legalize-fmad.s16.mir
63.8 KB
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
# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
# RUN: llc -mtriple=amdgcn-mesa-mesa3d -mcpu=tahiti -run-pass=legalizer %s -o - | FileCheck -check-prefix=GFX6 %s
# RUN: llc -mtriple=amdgcn-mesa-mesa3d -mcpu=fiji -run-pass=legalizer %s -o - | FileCheck -check-prefix=GFX7 %s
# RUN: llc -mtriple=amdgcn-mesa-mesa3d -mcpu=gfx1010 -run-pass=legalizer %s -o - | FileCheck -check-prefix=GFX10 %s
---
name: test_fmad_s16_flush
machineFunctionInfo:
mode:
fp64-fp16-input-denormals: false
fp64-fp16-output-denormals: false
body: |
bb.0:
liveins: $vgpr0, $vgpr1, $vgpr2
; GFX6-LABEL: name: test_fmad_s16_flush
; GFX6: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
; GFX6: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
; GFX6: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr1
; GFX6: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
; GFX6: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
; GFX6: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
; GFX6: [[FPEXT:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC]](s16)
; GFX6: [[FPEXT1:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC1]](s16)
; GFX6: [[FMUL:%[0-9]+]]:_(s32) = G_FMUL [[FPEXT]], [[FPEXT1]]
; GFX6: [[FPTRUNC:%[0-9]+]]:_(s16) = G_FPTRUNC [[FMUL]](s32)
; GFX6: [[FPEXT2:%[0-9]+]]:_(s32) = G_FPEXT [[FPTRUNC]](s16)
; GFX6: [[FPEXT3:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC2]](s16)
; GFX6: [[FADD:%[0-9]+]]:_(s32) = G_FADD [[FPEXT2]], [[FPEXT3]]
; GFX6: [[FPTRUNC1:%[0-9]+]]:_(s16) = G_FPTRUNC [[FADD]](s32)
; GFX6: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FPTRUNC1]](s16)
; GFX6: $vgpr0 = COPY [[ANYEXT]](s32)
; GFX7-LABEL: name: test_fmad_s16_flush
; GFX7: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
; GFX7: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
; GFX7: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr1
; GFX7: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
; GFX7: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
; GFX7: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
; GFX7: [[FMAD:%[0-9]+]]:_(s16) = G_FMAD [[TRUNC]], [[TRUNC1]], [[TRUNC2]]
; GFX7: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FMAD]](s16)
; GFX7: $vgpr0 = COPY [[ANYEXT]](s32)
; GFX10-LABEL: name: test_fmad_s16_flush
; GFX10: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
; GFX10: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
; GFX10: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr1
; GFX10: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
; GFX10: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
; GFX10: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
; GFX10: [[FMUL:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC]], [[TRUNC1]]
; GFX10: [[FADD:%[0-9]+]]:_(s16) = G_FADD [[FMUL]], [[TRUNC2]]
; GFX10: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD]](s16)
; GFX10: $vgpr0 = COPY [[ANYEXT]](s32)
%0:_(s32) = COPY $vgpr0
%1:_(s32) = COPY $vgpr1
%2:_(s32) = COPY $vgpr1
%3:_(s16) = G_TRUNC %0
%4:_(s16) = G_TRUNC %1
%5:_(s16) = G_TRUNC %2
%6:_(s16) = G_FMAD %3, %4, %5
%7:_(s32) = G_ANYEXT %6
$vgpr0 = COPY %7
...
---
name: test_fmad_v2s16_flush
machineFunctionInfo:
mode:
fp64-fp16-input-denormals: false
fp64-fp16-output-denormals: false
body: |
bb.0:
liveins: $vgpr0, $vgpr1, $vgpr2
; GFX6-LABEL: name: test_fmad_v2s16_flush
; GFX6: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
; GFX6: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
; GFX6: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
; GFX6: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>)
; GFX6: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32)
; GFX6: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
; GFX6: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
; GFX6: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
; GFX6: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[COPY1]](<2 x s16>)
; GFX6: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32)
; GFX6: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
; GFX6: [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32)
; GFX6: [[BITCAST2:%[0-9]+]]:_(s32) = G_BITCAST [[COPY2]](<2 x s16>)
; GFX6: [[TRUNC4:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST2]](s32)
; GFX6: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST2]], [[C]](s32)
; GFX6: [[TRUNC5:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR2]](s32)
; GFX6: [[FPEXT:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC]](s16)
; GFX6: [[FPEXT1:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC2]](s16)
; GFX6: [[FMUL:%[0-9]+]]:_(s32) = G_FMUL [[FPEXT]], [[FPEXT1]]
; GFX6: [[FPTRUNC:%[0-9]+]]:_(s16) = G_FPTRUNC [[FMUL]](s32)
; GFX6: [[FPEXT2:%[0-9]+]]:_(s32) = G_FPEXT [[FPTRUNC]](s16)
; GFX6: [[FPEXT3:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC4]](s16)
; GFX6: [[FADD:%[0-9]+]]:_(s32) = G_FADD [[FPEXT2]], [[FPEXT3]]
; GFX6: [[FPTRUNC1:%[0-9]+]]:_(s16) = G_FPTRUNC [[FADD]](s32)
; GFX6: [[FPEXT4:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC1]](s16)
; GFX6: [[FPEXT5:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC3]](s16)
; GFX6: [[FMUL1:%[0-9]+]]:_(s32) = G_FMUL [[FPEXT4]], [[FPEXT5]]
; GFX6: [[FPTRUNC2:%[0-9]+]]:_(s16) = G_FPTRUNC [[FMUL1]](s32)
; GFX6: [[FPEXT6:%[0-9]+]]:_(s32) = G_FPEXT [[FPTRUNC2]](s16)
; GFX6: [[FPEXT7:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC5]](s16)
; GFX6: [[FADD1:%[0-9]+]]:_(s32) = G_FADD [[FPEXT6]], [[FPEXT7]]
; GFX6: [[FPTRUNC3:%[0-9]+]]:_(s16) = G_FPTRUNC [[FADD1]](s32)
; GFX6: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[FPTRUNC1]](s16)
; GFX6: [[ZEXT1:%[0-9]+]]:_(s32) = G_ZEXT [[FPTRUNC3]](s16)
; GFX6: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[ZEXT1]], [[C]](s32)
; GFX6: [[OR:%[0-9]+]]:_(s32) = G_OR [[ZEXT]], [[SHL]]
; GFX6: [[BITCAST3:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32)
; GFX6: $vgpr0 = COPY [[BITCAST3]](<2 x s16>)
; GFX7-LABEL: name: test_fmad_v2s16_flush
; GFX7: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
; GFX7: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
; GFX7: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
; GFX7: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>)
; GFX7: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32)
; GFX7: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
; GFX7: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
; GFX7: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
; GFX7: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[COPY1]](<2 x s16>)
; GFX7: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32)
; GFX7: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
; GFX7: [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32)
; GFX7: [[BITCAST2:%[0-9]+]]:_(s32) = G_BITCAST [[COPY2]](<2 x s16>)
; GFX7: [[TRUNC4:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST2]](s32)
; GFX7: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST2]], [[C]](s32)
; GFX7: [[TRUNC5:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR2]](s32)
; GFX7: [[FMAD:%[0-9]+]]:_(s16) = G_FMAD [[TRUNC]], [[TRUNC2]], [[TRUNC4]]
; GFX7: [[FMAD1:%[0-9]+]]:_(s16) = G_FMAD [[TRUNC1]], [[TRUNC3]], [[TRUNC5]]
; GFX7: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[FMAD]](s16)
; GFX7: [[ZEXT1:%[0-9]+]]:_(s32) = G_ZEXT [[FMAD1]](s16)
; GFX7: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[ZEXT1]], [[C]](s32)
; GFX7: [[OR:%[0-9]+]]:_(s32) = G_OR [[ZEXT]], [[SHL]]
; GFX7: [[BITCAST3:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32)
; GFX7: $vgpr0 = COPY [[BITCAST3]](<2 x s16>)
; GFX10-LABEL: name: test_fmad_v2s16_flush
; GFX10: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
; GFX10: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
; GFX10: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
; GFX10: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>)
; GFX10: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32)
; GFX10: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
; GFX10: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
; GFX10: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
; GFX10: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[COPY1]](<2 x s16>)
; GFX10: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32)
; GFX10: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
; GFX10: [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32)
; GFX10: [[BITCAST2:%[0-9]+]]:_(s32) = G_BITCAST [[COPY2]](<2 x s16>)
; GFX10: [[TRUNC4:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST2]](s32)
; GFX10: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST2]], [[C]](s32)
; GFX10: [[TRUNC5:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR2]](s32)
; GFX10: [[FMUL:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC]], [[TRUNC2]]
; GFX10: [[FADD:%[0-9]+]]:_(s16) = G_FADD [[FMUL]], [[TRUNC4]]
; GFX10: [[FMUL1:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC1]], [[TRUNC3]]
; GFX10: [[FADD1:%[0-9]+]]:_(s16) = G_FADD [[FMUL1]], [[TRUNC5]]
; GFX10: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD]](s16)
; GFX10: [[ANYEXT1:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD1]](s16)
; GFX10: [[BUILD_VECTOR_TRUNC:%[0-9]+]]:_(<2 x s16>) = G_BUILD_VECTOR_TRUNC [[ANYEXT]](s32), [[ANYEXT1]](s32)
; GFX10: $vgpr0 = COPY [[BUILD_VECTOR_TRUNC]](<2 x s16>)
%0:_(<2 x s16>) = COPY $vgpr0
%1:_(<2 x s16>) = COPY $vgpr1
%2:_(<2 x s16>) = COPY $vgpr2
%3:_(<2 x s16>) = G_FMAD %0, %1, %2
$vgpr0 = COPY %3
...
