kernel                                                         instr  narrow   loads     ds     mac   vgpr  th_max     smem
kernel_flash_attn_ext_pf_amd_dk128_kf16_vf16                    4766      16      17    422    2018    223     256    43008
kernel_flash_attn_ext_pf_amd_dk128_kf16_vq4_0                   4834      17      18    422    2022    223     256    43008
kernel_flash_attn_ext_pf_amd_dk128_kf16_vq8_0                   4820      20      21    422    2018    223     256    43008
kernel_flash_attn_ext_pf_amd_dk128_kq4_0_vf16                   4831      17      18    422    2023    223     256    43008
kernel_flash_attn_ext_pf_amd_dk128_kq4_0_vq4_0                  4888      18      20    422    2028    223     256    43008
kernel_flash_attn_ext_pf_amd_dk128_kq4_0_vq8_0                  4870      21      23    422    2025    223     256    43008
kernel_flash_attn_ext_pf_amd_dk128_kq8_0_vf16                   4819      20      21    422    2019    223     256    43008
kernel_flash_attn_ext_pf_amd_dk128_kq8_0_vq4_0                  4871      21      24    422    2025    223     256    43008
kernel_flash_attn_ext_pf_amd_dk128_kq8_0_vq8_0                  4861      24      27    422    2019    223     256    43008
kernel_flash_attn_ext_pf_amd_dk256_kf16_vf16                    4005       8      13    462    1509    255     256    55808
kernel_flash_attn_ext_pf_amd_dk256_kf16_vq4_0                   4073       9      14    462    1513    255     256    55808
kernel_flash_attn_ext_pf_amd_dk256_kf16_vq8_0                   4061      12      17    462    1509    255     256    55808
kernel_flash_attn_ext_pf_amd_dk256_kq4_0_vf16                   4065       9      14    462    1514    255     256    55808
kernel_flash_attn_ext_pf_amd_dk256_kq4_0_vq4_0                  4114      10      16    462    1518    255     256    55808
kernel_flash_attn_ext_pf_amd_dk256_kq4_0_vq8_0                  4100      13      18    462    1514    255     256    55808
kernel_flash_attn_ext_pf_amd_dk256_kq8_0_vf16                   4053      12      17    462    1510    255     256    55808
kernel_flash_attn_ext_pf_amd_dk256_kq8_0_vq4_0                  4100      13      18    462    1514    255     256    55808
kernel_flash_attn_ext_pf_amd_dk256_kq8_0_vq8_0                  4087      16      21    462    1510    255     256    55808
kernel_flash_attn_ext_pf_amd_dk576_dv512_kf16_vf16              2513       2       4    308     586    183     256    56832
kernel_flash_attn_ext_pf_amd_dk64_kf16_vf16                     3243      16      17    246    1007    212     256    26112
kernel_flash_attn_ext_pf_amd_dk64_kf16_vq4_0                    3304      17      18    246    1012    244     256    26112
kernel_flash_attn_ext_pf_amd_dk64_kf16_vq8_0                    3291      20      21    246    1008    245     256    26112
kernel_flash_attn_ext_pf_amd_dk64_kq4_0_vf16                    3305      17      18    246    1011    244     256    26112
kernel_flash_attn_ext_pf_amd_dk64_kq4_0_vq4_0                   3374      18      19    246    1015    233     256    26112
kernel_flash_attn_ext_pf_amd_dk64_kq4_0_vq8_0                   3361      21      22    246    1011    230     256    26112
kernel_flash_attn_ext_pf_amd_dk64_kq8_0_vf16                    3292      20      21    246    1007    245     256    26112
kernel_flash_attn_ext_pf_amd_dk64_kq8_0_vq4_0                   3361      21      22    246    1011    230     256    26112
