aboutsummaryrefslogtreecommitdiff
path: root/modexpng_fpga_model.py
blob: 325f544b86001dad18e2f19038b957e55d9b6c0c (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
#!/usr/bin/python3
#
#
# ModExpNG core math model.
#
#
# Copyright (c) 2019, NORDUnet A/S
# All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are
# met:
# - Redistributions of source code must retain the above copyright notice,
#   this list of conditions and the following disclaimer.
#
# - Redistributions in binary form must reproduce the above copyright
#   notice, this list of conditions and the following disclaimer in the
#   documentation and/or other materials provided with the distribution.
#
# - Neither the name of the NORDUnet nor the names of its contributors may
#   be used to endorse or promote products derived from this software
#   without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS
# IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
# TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
# PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
# HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED
# TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
# LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
# NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
# SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#


# -------
# Imports
#--------

import sys
import importlib
from enum import Enum, auto


# --------------
# Model Settings
# --------------

# length of public key
KEY_LENGTH = 1024

# how many parallel multipliers to use
NUM_MULTS  = 8


# ---------------
# Internal Values
# ---------------

# half of key length
_KEY_LENGTH_HALF = KEY_LENGTH // 2

# width of internal math pipeline
_WORD_WIDTH     = 16
_WORD_WIDTH_EXT = 18

_WORD_MASK     = 2 ** _WORD_WIDTH     - 1
_WORD_MASK_EXT = 2 ** _WORD_WIDTH_EXT - 1
_CARRY_MASK    = _WORD_MASK ^ _WORD_MASK_EXT

# folder with test vector scripts
_VECTOR_PATH = "/vector"

# name of test vector class
_VECTOR_CLASS = "Vector"


# ------------------
# Debugging Settings
# ------------------
DUMP_LADDER_INDEX      = -1     # at which ladder step to print debug vector
DUMP_VECTORS           = False  # print entire debug vector components
DUMP_INDICES           = False  # print indices of words at MAC inputs
DUMP_MACS_INPUTS       = False  # print MAC input words
DUMP_MACS_CLEARING     = False  # print MAC clearing bitmaps
DUMP_MACS_ACCUMULATION = False  # print MAC accumulators contents
DUMP_MULT_PARTS        = False  # print multiplication output parts
DUMP_RECOMBINATION     = False  # print recombination internals
DUMP_REDUCTION         = False  # print reduction internals
FORCE_OVERFLOW         = False  # force rarely seen internal overflow situation to verify how its handler works
DUMP_PROGRESS_FACTOR   = 16     # once per how many ladder steps to update progress indicator

#
# Multi-Precision Integer
#
class ModExpNG_Operand():

    def __init__(self, number, length, words = None):

        if words is None:

            # length must be divisible by word width
            if (length % _WORD_WIDTH) > 0:
                raise Exception("Bad number length!")

            self._init_from_number(number, length)

        else:

            # length must match words count
            if len(words) != length:
                raise Exception("Bad words count!")

            self._init_from_words(words, length)

    def format_verilog_concat(self, name):

        for i in range(len(self.words)):
            if i > 0:
                if (i % 4) == 0: print("")
                else:            print(" ", end='')
            print("%s[%3d] = 18'h%05x;" % (name, i, self.words[i]), end='')
        print("")

    def _init_from_words(self, words, count):

        for i in range(count):

            # word must not exceed 18 bits
            if words[i] >= (2 ** (_WORD_WIDTH_EXT)):
                raise Exception("Word is too large!")

        self.words = list(words)

    def _init_from_number(self, number, length):

        num_hexchars_per_word = _WORD_WIDTH // 4
        num_hexchars_total = length // num_hexchars_per_word

        value_hex = format(number, 'x')

        # value must not be larger than specified, but it can be smaller, so
        # we may need to prepend it with zeroes
        if len(value_hex) > num_hexchars_total:
            raise Exception("Number is too large!")
        else:
            while len(value_hex) < num_hexchars_total:
                value_hex = "0" + value_hex

        # create empty list
        self.words = list()

        # fill in words
        while len(value_hex) > 0:
            value_hex_part = value_hex[-num_hexchars_per_word:]
            value_hex = value_hex[:-num_hexchars_per_word]
            self.words.append(int(value_hex_part, 16))

    def number(self):
        ret = 0
        shift = 0
        for word in self.words:
            ret += word << shift
            shift += _WORD_WIDTH
        return ret

    def _get_half(self, part):
        num_words = len(self.words)
        num_words_half = num_words // 2
        if not part: return ModExpNG_Operand(None, num_words_half, self.words[:num_words_half])
        else:        return ModExpNG_Operand(None, num_words_half, self.words[num_words_half:])

    def lower_half(self):
        return self._get_half(False)

    def upper_half(self):
        return self._get_half(True)

#
# Test Vector
#
class ModExpNG_TestVector():

    def __init__(self):

        # format target filename
        filename = "vector_" + str(KEY_LENGTH) + "_randomized"

        # add ./vector to import search path
        sys.path.insert(1, sys.path[0] + _VECTOR_PATH)

        # import from filename
        vector_module = importlib.import_module(filename)

        # get vector class
        vector_class = getattr(vector_module, _VECTOR_CLASS)

        # instantiate vector class
        vector_inst = vector_class()

        # obtain parts of vector
        self.m        = ModExpNG_Operand(vector_inst.m,         KEY_LENGTH)
        self.n        = ModExpNG_Operand(vector_inst.n,         KEY_LENGTH)
        self.d        = ModExpNG_Operand(vector_inst.d,         KEY_LENGTH)
        self.p        = ModExpNG_Operand(vector_inst.p,        _KEY_LENGTH_HALF)
        self.q        = ModExpNG_Operand(vector_inst.q,        _KEY_LENGTH_HALF)
        self.dp       = ModExpNG_Operand(vector_inst.dp,       _KEY_LENGTH_HALF)
        self.dq       = ModExpNG_Operand(vector_inst.dq,       _KEY_LENGTH_HALF)
        self.qinv     = ModExpNG_Operand(vector_inst.qinv,     _KEY_LENGTH_HALF)
        self.n_factor = ModExpNG_Operand(vector_inst.n_factor,  KEY_LENGTH)
        self.p_factor = ModExpNG_Operand(vector_inst.p_factor, _KEY_LENGTH_HALF)
        self.q_factor = ModExpNG_Operand(vector_inst.q_factor, _KEY_LENGTH_HALF)
        self.n_coeff  = ModExpNG_Operand(vector_inst.n_coeff,   KEY_LENGTH      + _WORD_WIDTH)
        self.p_coeff  = ModExpNG_Operand(vector_inst.p_coeff,  _KEY_LENGTH_HALF + _WORD_WIDTH)
        self.q_coeff  = ModExpNG_Operand(vector_inst.q_coeff,  _KEY_LENGTH_HALF + _WORD_WIDTH)
        self.x        = ModExpNG_Operand(vector_inst.x,         KEY_LENGTH)
        self.y        = ModExpNG_Operand(vector_inst.y,         KEY_LENGTH)

class ModExpNG_WideBankEnum(Enum):
    A   = auto()
    B   = auto()
    C   = auto()
    D   = auto()
    E   = auto()
    N   = auto()
    L   = auto()
    H   = auto()

class ModExpNG_NarrowBankEnum(Enum):
    A       = auto()
    B       = auto()
    C       = auto()
    D       = auto()
    E       = auto()
    N_COEFF = auto()
    I       = auto()

class ModExpNG_CoreInputEnum(Enum):
    M        = auto()

    N        = auto()
    P        = auto()
    Q        = auto()

    N_COEFF  = auto()
    P_COEFF  = auto()
    Q_COEFF  = auto()

    N_FACTOR = auto()
    P_FACTOR = auto()
    Q_FACTOR = auto()

    X        = auto()
    Y        = auto()

    QINV     = auto()

class ModExpNG_CoreOutputEnum(Enum):
    XM = auto()
    YM = auto()
    S  = auto()

class ModExpNG_WideBank():

    def __init__(self):
        self.a = None
        self.b = None
        self.c = None
        self.d = None
        self.e = None
        self.n = None
        self.l = None
        self.h = None

    def _get_value(self, sel):
        if   sel == ModExpNG_WideBankEnum.A:   return self.a
        elif sel == ModExpNG_WideBankEnum.B:   return self.b
        elif sel == ModExpNG_WideBankEnum.C:   return self.c
        elif sel == ModExpNG_WideBankEnum.D:   return self.d
        elif sel == ModExpNG_WideBankEnum.E:   return self.e
        elif sel == ModExpNG_WideBankEnum.N:   return self.n
        elif sel == ModExpNG_WideBankEnum.L:   return self.l
        elif sel == ModExpNG_WideBankEnum.H:   return self.h
        else: raise Exception("ModExpNG_WideBank._get_value(): Invalid selector!")

    def _set_value(self, sel, value):
        if   sel == ModExpNG_WideBankEnum.A:   self.a   = value
        elif sel == ModExpNG_WideBankEnum.B:   self.b   = value
        elif sel == ModExpNG_WideBankEnum.C:   self.c   = value
        elif sel == ModExpNG_WideBankEnum.D:   self.d   = value
        elif sel == ModExpNG_WideBankEnum.E:   self.e   = value
        elif sel == ModExpNG_WideBankEnum.N:   self.n   = value
        elif sel == ModExpNG_WideBankEnum.L:   self.l   = value
        elif sel == ModExpNG_WideBankEnum.H:   self.h   = value
        else: raise Exception("ModExpNG_WideBank._set_value(): Invalid selector!")