---
name: test_fmad_v4s16_flush
machineFunctionInfo:
mode:
fp64-fp16-input-denormals: false
fp64-fp16-output-denormals: false
body: |
bb.0:
liveins: $vgpr0_vgpr1, $vgpr2_vgpr3, $vgpr4_vgpr5
; GFX6-LABEL: name: test_fmad_v4s16_flush
; GFX6: [[COPY:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr0_vgpr1
; GFX6: [[COPY1:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr2_vgpr3
; GFX6: [[COPY2:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr4_vgpr5
; GFX6: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY]](<4 x s16>)
; GFX6: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[UV]](<2 x s16>)
; GFX6: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32)
; GFX6: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
; GFX6: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
; GFX6: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
; GFX6: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[UV1]](<2 x s16>)
; GFX6: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32)
; GFX6: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
; GFX6: [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32)
; GFX6: [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY1]](<4 x s16>)
; GFX6: [[BITCAST2:%[0-9]+]]:_(s32) = G_BITCAST [[UV2]](<2 x s16>)
; GFX6: [[TRUNC4:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST2]](s32)
; GFX6: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST2]], [[C]](s32)
; GFX6: [[TRUNC5:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR2]](s32)
; GFX6: [[BITCAST3:%[0-9]+]]:_(s32) = G_BITCAST [[UV3]](<2 x s16>)
; GFX6: [[TRUNC6:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST3]](s32)
; GFX6: [[LSHR3:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST3]], [[C]](s32)
; GFX6: [[TRUNC7:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR3]](s32)
; GFX6: [[UV4:%[0-9]+]]:_(<2 x s16>), [[UV5:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY2]](<4 x s16>)
; GFX6: [[BITCAST4:%[0-9]+]]:_(s32) = G_BITCAST [[UV4]](<2 x s16>)
; GFX6: [[TRUNC8:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST4]](s32)
; GFX6: [[LSHR4:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST4]], [[C]](s32)
; GFX6: [[TRUNC9:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR4]](s32)
; GFX6: [[BITCAST5:%[0-9]+]]:_(s32) = G_BITCAST [[UV5]](<2 x s16>)
; GFX6: [[TRUNC10:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST5]](s32)
; GFX6: [[LSHR5:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST5]], [[C]](s32)
; GFX6: [[TRUNC11:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR5]](s32)
; GFX6: [[FPEXT:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC]](s16)
; GFX6: [[FPEXT1:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC4]](s16)
; GFX6: [[FMUL:%[0-9]+]]:_(s32) = G_FMUL [[FPEXT]], [[FPEXT1]]
; GFX6: [[FPTRUNC:%[0-9]+]]:_(s16) = G_FPTRUNC [[FMUL]](s32)
; GFX6: [[FPEXT2:%[0-9]+]]:_(s32) = G_FPEXT [[FPTRUNC]](s16)
; GFX6: [[FPEXT3:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC8]](s16)
; GFX6: [[FADD:%[0-9]+]]:_(s32) = G_FADD [[FPEXT2]], [[FPEXT3]]
; GFX6: [[FPTRUNC1:%[0-9]+]]:_(s16) = G_FPTRUNC [[FADD]](s32)
; GFX6: [[FPEXT4:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC1]](s16)
; GFX6: [[FPEXT5:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC5]](s16)
; GFX6: [[FMUL1:%[0-9]+]]:_(s32) = G_FMUL [[FPEXT4]], [[FPEXT5]]
; GFX6: [[FPTRUNC2:%[0-9]+]]:_(s16) = G_FPTRUNC [[FMUL1]](s32)
; GFX6: [[FPEXT6:%[0-9]+]]:_(s32) = G_FPEXT [[FPTRUNC2]](s16)
; GFX6: [[FPEXT7:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC9]](s16)
; GFX6: [[FADD1:%[0-9]+]]:_(s32) = G_FADD [[FPEXT6]], [[FPEXT7]]
; GFX6: [[FPTRUNC3:%[0-9]+]]:_(s16) = G_FPTRUNC [[FADD1]](s32)
; GFX6: [[FPEXT8:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC2]](s16)
; GFX6: [[FPEXT9:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC6]](s16)
; GFX6: [[FMUL2:%[0-9]+]]:_(s32) = G_FMUL [[FPEXT8]], [[FPEXT9]]
; GFX6: [[FPTRUNC4:%[0-9]+]]:_(s16) = G_FPTRUNC [[FMUL2]](s32)
; GFX6: [[FPEXT10:%[0-9]+]]:_(s32) = G_FPEXT [[FPTRUNC4]](s16)
; GFX6: [[FPEXT11:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC10]](s16)
; GFX6: [[FADD2:%[0-9]+]]:_(s32) = G_FADD [[FPEXT10]], [[FPEXT11]]
; GFX6: [[FPTRUNC5:%[0-9]+]]:_(s16) = G_FPTRUNC [[FADD2]](s32)
; GFX6: [[FPEXT12:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC3]](s16)
; GFX6: [[FPEXT13:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC7]](s16)
; GFX6: [[FMUL3:%[0-9]+]]:_(s32) = G_FMUL [[FPEXT12]], [[FPEXT13]]
; GFX6: [[FPTRUNC6:%[0-9]+]]:_(s16) = G_FPTRUNC [[FMUL3]](s32)
; GFX6: [[FPEXT14:%[0-9]+]]:_(s32) = G_FPEXT [[FPTRUNC6]](s16)
; GFX6: [[FPEXT15:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC11]](s16)
; GFX6: [[FADD3:%[0-9]+]]:_(s32) = G_FADD [[FPEXT14]], [[FPEXT15]]
; GFX6: [[FPTRUNC7:%[0-9]+]]:_(s16) = G_FPTRUNC [[FADD3]](s32)
; GFX6: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[FPTRUNC1]](s16)
; GFX6: [[ZEXT1:%[0-9]+]]:_(s32) = G_ZEXT [[FPTRUNC3]](s16)
; GFX6: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[ZEXT1]], [[C]](s32)
; GFX6: [[OR:%[0-9]+]]:_(s32) = G_OR [[ZEXT]], [[SHL]]
; GFX6: [[BITCAST6:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32)
; GFX6: [[ZEXT2:%[0-9]+]]:_(s32) = G_ZEXT [[FPTRUNC5]](s16)
; GFX6: [[ZEXT3:%[0-9]+]]:_(s32) = G_ZEXT [[FPTRUNC7]](s16)
; GFX6: [[SHL1:%[0-9]+]]:_(s32) = G_SHL [[ZEXT3]], [[C]](s32)
; GFX6: [[OR1:%[0-9]+]]:_(s32) = G_OR [[ZEXT2]], [[SHL1]]
; GFX6: [[BITCAST7:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR1]](s32)
; GFX6: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[BITCAST6]](<2 x s16>), [[BITCAST7]](<2 x s16>)
; GFX6: $vgpr0_vgpr1 = COPY [[CONCAT_VECTORS]](<4 x s16>)
; GFX7-LABEL: name: test_fmad_v4s16_flush
; GFX7: [[COPY:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr0_vgpr1
; GFX7: [[COPY1:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr2_vgpr3
; GFX7: [[COPY2:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr4_vgpr5
; GFX7: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY]](<4 x s16>)
; GFX7: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[UV]](<2 x s16>)
; GFX7: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32)
; GFX7: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
; GFX7: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
; GFX7: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
; GFX7: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[UV1]](<2 x s16>)
; GFX7: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32)
; GFX7: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
; GFX7: [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32)
; GFX7: [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY1]](<4 x s16>)
; GFX7: [[BITCAST2:%[0-9]+]]:_(s32) = G_BITCAST [[UV2]](<2 x s16>)
; GFX7: [[TRUNC4:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST2]](s32)
; GFX7: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST2]], [[C]](s32)
; GFX7: [[TRUNC5:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR2]](s32)
; GFX7: [[BITCAST3:%[0-9]+]]:_(s32) = G_BITCAST [[UV3]](<2 x s16>)
; GFX7: [[TRUNC6:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST3]](s32)
; GFX7: [[LSHR3:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST3]], [[C]](s32)
; GFX7: [[TRUNC7:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR3]](s32)
; GFX7: [[UV4:%[0-9]+]]:_(<2 x s16>), [[UV5:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY2]](<4 x s16>)
; GFX7: [[BITCAST4:%[0-9]+]]:_(s32) = G_BITCAST [[UV4]](<2 x s16>)
; GFX7: [[TRUNC8:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST4]](s32)
; GFX7: [[LSHR4:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST4]], [[C]](s32)
; GFX7: [[TRUNC9:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR4]](s32)
; GFX7: [[BITCAST5:%[0-9]+]]:_(s32) = G_BITCAST [[UV5]](<2 x s16>)
; GFX7: [[TRUNC10:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST5]](s32)
; GFX7: [[LSHR5:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST5]], [[C]](s32)
