:
049516,000966: 7147 52162 DXCH MPAC +5 # MULTIPLY Z COMPONENTS.
049517,000967: 7150 52155 DXCH MPAC
049518,000968: 7151 30136 CA DOTINC
049519,000969: 7152 26116 ADS ADDRWD
049520,000970: 7153 07056 TC DMPSUB
049521,000971: 7154 52132 ENDDOT DXCH BUF +1 # LEAVE FINAL ACCUMULATION IN MPAC.
049522,000972: 7155 20156 DAS MPAC +1
049523,000973: 7156 60154 AD MPAC
049524,000974: 7157 60130 AD BUF
049525,000975: 7160 54154 TS MPAC
049526,000976: 7161 00137 TC DOTRET
049527,000977:
049528,000978: 7162 06766 TC OVERFLOW # ON OVERFLOW HERE.
049529,000979: 7163 00137 TC DOTRET
049530,000980:
|
|
Page 1139
|
049532,000982: # DOUBLE PRECISION POLYNOMIAL EVALUATOR
049533,000983: # N N-1
049534,000984: # THIS ROUTINE EVALUATES A X + A X + ... + A X + A LEAVING THE DP RESULT IN MPAC ON EXIT.
049535,000985: # N N-1 1 0
049536,000986:
049537,000987: # THE ROUTINE HAS TWO ENTRIES
049538,000988:
049539,000989: # 1. ENTRY THRU POWRSERS. THE COEFFICIENTS MAY BE EITHER IN FIXED OR ERASABLE, THE CALL IS BY
049540,000990: # TC POWRSERS, AND THE RETURN IS TO LOC(TC POWRSERS)+1. THE ENTERING DATA MUST BE AS FOLLOWS
049541,000991:
049542,000992: # A SP LOC-3 ADDRESS FOR REFERENCING COEF TABLE
049543,000993: # L SP N-1 N IS THE DEGREE OF THE POWER SERIES
049544,000994: # MPAC DP X ARGUMENT
049545,000995: # LOC-2N DP A(0)
049546,000996: # ...
049547,000997: # LOC DP A(N)
049548,000998:
049549,000999: # 2. ENTRY THRU POLY. THE CALL TO POLY AND THE ENTERING DATA MUST BE AS FOLLOWS
049550,001000:
049551,001001: # MPAC DP X ARGUMENT
049552,001002: # LOC TC POLY
049553,001003: # LOC+1 SP N-1
049554,001004: # LOC+2 DP A(0)
049555,001005: # ...
049556,001006: # LOC+2N+2 DP A(N) RETURN IS TO LOC+2N+4
049557,001007:
049558,001008: 7164 00006 POWRSERS EXTEND
049559,001009: 7165 22141 QXCH POLYRET # RETURN ADDRESS
049560,001010: 7166 54117 TS POLISH # POWER SERIES ADDRESS
049561,001011: 7167 22140 LXCH POLYCNT # N-1 TO COUNTER
049562,001012: 7170 17201 TCF POLYCOM # SKIP SET UP BY POLY
049563,001013:
049564,001014: 7171 50002 POLY INDEX Q
049565,001015: 7172 30000 CAF 0
049566,001016: 7173 54140 TS POLYCNT # N-1 TO COUNTER
049567,001017: 7174 60000 DOUBLE
049568,001018: 7175 60002 AD Q
049569,001019: 7176 54117 TS POLISH # L(A(N))-3 TO POLISH
049570,001020: 7177 65022 AD FIVE
049571,001021: 7200 54141 TS POLYRET # STORE RETURN ADDRESS
049572,001022:
049573,001023: 7201 36242 POLYCOM CAF LVBUF # INCOMING X WILL BE MOVED TO VBUF, SO
049574,001024: 7202 54116 TS ADDRWD # SET ADDRWD SO DMPSUB WILL MPY BY VBUF.
049575,001025:
049576,001026: 7203 00006 EXTEND
049577,001027: 7204 50117 INDEX POLISH
049578,001028: 7205 30004 DCA 3
049579,001029:
|
|
Page 1140
|
049581,001031: 7206 52155 DXCH MPAC # LOAD A(N) INTO MPAC,
049582,001032: 7207 52123 DXCH VBUF # SAVING X IN VBUF
049583,001033: 7210 17214 TCF POLY2
049584,001034:
049585,001035: 7211 54140 POLYLOOP TS POLYCNT # SAVE DECREMENTED LOOP COUNTER
049586,001036: 7212 45016 CS TWO
049587,001037: 7213 26117 ADS POLISH # REGRESS COEFFICIENT POINTER
049588,001038:
049589,001039: 7214 07056 POLY2 TC DMPSUB # MULTIPLY BY X
049590,001040: 7215 00006 EXTEND
049591,001041: 7216 50117 INDEX POLISH
049592,001042: 7217 30002 DCA 1 # ADD IN NEXT COEFFICIENT
049593,001043: 7220 20155 DAS MPAC # USERS RESPONSIBILITY TO ASSURE NO OVFLOW
049594,001044:
049595,001045: 7221 10140 CCS POLYCNT
049596,001046: 7222 17211 TCF POLYLOOP
049597,001047: 7223 00141 TC POLYRET # RETURN CALLER
049598,001048:
|
|
Page 1141
|
049600,001050: # MISCELLANEOUS MULTI-PRECISION ROUTINES REQUIRED IN FIXED-FIXED BUT NOT USED BY THE INTERPRETER.
049601,001051:
049602,001052: 7224 35021 DPAGREE CAF ZERO # DOUBLE PRECISION ENTRY -
049603,001053: 7225 54156 TS MPAC +2 # ZERO LOW-ORDER WORD
049604,001054:
049605,001055: 7226 22002 TPAGREE LXCH Q # FORCE SIGN AGREEMENT AMONG THE TRIPLE
049606,001056: 7227 06672 TC BRANCH # PRECISION CONTENTS OF MPAC. RETURNING
049607,001057: 7230 17234 TCF ARG+ # WITH SIGNUM OF THE INPUT IN A.
049608,001058: 7231 17254 TCF ARGZERO
049609,001059:
049610,001060: 7232 44777 CS POSMAX # IF NEGATIVE.
049611,001061: 7233 17235 TCF +2
049612,001062:
049613,001063: 7234 34777 ARG+ CAF POSMAX
049614,001064: 7235 54002 TS Q
049615,001065: 7236 00006 EXTEND
049616,001066: 7237 24000 AUG A # FORMS +-1.0.
049617,001067: 7240 60156 AD MPAC +2
049618,001068: 7241 54156 TS MPAC +2
049619,001069: 7242 35021 CAF ZERO
049620,001070: 7243 60002 AD Q
049621,001071: 7244 60155 AD MPAC +1
049622,001072: 7245 54155 TS MPAC +1
049623,001073: 7246 35021 CAF ZERO
049624,001074: 7247 60002 AD Q # Q STILL HAS POSMAX OR NEGMAX IN IT.
049625,001075: 7250 60154 AD MPAC
049626,001076: 7251 54154 ARGZERO2 TS MPAC # ALWAYS SKIPPING UNLESS ARGZERO.
049627,001077: 7252 54155 TS MPAC +1
049628,001078: 7253 00001 TC L # RETURN VIA L.
049629,001079:
049630,001080: 7254 54156 ARGZERO TS MPAC +2 # SET ALL THREE MPAC REGISTERS TO ZERO.
049631,001081: 7255 17251 TCF ARGZERO2
049632,001082:
049633,001083: # SHORTMP MULTIPLIES THE TP CONTENTS OF MPAC BY THE SINGLE PRECISION NUMBER ARRIVING IN A.
049634,001084:
049635,001085: 7256 54135 SHORTMP TS MPTEMP
049636,001086: 7257 00006 EXTEND
049637,001087: 7260 70156 MP MPAC +2
049638,001088: 7261 54156 TS MPAC +2
049639,001089: 7262 35021 SHORTMP2 CAF ZERO # SO SUBSEQUENT DAS WILL WORK.
049640,001090: 7263 56155 XCH MPAC +1
049641,001091: 7264 17075 TCF DMPSUB2
049642,001092:
|
|
Page 1142
|
049644,001094: # DMPNSUB MULTIPLIES THE DP FRACTION ARRIVING IN MPAC BY THE SP
049645,001095: # INTEGER ARRIVING IN A. THE DP PRODUCT DEPARTS BOTH IN MPAC AND IN
049646,001096: # A AND L. NOTE THAT DMPNSUB NORMALLY INCREASES THE MAGNITUDE OF THE
049647,001097: # CONTENTS OF MPAC. THE CUSTOMER MUST INSURE THAT B(A) X B(MPAC,MPAC+1)
049648,001098: # AND B(A) X B(MPAC) ARE LESS THAN 1 IN MAGNITUDE, WHERE B, AS IS OBVIOUS,
049649,001099: # INDICATES THE ARRIVING CONTENTS.
049650,001100:
049651,001101: 7265 54135 DMPNSUB TS DMPNTEMP
049652,001102: 7266 00006 EXTEND
049653,001103: 7267 70155 MP MPAC +1
049654,001104: 7270 52155 DXCH MPAC # LOW PRODUCT TO MPAC, HIGH FACTOR TO A
049655,001105: 7271 00006 EXTEND
049656,001106: 7272 70135 MP DMPNTEMP
049657,001107: 7273 30001 CA L
049658,001108: 7274 26154 ADS MPAC # COMPLETING THE PRODUCT IN MPAC
049659,001109: 7275 00006 EXTEND
049660,001110: 7276 30155 DCA MPAC # BRINGING THE PRODUCT INTO A AND L
049661,001111: 7277 00002 TC Q
049662,001112:
|
|
Page 1143
|
049664,001114: # MISCELLANEOUS VECTOR OPERATIONS. INCLUDED HERE ARE THE FOLLOWING:
049665,001115: # 1. DOT DP VECTOR DOT PRODUCT.
049666,001116: # 2. VXV DP VECTOR CROSS PRODUCT.
049667,001117: # 3. VXSC DP VECTOR TIMES SCALAR.
049668,001118: # 4. V/SC DP VECTOR DIVIDED BY SCALAR.
049669,001119: # 5. VPROJ DP VECTOR PROJECTION. ( (MPAC.X)MPAC ).
049670,001120: # 6. VXM DP VECTOR POST-MULTIPLIED BY MATRIX.
049671,001121: # 7. MXV DP VECTOR PRE-MULTIPLIED BY MATRIX.
049672,001122:
049673,001123: 7300 07123 DOT TC PREDOT # DO THE DOT PRODUCT AND EXIT, CHANGING
049674,001124: 7301 35021 DMODE CAF ZERO # THE MODE TO DP SCALAR.
049675,001125: 7302 16027 TCF NEWMODE
049676,001126:
049677,001127: 7303 35016 MXV CAF TWO # SET UP MATINC AND DOTINC FOR ROW
049678,001128: 7304 54140 TS MATINC # VECTORS.
049679,001129: 7305 17311 TCF VXM/MXV # GO TO COMMON PORTION.
049680,001130:
049681,001131: 7306 44377 VXM CS TEN # SET MATINC AND DOTINC TO REFER TO MATRIX
049682,001132: 7307 54140 TS MATINC # AS THREE COLUMN VECTORS.
049683,001133: 7310 36211 CAF SIX
049684,001134:
|
|
Page 1144
|
049686,001136: # COMMON PORTION OF MXV AND VXM.
049687,001137:
049688,001138: 7311 54136 VXM/MXV TS DOTINC
049689,001139:
049690,001140: # ITRACE (2) REFERS TO "VXM/MXV".
049691,001141:
049692,001142: 7312 07501 TC MPACVBUF # SAVE VECTOR IN MPAC FOR FURTHER USE.
049693,001143:
049694,001144: 7313 07125 TC DOTSUB # GO DOT TO GET X COMPONENT OF ANSWER.
049695,001145: 7314 00006 EXTEND
049696,001146: 7315 30123 DCA VBUF # MOVE MPAC VECTOR BACK INTO MPAC, SAVING
049697,001147: 7316 52155 DXCH MPAC # NEW X COMPONENT IN BUF2.
049698,001148: 7317 52134 DXCH BUF2
049699,001149: 7320 00006 EXTEND
049700,001150: 7321 30125 DCA VBUF +2
049701,001151: 7322 52160 DXCH MPAC +3
049702,001152: 7323 00006 EXTEND
049703,001153: 7324 30127 DCA VBUF +4
049704,001154: 7325 52162 DXCH MPAC +5
049705,001155: 7326 30140 CA MATINC # INITIALIZE ADDRWD FOR NEXT DOT PRODUCT.
049706,001156: 7327 26116 ADS ADDRWD # FORMS BASE ADDRESS OF NEXT COLUMN(ROW).
049707,001157:
049708,001158: 7330 07125 TC DOTSUB
049709,001159: 7331 52123 DXCH VBUF # MOVE GIVEN VECTOR BACK TO MPAC, SAVING Y
049710,001160: 7332 52155 DXCH MPAC # COMPONENT OF ANSWER IN VBUF +2.
049711,001161: 7333 52125 DXCH VBUF +2
049712,001162: 7334 52160 DXCH MPAC +3
049713,001163: 7335 52127 DXCH VBUF +4
049714,001164: 7336 52162 DXCH MPAC +5
049715,001165: 7337 30140 CA MATINC # FORM ADDRESS OF LAST COLUMN OR ROW.
049716,001166: 7340 26116 ADS ADDRWD
049717,001167:
049718,001168: 7341 07125 TC DOTSUB
049719,001169: 7342 52134 DXCH BUF2 # ANSWER NOW COMPLETE. PUT COMPONENTS INTO
049720,001170: 7343 52155 DXCH MPAC # PROPER MPAC REGISTERS.
049721,001171: 7344 52162 DXCH MPAC +5
049722,001172: 7345 52125 DXCH VBUF +2
049723,001173: 7346 52160 DXCH MPAC +3
049724,001174: 7347 16030 TCF DANZIG # EXIT.
049725,001175:
|
|
Page 1145
|
049727,001177: # VXSC - VECTOR TIMES SCALAR.
049728,001178:
049729,001179: 7350 10163 VXSC CCS MODE # TEST PRESENT MODE.
049730,001180: 7351 17377 TCF DVXSC # SEPARATE ROUTINE WHEN SCALAR IS IN MPAC.
049731,001181: 7352 17377 TCF DVXSC
049732,001182:
049733,001183: 7353 07056 VVXSC TC DMPSUB # COMPUTE X COMPONENT
049734,001184: 7354 07107 TC VROUND # AND ROUND IT.
049735,001185: 7355 52160 DXCH MPAC +3 # PUT Y COMPONENT INTO MPAC SAVING MPAC IN
049736,001186: 7356 52155 DXCH MPAC # MPAC +3.
049737,001187: 7357 52160 DXCH MPAC +3
049738,001188:
049739,001189: 7360 07056 TC DMPSUB # DO SAME FOR Y AND Z COMPONENTS.
049740,001190: 7361 07107 TC VROUND
049741,001191: 7362 52162 DXCH MPAC +5
049742,001192: 7363 52155 DXCH MPAC
049743,001193: 7364 52162 DXCH MPAC +5
049744,001194:
049745,001195: 7365 07056 TC DMPSUB
049746,001196: 7366 07107 TC VROUND
049747,001197: 7367 52155 VROTATEX DXCH MPAC # EXIT USED TO RESTORE MPAC AFTER THIS
049748,001198: 7370 52162 DXCH MPAC +5 # TYPE OF ROTATION. CALLED BY VECTOR SHIFT
049749,001199: 7371 52160 DXCH MPAC +3 # RIGHT, V/SC, ETC.
049750,001200: 7372 52155 DXCH MPAC
049751,001201: 7373 16030 TCF DANZIG
049752,001202:
|
|
Page 1146
|
049754,001204: # DP VECTOR PROJECTION ROUTINE.
049755,001205:
049756,001206: 7374 07123 VPROJ TC PREDOT # (MPAC.X)MPAC IS COMPUTED AND LEFT IN
049757,001207: 7375 45015 CS FOUR # MPAC. DO DOT AND FALL INTO DVXSC.
049758,001208: 7376 26116 ADS ADDRWD
049759,001209:
049760,001210: # VXSC WHEN SCALAR ARRIVES IN MPAC AND VECTOR IS AT X.
049761,001211:
049762,001212: 7377 00006 DVXSC EXTEND # SAVE SCALAR IN MPAC +3 AND GET X
049763,001213: 7400 30155 DCA MPAC # COMPONENT OF ANSWER.
049764,001214: 7401 52160 DXCH MPAC +3
049765,001215: 7402 07056 TC DMPSUB
049766,001216: 7403 07107 TC VROUND
049767,001217:
049768,001218: 7404 35016 CAF TWO # ADVANCE ADDRWD TO Y COMPONENT OF X.
049769,001219: 7405 26116 ADS ADDRWD
049770,001220: 7406 00006 EXTEND
049771,001221: 7407 30160 DCA MPAC +3 # PUT SCALAR BACK INTO MPAC AND SAVE
049772,001222: 7410 52155 DXCH MPAC # X RESULT IN MPAC +5.
049773,001223: 7411 52162 DXCH MPAC +5
049774,001224: 7412 07056 TC DMPSUB
049775,001225: 7413 07107 TC VROUND
049776,001226:
049777,001227: 7414 35016 CAF TWO
049778,001228: 7415 26116 ADS ADDRWD # TO Z COMPONENT.
049779,001229: 7416 52160 DXCH MPAC +3 # BRING SCALAR BACK, PUTTING Y RESULT IN
049780,001230: 7417 52155 DXCH MPAC # THE PROPER PLACE.
049781,001231: 7420 52160 DXCH MPAC +3
049782,001232: 7421 07056 TC DMPSUB
049783,001233: 7422 07107 TC VROUND
049784,001234:
049785,001235: 7423 52155 DXCH MPAC # PUT Z COMPONENT IN PROPER PLACE, ALSO
049786,001236: 7424 52162 DXCH MPAC +5 # POSITIONING X.
049787,001237: 7425 52155 DXCH MPAC
049788,001238:
049789,001239: 7426 16470 TCF VMODE # MODE HAS CHANGED TO VECTOR.
049790,001240:
|
|
Page 1147
|
049792,001242: # THE VECTOR CROSS PRODUCT ROUTINE CALCULATES (X M -X M ,X M -X M ,X M -X M ) WHERE M IS THE VECTOR IN
049793,001243: # 3 2 2 3 1 3 3 1 2 1 1 2
049794,001244: # MPAC AND X THE VECTOR AT THE GIVEN ADDRESS.