class ModExpNG_NarrowBank():

    def __init__(self, i):
        self.a       = None
        self.b       = None
        self.c       = None
        self.d       = None
        self.e       = None
        self.n_coeff = None
        self.i       = i

    def _get_value(self, sel):
        if   sel == ModExpNG_NarrowBankEnum.A:       return self.a
        elif sel == ModExpNG_NarrowBankEnum.B:       return self.b
        elif sel == ModExpNG_NarrowBankEnum.C:       return self.c
        elif sel == ModExpNG_NarrowBankEnum.D:       return self.d
        elif sel == ModExpNG_NarrowBankEnum.E:       return self.e
        elif sel == ModExpNG_NarrowBankEnum.N_COEFF: return self.n_coeff
        elif sel == ModExpNG_NarrowBankEnum.I:       return self.i
        else: raise Exception("ModExpNG_NarrowBank._get_value(): Invalid selector!")

    def _set_value(self, sel, value):
        if   sel == ModExpNG_NarrowBankEnum.A:       self.a       = value
        elif sel == ModExpNG_NarrowBankEnum.B:       self.b       = value
        elif sel == ModExpNG_NarrowBankEnum.C:       self.c       = value
        elif sel == ModExpNG_NarrowBankEnum.D:       self.d       = value
        elif sel == ModExpNG_NarrowBankEnum.E:       self.e       = value
        elif sel == ModExpNG_NarrowBankEnum.N_COEFF: self.n_coeff = value
        else: raise Exception("ModExpNG_NarrowBank._set_value(): Invalid selector!")

class ModExpNG_CoreInput():

    def __init__(self):
        self._m        = None

        self._n        = None
        self._p        = None
        self._q        = None

        self._n_coeff  = None
        self._p_coeff  = None
        self._q_coeff  = None

        self._n_factor = None
        self._p_factor = None
        self._q_factor = None

        self._x        = None
        self._y        = None

        self._qinv     = None

    def set_value(self, sel, value):
        if   sel == ModExpNG_CoreInputEnum.M:        self._m        = value

        elif sel == ModExpNG_CoreInputEnum.N:        self._n        = value
        elif sel == ModExpNG_CoreInputEnum.P:        self._p        = value
        elif sel == ModExpNG_CoreInputEnum.Q:        self._q        = value

        elif sel == ModExpNG_CoreInputEnum.N_COEFF:  self._n_coeff  = value
        elif sel == ModExpNG_CoreInputEnum.P_COEFF:  self._p_coeff  = value
        elif sel == ModExpNG_CoreInputEnum.Q_COEFF:  self._q_coeff  = value

        elif sel == ModExpNG_CoreInputEnum.N_FACTOR: self._n_factor = value
        elif sel == ModExpNG_CoreInputEnum.P_FACTOR: self._p_factor = value
        elif sel == ModExpNG_CoreInputEnum.Q_FACTOR: self._q_factor = value

        elif sel == ModExpNG_CoreInputEnum.X:        self._x        = value
        elif sel == ModExpNG_CoreInputEnum.Y:        self._y        = value

        elif sel == ModExpNG_CoreInputEnum.QINV:     self._qinv     = value

        else: raise Exception("ModExpNG_CoreInput.set_value(): invalid selector!")

    def _get_value(self, sel):
        if   sel == ModExpNG_CoreInputEnum.M:        return self._m

        elif sel == ModExpNG_CoreInputEnum.N:        return self._n
        elif sel == ModExpNG_CoreInputEnum.P:        return self._p
        elif sel == ModExpNG_CoreInputEnum.Q:        return self._q

        elif sel == ModExpNG_CoreInputEnum.N_COEFF:  return self._n_coeff
        elif sel == ModExpNG_CoreInputEnum.P_COEFF:  return self._p_coeff
        elif sel == ModExpNG_CoreInputEnum.Q_COEFF:  return self._q_coeff

        elif sel == ModExpNG_CoreInputEnum.N_FACTOR: return self._n_factor
        elif sel == ModExpNG_CoreInputEnum.P_FACTOR: return self._p_factor
        elif sel == ModExpNG_CoreInputEnum.Q_FACTOR: return self._q_factor

        elif sel == ModExpNG_CoreInputEnum.X:        return self._x
        elif sel == ModExpNG_CoreInputEnum.Y:        return self._y

        elif sel == ModExpNG_CoreInputEnum.QINV:     return self._qinv

        else: raise Exception("ModExpNG_CoreInput._get_value(): invalid selector!")

class ModExpNG_CoreOutput():

    def __init__(self):
        self._xm = None
        self._ym = None
        self._s  = None

    def _set_value(self, sel, value):
        if   sel == ModExpNG_CoreOutputEnum.XM: self._xm = value
        elif sel == ModExpNG_CoreOutputEnum.YM: self._ym = value
        elif sel == ModExpNG_CoreOutputEnum.S:  self._s  = value
        else: raise Exception("ModExpNG_CoreOutput._set_value(): invalid selector!")

    def get_value(self, sel):
        if   sel == ModExpNG_CoreOutputEnum.XM: return self._xm
        elif sel == ModExpNG_CoreOutputEnum.YM: return self._ym
        elif sel == ModExpNG_CoreOutputEnum.S:  return self._s
        else: raise Exception("ModExpNG_CoreOutput.get_value(): invalid selector!")

class ModExpNG_BanksPair():

    def __init__(self, i):
        self.wide = ModExpNG_WideBank()
        self.narrow = ModExpNG_NarrowBank(i)

    def _get_wide(self, sel):
        return self.wide._get_value(sel)

    def _get_narrow(self, sel):
        return self.narrow._get_value(sel)

    def _set_wide(self, sel, value):
        self.wide._set_value(sel, value)

    def _set_narrow(self, sel, value):
        self.narrow._set_value(sel, value)

class ModExpNG_BanksLadder():

    def __init__(self, i):
        self.ladder_x = ModExpNG_BanksPair(i)
        self.ladder_y = ModExpNG_BanksPair(i)

class ModExpNG_BanksCRT():

    def __init__(self, i):
        self.crt_x = ModExpNG_BanksLadder(i)
        self.crt_y = ModExpNG_BanksLadder(i)

class ModExpNG_PartRecombinator():

    def _bit_select(self, x, msb, lsb):
        y = 0
        for pos in range(lsb, msb+1):
            y |= (x & (1 << pos)) >> lsb
        return y

    def _flush_pipeline(self, dump):
        self.z0, self.y0, self.x0 = 0, 0, 0
        if dump and DUMP_RECOMBINATION:
            print("RCMB -> flush()")

    def _push_pipeline(self, part, dump):

        # split next part into 16-bit words
        z = self._bit_select(part, 46, 32)
        y = self._bit_select(part, 31, 16)
        x = self._bit_select(part, 15,  0)

        # shift to the right
        z1 = z
        y1 = y + self.z0
        x1 = x + self.y0 + (self.x0 >> _WORD_WIDTH) # IMPORTANT: This carry can be up to two bits wide!!

        # save lower 16 bits of the rightmost cell
        t = self.x0 & _WORD_MASK

        # update internal latches
        self.z0, self.y0, self.x0 = z1, y1, x1

        # dump
        if dump and DUMP_RECOMBINATION:
            print("RCMB -> push(): part = 0x%012x, word = 0x%04x" % (part, t))

        # done
        return t

    def recombine_square(self, parts, ab_num_words, dump):

        # empty results so far
        words_lsb = list()  # n words
        words_msb = list()  # n words

        # recombine the lower half (n parts)
        # the first tick produces null result, the last part
        # produces three words and needs two extra ticks
        self._flush_pipeline(dump)
        for i in range(ab_num_words + 1 + 2):
            next_part = parts[i] if i < ab_num_words else 0
            next_word = self._push_pipeline(next_part, dump)

            if i > 0:
                words_lsb.append(next_word)

        # recombine the upper half (n-1 parts)
        # the first tick produces null result
        self._flush_pipeline(dump)
        for i in range(ab_num_words + 1):
            next_part = parts[i + ab_num_words] if i < (ab_num_words - 1) else 0
            next_word = self._push_pipeline(next_part, dump)

            if i > 0:
                words_msb.append(next_word)

        # merge words
        words = list()

        # merge lower half
        for x in range(ab_num_words):
            next_word = words_lsb[x]
            words.append(next_word)

        # merge upper half adding the two overlapping words
        for x in range(ab_num_words):
            next_word = words_msb[x]
            if x < 2:
                next_word += words_lsb[x + ab_num_words]
            words.append(next_word)

        return words

    def recombine_triangle(self, parts, ab_num_words, dump):

        # empty result so far
        words_lsb = list()

        # recombine the lower half (n+1 parts)
        # the first tick produces null result, so we need n + 1 + 1 = n + 2
        # ticks total and should only save the result word during the last
        # n + 1 ticks
        self._flush_pipeline(dump)
        for i in range(ab_num_words + 2):

            next_part = parts[i] if i < (ab_num_words + 1) else 0
            next_word = self._push_pipeline(next_part, dump)

            if i > 0:
                words_lsb.append(next_word)

        return words_lsb

    def recombine_rectangle(self, parts, ab_num_words, dump):

        # empty result so far
        words_lsb = list()  # n words
        words_msb = list()  # n+1 words

        # recombine the lower half (n parts)
        # the first tick produces null result, the last part
        # produces three words and needs two extra ticks
        self._flush_pipeline(dump)
        for i in range(ab_num_words + 1 + 2):
            next_part = parts[i] if i < ab_num_words else 0
            next_word = self._push_pipeline(next_part, dump)

            if i > 0:
                words_lsb.append(next_word)

        # recombine the upper half (n parts)
        # the first tick produces null result, the last part
        # produces two words and needs an extra tick
        self._flush_pipeline(dump)
        for i in range(ab_num_words + 2):
            next_part = parts[i + ab_num_words] if i < ab_num_words else 0
            next_word = self._push_pipeline(next_part, dump)

            if i > 0:
                words_msb.append(next_word)

        # merge words
        words = list()

        # merge lower half
        for x in range(ab_num_words):
            next_word = words_lsb[x]
            words.append(next_word)