; GFX7: [[TRUNC11:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR5]](s32)
; GFX7: [[FMAD:%[0-9]+]]:_(s16) = G_FMAD [[TRUNC]], [[TRUNC4]], [[TRUNC8]]
; GFX7: [[FMAD1:%[0-9]+]]:_(s16) = G_FMAD [[TRUNC1]], [[TRUNC5]], [[TRUNC9]]
; GFX7: [[FMAD2:%[0-9]+]]:_(s16) = G_FMAD [[TRUNC2]], [[TRUNC6]], [[TRUNC10]]
; GFX7: [[FMAD3:%[0-9]+]]:_(s16) = G_FMAD [[TRUNC3]], [[TRUNC7]], [[TRUNC11]]
; GFX7: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[FMAD]](s16)
; GFX7: [[ZEXT1:%[0-9]+]]:_(s32) = G_ZEXT [[FMAD1]](s16)
; GFX7: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[ZEXT1]], [[C]](s32)
; GFX7: [[OR:%[0-9]+]]:_(s32) = G_OR [[ZEXT]], [[SHL]]
; GFX7: [[BITCAST6:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32)
; GFX7: [[ZEXT2:%[0-9]+]]:_(s32) = G_ZEXT [[FMAD2]](s16)
; GFX7: [[ZEXT3:%[0-9]+]]:_(s32) = G_ZEXT [[FMAD3]](s16)
; GFX7: [[SHL1:%[0-9]+]]:_(s32) = G_SHL [[ZEXT3]], [[C]](s32)
; GFX7: [[OR1:%[0-9]+]]:_(s32) = G_OR [[ZEXT2]], [[SHL1]]
; GFX7: [[BITCAST7:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR1]](s32)
; GFX7: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[BITCAST6]](<2 x s16>), [[BITCAST7]](<2 x s16>)
; GFX7: $vgpr0_vgpr1 = COPY [[CONCAT_VECTORS]](<4 x s16>)
; GFX10-LABEL: name: test_fmad_v4s16_flush
; GFX10: [[COPY:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr0_vgpr1
; GFX10: [[COPY1:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr2_vgpr3
; GFX10: [[COPY2:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr4_vgpr5
; GFX10: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY]](<4 x s16>)
; GFX10: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[UV]](<2 x s16>)
; GFX10: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32)
; GFX10: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
; GFX10: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
; GFX10: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
; GFX10: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[UV1]](<2 x s16>)
; GFX10: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32)
; GFX10: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
; GFX10: [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32)
; GFX10: [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY1]](<4 x s16>)
; GFX10: [[BITCAST2:%[0-9]+]]:_(s32) = G_BITCAST [[UV2]](<2 x s16>)
; GFX10: [[TRUNC4:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST2]](s32)
; GFX10: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST2]], [[C]](s32)
; GFX10: [[TRUNC5:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR2]](s32)
; GFX10: [[BITCAST3:%[0-9]+]]:_(s32) = G_BITCAST [[UV3]](<2 x s16>)
; GFX10: [[TRUNC6:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST3]](s32)
; GFX10: [[LSHR3:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST3]], [[C]](s32)
; GFX10: [[TRUNC7:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR3]](s32)
; GFX10: [[UV4:%[0-9]+]]:_(<2 x s16>), [[UV5:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY2]](<4 x s16>)
; GFX10: [[BITCAST4:%[0-9]+]]:_(s32) = G_BITCAST [[UV4]](<2 x s16>)
; GFX10: [[TRUNC8:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST4]](s32)
; GFX10: [[LSHR4:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST4]], [[C]](s32)
; GFX10: [[TRUNC9:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR4]](s32)
; GFX10: [[BITCAST5:%[0-9]+]]:_(s32) = G_BITCAST [[UV5]](<2 x s16>)
; GFX10: [[TRUNC10:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST5]](s32)
; GFX10: [[LSHR5:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST5]], [[C]](s32)
; GFX10: [[TRUNC11:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR5]](s32)
; GFX10: [[FMUL:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC]], [[TRUNC4]]
; GFX10: [[FADD:%[0-9]+]]:_(s16) = G_FADD [[FMUL]], [[TRUNC8]]
; GFX10: [[FMUL1:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC1]], [[TRUNC5]]
; GFX10: [[FADD1:%[0-9]+]]:_(s16) = G_FADD [[FMUL1]], [[TRUNC9]]
; GFX10: [[FMUL2:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC2]], [[TRUNC6]]
; GFX10: [[FADD2:%[0-9]+]]:_(s16) = G_FADD [[FMUL2]], [[TRUNC10]]
; GFX10: [[FMUL3:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC3]], [[TRUNC7]]
; GFX10: [[FADD3:%[0-9]+]]:_(s16) = G_FADD [[FMUL3]], [[TRUNC11]]
; GFX10: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD]](s16)
; GFX10: [[ANYEXT1:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD1]](s16)
; GFX10: [[BUILD_VECTOR_TRUNC:%[0-9]+]]:_(<2 x s16>) = G_BUILD_VECTOR_TRUNC [[ANYEXT]](s32), [[ANYEXT1]](s32)
; GFX10: [[ANYEXT2:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD2]](s16)
; GFX10: [[ANYEXT3:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD3]](s16)
; GFX10: [[BUILD_VECTOR_TRUNC1:%[0-9]+]]:_(<2 x s16>) = G_BUILD_VECTOR_TRUNC [[ANYEXT2]](s32), [[ANYEXT3]](s32)
; GFX10: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[BUILD_VECTOR_TRUNC]](<2 x s16>), [[BUILD_VECTOR_TRUNC1]](<2 x s16>)
; GFX10: $vgpr0_vgpr1 = COPY [[CONCAT_VECTORS]](<4 x s16>)
%0:_(<4 x s16>) = COPY $vgpr0_vgpr1
%1:_(<4 x s16>) = COPY $vgpr2_vgpr3
%2:_(<4 x s16>) = COPY $vgpr4_vgpr5
%3:_(<4 x s16>) = G_FMAD %0, %1, %2
$vgpr0_vgpr1 = COPY %3
...
---
name: test_fmad_s16_denorm
machineFunctionInfo:
mode:
fp64-fp16-input-denormals: true
fp64-fp16-output-denormals: true
body: |
bb.0:
liveins: $vgpr0, $vgpr1, $vgpr2
; GFX6-LABEL: name: test_fmad_s16_denorm
; GFX6: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
; GFX6: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
; GFX6: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr1
; GFX6: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
; GFX6: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
; GFX6: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
; GFX6: [[FPEXT:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC]](s16)
; GFX6: [[FPEXT1:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC1]](s16)
; GFX6: [[FMUL:%[0-9]+]]:_(s32) = G_FMUL [[FPEXT]], [[FPEXT1]]
; GFX6: [[FPTRUNC:%[0-9]+]]:_(s16) = G_FPTRUNC [[FMUL]](s32)
; GFX6: [[FPEXT2:%[0-9]+]]:_(s32) = G_FPEXT [[FPTRUNC]](s16)
; GFX6: [[FPEXT3:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC2]](s16)
; GFX6: [[FADD:%[0-9]+]]:_(s32) = G_FADD [[FPEXT2]], [[FPEXT3]]
; GFX6: [[FPTRUNC1:%[0-9]+]]:_(s16) = G_FPTRUNC [[FADD]](s32)
; GFX6: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FPTRUNC1]](s16)
; GFX6: $vgpr0 = COPY [[ANYEXT]](s32)
; GFX7-LABEL: name: test_fmad_s16_denorm
; GFX7: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
; GFX7: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
; GFX7: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr1
; GFX7: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
; GFX7: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
; GFX7: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
; GFX7: [[FMUL:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC]], [[TRUNC1]]
; GFX7: [[FADD:%[0-9]+]]:_(s16) = G_FADD [[FMUL]], [[TRUNC2]]
; GFX7: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD]](s16)
; GFX7: $vgpr0 = COPY [[ANYEXT]](s32)
; GFX10-LABEL: name: test_fmad_s16_denorm
; GFX10: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
; GFX10: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
; GFX10: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr1
; GFX10: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
; GFX10: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
; GFX10: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
; GFX10: [[FMUL:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC]], [[TRUNC1]]
; GFX10: [[FADD:%[0-9]+]]:_(s16) = G_FADD [[FMUL]], [[TRUNC2]]
; GFX10: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD]](s16)
; GFX10: $vgpr0 = COPY [[ANYEXT]](s32)
%0:_(s32) = COPY $vgpr0
%1:_(s32) = COPY $vgpr1
%2:_(s32) = COPY $vgpr1
%3:_(s16) = G_TRUNC %0
%4:_(s16) = G_TRUNC %1
%5:_(s16) = G_TRUNC %2
%6:_(s16) = G_FMAD %3, %4, %5
%7:_(s32) = G_ANYEXT %6
$vgpr0 = COPY %7
...