049795,001245:
049796,001246: 7427 00006 VXV EXTEND
049797,001247: 7430 30162 DCA MPAC +5 # FORM UP M3X1, LEAVING M1 IN VBUF.
049798,001248: 7431 52155 DXCH MPAC
049799,001249: 7432 52123 DXCH VBUF
049800,001250: 7433 07056 TC DMPSUB # BY X1.
049801,001251:
049802,001252: 7434 00006 EXTEND
049803,001253: 7435 40160 DCS MPAC +3 # CALCULATE -X1M2, SAVING X1M3 IN VBUF +2.
049804,001254: 7436 52155 DXCH MPAC
049805,001255: 7437 52125 DXCH VBUF +2
049806,001256: 7440 07056 TC DMPSUB
049807,001257:
049808,001258: 7441 35016 CAF TWO # ADVANCE ADDRWD TO X2.
049809,001259: 7442 26116 ADS ADDRWD
049810,001260: 7443 00006 EXTEND
049811,001261: 7444 40162 DCS MPAC +5 # PREPARE TO GET -X2M3, SAVING -X1M2 IN
049812,001262: 7445 52155 DXCH MPAC # MPAC +5.
049813,001263: 7446 52162 DXCH MPAC +5
049814,001264: 7447 07056 TC DMPSUB
049815,001265:
049816,001266: 7450 00006 EXTEND
049817,001267: 7451 30123 DCA VBUF # GET X2M1, SAVING -X2M3 IN VBUF +4.
049818,001268: 7452 52155 DXCH MPAC
049819,001269: 7453 52127 DXCH VBUF +4
049820,001270: 7454 07056 TC DMPSUB
049821,001271:
049822,001272: 7455 35016 CAF TWO # ADVANCE ADDRWD TO X3.
049823,001273: 7456 26116 ADS ADDRWD
049824,001274: 7457 00006 EXTEND
049825,001275: 7460 40123 DCS VBUF # GET -X3M1, ADDING X2M1 TO MPAC +5 TO
049826,001276: 7461 52155 DXCH MPAC # COMPLETE THE Z COMPONENT OF THE ANSWER.
049827,001277: 7462 20162 DAS MPAC +5
049828,001278:
049829,001279: 7463 00006 EXTEND
049830,001280: 7464 17466 BZF +2
049831,001281: 7465 06760 TC OVERFLWZ
049832,001282:
049833,001283: 7466 07056 TC DMPSUB
049834,001284: 7467 52125 DXCH VBUF +2 # MOVE X1M3 TO MPAC +3 SETTING UP FOR X3M2
049835,001285: 7470 52160 DXCH MPAC +3 # AND ADD -X3M1 TO MPAC +3 TO COMPLETE THE
049836,001286: 7471 52155 DXCH MPAC # Y COMPONENT OF THE RESULT.
049837,001287: 7472 20160 DAS MPAC +3
049838,001288:
049839,001289: 7473 00006 EXTEND
049840,001290: 7474 17476 BZF +2
|
|
Page 1148
|
049842,001292: 7475 06763 TC OVERFLWY
049843,001293:
049844,001294: 7476 07056 TC DMPSUB
049845,001295: 7477 52127 DXCH VBUF +4 # GO ADD -X2M3 TO X3M2 TO COMPLETE THE X
049846,001296: 7500 16747 TCF ENDVXV # COMPONENT (TAIL END OF DAD).
049847,001297:
049848,001298: # THE MPACVBUF SUBROUTINE SAVES THE VECTOR IN MPAC IN VBUF WITHOUT CLOBBERING MPAC.
049849,001299:
049850,001300: 7501 00006 MPACVBUF EXTEND # CALLED BY MXV, VXM, AND UNIT.
049851,001301: 7502 30155 DCA MPAC
049852,001302: 7503 52123 DXCH VBUF
049853,001303: 7504 00006 EXTEND
049854,001304: 7505 30160 DCA MPAC +3
049855,001305: 7506 52125 DXCH VBUF +2
049856,001306: 7507 00006 EXTEND
049857,001307: 7510 30162 DCA MPAC +5
049858,001308: 7511 52127 DXCH VBUF +4
049859,001309: 7512 00002 TC Q # RETURN TO CALLER.
049860,001310:
049861,001311: # DOUBLE PRECISION SIGN AGREE ROUTINE. ARRIVE WITH INPUT IN A+L. OUTPUT IS IN A + L.
049862,001312:
049863,001313: 7513 10000 ALSIGNAG CCS A # TEST UPPER PART.
049864,001314: 7514 17520 TCF UPPOS # IT IS POSITIVE
049865,001315: 7515 00002 TC Q # ZERO
049866,001316: 7516 17530 TCF UPNEG # NEGATIVE
049867,001317: 7517 00002 TC Q # ZERO
049868,001318:
049869,001319: 7520 56001 UPPOS XCH L # SAVE DECREMENTED UPPER PART.
049870,001320: 7521 65002 AD HALF
049871,001321: 7522 65002 AD HALF
049872,001322: 7523 54000 TS A # SKIPS ON OVERFLOW
049873,001323: 7524 17526 TCF +2
049874,001324: 7525 24001 INCR L # RESTORE UPPER TO ROIGINAL VALUE
049875,001325: 7526 56001 XCH L # SWAP A + L BACK.
049876,001326: 7527 00002 TC Q
049877,001327:
049878,001328: 7530 56001 UPNEG XCH L # SAVE COMPLEMENTED + DECREMENTED UPPER PT
049879,001329: 7531 65001 AD NEGMAX
049880,001330: 7532 67716 AD NEGONE
049881,001331: 7533 54000 TS A
049882,001332: 7534 17536 TCF +2 # DONT INCREMENT IF NO OVERFLOW.
049883,001333: 7535 24001 INCR L
049884,001334: 7536 56001 XCH L
049885,001335: 7537 40000 COM # MAKE NEGATIVE AGAIN.
049886,001336: 7540 00002 TC Q
049887,001337:
|
|
Page 1149
|
049889,001339: # INTERPRETIVE INSTRUCTIONS WHOSE EXECUTION CONSISTS OF PRINCIPALLY CALLING SUBROUTINES.
049890,001340:
049891,001341: 7541 07056 DMP1 TC DMPSUB # DMP INSTRUCTION.
049892,001342: 7542 16030 TCF DANZIG
049893,001343:
049894,001344: 7543 07056 DMPR TC DMPSUB
049895,001345: 7544 07106 TC ROUNDSUB +1 # (C(A) = +0).
049896,001346: 7545 16030 TCF DANZIG
049897,001347:
049898,001348: 7546 00006 DDV EXTEND
049899,001349: 7547 50116 INDEX ADDRWD # MOVE DIVIDEND INTO BUF.
049900,001350: 7550 30001 DCA 0
049901,001351: 7551 17556 TCF BDDV +4
049902,001352:
049903,001353: 7552 00006 BDDV EXTEND # MOVE DIVISOR INTO MPAC SAVING MPAC, THE
049904,001354: 7553 50116 INDEX ADDRWD # DIVIDEND, IN BUF.
049905,001355: 7554 30001 DCA 0
049906,001356: 7555 52155 DXCH MPAC
049907,001357: 7556 52131 +4 DXCH BUF
049908,001358: 7557 35021 CAF ZERO # DIVIDE ROUTINES IN BANK 0.
049909,001359: 7560 54004 TS FBANK
049910,001360: 7561 12353 TCF DDV/BDDV
049911,001361:
049912,001362: 7562 30116 SETPD CA ADDRWD # MUST SET TO WORK AREA, OR EBANK TROUBLE.
049913,001363: 7563 54166 TS PUSHLOC
049914,001364: 7564 16032 TCF NOIBNKSW # NO FBANK SWITCH REQUIRED.
049915,001365:
049916,001366: 7565 35021 TSLC CAF ZERO # SHIFTING ROUTINES LOCATED IN BANK 00.
049917,001367: 7566 54004 TS FBANK
049918,001368: 7567 12172 TCF TSLC2
049919,001369:
049920,001370: 7570 36043 GSHIFT CAF LOW7 # USED AS MASK AT GENSHIFT. THIS PROCESSES
049921,001371: 7571 54004 TS FBANK # ANY SHIFT INSTRUCTION (EXCEPT TSLC) WITH
049922,001372: 7572 12214 TCF GENSHIFT # AN ADDRESS (ROUTINES IN BANK 0).
049923,001373:
|
|
Page 1150
|
049925,001375: # THE FOLLOWING IS THE PROLOGUE TO V/SC. IF THE PRESENT MODE IS VECTOR, IT SAVES THE SCALAR AT X IN BUF
049926,001376: # AND CALLS THE V/SC ROUTINE IN BANK 0. IF THE PRESENT MODE IS SCALAR, IT MOVES THE VECTOR AT X INTO MPAC, SAVING
049927,001377: # THE SCALAR IN MPAC IN BUF BEFORE CALLING THE V/SC ROUTINE IN BANK 0.
049928,001378:
049929,001379: 7573 10163 V/SC CCS MODE
049930,001380: 7574 17605 TCF DV/SC # MOVE VECTOR INTO MPAC.
049931,001381: 7575 17605 TCF DV/SC
049932,001382:
049933,001383: 7576 00006 VV/SC EXTEND
049934,001384: 7577 50116 INDEX ADDRWD
049935,001385: 7600 30001 DCA 0
049936,001386: 7601 52131 V/SC1 DXCH BUF # IN BOTH CASES, VECTOR IS NOW IN MPAC AND
049937,001387: 7602 35021 CAF ZERO # SCALAR IN BUF.
049938,001388: 7603 54004 TS FBANK
049939,001389: 7604 12654 TCF V/SC2
049940,001390:
049941,001391: 7605 00006 DV/SC EXTEND
049942,001392: 7606 50116 INDEX ADDRWD
049943,001393: 7607 30003 DCA 2
049944,001394: 7610 52160 DXCH MPAC +3
049945,001395: 7611 00006 EXTEND
049946,001396: 7612 50116 INDEX ADDRWD
049947,001397: 7613 30005 DCA 4
049948,001398: 7614 52162 DXCH MPAC +5
049949,001399:
049950,001400: 7615 45017 CS ONE # CHANGE MODE TO VECTOR.
049951,001401: 7616 54163 TS MODE
049952,001402:
049953,001403: 7617 00006 EXTEND
049954,001404: 7620 50116 INDEX ADDRWD
049955,001405: 7621 30001 DCA 0
049956,001406: 7622 52155 DXCH MPAC
049957,001407: 7623 17601 TCF V/SC1 # FINISH PROLOGUE AT COMMON SECTION.
049958,001408:
|
|
Page 1151
|
049960,001410: # SIGN AND COMPLEMENT INSTRUCTIONS.
049961,001411:
049962,001412: 7624 50116 SIGN INDEX ADDRWD # CALL COMP INSTRUCTION IF WORD AT X IS
049963,001413: 7625 10000 CCS 0 # NEGATIVE NON-ZERO.
049964,001414: 7626 16030 TCF DANZIG
049965,001415: 7627 17631 TCF +2
049966,001416: 7630 17637 TCF COMP # DO THE COMPLEMENT.
049967,001417:
049968,001418: 7631 50116 INDEX ADDRWD
049969,001419: 7632 10001 CCSL CCS 1
049970,001420: 7633 16030 TCF DANZIG
049971,001421: 7634 16030 TCF DANZIG
049972,001422: 7635 17637 TCF COMP
049973,001423: 7636 16030 TCF DANZIG
049974,001424: 7637 00006 COMP EXTEND # COMPLEMENT DP MPAC IN EVERY CASE.
049975,001425: 7640 40155 DCS MPAC
049976,001426: 7641 52155 DXCH MPAC
049977,001427:
049978,001428: 7642 10163 CCS MODE # EITHER COMPLEMENT MPAC +3 OR THE REST OF
049979,001429: 7643 17654 TCF DCOMP # THE VECTOR ACCUMULATOR.
049980,001430: 7644 17654 TCF DCOMP
049981,001431:
049982,001432: 7645 00006 EXTEND # VECTOR COMPLEMENT.
049983,001433: 7646 40160 DCS MPAC +3
049984,001434: 7647 52160 DXCH MPAC +3
049985,001435: 7650 00006 EXTEND
049986,001436: 7651 40162 DCS MPAC +5
049987,001437: 7652 52162 DXCH MPAC +5
049988,001438: 7653 16030 TCF DANZIG
049989,001439:
049990,001440: 7654 40156 DCOMP CS MPAC +2
049991,001441: 7655 54156 TS MPAC +2
049992,001442: 7656 16030 TCF DANZIG
049993,001443:
|
|
Page 1152
|
049995,001445: # THE FOLLOWING SHORT SHIFT CODES REQUIRE NO ADDRESS WORD:
049996,001446: # 1. SR1 TO SR4 SCALAR SHIFT RIGHT.
049997,001447: # 2. SR1R TO SR4R SCALAR SHIFT RIGHT AND ROUND.
049998,001448: # 3. SL1 TO SL4 SCALAR SHIFT LEFT.
049999,001449: # 4. SL1R TO SL4R SCALAR SHIFT LEFT AND ROUND.
050000,001450: # 5. VSR1 TO VSR8 VECTOR SHIFT RIGHT (ALWAYS ROUNDS).
050001,001451: # 6. VSL1 TO VSL8 VECTOR SHIFT LEFT (NEVER ROUNDS).
050002,001452: # THE FOLLOWING CODES REQUIRE AN ADDRESS WHICH MAY BE INDEXED:*
050003,001453: # 1. SR SCALAR SHIFT RIGHT.
050004,001454: # 2. SRR SCALAR SHIFT RIGHT AND ROUND.
050005,001455: # 3. SL SCALAR SHIFT LEFT.
050006,001456: # 4. SLR SCALAR SHIFT LEFT AND ROUND.
050007,001457: # 5. VSR VECTOR SHIFT RIGHT.
050008,001458: # 6. VSL VECTOR SHIFT LEFT.
050009,001459: # * IF THE ADDRESS IS INDEXED, AND THE INDEX MODIFICATION RESULTS IN A NEGATIVE SHIFT COUNT, A SHIFT OF THE
050010,001460: # ABSOLUTE VALUE OF THE COUNT IS DONE IN THE OPPOSITE DIRECTION.
050011,001461:
050012,001462: 00,2017 BANK 00
050013,001463:
050014,001464: 00,2017 COUNT 00/INTER
050015,001465:
050016,001466: 00,2017 36211 SHORTT CAF SIX # SCALAR SHORT SHIFTS COME HERE. THE SHIFT
050017,001467: 00,2020 70020 MASK CYR # COUNT-1 IS NOW IN BITS 2-3 OF CYR. THE
050018,001468: 00,2021 54021 TS SR # ROUNDING BIT IS IN BIT1 AT THIS POINT.
050019,001469:
050020,001470: 00,2022 10020 CCS CYR # SEE IF RIGHT OR LEFT SHIFT DESIRED.
050021,001471: 00,2023 12101 TCF TSSL # SHIFT LEFT.
050022,001472:
050023,001473: 00,2024 00024 SRDDV DEC 20 B-14 # MPTEMP SETTING FOR SR BEFORE DDV.
050024,001474:
050025,001475: 00,2025 50021 TSSR INDEX SR # GET SHIFTING BIT.
050026,001476: 00,2026 35002 CAF BIT14
050027,001477: 00,2027 54135 TS MPTEMP
050028,001478:
050029,001479: 00,2030 10020 CCS CYR # SEE IF A ROUND IS DESIRED.
050030,001480: 00,2031 02050 RIGHTR TC MPACSRND # YES - SHIFT RIGHT AND ROUND.
050031,001481: 00,2032 16027 TCF NEWMODE # SET MODE TO DP (C(A) = 0).
050032,001482: 00,2033 30135 MPACSHR CA MPTEMP # DO A TRIPLE PRECISION SHIFT RIGHT.
050033,001483: 00,2034 00006 EXTEND
050034,001484: 00,2035 70156 MP MPAC +2
050035,001485: 00,2036 54156 +3 TS MPAC +2 # (EXIT FROM SQRT AND ABVAL).
050036,001486: 00,2037 30135 CA MPTEMP
050037,001487: 00,2040 00006 EXTEND
|
|
Page 1153
|
050039,001489: 00,2041 70154 MP MPAC # SHIFT MAJOR PART INTO A,L AND PLACE IN
050040,001490: 00,2042 52155 DXCH MPAC # MPAC,+1.
050041,001491: 00,2043 30135 CA MPTEMP
050042,001492: 00,2044 00006 EXTEND
050043,001493: 00,2045 70001 MP L # ORIGINAL C(MPAC +1).
050044,001494: 00,2046 20156 DAS MPAC +1 # GUARANTEED NO OVERFLOW.
050045,001495: 00,2047 16030 TCF DANZIG
050046,001496:
050047,001497: # MPAC SHIFT RIGHT AND ROUND SUBROUTINES.
050048,001498:
050049,001499: 00,2050 30156 MPACSRND CA MPAC +2 # WE HAVE TO DO ALL THREE MULTIPLIES SINCE
050050,001500: 00,2051 00006 EXTEND # MPAC +1 AND MPAC +2 MIGHT HAVE SIGN
050051,001501: 00,2052 70135 MP MPTEMP # DISAGREEMENT WITH A SHIFT RIGHT OF 1.
050052,001502: 00,2053 56155 XCH MPAC +1
050053,001503: 00,2054 00006 EXTEND
050054,001504: 00,2055 70135 MP MPTEMP
050055,001505: 00,2056 56155 XCH MPAC +1 # TRIAL MINOR PART.
050056,001506: 00,2057 60001 AD L
050057,001507:
050058,001508: 00,2060 60000 VSHR2 DOUBLE # (FINISH VECTOR COMPONENT SHIFT RIGHT
050059,001509: 00,2061 54156 TS MPAC +2 # AND ROUND.
050060,001510: 00,2062 12064 TCF +2
050061,001511: 00,2063 26155 ADS MPAC +1 # GUARANTEED NO OVERFLOW.
050062,001512:
050063,001513: 00,2064 35021 CAF ZERO
050064,001514: 00,2065 54156 TS MPAC +2
050065,001515: 00,2066 56154 XCH MPAC # SETTING TO ZERO SO FOLLOWING DAS WORKS.
050066,001516: 00,2067 00006 EXTEND
050067,001517: 00,2070 70135 MP MPTEMP
050068,001518: 00,2071 20155 DAS MPAC # AGAIN NO OVERFLOW.
050069,001519: 00,2072 00002 TC Q
050070,001520:
050071,001521: 00,2073 30135 VSHRRND CA MPTEMP # ENTRY TO SHIFT RIGHT AND ROUND MPAC WHEN
050072,001522: 00,2074 00006 EXTEND # MPAC CONTAINS A VECTOR COMPONENT.