        # merge upper half adding the two overlapping words
        for x in range(ab_num_words + 1):
            next_word = words_msb[x]
            if x < 2:
                next_word += words_lsb[x + ab_num_words]
            words.append(next_word)

        return words

class ModExpNG_WordMultiplier():

    def __init__(self):

        self._macs = list()
        self._indices = list()

        self._mac_aux = list()
        self._index_aux = list()

        for x in range(NUM_MULTS):
            self._macs.append(0)
            self._indices.append(0)

        self._mac_aux.append(0)
        self._index_aux.append(0)

    def _clear_all_macs(self, t, col, dump):
        for x in range(NUM_MULTS):
            self._macs[x] = 0
        if dump and DUMP_MACS_CLEARING:
            print("t=%2d, col=%2d > clear > all" % (t, col))

    def _clear_one_mac(self, x, t, col, dump):
        self._macs[x] = 0
        if dump and DUMP_MACS_CLEARING:
            print("t=%2d, col=%2d > clear > x=%d" % (t, col, x))

    def _clear_mac_aux(self, t, col, dump):
        self._mac_aux[0] = 0
        if dump and DUMP_MACS_CLEARING:
            print("t= 0, col=%2d > clear > aux" % (col))

    def _update_one_mac(self, x, t, col, a, b, dump, need_aux=False):

        if a >= (2 ** _WORD_WIDTH_EXT):
            raise Exception("a > 0x3FFFF!")

        if b >= (2 ** _WORD_WIDTH):
            raise Exception("b > 0xFFFF!")

        p = a * b
        if dump and DUMP_MACS_INPUTS:
            if x == 0: print("t=%2d, col=%2d > b=%05x > " % (t, col, b), end='')
            if x > 0: print("; ", end='')
            print("MAC[%d]: a=%05x" % (x, a), end='')
            if x == (NUM_MULTS-1) and not need_aux: print("")

        self._macs[x] += p

    def _update_mac_aux(self, y, col, a, b, dump):

        if a >= (2 ** _WORD_WIDTH_EXT):
            raise Exception("a > 0x3FFFF!")

        if b >= (2 ** _WORD_WIDTH):
            raise Exception("b > 0xFFFF!")

        p = a * b
        if dump and DUMP_MACS_INPUTS:
            print("; AUX: a=%05x" % a)
            
        self._mac_aux[0] += p

    def _preset_indices(self, col):
        for x in range(len(self._indices)):
            self._indices[x] = col * len(self._indices) + x

    def _preset_index_aux(self, num_cols):
        self._index_aux[0] = num_cols * len(self._indices)

    def _dump_macs_helper(self, t, col, aux=False):
        print("t=%2d, col=%2d > "% (t, col), end='')
        for i in range(NUM_MULTS):
            if i > 0: print(" | ", end='')
            print("mac[%d]: 0x%012x" % (i, self._macs[i]), end='')
        if aux:
            print(" | mac_aux[ 0]: 0x%012x" % (self._mac_aux[0]), end='')
        print("")

    def _dump_macs(self, t, col):
        self._dump_macs_helper(t, col)

    def _dump_macs_with_aux(self, t, col):
        self._dump_macs_helper(t, col, True)

    def _dump_indices_helper(self, t, col, aux=False):
        print("t=%2d, col=%2d > indices:" % (t, col), end='')
        for i in range(NUM_MULTS):
            print(" %2d" % self._indices[i], end='')
        if aux:
            print(" %2d" % self._index_aux[0], end='')
        print("")

    def _dump_indices(self, t, col):
        self._dump_indices_helper(t, col)

    def _dump_indices_with_aux(self, t, col):
        self._dump_indices_helper(t, col, True)

    def _rotate_indices(self, num_words):
        for x in range(len(self._indices)):
            if self._indices[x] > 0:
                self._indices[x] -= 1
            else:
                self._indices[x] = num_words - 1

    def _rotate_index_aux(self):
        self._index_aux[0] -= 1

    def _mult_store_part(self, parts, time, column, part_index, mac_index, dump):
        parts[part_index] = self._macs[mac_index]
        if dump and DUMP_MULT_PARTS:
            print("t=%2d, col=%2d > parts[%2d]: mac[%d] = 0x%012x" %
                (time, column, part_index, mac_index, parts[part_index]))

    def _mult_store_part_aux(self, parts, time, column, part_index, dump):
        parts[part_index] = self._mac_aux[0]
        if dump and DUMP_MULT_PARTS:
            print("t=%2d, col=%2d > parts[%2d]: mac_aux[%d] = 0x%012x" %
                (time, column, part_index, 0, parts[part_index]))

    def multiply_square(self, a_wide, b_narrow, ab_num_words, dump=False):

        num_cols = ab_num_words // NUM_MULTS

        parts = list()
        for i in range(2 * ab_num_words - 1):
            parts.append(0)

        for col in range(num_cols):

            b_carry = 0

            for t in range(ab_num_words):

                # take care of indices
                if t == 0: self._preset_indices(col)
                else:      self._rotate_indices(ab_num_words)

                # take care of macs
                if t == 0:
                    self._clear_all_macs(t, col, dump)
                else:
                    t1 = t - 1
                    if (t1 // 8) == col:
                        self._clear_one_mac(t1 % NUM_MULTS, t, col, dump)

                # debug output
                if dump and DUMP_INDICES: self._dump_indices(t, col)

                # current b-word
                # multiplier's b-input is limited to 16-bit words, so we need to propagate
                # carries on the fly here, carry can be up to two bits
                bt = b_narrow.words[t] + b_carry
                b_carry = (bt & _CARRY_MASK) >> _WORD_WIDTH
                if dump and b_carry > 1:
                    print("Rare overflow case was detected and then successfully corrected.")
                bt &= _WORD_MASK

                # multiply by a-words
                for x in range(NUM_MULTS):
                    ax = a_wide.words[self._indices[x]]
                    self._update_one_mac(x, t, col, ax, bt, dump)

                    if t == (col * NUM_MULTS + x):
                        part_index = t
                        self._mult_store_part(parts, t, col, part_index, x, dump)

                # debug output
                if dump and DUMP_MACS_ACCUMULATION: self._dump_macs(t, col)

                # save the uppers part of product at end of column,
                # for the last column don't save the very last part
                if t == (ab_num_words - 1):
                    for x in range(NUM_MULTS):
                        if not (col == (num_cols - 1) and x == (NUM_MULTS - 1)):
                            part_index = ab_num_words + col * NUM_MULTS + x
                            self._mult_store_part(parts, t, col, part_index, x, dump)

        return parts

    def multiply_triangle(self, a_wide, b_narrow, ab_num_words, dump=False):

        num_cols = ab_num_words // NUM_MULTS

        parts = list()
        for i in range(ab_num_words + 1):
            parts.append(0)

        for col in range(num_cols):

            last_col = col == (num_cols - 1)

            for t in range(ab_num_words + 1):

                # take care of indices
                if t == 0: self._preset_indices(col)
                else:      self._rotate_indices(ab_num_words)

                # take care of auxilary index
                if last_col:
                    if t == 0: self._preset_index_aux(num_cols)
                    else:      self._rotate_index_aux()

                # take care of macs
                if t == 0: self._clear_all_macs(t, col, dump)

                # take care of auxilary mac
                if last_col:
                    if t == 0: self._clear_mac_aux(t, col, dump)

                # debug output
                if dump and DUMP_INDICES: self._dump_indices_with_aux(t, col)

                # current b-word
                bt = b_narrow.words[t]

                # multiply by a-words
                for x in range(NUM_MULTS):
                    ax = a_wide.words[self._indices[x]]
                    self._update_one_mac(x, t, col, ax, bt, dump, last_col)

                    if t == (col * NUM_MULTS + x):
                        part_index = t
                        self._mult_store_part(parts, t, col, part_index, x, dump)

                # aux multiplier
                if last_col:
                    ax = a_wide.words[self._index_aux[0]]
                    self._update_mac_aux(t, col, ax, bt, dump)

                    if t == ab_num_words:
                        part_index = t
                        self._mult_store_part_aux(parts, t, col, part_index, dump)

                # debug output
                if dump and DUMP_MACS_ACCUMULATION: self._dump_macs_with_aux(t, col)

                # shortcut
                if not last_col:
                    if t == (NUM_MULTS * (col + 1) - 1): break

        return parts

    def multiply_rectangle(self, a_wide, b_narrow, ab_num_words, dump=False):

        num_cols = ab_num_words // NUM_MULTS

        parts = list()
        for i in range(2 * ab_num_words):
            parts.append(0)

        for col in range(num_cols):

            for t in range(ab_num_words + 1):

                # take care of indices
                if t == 0: self._preset_indices(col)
                else:      self._rotate_indices(ab_num_words)

                # take care of macs
                if t == 0:
                    self._clear_all_macs(t, col, dump)
                else:
                    t1 = t - 1
                    if (t1 // 8) == col:
                        self._clear_one_mac(t1 % NUM_MULTS, t, col, dump)

                # debug output
                if dump and DUMP_INDICES: self._dump_indices(t, col)

                # current b-word
                bt = b_narrow.words[t]

                # multiply by a-words
                for x in range(NUM_MULTS):
                    ax = a_wide.words[self._indices[x]]
                    self._update_one_mac(x, t, col, ax, bt, dump)

                    # don't save one value for the very last time instant per column
                    if t < ab_num_words and t == (col * NUM_MULTS + x):
                        part_index = t
                        self._mult_store_part(parts, t, col, part_index, x, dump)

                # debug output
                if dump and DUMP_MACS_ACCUMULATION: self._dump_macs(t, col)

                # save the upper parts of product at end of column
                if t == ab_num_words:
                    for x in range(NUM_MULTS):
                        part_index = ab_num_words + col * NUM_MULTS + x
                        self._mult_store_part(parts, t, col, part_index, x, dump)

        return parts

class ModExpNG_LowlevelOperator():

    def _check_word(self, a):
        if a < 0 or a > _WORD_MASK:
            raise Exception("Word out of range!")