---
name: test_fmad_s16_denorm_flags
machineFunctionInfo:
mode:
fp64-fp16-input-denormals: true
fp64-fp16-output-denormals: true
body: |
bb.0:
liveins: $vgpr0, $vgpr1, $vgpr2
; GFX6-LABEL: name: test_fmad_s16_denorm_flags
; GFX6: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
; GFX6: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
; GFX6: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr1
; GFX6: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
; GFX6: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
; GFX6: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
; GFX6: [[FPEXT:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC]](s16)
; GFX6: [[FPEXT1:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC1]](s16)
; GFX6: [[FMUL:%[0-9]+]]:_(s32) = nnan G_FMUL [[FPEXT]], [[FPEXT1]]
; GFX6: [[FPTRUNC:%[0-9]+]]:_(s16) = G_FPTRUNC [[FMUL]](s32)
; GFX6: [[FPEXT2:%[0-9]+]]:_(s32) = G_FPEXT [[FPTRUNC]](s16)
; GFX6: [[FPEXT3:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC2]](s16)
; GFX6: [[FADD:%[0-9]+]]:_(s32) = nnan G_FADD [[FPEXT2]], [[FPEXT3]]
; GFX6: [[FPTRUNC1:%[0-9]+]]:_(s16) = G_FPTRUNC [[FADD]](s32)
; GFX6: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FPTRUNC1]](s16)
; GFX6: $vgpr0 = COPY [[ANYEXT]](s32)
; GFX7-LABEL: name: test_fmad_s16_denorm_flags
; GFX7: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
; GFX7: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
; GFX7: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr1
; GFX7: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
; GFX7: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
; GFX7: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
; GFX7: [[FMUL:%[0-9]+]]:_(s16) = nnan G_FMUL [[TRUNC]], [[TRUNC1]]
; GFX7: [[FADD:%[0-9]+]]:_(s16) = nnan G_FADD [[FMUL]], [[TRUNC2]]
; GFX7: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD]](s16)
; GFX7: $vgpr0 = COPY [[ANYEXT]](s32)
; GFX10-LABEL: name: test_fmad_s16_denorm_flags
; GFX10: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
; GFX10: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
; GFX10: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr1
; GFX10: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
; GFX10: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
; GFX10: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
; GFX10: [[FMUL:%[0-9]+]]:_(s16) = nnan G_FMUL [[TRUNC]], [[TRUNC1]]
; GFX10: [[FADD:%[0-9]+]]:_(s16) = nnan G_FADD [[FMUL]], [[TRUNC2]]
; GFX10: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD]](s16)
; GFX10: $vgpr0 = COPY [[ANYEXT]](s32)
%0:_(s32) = COPY $vgpr0
%1:_(s32) = COPY $vgpr1
%2:_(s32) = COPY $vgpr1
%3:_(s16) = G_TRUNC %0
%4:_(s16) = G_TRUNC %1
%5:_(s16) = G_TRUNC %2
%6:_(s16) = nnan G_FMAD %3, %4, %5
%7:_(s32) = G_ANYEXT %6
$vgpr0 = COPY %7
...
---
name: test_fmad_v2s16_denorm
machineFunctionInfo:
mode:
fp64-fp16-input-denormals: true
fp64-fp16-output-denormals: true
body: |
bb.0:
liveins: $vgpr0, $vgpr1, $vgpr2
; GFX6-LABEL: name: test_fmad_v2s16_denorm
; GFX6: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
; GFX6: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
; GFX6: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
; GFX6: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>)
; GFX6: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32)
; GFX6: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
; GFX6: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
; GFX6: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
; GFX6: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[COPY1]](<2 x s16>)
; GFX6: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32)
; GFX6: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
; GFX6: [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32)
; GFX6: [[BITCAST2:%[0-9]+]]:_(s32) = G_BITCAST [[COPY2]](<2 x s16>)
; GFX6: [[TRUNC4:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST2]](s32)
; GFX6: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST2]], [[C]](s32)
; GFX6: [[TRUNC5:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR2]](s32)
; GFX6: [[FPEXT:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC]](s16)
; GFX6: [[FPEXT1:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC2]](s16)
; GFX6: [[FMUL:%[0-9]+]]:_(s32) = G_FMUL [[FPEXT]], [[FPEXT1]]
; GFX6: [[FPTRUNC:%[0-9]+]]:_(s16) = G_FPTRUNC [[FMUL]](s32)
; GFX6: [[FPEXT2:%[0-9]+]]:_(s32) = G_FPEXT [[FPTRUNC]](s16)
; GFX6: [[FPEXT3:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC4]](s16)
; GFX6: [[FADD:%[0-9]+]]:_(s32) = G_FADD [[FPEXT2]], [[FPEXT3]]
; GFX6: [[FPTRUNC1:%[0-9]+]]:_(s16) = G_FPTRUNC [[FADD]](s32)
; GFX6: [[FPEXT4:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC1]](s16)
; GFX6: [[FPEXT5:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC3]](s16)
; GFX6: [[FMUL1:%[0-9]+]]:_(s32) = G_FMUL [[FPEXT4]], [[FPEXT5]]
; GFX6: [[FPTRUNC2:%[0-9]+]]:_(s16) = G_FPTRUNC [[FMUL1]](s32)
; GFX6: [[FPEXT6:%[0-9]+]]:_(s32) = G_FPEXT [[FPTRUNC2]](s16)
; GFX6: [[FPEXT7:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC5]](s16)
; GFX6: [[FADD1:%[0-9]+]]:_(s32) = G_FADD [[FPEXT6]], [[FPEXT7]]
; GFX6: [[FPTRUNC3:%[0-9]+]]:_(s16) = G_FPTRUNC [[FADD1]](s32)
; GFX6: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[FPTRUNC1]](s16)
; GFX6: [[ZEXT1:%[0-9]+]]:_(s32) = G_ZEXT [[FPTRUNC3]](s16)
; GFX6: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[ZEXT1]], [[C]](s32)
; GFX6: [[OR:%[0-9]+]]:_(s32) = G_OR [[ZEXT]], [[SHL]]
; GFX6: [[BITCAST3:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32)
; GFX6: $vgpr0 = COPY [[BITCAST3]](<2 x s16>)
; GFX7-LABEL: name: test_fmad_v2s16_denorm
; GFX7: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
; GFX7: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
; GFX7: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
; GFX7: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>)
; GFX7: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32)
; GFX7: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
; GFX7: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
; GFX7: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
; GFX7: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[COPY1]](<2 x s16>)
; GFX7: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32)
; GFX7: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
; GFX7: [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32)
; GFX7: [[BITCAST2:%[0-9]+]]:_(s32) = G_BITCAST [[COPY2]](<2 x s16>)
; GFX7: [[TRUNC4:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST2]](s32)
; GFX7: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST2]], [[C]](s32)
; GFX7: [[TRUNC5:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR2]](s32)
; GFX7: [[FMUL:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC]], [[TRUNC2]]
; GFX7: [[FADD:%[0-9]+]]:_(s16) = G_FADD [[FMUL]], [[TRUNC4]]
; GFX7: [[FMUL1:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC1]], [[TRUNC3]]
; GFX7: [[FADD1:%[0-9]+]]:_(s16) = G_FADD [[FMUL1]], [[TRUNC5]]
; GFX7: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[FADD]](s16)
; GFX7: [[ZEXT1:%[0-9]+]]:_(s32) = G_ZEXT [[FADD1]](s16)
; GFX7: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[ZEXT1]], [[C]](s32)
; GFX7: [[OR:%[0-9]+]]:_(s32) = G_OR [[ZEXT]], [[SHL]]
; GFX7: [[BITCAST3:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32)
; GFX7: $vgpr0 = COPY [[BITCAST3]](<2 x s16>)
; GFX10-LABEL: name: test_fmad_v2s16_denorm
; GFX10: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
; GFX10: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
; GFX10: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
; GFX10: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>)
; GFX10: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32)
; GFX10: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
; GFX10: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
; GFX10: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
; GFX10: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[COPY1]](<2 x s16>)
; GFX10: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32)
; GFX10: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
; GFX10: [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32)
; GFX10: [[BITCAST2:%[0-9]+]]:_(s32) = G_BITCAST [[COPY2]](<2 x s16>)
; GFX10: [[TRUNC4:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST2]](s32)
; GFX10: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST2]], [[C]](s32)
; GFX10: [[TRUNC5:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR2]](s32)
; GFX10: [[FMUL:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC]], [[TRUNC2]]
; GFX10: [[FADD:%[0-9]+]]:_(s16) = G_FADD [[FMUL]], [[TRUNC4]]
; GFX10: [[FMUL1:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC1]], [[TRUNC3]]
; GFX10: [[FADD1:%[0-9]+]]:_(s16) = G_FADD [[FMUL1]], [[TRUNC5]]
; GFX10: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD]](s16)
; GFX10: [[ANYEXT1:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD1]](s16)
; GFX10: [[BUILD_VECTOR_TRUNC:%[0-9]+]]:_(<2 x s16>) = G_BUILD_VECTOR_TRUNC [[ANYEXT]](s32), [[ANYEXT1]](s32)
; GFX10: $vgpr0 = COPY [[BUILD_VECTOR_TRUNC]](<2 x s16>)
%0:_(<2 x s16>) = COPY $vgpr0
%1:_(<2 x s16>) = COPY $vgpr1
%2:_(<2 x s16>) = COPY $vgpr2
%3:_(<2 x s16>) = G_FMAD %0, %1, %2
$vgpr0 = COPY %3
...