050073,001523: 00,2075 70155 MP MPAC +1
050074,001524: 00,2076 54155 TS MPAC +1
050075,001525: 00,2077 56001 XCH L
050076,001526: 00,2100 12060 TCF VSHR2 # GO ADD ONE IF NECESSARY AND FINISH.
050077,001527:
|
|
Page 1154
|
050079,001529: # ROUTINE FOR SHORT SCALAR SHIFT LEFT (AND MAYBE ROUND).
050080,001530:
050081,001531: 00,2101 30021 TSSL CA SR # GET SHIFT COUNT FOR SR.
050082,001532: 00,2102 54135 +1 TS MPTEMP
050083,001533:
050084,001534: 00,2103 00006 +2 EXTEND # ENTRY HERE FROM SL FOR SCALARS.
050085,001535: 00,2104 30156 DCA MPAC +1 # SHIFTING LEFT ONE PLACE AT A TIME IS
050086,001536: 00,2105 20156 DAS MPAC +1 # FASTER THAN DOING THE WHOLE SHIFT WITH
050087,001537: 00,2106 60154 AD MPAC # MULTIPLIES ASSUMING THAT FREQUENCY OF
050088,001538: 00,2107 60154 AD MPAC # SHIFT COUNTS GOES DOWN RAPIDLY AS A
050089,001539: 00,2110 54154 TS MPAC # FUNCTION OF THEIR MAGNITUDE.
050090,001540: 00,2111 12113 TCF +2
050091,001541: 00,2112 54121 TS OVFIND # OVERFLOW. (LEAVES OVERFLOW-CORRECTED
050092,001542: # RESULT ANYWAY).
050093,001543: 00,2113 10135 CCS MPTEMP # LOOP ON DECREMENTED SHIFT COUNT.
050094,001544: 00,2114 12102 TCF TSSL +1
050095,001545:
050096,001546: 00,2115 10020 CCS CYR # SEE IF ROUND WANTED.
050097,001547: 00,2116 07105 ROUND TC ROUNDSUB # YES - ROUND AND EXIT.
050098,001548: 00,2117 16030 TCF DANZIG # SL LEAVES A ZERO IN CYR FOR NO ROUND.
050099,001549: 00,2120 16030 TCF DANZIG # NO - EXIT IMMEDIATL
050100,001550:
|
|
Page 1155
|
050102,001552: # VECTOR SHIFTING ROUTINES.
050103,001553:
050104,001554: 00,2121 35023 SHORTV CAF LOW3 # SAVE 3 BIT SHIFT COUNT - 1 WITHOUT
050105,001555: 00,2122 70020 MASK CYR # EDITING CYR.
050106,001556: 00,2123 54135 TS MPTEMP
050107,001557: 00,2124 10020 CCS CYR # SEE IF LEFT OR RIGHT SHIFT.
050108,001558: 00,2125 12145 TCF VSSL # VECTOR SHIFT LEFT.
050109,001559: 00,2126 00176 OCT176 OCT 176 # USED IN PROCESSED SHIFTS WITH - COUNT.
050110,001560:
050111,001561: 00,2127 50135 VSSR INDEX MPTEMP # (ENTRY FROM SR). PICK UP SHIFTING BIT.
050112,001562: 00,2130 35002 CAF BIT14 # MPTEMP CONTAINS THE SHIFT COUNT - 1.
050113,001563: 00,2131 54135 TS MPTEMP
050114,001564: 00,2132 02073 TC VSHRRND # SHIFT X COMPONENT.
050115,001565:
050116,001566: 00,2133 52155 DXCH MPAC # SWAP X AND Y COMPONENTS.
050117,001567: 00,2134 52160 DXCH MPAC +3
050118,001568: 00,2135 52155 DXCH MPAC
050119,001569: 00,2136 02073 TC VSHRRND # SHIFT Y COMPONENT.
050120,001570:
050121,001571: 00,2137 52155 DXCH MPAC # SWAP Y AND Z COMPONENTS.
050122,001572: 00,2140 52162 DXCH MPAC +5
050123,001573: 00,2141 52155 DXCH MPAC
050124,001574: 00,2142 02073 TC VSHRRND # SHIFT Z COMPONENT.
050125,001575:
050126,001576: 00,2143 17367 TCF VROTATEX # RESTORE COMPONENTS TO PROPER PLACES.
050127,001577:
|
|
Page 1156
|
050129,001579: # VECTOR SHIFT LEFT - DONE ONE PLACE AT A TIME.
050130,001580:
050131,001581: 00,2144 54135 -1 TS MPTEMP # SHIFTING LOOP.
050132,001582:
050133,001583: 00,2145 00006 VSSL EXTEND
050134,001584: 00,2146 30155 DCA MPAC
050135,001585: 00,2147 20155 DAS MPAC
050136,001586: 00,2150 00006 EXTEND
050137,001587: 00,2151 12153 BZF +2
050138,001588: 00,2152 06766 TC OVERFLOW
050139,001589:
050140,001590: 00,2153 00006 EXTEND
050141,001591: 00,2154 30160 DCA MPAC +3
050142,001592: 00,2155 20160 DAS MPAC +3
050143,001593: 00,2156 00006 EXTEND
050144,001594: 00,2157 12161 BZF +2
050145,001595: 00,2160 06763 TC OVERFLWY
050146,001596:
050147,001597: 00,2161 00006 EXTEND
050148,001598: 00,2162 30162 DCA MPAC +5
050149,001599: 00,2163 20162 DAS MPAC +5
050150,001600: 00,2164 00006 EXTEND
050151,001601: 00,2165 12167 BZF +2
050152,001602: 00,2166 06760 TC OVERFLWZ
050153,001603:
050154,001604: 00,2167 10135 CCS MPTEMP # LOOP ON DECREMENTED SHIFT COUNTER.
050155,001605: 00,2170 12144 TCF VSSL -1
050156,001606: 00,2171 16030 TCF DANZIG # EXIT.
050157,001607:
|
|
Page 1157
|
050159,001609: # TSLC - TRIPLE SHIFT LEFT AND COUNT. SHIFTS MPAC LEFT UNTIL GREATER THAN .5 IN MAGNITUDE, LEAVING
050160,001610: # THE COMPLEMENT OF THE NUMBER OF SHIFTS REQUIRED IN X.
050161,001611:
050162,001612: 00,2172 54135 TSLC2 TS MPTEMP # START BY ZEROING SHIFT COUNT (IN A NOW).
050163,001613: 00,2173 06672 TC BRANCH # EXIT WITH NO SHIFTING IF ARGUMENT ZERO.
050164,001614: 00,2174 12176 TCF +2
050165,001615: 00,2175 12212 TCF ENDTSLC # STORES ZERO SHIFT COUNT IN THIS CASE.
050166,001616:
050167,001617: 00,2176 07226 TC TPAGREE # MAY CAUSE UPSHIFT OF ONE EXTRA PLACE.
050168,001618:
050169,001619: 00,2177 30154 CA MPAC # BEGIN NORMALIZATION LOOP.
050170,001620: 00,2200 12207 TCF TSLCTEST
050171,001621:
050172,001622: 00,2201 24135 TSLCLOOP INCR MPTEMP # INCREMENT SHIFT COUNTER.
050173,001623: 00,2202 00006 EXTEND
050174,001624: 00,2203 30156 DCA MPAC +1
050175,001625: 00,2204 20156 DAS MPAC +1
050176,001626: 00,2205 60154 AD MPAC
050177,001627: 00,2206 26154 ADS MPAC
050178,001628: 00,2207 60000 TSLCTEST DOUBLE # SEE IF (ANOTHER) SHIFT IS REQUIRED.
050179,001629: 00,2210 54000 OVSK
050180,001630: 00,2211 12201 TCF TSLCLOOP # YES - INCREMENT COUNT AND SHIFT AGAIN.
050181,001631:
050182,001632: 00,2212 40135 ENDTSLC CS MPTEMP
050183,001633: 00,2213 16572 TCF STORE1 # STORE SHIFT COUNT AND RETURN TO DANZIG.
050184,001634:
|
|
Page 1158
|
050186,001636: # THE FOLLOWING ROUTINES PROCESSES THE GENERAL SHIFT INSTRUCTIONS SR, SRR, SL, AND SLR.
050187,001637: # THE GIVEN ADDRESS IS DECODED AS FOLLOWS:
050188,001638: # BITS 1-7 SHIFT COUNT (SUBADDRESS) LESS THAN 125 DECIMAL.
050189,001639: # BIT 8 PSEUDO SIGN BIT (DETECTS CHANGE IN SIGN IN INDEXED SHIFTS).
050190,001640: # BIT 9 0 FOR LEFT SHIFT, AND 1 FOR RIGHT SHIFT.
050191,001641: # BIT 10 1 FOR TERMINAL ROUND ON SCALAR SHIFTS, 0 OTHERWISE.
050192,001642: # BITS 11-13 0.
050193,001643: # BIT 14 1.
050194,001644: # BIT 15 0.
050195,001645: # THE ABOVE ENCODING IS DONE BY THE YUL SYSTEM.
050196,001646:
050197,001647: 00,2214 70116 GENSHIFT MASK ADDRWD # GET SHIFT COUNT, TESTING FOR ZERO.
050198,001648: 00,2215 10000 CCS A # (ARRIVES WITH C(A) = LOW7).
050199,001649: 00,2216 12224 TCF GENSHFT2 # IF NON-ZERO, PROCEED WITH DECREMENTED CT
050200,001650:
050201,001651: 00,2217 35006 CAF BIT10 # ZERO SHIFT COUNT. NO SHIFTS NEEDED BUT
050202,001652: 00,2220 70116 MASK ADDRWD # WE MIGHT HAVE TO ROUND MPAC ON SLR AND
050203,001653: 00,2221 10000 CCS A # SRR (SCALAR ONLY).
050204,001654: 00,2222 07105 TC ROUNDSUB
050205,001655: 00,2223 16030 TCF DANZIG
050206,001656:
050207,001657: 00,2224 54135 GENSHFT2 TS MPTEMP # DECREMENTED SHIFT COUNT TO MPTEMP.
050208,001658: 00,2225 35010 CAF BIT8 # TEST MEANING OF LOW SEVEN BIT COUNT IN
050209,001659: 00,2226 00006 EXTEND # MPTEMP NOW.
050210,001660: 00,2227 70116 MP ADDRWD
050211,001661: 00,2230 76214 MASK LOW2 # JUMPS ON SHIFT DIRECTION (BIT8) AND
050212,001662: 00,2231 50000 INDEX A
050213,001663: 00,2232 12233 TCF +1 # ORIGINAL SHIFT DIRECTION (BIT 9).
050214,001664: 00,2233 12332 TCF RIGHT- # NEGATIVE SHIFT COUNT FOR SL OR SLR.
050215,001665: 00,2234 12342 TCF LEFT # SL OR SLR.
050216,001666: 00,2235 12336 TCF LEFT- # NEGATIVE SHIFT COUNT WITH SR OR SRR.
050217,001667:
|
|
Page 1159
|
050219,001669: # GENERAL SHIFT RIGHT.
050220,001670:
050221,001671: 00,2236 10163 RIGHT CCS MODE # SEE IF VECTOR OR SCALAR.
050222,001672: 00,2237 12277 TCF GENSCR
050223,001673: 00,2240 12277 TCF GENSCR
050224,001674:
050225,001675: 00,2241 30135 CA MPTEMP # SEE IF SHIFT COUNT LESS THAN 14D.
050226,001676: 00,2242 63730 VRIGHT2 AD NEG12
050227,001677: 00,2243 00006 EXTEND
050228,001678: 00,2244 62127 BZMF VSSR # IF SO, BRANCH AND SHIFT IMMEDIATELY.
050229,001679:
050230,001680: 00,2245 67716 AD NEGONE # IF NOT, REDUCE MPTEMP BY A TOTAL OF 14,
050231,001681: 00,2246 54135 TS MPTEMP # AND DO A SHIFT RIGHT AND ROUND BY 14.
050232,001682: 00,2247 35021 CAF ZERO # THE ROUND AT THIS STAGE MAY INTRODUCE A
050233,001683: 00,2250 54001 TS L # ONE BIT ERROR IN A SHIFT RIGHT 15D.
050234,001684: 00,2251 56154 XCH MPAC
050235,001685: 00,2252 56155 XCH MPAC +1
050236,001686: 00,2253 02272 TC SETROUND # X COMPONENT NOW SHIFTED, SO MAKE UP THE
050237,001687: 00,2254 20155 DAS MPAC # ROUNDING QUANTITY (0 IN A AND 0 OR +-1
050238,001688: # IN L).
050239,001689: 00,2255 56157 XCH MPAC +3 # REPEAT THE ABOVE PROCESS FOR Y AND Z.
050240,001690: 00,2256 56160 XCH MPAC +4
050241,001691: 00,2257 02272 TC SETROUND
050242,001692: 00,2260 20160 DAS MPAC +3 # NO OVERFLOW ON THESE ADDS.
050243,001693:
050244,001694: 00,2261 56161 XCH MPAC +5
050245,001695: 00,2262 56162 XCH MPAC +6
050246,001696: 00,2263 02272 TC SETROUND
050247,001697: 00,2264 20162 DAS MPAC +5
050248,001698:
050249,001699: 00,2265 10135 CCS MPTEMP # SEE IF DONE, DOING FINAL DECREMENT.
050250,001700: 00,2266 54135 TS MPTEMP
050251,001701: 00,2267 12242 TCF VRIGHT2
050252,001702: 00,2270 04604 BIASLO DEC .2974 B-1 # SQRT CONSTANT
050253,001703:
050254,001704: 00,2271 16030 TCF DANZIG
050255,001705:
050256,001706: 00,2272 60000 SETROUND DOUBLE # MAKES UP ROUNDING QUANTITY FROM ARRIVING
050257,001707: 00,2273 54156 TS MPAC +2 # C(A). L IS ZERO INITIALLY.
050258,001708: 00,2274 35021 CAF ZERO
050259,001709: 00,2275 56001 XCH L
050260,001710: 00,2276 00002 TC Q # RETURN AND DO THE DAS, RESETTING L TO 0.
050261,001711:
|
|
Page 1160
|
050263,001713: # PROCESS SR AND SRR FOR SCALARS.
050264,001714:
050265,001715: 00,2277 30135 GENSCR CA MPTEMP # SEE IF THE ORIGINAL SHIFT COUNT WAS LESS
050266,001716: 00,2300 63730 +1 AD NEG12 # THAN 14D.
050267,001717: 00,2301 00006 EXTEND
050268,001718: 00,2302 62322 BZMF DOSSHFT # DO THE SHIFT IMMEDIATELY IF SO.
050269,001719:
050270,001720: 00,2303 67716 +4 AD NEGONE # IF NOT, DECREMENT SHIFT COUNT BY 14D AND
050271,001721: 00,2304 54135 TS MPTEMP # SHIFT MPAC RIGHT 14 PLACES.
050272,001722: 00,2305 35021 CAF ZERO
050273,001723: 00,2306 56154 XCH MPAC
050274,001724: 00,2307 56155 XCH MPAC +1
050275,001725: 00,2310 54156 TS MPAC +2
050276,001726: 00,2311 10135 CCS MPTEMP # SEE IF FINISHED, DO FINAL DECREMENT.
050277,001727: 00,2312 54135 TS MPTEMP
050278,001728: 00,2313 02300 TC GENSCR +1
050279,001729: 00,2314 22650 SLOPEHI DEC .5884 # SQRT CONSTANT.
050280,001730: 00,2315 35006 CAF BIT10 # FINISHED WITH SHIFT. SEE IF ROUND
050281,001731: 00,2316 70116 MASK ADDRWD # WANTED.
050282,001732: 00,2317 10000 CCS A
050283,001733: 00,2320 07105 TC ROUNDSUB
050284,001734: 00,2321 16030 TCF DANZIG # DO SO AND/OR EXIT.
050285,001735:
050286,001736: 00,2322 50135 DOSSHFT INDEX MPTEMP # PICK UP SHIFTING BIT.
050287,001737: 00,2323 35002 CAF BIT14
050288,001738: 00,2324 54135 TS MPTEMP
050289,001739: 00,2325 35006 CAF BIT10 # SEE IF TERMINAL ROUND DESIRED.
050290,001740: 00,2326 70116 MASK ADDRWD
050291,001741: 00,2327 10000 CCS A
050292,001742: 00,2330 12031 TCF RIGHTR # YES.
050293,001743: 00,2331 12033 TCF MPACSHR # JUST SHIFT RIGHT.
050294,001744:
|
|
Page 1161
|
050296,001746: # PROCESS THE RIGHT- (SL(R) WITH A NEGATIVE COUNT), LEFT-, AND LEFT OPTIONS.
050297,001747:
050298,001748: 00,2332 40135 RIGHT- CS MPTEMP # GET ABSOLUTE VALUE - 1 OF SHIFT COUNT
050299,001749: 00,2333 62126 AD OCT176 # UNDERSTANDING THAT BIT8 (PSEUDO-SIGN)
050300,001750: 00,2334 54135 TS MPTEMP # WAS 1 INITIALLY.
050301,001751: 00,2335 12236 TCF RIGHT # DO NORMAL SHIFT RIGHT.
050302,001752:
050303,001753: 00,2336 42126 LEFT- CS OCT176 # SAME PROLOGUE TO LEFT FOR INDEXED RIGHT
050304,001754: 00,2337 60135 AD MPTEMP # SHIFTS WHOSE NET SHIFT COUNT IS NEGATIVE
050305,001755: 00,2340 40000 COM
050306,001756: 00,2341 54135 TS MPTEMP
050307,001757:
050308,001758: 00,2342 10163 LEFT CCS MODE # SINCE LEFT SHIFTING IS SONE ONE PLACE AT
050309,001759: 00,2343 12346 TCF GENSCL # A TIME, NO COMPARISON WITH 14 NEED BE
050310,001760: 00,2344 12346 TCF GENSCL # DONE. FOR SCALARS, SEE IF TERMINAL ROUND
050311,001761: 00,2345 12145 TCF VSSL # DESIRED. FOR VECTORS, SHIFT IMMEDIATELY.
050312,001762:
050313,001763: 00,2346 40116 GENSCL CS ADDRWD # PUT ROUNDING BIT (BIT 10 OF ADDRWD) INTO
050314,001764: 00,2347 00006 EXTEND # BIT 15 OF CYR WHERE THE ROUNDING BIT OF
050315,001765: 00,2350 75012 MP BIT6 # A SHORT SHIFT LEFT WOULD BE
050316,001766: 00,2351 54020 TS CYR
050317,001767: 00,2352 12103 TCF TSSL +2 # DO THE SHIFT.
