@@ -417,7 +417,7 @@ body: |
417
417
# GCN-LABEL: name: xdl_sgemm16x16_mfma_write_agpr_mfma_srca_read_overlap
418
418
# GCN: V_MFMA
419
419
# GCN-NEXT: S_NOP 7
420
- # GFX940 -NEXT: S_NOP 2
420
+ # GFX942 -NEXT: S_NOP 2
421
421
# GFX950-NEXT: S_NOP 3
422
422
# GCN-NEXT: V_MFMA
423
423
name : xdl_sgemm16x16_mfma_write_agpr_mfma_srca_read_overlap
@@ -440,7 +440,7 @@ body: |
440
440
# GCN-LABEL: name: smfmac32x32_write_agpr_mfma_srca_read_overlap
441
441
# GCN: V_SMFMAC
442
442
# GCN-NEXT: S_NOP 7
443
- # GFX940 -NEXT: S_NOP 2
443
+ # GFX942 -NEXT: S_NOP 2
444
444
# GFX950-NEXT: S_NOP 3
445
445
# GCN-NEXT: V_MFMA
446
446
name : smfmac32x32_write_agpr_mfma_srca_read_overlap
@@ -452,7 +452,7 @@ body: |
452
452
# GCN-LABEL: name: smfmac32x32_write_agpr_smfmac_srcc_read_overlap
453
453
# GCN: V_SMFMAC
454
454
# GCN-NEXT: S_NOP 7
455
- # GFX940 -NEXT: S_NOP 2
455
+ # GFX942 -NEXT: S_NOP 2
456
456
# GFX950-NEXT: S_NOP 3
457
457
# GCN-NEXT: V_SMFMAC
458
458
name : smfmac32x32_write_agpr_smfmac_srcc_read_overlap
@@ -465,7 +465,7 @@ body: |
465
465
# GCN: V_MFMA
466
466
# GCN-NEXT: S_NOP 7
467
467
# GCN-NEXT: S_NOP 7
468
- # GFX940 -NEXT: S_NOP 2
468
+ # GFX942 -NEXT: S_NOP 2
469
469
# GFX950-NEXT: S_NOP 3
470
470
# GCN-NEXT: V_MFMA
471
471
name : xdl_sgemm32x32_mfma_write_agpr_mfma_srca_read_overlap
@@ -734,7 +734,7 @@ body: |
734
734
...
735
735
# GCN-LABEL: name: smfmac16x16_write_vgpr_flat_read
736
736
# GCN: V_SMFMAC
737
- # GFX940 -NEXT: S_NOP 6
737
+ # GFX942 -NEXT: S_NOP 6
738
738
# GFX950-NEXT: S_NOP 7
739
739
# GCN-NEXT: FLAT_STORE_DWORD
740
740
name : smfmac16x16_write_vgpr_flat_read
@@ -746,7 +746,7 @@ body: |
746
746
# GCN-LABEL: name: xdl_smfma16x16_write_vgpr_flat_read
747
747
# GCN: V_MFMA
748
748
# GCN-NEXT: S_NOP 7
749
- # GFX940 -NEXT: S_NOP 2
749
+ # GFX942 -NEXT: S_NOP 2
750
750
# GFX950-NEXT: S_NOP 3
751
751
# GCN-NEXT: FLAT_STORE_DWORD
752
752
name : xdl_smfma16x16_write_vgpr_flat_read
@@ -758,7 +758,7 @@ body: |
758
758
# GCN-LABEL: name: smfmac32x32_write_vgpr_flat_read
759
759
# GCN: V_SMFMAC
760
760
# GCN-NEXT: S_NOP 7
761
- # GFX940 -NEXT: S_NOP 2
761
+ # GFX942 -NEXT: S_NOP 2
762
762
# GFX950-NEXT: S_NOP 3
763
763
# GCN-NEXT: FLAT_STORE_DWORD
764
764
name : smfmac32x32_write_vgpr_flat_read