kernel_flash_attn_ext_pf_amd_dk64_kq8_0_vq8_0                   3348      24      25    246    1007    228     256    26112
kernel_flash_attn_ext_pf_amd_dk72_kf16_vf16                     4547      16     197    340    1197    255     256    28224
kernel_flash_attn_ext_tile_amd_dk128_kf16_vf16_w64               786       2       6      7       8    186    1024    16385
kernel_flash_attn_ext_tile_amd_dk128_kf16_vq4_0_w64              845       3       5      7      10    232    1024    16385
kernel_flash_attn_ext_tile_amd_dk128_kf16_vq8_0_w64              830       6       8      7       7    204    1024    16385
kernel_flash_attn_ext_tile_amd_dk128_kq4_0_vf16_w64              853       3       5      7       9    237    1024    16385
kernel_flash_attn_ext_tile_amd_dk128_kq4_0_vq4_0_w64             911       4       5      7      13    248    1024    16385
kernel_flash_attn_ext_tile_amd_dk128_kq4_0_vq8_0_w64             897       7       8      7      10    219    1024    16385
kernel_flash_attn_ext_tile_amd_dk128_kq8_0_vf16_w64              837       6       8      7       6    240    1024    16385
kernel_flash_attn_ext_tile_amd_dk128_kq8_0_vq4_0_w64             897       7       8      7      10    219    1024    16385
kernel_flash_attn_ext_tile_amd_dk128_kq8_0_vq8_0_w64             882      10      11      7       7    224    1024    16385
kernel_flash_attn_ext_tile_amd_dk128_kturbo3_vturbo3_w64         902       8      17      7       7    248    1024    16385
kernel_flash_attn_ext_tile_amd_dk128_kturbo4_vturbo4_w64         873       8      17      7       9    252    1024    16385
kernel_flash_attn_ext_tile_amd_dk256_kf16_vf16_w64               786       2       6      7       8    186    1024    16385
kernel_flash_attn_ext_tile_amd_dk256_kf16_vq4_0_w64              845       3       5      7      10    232    1024    16385
kernel_flash_attn_ext_tile_amd_dk256_kf16_vq8_0_w64              830       6       8      7       7    204    1024    16385
kernel_flash_attn_ext_tile_amd_dk256_kq4_0_vf16_w64              853       3       5      7       9    237    1024    16385
kernel_flash_attn_ext_tile_amd_dk256_kq4_0_vq4_0_w64             911       4       5      7      13    248    1024    16385
kernel_flash_attn_ext_tile_amd_dk256_kq4_0_vq8_0_w64             897       7       8      7      10    219    1024    16385
kernel_flash_attn_ext_tile_amd_dk256_kq8_0_vf16_w64              837       6       8      7       6    240    1024    16385
kernel_flash_attn_ext_tile_amd_dk256_kq8_0_vq4_0_w64             897       7       8      7      10    219    1024    16385
kernel_flash_attn_ext_tile_amd_dk256_kq8_0_vq8_0_w64             882      10      11      7       7    224    1024    16385
kernel_flash_attn_ext_tile_amd_dk384_kturbo3_vturbo3_w64        1012       6      18      9       8    248    1024    15361
kernel_flash_attn_ext_tile_amd_dk384_kturbo4_vturbo4_w64         987       6      17      9       9    252    1024    15361
kernel_flash_attn_ext_tile_amd_dk512_kf16_vf16_w64               864       2       6      9      12    248    1024    16385
kernel_flash_attn_ext_tile_amd_dk512_kf16_vq4_0_w64              930       3       6      9      11    216    1024    16385
kernel_flash_attn_ext_tile_amd_dk512_kf16_vq8_0_w64              917       6       9      9       7    201    1024    16385
kernel_flash_attn_ext_tile_amd_dk512_kq4_0_vf16_w64              939       3       6      9      10    248    1024    16385