050318,001768:
|
|
Page 1162
|
050320,001770: # SCALAR DIVISION INSTRUCTIONS, DDV AND BDDV, ARE EXECUTED HERE. AT THIS POINT, THE DIVIDEND IS IN MPAC
050321,001771: # AND THE DIVISOR IN BUF.
050322,001772:
050323,001773: 00,2353 45017 DDV/BDDV CS ONE # INITIALIZATION.
050324,001774: 00,2354 54136 TS DVSIGN # +-1 FOR POSITIVE QUOTIENT - -0 FOR NEG.
050325,001775: 00,2355 54137 TS DVNORMCT # DIVIDEND NORMALIZATION COUNT.
050326,001776: 00,2356 54140 TS MAXDVSW # NEAR-ONE DIVIDE FLAG.
050327,001777:
050328,001778: 00,2357 10130 CCS BUF # FORCE BUF POSITIVE WITH THE MAJOR PART
050329,001779: 00,2360 12516 TCF BUFPOS # NON-ZERO.
050330,001780: 00,2361 12363 TCF +2
050331,001781: 00,2362 12531 TCF BUFNEG
050332,001782:
050333,001783: 00,2363 54156 BUFZERO TS MPAC +2 # ZERO THIS.
050334,001784: 00,2364 07226 TC TPAGREE # FORCE SIGN AGREEMENT BEFORE OVERFLOW
050335,001785:
050336,001786: 00,2365 10154 CCS MPAC # TEST TO SEE IF MPAC NON-ZERO. (TOO BIG)
050337,001787: 00,2366 12414 TCF OVF+ # MAJOR PART OF DIVIDEND IS POSITIVE NON-0
050338,001788: 00,2367 12371 TCF +2
050339,001789: 00,2370 12413 TCF OVF+ -1 # MAJOR PART OF DIVIDEND IS NEG. NON-ZERO
050340,001790:
050341,001791: 00,2371 56131 XCH BUF +1 # SHIFT DIVIDEND AND DIVISOR LEFT 14.
050342,001792: 00,2372 56130 XCH BUF
050343,001793: 00,2373 56155 XCH MPAC +1
050344,001794: 00,2374 56154 XCH MPAC
050345,001795: 00,2375 10130 CCS BUF # TRY AGAIN ON FORMER MINOR PART.
050346,001796: 00,2376 12422 TCF BUF+
050347,001797: 00,2377 12401 TCF +2 # OVERFLOW ON ZERO DIVISOR.
050348,001798: 00,2400 12416 TCF BUF-
050349,001799:
050350,001800: 00,2401 40154 CS MPAC # SIGN OF MPAC DETERMINES SIGN OF RESULT.
050351,001801: 00,2402 00006 SGNDVOVF EXTEND
050352,001802: 00,2403 62405 BZMF +2
050353,001803: 00,2404 24136 INCR DVSIGN # NEGMAX IN MPAC PERHAPS.
050354,001804: 00,2405 34777 DVOVF CAF POSMAX # ON DIVISION OVERFLOW OF ANY SORT, SET
050355,001805: 00,2406 54154 TS MPAC # SET DP MPAC TO +-POSMAX.
050356,001806: 00,2407 02630 TC FINALDV +3
050357,001807: 00,2410 35017 CAF ONE # SET OVERFLOW INDICATOR AND EXIT.
050358,001808: 00,2411 54121 TS OVFIND
050359,001809: 00,2412 06030 TC DANZIG
050360,001810:
050361,001811: 00,2413 24136 -1 INCR DVSIGN
050362,001812: 00,2414 40131 OVF+ CS BUF +1 # LOAD LOWER ORDER PART OF DIVISOR.
050363,001813: 00,2415 12402 TCF SGNDVOVF # GET SIGN OF RESULT.
050364,001814:
050365,001815: 00,2416 00006 BUF- EXTEND # IF BUF IS NEGATIVE, COMPLEMENT IT AND
050366,001816: 00,2417 40131 DCS BUF # MAINTAIN DVSIGN FOR FINAL QUOTIENT SIGN.
050367,001817: 00,2420 52131 DXCH BUF
050368,001818: 00,2421 24136 INCR DVSIGN # NOW -0.
050369,001819:
|
|
Page 1163
|
050371,001821: 00,2422 10154 BUF+ CCS MPAC # FORCE MPAC POSITIVE, CHECKING FOR ZERO
050372,001822: 00,2423 12437 TCF MPAC+ # DIVIDEND IN THE PROCESS.
050373,001823: 00,2424 12426 TCF +2
050374,001824: 00,2425 12433 TCF MPAC-
050375,001825: 00,2426 10155 CCS MPAC +1
050376,001826: 00,2427 12437 TCF MPAC+
050377,001827: 00,2430 16030 TCF DANZIG # EXIT IMMEDIATELY ON ZERO DIVIDEND.
050378,001828: 00,2431 12433 TCF MPAC-
050379,001829: 00,2432 16030 TCF DANZIG
050380,001830:
050381,001831: 00,2433 00006 MPAC- EXTEND # FORCE MPAC POSITIVE AS BUF IN BUF-.
050382,001832: 00,2434 40155 DCS MPAC
050383,001833: 00,2435 52155 DXCH MPAC
050384,001834: 00,2436 24136 INCR DVSIGN # NOW +1 OR -0.
050385,001835:
|
|
Page 1164
|
050387,001837: 00,2437 40154 MPAC+ CS MPAC # CHECK FOR DIVISION OVERFLOW. IF THE
050388,001838: 00,2440 67716 AD NEGONE # MAJOR PART OF THE DIVIDEND IS LESS THAN
050389,001839: 00,2441 60130 AD BUF # THE MAJOR PART OF THE DIVISOR BY AT
050390,001840: 00,2442 10000 CCS A # LEAST TWO, WE CAN PROCEED IMMEDIATELY
050391,001841: 00,2443 12505 TCF DVNORM # WITHOUT NORMALIZATION PRODUCING A DVMAX.
050392,001842: 00,2444 60001 -1/2+2 OCT 60001 # USED IN SQRTSUB.
050393,001843:
050394,001844: 00,2445 12446 TCF +1 # IF THE ABOVE DOES NOT HOLD, FORCE SIGN
050395,001845: 00,2446 35002 CAF HALF # AGREEMENT IN NUMERATOR AND DENOMINATOR
050396,001846: 00,2447 60000 DOUBLE # TO FACILITATE OVERFLOW AND NEAR-ONE
050397,001847: 00,2450 60155 AD MPAC +1 # CHECKING.
050398,001848: 00,2451 54155 TS MPAC +1
050399,001849: 00,2452 35021 CAF ZERO
050400,001850: 00,2453 64777 AD POSMAX
050401,001851: 00,2454 26154 ADS MPAC
050402,001852:
050403,001853: 00,2455 35002 CAF HALF # SAME FOR BUF.
050404,001854: 00,2456 60000 DOUBLE
050405,001855: 00,2457 60131 AD BUF +1
050406,001856: 00,2460 54131 TS BUF +1
050407,001857: 00,2461 35021 CAF ZERO
050408,001858: 00,2462 64777 AD POSMAX
050409,001859: 00,2463 26130 ADS BUF
050410,001860:
050411,001861: 00,2464 40154 CS MPAC # CHECK MAGNITUDE OF SIGN-CORRECTED
050412,001862: 00,2465 60130 AD BUF # OPERANDS.
050413,001863: 00,2466 10000 CCS A
050414,001864: 00,2467 12505 TCF DVNORM # DIVIDE OK - WILL NOT BECOME MAXDV CASE.
050415,001865: 00,2470 00133 LBUF2 ADRES BUF2
050416,001866: 00,2471 12405 TCF DVOVF # DIVISOR NOT LESS THAN DIVIDEND - OVF.
050417,001867:
050418,001868: 00,2472 54140 TS MAXDVSW # IF THE MAJOR PARTS OF THE DIVIDEND AND
050419,001869: 00,2473 40155 CS MPAC +1 # DIVISOR ARE EQUAL, A SPECIAL APPROXIMA-
050420,001870: 00,2474 60131 AD BUF +1 # TION IS USED (PROVIDED THE DIVISION IS
050421,001871: 00,2475 00006 EXTEND # POSSIBLE, OF COURSE).
050422,001872: 00,2476 62405 BZMF DVOVF
050423,001873: 00,2477 12505 TCF DVNORM # IF NO OVERFLOW.
050424,001874:
|
|
Page 1165
|
050426,001876: 00,2500 00006 BUFNORM EXTEND # ADD -1 TO AUGMENT SHIFT COUNT AND SHIFT
050427,001877: 00,2501 24137 AUG DVNORMCT # LEFT ONE PLACE.
050428,001878: 00,2502 00006 EXTEND
050429,001879: 00,2503 30131 DCA BUF
050430,001880: 00,2504 20131 DAS BUF
050431,001881:
050432,001882: 00,2505 30130 DVNORM CA BUF # SEE IF DIVISOR NORMALIZED YET.
050433,001883: 00,2506 60000 DOUBLE
050434,001884: 00,2507 54000 OVSK
050435,001885: 00,2510 12500 TCF BUFNORM # NO - SHIFT LEFT ONE AND TRY AGAIN.
050436,001886:
050437,001887: 00,2511 52155 DXCH MPAC # CALL DIVIDEND NORMALIZATION SEQUENCE
050438,001888: 00,2512 50137 INDEX DVNORMCT # PRIOR TO DOING THE DIVIDE.
050439,001889: 00,2513 02565 TC MAXTEST
050440,001890:
050441,001891: 00,2514 54156 TS MPAC +2 # RETURNS WITH DIVISION DONE AND C(A) = 0.
050442,001892: 00,2515 16030 TCF DANZIG
050443,001893:
050444,001894: 00,2516 10000 BUFPOS CCS A
050445,001895: 00,2517 12422 TCF BUF+ # TO BUF+ IF BUF IS GREATER THAN +1.
050446,001896:
050447,001897: 00,2520 40131 CS BUF +1 # IF BUF IS +1, FORCING SIGN AGREEMENT
050448,001898: 00,2521 00006 EXTEND # MAY CAUSE BUF TO BECOME ZERO.
050449,001899: 00,2522 62422 BZMF BUF+ # BRANCH IF SIGNS AGREE.
050450,001900:
050451,001901: 00,2523 35002 CA HALF # SIGNS DISAGREE. FORCE AGREEMENT.
050452,001902: 00,2524 60000 +6 DOUBLE
050453,001903: 00,2525 26131 ADS BUF +1
050454,001904: 00,2526 35021 CA ZERO
050455,001905: 00,2527 54130 TS BUF
050456,001906: 00,2530 12363 TCF BUFZERO
050457,001907:
050458,001908: 00,2531 10000 BUFNEG CCS A
050459,001909: 00,2532 12416 TCF BUF- # TO BUF- IF BUF IS LESS THAN -1.
050460,001910:
050461,001911: 00,2533 30131 CA BUF +1 # IF BUF IS -1, FORCING SIGN AGREEMENT
050462,001912: 00,2534 00006 EXTEND # MAY CAUSE BUF TO BECOME ZERO.
050463,001913: 00,2535 62416 BZMF BUF- # BRANCH IF SIGNS AGREE.
050464,001914:
050465,001915: 00,2536 45002 CS HALF # SIGNS DISAGREE. FORCE AGREEMENT.
050466,001916: 00,2537 12524 TCF BUFPOS +6
050467,001917:
|
|
Page 1166
|
050469,001919: # THE FOLLOWING ARE PROLOGUES TO SHIFT THE DIVIDEND ARRIVING IN A AND L BEFORE THE DIVIDE.
050470,001920:
050471,001921: 00,2540 22021 -21D LXCH SR # SPECIAL PROLOGUE FOR UNIT WHEN THE
050472,001922: 00,2541 00006 EXTEND # LENGTH OF THE ARGUMENT WAS NOT LESS THAN
050473,001923: 00,2542 75002 MP HALF # .5. IN THIS CASE, EACH COMPONENT MUST BE
050474,001924: 00,2543 56001 XCH L # SHIFTED RIGHT ONE TO PRODUCE A HALF-UNIT
050475,001925: 00,2544 60021 AD SR # VECTOR.
050476,001926: 00,2545 56001 XCH L
050477,001927: 00,2546 12571 TCF GENDDV +1 # WITH DP DIVIDEND IN A,L.
050478,001928:
050479,001929: 00,2547 20001 DDOUBL # PROLOGUE WHICH NORMALIZES THE DIVIDEND
050480,001930: 00,2550 20001 DDOUBL # WHEN IT IS KNOWN THAT NO DIVISION
050481,001931: 00,2551 20001 DDOUBL # OVERFLOW WILL OCCUR.
050482,001932: 00,2552 20001 DDOUBL
050483,001933: 00,2553 20001 DDOUBL
050484,001934: 00,2554 20001 DDOUBL
050485,001935: 00,2555 20001 DDOUBL
050486,001936: 00,2556 20001 DDOUBL
050487,001937: 00,2557 20001 DDOUBL
050488,001938: 00,2560 20001 DDOUBL
050489,001939: 00,2561 20001 DDOUBL
050490,001940: 00,2562 20001 DDOUBL
050491,001941: 00,2563 20001 DDOUBL
050492,001942: 00,2564 52155 DXCH MPAC
050493,001943:
050494,001944: 00,2565 10140 MAXTEST CCS MAXDVSW # 0 IF MAJORS MIGHT BE =, -1 OTHERWISE.
050495,001945: 00,2566 06552 BIASHI DEC .4192 B-1 # SQRT CONSTANTS
050496,001946:
050497,001947: 00,2567 12642 TCF MAXDV # CHECK TO SEE IF THEY ARE NOW EQUAL.
050498,001948:
|
|
Page 1167
|
050500,001950: # THE FOLLOWING IS A GENERAL PURPOSE DOUBLE PRECISION DIVISION ROUTINE. IT DIVIDES MPAC BY BUF AND LEAVES
050501,001951: # THE RESULT IN MPAC. THE FOLLOWING CONDITIONS MUST BE SATISFIED:
050502,001952:
050503,001953: # 1. THE DIVISOR (BUF) MUST BE POSITIVE AND NOT LESS THAN .5.
050504,001954:
050505,001955: # 2. THE DIVIDEND (MPAC) MUST BE POSITIVE WITH THE MAJOR PART OF MPAC STRICTLY LESS THAN THAT OF BUF
050506,001956: # (A SPECIAL APPROXIMATION, MAXDV, IS USED WHEN THE MAJOR PARTS ARE EQUAL).
050507,001957:
050508,001958: # UNDERSTANDING THAT A/B = Q + S(R/B) WHERE S = 2(-14) AND Q AND R ARE QUOTIENT AND REMAINDER, RESPEC-
050509,001959: # TIVELY, THE FOLLOWING APPROXIMATION IS OBTAINED BY MULTIPLYING ABOVE AND BELOW BY C - SD AND NEGLECTING TERMS OF
050510,001960: # ORDER S-SQUARED (POSSIBLY INTRODUCING ERROR INTO THE LOW TWO BITS OF THE RESULT). SIGN AGREEMENT IS UNNECESSARY.
050511,001961:
050512,001962: # A + SB . (R - QD) A + SB
050513,001963: # ------ = Q + S(------) WHERE Q AND R ARE QUOTIENT AND REMAINDER OF ------ RESPECTIVELY.
050514,001964: # C + SD ( C ) C
050515,001965:
050516,001966: 00,2570 52155 GENDDV DXCH MPAC # WE NEED A AND B ONLY FOR FIRST DV.
050517,001967: 00,2571 00006 +1 EXTEND # (SPECIAL UNIT PROLOGUE ENTERS HERE).
050518,001968: 00,2572 10130 DV BUF # A NOW CONTAINS Q AND L, R.
050519,001969: 00,2573 52155 DXCH MPAC
050520,001970:
050521,001971: 00,2574 40154 CS MPAC # FORM DIVIDEND FOR MINOR PART OF RESULT.
050522,001972: 00,2575 00006 EXTEND
050523,001973: 00,2576 70131 MP BUF +1
050524,001974: 00,2577 60155 AD MPAC +1 # OVERFLOW AT THIS POINT IS POSITIVE SINCE
050525,001975: 00,2600 54000 OVSK # R IS POSITIVE IN EVERY CASE.
050526,001976: 00,2601 12606 TCF +5
050527,001977:
050528,001978: 00,2602 00006 EXTEND # OVERFLOW CAN BE REMOVED BY SUBTRACTING C
050529,001979: 00,2603 60130 SU BUF # (BUF) ONCE SINCE R IS ALWAYS LESS THAN C
050530,001980: 00,2604 24154 INCR MPAC # IN THIS CASE. INCR COMPENSATES SUBTRACT.
050531,001981: 00,2605 12610 TCF +DOWN # (SINCE C(A) IS STILL POSITIVE).
050532,001982:
050533,001983: 00,2606 00006 +5 EXTEND # C(A) CAN BE MADE LESS THAN C IN MAGNI-
050534,001984: 00,2607 62620 BZMF -UP # TUDE BY DIMINISHING IT BY C (SINCE C IS
050535,001985: # NOT LESS THAN .5) UNLESS C(A) = 0.
050536,001986:
|
|
Page 1168
|
050538,001988: 00,2610 00006 +DOWN EXTEND
050539,001989: 00,2611 60130 SU BUF # IF POSITIVE, REDUCE ONLY IF NECESSARY
050540,001990: 00,2612 00006 EXTEND # SINCE THE COMPENSATING INCR MIGHT CAUSE
050541,001991: 00,2613 12616 BZF +3 # OVERFLOW.
050542,001992: 00,2614 00006 EXTEND # DONT SUBTRACT UNLESS RESULT IS POSITIVE
050543,001993: 00,2615 62624 BZMF ENDMAXDV # OR ZERO.
050544,001994:
050545,001995: 00,2616 24154 +3 INCR MPAC # KEEP SUBTRACT HERE AND COMPENSATE.
050546,001996: 00,2617 12625 TCF FINALDV
050547,001997:
050548,001998: 00,2620 00006 -UP EXTEND # IF ZERO, SET MINOR PART OF RESULT TO
050549,001999: 00,2621 12630 BZF FINALDV +3 # ZERO.