    def _check_carry_borrow(self, cb):
        if cb < 0 or cb > 1:
            raise Exception("Carry or borrow out of range!")

    def add_words(self, a, b, c_in):

        self._check_word(a)
        self._check_word(b)
        self._check_carry_borrow(c_in)

        sum = a + b + c_in

        sum_s = sum & _WORD_MASK
        sum_c = sum >> _WORD_WIDTH

        return (sum_c, sum_s)

    def sub_words(self, a, b, b_in):

        self._check_word(a)
        self._check_word(b)
        self._check_carry_borrow(b_in)

        dif = a - b - b_in

        if dif < 0:
            dif_b = 1
            dif_d = dif + 2 ** _WORD_WIDTH
        else:
            dif_b = 0
            dif_d = dif

        return (dif_b, dif_d)

class ModExpNG_Worker():

    def __init__(self):
        self.lowlevel     = ModExpNG_LowlevelOperator()
        self.multiplier   = ModExpNG_WordMultiplier()
        self.recombinator = ModExpNG_PartRecombinator()

    def serial_subtract_modular(self, a, b, n, ab_num_words):
        c_in = 0
        b_in = 0
        ab = list()
        ab_n = list()
        for x in range(ab_num_words):
            a_word = a.words[x]
            b_word = b.words[x]
            (b_out, d_out) = self.lowlevel.sub_words(a_word, b_word, b_in)
            (c_out, s_out) = self.lowlevel.add_words(d_out, n.words[x], c_in)
            ab.append(d_out)
            ab_n.append(s_out)
            (c_in, b_in) = (c_out, b_out)
        d = ab if not b_out else ab_n
        return ModExpNG_Operand(None, ab_num_words, d)

    def serial_add_uneven(self, a, b, ab_num_words):
        c_in = 0
        ab = list()
        for x in range(2 * ab_num_words):
            a_word = a.words[x] if x < ab_num_words else 0
            b_word = b.words[x]
            (c_out, s_out) = self.lowlevel.add_words(a_word, b_word, c_in)
            ab.append(s_out)
            c_in = c_out
        return ModExpNG_Operand(None, 2*ab_num_words, ab)

    def multipurpose_multiply(self, a, b, n, n_coeff, ab_num_words, reduce_only=False, multiply_only=False, dump=False, dump_crt="", dump_ladder=""):

        #
        # 1. AB = A * B
        #
        if dump: print("multiply_square(%s_%s)" % (dump_crt, dump_ladder))

        if reduce_only:
            ab = b
        else:
            ab_parts = self.multiplier.multiply_square(a, b, ab_num_words, dump)
            ab_words = self.recombinator.recombine_square(ab_parts, ab_num_words, dump)
            ab = ModExpNG_Operand(None, 2 * ab_num_words, ab_words)

        if dump and DUMP_VECTORS:
            ab.format_verilog_concat("%s_%s_AB" % (dump_crt, dump_ladder))

        if multiply_only:
            return ModExpNG_Operand(None, 2*ab_num_words, ab_words)

        #
        # 2. Q = LSB(AB) * N_COEFF
        #
        if dump: print("multiply_triangle(%s_%s)" % (dump_crt, dump_ladder))

        q_parts = self.multiplier.multiply_triangle(ab, n_coeff, ab_num_words, dump)
        q_words = self.recombinator.recombine_triangle(q_parts, ab_num_words, dump)
        q = ModExpNG_Operand(None, ab_num_words + 1, q_words)

        if dump and DUMP_VECTORS:
            q.format_verilog_concat("%s_%s_Q" % (dump_crt, dump_ladder))

        #
        # 3. M = Q * N
        #
        if dump: print("multiply_rectangle(%s_%s)" % (dump_crt, dump_ladder))

        m_parts = self.multiplier.multiply_rectangle(n, q, ab_num_words, dump)
        m_words = self.recombinator.recombine_rectangle(m_parts, ab_num_words, dump)
        m = ModExpNG_Operand(None, 2 * ab_num_words + 1, m_words)

        if dump and DUMP_VECTORS:
            m.format_verilog_concat("%s_%s_M" % (dump_crt, dump_ladder))

        #
        # 4. R = AB + M
        #

        #
        # 4a. compute carry (actual sum is all zeroes and need not be stored)
        #
        
        r_cy = 0 # this can be up to two bits, since we're adding extended words!!
        for i in range(ab_num_words + 1):
            s = ab.words[i] + m.words[i] + r_cy
            r_cy_new = s >> _WORD_WIDTH

            if dump and DUMP_REDUCTION:
                print("[%2d] 0x%05x + 0x%05x + 0x%x => {0x%x, [0x%05x]}" %
                    (i, ab.words[i], m.words[i], r_cy, r_cy_new, s & 0xffff))   # ???

            r_cy = r_cy_new


        #
        # 4b. Initialize empty result
        #
        
        R = list()
        for i in range(ab_num_words):
            R.append(0)

        #
        # 4c. compute the actual upper part of sum (take carry into account)
        #
        
        for i in range(ab_num_words):

            if dump and DUMP_REDUCTION:
                print("[%2d]" % i, end='')

            ab_word = ab.words[ab_num_words + i + 1] if i < (ab_num_words - 1) else 0
            if dump and DUMP_REDUCTION:
                print(" 0x%05x" % ab_word, end='')

            m_word = m.words[ab_num_words + i + 1]
            if dump and DUMP_REDUCTION:
                print(" + 0x%05x" % m_word, end='')

            if i == 0: R[i] = r_cy
            else:      R[i] = 0

            if dump and DUMP_REDUCTION:
                print(" + 0x%x" % R[i], end='')

            R[i] += ab_word
            R[i] += m_word
            if dump and DUMP_REDUCTION:
                print(" = 0x%05x" % R[i])

        return ModExpNG_Operand(None, ab_num_words, R)

    def convert_nonredundant(self, a, num_words):
        carry = 0
        for x in range(num_words):
            a.words[x] += carry
            carry = a.words[x] >> _WORD_WIDTH
            a.words[x] &= _WORD_MASK
        return carry

class ModExpNG_Core():

    def __init__(self, i):
        self.wrk = ModExpNG_Worker()
        self.bnk = ModExpNG_BanksCRT(i)
        self.inp = ModExpNG_CoreInput()
        self.out = ModExpNG_CoreOutput()

    #
    # CRT_(X|Y) means either CRT_X or CRT_Y
    # LADDER_{X,Y} means both LADDER_X and LADDER_Y
    #

    #
    # copy from CRT_(X|Y).LADDER_X.NARROW to OUTPUT
    #
    def set_output_from_narrow(self, sel_output, bank_crt, sel_narrow):
        self.out._set_value(sel_output, bank_crt.ladder_x._get_narrow(sel_narrow))

    #
    # copy from INPUT to CRT_(X|Y).LADDER_{X,Y}.NARROW
    #
    def set_narrow_from_input(self, bank_crt, sel_narrow, sel_input):
        bank_crt.ladder_x._set_narrow(sel_narrow, self.inp._get_value(sel_input))
        bank_crt.ladder_y._set_narrow(sel_narrow, self.inp._get_value(sel_input))

    #
    # copy from INPUT to CRT_(X|Y).LADDER_{X,Y}.WIDE
    #
    def set_wide_from_input(self, bank_crt, sel_wide, sel_input):
        bank_crt.ladder_x._set_wide(sel_wide, self.inp._get_value(sel_input))
        bank_crt.ladder_y._set_wide(sel_wide, self.inp._get_value(sel_input))

    #
    # copy from CRT_Y.LADDER_{X,Y}.{WIDE,NARROW} to CRT_X.LADDER_{X,Y}.{WIDE,NARROW}
    #
    def copy_crt_y2x(self, sel_wide, sel_narrow):

        self.bnk.crt_x.ladder_x._set_wide(sel_wide, self.bnk.crt_y.ladder_x._get_wide(sel_wide))
        self.bnk.crt_x.ladder_y._set_wide(sel_wide, self.bnk.crt_y.ladder_y._get_wide(sel_wide))

        self.bnk.crt_x.ladder_x._set_narrow(sel_narrow, self.bnk.crt_y.ladder_x._get_narrow(sel_narrow))
        self.bnk.crt_x.ladder_y._set_narrow(sel_narrow, self.bnk.crt_y.ladder_y._get_narrow(sel_narrow))

    #
    # copy from CRT_{X,Y}.LADDER_X.{WIDE,NARROW} to CRT_{X,Y}.LADDER_Y.{WIDE,NARROW}
    #
    def copy_ladders_x2y(self, sel_wide_in, sel_narrow_in, sel_wide_out, sel_narrow_out):

        self.bnk.crt_x.ladder_y._set_wide(sel_wide_out, self.bnk.crt_x.ladder_x._get_wide(sel_wide_in))
        self.bnk.crt_y.ladder_y._set_wide(sel_wide_out, self.bnk.crt_y.ladder_x._get_wide(sel_wide_in))

        self.bnk.crt_x.ladder_y._set_narrow(sel_narrow_out, self.bnk.crt_x.ladder_x._get_narrow(sel_narrow_in))
        self.bnk.crt_y.ladder_y._set_narrow(sel_narrow_out, self.bnk.crt_y.ladder_x._get_narrow(sel_narrow_in))

