kernel                                                 instr  narrow   loads     ds     mac   vgpr  th_max     smem
kernel_mul_mm_bf16_f32                                  1157       0       4     38     263    183     512        0
kernel_mul_mm_f16_f16                                    945       0       3     70     279    208     640        0
kernel_mul_mm_f16_f32                                    956       0       4     70     279    226     640        0
kernel_mul_mm_f16_f32_w                                 1538       0       6     92     552    182     512        0
kernel_mul_mm_f32_f32                                    975       0       6     70     279    208     640        0
kernel_mul_mm_id_bf16_f32                               1198       0       6     51     203    248     640        0
kernel_mul_mm_id_f16_f32                                 990       0       6     84     277    182     640        0
kernel_mul_mm_id_f16_f32_w                              1561       0       9    127     558    183     512        0
kernel_mul_mm_id_f32_f32                                1017       0       8     84     277    182     640        0
kernel_mul_mm_id_iq1_m_f32                              2810      12      16     84     303    234     640        0
kernel_mul_mm_id_iq1_m_f32_w                            3352      12      19    127     571    184     512        0
kernel_mul_mm_id_iq1_s_f32                              2770       9      13     84     299    202     640        0
kernel_mul_mm_id_iq2_s_f32                              2829       6      10     84     292    185     640        0
kernel_mul_mm_id_iq2_s_f32_w                            3398       6      13    127     569    185     512        0
kernel_mul_mm_id_iq2_xs_f32                             2023       5       9     84     287    251     640        0
kernel_mul_mm_id_iq2_xs_f32_w                           2553       5      12    127     556    251     512        0
kernel_mul_mm_id_iq2_xxs_f32                            1618       7      11     84     284    251     640        0
kernel_mul_mm_id_iq2_xxs_f32_w                          2178       7      14    127     560    251     512        0
kernel_mul_mm_id_iq3_s_f32                              1616       8      12     84     408    185     640        0
kernel_mul_mm_id_iq3_s_f32_w                            2193       8      15    127     682    185     512        0
kernel_mul_mm_id_iq3_xxs_f32                            1419      10      14     84     381    251     640        0
kernel_mul_mm_id_iq3_xxs_f32_w                          1976      10      17    127     656    251     512        0
kernel_mul_mm_id_iq4_nl_f32                             1097       1      21     84     294    186     640        0
kernel_mul_mm_id_iq4_xs_f32                             1130       1      24     84     294    188     640        0
kernel_mul_mm_id_map0_ne20_1                             177       1       1      2       4    182    1024        0
kernel_mul_mm_id_map0_ne20_16                            298      16      16     32       6    192    1024        0
kernel_mul_mm_id_map0_ne20_2                             188       2       2      4       4    182    1024        0
kernel_mul_mm_id_map0_ne20_4                             203       4       4      8       4    182    1024        0
kernel_mul_mm_id_mxfp4_f32                              1119      17      37     84     297    187     640        0
kernel_mul_mm_id_mxfp4_f32_w                            1674      17      40    127     572    202     512        0
kernel_mul_mm_id_q1_0_f32                               1102       2       6     84     280    247     640        0
kernel_mul_mm_id_q2_0_f32                               1090       4       8     84     296    241     640        0
kernel_mul_mm_id_q2_K_f32                               1124       4       8     84     301    202     640        0
kernel_mul_mm_id_q2_K_f32_w                             1654       4      11    127     569    182     512        0
kernel_mul_mm_id_q3_K_f32                               1220       4       8     84     300    202     640        0
kernel_mul_mm_id_q3_K_f32_w                             1744       4      11    127     567    184     512        0
kernel_mul_mm_id_q4_0_f32                               1066       9      13     84     296    229     576        0
kernel_mul_mm_id_q4_0_f32_w                             1614       9      16    127     567    182     512        0
kernel_mul_mm_id_q4_1_f32                               1065      10      14     84     296    229     576        0
kernel_mul_mm_id_q4_1_f32_w                             1613      10      17    127     567    182     512        0
kernel_mul_mm_id_q4_K_f32                               1118       0       6     84     297    255     640        0
kernel_mul_mm_id_q4_K_f32_w                             1648       0       9    127     568    182     512        0
kernel_mul_mm_id_q5_0_f32                               1169       4       8     84     298    202     576        0
kernel_mul_mm_id_q5_0_f32_w                             1661       4      11    127     567    184     320        0
kernel_mul_mm_id_q5_1_f32                               1162       4       9     84     298    202     576        0