050550,002000:
050551,002001: 00,2622 00006 EXTEND # IF NEGATIVE, ADD C TO A, SUBTRACTING ONE
050552,002002: 00,2623 26154 DIM MPAC # TO COMPENSATE. DIM IS OK HERE SINCE THE
050553,002003: 00,2624 60130 ENDMAXDV AD BUF # MAJOR PART NEVER GOES NEGATIVE.
050554,002004:
|
|
Page 1169
|
050556,002006: 00,2625 22007 FINALDV ZL # DO DV TO OBTAIN MINOR PART OF RESULT.
050557,002007: 00,2626 00006 EXTEND
050558,002008: 00,2627 10130 DV BUF
050559,002009: 00,2630 54155 +3 TS MPAC +1
050560,002010:
050561,002011: 00,2631 10136 CCS DVSIGN # LEAVE RESULT POSITIVE UNLESS C(DVSIGN)=
050562,002012: 00,2632 00002 TC Q # -0.
050563,002013: 00,2633 00002 TC Q
050564,002014: 00,2634 00002 TC Q
050565,002015:
050566,002016: 00,2635 00006 EXTEND
050567,002017: 00,2636 40155 DCS MPAC
050568,002018: 00,2637 52155 DXCH MPAC
050569,002019: 00,2640 35021 CAF ZERO # SO WE ALWAYS RETURN WITH C(A) = 0.
050570,002020: 00,2641 00002 TC Q
050571,002021:
|
|
Page 1170
|
050573,002023: # IF THE MAJOR PARTS OF THE DIVISOR AND DIVIDEND ARE EQUAL, BUT THE MINOR PARTS ARE SUCH THAT THE
050574,002024: # DIVIDEND IS STRICTLY LESS THAN THE DIVISOR IN MAGNITUDE, THE FOLLOWING APPROXIMATION IS USED. THE ASSUMPTIONS
050575,002025: # ARE THE SAME AS THE GENERAL ROUTINE WITH THE ADDITION THAT SIGN AGREEMENT IS NECESSARY (B, C, & D POSITIVE).
050576,002026:
050577,002027: # C + SB . (C + B - D)
050578,002028: # ------ = 37777 + S(---------)
050579,002029: # C + SD ( C )
050580,002030:
050581,002031: # THE DIVISION MAY BE PERFORMED IMMEDIATELY SINCE B IS STRICTLY LESS THAN D AND C IS NOT LESS THAN .5.
050582,002032:
050583,002033: 00,2642 40154 MAXDV CS MPAC # SEE IF MAXDV CASE STILL HOLDS AFTER
050584,002034: 00,2643 60130 AD BUF # NORMALIZATION.
050585,002035: 00,2644 00006 EXTEND
050586,002036: 00,2645 12647 BZF +2
050587,002037: 00,2646 12570 TCF GENDDV # MPAC NOW LESS THAN BUF - DIVIDE AS USUAL
050588,002038:
050589,002039: 00,2647 34777 +2 CAF POSMAX # SET MAJOR PART OF RESULT.
050590,002040: 00,2650 54154 TS MPAC
050591,002041:
050592,002042: 00,2651 40131 CS BUF +1 # FORM DIVIDEND OF MINOR PART OF RESULT.
050593,002043: 00,2652 60155 AD MPAC +1
050594,002044: 00,2653 12624 TCF ENDMAXDV # GO ADD C AND DO DIVIDE, ATTACHING SIGN
050595,002045: # BEFORE EXITING.
050596,002046:
|
|
Page 1171
|
050598,002048: # VECTOR DIVIDED BY SCALAR, V/SC, IS EXECUTED HERE. THE VECTOR IS NOW IN MPAC WITH SCALAR IN BUF.
050599,002049:
050600,002050: 00,2654 45017 V/SC2 CS ONE # INITIALIZE DIVIDEND NORMALIZATION COUNT
050601,002051: 00,2655 54137 TS DVNORMCT # AND DIVISION SIGN REGISTER.
050602,002052: 00,2656 54127 TS VBUF +5
050603,002053:
050604,002054: 00,2657 03010 TC VECAGREE # FORCE SIGN AGREEMENT IN VECTOR
050605,002055:
050606,002056: 00,2660 52131 DXCH BUF
050607,002057: 00,2661 07513 TC ALSIGNAG # SIGN AGREE BUF
050608,002058: 00,2662 52131 DXCH BUF
050609,002059: 00,2663 10130 CCS BUF # FORCE DIVISOR POSITIVE WITH MAJOR PART
050610,002060: 00,2664 12721 TCF /BUF+ # NON-ZERO (IF POSSIBLE).
050611,002061: 00,2665 12667 TCF +2
050612,002062: 00,2666 12715 TCF /BUF-
050613,002063:
050614,002064: 00,2667 56131 XCH BUF +1 # SHIFT VECTOR AND SCALAR LEFT 14.
050615,002065: 00,2670 56130 XCH BUF
050616,002066: 00,2671 56155 XCH MPAC +1
050617,002067: 00,2672 56154 XCH MPAC
050618,002068: 00,2673 00006 EXTEND # CHECK FOR OVERFLOW IN EACH CASE.
050619,002069: 00,2674 12676 BZF +2
050620,002070: 00,2675 12405 TCF DVOVF
050621,002071:
050622,002072: 00,2676 56160 XCH MPAC +4
050623,002073: 00,2677 56157 XCH MPAC +3
050624,002074: 00,2700 00006 EXTEND
050625,002075: 00,2701 12703 BZF +2
050626,002076: 00,2702 12405 TCF DVOVF
050627,002077:
050628,002078: 00,2703 56162 XCH MPAC +6
050629,002079: 00,2704 56161 XCH MPAC +5
050630,002080: 00,2705 00006 EXTEND
050631,002081: 00,2706 12710 BZF +2
050632,002082: 00,2707 12405 TCF DVOVF
050633,002083:
050634,002084: 00,2710 10130 CCS BUF
050635,002085: 00,2711 12721 TCF /BUF+
050636,002086: 00,2712 12405 TCF DVOVF # ZERO DIVISOR - OVERFLOW.
050637,002087: 00,2713 12715 TCF /BUF-
050638,002088: 00,2714 12405 TCF DVOVF
050639,002089:
050640,002090: 00,2715 00006 /BUF- EXTEND # ON NEGATIVE, COMPLEMENT BUF AND MAINTAIN
050641,002091: 00,2716 40131 DCS BUF # DVSIGN IN VBUF +5.
050642,002092: 00,2717 52131 DXCH BUF
050643,002093: 00,2720 24127 INCR VBUF +5
050644,002094:
|
|
Page 1172
|
050646,002096: 00,2721 00006 /BUF+ EXTEND
050647,002097: 00,2722 30131 DCA BUF # LEAVE ABS(ORIG DIVISOR) IN BUF2
050648,002098: 00,2723 52134 DXCH BUF2 # FOR OVERFLOW TESTING
050649,002099: 00,2724 12732 TCF /NORM # NORMALIZE DIVISOR IN BUF.
050650,002100:
050651,002101: 00,2725 00006 /NORM2 EXTEND # IF LESS THAN .5, AUGMENT DVNORMCT AND
050652,002102: 00,2726 24137 AUG DVNORMCT # DOUBLE DIVISOR.
050653,002103: 00,2727 00006 EXTEND
050654,002104: 00,2730 30131 DCA BUF
050655,002105: 00,2731 20131 DAS BUF
050656,002106:
050657,002107: 00,2732 30130 /NORM CA BUF # SEE IF DIVISOR NORMALIZED.
050658,002108: 00,2733 60000 DOUBLE
050659,002109: 00,2734 54000 OVSK
050660,002110: 00,2735 12725 TCF /NORM2 # DOUBLE AND TRY AGAIN IF NOT.
050661,002111:
050662,002112: 00,2736 02750 TC V/SCDV # DO X COMPONENT DIVIDE.
050663,002113: 00,2737 52160 DXCH MPAC +3 # SUPPLY ARGUMENTS IN USUAL SEQUENCE.
050664,002114: 00,2740 52155 DXCH MPAC
050665,002115: 00,2741 52160 DXCH MPAC +3
050666,002116:
050667,002117: 00,2742 02750 TC V/SCDV # Y COMPONENT.
050668,002118: 00,2743 52162 DXCH MPAC +5
050669,002119: 00,2744 52155 DXCH MPAC
050670,002120: 00,2745 52162 DXCH MPAC +5
050671,002121:
050672,002122: 00,2746 02750 TC V/SCDV # Z COMPONENT.
050673,002123: 00,2747 17367 TCF VROTATEX # GO RE-ARRANGE COMPONENTS BEFORE EXIT.
050674,002124:
|
|
Page 1173
|
050676,002126: # SUBROUTINE USED BY V/SC TO DIVIDE VECTOR COMPONENT IN MPAC,+1 BY THE SCALAR GIVEN IN BUF.
050677,002127:
050678,002128: 00,2750 30127 V/SCDV CA VBUF +5 # REFLECTS SIGN OF SCALAR.
050679,002129: 00,2751 54136 TS DVSIGN
050680,002130:
050681,002131: 00,2752 10154 CCS MPAC # FORCE MPAC POSITIVE, EXITING ON ZERO.
050682,002132: 00,2753 12767 TCF /MPAC+
050683,002133: 00,2754 12756 TCF +2
050684,002134: 00,2755 12763 TCF /MPAC-
050685,002135:
050686,002136: 00,2756 10155 CCS MPAC +1
050687,002137: 00,2757 12767 TCF /MPAC+
050688,002138: 00,2760 00002 TC Q
050689,002139: 00,2761 12763 TCF /MPAC-
050690,002140: 00,2762 00002 TC Q
050691,002141:
050692,002142: 00,2763 00006 /MPAC- EXTEND # USUAL COMPLEMENTING AND SETTING OF SIGN.
050693,002143: 00,2764 40155 DCS MPAC
050694,002144: 00,2765 52155 DXCH MPAC
050695,002145: 00,2766 24136 INCR DVSIGN
050696,002146:
050697,002147: 00,2767 45017 /MPAC+ CS ONE # INITIALIZE NEAR-ONE SWITCH.
050698,002148: 00,2770 54140 TS MAXDVSW
050699,002149:
050700,002150: 00,2771 40154 CS MPAC # CHECK POSSIBLE OVERFLOW.
050701,002151: 00,2772 60133 AD BUF2 # UNNORMALIZED INPUT DIVISOR.
050702,002152: 00,2773 10000 CCS A
050703,002153: 00,2774 13004 TCF DDVCALL # NOT NEAR-ONE
050704,002154: 00,2775 12777 TCF +2 # +0 IS JUST POSSIBLE
050705,002155: 00,2776 12405 TCF DVOVF # NO HOPE
050706,002156: 00,2777 54140 TS MAXDVSW # SIGNAL POSSIBLE NEAR-ONE CASE
050707,002157: 00,3000 40155 CS MPAC +1 # SEE IF DIVISION CAN BE DONE
050708,002158: 00,3001 60134 AD BUF2 +1
050709,002159: 00,3002 00006 EXTEND
050710,002160: 00,3003 62405 BZMF DVOVF
050711,002161:
050712,002162: 00,3004 52155 DDVCALL DXCH MPAC # CALL PRE-DIVIDE NORMALIZATION.
050713,002163: 00,3005 50137 INDEX DVNORMCT
050714,002164: 00,3006 12565 TCF MAXTEST
050715,002165:
|
|
Page 1174
|
050717,002167: 00,3007 32506 SLOPELO DEC .8324
050718,002168:
050719,002169: 00,3010 56002 VECAGREE XCH Q # SAVE Q IN A
050720,002170: 00,3011 52155 DXCH MPAC
050721,002171: 00,3012 07513 TC ALSIGNAG # SIGNAGREE MPAC
050722,002172: 00,3013 52155 DXCH MPAC
050723,002173: 00,3014 52160 DXCH MPAC +3
050724,002174: 00,3015 07513 TC ALSIGNAG # SIGN AGREE MPAC +3
050725,002175: 00,3016 52160 DXCH MPAC +3
050726,002176: 00,3017 52162 DXCH MPAC +5
050727,002177: 00,3020 07513 TC ALSIGNAG # SIGNAGREE MPAC +5
050728,002178: 00,3021 52162 DXCH MPAC +5
050729,002179: 00,3022 00000 TC A
050730,002180:
|
|
Page 1175
|
050732,002182: # THE FOLLOWING ROUTINE EXECUTES THE UNIT INSTRUCTION, WHICH TAKES THE UNIT OF THE VECTOR IN MPAC.
050733,002183:
050734,002184: 00,3023 03010 UNIT TC VECAGREE # FORCE SIGN AGREEMENT IN VECTOR
050735,002185: 00,3024 07501 TC MPACVBUF # SAVE ARGUMENT IN VBUF
050736,002186: 00,3025 35021 CAF ZERO # MUST SENSE OVERFLOW IN FOLLOWING DOT.
050737,002187: 00,3026 56121 XCH OVFIND
050738,002188: 00,3027 54141 TS TEM1
050739,002189: 00,3030 03317 TC VSQSUB # DOT MPAC WITH ITSELF.
050740,002190: 00,3031 30141 CA TEM1
050741,002191: 00,3032 56121 XCH OVFIND
050742,002192: 00,3033 00006 EXTEND
050743,002193: 00,3034 13036 BZF +2
050744,002194: 00,3035 12405 TCF DVOVF
050745,002195: 00,3036 00006 EXTEND
050746,002196: 00,3037 30155 DCA MPAC # LEAVE THE SQUARE OF THE LENGTH OF THE
050747,002197: 00,3040 50120 INDEX FIXLOC # ARGUMENT IN LVSQUARE.
050748,002198: 00,3041 52043 DXCH LVSQUARE
050749,002199:
050750,002200: 00,3042 03343 TC SQRTSUB # GO TAKE THE NORMALIZED SQUARE ROOT.
050751,002201:
050752,002202: 00,3043 10154 CCS MPAC # CHECK FOR UNIT OVERFLOW.
050753,002203: 00,3044 13051 TCF +5 # MPAC IS NOT LESS THAN .5 UNLESS
050754,002204: 00,3045 54001 TS L
050755,002205: 00,3046 50120 INDEX FIXLOC
050756,002206: 00,3047 52045 DXCH LV
050757,002207: 00,3050 12405 TCF DVOVF # INPUT TO SQRTSUB WAS 0.
050758,002208:
050759,002209: 00,3051 44333 CS FOURTEEN # SEE IF THE INPUT WAS SO SMALL THE THE
050760,002210: 00,3052 60135 AD MPTEMP # FIRST TWO REGISTERS OF THE SQUARE WERE 0
050761,002211: 00,3053 10000 CCS A
050762,002212: 00,3054 40000 COM # IF SO, SAVE THE NEGATIVE OF THE SHIFT
050763,002213: 00,3055 13133 TCF SMALL # COUNT -15D.
050764,002214:
050765,002215: 00,3056 13065 TCF LARGE # (THIS IS USUALLY THE CASE.)
050766,002216:
050767,002217: 00,3057 45025 CS THIRTEEN # IF THE SHIFT COUNT WAS EXACTLY 14, SET
050768,002218: 00,3060 54135 TS MPTEMP # THE PRE-DIVIDE NORM COUNT TO -13D.
050769,002219:
050770,002220: 00,3061 30154 CA MPAC # SHIFT THE LENGTH RIGHT 14 BEFORE STORING
050771,002221: 00,3062 54001 SMALL2 TS L # (SMALL EXITS TO THIS POINT).
050772,002222: 00,3063 35021 CAF ZERO
050773,002223: 00,3064 13112 TCF LARGE2 # GO TO STORE LENGTH AND PROCEED.
050774,002224:
050775,002225: 00,3065 10135 LARGE CCS MPTEMP # MOST ALL CASES COME HERE.
050776,002226: 00,3066 13074 TCF LARGE3 # SEE IF NO NORMALIZATION WAS REQUIRED BY
050777,002227: 00,3067 42024 CS SRDDV # SQRT, AND IF SO, SET UP FOR A SHIFT
050778,002228: 00,3070 54135 TS MPTEMP # RIGHT 1 BEFORE DIVIDING TO PRODUCE
050779,002229: 00,3071 00006 EXTEND # THE DESIRED HALF UNIT VECTOR.
050780,002230: 00,3072 30155 DCA MPAC
|
|
Page 1176
|
050782,002232: 00,3073 13112 TCF LARGE2
050783,002233:
|
|
Page 1177
|
050785,002235: 00,3074 40000 LARGE3 COM # LEAVE NEGATIVE OF SHIFT COUNT-1 FOR
050786,002236: 00,3075 54135 TS MPTEMP # PREDIVIDE LEFT SHIFT.
050787,002237:
050788,002238: 00,3076 40000 COM # PICK UP REQUIRED SHIFTING BIT TO UNNORM-
050789,002239: 00,3077 50000 INDEX A # ALIZE THE SQRT RESULT.
050790,002240: 00,3100 35002 CAF BIT14
050791,002241: 00,3101 54130 TS BUF
050792,002242: 00,3102 00006 EXTEND
050793,002243: 00,3103 70155 MP MPAC +1
050794,002244: 00,3104 56130 XCH BUF
050795,002245: 00,3105 00006 EXTEND # (UNNORMALIZE THE SQRT FOR LV).
050796,002246: 00,3106 70154 MP MPAC
050797,002247: 00,3107 56001 XCH L
050798,002248: 00,3110 60130 AD BUF
050799,002249: 00,3111 56001 XCH L
050800,002250:
050801,002251: 00,3112 50120 LARGE2 INDEX FIXLOC
050802,002252: 00,3113 52045 DXCH LV # LENGTH NOW STORED IN WORK AREA.
050803,002253:
050804,002254: 00,3114 45017 CS ONE
050805,002255: 00,3115 54140 TS MAXDVSW # NO MAXDV CASES IN UNIT.
050806,002256:
050807,002257: 00,3116 52123 DXCH VBUF # PREPARE X COMPONENT FOR DIVIDE, SETTING
050808,002258: 00,3117 52155 DXCH MPAC # LENGTH OF VECTOR AS DIVISOR IN BUF.
050809,002259: 00,3120 52131 DXCH BUF
050810,002260: 00,3121 03151 TC UNITDV
050811,002261:
050812,002262: 00,3122 52125 DXCH VBUF +2 # DO Y AND Z IN USUAL FASHION SO WE CAN
050813,002263: 00,3123 52155 DXCH MPAC # EXIT THROUGH VROTATEX.
050814,002264: 00,3124 52160 DXCH MPAC +3
050815,002265: 00,3125 03151 TC UNITDV
050816,002266:
050817,002267: 00,3126 52127 DXCH VBUF +4
050818,002268: 00,3127 52155 DXCH MPAC
050819,002269: 00,3130 52162 DXCH MPAC +5
050820,002270: 00,3131 03151 TC UNITDV
050821,002271: 00,3132 17367 TCF VROTATEX # AND EXIT.