    #
    # copy from CRT_{X,Y}.LADDER_Y.{WIDE,NARROW} to CRT_{X,Y}.LADDER_X.{WIDE,NARROW}
    #
    def copy_ladders_y2x(self, sel_wide_in, sel_narrow_in, sel_wide_out, sel_narrow_out):

        self.bnk.crt_x.ladder_x._set_wide(sel_wide_out, self.bnk.crt_x.ladder_y._get_wide(sel_wide_in))
        self.bnk.crt_y.ladder_x._set_wide(sel_wide_out, self.bnk.crt_y.ladder_y._get_wide(sel_wide_in))

        self.bnk.crt_x.ladder_x._set_narrow(sel_narrow_out, self.bnk.crt_x.ladder_y._get_narrow(sel_narrow_in))
        self.bnk.crt_y.ladder_x._set_narrow(sel_narrow_out, self.bnk.crt_y.ladder_y._get_narrow(sel_narrow_in))

    #
    # copy from CRT_{X,Y}.LADDER_X.{WIDE,NARROW} to CRT_{Y,X}.LADDER_Y.{WIDE,NARROW}
    #
    def cross_ladders_x2y(self, sel_wide_in, sel_narrow_in, sel_wide_out, sel_narrow_out):

        self.bnk.crt_x.ladder_y._set_wide(sel_wide_out, self.bnk.crt_y.ladder_x._get_wide(sel_wide_in))
        self.bnk.crt_y.ladder_y._set_wide(sel_wide_out, self.bnk.crt_x.ladder_x._get_wide(sel_wide_in))
        
        self.bnk.crt_x.ladder_y._set_narrow(sel_narrow_out, self.bnk.crt_y.ladder_x._get_narrow(sel_narrow_in))
        self.bnk.crt_y.ladder_y._set_narrow(sel_narrow_out, self.bnk.crt_x.ladder_x._get_narrow(sel_narrow_in))

    #
    # modular multiply sel_wide_in by sel_narrow_in
    # stores intermediate result in WIDE.L and WIDE.H
    # needs modulus WIDE.N and speed-up coefficients NARROW.N_COEFF to be filled
    # places two copies of resulting quantity in sel_wide_out and sel_narrow_out
    # sel_*_in and sel_*_out can overlap (overwriting of input operands is ok)
    #
    def modular_multiply(self, sel_wide_in, sel_narrow_in, sel_wide_out, sel_narrow_out, num_words, mode=(True, True), d=False):

        xn       = self.bnk.crt_x.ladder_x._get_wide(ModExpNG_WideBankEnum.N)
        yn       = self.bnk.crt_y.ladder_x._get_wide(ModExpNG_WideBankEnum.N)

        xn_coeff = self.bnk.crt_x.ladder_x._get_narrow(ModExpNG_NarrowBankEnum.N_COEFF)
        yn_coeff = self.bnk.crt_y.ladder_x._get_narrow(ModExpNG_NarrowBankEnum.N_COEFF)

        xxa       = self.bnk.crt_x.ladder_x._get_wide(sel_wide_in)
        xya       = self.bnk.crt_x.ladder_y._get_wide(sel_wide_in)

        yxa       = self.bnk.crt_y.ladder_x._get_wide(sel_wide_in)
        yya       = self.bnk.crt_y.ladder_y._get_wide(sel_wide_in)

        xxb       = self.bnk.crt_x.ladder_x._get_narrow(sel_narrow_in)
        xyb       = self.bnk.crt_x.ladder_y._get_narrow(sel_narrow_in)

        yxb       = self.bnk.crt_y.ladder_x._get_narrow(sel_narrow_in)
        yyb       = self.bnk.crt_y.ladder_y._get_narrow(sel_narrow_in)

        if not mode[0]: xb = xxb
        else:           xb = xyb

        if not mode[1]: yb = yxb
        else:           yb = yyb

        xxp = self.wrk.multipurpose_multiply(xxa, xb, xn, xn_coeff, num_words, dump=d, dump_crt="X", dump_ladder="X")
        xyp = self.wrk.multipurpose_multiply(xya, xb, xn, xn_coeff, num_words, dump=d, dump_crt="X", dump_ladder="Y")

        yxp = self.wrk.multipurpose_multiply(yxa, yb, yn, yn_coeff, num_words, dump=d, dump_crt="Y", dump_ladder="X")
        yyp = self.wrk.multipurpose_multiply(yya, yb, yn, yn_coeff, num_words, dump=d, dump_crt="Y", dump_ladder="Y")

        self.bnk.crt_x.ladder_x._set_wide(sel_wide_out, xxp)
        self.bnk.crt_x.ladder_y._set_wide(sel_wide_out, xyp)
        self.bnk.crt_y.ladder_x._set_wide(sel_wide_out, yxp)
        self.bnk.crt_y.ladder_y._set_wide(sel_wide_out, yyp)

        self.bnk.crt_x.ladder_x._set_narrow(sel_narrow_out, xxp)
        self.bnk.crt_x.ladder_y._set_narrow(sel_narrow_out, xyp)
        self.bnk.crt_y.ladder_x._set_narrow(sel_narrow_out, yxp)
        self.bnk.crt_y.ladder_y._set_narrow(sel_narrow_out, yyp)

    #
    # modular subtract values in sel_narrow_in (X-Y)
    # stores two copies of the result in sel_*_out
    #
    def modular_subtract(self, sel_narrow_in, sel_narrow_out, sel_wide_out, num_words):

        xa = self.bnk.crt_x.ladder_x._get_narrow(sel_narrow_in)
        xb = self.bnk.crt_x.ladder_y._get_narrow(sel_narrow_in)
        xn = self.bnk.crt_x.ladder_x._get_wide(ModExpNG_WideBankEnum.N)

        ya = self.bnk.crt_y.ladder_x._get_narrow(sel_narrow_in)
        yb = self.bnk.crt_y.ladder_y._get_narrow(sel_narrow_in)
        yn = self.bnk.crt_y.ladder_x._get_wide(ModExpNG_WideBankEnum.N)

        xd = self.wrk.serial_subtract_modular(xa, xb, xn, num_words)
        yd = self.wrk.serial_subtract_modular(ya, yb, yn, num_words)

        self.bnk.crt_x.ladder_x._set_narrow(sel_narrow_out, xd)
        self.bnk.crt_y.ladder_x._set_narrow(sel_narrow_out, yd)

        self.bnk.crt_x.ladder_x._set_wide(sel_wide_out, xd)
        self.bnk.crt_y.ladder_x._set_wide(sel_wide_out, yd)
    
    #
    # modular reduce sel_narrow_in
    # stores two copies of the result in sel_*_out
    #
    def modular_reduce(self, sel_narrow_in, sel_wide_out, sel_narrow_out, num_words):

        xn       = self.bnk.crt_x.ladder_x._get_wide(ModExpNG_WideBankEnum.N)
        yn       = self.bnk.crt_y.ladder_x._get_wide(ModExpNG_WideBankEnum.N)

        xn_coeff = self.bnk.crt_x.ladder_x._get_narrow(ModExpNG_NarrowBankEnum.N_COEFF)
        yn_coeff = self.bnk.crt_y.ladder_x._get_narrow(ModExpNG_NarrowBankEnum.N_COEFF)

        xb       = self.bnk.crt_x.ladder_x._get_narrow(sel_narrow_in)
        yb       = self.bnk.crt_y.ladder_x._get_narrow(sel_narrow_in)

        xp = self.wrk.multipurpose_multiply(None, xb, xn, xn_coeff, num_words, reduce_only=True)
        yp = self.wrk.multipurpose_multiply(None, yb, yn, yn_coeff, num_words, reduce_only=True)

        self.bnk.crt_x.ladder_x._set_wide(sel_wide_out, xp)
        self.bnk.crt_x.ladder_y._set_wide(sel_wide_out, xp)
        self.bnk.crt_y.ladder_x._set_wide(sel_wide_out, yp)
        self.bnk.crt_y.ladder_y._set_wide(sel_wide_out, yp)

        self.bnk.crt_x.ladder_x._set_narrow(sel_narrow_out, xp)
        self.bnk.crt_x.ladder_y._set_narrow(sel_narrow_out, xp)
        self.bnk.crt_y.ladder_x._set_narrow(sel_narrow_out, yp)
        self.bnk.crt_y.ladder_y._set_narrow(sel_narrow_out, yp)

    #
    # propagate carries (convert to non-redundant representation) content in sel_narrow
    # overwrites input value
    #
    def propagate_carries(self, sel_narrow, num_words):
        self.wrk.convert_nonredundant(self.bnk.crt_x.ladder_x._get_narrow(sel_narrow), num_words)
        self.wrk.convert_nonredundant(self.bnk.crt_x.ladder_y._get_narrow(sel_narrow), num_words)
        self.wrk.convert_nonredundant(self.bnk.crt_y.ladder_x._get_narrow(sel_narrow), num_words)
        self.wrk.convert_nonredundant(self.bnk.crt_y.ladder_y._get_narrow(sel_narrow), num_words)

    #
    # copy from CRT_{X,Y}.LADDER_{X,Y}.WIDE.{H,L} to CRT_{X,Y}.LADDER_{X,Y}.NARROW
    #
    def merge_lha(self, sel_narrow, num_words):
        xx_lsb = self.bnk.crt_x.ladder_x._get_wide(ModExpNG_WideBankEnum.L)
        xy_lsb = self.bnk.crt_x.ladder_y._get_wide(ModExpNG_WideBankEnum.L)
        yx_lsb = self.bnk.crt_y.ladder_x._get_wide(ModExpNG_WideBankEnum.L)
        yy_lsb = self.bnk.crt_y.ladder_y._get_wide(ModExpNG_WideBankEnum.L)

        xx_msb = self.bnk.crt_x.ladder_x._get_wide(ModExpNG_WideBankEnum.H)
        xy_msb = self.bnk.crt_x.ladder_y._get_wide(ModExpNG_WideBankEnum.H)
        yx_msb = self.bnk.crt_y.ladder_x._get_wide(ModExpNG_WideBankEnum.H)
        yy_msb = self.bnk.crt_y.ladder_y._get_wide(ModExpNG_WideBankEnum.H)

        xx = xx_lsb.words + xx_msb.words
        xy = xy_lsb.words + xy_msb.words
        yx = yx_lsb.words + yx_msb.words
        yy = yy_lsb.words + yy_msb.words

        self.bnk.crt_x.ladder_x._set_narrow(sel_narrow, ModExpNG_Operand(None, 2*num_words, xx))
        self.bnk.crt_x.ladder_y._set_narrow(sel_narrow, ModExpNG_Operand(None, 2*num_words, xy))
        self.bnk.crt_y.ladder_x._set_narrow(sel_narrow, ModExpNG_Operand(None, 2*num_words, yx))
        self.bnk.crt_y.ladder_y._set_narrow(sel_narrow, ModExpNG_Operand(None, 2*num_words, yy))