kernel_mul_mm_id_q5_1_f32_w                             1654       4      12    127     567    184     320        0
kernel_mul_mm_id_q5_K_f32                               1205       7      13     84     318    202     640        0
kernel_mul_mm_id_q5_K_f32_w                             1728       7      16    127     585    184     512        0
kernel_mul_mm_id_q6_K_f32                               1180       1       5     84     299    240     640        0
kernel_mul_mm_id_q6_K_f32_w                             1710       1       8    127     567    182     512        0
kernel_mul_mm_id_q8_0_f32                               1050      16      20     84     296    222     640        0
kernel_mul_mm_id_q8_0_f32_w                             1575      16      23    127     567    182     512        0
kernel_mul_mm_iq1_m_f32                                 2794      12      14     70     301    184     576        0
kernel_mul_mm_iq1_m_f32_w                               3364      12      16     92     572    185     512        0
kernel_mul_mm_iq1_s_f32                                 2760       9      11     70     300    185     576        0
kernel_mul_mm_iq2_s_f32                                 2794       6       8     70     296    208     576        0
kernel_mul_mm_iq2_s_f32_w                               3385       6      10     92     569    185     512        0
kernel_mul_mm_iq2_xs_f32                                2021       5       8     70     285    251     576        0
kernel_mul_mm_iq2_xs_f32_w                              2582       5       9     92     555    251     512        0
kernel_mul_mm_iq2_xxs_f32                               1590       7       9     70     283    251     576        0
kernel_mul_mm_iq2_xxs_f32_w                             2220       7      11     92     556    251     512        0
kernel_mul_mm_iq3_s_f32                                 1585       8      10     70     413    208     576        0
kernel_mul_mm_iq3_s_f32_w                               2166       8      12     92     685    185     512        0
kernel_mul_mm_iq3_xxs_f32                               1389      10      12     70     380    251     576        0
kernel_mul_mm_iq3_xxs_f32_w                             2020      10      14     92     654    251     512        0
kernel_mul_mm_iq4_nl_f32                                1054       1      19     70     298    218     576        0
kernel_mul_mm_iq4_xs_f32                                1084       1      22     70     297    233     576        0
kernel_mul_mm_mxfp4_f32                                 1085      17      35     71     295    186     576        0
kernel_mul_mm_mxfp4_f32_w                               1710      17      37     92     573    185     512        0
kernel_mul_mm_q1_0_f32                                  1088       2       4     72     281    184     576        0
kernel_mul_mm_q2_0_f32                                  1076       4       6     70     296    192     576        0
kernel_mul_mm_q2_K_f32                                  1106       4       6     70     298    183     576        0
kernel_mul_mm_q2_K_f32_w                                1679       4       8     92     569    182     512        0
kernel_mul_mm_q3_K_f32                                  1198       4       6     70     295    184     576        0
kernel_mul_mm_q3_K_f32_w                                1771       4       8     92     567    182     512        0
kernel_mul_mm_q4_0_f32                                  1066       9      11     70     296    192     640        0
kernel_mul_mm_q4_0_f32_w                                1641       9      13     92     567    182     512        0
kernel_mul_mm_q4_1_f32                                  1062      10      12     70     296    192     640        0
kernel_mul_mm_q4_1_f32_w                                1641      10      14     92     567    182     512        0
kernel_mul_mm_q4_K_f32                                  1115       0       6     70     297    192     576        0
kernel_mul_mm_q4_K_f32_w                                1681       0       6     92     568    182     512        0
kernel_mul_mm_q5_0_f32                                  1149       4       6     70     297    184     512        0
kernel_mul_mm_q5_0_f32_w                                1731       4       8     92     567    184     448        0
kernel_mul_mm_q5_1_f32                                  1146       4       7     70     295    184     512        0
kernel_mul_mm_q5_1_f32_w                                1723       4       9     92     567    184     448        0
kernel_mul_mm_q5_K_f32                                  1186       7      11     70     317    184     576        0
kernel_mul_mm_q5_K_f32_w                                1762       7      13     92     585    184     512        0
kernel_mul_mm_q6_K_f32                                  1168       1       3     70     295    183     576        0
kernel_mul_mm_q6_K_f32_w                                1740       1       5     92     567    182     512        0
kernel_mul_mm_q8_0_f32                                  1047      16      18     70     296    183     576        0