@@ -771,7 +771,7 @@ body: |
771
771
# GCN: V_MFMA
772
772
# GCN-NEXT: S_NOP 7
773
773
# GCN-NEXT: S_NOP 7
774
- # GFX940 -NEXT: S_NOP 2
774
+ # GFX942 -NEXT: S_NOP 2
775
775
# GFX950-NEXT: S_NOP 3
776
776
# GCN-NEXT: FLAT_STORE_DWORD
777
777
name : xdl_smfma32x32_write_vgpr_flat_read
@@ -827,7 +827,7 @@ body: |
827
827
# GCN-LABEL: name: xdl_smfma16x16_write_vgpr_valu_read
828
828
# GCN: V_MFMA
829
829
# GCN-NEXT: S_NOP 7
830
- # GFX940 -NEXT: S_NOP 2
830
+ # GFX942 -NEXT: S_NOP 2
831
831
# GFX950-NEXT: S_NOP 3
832
832
# GCN-NEXT: V_MOV_B32
833
833
name : xdl_smfma16x16_write_vgpr_valu_read
@@ -840,7 +840,7 @@ body: |
840
840
# GCN: V_MFMA
841
841
# GCN-NEXT: S_NOP 7
842
842
# GCN-NEXT: S_NOP 7
843
- # GFX940 -NEXT: S_NOP 2
843
+ # GFX942 -NEXT: S_NOP 2
844
844
# GFX950-NEXT: S_NOP 3
845
845
# GCN-NEXT: V_MOV_B32
846
846
name : xdl_smfma32x32_write_vgpr_valu_read
@@ -887,7 +887,7 @@ body: |
887
887
# GCN-LABEL: name: xdl_smfma16x16_write_vgpr_accv_read
888
888
# GCN: V_MFMA
889
889
# GCN-NEXT: S_NOP 7
890
- # GFX940 -NEXT: S_NOP 2
890
+ # GFX942 -NEXT: S_NOP 2
891
891
# GFX950-NEXT: S_NOP 3
892
892
# GCN-NEXT: V_ACCVGPR_WRITE_B32_e64
893
893
name : xdl_smfma16x16_write_vgpr_accv_read
@@ -900,7 +900,7 @@ body: |
900
900
# GCN: V_MFMA
901
901
# GCN-NEXT: S_NOP 7
902
902
# GCN-NEXT: S_NOP 7
903
- # GFX940 -NEXT: S_NOP 2
903
+ # GFX942 -NEXT: S_NOP 2
904
904
# GFX950-NEXT: S_NOP 3
905
905
# GCN-NEXT: V_ACCVGPR_WRITE_B32_e64
906
906
name : xdl_smfma32x32_write_vgpr_accv_read
@@ -958,7 +958,7 @@ body: |
958
958
# GCN-LABEL: name: xdl_smfma16x16_write_vgpr_valu_write
959
959
# GCN: V_MFMA
960
960
# GCN-NEXT: S_NOP 7
961
- # GFX940 -NEXT: S_NOP 2
961
+ # GFX942 -NEXT: S_NOP 2
962
962
# GFX950-NEXT: S_NOP 3
963
963
# GCN-NEXT: V_MOV_B32
964
964
name : xdl_smfma16x16_write_vgpr_valu_write
@@ -971,7 +971,7 @@ body: |
971
971
# GCN: V_MFMA
972
972
# GCN-NEXT: S_NOP 7
973
973
# GCN-NEXT: S_NOP 7
974
- # GFX940 -NEXT: S_NOP 2
974
+ # GFX942 -NEXT: S_NOP 2
975
975
# GFX950-NEXT: S_NOP 3
976
976
# GCN-NEXT: V_MOV_B32
977
977
name : xdl_smfma32x32_write_vgpr_valu_write
@@ -993,7 +993,7 @@ body: |
993
993
# GCN-LABEL: name: xdl_smfma16x16_write_vgpr_valu_f16_write
994
994
# GCN: V_MFMA
995
995
# GCN-NEXT: S_NOP 7
996