kernel_flash_attn_ext_tile_amd_dk512_kq4_0_vq4_0_w64             996       4       6      9      15    254    1024    16385
kernel_flash_attn_ext_tile_amd_dk512_kq4_0_vq8_0_w64             982       7       9      9      12    255    1024    16385
kernel_flash_attn_ext_tile_amd_dk512_kq8_0_vf16_w64              925       6       9      9       6    201    1024    16385
kernel_flash_attn_ext_tile_amd_dk512_kq8_0_vq4_0_w64             982       7       9      9      12    255    1024    16385
kernel_flash_attn_ext_tile_amd_dk512_kq8_0_vq8_0_w64             970      10      12      9       7    241    1024    16385
kernel_flash_attn_ext_tile_amd_dk576_dv512_kf16_vf16_w64         931       2       7     10       8    229    1024    15233
kernel_flash_attn_ext_tile_amd_dk576_dv512_kq8_0_vq8_0_w64      1011      10      13     10      10    187    1024    15233
kernel_flash_attn_ext_tile_amd_dk640_kturbo3_vturbo3_w64        1088       6      18     11      10    248    1024    15361
kernel_flash_attn_ext_tile_amd_dk640_kturbo4_vturbo4_w64        1061       6      17     11      11    252    1024    15361
kernel_flash_attn_ext_tile_amd_dk64_kf16_vf16_w64                797       2       5      7       7    240    1024    16385
kernel_flash_attn_ext_tile_amd_dk64_kf16_vq4_0_w64               845       3       5      7      10    208    1024    16385
kernel_flash_attn_ext_tile_amd_dk64_kf16_vq8_0_w64               826       6       8      7       6    187    1024    16385
kernel_flash_attn_ext_tile_amd_dk64_kq4_0_vf16_w64               845       3       5      7      10    219    1024    16385
kernel_flash_attn_ext_tile_amd_dk64_kq4_0_vq4_0_w64              913       4       5      7      12    240    1024    16385
kernel_flash_attn_ext_tile_amd_dk64_kq4_0_vq8_0_w64              898       7       8      7       9    186    1024    16385
kernel_flash_attn_ext_tile_amd_dk64_kq8_0_vf16_w64               827       6       8      7       6    186    1024    16385
kernel_flash_attn_ext_tile_amd_dk64_kq8_0_vq4_0_w64              898       7       8      7       9    186    1024    16385
kernel_flash_attn_ext_tile_amd_dk64_kq8_0_vq8_0_w64              884      10      11      7       6    192    1024    16385
kernel_flash_attn_ext_tile_amd_dk72_kf16_vf16_w64                801       2       5      7       7    200    1024    16129
kernel_flash_attn_ext_vec_amd_dk128_kf16_vf16_w64                896       1       5      8      22    218     512    34304
kernel_flash_attn_ext_vec_amd_dk128_kf16_vq4_0_w64               977       5       9      8      35    229     768    34304
kernel_flash_attn_ext_vec_amd_dk128_kf16_vq8_0_w64               949      17      21      8      24    192     512    34304
kernel_flash_attn_ext_vec_amd_dk128_kq4_0_vf16_w64               977       5       9      8      35    187     512    34304
kernel_flash_attn_ext_vec_amd_dk128_kq4_0_vq4_0_w64             1069       9      13      8      45    226     512    34304
kernel_flash_attn_ext_vec_amd_dk128_kq4_0_vq8_0_w64             1044      21      25      8      33    201     512    34304
kernel_flash_attn_ext_vec_amd_dk128_kq8_0_vf16_w64               949      17      21      8      24    192     512    34304
kernel_flash_attn_ext_vec_amd_dk128_kq8_0_vq4_0_w64             1044      21      25      8      33    201     768    34304
kernel_flash_attn_ext_vec_amd_dk128_kq8_0_vq8_0_w64             1007      33      37      8      21    233     512    34304
kernel_flash_attn_ext_vec_amd_dk128_kturbo3_vturbo3_w64         1221      17      53      8      21    248     512    34304