---
name: test_fmad_v2s16_denorm_flags
machineFunctionInfo:
mode:
fp64-fp16-input-denormals: true
fp64-fp16-output-denormals: true
body: |
bb.0:
liveins: $vgpr0, $vgpr1, $vgpr2
; GFX6-LABEL: name: test_fmad_v2s16_denorm_flags
; GFX6: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
; GFX6: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
; GFX6: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
; GFX6: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>)
; GFX6: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32)
; GFX6: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
; GFX6: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
; GFX6: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
; GFX6: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[COPY1]](<2 x s16>)
; GFX6: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32)
; GFX6: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
; GFX6: [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32)
; GFX6: [[BITCAST2:%[0-9]+]]:_(s32) = G_BITCAST [[COPY2]](<2 x s16>)
; GFX6: [[TRUNC4:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST2]](s32)
; GFX6: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST2]], [[C]](s32)
; GFX6: [[TRUNC5:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR2]](s32)
; GFX6: [[FPEXT:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC]](s16)
; GFX6: [[FPEXT1:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC2]](s16)
; GFX6: [[FMUL:%[0-9]+]]:_(s32) = nnan G_FMUL [[FPEXT]], [[FPEXT1]]
; GFX6: [[FPTRUNC:%[0-9]+]]:_(s16) = G_FPTRUNC [[FMUL]](s32)
; GFX6: [[FPEXT2:%[0-9]+]]:_(s32) = G_FPEXT [[FPTRUNC]](s16)
; GFX6: [[FPEXT3:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC4]](s16)
; GFX6: [[FADD:%[0-9]+]]:_(s32) = nnan G_FADD [[FPEXT2]], [[FPEXT3]]
; GFX6: [[FPTRUNC1:%[0-9]+]]:_(s16) = G_FPTRUNC [[FADD]](s32)
; GFX6: [[FPEXT4:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC1]](s16)
; GFX6: [[FPEXT5:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC3]](s16)
; GFX6: [[FMUL1:%[0-9]+]]:_(s32) = nnan G_FMUL [[FPEXT4]], [[FPEXT5]]
; GFX6: [[FPTRUNC2:%[0-9]+]]:_(s16) = G_FPTRUNC [[FMUL1]](s32)
; GFX6: [[FPEXT6:%[0-9]+]]:_(s32) = G_FPEXT [[FPTRUNC2]](s16)
; GFX6: [[FPEXT7:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC5]](s16)
; GFX6: [[FADD1:%[0-9]+]]:_(s32) = nnan G_FADD [[FPEXT6]], [[FPEXT7]]
; GFX6: [[FPTRUNC3:%[0-9]+]]:_(s16) = G_FPTRUNC [[FADD1]](s32)
; GFX6: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[FPTRUNC1]](s16)
; GFX6: [[ZEXT1:%[0-9]+]]:_(s32) = G_ZEXT [[FPTRUNC3]](s16)
; GFX6: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[ZEXT1]], [[C]](s32)
; GFX6: [[OR:%[0-9]+]]:_(s32) = G_OR [[ZEXT]], [[SHL]]
; GFX6: [[BITCAST3:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32)
; GFX6: $vgpr0 = COPY [[BITCAST3]](<2 x s16>)
; GFX7-LABEL: name: test_fmad_v2s16_denorm_flags
; GFX7: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
; GFX7: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
; GFX7: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
; GFX7: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>)
; GFX7: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32)
; GFX7: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
; GFX7: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
; GFX7: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
; GFX7: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[COPY1]](<2 x s16>)
; GFX7: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32)
; GFX7: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
; GFX7: [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32)
; GFX7: [[BITCAST2:%[0-9]+]]:_(s32) = G_BITCAST [[COPY2]](<2 x s16>)
; GFX7: [[TRUNC4:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST2]](s32)
; GFX7: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST2]], [[C]](s32)
; GFX7: [[TRUNC5:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR2]](s32)
; GFX7: [[FMUL:%[0-9]+]]:_(s16) = nnan G_FMUL [[TRUNC]], [[TRUNC2]]
; GFX7: [[FADD:%[0-9]+]]:_(s16) = nnan G_FADD [[FMUL]], [[TRUNC4]]
; GFX7: [[FMUL1:%[0-9]+]]:_(s16) = nnan G_FMUL [[TRUNC1]], [[TRUNC3]]
; GFX7: [[FADD1:%[0-9]+]]:_(s16) = nnan G_FADD [[FMUL1]], [[TRUNC5]]
; GFX7: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[FADD]](s16)
; GFX7: [[ZEXT1:%[0-9]+]]:_(s32) = G_ZEXT [[FADD1]](s16)
; GFX7: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[ZEXT1]], [[C]](s32)
; GFX7: [[OR:%[0-9]+]]:_(s32) = G_OR [[ZEXT]], [[SHL]]
; GFX7: [[BITCAST3:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32)
; GFX7: $vgpr0 = COPY [[BITCAST3]](<2 x s16>)
; GFX10-LABEL: name: test_fmad_v2s16_denorm_flags
; GFX10: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
; GFX10: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
; GFX10: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
; GFX10: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>)
; GFX10: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32)
; GFX10: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
; GFX10: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
; GFX10: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
; GFX10: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[COPY1]](<2 x s16>)
; GFX10: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32)
; GFX10: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
; GFX10: [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32)
; GFX10: [[BITCAST2:%[0-9]+]]:_(s32) = G_BITCAST [[COPY2]](<2 x s16>)
; GFX10: [[TRUNC4:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST2]](s32)
; GFX10: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST2]], [[C]](s32)
; GFX10: [[TRUNC5:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR2]](s32)
; GFX10: [[FMUL:%[0-9]+]]:_(s16) = nnan G_FMUL [[TRUNC]], [[TRUNC2]]
; GFX10: [[FADD:%[0-9]+]]:_(s16) = nnan G_FADD [[FMUL]], [[TRUNC4]]
; GFX10: [[FMUL1:%[0-9]+]]:_(s16) = nnan G_FMUL [[TRUNC1]], [[TRUNC3]]
; GFX10: [[FADD1:%[0-9]+]]:_(s16) = nnan G_FADD [[FMUL1]], [[TRUNC5]]
; GFX10: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD]](s16)
; GFX10: [[ANYEXT1:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD1]](s16)
; GFX10: [[BUILD_VECTOR_TRUNC:%[0-9]+]]:_(<2 x s16>) = G_BUILD_VECTOR_TRUNC [[ANYEXT]](s32), [[ANYEXT1]](s32)
; GFX10: $vgpr0 = COPY [[BUILD_VECTOR_TRUNC]](<2 x s16>)
%0:_(<2 x s16>) = COPY $vgpr0
%1:_(<2 x s16>) = COPY $vgpr1
%2:_(<2 x s16>) = COPY $vgpr2
%3:_(<2 x s16>) = nnan G_FMAD %0, %1, %2
$vgpr0 = COPY %3
...
---
name: test_fmad_v4s16_denorm
machineFunctionInfo:
mode:
fp64-fp16-input-denormals: true
fp64-fp16-output-denormals: true
body: |
bb.0:
liveins: $vgpr0_vgpr1, $vgpr2_vgpr3, $vgpr4_vgpr5
; GFX6-LABEL: name: test_fmad_v4s16_denorm
; GFX6: [[COPY:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr0_vgpr1
; GFX6: [[COPY1:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr2_vgpr3
; GFX6: [[COPY2:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr4_vgpr5
; GFX6: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY]](<4 x s16>)
; GFX6: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[UV]](<2 x s16>)
; GFX6: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32)
; GFX6: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
; GFX6: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
; GFX6: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
; GFX6: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[UV1]](<2 x s16>)
; GFX6: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32)
; GFX6: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
; GFX6: [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32)
; GFX6: [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY1]](<4 x s16>)
; GFX6: [[BITCAST2:%[0-9]+]]:_(s32) = G_BITCAST [[UV2]](<2 x s16>)
; GFX6: [[TRUNC4:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST2]](s32)
; GFX6: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST2]], [[C]](s32)
; GFX6: [[TRUNC5:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR2]](s32)
; GFX6: [[BITCAST3:%[0-9]+]]:_(s32) = G_BITCAST [[UV3]](<2 x s16>)
; GFX6: [[TRUNC6:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST3]](s32)
; GFX6: [[LSHR3:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST3]], [[C]](s32)
; GFX6: [[TRUNC7:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR3]](s32)