- # GFX940 -NEXT: S_NOP 2
996
+ # GFX942 -NEXT: S_NOP 2
997
997
# GFX950-NEXT: S_NOP 3
998
998
# GCN-NEXT: V_FMA_F16_e64
999
999
name : xdl_smfma16x16_write_vgpr_valu_f16_write
@@ -1006,7 +1006,7 @@ body: |
1006
1006
# GCN: V_MFMA
1007
1007
# GCN-NEXT: S_NOP 7
1008
1008
# GCN-NEXT: S_NOP 7
1009
- # GFX940 -NEXT: S_NOP 2
1009
+ # GFX942 -NEXT: S_NOP 2
1010
1010
# GFX950-NEXT: S_NOP 3
1011
1011
# GCN-NEXT: V_FMA_F16_e64
1012
1012
name : xdl_smfma32x32_write_vgpr_valu_f16_write
@@ -1028,7 +1028,7 @@ body: |
1028
1028
# GCN-LABEL: name: xdl_smfma16x16_write_vgpr_valu_sdwa_write
1029
1029
# GCN: V_MFMA
1030
1030
# GCN-NEXT: S_NOP 7
1031
- # GFX940 -NEXT: S_NOP 2
1031
+ # GFX942 -NEXT: S_NOP 2
1032
1032
# GFX950-NEXT: S_NOP 3
1033
1033
# GCN-NEXT: V_MOV_B32_sdwa
1034
1034
name : xdl_smfma16x16_write_vgpr_valu_sdwa_write
@@ -1041,7 +1041,7 @@ body: |
1041
1041
# GCN: V_MFMA
1042
1042
# GCN-NEXT: S_NOP 7
1043
1043
# GCN-NEXT: S_NOP 7
1044
- # GFX940 -NEXT: S_NOP 2
1044
+ # GFX942 -NEXT: S_NOP 2
1045
1045
# GFX950-NEXT: S_NOP 3
1046
1046
# GCN-NEXT: V_MOV_B32_sdwa
1047
1047
name : xdl_smfma32x32_write_vgpr_valu_sdwa_write
@@ -1733,7 +1733,7 @@ body: |
1733
1733
...
1734
1734
# GCN-LABEL: name: xdl_sgemm16X16X16_mfma_write_agpr_mfma_srca_read_overlap
1735
1735
# GCN: V_MFMA
1736
- # GFX940 -NEXT: S_NOP 6
1736
+ # GFX942 -NEXT: S_NOP 6
1737
1737
# GFX950-NEXT: S_NOP 7
1738
1738
# GCN-NEXT: V_MFMA
1739
1739
name : xdl_sgemm16X16X16_mfma_write_agpr_mfma_srca_read_overlap
@@ -1744,7 +1744,7 @@ body: |
1744
1744
...
1745
1745
# GCN-LABEL: name: xdl_sgemm16X16X32_mfma_write_agpr_mfma_srcb_read_overlap
1746
1746
# GCN: V_MFMA
1747
- # GFX940 -NEXT: S_NOP 6
1747
+ # GFX942 -NEXT: S_NOP 6
1748
1748
# GFX950-NEXT: S_NOP 7
1749
1749
# GCN-NEXT: V_MFMA
1750
1750
name : xdl_sgemm16X16X32_mfma_write_agpr_mfma_srcb_read_overlap
@@ -1755,7 +1755,7 @@ body: |
1755
1755
...
1756
1756
# GCN-LABEL: name: xdl_sgemm16X16X16_mfma_write_vgpr_dmfma16x16_srca_read_overlap
1757
1757
# GCN: V_MFMA
1758
- # GFX940 -NEXT: S_NOP 6
1758
+ # GFX942 -NEXT: S_NOP 6
1759
1759
# GFX950-NEXT: S_NOP 7
1760
1760
# GCN-NEXT: V_MFMA
1761
1761
name : xdl_sgemm16X16X16_mfma_write_vgpr_dmfma16x16_srca_read_overlap
@@ -1766,7 +1766,7 @@ body: |
1766
1766
...