kernel_mul_mm_q8_0_f32_w                                1617      16      20     92     567    182     512        0
kernel_mul_mv_bf16_f32                                   467       0       3      5      24    183    1024        0
kernel_mul_mv_bf16_f32_4                                 467       0       5      5      29    182    1024        0
kernel_mul_mv_bf16_f32_short                             163       0       0      0       3    182    1024        0
kernel_mul_mv_ext_bf16_f32_r1_2                          437       0       9      0      31    182    1024        0
kernel_mul_mv_ext_bf16_f32_r1_3                          501       0      13      0      43    182     768        0
kernel_mul_mv_ext_bf16_f32_r1_4                          570       0      17      0      55    182     576        0
kernel_mul_mv_ext_bf16_f32_r1_5                          646       0      21      0      67    182     576        0
kernel_mul_mv_ext_f16_f32_r1_2                           454       0       9      0      31    185    1024        0
kernel_mul_mv_ext_f16_f32_r1_3                           511       0      13      0      48    185     768        0
kernel_mul_mv_ext_f16_f32_r1_4                           587       0      17      0      66    185     576        0
kernel_mul_mv_ext_f16_f32_r1_5                           681       0      21      0      73    185    1024        0
kernel_mul_mv_ext_f32_f32_r1_2                           451       0      12      0      30    185    1024        0
kernel_mul_mv_ext_f32_f32_r1_3                           519       0      16      0      44    185     768        0
kernel_mul_mv_ext_f32_f32_r1_4                           584       0      20      0      62    185     576        0
kernel_mul_mv_ext_f32_f32_r1_5                           683       0      24      0      68    185    1024        0
kernel_mul_mv_ext_iq4_nl_f32_r1_2                        571       0      24      0      33    185     768        0
kernel_mul_mv_ext_iq4_nl_f32_r1_3                        622       0      28      0      46    185     640        0
kernel_mul_mv_ext_iq4_nl_f32_r1_4                        719       0      32      0      68    185     576        0
kernel_mul_mv_ext_iq4_nl_f32_r1_5                        793       0      36      0      71    185     640        0
kernel_mul_mv_ext_mxfp4_f32_r1_2                         573      20      44      0      34    206     768        0
kernel_mul_mv_ext_mxfp4_f32_r1_3                         629      20      48      0      47    206     640        0
kernel_mul_mv_ext_mxfp4_f32_r1_4                         706      20      52      0      57    206     576        0
kernel_mul_mv_ext_mxfp4_f32_r1_5                         797      20      56      0      69    254     640        0
kernel_mul_mv_ext_q1_0_f32_r1_2                          607       4      12      0      31    184     768        0
kernel_mul_mv_ext_q1_0_f32_r1_3                          668       4      16      0      43    184     640        0
kernel_mul_mv_ext_q1_0_f32_r1_4                          750       4      20      0      55    184     576        0
kernel_mul_mv_ext_q1_0_f32_r1_5                          818       4      24      0      67    184     576        0
kernel_mul_mv_ext_q2_0_f32_r1_2                          495       4      12      0      31    182    1024        0
kernel_mul_mv_ext_q2_0_f32_r1_3                          600       4      16      0      43    182     640        0
kernel_mul_mv_ext_q2_0_f32_r1_4                          676       4      20      0      55    182     576        0
kernel_mul_mv_ext_q2_0_f32_r1_5                          753       4      24      0      67    182     576        0
kernel_mul_mv_ext_q2_K_hdr_f32_r1_2_nr0_2                702       8      16      0      90    206     640        0
kernel_mul_mv_ext_q2_K_hdr_f32_r1_3_nr0_2                828       8      20      0     114    206     576        0
kernel_mul_mv_ext_q2_K_hdr_f32_r1_4_nr0_2                973       8      24      0     138    230     640        0
kernel_mul_mv_ext_q2_K_hdr_f32_r1_5_nr0_2               1069       8      28      0     163    242     576        0
kernel_mul_mv_ext_q3_K_hdr_f32_r1_2_nr0_2                855       8      16      0      86    224     640        0
kernel_mul_mv_ext_q3_K_hdr_f32_r1_3_nr0_2                974      10      22      0     112    244     640        0
kernel_mul_mv_ext_q3_K_hdr_f32_r1_4_nr0_2               1127      10      26      0     138    206     640        0
kernel_mul_mv_ext_q3_K_hdr_f32_r1_5_nr0_2               1223      10      30      0     163    206     576        0
kernel_mul_mv_ext_q4_0_f32_r1_2                          522       4      12      0      47    182     768        0
kernel_mul_mv_ext_q4_0_f32_r1_3                          572       4      16      0      59    182     640        0
kernel_mul_mv_ext_q4_0_f32_r1_4                          659       4      20      0      71    182     576        0
kernel_mul_mv_ext_q4_0_f32_r1_5                          740       4      24      0      83    182     768        0
kernel_mul_mv_ext_q4_1_f32_r1_2                          513       8      16      0      47    182     768        0
kernel_mul_mv_ext_q4_1_f32_r1_3                          572       8      20      0      59    182     640        0