050822,002272:
|
|
Page 1178
|
050824,002274: # IF THE LENGTH OF THE ARGUMENT VECTOR WAS LESS THAN 2(-28), EACH COMPONENT MUST BE SHIFTED LEFT AT LEAST
050825,002275: # 14 PLACES BEFORE THE DIVIDE. NOTE THAT IN THIS CASE, THE MAJOR PART OF EACH COMPONENT IS ZERO.
050826,002276:
050827,002277: 00,3133 54135 SMALL TS MPTEMP # NEGATIVE OF PRE-DIVIDE SHIFT COUNT.
050828,002278:
050829,002279: 00,3134 35021 CAF ZERO # SHIFT EACH COMPONENT LEFT 14.
050830,002280: 00,3135 56123 XCH VBUF +1
050831,002281: 00,3136 56122 XCH VBUF
050832,002282: 00,3137 56125 XCH VBUF +3
050833,002283: 00,3140 56124 XCH VBUF +2
050834,002284: 00,3141 56127 XCH VBUF +5
050835,002285: 00,3142 56126 XCH VBUF +4
050836,002286:
050837,002287: 00,3143 40135 CS MPTEMP
050838,002288: 00,3144 50000 INDEX A
050839,002289: 00,3145 35002 CAF BIT14
050840,002290: 00,3146 00006 EXTEND
050841,002291: 00,3147 70154 MP MPAC
050842,002292: 00,3150 13062 TCF SMALL2
050843,002293:
050844,002294: 00,3151 5025 THIRTEEN = OCT15
050845,002295: 00,3151 4333 FOURTEEN = OCT16
050846,002296: 00,3151 4333 OCT16 = R1D1
050847,002297:
|
|
Page 1179
|
050849,002299: # THE FOLLOWING ROUTINE SETS UP THE CALL TO THE DIVIDE ROUTINES.
050850,002300:
050851,002301: 00,3151 10154 UNITDV CCS MPAC # FORCE MPAC POSITIVE IF POSSIBLE, SETTING
050852,002302: 00,3152 13170 TCF UMPAC+ # DVSIGN ACCORDING TO THE SIGN OF MPAC
050853,002303: 00,3153 13155 TCF +2 # SINCE THE DIVISOR IS ALWAYS POSITIVE
050854,002304: 00,3154 13162 TCF UMPAC- # HERE.
050855,002305:
050856,002306: 00,3155 10155 CCS MPAC +1
050857,002307: 00,3156 13170 TCF UMPAC+
050858,002308: 00,3157 00002 TC Q # EXIT IMMEDIATELY ON ZERO.
050859,002309: 00,3160 13162 TCF UMPAC-
050860,002310: 00,3161 00002 TC Q
050861,002311:
050862,002312: 00,3162 45021 UMPAC- CS ZERO # IF NEGATIVE, SET -0 IN DVSIGN FOR FINAL
050863,002313: 00,3163 54136 TS DVSIGN # COMPLEMENT.
050864,002314: 00,3164 00006 EXTEND
050865,002315: 00,3165 40155 DCS MPAC # PICK UP ABSOLUTE VALUE OF ARG AND JUMP.
050866,002316: 00,3166 50135 INDEX MPTEMP
050867,002317: 00,3167 12564 TCF MAXTEST -1
050868,002318:
050869,002319: 00,3170 54136 UMPAC+ TS DVSIGN # SET DVSIGN FOR POSITIVE QUOTIENT.
050870,002320: 00,3171 52155 DXCH MPAC
050871,002321: 00,3172 50135 INDEX MPTEMP
050872,002322: 00,3173 12564 TCF MAXTEST -1
050873,002323:
|
|
Page 1180
|
050875,002325: # MISCELLANEOUS UNARY OPERATIONS.
050876,002326:
050877,002327: 00,3174 03300 DSQ TC DSQSUB # SQUARE THE DP CONTENTS OF MPAC.
050878,002328: 00,3175 16030 TCF DANZIG
050879,002329:
050880,002330: 00,3176 10163 ABVALABS CCS MODE # ABVAL OR ABS INSTRUCTION.
050881,002331: 00,3177 13226 TCF ABS # DO ABS ON SCALAR.
050882,002332: 00,3200 13226 TCF ABS
050883,002333:
050884,002334: 00,3201 03317 ABVAL TC VSQSUB # DOT MPAC WITH ITSELF.
050885,002335: 00,3202 22163 LXCH MODE # MODE IS NOW DP (L ZERO AFTER DAS).
050886,002336:
050887,002337: 00,3203 00006 EXTEND # STORE SQUARE OF LENGTH IN WORK AREA.
050888,002338: 00,3204 30155 DCA MPAC
050889,002339: 00,3205 50120 INDEX FIXLOC
050890,002340: 00,3206 52043 DXCH LVSQUARE
050891,002341:
|
|
Page 1181
|
050893,002343: # PROGRAM DESCRIPTION - SUBROUTINE SQRT
050894,002344:
050895,002345: # FUNCTIONAL DESCRIPTION - DOUBLE PRECISION SQUARE ROOT ROUTINE
050896,002346: # THIS PROGRAM TAKES THE SQUARE ROOT OF THE 27 OR 28 MOST SIGNIFICANT BITS IN THE TRIPLE PRECISION SET OF
050897,002347: # NUMBERS - MPAC, MPAC+1, AND MPAC+2. THE ROOT IS RETURNED DOUBLE PRECISION IN MPAC AND MPAC+1.
050898,002348:
050899,002349: # WARNING - THIS SUBROUTINE USES A TRIPLE PRECISION INPUT. THE PROGRAMMER MUST ASSURE THE CONTENTS OF MPAC+2
050900,002350: # ESPECIALLY IF THE CONTENTS OF MPAC IS SMALL OR ZERO. FOR DETAILS SEE STG MEMO NO.949.
050901,002351:
050902,002352: # CALLING SEQUENCE - IN INTERPRETIVE MODE I.E., FOLLOWING TC INTPRET,SQRT NO ADDRESS IS ALLOWED
050903,002353: # INPUT SCALING THE BINARY POINT IS ASSUMED TO THE RIGHT OF BIT 15. THE ANSWER IS RETURNED WITH THE SAME SCALING
050904,002354:
050905,002355: # SUBROUTINES - GENSCR, MPACSHR, SQRTSUB, ABORT
050906,002356:
050907,002357: # ABORT EXIT MODE - ABORTS ON NEGATIVE INPUT -1.2X10E-4 (77775 OCTAL) OR LESS.
050908,002358: # DISPLAYS ERROR CODE 1302
050909,002359: # TC ABORT
050910,002360: # OCT 1302
050911,002361:
050912,002362: # DEBRIS - LOCATIONS BUF, MPTEMP, ADDRWD ARE USED
050913,002363:
050914,002364: 00,3207 03343 SQRT TC SQRTSUB # TAKE THE SQUARE ROOT OF MPAC.
050915,002365: 00,3210 10135 CCS MPTEMP # RETURNED NORMALIZED SQUARE ROOT. SEE IF
050916,002366: 00,3211 13213 TCF +2 # ANY UN-NORMALIZATION REQUIRED AND EXIT
050917,002367: 00,3212 16030 TCF DANZIG # IF NOT.
050918,002368:
050919,002369: 00,3213 63730 AD NEG12 # A RIGHT SHIFT OF MORE THAN 13 COULD BE
050920,002370: 00,3214 00006 EXTEND # REQUIRED IF INPUT WAS ZERO IN MPAC,+1.
050921,002371: 00,3215 63221 BZMF SQRTSHFT # GOES HERE IN MOST CASES.
050922,002372: 00,3216 22007 ZL # IF A LONG SHIFT IS REQUIRED, GO TO
050923,002373: 00,3217 22116 LXCH ADDRWD # GENERAL RIGHT SHIFT ROUTINES.
050924,002374: 00,3220 12303 TCF GENSCR +4 # ADDRWD WAS ZERO TO PREVENT ROUND.
050925,002375:
050926,002376: 00,3221 50135 SQRTSHFT INDEX MPTEMP # SELECT SHIFTING BIT AND EXIT THROUGH
050927,002377: 00,3222 35001 CAF BIT15 # SHIFT ROUTINES.
050928,002378: 00,3223 54135 TS MPTEMP
050929,002379: 00,3224 35021 CAF ZERO # TO ZERO MPAC +2 IN THE PROCESS.
050930,002380: 00,3225 12036 TCF MPACSHR +3
050931,002381:
050932,002382: 00,3226 06672 ABS TC BRANCH # TEST SIGN OF MPAC AND COMPLEMENT IF
050933,002383: 00,3227 16030 TCF DANZIG
050934,002384: 00,3230 16030 TCF DANZIG
050935,002385: 00,3231 17637 TCF COMP
050936,002386:
|
|
Page 1182
|
050938,002388: 00,3232 45015 VDEF CS FOUR # VECTOR DEFINE - ESSENTIALLY TREATS
050939,002389: 00,3233 26166 ADS PUSHLOC # SCALAR IN MPAC AS X COMPONENT, PUSHES UP
050940,002390: 00,3234 00006 EXTEND # FOR Y AND THEN AGAIN FOR Z.
050941,002391: 00,3235 50000 INDEX A
050942,002392: 00,3236 30003 DCA 2
050943,002393: 00,3237 52160 DXCH MPAC +3
050944,002394: 00,3240 00006 EXTEND
050945,002395: 00,3241 50166 INDEX PUSHLOC
050946,002396: 00,3242 30001 DCA 0
050947,002397: 00,3243 52162 DXCH MPAC +5
050948,002398: 00,3244 16470 TCF VMODE # MODE IS NON VECTOR.
050949,002399:
050950,002400: 00,3245 03317 VSQ TC VSQSUB # DOT MPAC WITH ITSELF.
050951,002401: 00,3246 17301 TCF DMODE # MODE IS NOW DP.
050952,002402:
050953,002403: 00,3247 00006 PUSH EXTEND # PUSH DOWN MPAC LEAVING IT LOADED.
050954,002404: 00,3250 30155 DCA MPAC
050955,002405: 00,3251 50166 INDEX PUSHLOC # PUSH DOWN FIRST TWO REGISTERS IN EACH
050956,002406: 00,3252 52001 DXCH 0
050957,002407:
050958,002408: 00,3253 50163 INDEX MODE # INCREMENT PUSHDOWN POINTER.
050959,002409: 00,3254 36213 CAF NO.WDS
050960,002410: 00,3255 26166 ADS PUSHLOC
050961,002411:
050962,002412: 00,3256 10163 CCS MODE
050963,002413: 00,3257 13272 TCF TPUSH # PUSH DOWN MPAC +2.
050964,002414: 00,3260 16030 TCF DANZIG # DONE FOR DP.
050965,002415:
050966,002416: 00,3261 00006 EXTEND # ON VECTOR, PUSH DOWN Y AND Z COMPONENTS.
050967,002417: 00,3262 30160 DCA MPAC +3
050968,002418: 00,3263 50166 INDEX PUSHLOC
050969,002419: 00,3264 51775 DXCH 0 -4
050970,002420: 00,3265 00006 EXTEND
050971,002421: 00,3266 30162 DCA MPAC +5
050972,002422: 00,3267 50166 INDEX PUSHLOC
050973,002423: 00,3270 51777 DXCH 0 -2
050974,002424: 00,3271 16030 TCF DANZIG
050975,002425:
050976,002426: 00,3272 30156 TPUSH CA MPAC +2
050977,002427: 00,3273 16523 TCF ENDTPUSH +2
050978,002428:
050979,002429: 00,3274 50120 RVQ INDEX FIXLOC # RVQ - RETURN IVA QPRET.
050980,002430: 00,3275 30052 CA QPRET
050981,002431: 00,3276 54117 TS POLISH
050982,002432: 00,3277 16621 TCF GOTO +4 # (ASSUME QPRET POINTS TO FIXED ONLY.)
050983,002433:
|
|
Page 1183
|
050985,002435: # THE FOLLOWING SUBROUTINES ARE USED IN SQUARING MPAC, IN BOTH THE SCALAR AND VECTOR SENSE. THEY ARE
050986,002436: # SPECIAL CASES OF DMPSUB AND DOTSUB, PUT IN TO SAVE SOME TIME.
050987,002437:
050988,002438: 00,3300 30155 DSQSUB CA MPAC +1 # SQUARES THE SCALAR CONTENTS OF MPAC.
050989,002439: 00,3301 00006 EXTEND
050990,002440: 00,3302 70000 SQUARE
050991,002441: 00,3303 54156 TS MPAC +2
050992,002442: 00,3304 35021 CAF ZERO # FORM 2(CROSS TERM).
050993,002443: 00,3305 56155 XCH MPAC +1
050994,002444: 00,3306 00006 EXTEND
050995,002445: 00,3307 70154 MP MPAC
050996,002446: 00,3310 20001 DDOUBL # AND MAYBE OVERFLOW.
050997,002447: 00,3311 20156 DAS MPAC +1 # AND SET A TO NET OVERFLOW.
050998,002448: 00,3312 56154 XCH MPAC
050999,002449: 00,3313 00006 EXTEND
051000,002450: 00,3314 70000 SQUARE
051001,002451: 00,3315 20155 DAS MPAC
051002,002452: 00,3316 00002 TC Q
051003,002453:
051004,002454: 00,3317 00006 VSQSUB EXTEND # DOTS THE VECTOR IN MPAC WITH ITSELF.
051005,002455: 00,3320 22137 QXCH DOTRET
051006,002456: 00,3321 03300 TC DSQSUB # SQUARE THE X COMPONENT.
051007,002457: 00,3322 52160 DXCH MPAC +3
051008,002458: 00,3323 52155 DXCH MPAC
051009,002459: 00,3324 52131 DXCH BUF # SO WE CAN END IN DOTSUB.
051010,002460: 00,3325 30156 CA MPAC +2
051011,002461: 00,3326 54132 TS BUF +2
051012,002462:
051013,002463: 00,3327 03300 TC DSQSUB # SQUARE Y COMPONENT.
051014,002464: 00,3330 52156 DXCH MPAC +1
051015,002465: 00,3331 20132 DAS BUF +1
051016,002466: 00,3332 60154 AD MPAC
051017,002467: 00,3333 60130 AD BUF
051018,002468: 00,3334 54130 TS BUF
051019,002469: 00,3335 13337 TCF +2
051020,002470: 00,3336 54121 TS OVFIND # IF OVERFLOW.
051021,002471:
051022,002472: 00,3337 52162 DXCH MPAC +5
051023,002473: 00,3340 52155 DXCH MPAC
051024,002474: 00,3341 03300 TC DSQSUB # SQUARE Z COMPONENT.
051025,002475: 00,3342 17154 TCF ENDDOT # END AS IN DOTSUB.
051026,002476:
|
|
Page 1184
|
051028,002478: # DOUBLE PRECISION SQUARE ROOT ROUTINE. TAKE THE SQUARE ROOT OF THE TRIPLE PRECISION (MPAC +2 USED ONLY
051029,002479: # IN NORMALIZATION) CONTENTS OF MPAC AND LEAVE THE NORMALIZED RESULT IN MPAC (C(MPAC) GREATER THAN OR EQUAL TO
051030,002480: # .5). THE RIGHT SHIFT COUNT (TO UNNORMALIZE) IS LEFT IN MPTEMP.
051031,002481:
051032,002482: 00,3343 35021 SQRTSUB CAF ZERO # START BY ZEROING RIGHT SHIFT COUNT.
051033,002483: 00,3344 54135 TS MPTEMP
051034,002484:
051035,002485: 00,3345 10154 CCS MPAC # CHECK FOR POSITIVE ARGUMENT, SHIFTING
051036,002486: 00,3346 13404 TCF SMPAC+ # FIRST SIGNIFICANT MPAC REGISTER INTO
051037,002487: 00,3347 13351 TCF +2 # MPAC ITSELF.
051038,002488: 00,3350 13373 TCF SQRTNEG # SEE IF MAG OF ARGUMENT LESS THAN 10(-4).
051039,002489:
051040,002490: 00,3351 56156 XCH MPAC +2 # MPAC IS ZERO - SHIFT LEFT 14.
051041,002491: 00,3352 56155 XCH MPAC +1
051042,002492: 00,3353 54154 TS MPAC
051043,002493: 00,3354 35023 CAF SEVEN # AUGMENT RIGHT SHIFT COUNTER.
051044,002494: 00,3355 54135 TS MPTEMP
051045,002495:
051046,002496: 00,3356 10154 CCS MPAC # SEE IF MPAC NOW PNZ.
051047,002497: 00,3357 13404 TCF SMPAC+
051048,002498: 00,3360 13362 TCF +2
051049,002499: 00,3361 13376 TCF ZEROANS # NEGATIVE BUT LESS THAN 10(-4) IN MAG.
051050,002500:
051051,002501: 00,3362 56155 XCH MPAC +1 # ZERO - SHIFT LEFT 14 AGAIN.
051052,002502: 00,3363 54154 TS MPAC
051053,002503: 00,3364 35023 CAF SEVEN # AUGMENT RIGHT SHIFT COUNTER.
051054,002504: 00,3365 26135 ADS MPTEMP
051055,002505:
051056,002506: 00,3366 10154 CCS MPAC
051057,002507: 00,3367 13404 TCF SMPAC+
051058,002508: 00,3370 00002 TC Q # SQRT(0) = 0.
051059,002509: 00,3371 13376 TCF ZEROANS
051060,002510: 00,3372 13452 TCF FIXROOT # DO NOT LEAVE SQRTSUB WITH -0 IN MPAC.
051061,002511:
051062,002512: 00,3373 10000 SQRTNEG CCS A # ARGUMENT IS NEGATIVE, BUT SEE IF SIGN-
051063,002513: 00,3374 13402 TCF SQRTABRT # CORRECTED ARGUMENT IS LESS THAN 10(-4)
051064,002514:
051065,002515: 00,3375 10155 CCS MPAC +1 # IN MAGNITUDE. IF SO, CALL ANSWER ZERO.
051066,002516: 00,3376 35021 ZEROANS CAF ZERO # FORCE ANSWER TO ZERO HERE.
051067,002517: 00,3377 13452 TCF FIXROOT
051068,002518: 00,3400 13402 TCF SQRTABRT
051069,002519: 00,3401 13452 TCF FIXROOT
051070,002520:
051071,002521: 00,3402 05721 SQRTABRT TC POODOO
051072,002522: 00,3403 01302 OCT 1302
051073,002523:
|
|
Page 1185
|
051075,002525: 00,3404 62444 SMPAC+ AD -1/2+2 # SEE IF ARGUMENT GREATER THAN OR EQUAL TO
051076,002526: 00,3405 00006 EXTEND # .5.