1767
1767
# GCN-LABEL: name: xdl_sgemm16X16X16_mfma_write_vgpr_valu_write
1768
1768
# GCN: V_MFMA
1769
- # GFX940 -NEXT: S_NOP 6
1769
+ # GFX942 -NEXT: S_NOP 6
1770
1770
# GFX950-NEXT: S_NOP 7
1771
1771
# GCN-NEXT: V_MOV_B32
1772
1772
name : xdl_sgemm16X16X16_mfma_write_vgpr_valu_write
@@ -1777,7 +1777,7 @@ body: |
1777
1777
...
1778
1778
# GCN-LABEL: name: xdl_sgemm16X16X16_mfma_write_vgpr_vm_read
1779
1779
# GCN: V_MFMA
1780
- # GFX940 -NEXT: S_NOP 6
1780
+ # GFX942 -NEXT: S_NOP 6
1781
1781
# GFX950-NEXT: S_NOP 7
1782
1782
# GCN-NEXT: BUFFER_STORE_DWORD
1783
1783
name : xdl_sgemm16X16X16_mfma_write_vgpr_vm_read
@@ -1788,7 +1788,7 @@ body: |
1788
1788
...
1789
1789
# GCN-LABEL: name: xdl_sgemm16X16X16_mfma_write_vgpr_valu_read
1790
1790
# GCN: V_MFMA
1791
- # GFX940 -NEXT: S_NOP 6
1791
+ # GFX942 -NEXT: S_NOP 6
1792
1792
# GFX950-NEXT: S_NOP 7
1793
1793
# GCN-NEXT: V_MOV_B32
1794
1794
name : xdl_sgemm16X16X16_mfma_write_vgpr_valu_read
@@ -1799,7 +1799,7 @@ body: |
1799
1799
...
1800
1800
# GCN-LABEL: name: xdl_sgemm16X16X16_mfma_write_vgpr_dot_read
1801
1801
# GCN: V_MFMA
1802
- # GFX940 -NEXT: S_NOP 6
1802
+ # GFX942 -NEXT: S_NOP 6
1803
1803
# GFX950-NEXT: S_NOP 7
1804
1804
# GCN-NEXT: V_DOT
1805
1805
name : xdl_sgemm16X16X16_mfma_write_vgpr_dot_read
@@ -1851,7 +1851,7 @@ body: |
1851
1851
...
1852
1852
# GCN-LABEL: name: smfmac16x16x32_mfma_write_vgpr_smfmac_read_idx
1853
1853
# GCN: V_SMFMAC
1854
- # GFX940 -NEXT: S_NOP 6
1854
+ # GFX942 -NEXT: S_NOP 6
1855
1855
# GFX950-NEXT: S_NOP 7
1856
1856
# GCN-NEXT: V_SMFMAC
1857
1857
name : smfmac16x16x32_mfma_write_vgpr_smfmac_read_idx
@@ -2078,7 +2078,7 @@ body: |
2078
2078
...