; GFX6: [[UV4:%[0-9]+]]:_(<2 x s16>), [[UV5:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY2]](<4 x s16>)
; GFX6: [[BITCAST4:%[0-9]+]]:_(s32) = G_BITCAST [[UV4]](<2 x s16>)
; GFX6: [[TRUNC8:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST4]](s32)
; GFX6: [[LSHR4:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST4]], [[C]](s32)
; GFX6: [[TRUNC9:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR4]](s32)
; GFX6: [[BITCAST5:%[0-9]+]]:_(s32) = G_BITCAST [[UV5]](<2 x s16>)
; GFX6: [[TRUNC10:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST5]](s32)
; GFX6: [[LSHR5:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST5]], [[C]](s32)
; GFX6: [[TRUNC11:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR5]](s32)
; GFX6: [[FPEXT:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC]](s16)
; GFX6: [[FPEXT1:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC4]](s16)
; GFX6: [[FMUL:%[0-9]+]]:_(s32) = G_FMUL [[FPEXT]], [[FPEXT1]]
; GFX6: [[FPTRUNC:%[0-9]+]]:_(s16) = G_FPTRUNC [[FMUL]](s32)
; GFX6: [[FPEXT2:%[0-9]+]]:_(s32) = G_FPEXT [[FPTRUNC]](s16)
; GFX6: [[FPEXT3:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC8]](s16)
; GFX6: [[FADD:%[0-9]+]]:_(s32) = G_FADD [[FPEXT2]], [[FPEXT3]]
; GFX6: [[FPTRUNC1:%[0-9]+]]:_(s16) = G_FPTRUNC [[FADD]](s32)
; GFX6: [[FPEXT4:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC1]](s16)
; GFX6: [[FPEXT5:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC5]](s16)
; GFX6: [[FMUL1:%[0-9]+]]:_(s32) = G_FMUL [[FPEXT4]], [[FPEXT5]]
; GFX6: [[FPTRUNC2:%[0-9]+]]:_(s16) = G_FPTRUNC [[FMUL1]](s32)
; GFX6: [[FPEXT6:%[0-9]+]]:_(s32) = G_FPEXT [[FPTRUNC2]](s16)
; GFX6: [[FPEXT7:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC9]](s16)
; GFX6: [[FADD1:%[0-9]+]]:_(s32) = G_FADD [[FPEXT6]], [[FPEXT7]]
; GFX6: [[FPTRUNC3:%[0-9]+]]:_(s16) = G_FPTRUNC [[FADD1]](s32)
; GFX6: [[FPEXT8:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC2]](s16)
; GFX6: [[FPEXT9:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC6]](s16)
; GFX6: [[FMUL2:%[0-9]+]]:_(s32) = G_FMUL [[FPEXT8]], [[FPEXT9]]
; GFX6: [[FPTRUNC4:%[0-9]+]]:_(s16) = G_FPTRUNC [[FMUL2]](s32)
; GFX6: [[FPEXT10:%[0-9]+]]:_(s32) = G_FPEXT [[FPTRUNC4]](s16)
; GFX6: [[FPEXT11:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC10]](s16)
; GFX6: [[FADD2:%[0-9]+]]:_(s32) = G_FADD [[FPEXT10]], [[FPEXT11]]
; GFX6: [[FPTRUNC5:%[0-9]+]]:_(s16) = G_FPTRUNC [[FADD2]](s32)
; GFX6: [[FPEXT12:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC3]](s16)
; GFX6: [[FPEXT13:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC7]](s16)
; GFX6: [[FMUL3:%[0-9]+]]:_(s32) = G_FMUL [[FPEXT12]], [[FPEXT13]]
; GFX6: [[FPTRUNC6:%[0-9]+]]:_(s16) = G_FPTRUNC [[FMUL3]](s32)
; GFX6: [[FPEXT14:%[0-9]+]]:_(s32) = G_FPEXT [[FPTRUNC6]](s16)
; GFX6: [[FPEXT15:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC11]](s16)
; GFX6: [[FADD3:%[0-9]+]]:_(s32) = G_FADD [[FPEXT14]], [[FPEXT15]]
; GFX6: [[FPTRUNC7:%[0-9]+]]:_(s16) = G_FPTRUNC [[FADD3]](s32)
; GFX6: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[FPTRUNC1]](s16)
; GFX6: [[ZEXT1:%[0-9]+]]:_(s32) = G_ZEXT [[FPTRUNC3]](s16)
; GFX6: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[ZEXT1]], [[C]](s32)
; GFX6: [[OR:%[0-9]+]]:_(s32) = G_OR [[ZEXT]], [[SHL]]
; GFX6: [[BITCAST6:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32)
; GFX6: [[ZEXT2:%[0-9]+]]:_(s32) = G_ZEXT [[FPTRUNC5]](s16)
; GFX6: [[ZEXT3:%[0-9]+]]:_(s32) = G_ZEXT [[FPTRUNC7]](s16)
; GFX6: [[SHL1:%[0-9]+]]:_(s32) = G_SHL [[ZEXT3]], [[C]](s32)
; GFX6: [[OR1:%[0-9]+]]:_(s32) = G_OR [[ZEXT2]], [[SHL1]]
; GFX6: [[BITCAST7:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR1]](s32)
; GFX6: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[BITCAST6]](<2 x s16>), [[BITCAST7]](<2 x s16>)
; GFX6: $vgpr0_vgpr1 = COPY [[CONCAT_VECTORS]](<4 x s16>)
; GFX7-LABEL: name: test_fmad_v4s16_denorm
; GFX7: [[COPY:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr0_vgpr1
; GFX7: [[COPY1:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr2_vgpr3
; GFX7: [[COPY2:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr4_vgpr5
; GFX7: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY]](<4 x s16>)
; GFX7: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[UV]](<2 x s16>)
; GFX7: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32)
; GFX7: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
; GFX7: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
; GFX7: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
; GFX7: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[UV1]](<2 x s16>)
; GFX7: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32)
; GFX7: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
; GFX7: [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32)
; GFX7: [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY1]](<4 x s16>)
; GFX7: [[BITCAST2:%[0-9]+]]:_(s32) = G_BITCAST [[UV2]](<2 x s16>)
; GFX7: [[TRUNC4:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST2]](s32)
; GFX7: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST2]], [[C]](s32)
; GFX7: [[TRUNC5:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR2]](s32)
; GFX7: [[BITCAST3:%[0-9]+]]:_(s32) = G_BITCAST [[UV3]](<2 x s16>)
; GFX7: [[TRUNC6:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST3]](s32)
; GFX7: [[LSHR3:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST3]], [[C]](s32)
; GFX7: [[TRUNC7:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR3]](s32)
; GFX7: [[UV4:%[0-9]+]]:_(<2 x s16>), [[UV5:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY2]](<4 x s16>)
; GFX7: [[BITCAST4:%[0-9]+]]:_(s32) = G_BITCAST [[UV4]](<2 x s16>)
; GFX7: [[TRUNC8:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST4]](s32)
; GFX7: [[LSHR4:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST4]], [[C]](s32)
; GFX7: [[TRUNC9:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR4]](s32)
; GFX7: [[BITCAST5:%[0-9]+]]:_(s32) = G_BITCAST [[UV5]](<2 x s16>)
; GFX7: [[TRUNC10:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST5]](s32)
; GFX7: [[LSHR5:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST5]], [[C]](s32)
; GFX7: [[TRUNC11:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR5]](s32)
; GFX7: [[FMUL:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC]], [[TRUNC4]]
; GFX7: [[FADD:%[0-9]+]]:_(s16) = G_FADD [[FMUL]], [[TRUNC8]]
; GFX7: [[FMUL1:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC1]], [[TRUNC5]]
; GFX7: [[FADD1:%[0-9]+]]:_(s16) = G_FADD [[FMUL1]], [[TRUNC9]]
; GFX7: [[FMUL2:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC2]], [[TRUNC6]]
; GFX7: [[FADD2:%[0-9]+]]:_(s16) = G_FADD [[FMUL2]], [[TRUNC10]]
; GFX7: [[FMUL3:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC3]], [[TRUNC7]]
; GFX7: [[FADD3:%[0-9]+]]:_(s16) = G_FADD [[FMUL3]], [[TRUNC11]]
; GFX7: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[FADD]](s16)
; GFX7: [[ZEXT1:%[0-9]+]]:_(s32) = G_ZEXT [[FADD1]](s16)
; GFX7: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[ZEXT1]], [[C]](s32)
; GFX7: [[OR:%[0-9]+]]:_(s32) = G_OR [[ZEXT]], [[SHL]]
; GFX7: [[BITCAST6:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32)
; GFX7: [[ZEXT2:%[0-9]+]]:_(s32) = G_ZEXT [[FADD2]](s16)
; GFX7: [[ZEXT3:%[0-9]+]]:_(s32) = G_ZEXT [[FADD3]](s16)
; GFX7: [[SHL1:%[0-9]+]]:_(s32) = G_SHL [[ZEXT3]], [[C]](s32)
; GFX7: [[OR1:%[0-9]+]]:_(s32) = G_OR [[ZEXT2]], [[SHL1]]
; GFX7: [[BITCAST7:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR1]](s32)
; GFX7: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[BITCAST6]](<2 x s16>), [[BITCAST7]](<2 x s16>)
; GFX7: $vgpr0_vgpr1 = COPY [[CONCAT_VECTORS]](<4 x s16>)
; GFX10-LABEL: name: test_fmad_v4s16_denorm
; GFX10: [[COPY:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr0_vgpr1
; GFX10: [[COPY1:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr2_vgpr3
; GFX10: [[COPY2:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr4_vgpr5
; GFX10: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY]](<4 x s16>)
; GFX10: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[UV]](<2 x s16>)
; GFX10: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32)
; GFX10: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
; GFX10: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
; GFX10: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