    #
    # multiply sel_wide_in by sel_narrow_in
    # stores twice larger product in WIDE.L and WIDE.H
    #
    def regular_multiply(self, sel_wide_in, sel_narrow_in, num_words):

        xn       = self.bnk.crt_x.ladder_x._get_wide(ModExpNG_WideBankEnum.N)
        yn       = self.bnk.crt_y.ladder_x._get_wide(ModExpNG_WideBankEnum.N)

        xn_coeff = self.bnk.crt_x.ladder_x._get_narrow(ModExpNG_NarrowBankEnum.N_COEFF)
        yn_coeff = self.bnk.crt_y.ladder_x._get_narrow(ModExpNG_NarrowBankEnum.N_COEFF)

        xxa       = self.bnk.crt_x.ladder_x._get_wide(sel_wide_in)
        xya       = self.bnk.crt_x.ladder_y._get_wide(sel_wide_in)

        yxa       = self.bnk.crt_y.ladder_x._get_wide(sel_wide_in)
        yya       = self.bnk.crt_y.ladder_y._get_wide(sel_wide_in)

        xb       = self.bnk.crt_x.ladder_x._get_narrow(sel_narrow_in)
        yb       = self.bnk.crt_y.ladder_x._get_narrow(sel_narrow_in)

        xxp = self.wrk.multipurpose_multiply(xxa, xb, None, None, num_words, multiply_only=True)
        xyp = self.wrk.multipurpose_multiply(xya, xb, None, None, num_words, multiply_only=True)

        yxp = self.wrk.multipurpose_multiply(yxa, yb, None, None, num_words, multiply_only=True)
        yyp = self.wrk.multipurpose_multiply(yya, yb, None, None, num_words, multiply_only=True)

        xxp_lsb = xxp.lower_half()
        xxp_msb = xxp.upper_half()

        xyp_lsb = xyp.lower_half()
        xyp_msb = xyp.upper_half()

        yxp_lsb = yxp.lower_half()
        yxp_msb = yxp.upper_half()

        yyp_lsb = yyp.lower_half()
        yyp_msb = yyp.upper_half()

        self.bnk.crt_x.ladder_x._set_wide(ModExpNG_WideBankEnum.L, xxp_lsb)
        self.bnk.crt_x.ladder_y._set_wide(ModExpNG_WideBankEnum.L, xyp_lsb)
        self.bnk.crt_y.ladder_x._set_wide(ModExpNG_WideBankEnum.L, yxp_lsb)
        self.bnk.crt_y.ladder_y._set_wide(ModExpNG_WideBankEnum.L, yyp_lsb)

        self.bnk.crt_x.ladder_x._set_wide(ModExpNG_WideBankEnum.H, xxp_msb)
        self.bnk.crt_x.ladder_y._set_wide(ModExpNG_WideBankEnum.H, xyp_msb)
        self.bnk.crt_y.ladder_x._set_wide(ModExpNG_WideBankEnum.H, yxp_msb)
        self.bnk.crt_y.ladder_y._set_wide(ModExpNG_WideBankEnum.H, yyp_msb)

    #
    # adds sel_narrow_a_in to sel_narrow_b_in
    # stores result in sel_narrow_out
    #
    def regular_add(self, sel_narrow_a_in, sel_narrow_b_in, sel_narrow_out, num_words):
        xxa = self.bnk.crt_x.ladder_x._get_narrow(sel_narrow_a_in)
        xya = self.bnk.crt_x.ladder_y._get_narrow(sel_narrow_a_in)
        yxa = self.bnk.crt_y.ladder_x._get_narrow(sel_narrow_a_in)
        yya = self.bnk.crt_y.ladder_y._get_narrow(sel_narrow_a_in)

        xxb = self.bnk.crt_x.ladder_x._get_narrow(sel_narrow_b_in)
        xyb = self.bnk.crt_x.ladder_y._get_narrow(sel_narrow_b_in)
        yxb = self.bnk.crt_y.ladder_x._get_narrow(sel_narrow_b_in)
        yyb = self.bnk.crt_y.ladder_y._get_narrow(sel_narrow_b_in)

        xxc = self.wrk.serial_add_uneven(xxa, xxb, num_words)
        xyc = self.wrk.serial_add_uneven(xya, xyb, num_words)
        yxc = self.wrk.serial_add_uneven(yxa, yxb, num_words)
        yyc = self.wrk.serial_add_uneven(yya, yyb, num_words)

        self.bnk.crt_x.ladder_x._set_narrow(sel_narrow_out, xxc)
        self.bnk.crt_x.ladder_y._set_narrow(sel_narrow_out, xyc)
        self.bnk.crt_y.ladder_x._set_narrow(sel_narrow_out, yxc)
        self.bnk.crt_y.ladder_y._set_narrow(sel_narrow_out, yyc)

    #
    # dump working variables before ladder step
    #
    def dump_before_step_using_crt(self, pq, m):
        print("num_words = %d" % pq)
        print("\rladder_mode_x = %d" % m[0])
        print("\rladder_mode_y = %d" % m[1])
        self.bnk.crt_x.ladder_x._get_narrow(N.C).format_verilog_concat("X_X")
        self.bnk.crt_x.ladder_y._get_narrow(N.C).format_verilog_concat("X_Y")
        self.bnk.crt_y.ladder_x._get_narrow(N.C).format_verilog_concat("Y_X")
        self.bnk.crt_y.ladder_y._get_narrow(N.C).format_verilog_concat("Y_Y")
        self.bnk.crt_x.ladder_x._get_wide(W.N).format_verilog_concat("X_N")
        self.bnk.crt_x.ladder_x._get_wide(W.N).format_verilog_concat("Y_N")
        self.bnk.crt_x.ladder_x._get_narrow(N.N_COEFF).format_verilog_concat("X_N_COEFF")
        self.bnk.crt_x.ladder_x._get_narrow(N.N_COEFF).format_verilog_concat("Y_N_COEFF")

    #
    # dump working variables after ladder step
    #
    def dump_after_step_using_crt(self):
        self.bnk.crt_x.ladder_x._get_narrow(N.C).format_verilog_concat("X_X")
        self.bnk.crt_x.ladder_y._get_narrow(N.C).format_verilog_concat("X_Y")
        self.bnk.crt_y.ladder_x._get_narrow(N.C).format_verilog_concat("Y_X")
        self.bnk.crt_y.ladder_y._get_narrow(N.C).format_verilog_concat("Y_Y")

    #
    # this deliberately converts narrow operand into redundant representation
    #
    def _force_overflow(self, bank_crt, sel_narrow):

        # original words
        T = bank_crt.ladder_x._get_narrow(sel_narrow).words

        # loop through upper N-1 words
        for i in range(1, len(T)):

            # get msbs of the previous word
            upper_bits = T[i-1] & _CARRY_MASK

            # if the previous msbs are empty, force lsbs of the current word
            # into them and then wipe the current lsbs
            if upper_bits == 0:
                lower_bits = T[i] & (_CARRY_MASK >> _WORD_WIDTH)
                T[i] ^= lower_bits
                T[i-1] |= (lower_bits << _WORD_WIDTH)

        # overwrite original words
        bank_crt.ladder_x._set_narrow(sel_narrow, ModExpNG_Operand(None, len(T), T))

        print("Forced overflow.")

#
# read content of core's output bank and compare it against known good values
#
def compare_signature():

    c  = core
    s  = s_known
    xm = xm_known
    ym = ym_known

    core_s  = c.out.get_value(O.S)
    core_xm = c.out.get_value(O.XM)
    core_ym = c.out.get_value(O.YM)

    if core_s.number()  != s:  print("ERROR: core_s != s!")
    else:                      print("s is OK")

    if core_xm.number() != xm: print("ERROR: core_xm != xm!")
    else:                      print("x_mutated is OK")

    if core_ym.number() != ym: print("ERROR: core_ym != ym!")
    else:                      print("y_mutated is OK")

#
# get current ladder mode based on two exponents' bits
#
def get_ladder_mode_using_crt(v, bit):

    bit_value_p = (v.dp.number() & (1 << bit)) >> bit
    bit_value_q = (v.dq.number() & (1 << bit)) >> bit

    bit_value_p = bit_value_p > 0
    bit_value_q = bit_value_q > 0

    return (bit_value_p, bit_value_q)

#
# get current ladder mode based on private exponent's bit
#
def get_ladder_mode_without_crt(v, bit):

    bit_value_d = (v.d.number() & (1 << bit)) >> bit

    bit_value_d = bit_value_d > 0

    return (not bit_value_d, bit_value_d)

#
# print current exponentiation progress
#
def print_ladder_progress(current, total):

    # this will always print "100.0%" at the very last iteration, since we're
    # counting bits from msb to lsb and the very last index is zero, which
    # is congruent to 0 mod DUMP_PROGRESS_FACTOR
    if (current % DUMP_PROGRESS_FACTOR) == 0:
        pct = float((_WORD_WIDTH * total - current) / (_WORD_WIDTH * total)) * 100.0
        print("\rdone: %5.1f%%" % pct, end='')