2079
2079
# GCN-LABEL: name: smfmac16x16_read_vgpr_srcc_valu_write
2080
2080
# GCN: V_SMFMAC
2081
- # GFX940 -NEXT: S_NOP 6
2081
+ # GFX942 -NEXT: S_NOP 6
2082
2082
# GFX950-NEXT: S_NOP 7
2083
2083
# GCN-NEXT: V_MOV_B32
2084
2084
name : smfmac16x16_read_vgpr_srcc_valu_write
@@ -2109,7 +2109,7 @@ body: |
2109
2109
# GCN-LABEL: name: smfmac32x32_read_vgpr_srcc_valu_write
2110
2110
# GCN: V_SMFMAC
2111
2111
# GCN-NEXT: S_NOP 7
2112
- # GFX940 -NEXT: S_NOP 2
2112
+ # GFX942 -NEXT: S_NOP 2
2113
2113
# GFX950-NEXT: S_NOP 3
2114
2114
# GCN-NEXT: V_MOV_B32
2115
2115
name : smfmac32x32_read_vgpr_srcc_valu_write
@@ -2216,7 +2216,7 @@ body: |
2216
2216
# 4 pass source
2217
2217
# GCN-LABEL: name: xdl_mfma_4pass_write_vgpr_xdl_mfma_read_overlap_srca
2218
2218
# GCN: V_MFMA
2219
- # GFX940 -NEXT: S_NOP 6
2219
+ # GFX942 -NEXT: S_NOP 6
2220
2220
# GFX950-NEXT: S_NOP 7
2221
2221
# GCN-NEXT: V_MFMA
2222
2222
name : xdl_mfma_4pass_write_vgpr_xdl_mfma_read_overlap_srca
@@ -2231,7 +2231,7 @@ body: |
2231
2231
# 4 pass source
2232
2232
# GCN-LABEL: name: xdl_mfma_4pass_write_vgpr_xdl_mfma_read_overlap_srcb
2233
2233
# GCN: V_MFMA
2234
- # GFX940 -NEXT: S_NOP 6
2234
+ # GFX942 -NEXT: S_NOP 6
2235
2235
# GFX950-NEXT: S_NOP 7
2236
2236
# GCN-NEXT: V_MFMA
2237
2237
name : xdl_mfma_4pass_write_vgpr_xdl_mfma_read_overlap_srcb
@@ -2306,7 +2306,7 @@ body: |
2306
2306
# 4 pass source
2307
2307
# GCN-LABEL: name: xdl_mfma_4pass_write_vgpr_sgemm_mfma_read_overlap_srca
2308
2308
# GCN: V_MFMA
2309
- # GFX940 -NEXT: S_NOP 6
2309
+ # GFX942 -NEXT: S_NOP 6
2310
2310
# GFX950-NEXT: S_NOP 7
2311
2311
# GCN-NEXT: V_MFMA
2312
2312
name : xdl_mfma_4pass_write_vgpr_sgemm_mfma_read_overlap_srca
@@ -2321,7 +2321,7 @@ body: |
2321
2321
# 4 pass source
2322
2322
# GCN-LABEL: name: xdl_mfma_4pass_write_vgpr_sgemm_mfma_read_overlap_srcb
2323
2323
# GCN: V_MFMA
2324
- # GFX940 -NEXT: S_NOP 6
2324
+ # GFX942 -NEXT: S_NOP 6
2325
2325
# GFX950-NEXT: S_NOP 7
2326
2326
# GCN-NEXT: V_MFMA
2327
2327
name : xdl_mfma_4pass_write_vgpr_sgemm_mfma_read_overlap_srcb
@@ -2353,7 +2353,7 @@ body: |
2353
2353
# GCN-LABEL: name: xdl_mfma_8pass_write_vgpr_nonxdl_sgemm_mfma_read_overlap_srca
2354
2354
# GCN: V_MFMA
2355
2355
# GCN-NEXT: S_NOP 7
2356
- # GFX940 -NEXT: S_NOP 2
2356
+ # GFX942 -NEXT: S_NOP 2
2357
2357
# GFX950-NEXT: S_NOP 3
2358
2358
# GCN-NEXT: V_MFMA
2359
2359