; GFX10: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[UV1]](<2 x s16>)
; GFX10: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32)
; GFX10: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
; GFX10: [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32)
; GFX10: [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY1]](<4 x s16>)
; GFX10: [[BITCAST2:%[0-9]+]]:_(s32) = G_BITCAST [[UV2]](<2 x s16>)
; GFX10: [[TRUNC4:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST2]](s32)
; GFX10: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST2]], [[C]](s32)
; GFX10: [[TRUNC5:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR2]](s32)
; GFX10: [[BITCAST3:%[0-9]+]]:_(s32) = G_BITCAST [[UV3]](<2 x s16>)
; GFX10: [[TRUNC6:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST3]](s32)
; GFX10: [[LSHR3:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST3]], [[C]](s32)
; GFX10: [[TRUNC7:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR3]](s32)
; GFX10: [[UV4:%[0-9]+]]:_(<2 x s16>), [[UV5:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY2]](<4 x s16>)
; GFX10: [[BITCAST4:%[0-9]+]]:_(s32) = G_BITCAST [[UV4]](<2 x s16>)
; GFX10: [[TRUNC8:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST4]](s32)
; GFX10: [[LSHR4:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST4]], [[C]](s32)
; GFX10: [[TRUNC9:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR4]](s32)
; GFX10: [[BITCAST5:%[0-9]+]]:_(s32) = G_BITCAST [[UV5]](<2 x s16>)
; GFX10: [[TRUNC10:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST5]](s32)
; GFX10: [[LSHR5:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST5]], [[C]](s32)
; GFX10: [[TRUNC11:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR5]](s32)
; GFX10: [[FMUL:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC]], [[TRUNC4]]
; GFX10: [[FADD:%[0-9]+]]:_(s16) = G_FADD [[FMUL]], [[TRUNC8]]
; GFX10: [[FMUL1:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC1]], [[TRUNC5]]
; GFX10: [[FADD1:%[0-9]+]]:_(s16) = G_FADD [[FMUL1]], [[TRUNC9]]
; GFX10: [[FMUL2:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC2]], [[TRUNC6]]
; GFX10: [[FADD2:%[0-9]+]]:_(s16) = G_FADD [[FMUL2]], [[TRUNC10]]
; GFX10: [[FMUL3:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC3]], [[TRUNC7]]
; GFX10: [[FADD3:%[0-9]+]]:_(s16) = G_FADD [[FMUL3]], [[TRUNC11]]
; GFX10: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD]](s16)
; GFX10: [[ANYEXT1:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD1]](s16)
; GFX10: [[BUILD_VECTOR_TRUNC:%[0-9]+]]:_(<2 x s16>) = G_BUILD_VECTOR_TRUNC [[ANYEXT]](s32), [[ANYEXT1]](s32)
; GFX10: [[ANYEXT2:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD2]](s16)
; GFX10: [[ANYEXT3:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD3]](s16)
; GFX10: [[BUILD_VECTOR_TRUNC1:%[0-9]+]]:_(<2 x s16>) = G_BUILD_VECTOR_TRUNC [[ANYEXT2]](s32), [[ANYEXT3]](s32)
; GFX10: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[BUILD_VECTOR_TRUNC]](<2 x s16>), [[BUILD_VECTOR_TRUNC1]](<2 x s16>)
; GFX10: $vgpr0_vgpr1 = COPY [[CONCAT_VECTORS]](<4 x s16>)
%0:_(<4 x s16>) = COPY $vgpr0_vgpr1
%1:_(<4 x s16>) = COPY $vgpr2_vgpr3
%2:_(<4 x s16>) = COPY $vgpr4_vgpr5
%3:_(<4 x s16>) = G_FMAD %0, %1, %2
$vgpr0_vgpr1 = COPY %3
...
---
name: test_fmad_v4s16_denorm_flags
machineFunctionInfo:
mode:
fp64-fp16-input-denormals: true
fp64-fp16-output-denormals: true
body: |
bb.0:
liveins: $vgpr0_vgpr1, $vgpr2_vgpr3, $vgpr4_vgpr5
; GFX6-LABEL: name: test_fmad_v4s16_denorm_flags
; GFX6: [[COPY:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr0_vgpr1
; GFX6: [[COPY1:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr2_vgpr3
; GFX6: [[COPY2:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr4_vgpr5
; GFX6: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY]](<4 x s16>)
; GFX6: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[UV]](<2 x s16>)
; GFX6: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32)
; GFX6: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
; GFX6: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
; GFX6: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
; GFX6: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[UV1]](<2 x s16>)
; GFX6: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32)
; GFX6: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
; GFX6: [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32)
; GFX6: [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY1]](<4 x s16>)
; GFX6: [[BITCAST2:%[0-9]+]]:_(s32) = G_BITCAST [[UV2]](<2 x s16>)
; GFX6: [[TRUNC4:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST2]](s32)
; GFX6: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST2]], [[C]](s32)
; GFX6: [[TRUNC5:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR2]](s32)
; GFX6: [[BITCAST3:%[0-9]+]]:_(s32) = G_BITCAST [[UV3]](<2 x s16>)
; GFX6: [[TRUNC6:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST3]](s32)
; GFX6: [[LSHR3:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST3]], [[C]](s32)
; GFX6: [[TRUNC7:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR3]](s32)
; GFX6: [[UV4:%[0-9]+]]:_(<2 x s16>), [[UV5:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY2]](<4 x s16>)
; GFX6: [[BITCAST4:%[0-9]+]]:_(s32) = G_BITCAST [[UV4]](<2 x s16>)
; GFX6: [[TRUNC8:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST4]](s32)
; GFX6: [[LSHR4:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST4]], [[C]](s32)
; GFX6: [[TRUNC9:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR4]](s32)
; GFX6: [[BITCAST5:%[0-9]+]]:_(s32) = G_BITCAST [[UV5]](<2 x s16>)
; GFX6: [[TRUNC10:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST5]](s32)
; GFX6: [[LSHR5:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST5]], [[C]](s32)
; GFX6: [[TRUNC11:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR5]](s32)
; GFX6: [[FPEXT:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC]](s16)
; GFX6: [[FPEXT1:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC4]](s16)
; GFX6: [[FMUL:%[0-9]+]]:_(s32) = nnan G_FMUL [[FPEXT]], [[FPEXT1]]
; GFX6: [[FPTRUNC:%[0-9]+]]:_(s16) = G_FPTRUNC [[FMUL]](s32)
; GFX6: [[FPEXT2:%[0-9]+]]:_(s32) = G_FPEXT [[FPTRUNC]](s16)
; GFX6: [[FPEXT3:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC8]](s16)
; GFX6: [[FADD:%[0-9]+]]:_(s32) = nnan G_FADD [[FPEXT2]], [[FPEXT3]]
; GFX6: [[FPTRUNC1:%[0-9]+]]:_(s16) = G_FPTRUNC [[FADD]](s32)
; GFX6: [[FPEXT4:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC1]](s16)
; GFX6: [[FPEXT5:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC5]](s16)
; GFX6: [[FMUL1:%[0-9]+]]:_(s32) = nnan G_FMUL [[FPEXT4]], [[FPEXT5]]
; GFX6: [[FPTRUNC2:%[0-9]+]]:_(s16) = G_FPTRUNC [[FMUL1]](s32)
; GFX6: [[FPEXT6:%[0-9]+]]:_(s32) = G_FPEXT [[FPTRUNC2]](s16)
; GFX6: [[FPEXT7:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC9]](s16)
; GFX6: [[FADD1:%[0-9]+]]:_(s32) = nnan G_FADD [[FPEXT6]], [[FPEXT7]]
; GFX6: [[FPTRUNC3:%[0-9]+]]:_(s16) = G_FPTRUNC [[FADD1]](s32)
; GFX6: [[FPEXT8:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC2]](s16)
; GFX6: [[FPEXT9:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC6]](s16)
; GFX6: [[FMUL2:%[0-9]+]]:_(s32) = nnan G_FMUL [[FPEXT8]], [[FPEXT9]]
; GFX6: [[FPTRUNC4:%[0-9]+]]:_(s16) = G_FPTRUNC [[FMUL2]](s32)
; GFX6: [[FPEXT10:%[0-9]+]]:_(s32) = G_FPEXT [[FPTRUNC4]](s16)
; GFX6: [[FPEXT11:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC10]](s16)
; GFX6: [[FADD2:%[0-9]+]]:_(s32) = nnan G_FADD [[FPEXT10]], [[FPEXT11]]
; GFX6: [[FPTRUNC5:%[0-9]+]]:_(s16) = G_FPTRUNC [[FADD2]](s32)
; GFX6: [[FPEXT12:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC3]](s16)
; GFX6: [[FPEXT13:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC7]](s16)
; GFX6: [[FMUL3:%[0-9]+]]:_(s32) = nnan G_FMUL [[FPEXT12]], [[FPEXT13]]
; GFX6: [[FPTRUNC6:%[0-9]+]]:_(s16) = G_FPTRUNC [[FMUL3]](s32)
; GFX6: [[FPEXT14:%[0-9]+]]:_(s32) = G_FPEXT [[FPTRUNC6]](s16)
; GFX6: [[FPEXT15:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC11]](s16)
; GFX6: [[FADD3:%[0-9]+]]:_(s32) = nnan G_FADD [[FPEXT14]], [[FPEXT15]]
; GFX6: [[FPTRUNC7:%[0-9]+]]:_(s16) = G_FPTRUNC [[FADD3]](s32)
; GFX6: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[FPTRUNC1]](s16)
; GFX6: [[ZEXT1:%[0-9]+]]:_(s32) = G_ZEXT [[FPTRUNC3]](s16)
; GFX6: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[ZEXT1]], [[C]](s32)