051077,002527: 00,3406 63455 BZMF SRTEST # IF SO, SEE IF LESS THAN .25.
051078,002528:
051079,002529: 00,3407 52155 DXCH MPAC # WE WILL TAKE THE SQUARE ROOT OF MPAC/2.
051080,002530: 00,3410 22021 LXCH SR # SHIFT RIGHT 1 AND GO TO THE SQRT ROUTINE
051081,002531: 00,3411 00006 EXTEND
051082,002532: 00,3412 75002 MP HALF
051083,002533: 00,3413 52155 DXCH MPAC
051084,002534: 00,3414 56021 XCH SR
051085,002535: 00,3415 26155 ADS MPAC +1 # GUARANTEED NO OVERFLOW.
051086,002536:
051087,002537: 00,3416 32314 ARGHI CAF SLOPEHI # ARGUMENT BETWEEN .25 AND .5. GET A
051088,002538: 00,3417 00006 EXTEND # LINEAR APPROXIMATION FOR THIS RANGE.
051089,002539: 00,3420 70154 MP MPAC
051090,002540: 00,3421 62566 AD BIASHI # X0/2 = (MPAC/2)(SLOPEHI) + BIASHI/2.
051091,002541:
051092,002542: 00,3422 54130 +4 TS BUF # X0/2 (ARGLO ENTERS HERE).
051093,002543: 00,3423 30154 CA MPAC # SINGLE-PRECISION THROUGHOUT.
051094,002544: 00,3424 22007 ZL
051095,002545: 00,3425 00006 EXTEND
051096,002546: 00,3426 10130 DV BUF # (MPAC/2)/(X0/2)
051097,002547: 00,3427 00006 EXTEND
051098,002548: 00,3430 75002 MP HALF
051099,002549: 00,3431 26130 ADS BUF # X1 = X0/2 + .5(MPAC/2)/(X0/2).
051100,002550:
051101,002551: 00,3432 00006 EXTEND
051102,002552: 00,3433 75002 MP HALF # FORM UP X1/2.
051103,002553: 00,3434 52155 DXCH MPAC # SAVE AND BRING OUT ARGUMENT.
051104,002554: 00,3435 00006 EXTEND # TAKE DP QUOTIENT WITH X1.
051105,002555: 00,3436 10130 DV BUF
051106,002556: 00,3437 54131 TS BUF +1 # SAVE MAJOR PART OF QUOTIENT.
051107,002557: 00,3440 35021 CAF ZERO # FORM MINOR PART OF QUOTIENT USING
051108,002558: 00,3441 56001 XCH L # (REMAINDER,0).
051109,002559: 00,3442 00006 EXTEND
051110,002560: 00,3443 10130 DV BUF
051111,002561: 00,3444 54001 TS L # IN PREPARATION FOR DAS.
051112,002562: 00,3445 30131 CA BUF +1
051113,002563: 00,3446 20155 DAS MPAC # X2 = X1/2 + (MPAC/2)X1
051114,002564:
051115,002565: 00,3447 00006 EXTEND # OVERFLOWS IF ARG. NEAR POSMAX.
051116,002566: 00,3450 13454 BZF TCQBNK00
051117,002567: 00,3451 34777 CAF POSMAX
051118,002568: 00,3452 54154 FIXROOT TS MPAC
051119,002569: 00,3453 54155 TS MPAC +1
051120,002570: 00,3454 00002 TCQBNK00 TC Q # RETURN TO CALLER TO UNNORMALIZE, ETC.
051121,002571:
|
|
Page 1186
|
051123,002573: 00,3455 65003 SRTEST AD QUARTER # ARGUMENT WAS LESS THAN .5, SEE IF LESS
051124,002574: 00,3456 00006 EXTEND # THAN .25.
051125,002575: 00,3457 63501 BZMF SQRTNORM # IF SO, BEGIN NORMALIZATION.
051126,002576:
051127,002577: 00,3460 52155 DXCH MPAC # IF BETWEEN .5 AND .25, SHIFT RIGHT 1 AND
051128,002578: 00,3461 22021 LXCH SR # START AT ARGLO.
051129,002579: 00,3462 00006 EXTEND
051130,002580: 00,3463 75002 MP HALF
051131,002581: 00,3464 52155 DXCH MPAC
051132,002582: 00,3465 56021 XCH SR
051133,002583: 00,3466 26155 ADS MPAC +1 # NO OVERFLOW.
051134,002584:
051135,002585: 00,3467 33007 ARGLO CAF SLOPELO # (NORMALIZED) ARGUMENT BETWEEN .125 AND
051136,002586: 00,3470 00006 EXTEND # .25
051137,002587: 00,3471 70154 MP MPAC
051138,002588: 00,3472 62270 AD BIASLO
051139,002589: 00,3473 13422 TCF ARGHI +4 # BEGIN SQUARE ROOT.
051140,002590:
051141,002591: 00,3474 00006 SQRTNM2 EXTEND # SHIFT LEFT 2 AND INCREMENT RIGHT SHIFT
051142,002592: 00,3475 30156 DCA MPAC +1 # COUNT (FOR TERMINAL UNNORMALIZATION).
051143,002593: 00,3476 20156 DAS MPAC +1
051144,002594: 00,3477 60154 AD MPAC
051145,002595: 00,3500 26154 ADS MPAC # (NO OVERFLOW).
051146,002596:
051147,002597: 00,3501 24135 SQRTNORM INCR MPTEMP # FIRST TIME THROUGH, JUST SHIFT LEFT 1
051148,002598: 00,3502 00006 EXTEND # (PUTS IN EFFECTIVE RIGHT SHIFT SINCE
051149,002599: 00,3503 30156 DCA MPAC +1 # WE WANT MPAC/2).
051150,002600: 00,3504 20156 DAS MPAC +1
051151,002601: 00,3505 60154 AD MPAC
051152,002602: 00,3506 26154 ADS MPAC # (AGAIN NO OVERFLOW).
051153,002603: 00,3507 60000 DOUBLE
051154,002604: 00,3510 54022 TS CYL
051155,002605:
051156,002606: 00,3511 10022 NORMTEST CCS CYL # SEE IF ARGUMENT NOW NORMALIZED AT
051157,002607: 00,3512 10022 CCS CYL # GREATER THAN .125.
051158,002608: 00,3513 13474 TCF SQRTNM2 # NO - SHIFT LEFT 2 MORE AND TRY AGAIN.
051159,002609: 00,3514 13416 TCF ARGHI # YES - NOW BETWEEN .5 AND .25.
051160,002610: 00,3515 13467 TCF ARGLO # ARGUMENT NOW BETWEEN .25 AND .125.
051161,002611:
|
|
Page 1187
|
051163,002613: # TRIGONOMETRIC FUNCTION PACKAGE.
051164,002614: # THE FOLLOWING TRIGONOMETRIC FUNCTIONS ARE AVAILABLE AS INTERPRETIVE OPERATIONS:
051165,002615: # 1. SIN COMPUTES (1/2)SINE(2 PI MPAC).
051166,002616: # 2. COS COMPUTES (1/2)COSINE(2 PI MPAC).
051167,002617: # 3. ASIN COMPUTES (1/2PI)ARCSINE(2 MPAC).
051168,002618: # 4. ACOS COMPUTES (1/2PI)ARCCOSINE(2 MPAC).
051169,002619:
051170,002620: # SIN-ASIN AND COS-ACOS ARE MUTUALLY INVERSE, IE SIN(ASIN(X)) = X.
051171,002621:
051172,002622: 00,3516 06672 COSINE TC BRANCH # FINDS COSINE USING THE IDENTITY
051173,002623: 00,3517 13522 TCF +3 # COS(X) = SIN(PI/2 - ABS(X)).
051174,002624: 00,3520 13525 TCF PRESINE
051175,002625: 00,3521 13525 TCF PRESINE
051176,002626:
051177,002627: 00,3522 00006 +3 EXTEND
051178,002628: 00,3523 40155 DCS MPAC
051179,002629: 00,3524 52155 DXCH MPAC
051180,002630:
051181,002631: 00,3525 35003 PRESINE CAF QUARTER # PI/2 SCALED.
051182,002632: 00,3526 26154 ADS MPAC
051183,002633:
051184,002634: 00,3527 52155 SINE DXCH MPAC # DOUBLE ARGUMENT.
051185,002635: 00,3530 20001 DDOUBL
051186,002636: 00,3531 54000 OVSK # SEE IF OVERFLOW PRESENT.
051187,002637: 00,3532 13535 TCF +3 # IF NOT, ARGUMENT OK AS IS.
051188,002638:
051189,002639: 00,3533 00006 EXTEND # IF SO, WE LOST (OR GAINED) PI, SO
051190,002640: 00,3534 40001 DCOM # COMPLEMENT MPAC USING THE IDENTITY
051191,002641: # SIN(X-(+)PI) = SIN(-X).
051192,002642: 00,3535 52155 +3 DXCH MPAC
051193,002643: 00,3536 30154 CA MPAC # SEE IF ARGUMENT GREATER THAN .5 IN
051194,002644: 00,3537 60000 DOUBLE # MAGNITUDE. IF SO, REDUCE IT TO LESS THAN
051195,002645: 00,3540 54001 TS L # .5 (+-PI/2 SCALED) AS FOLLOWS:
051196,002646: 00,3541 13552 TCF SN1
051197,002647:
051198,002648: 00,3542 50000 INDEX A # IF POSITIVE, FORM PI - X, IF NEGATIVE
051199,002649: 00,3543 35001 CAF NEG1/2 +1 # USE -PI -X.
051200,002650: 00,3544 60000 DOUBLE
051201,002651: 00,3545 00006 EXTEND
051202,002652: 00,3546 60154 SU MPAC # GUARANTEED NO OVERFLOW.
051203,002653: 00,3547 54154 TS MPAC
051204,002654: 00,3550 40155 CS MPAC +1
051205,002655: 00,3551 54155 TS MPAC +1
051206,002656:
|
|
Page 1188
|
051208,002658: 00,3552 00006 SN1 EXTEND # SET UP TO EVALUATE HASTINGS POLYNOMIAL
051209,002659: 00,3553 30155 DCA MPAC
051210,002660: 00,3554 52134 DXCH BUF2
051211,002661: 00,3555 03300 TC DSQSUB # SQUARE MPAC.
051212,002662:
051213,002663: 00,3556 07171 TC POLY # EVALUATE FOURTH ORDER POLYNOMIAL.
051214,002664: 00,3557 00003 DEC 3 B-14
051215,002665: 00,3560 14441 37325 2DEC +.3926990796
051216,002666: 00,3562 53250 60764 2DEC -.6459637111
051217,002667: 00,3564 12146 21276 2DEC +.318758717
051218,002668: 00,3566 75466 71471 2DEC -.074780249
051219,002669: 00,3570 00236 32757 2DEC +.009694988
051220,002670:
051221,002671: 00,3572 32470 CAF LBUF2 # MULTIPLY BY ARGUMENT AND SHIFT LEFT 2.
051222,002672: 00,3573 07055 TC DMPSUB -1
051223,002673:
051224,002674: 00,3574 00006 EXTEND
051225,002675: 00,3575 30156 DCA MPAC +1
051226,002676: 00,3576 20156 DAS MPAC +1
051227,002677: 00,3577 60154 AD MPAC
051228,002678: 00,3600 26154 ADS MPAC # NEITHER SHIFT OVERFLOWS.
051229,002679: 00,3601 00006 EXTEND
051230,002680: 00,3602 30156 DCA MPAC +1
051231,002681: 00,3603 20156 DAS MPAC +1
051232,002682: 00,3604 60154 AD MPAC
051233,002683: 00,3605 26154 ADS MPAC
051234,002684: 00,3606 16030 TCF DANZIG
051235,002685:
|
|
Page 1189
|
051237,002687: # ARCSIN/ARCCOS ROUTINE.
051238,002688:
051239,002689: 00,3607 33630 ARCSIN CAF LASINEX # COMPUTE ARCSIN BY USING THE IDENTITY
051240,002690: 00,3610 13612 TCF +2 # ARCSIN(X) = PI/2 - ARCCOS(X).
051241,002691:
051242,002692: 00,3611 33712 ARCCOS CAF LDANZIG # (EXITS IMMEDIATELY).
051243,002693: 00,3612 54136 TS ESCAPE
051244,002694: 00,3613 06672 TC BRANCH # TEST SIGN OF INPUT.
051245,002695: 00,3614 13624 TCF ACOSST # START IMMEDIATELY IF POSITIVE.
051246,002696: 00,3615 13726 TCF ACOSZERO # ARCCOS(0) = PI/2 = .25.
051247,002697: 00,3616 00006 EXTEND # IF NEGATIVE, USE THE IDENTITY
051248,002698: 00,3617 40155 DCS MPAC # ARCCOS(X) = PI - ARCCOS(-X), FORCING
051249,002699: 00,3620 52155 DXCH MPAC # ARGUMENT POSITIVE.
051250,002700: 00,3621 33731 CAF TCSUBTR # SET EXIT TO DO ABOVE BEFORE
051251,002701: 00,3622 56136 XCH ESCAPE # ARCSIN/ARCCOS CONSIDERATIONS.
051252,002702: 00,3623 54137 TS ESCAPE2
051253,002703:
051254,002704: 00,3624 45002 ACOSST CS HALF # TEST MAGNITUDE OF INPUT.
051255,002705: 00,3625 60154 AD MPAC
051256,002706: 00,3626 10000 CCS A
051257,002707: 00,3627 13720 TCF ACOSOVF # THIS IS PROBABLY AN OVERFLOW CASE.
051258,002708:
051259,002709: 00,3630 13706 LASINEX TCF ASINEX
051260,002710:
051261,002711: 00,3631 13641 TCF ACOSST2 # NO OVERFLOW - PROCEED.
051262,002712:
051263,002713: 00,3632 10155 CCS MPAC +1 # IF MAJOR PART IS .5, CALL ANSWER 0
051264,002714: 00,3633 35021 CAF ZERO # UNLESS MINOR PART NEGATIVE.
051265,002715: 00,3634 13636 TCF ACOS=0
051266,002716:
051267,002717: 00,3635 13641 TCF ACOSST2
051268,002718:
051269,002719: 00,3636 54155 ACOS=0 TS MPAC +1
051270,002720: 00,3637 54154 TS MPAC
051271,002721: 00,3640 00136 TC ESCAPE
051272,002722:
051273,002723: 00,3641 00006 ACOSST2 EXTEND # NOW THAT ARGUMENT IS IN PROPER RANGE,
051274,002724: 00,3642 40155 DCS MPAC # BEGIN COMPUTATION. USE HASTINGS
051275,002725: 00,3643 65002 AD HALF # APPROXIMATION ARCCOS(X) = SQRT(1-X)P(X)
051276,002726: 00,3644 52155 DXCH MPAC # IN A SCALED VERSION WHERE P(X) IS A
051277,002727: 00,3645 52134 DXCH BUF2 # SEVENTH ORDER POLYNOMIAL.
051278,002728:
051279,002729: 00,3646 03343 TC SQRTSUB # RETURNS WITH NORMALIZED SQUARE ROOT.
051280,002730:
051281,002731: 00,3647 10135 CCS MPTEMP # SEE IF UN-NORMALIZATION REQUIRED.
051282,002732: 00,3650 13713 TCF ACOSSHR # IF SO.
051283,002733:
|
|
Page 1190
|
051285,002735: 00,3651 52155 ACOS3 DXCH MPAC # SET UP FOR POLYNOMIAL EVALUATION.
051286,002736: 00,3652 52134 DXCH BUF2
051287,002737: 00,3653 52155 DXCH MPAC
051288,002738:
051289,002739: 00,3654 07171 TC POLY
051290,002740: 00,3655 00006 DEC 6 B-14
051291,002741: 00,3656 13240 23630 2DEC +.353553385 # COEFFICIENTS ARE C 2(+I)/PISQRT(2) WHERE
051292,002742: 00,3660 74721 47775 2DEC* -.0483017006 B+1* # I
051293,002743: 00,3662 02440 20237 2DEC* +.0200273085 B+2* # WHERE C STANDS FOR ORIGINAL COEFFS.
051294,002744: 00,3664 75067 70742 2DEC* -.0112931863 B+3*
051295,002745: 00,3666 03436 26756 2DEC* +.00695311612 B+4*
051296,002746: 00,3670 74037 57640 2DEC* -.00384617957 B+5*
051297,002747: 00,3672 03046 07143 2DEC* +.001501297736 B+6*
051298,002748: 00,3674 76654 42244 2DEC* -.000284160334 B+7*
051299,002749:
051300,002750: 00,3676 32470 CAF LBUF2 # DO FINAL MULTIPLY AND GO TO ANY
051301,002751: 00,3677 07055 TC DMPSUB -1 # EPILOGUE SEQUENCES.
051302,002752: 00,3700 00136 TC ESCAPE
051303,002753:
051304,002754: 00,3701 00006 SUBTR EXTEND # EPILOGUE FOR NEGATIVE INPUTS TO ARCCOS.
051305,002755: 00,3702 40155 DCS MPAC
051306,002756: 00,3703 65002 AD HALF # FORMS PI - ARCCOS(-X) = ARCCOS(X).
051307,002757: 00,3704 52155 DXCH MPAC
051308,002758: 00,3705 00137 TC ESCAPE2 # GO TO POSSIBLE ARCSIN EPILOGUE.
051309,002759:
051310,002760: 00,3706 00006 ASINEX EXTEND
051311,002761: 00,3707 40155 DCS MPAC # ARCSIN EPILOGUE - GET ARCSIN(X)
051312,002762: 00,3710 65003 AD QUARTER # = PI/2 - ARCCOS(X).
051313,002763: 00,3711 52155 DXCH MPAC
051314,002764: 00,3712 16030 LDANZIG TCF DANZIG
051315,002765:
|
|
Page 1191
|
051317,002767: 00,3713 50000 ACOSSHR INDEX A # THE SHIFT RIGHT IS LESS THAN 14 SINCE
051318,002768: 00,3714 35002 CAF BIT14 # THE INPUT WAS NON-ZERO DP.
051319,002769: 00,3715 54135 TS MPTEMP
051320,002770: 00,3716 02073 TC VSHRRND # DP SHIFT RIGHT AND ROUND.
051321,002771: 00,3717 13651 TCF ACOS3 # PROCEED.