kernel_flash_attn_ext_vec_amd_dk128_kturbo4_vturbo4_w64         1140      17      53      8       5    252     512    34304
kernel_flash_attn_ext_vec_amd_dk256_kf16_vf16_w64                901       1       5      8      22    223     512    33536
kernel_flash_attn_ext_vec_amd_dk256_kf16_vf16_w64_wg32           943       1       6      8      24    187     512    33536
kernel_flash_attn_ext_vec_amd_dk256_kf16_vq4_0_w64              1025       5      10      8      35    187     512    33536
kernel_flash_attn_ext_vec_amd_dk256_kf16_vq4_0_w64_wg32         1074       5       9      8      34    230     512    33536
kernel_flash_attn_ext_vec_amd_dk256_kf16_vq8_0_w64               975      17      21      8      24    187     512    33536
kernel_flash_attn_ext_vec_amd_dk256_kf16_vq8_0_w64_wg32         1034      17      21      8      22    189     512    33536
kernel_flash_attn_ext_vec_amd_dk256_kq4_0_vf16_w64              1025       5      10      8      35    187     512    33536
kernel_flash_attn_ext_vec_amd_dk256_kq4_0_vf16_w64_wg32         1074       5       9      8      34    230     512    33536
kernel_flash_attn_ext_vec_amd_dk256_kq4_0_vq4_0_w64             1125       9      13      8      45    210     512    33536
kernel_flash_attn_ext_vec_amd_dk256_kq4_0_vq4_0_w64_wg32        1167       9      13      8      46    206     512    33536
kernel_flash_attn_ext_vec_amd_dk256_kq4_0_vq8_0_w64             1124      21      25      8      33    187     512    33536
kernel_flash_attn_ext_vec_amd_dk256_kq4_0_vq8_0_w64_wg32        1166      21      25      8      34    206     512    33536
kernel_flash_attn_ext_vec_amd_dk256_kq8_0_vf16_w64               975      17      21      8      24    187     512    33536
kernel_flash_attn_ext_vec_amd_dk256_kq8_0_vf16_w64_wg32         1034      17      21      8      22    189     512    33536
kernel_flash_attn_ext_vec_amd_dk256_kq8_0_vq4_0_w64             1124      21      25      8      33    187     512    33536
kernel_flash_attn_ext_vec_amd_dk256_kq8_0_vq4_0_w64_wg32        1167      21      25      8      34    206     512    33536
kernel_flash_attn_ext_vec_amd_dk256_kq8_0_vq8_0_w64             1050      33      37      8      21    205     512    33536
kernel_flash_attn_ext_vec_amd_dk256_kq8_0_vq8_0_w64_wg32        1092      33      37      8      22    252     512    33536
kernel_flash_attn_ext_vec_amd_dk384_kturbo3_vturbo3_w64         1144      13      40      7      17    248     512    24960
kernel_flash_attn_ext_vec_amd_dk384_kturbo4_vturbo4_w64         1053      13      40      7       6    252     768    24960
kernel_flash_attn_ext_vec_amd_dk512_kf16_vf16_w64                924       1       5      8      22    227     512    33152
kernel_flash_attn_ext_vec_amd_dk512_kf16_vf16_w64_wg32           958       1       5      8      23    240     512    33152
kernel_flash_attn_ext_vec_amd_dk512_kf16_vq4_0_w64              1044       5       9      8      35    187     512    33152
kernel_flash_attn_ext_vec_amd_dk512_kf16_vq4_0_w64_wg32         1089       5       9      8      34    245     512    33152
kernel_flash_attn_ext_vec_amd_dk512_kf16_vq8_0_w64               991      17      21      8      25    187     512    33152
kernel_flash_attn_ext_vec_amd_dk512_kf16_vq8_0_w64_wg32         1041      17      21      8      22    196     512    33152
kernel_flash_attn_ext_vec_amd_dk512_kq4_0_vf16_w64              1036       5       9      8      35    187     512    33152