name : xdl_mfma_8pass_write_vgpr_nonxdl_sgemm_mfma_read_overlap_srca
@@ -2369,7 +2369,7 @@ body: |
2369
2369
# GCN-LABEL: name: xdl_mfma_8pass_write_vgpr_nonxdl_sgemm_mfma_read_overlap_srcb
2370
2370
# GCN: V_MFMA
2371
2371
# GCN-NEXT: S_NOP 7
2372
- # GFX940 -NEXT: S_NOP 2
2372
+ # GFX942 -NEXT: S_NOP 2
2373
2373
# GFX950-NEXT: S_NOP 3
2374
2374
# GCN-NEXT: V_MFMA
2375
2375
name : xdl_mfma_8pass_write_vgpr_nonxdl_sgemm_mfma_read_overlap_srcb
@@ -2404,7 +2404,7 @@ body: |
2404
2404
# GCN: V_MFMA
2405
2405
# GCN-NEXT: S_NOP 7
2406
2406
# GCN-NEXT: S_NOP 7
2407
- # GFX940 -NEXT: S_NOP 2
2407
+ # GFX942 -NEXT: S_NOP 2
2408
2408
# GFX950-NEXT: S_NOP 3
2409
2409
# GCN-NEXT: V_MFMA
2410
2410
name : xdl_16pass_write_vgpr_nonxdl_sgemm_mfma_read_overlap_srca
@@ -2421,7 +2421,7 @@ body: |
2421
2421
# GCN: V_MFMA
2422
2422
# GCN-NEXT: S_NOP 7
2423
2423
# GCN-NEXT: S_NOP 7
2424
- # GFX940 -NEXT: S_NOP 2
2424
+ # GFX942 -NEXT: S_NOP 2
2425
2425
# GFX950-NEXT: S_NOP 3
2426
2426
# GCN-NEXT: V_MFMA
2427
2427
name : xdl_16pass_write_vgpr_nonxdl_sgemm_mfma_read_overlap_srcb
@@ -2492,7 +2492,7 @@ body: |
2492
2492
# GCN-LABEL: name: xdl_mfma_8pass_write_vgpr_xdl_mfma_read_overlap_srca
2493
2493
# GCN: V_MFMA
2494
2494
# GCN-NEXT: S_NOP 7
2495
- # GFX940 -NEXT: S_NOP 2
2495
+ # GFX942 -NEXT: S_NOP 2
2496
2496
# GFX950-NEXT: S_NOP 3
2497
2497
# GCN-NEXT: V_MFMA
2498
2498
name : xdl_mfma_8pass_write_vgpr_xdl_mfma_read_overlap_srca
@@ -2507,7 +2507,7 @@ body: |
2507
2507
# GCN-LABEL: name: xdl_mfma_8pass_write_vgpr_xdl_mfma_read_overlap_srcb
2508
2508
# GCN: V_MFMA
2509
2509
# GCN-NEXT: S_NOP 7
2510
- # GFX940 -NEXT: S_NOP 2
2510
+ # GFX942 -NEXT: S_NOP 2
2511
2511
# GFX950-NEXT: S_NOP 3
2512
2512
# GCN-NEXT: V_MFMA
2513
2513
name : xdl_mfma_8pass_write_vgpr_xdl_mfma_read_overlap_srcb
@@ -2540,7 +2540,7 @@ body: |
2540
2540
# GCN: V_MFMA
2541
2541
# GCN-NEXT: S_NOP 7
2542
2542
# GCN-NEXT: S_NOP 7
2543
- # GFX940 -NEXT: S_NOP 2
2543
+ # GFX942 -NEXT: S_NOP 2
2544
2544
# GFX950-NEXT: S_NOP 3
2545
2545
# GCN-NEXT: V_MFMA
2546
2546
name : xdl_16pass_write_vgpr_xdl_mfma_read_overlap_srca
@@ -2558,7 +2558,7 @@ body: |
2558
2558
# GCN: V_MFMA
2559
2559
# GCN-NEXT: S_NOP 7
2560
2560
# GCN-NEXT: S_NOP 7
2561
- # GFX940 -NEXT: S_NOP 2
2561
+ # GFX942 -NEXT: S_NOP 2
2562
2562
# GFX950-NEXT: S_NOP 3
2563
2563
# GCN-NEXT: V_MFMA
2564
2564
name : xdl_16pass_write_vgpr_xdl_mfma_read_overlap_srcb
0 commit comments