051322,002772:
051323,002773: 00,3720 00006 ACOSOVF EXTEND # IF MAJOR PART WAS ONLY 1 MORE THAN .5,
051324,002774: 00,3721 13636 BZF ACOS=0 # CALL ANSWER ZERO.
051325,002775:
051326,002776: 00,3722 05650 ACOSABRT TC ALARM # IF OVERFLOW, CALL ANSWER ZERO BUT
051327,002777: 00,3723 01301 OCT 1301 # SOUND AN ALARM.
051328,002778:
051329,002779: 00,3724 35021 CAF ZERO
051330,002780: 00,3725 13636 TCF ACOS=0
051331,002781:
051332,002782: 00,3726 35003 ACOSZERO CAF QUARTER # ACOS(0) = PI/2.
051333,002783: 00,3727 13637 TCF ACOS=0 +1 # SET MPAC AND EXIT VIA ESCAPE.
051334,002784:
051335,002785: 00,3730 77763 NEG12 DEC -12 B-14
051336,002786: 00,3731 13701 TCSUBTR TCF SUBTR
051337,002787:
|
|
Page 1192
|
051339,002789: # THE FOLLOWING INSTRUCTIONS ARE AVAILABLE FOR SETTING, MODIFYING, AND BRANCHING ON INDEX REGISTERS:
051340,002790:
051341,002791: # 1. AXT ADDRESS TO INDEX TRUE.
051342,002792: # 1. AXC ADDRESS TO INDEX COMPLEMENTED.
051343,002793: # 3. LXA LOAD INDEX FROM ERASABLE.
051344,002794: # 4. LXC LOAD INDEX COMPLEMENTED FROM ERASABLE.
051345,002795: # 5. SXA STORE INDEX IN ERASABLE.
051346,002796: # 6. XCHX EXCHANGE INDEX REIGSTER WITH ERASABLE.
051347,002797:
051348,002798: # 7. INCR INCREMENT INDEX REGISTER.
051349,002799: # 8. XAD ERASABLE ADD TO INDEX REGISTER.
051350,002800: # 9. XSU ERASABLE SUBTRACT FROM INDEX REGISTER.
051351,002801: # 10. TIX BRANCH ON INDEX REGISTER AND DECREMENT.
051352,002802:
051353,002803: 01,2371 BANK 01
051354,002804:
051355,002805: 01,2371 COUNT 01/INTER
051356,002806:
051357,002807: 01,2371 02466 AXT TC TAGSUB # SELECT APPROPRIATE INDEX REGISTER.
051358,002808: 01,2372 30117 CA POLISH
051359,002809: 01,2373 50130 XSTORE INDEX INDEXLOC # CONTAINS C(FIXLOC) OR C(FIXLOC)+1.
051360,002810: 01,2374 54046 TS X1
051361,002811: 01,2375 16030 TCF DANZIG
051362,002812:
051363,002813: 01,2376 02466 AXC TC TAGSUB
051364,002814: 01,2377 40117 CS POLISH
051365,002815: 01,2400 02373 TC XSTORE
051366,002816:
051367,002817: 01,2401 02454 LXA TC 15ADRERS # LOAD INDEX REGISTER FROM ERASABLE.
051368,002818: 01,2402 50117 INDEX POLISH
051369,002819: 01,2403 30000 CA 0
051370,002820: 01,2404 12373 TCF XSTORE
051371,002821:
051372,002822: 01,2405 02454 LXC TC 15ADRERS # LOAD NDX REG FROM ERASABLE COMPLEMENTED.
051373,002823: 01,2406 50117 INDEX POLISH
051374,002824: 01,2407 40000 CS 0
051375,002825: 01,2410 12373 TCF XSTORE
051376,002826:
051377,002827: 01,2411 02454 SXA TC 15ADRERS # STORE INDEX REGISTER IN ERASABLE.
051378,002828: 01,2412 50130 INDEX INDEXLOC
051379,002829: 01,2413 30046 CA X1
051380,002830: 01,2414 50117 MSTORE1 INDEX POLISH
051381,002831: 01,2415 54000 TS 0
051382,002832: 01,2416 16030 TCF DANZIG
051383,002833:
|
|
Page 1193
|
051385,002835: 01,2417 02454 XCHX TC 15ADRERS # EXCHANGE INDEX REGISTER WITH ERASABLE.
051386,002836: 01,2420 50117 INDEX POLISH
051387,002837: 01,2421 30000 CA 0
051388,002838: 01,2422 50130 INDEX INDEXLOC
051389,002839: 01,2423 56046 XCH X1
051390,002840: 01,2424 12414 TCF MSTORE1
051391,002841:
051392,002842: 01,2425 02454 XAD TC 15ADRERS # ADD ERASABLE TO INDEX REGISTER.
051393,002843: 01,2426 50117 INDEX POLISH
051394,002844: 01,2427 30000 CA 0
051395,002845: 01,2430 50130 XAD2 INDEX INDEXLOC
051396,002846: 01,2431 26046 ADS X1 # IGNORING OVERFLOWS.
051397,002847: 01,2432 16030 TCF DANZIG
051398,002848:
051399,002849: 01,2433 02466 INCR TC TAGSUB # INCREMENT INDEX REGISTER.
051400,002850: 01,2434 30117 CA POLISH
051401,002851: 01,2435 12430 TCF XAD2
051402,002852:
051403,002853: 01,2436 02454 XSU TC 15ADRERS # SUBTRACT ERASABLE FROM INDEX REGISTER.
051404,002854: 01,2437 50117 INDEX POLISH
051405,002855: 01,2440 40000 CS 0
051406,002856: 01,2441 12430 TCF XAD2
051407,002857:
051408,002858: 01,2442 02466 TIX TC TAGSUB # BRANCH AND DECREMENT ON INDEX.
051409,002859: 01,2443 50130 INDEX INDEXLOC
051410,002860: 01,2444 40050 CS S1
051411,002861: 01,2445 50130 INDEX INDEXLOC
051412,002862: 01,2446 60046 AD X1
051413,002863: 01,2447 00006 EXTEND # NO OPERATION IF DECREMENTED INDEX IS
051414,002864: 01,2450 66030 BZMF DANZIG # NEGATIVE OR ZERO.
051415,002865:
051416,002866: 01,2451 50130 DOTIXBR INDEX INDEXLOC
051417,002867: 01,2452 56046 XCH X1 # IGNORING OVERFLOWS.
051418,002868:
051419,002869: 01,2453 16615 TCF GOTO # DO THE BRANCH USING THE CADR IN POLISH.
051420,002870:
|
|
Page 1194
|
051422,002872: # SUBROUTINE TO CONVERT AN ERASABLE ADDRESS (11 BITS) TO AN EBANK SETTING AND SUBADDRESS.
051423,002873:
051424,002874: 01,2454 40117 15ADRERS CS POLISH
051425,002875: 01,2455 65034 AD DEC45
051426,002876: 01,2456 10000 CCS A # DOES THE ADDRESS POINT TO THE WORK AREA?
051427,002877: 01,2457 30120 CA FIXLOC # YES. ADD FIXLOC. EBANK OK AS IS.
051428,002878: 01,2460 12465 TCF +5
051429,002879:
051430,002880: 01,2461 35051 CA OCT1400 # NO. SET EBANK & MAKE UP SUBADDRESS.
051431,002881: 01,2462 56117 XCH POLISH
051432,002882: 01,2463 54003 TS EBANK
051433,002883: 01,2464 74373 MASK LOW8
051434,002884: 01,2465 26117 +5 ADS POLISH # FALL INTO TAGSUB, AND RETURN VIA Q.
051435,002885:
051436,002886: # SUBROUTINE WHICH SETS THE ADDRESS OF THE SPECIFIED INDEX IN INDEXLOC. (ACTUALLY, THE ADDRESS -38D.)
051437,002887:
051438,002888: 01,2466 30120 TAGSUB CA FIXLOC
051439,002889: 01,2467 54130 TS INDEXLOC
051440,002890:
051441,002891: 01,2470 10020 CCS CYR # BIT 15 SPECIFIES INDEX.
051442,002892: 01,2471 24130 INCR INDEXLOC # 0 MEANS USE X2.
051443,002893: 01,2472 00002 TC Q
051444,002894: 01,2473 00002 TC Q # 1 FOR X1.
051445,002895:
|
|
Page 1195
|
051447,002897: # MISCELLANEOUS OPERATION CODES WITH DIRECT ADDRESSES. INCLUDED HERE ARE:
051448,002898: # 1. ITA STORE QPRET (RETURN ADDRESS) IN ERASABLE.
051449,002899: # 2. CALL CALL A SUBROUTINE, LEAVING RETURN IN QPRET.
051450,002900: # 3. RTB RETURN TO BASIC LANGUAGE AT THE GIVEN ADDRESS.
051451,002901: # 4. BHIZ BRANCH IF THE HIGHORDER OF MPAC IS ZERO (SINGLE PRECISION).
051452,002902: # 5. BOV BRANCH ON OVERFLOW.
051453,002903: # 6. GOTO SIMPLE SEQUENCE CHANGE.
051454,002904:
051455,002905: 01,2474 10020 RTB/BHIZ CCS CYR
051456,002906: 01,2475 30117 RTB CA POLISH
051457,002907: 01,2476 04665 TC SWCALL -1 # SO A "TC Q" FROM ROUTINE LEADS TO DANZIG
051458,002908:
051459,002909: 01,2477 10154 BHIZ CCS MPAC
051460,002910: 01,2500 16030 TCF DANZIG
051461,002911: 01,2501 16615 TCF GOTO
051462,002912: 01,2502 16030 TCF DANZIG
051463,002913: 01,2503 16615 TCF GOTO
051464,002914:
051465,002915: 01,2504 10121 BOV(B) CCS OVFIND # BRANCH ON OVERFLOW TO BASIC OR INTERP.
051466,002916: 01,2505 12507 TCF +2
051467,002917: 01,2506 16030 TCF DANZIG
051468,002918: 01,2507 54121 TS OVFIND
051469,002919: 01,2510 10020 CCS CYR
051470,002920: 01,2511 12475 TCF RTB # IF BASIC.
051471,002921: 01,2512 00360 B5TOBB OCT 360
051472,002922: 01,2513 16615 TCF GOTO
051473,002923:
|
|
Page 1196
|
051475,002925: 01,2514 10020 BZE/GOTO CCS CYR # SEE WHICH OP-CODE IS DESIRED.
051476,002926: 01,2515 06672 TC BRANCH # DO BZE.
051477,002927: 01,2516 16030 TCF DANZIG
051478,002928: 01,2517 16615 TCF GOTO # DO GOTO.
051479,002929: 01,2520 16030 TCF DANZIG
051480,002930:
051481,002931: 01,2521 10020 BPL/BMN CCS CYR
051482,002932: 01,2522 12530 TCF BPL
051483,002933: 01,2523 12000 5B10 DEC 5 B+10 # SHIFTS OP CODE IN SWITCH INSTRUCTION ADR
051484,002934:
051485,002935: 01,2524 06672 TC BRANCH # DO BMN.
051486,002936: 01,2525 16030 TCF DANZIG
051487,002937: 01,2526 16030 TCF DANZIG
051488,002938: 01,2527 16615 TCF GOTO # ONLY IF NNZ.
051489,002939:
051490,002940: 01,2530 06672 BPL TC BRANCH
051491,002941: 01,2531 16615 TCF GOTO # IF POSITIVE OR ZERO.
051492,002942: 01,2532 16615 TCF GOTO
051493,002943: 01,2533 16030 TCF DANZIG
051494,002944:
051495,002945: 01,2534 10020 CALL/ITA CCS CYR
051496,002946: 01,2535 16607 TCF CALL
051497,002947:
051498,002948: 01,2536 05741 TC CCSHOLE
051499,002949: 01,2537 02454 TC 15ADRERS # STORE QPRET. (TAGSUB AFTER 15ADRERS IS
051500,002950: 01,2540 50120 INDEX FIXLOC # SLOW IN THIS CASE, BUT SAVES STORAGE.)
051501,002951: 01,2541 30052 CA QPRET
051502,002952: 01,2542 12414 TCF MSTORE1
051503,002953:
|
|
Page 1197
|
051505,002955: # THE FOLLOWING OPERATIONS ARE AVAILABLE FOR ALTERING AND TESTING INTERPRETIVE SWITCHES:
051506,002956: # 00 BONSET SET A SWITCH AND DO A GOTO IF IT WAS ON.
051507,002957: # 01 SETGO SET A SWITCH AND DO A GOTO.
051508,002958: # 02 BOFSET SET A SWITCH AND DOA GOTO IF IT WAS OFF
051509,002959: # 03 SET SET A SWITCH.
051510,002960: # 04 BONINV INVERT A SWITCH AND BRANCH IF IT WAS ON.
051511,002961: # 05 INVGO INVERT A SWITCH AND DO A GOTO.
051512,002962: # 06 BOFINV INVERT A SWITCH AND BRANCH IF IT WAS OFF
051513,002963: # 07 INVERT INVERT A SWITCH.
051514,002964: # 10 BONCLR CLEAR A SWITCH AND BRANCH IF IT WAS ON.
051515,002965: # 11 CLRGO CLEAR A SWITCH AND DO A GOTO.
051516,002966: # 12 BOFCLR CLEAR A SWITCH AND BRANCH IF IT WAS OFF.
051517,002967: # 13 CLEAR CLEAR A SWITCH.
051518,002968: # 14 BON BRANCH IF A SWITCH WAS ON.
051519,002969: # 16 BOFF BRANCH IF A SWITCH WAS OFF.
051520,002970: # THE ADDRESS SUPPLIED WITH THE SWITCH INSTRUCTION IS INTERPRETED AS FOLLOWS:
051521,002971: # BITS 1-4 SWITCH BIT NUMBER (1-15).
051522,002972: # BITS 5-8 SWITCH OPERATION NUMBER.
051523,002973: # BITS 9- SWITCH WORD NUMBER (UP TO 64 SWITCH WORDS).
051524,002974: # THE ADDRESS ITSELF IS MADE UP BY THE YUL SYSTEM ASSEMBLER. THE BRANCH INSTRUCTIONS REQUIRE TWO
051525,002975: # ADDRESSES, THE SECOND TAKEN AS THE DIRECT (OR INDIRECT IF IN ERASABLE) ADDRESS OF THE BRANCH.
051526,002976:
051527,002977: 01,2543 35026 SWITCHES CAF LOW4 # LEAVE THE SWITCH BIT IN SWBIT.
051528,002978: 01,2544 70117 MASK POLISH
051529,002979: 01,2545 50000 INDEX A
051530,002980: 01,2546 35001 CAF BIT15 # (NUMBER FROM LEFT TO RIGHT.)
051531,002981: 01,2547 54131 TS SWBIT
051532,002982:
051533,002983: 01,2550 35011 CAF BIT7 # LEAVE THE SWITCH NUMBER IN SWWORD.
051534,002984: 01,2551 00006 EXTEND
051535,002985: 01,2552 70117 MP POLISH
051536,002986: 01,2553 54130 TS SWWORD
051537,002987:
051538,002988: 01,2554 00004 INHINT # DURING SWITCH CHANGE SO RUPT CAN USE TOO
051539,002989: 01,2555 50000 INDEX A # LEAVE THE SWITCH WORD ITSELF IN L.
051540,002990: 01,2556 30074 CA STATE
051541,002991: 01,2557 54002 TS Q # Q WILL BE USED AS A CHANNEL.
|
|
Page 1198
|
051543,002993: 01,2560 35005 CAF BIT11
051544,002994: 01,2561 00006 EXTEND # DISPATCH SWITCH BIT OPERATION AS IN BITS
051545,002995: 01,2562 70117 MP POLISH # 7-8 OF POLISH.
051546,002996: 01,2563 72623 MASK B3TOB4 # GETS 4X2-BIT CODE.
051547,002997: 01,2564 50000 INDEX A
051548,002998: 01,2565 12566 TCF +1
051549,002999:
051550,003000: 01,2566 30131 +1 CA SWBIT # 00 - SET SWITCH IN QUESTION.
051551,003001: 01,2567 00006 EXTEND
051552,003002: 01,2570 04002 ROR QCHAN
051553,003003: 01,2571 12600 TCF SWSTORE
051554,003004:
051555,003005: 01,2572 30131 +5 CA SWBIT # 01 - INVERT SWITCH.
051556,003006: 01,2573 00006 EXTEND
051557,003007: 01,2574 06002 RXOR QCHAN
051558,003008: 01,2575 12600 TCF SWSTORE
051559,003009:
051560,003010: 01,2576 40131 +9D CS SWBIT # 10 - CLEAR.
051561,003011: 01,2577 70002 MASK Q
051562,003012: 01,2600 50130 SWSTORE INDEX SWWORD
051563,003013: 01,2601 54074 TS STATE # NEW SWITCH WORD.
051564,003014:
|
|
Page 1199
|
051566,003016: 01,2602 00003 +13D RELINT # 11 - NOOP.
051567,003017: 01,2603 35003 CAF BIT13
051568,003018: 01,2604 00006 EXTEND # DISPATCH SEQUCE CHANGING OR BRANCHING
051569,003019: 01,2605 70117 MP POLISH # CODE.
051570,003020: 01,2606 72623 MASK B3TOB4
051571,003021: 01,2607 50000 INDEX A
051572,003022: 01,2610 12611 TCF +1 # ORIGINALLY STORED IN BITS 5-6.
051573,003023:
051574,003024: 01,2611 40002 +1 CS Q # 00 - BRANCH IF ON.
051575,003025: 01,2612 70131 TEST MASK SWBIT
051576,003026: 01,2613 10000 CCS A
051577,003027: 01,2614 12624 TCF SWSKIP
051578,003028:
051579,003029: 01,2615 16664 +5 TCF SWBRANCH # 01 - GO TO.
051580,003030:
051581,003031: 01,2616 12624 TCF SWSKIP # HERE ONLY ON BIT 15.
051582,003032:
051583,003033: 01,2617 05741 TC CCSHOLE
051584,003034: 01,2620 05741 TC CCSHOLE
051585,003035:
051586,003036: 01,2621 30002 +9D CA Q # 10 - BRANCH IF OFF.
051587,003037: 01,2622 12612 TCF TEST
051588,003038:
051589,003039: 01,2623 00014 B3TOB4 OCT 0014
051590,003040: 01,2624 24164 SWSKIP INCR LOC
051591,003041:
051592,003042: 01,2625 01,2543 SW/ EQUALS SWITCHES
051593,003043:
051594,003044: 01,2625 16030 +13D TCF DANZIG # 11 - NOOP.
End of include-file INTERPRETER.agc. Parent file is MAIN.agc