    # move to next line after the very last iteration
    if current == 0: print("")

#
# try to exponentiate using the quad-multiplier (dual-core, dual-ladder) scheme
#
def sign_using_crt():

    c  = core
    v  = vector
    n  = n_num_words
    pq = pq_num_words

    ff = (False, False)
                                                                   #
                                                                   # A / B => different content in banks (A in WIDE, B in NARROW)
                                                                   # [XY]Z => different content in ladders (XZ in X, YZ in Y)
                                                                   # ..    => temporarily half-filled bank (omitted to save space)
                                                                   # *     => "crossed" content (X.Y == Y.X and Y.Y == X.X)
                                                                   #
                                                                   # +------------------------+-------+------------------+---------+-----------+
                                                                   # |  A                     |  B    |  C               |  D      |  E        |
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.set_wide_from_input   (c.bnk.crt_x, W.N,       I.N)          # |  ?                     |  ?    |  ?               |  ?      | ?         |
    c.set_wide_from_input   (c.bnk.crt_y, W.N,       I.N)          # |  ?                     |  ?    |  ?               |  ?      | ?         |
    c.set_wide_from_input   (c.bnk.crt_x, W.A,       I.X)          # |  ..                    |  ?    |  ?               |  ?      | ?         |
    c.set_wide_from_input   (c.bnk.crt_y, W.A,       I.Y)          # | [XY] / ?               |  ?    |  ?               |  ?      | ?         |
    c.set_wide_from_input   (c.bnk.crt_x, W.E,       I.M)          # | [XY] / ?               |  ?    |  ?               |  ?      | .. / ?    |
    c.set_wide_from_input   (c.bnk.crt_y, W.E,       I.M)          # | [XY] / ?               |  ?    |  ?               |  ?      | M  / ?    |
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.set_narrow_from_input (c.bnk.crt_x, N.N_COEFF, I.N_COEFF)    # | [XY] / ?               |  ?    |  ?               |  ?      | M  / ?    |
    c.set_narrow_from_input (c.bnk.crt_y, N.N_COEFF, I.N_COEFF)    # | [XY] / ?               |  ?    |  ?               |  ?      | M  / ?    |
    c.set_narrow_from_input (c.bnk.crt_x, N.A,       I.N_FACTOR)   # | [XY] / ..              |  ?    |  ?               |  ?      | M  / ?    |
    c.set_narrow_from_input (c.bnk.crt_y, N.A,       I.N_FACTOR)   # | [XY] / N_FACTOR        |  ?    |  ?               |  ?      | M  / ?    |
    c.set_narrow_from_input (c.bnk.crt_x, N.E,       I.M)          # | [XY] / N_FACTOR        |  ?    |  ?               |  ?      | M  / ..   |
    c.set_narrow_from_input (c.bnk.crt_y, N.E,       I.M)          # | [XY] / N_FACTOR        |  ?    |  ?               |  ?      | M         |
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.modular_multiply(W.A, N.A, W.B, N.B, n)                      # | [XY] / N_FACTOR        | [XY]F |  ?               |  ?      | M         | [XY]F = [XY] * N_FACTOR
    c.modular_multiply(W.B, N.B, W.C, N.C, n, mode=ff)             # | [XY] / N_FACTOR        | [XY]F | [XY]YM           |  ?      | M         | [XY]MF = [XY]F * [XY]F
    c.modular_multiply(W.C, N.I, W.D, N.D, n)                      # | [XY] / N_FACTOR        | [XY]F | [XY]YM           | [XY]M   | M         | [XY]M = [XY]MF * 1
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.propagate_carries(N.D, n_num_words)                          # | [XY] / N_FACTOR        | [XY]F | [XY]YM           | [XY]M   | M         |
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.set_output_from_narrow(O.XM, c.bnk.crt_x, N.D)               # | [XY] / N_FACTOR        | [XY]F | [XY]YM           | [XY]M   | M         |
    c.set_output_from_narrow(O.YM, c.bnk.crt_y, N.D)               # | [XY] / N_FACTOR        | [XY]F | [XY]YM           | [XY]M   | M         |
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.modular_multiply(W.E, N.B, W.C, N.C, n)                      # | [XY] / N_FACTOR        | [XY]F | [XY]MB           | [XY]M   | M         | [XY]MB = M * [XY]F
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.propagate_carries(N.C, n_num_words)                          # | [XY] / N_FACTOR        | [XY]F | [XY]MB           | [XY]M   | M         |
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.copy_crt_y2x(W.C, N.C)                                       # | [XY] / N_FACTOR        | [XY]F |  YMB             | [XY]M   | M         |
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.set_wide_from_input  (c.bnk.crt_x, W.N,       I.P)           # | [XY] / N_FACTOR        | [XY]F |  YMB             | [XY]M   | M         |
    c.set_wide_from_input  (c.bnk.crt_y, W.N,       I.Q)           # | [XY] / N_FACTOR        | [XY]F |  YMB             | [XY]M   | M         |
    c.set_wide_from_input  (c.bnk.crt_x, W.A,       I.P_FACTOR)    # | ...         / N_FACTOR | [XY]F |  YMB             | [XY]M   | M         |
    c.set_wide_from_input  (c.bnk.crt_y, W.A,       I.Q_FACTOR)    # | [PQ]_FACTOR / N_FACTOR | [XY]F |  YMB             | [XY]M   | M         |
    c.set_wide_from_input  (c.bnk.crt_x, W.E,       I.QINV)        # | [PQ]_FACTOR / N_FACTOR | [XY]F |  YMB             | [XY]M   | ..        |
    c.set_wide_from_input  (c.bnk.crt_x, W.E,       I.QINV)        # | [PQ]_FACTOR / N_FACTOR | [XY]F |  YMB             | [XY]M   | QINV / M  |
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.set_narrow_from_input(c.bnk.crt_x, N.N_COEFF, I.P_COEFF)     # | [PQ]_FACTOR / N_FACTOR | [XY]F |  YMB             | [XY]M   | QINV / M  |
    c.set_narrow_from_input(c.bnk.crt_y, N.N_COEFF, I.Q_COEFF)     # | [PQ]_FACTOR / N_FACTOR | [XY]F |  YMB             | [XY]M   | QINV / M  |
    c.set_narrow_from_input(c.bnk.crt_x, N.A,       I.P_FACTOR)    # | [PQ]_FACTOR / ...      | [XY]F |  YMB             | [XY]M   | QINV / M  |
    c.set_narrow_from_input(c.bnk.crt_y, N.A,       I.Q_FACTOR)    # | [PQ]_FACTOR            | [XY]F |  YMB             | [XY]M   | QINV / M  |
    c.set_narrow_from_input(c.bnk.crt_x, N.E,       I.QINV)        # | [PQ]_FACTOR            | [XY]F |  YMB             | [XY]M   | QINV / .. |
    c.set_narrow_from_input(c.bnk.crt_x, N.E,       I.QINV)        # | [PQ]_FACTOR            | [XY]F |  YMB             | [XY]M   | QINV      |
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.modular_reduce(N.C, W.D, N.D, pq)                            # | [PQ]_FACTOR            | [XY]F |  YMB             | [PQ]MBZ | QINV      | [PQ]MBZ = YMB mod [PQ]
    c.modular_multiply(W.D, N.A, W.C, N.C, pq)                     # | [PQ]_FACTOR            | [XY]F | [PQ]MB           | [PQ]MBZ | QINV      | [PQ]MB = [PQ]MBZ * [PQ]_FACTOR
    c.modular_multiply(W.C, N.A, W.D, N.D, pq)                     # | [PQ]_FACTOR            | [XY]F | [PQ]MB           | [PQ]MBF | QINV      | [PQ]MBF = [PQ]MB * [PQ]_FACTOR
    c.modular_multiply(W.A, N.I, W.C, N.C, pq)                     # | [PQ]_FACTOR            | [XY]F | [PQ]IF           | [PQ]MBF | QINV      | [PQ]IF = 1 * [PQ]_FACTOR
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.copy_ladders_x2y(W.D, N.D, W.C, N.C)                         # | [PQ]_FACTOR            | [XY]F | [PQ]IF / [PQ]MBF | [PQ]MBF | QINV      |
                                                                   # +------------------------+-------+------------------+---------+-----------+
    ###########################                                    # |                        |       |                  |         |           |
    # Begin Montgomery Ladder #                                    # |                        |       |                  |         |           |
    ###########################                                    # |                        |       |                  |         |           |
                                                                   # |                        |       |                  |         |           |
    for bit in range(_WORD_WIDTH * pq - 1, -1, -1):                # |                        |       |                  |         |           |
                                                                   # |                        |       |                  |         |           |
        m  = get_ladder_mode_using_crt(v, bit)                     # |                        |       |                  |         |           |
        dbg = bit == DUMP_LADDER_INDEX                             # |                        |       |                  |         |           |
                                                                   # |                        |       |                  |         |           |
        if dbg:                                                    # |                        |       |                  |         |           |
            if FORCE_OVERFLOW: c._force_overflow(c.bnk.crt_x, N.C) # |                        |       |                  |         |           |
            if DUMP_VECTORS: c.dump_before_step_using_crt(pq, m)   # |                        |       |                  |         |           |
                                                                   # +------------------------+-------+------------------+---------+-----------+
        c.modular_multiply(W.C, N.C, W.C, N.C, pq, mode=m, d=dbg)  # | [PQ]_FACTOR            | [XY]F | [PQ]SBF          | [PQ]MBF | QINV      | <LADDER>
                                                                   # +------------------------+-------+------------------+---------+-----------+
        if dbg and DUMP_VECTORS: c.dump_after_step_using_crt()     # |                        |       |                  |         |           |
        print_ladder_progress(bit, pq)                             # |                        |       |                  |         |           |
                                                                   # |                        |       |                  |         |           |
    #########################                                      # |                        |       |                  |         |           |
    # End Montgomery Ladder #                                      # |                        |       |                  |         |           |
    #########################                                      # |                        |       |                  |         |           |
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.modular_multiply(W.C, N.I, W.D, N.D, pq)                     # | [PQ]_FACTOR            | [XY]F | [PQ]SBF          | [PQ]SB  | QINV      | [PQ]SB = [PQ]SBF * 1
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.propagate_carries(N.D, pq)                                   # | [PQ]_FACTOR            | [XY]F | [PQ]SBF          | [PQ]SB  | QINV      |
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.cross_ladders_x2y(W.D, N.D, W.D, N.D)                        # | [PQ]_FACTOR            | [XY]F | [PQ]SBF          | [PQ]SB* | QINV      |
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.modular_subtract(N.D, N.C, W.C, pq)                          # | [PQ]_FACTOR            | [XY]F |  RSB             | [PQ]SB* | QINV      | RSB = PSB - QSB
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.modular_multiply(W.C, N.E, W.C, N.C, pq)                     # | [PQ]_FACTOR            | [XY]F |  RSBIZ           | [PQ]SB* | QINV      | RSBIZ = RSB * QINV
    c.modular_multiply(W.C, N.A, W.C, N.C, pq)                     # | [PQ]_FACTOR            | [XY]F |  RSBI            | [PQ]SB* | QINV      | RSBI = RSBIZ * P_FACTOR
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.set_wide_from_input  (c.bnk.crt_x, W.E, I.Q)                 # | [PQ]_FACTOR / N_FACTOR | [XY]F |  RSBI            | [PQ]SB* | ..        |
    c.set_wide_from_input  (c.bnk.crt_x, W.E, I.Q)                 # | [PQ]_FACTOR / N_FACTOR | [XY]F |  RSBI            | [PQ]SB* | Q / QINV  |
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.set_narrow_from_input(c.bnk.crt_x, N.E, I.Q)                 # | [PQ]_FACTOR            | [XY]F |  RSBI            | [PQ]SB* | Q / ..    |
    c.set_narrow_from_input(c.bnk.crt_x, N.E, I.Q)                 # | [PQ]_FACTOR            | [XY]F |  RSBI            | [PQ]SB* | Q         |
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.regular_multiply(W.E, N.C, pq)                               # | [PQ]_FACTOR            | [XY]F |  RSBI            | [PQ]SB* | Q         | = RSBI * Q
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.merge_lha(N.A, pq)                                           # | [PQ]_FACTOR / QRSBI    | [XY]F |  RSBI            | [PQ]SB* | Q         |
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.propagate_carries(N.A, n)                                    # | [PQ]_FACTOR / QRSBI    | [XY]F |  RSBI            | [PQ]SB* | Q         |
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.copy_crt_y2x(W.D, N.D)                                       # | [PQ]_FACTOR / QRSBI    | [XY]F |  RSBI            |  QSB*   | Q         |
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.regular_add(N.D, N.A, N.C, pq)                               # | [PQ]_FACTOR / QRSBI    | [XY]F |  SB              |  QSB*   | Q         | SB = QSB + RSBI
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.set_wide_from_input  (c.bnk.crt_x, W.N, I.N)                 # |                        |       |                  |         |           |
    c.set_wide_from_input  (c.bnk.crt_y, W.N, I.N)                 # |                        |       |                  |         |           |
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.set_narrow_from_input(c.bnk.crt_x, N.N_COEFF, I.N_COEFF)     # |                        |       |                  |         |           |
    c.set_narrow_from_input(c.bnk.crt_y, N.N_COEFF, I.N_COEFF)     # |                        |       |                  |         |           |
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.modular_multiply(W.B, N.C, W.A, N.A, n, ff)                  # |  S                     |       |                  |         |           | S = XF * SB
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.propagate_carries(N.A, n)                                    # |  S                     |       |                  |         |           |
                                                                   # +------------------------+-------+------------------+---------+-----------+
    c.set_output_from_narrow(O.S, c.bnk.crt_x, N.A)                # |  S                     |       |                  |         |           |
                                                                   # +------------------------+-------+------------------+---------+-----------+