kernel_flash_attn_ext_vec_amd_dk512_kq4_0_vf16_w64_wg32         1089       5       9      8      34    245     512    33152
kernel_flash_attn_ext_vec_amd_dk512_kq4_0_vq4_0_w64             1140       9      13      8      45    187     512    33152
kernel_flash_attn_ext_vec_amd_dk512_kq4_0_vq4_0_w64_wg32        1183       9      13      8      46    250     512    33152
kernel_flash_attn_ext_vec_amd_dk512_kq4_0_vq8_0_w64             1139      21      25      8      33    187     512    33152
kernel_flash_attn_ext_vec_amd_dk512_kq4_0_vq8_0_w64_wg32        1182      21      25      8      34    250     512    33152
kernel_flash_attn_ext_vec_amd_dk512_kq8_0_vf16_w64               981      17      21      8      26    187     512    33152
kernel_flash_attn_ext_vec_amd_dk512_kq8_0_vf16_w64_wg32         1041      17      21      8      22    235     512    33152
kernel_flash_attn_ext_vec_amd_dk512_kq8_0_vq4_0_w64             1139      21      25      8      33    187     512    33152
kernel_flash_attn_ext_vec_amd_dk512_kq8_0_vq4_0_w64_wg32        1181      21      25      8      34    250     512    33152
kernel_flash_attn_ext_vec_amd_dk512_kq8_0_vq8_0_w64             1057      33      37      8      21    212     512    33152
kernel_flash_attn_ext_vec_amd_dk512_kq8_0_vq8_0_w64_wg32        1107      33      37      8      22    218     512    33152
kernel_flash_attn_ext_vec_amd_dk576_dv512_kf16_vf16_w64          843       1       4      6      13    240    1024    16576
kernel_flash_attn_ext_vec_amd_dk576_dv512_kf16_vf16_w64_wg32     885       1       4      6      14    203    1024    16576
kernel_flash_attn_ext_vec_amd_dk576_dv512_kq8_0_vq8_0_w64        935      21      24      6      13    232    1024    16576
kernel_flash_attn_ext_vec_amd_dk576_dv512_kq8_0_vq8_0_w64_wg     978      21      24      6      14    212    1024    16576
kernel_flash_attn_ext_vec_amd_dk640_kturbo3_vturbo3_w64         1366      21      66      9      24    248     512    41344
kernel_flash_attn_ext_vec_amd_dk640_kturbo4_vturbo4_w64         1220      21      66      9       5    252     512    41344
kernel_flash_attn_ext_vec_amd_dk64_kf16_vf16_w64                 863       1       3     10      13    200    1024    35840
kernel_flash_attn_ext_vec_amd_dk64_kf16_vq4_0_w64                939       3       5     10      19    236    1024    35840
kernel_flash_attn_ext_vec_amd_dk64_kf16_vq8_0_w64                911       9      11     10      14    213    1024    35840
kernel_flash_attn_ext_vec_amd_dk64_kq4_0_vf16_w64                937       3       5     10      20    236    1024    35840
kernel_flash_attn_ext_vec_amd_dk64_kq4_0_vq4_0_w64               961       5       7     10      28    192    1024    35840
kernel_flash_attn_ext_vec_amd_dk64_kq4_0_vq8_0_w64               955      11      13     10      23    187    1024    35840
kernel_flash_attn_ext_vec_amd_dk64_kq8_0_vf16_w64                911       9      11     10      14    213    1024    35840
kernel_flash_attn_ext_vec_amd_dk64_kq8_0_vq4_0_w64               956      11      13     10      23    187    1024    35840
kernel_flash_attn_ext_vec_amd_dk64_kq8_0_vq8_0_w64               927      17      19     10      16    187    1024    35840
kernel_flash_attn_ext_vec_amd_dk72_kf16_vf16_w64                 842       1       3      6      14    228    1024    19968
kernel_flash_attn_ext_vec_reduce_dv=256_nwg=32                 FALLA
kernel_flash_attn_ext_vec_reduce_dv=512_nwg=32                 FALLA