; GFX6: [[OR:%[0-9]+]]:_(s32) = G_OR [[ZEXT]], [[SHL]]
; GFX6: [[BITCAST6:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32)
; GFX6: [[ZEXT2:%[0-9]+]]:_(s32) = G_ZEXT [[FPTRUNC5]](s16)
; GFX6: [[ZEXT3:%[0-9]+]]:_(s32) = G_ZEXT [[FPTRUNC7]](s16)
; GFX6: [[SHL1:%[0-9]+]]:_(s32) = G_SHL [[ZEXT3]], [[C]](s32)
; GFX6: [[OR1:%[0-9]+]]:_(s32) = G_OR [[ZEXT2]], [[SHL1]]
; GFX6: [[BITCAST7:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR1]](s32)
; GFX6: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[BITCAST6]](<2 x s16>), [[BITCAST7]](<2 x s16>)
; GFX6: $vgpr0_vgpr1 = COPY [[CONCAT_VECTORS]](<4 x s16>)
; GFX7-LABEL: name: test_fmad_v4s16_denorm_flags
; GFX7: [[COPY:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr0_vgpr1
; GFX7: [[COPY1:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr2_vgpr3
; GFX7: [[COPY2:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr4_vgpr5
; GFX7: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY]](<4 x s16>)
; GFX7: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[UV]](<2 x s16>)
; GFX7: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32)
; GFX7: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
; GFX7: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
; GFX7: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
; GFX7: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[UV1]](<2 x s16>)
; GFX7: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32)
; GFX7: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
; GFX7: [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32)
; GFX7: [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY1]](<4 x s16>)
; GFX7: [[BITCAST2:%[0-9]+]]:_(s32) = G_BITCAST [[UV2]](<2 x s16>)
; GFX7: [[TRUNC4:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST2]](s32)
; GFX7: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST2]], [[C]](s32)
; GFX7: [[TRUNC5:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR2]](s32)
; GFX7: [[BITCAST3:%[0-9]+]]:_(s32) = G_BITCAST [[UV3]](<2 x s16>)
; GFX7: [[TRUNC6:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST3]](s32)
; GFX7: [[LSHR3:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST3]], [[C]](s32)
; GFX7: [[TRUNC7:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR3]](s32)
; GFX7: [[UV4:%[0-9]+]]:_(<2 x s16>), [[UV5:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY2]](<4 x s16>)
; GFX7: [[BITCAST4:%[0-9]+]]:_(s32) = G_BITCAST [[UV4]](<2 x s16>)
; GFX7: [[TRUNC8:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST4]](s32)
; GFX7: [[LSHR4:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST4]], [[C]](s32)
; GFX7: [[TRUNC9:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR4]](s32)
; GFX7: [[BITCAST5:%[0-9]+]]:_(s32) = G_BITCAST [[UV5]](<2 x s16>)
; GFX7: [[TRUNC10:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST5]](s32)
; GFX7: [[LSHR5:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST5]], [[C]](s32)
; GFX7: [[TRUNC11:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR5]](s32)
; GFX7: [[FMUL:%[0-9]+]]:_(s16) = nnan G_FMUL [[TRUNC]], [[TRUNC4]]
; GFX7: [[FADD:%[0-9]+]]:_(s16) = nnan G_FADD [[FMUL]], [[TRUNC8]]
; GFX7: [[FMUL1:%[0-9]+]]:_(s16) = nnan G_FMUL [[TRUNC1]], [[TRUNC5]]
; GFX7: [[FADD1:%[0-9]+]]:_(s16) = nnan G_FADD [[FMUL1]], [[TRUNC9]]
; GFX7: [[FMUL2:%[0-9]+]]:_(s16) = nnan G_FMUL [[TRUNC2]], [[TRUNC6]]
; GFX7: [[FADD2:%[0-9]+]]:_(s16) = nnan G_FADD [[FMUL2]], [[TRUNC10]]
; GFX7: [[FMUL3:%[0-9]+]]:_(s16) = nnan G_FMUL [[TRUNC3]], [[TRUNC7]]
; GFX7: [[FADD3:%[0-9]+]]:_(s16) = nnan G_FADD [[FMUL3]], [[TRUNC11]]
; GFX7: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[FADD]](s16)
; GFX7: [[ZEXT1:%[0-9]+]]:_(s32) = G_ZEXT [[FADD1]](s16)
; GFX7: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[ZEXT1]], [[C]](s32)
; GFX7: [[OR:%[0-9]+]]:_(s32) = G_OR [[ZEXT]], [[SHL]]
; GFX7: [[BITCAST6:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32)
; GFX7: [[ZEXT2:%[0-9]+]]:_(s32) = G_ZEXT [[FADD2]](s16)
; GFX7: [[ZEXT3:%[0-9]+]]:_(s32) = G_ZEXT [[FADD3]](s16)
; GFX7: [[SHL1:%[0-9]+]]:_(s32) = G_SHL [[ZEXT3]], [[C]](s32)
; GFX7: [[OR1:%[0-9]+]]:_(s32) = G_OR [[ZEXT2]], [[SHL1]]
; GFX7: [[BITCAST7:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR1]](s32)
; GFX7: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[BITCAST6]](<2 x s16>), [[BITCAST7]](<2 x s16>)
; GFX7: $vgpr0_vgpr1 = COPY [[CONCAT_VECTORS]](<4 x s16>)
; GFX10-LABEL: name: test_fmad_v4s16_denorm_flags
; GFX10: [[COPY:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr0_vgpr1
; GFX10: [[COPY1:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr2_vgpr3
; GFX10: [[COPY2:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr4_vgpr5
; GFX10: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY]](<4 x s16>)
; GFX10: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[UV]](<2 x s16>)
; GFX10: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32)
; GFX10: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
; GFX10: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
; GFX10: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
; GFX10: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[UV1]](<2 x s16>)
; GFX10: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32)
; GFX10: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
; GFX10: [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32)
; GFX10: [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY1]](<4 x s16>)
; GFX10: [[BITCAST2:%[0-9]+]]:_(s32) = G_BITCAST [[UV2]](<2 x s16>)
; GFX10: [[TRUNC4:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST2]](s32)
; GFX10: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST2]], [[C]](s32)
; GFX10: [[TRUNC5:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR2]](s32)
; GFX10: [[BITCAST3:%[0-9]+]]:_(s32) = G_BITCAST [[UV3]](<2 x s16>)
; GFX10: [[TRUNC6:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST3]](s32)
; GFX10: [[LSHR3:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST3]], [[C]](s32)
; GFX10: [[TRUNC7:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR3]](s32)
; GFX10: [[UV4:%[0-9]+]]:_(<2 x s16>), [[UV5:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY2]](<4 x s16>)
; GFX10: [[BITCAST4:%[0-9]+]]:_(s32) = G_BITCAST [[UV4]](<2 x s16>)
; GFX10: [[TRUNC8:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST4]](s32)
; GFX10: [[LSHR4:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST4]], [[C]](s32)
; GFX10: [[TRUNC9:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR4]](s32)
; GFX10: [[BITCAST5:%[0-9]+]]:_(s32) = G_BITCAST [[UV5]](<2 x s16>)
; GFX10: [[TRUNC10:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST5]](s32)
; GFX10: [[LSHR5:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST5]], [[C]](s32)
; GFX10: [[TRUNC11:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR5]](s32)
; GFX10: [[FMUL:%[0-9]+]]:_(s16) = nnan G_FMUL [[TRUNC]], [[TRUNC4]]
; GFX10: [[FADD:%[0-9]+]]:_(s16) = nnan G_FADD [[FMUL]], [[TRUNC8]]
; GFX10: [[FMUL1:%[0-9]+]]:_(s16) = nnan G_FMUL [[TRUNC1]], [[TRUNC5]]
; GFX10: [[FADD1:%[0-9]+]]:_(s16) = nnan G_FADD [[FMUL1]], [[TRUNC9]]
; GFX10: [[FMUL2:%[0-9]+]]:_(s16) = nnan G_FMUL [[TRUNC2]], [[TRUNC6]]
; GFX10: [[FADD2:%[0-9]+]]:_(s16) = nnan G_FADD [[FMUL2]], [[TRUNC10]]
; GFX10: [[FMUL3:%[0-9]+]]:_(s16) = nnan G_FMUL [[TRUNC3]], [[TRUNC7]]
; GFX10: [[FADD3:%[0-9]+]]:_(s16) = nnan G_FADD [[FMUL3]], [[TRUNC11]]
; GFX10: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD]](s16)
; GFX10: [[ANYEXT1:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD1]](s16)
; GFX10: [[BUILD_VECTOR_TRUNC:%[0-9]+]]:_(<2 x s16>) = G_BUILD_VECTOR_TRUNC [[ANYEXT]](s32), [[ANYEXT1]](s32)
; GFX10: [[ANYEXT2:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD2]](s16)
; GFX10: [[ANYEXT3:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD3]](s16)
; GFX10: [[BUILD_VECTOR_TRUNC1:%[0-9]+]]:_(<2 x s16>) = G_BUILD_VECTOR_TRUNC [[ANYEXT2]](s32), [[ANYEXT3]](s32)
; GFX10: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[BUILD_VECTOR_TRUNC]](<2 x s16>), [[BUILD_VECTOR_TRUNC1]](<2 x s16>)
; GFX10: $vgpr0_vgpr1 = COPY [[CONCAT_VECTORS]](<4 x s16>)
%0:_(<4 x s16>) = COPY $vgpr0_vgpr1
%1:_(<4 x s16>) = COPY $vgpr2_vgpr3
%2:_(<4 x s16>) = COPY $vgpr4_vgpr5
%3:_(<4 x s16>) = nnan G_FMAD %0, %1, %2
$vgpr0_vgpr1 = COPY %3
...