#
# try to exponentiate using only half of the quad-multiplier (one dual-ladder core)
#
def sign_without_crt():

    c  = core
    v  = vector
    n  = n_num_words

    ff = (False, False)

    c.set_wide_from_input   (c.bnk.crt_x, W.N,       I.N)
    c.set_wide_from_input   (c.bnk.crt_y, W.N,       I.N)
    c.set_wide_from_input   (c.bnk.crt_x, W.A,       I.X)
    c.set_wide_from_input   (c.bnk.crt_y, W.A,       I.Y)
    c.set_wide_from_input   (c.bnk.crt_x, W.E,       I.M)
    c.set_wide_from_input   (c.bnk.crt_y, W.E,       I.M)

    c.set_narrow_from_input (c.bnk.crt_x, N.N_COEFF, I.N_COEFF)
    c.set_narrow_from_input (c.bnk.crt_y, N.N_COEFF, I.N_COEFF)
    c.set_narrow_from_input (c.bnk.crt_x, N.A,       I.N_FACTOR)
    c.set_narrow_from_input (c.bnk.crt_y, N.A,       I.N_FACTOR)
    c.set_narrow_from_input (c.bnk.crt_x, N.E,       I.M)
    c.set_narrow_from_input (c.bnk.crt_y, N.E,       I.M)

    c.modular_multiply(W.A, N.A, W.B, N.B, n)           # [XY]F = [XY] * N_FACTOR
    c.modular_multiply(W.B, N.B, W.C, N.C, n, mode=ff)  # [XY]MF = [XY]F * [XY]F
    c.modular_multiply(W.C, N.I, W.D, N.D, n)           # [XY]M = [XY]MF * 1

    c.propagate_carries(N.D, n)

    c.set_output_from_narrow(O.XM, c.bnk.crt_x, N.D)
    c.set_output_from_narrow(O.YM, c.bnk.crt_y, N.D)

    c.modular_multiply(W.E, N.B, W.C, N.C, n)   # [XY]MB = M * [XY]F

    XF = c.bnk.crt_x.ladder_x._get_narrow(N.B)

    c.set_wide_from_input(c.bnk.crt_x, W.A, I.N_FACTOR)
    c.set_wide_from_input(c.bnk.crt_y, W.A, I.N_FACTOR)

    c.modular_multiply(W.C, N.A, W.D, N.D, n)   # MBF = MB * N_FACTOR
    c.modular_multiply(W.A, N.I, W.C, N.C, n)   # IF = 1 * N_FACTOR    
    
    c.copy_ladders_x2y(W.D, N.D, W.C, N.C)

    ###########################
    # Begin Montgomery Ladder #
    ###########################

    for bit in range(_WORD_WIDTH * n - 1, -1, -1):

        m  = get_ladder_mode_without_crt(v, bit)
        dbg = bit == DUMP_LADDER_INDEX

        if dbg:
            if FORCE_OVERFLOW: c._force_overflow(c.bnk.crt_x, N.C)
            if DUMP_VECTORS: c.dump_before_step_without_crt(n, m)

        c.modular_multiply(W.C, N.C, W.C, N.C, n, mode=m, d=dbg)

        if dbg and DUMP_VECTORS: c.dump_after_step_without_crt()
        print_ladder_progress(bit, n)
            
    #########################
    # End Montgomery Ladder #
    #########################

    c.cross_ladders_x2y(W.B, N.B, W.B, N.B)

    c.modular_multiply(W.C, N.I, W.D, N.D, n)           # SB = SBF * 1    
    c.modular_multiply(W.B, N.D, W.A, N.A, n, mode=ff)  # S = XF * SB

    c.copy_ladders_y2x(W.A, N.A, W.B, N.B)
    
    c.propagate_carries(N.B, n)
    
    c.set_output_from_narrow(O.S, c.bnk.crt_y, N.B)


#
# main()
#
if __name__ == "__main__":

    # handy shortcuts
    W = ModExpNG_WideBankEnum
    N = ModExpNG_NarrowBankEnum
    I = ModExpNG_CoreInputEnum
    O = ModExpNG_CoreOutputEnum

    # set helper quantity
    # instantiate core
    # load test vector
    # transfer numbers from vector to core
    # set numbers of words
    # obtain known good reference value with built-in math
    # mutate blinding quantities with built-in math

    i = ModExpNG_Operand(1, KEY_LENGTH)

    core   = ModExpNG_Core(i)
    vector = ModExpNG_TestVector()

    core.inp.set_value(I.M,        vector.m)

    core.inp.set_value(I.N,        vector.n)
    core.inp.set_value(I.P,        vector.p)
    core.inp.set_value(I.Q,        vector.q)

    core.inp.set_value(I.N_COEFF,  vector.n_coeff)
    core.inp.set_value(I.P_COEFF,  vector.p_coeff)
    core.inp.set_value(I.Q_COEFF,  vector.q_coeff)

    core.inp.set_value(I.N_FACTOR, vector.n_factor)
    core.inp.set_value(I.P_FACTOR, vector.p_factor)
    core.inp.set_value(I.Q_FACTOR, vector.q_factor)

    core.inp.set_value(I.X,        vector.x)
    core.inp.set_value(I.Y,        vector.y)

    core.inp.set_value(I.QINV,     vector.qinv)

    n_num_words  = KEY_LENGTH  // _WORD_WIDTH
    pq_num_words = n_num_words // 2

    s_known = pow(vector.m.number(), vector.d.number(), vector.n.number())

    xm_known = pow(vector.x.number(), 2, vector.n.number())
    ym_known = pow(vector.y.number(), 2, vector.n.number())

    # sign using CRT and check
    print("Signing using CRT...")
    sign_using_crt()
    compare_signature()

    # sign without CRT and check
    print("Signing without CRT...")
    sign_without_crt()
    compare_signature()


#
# End-of-File
#