kernel_mul_mv_ext_q4_1_f32_r1_4                          655       8      24      0      71    182     576        0
kernel_mul_mv_ext_q4_1_f32_r1_5                          737       8      28      0      83    182     640        0
kernel_mul_mv_ext_q4_K_raw_f32_r1_2_nr0_2                754       0      12      0      72    206     768        0
kernel_mul_mv_ext_q4_K_raw_f32_r1_3_nr0_2                896       0      16      0     106    220     576        0
kernel_mul_mv_ext_q4_K_raw_f32_r1_4_nr0_2               1032       0      20      0     137    209     512        0
kernel_mul_mv_ext_q4_K_raw_f32_r1_5_nr0_2               1161       0      24      0     170    224     576        0
kernel_mul_mv_ext_q5_0_f32_r1_2                          592      12      20      0      47    184     640        0
kernel_mul_mv_ext_q5_0_f32_r1_3                          646      12      24      0      59    184     576        0
kernel_mul_mv_ext_q5_0_f32_r1_4                          729      12      28      0      71    184     640        0
kernel_mul_mv_ext_q5_0_f32_r1_5                          812      12      32      0      83    184     640        0
kernel_mul_mv_ext_q5_1_f32_r1_2                          596       0      12      0      47    184     768        0
kernel_mul_mv_ext_q5_1_f32_r1_3                          635       0      16      0      59    184     576        0
kernel_mul_mv_ext_q5_1_f32_r1_4                          720       0      20      0      71    184     640        0
kernel_mul_mv_ext_q5_1_f32_r1_5                          802       0      24      0      83    184     640        0
kernel_mul_mv_ext_q5_K_hdr_f32_r1_2_nr0_2                878       0      14      0      88    242     576        0
kernel_mul_mv_ext_q5_K_hdr_f32_r1_3_nr0_2               1012       0      18      0     112    206     576        0
kernel_mul_mv_ext_q5_K_hdr_f32_r1_4_nr0_2               1148       0      22      0     138    244     576        0
kernel_mul_mv_ext_q5_K_hdr_f32_r1_5_nr0_2               1261       0      26      0     164    206     576        0
kernel_mul_mv_ext_q6_K_f32_r1_2_nr0_2                    818       2      10      0      78    206     640        0
kernel_mul_mv_ext_q6_K_f32_r1_3_nr0_2                    935       2      14      0     103    228     576        0
kernel_mul_mv_ext_q6_K_f32_r1_4_nr0_2                   1055       2      18      0     126    221     512        0
kernel_mul_mv_ext_q6_K_f32_r1_5_nr0_2                   1182       2      22      0     152    206     576        0
kernel_mul_mv_ext_q8_0_f32_r1_2                          479      16      24      0      31    182     768        0
kernel_mul_mv_ext_q8_0_f32_r1_3                          535      16      28      0      43    182     640        0
kernel_mul_mv_ext_q8_0_f32_r1_4                          615      16      32      0      55    182     576        0
kernel_mul_mv_ext_q8_0_f32_r1_5                          694      16      36      0      67    182     640        0
kernel_mul_mv_f16_f16_4                                  414       3       5      5      39    182    1024        0
kernel_mul_mv_f16_f16_short                              154       0       0      0       5    182    1024        0
kernel_mul_mv_f16_f32                                    407       0       3      5      21    208    1024        0
kernel_mul_mv_f16_f32_4                                  420       0       5      5      31    182    1024        0
kernel_mul_mv_f16_f32_short                              151       0       0      0       5    182    1024        0
kernel_mul_mv_f32_f32                                    406       0      21      5      21    208    1024        0
kernel_mul_mv_f32_f32_4                                  418       0      15      5      31    182    1024        0
kernel_mul_mv_f32_f32_short                              149       0       1      0       5    182    1024        0
kernel_mul_mv_id_bf16_f32                                538       0       3      5      24    202    1024        0
kernel_mul_mv_id_bf16_f32_4                              565       0       5      5      31    182     768        0
kernel_mul_mv_id_f16_f32                                 519       0       3      5      19    187    1024        0
kernel_mul_mv_id_f16_f32_4                               570       0       5      5      35    183     768        0
kernel_mul_mv_id_f32_f32                                 504       0       9      5      19    187    1024        0
kernel_mul_mv_id_f32_f32_4                               568       0      15      5      35    183     768        0
kernel_mul_mv_id_iq1_m_f32                              2936      40      58      0     146    237     768        0
kernel_mul_mv_id_iq1_s_f32                              2712      28      36      0     142    240     512        0
kernel_mul_mv_id_iq2_s_f32                              3460      32      41    129     132    187     768        0
kernel_mul_mv_id_iq2_xs_f32                             2066       7      16     74      17    251     576        0
kernel_mul_mv_id_iq2_xxs_f32                            1646      11      20     74      14    251     640        0
kernel_mul_mv_id_iq3_s_f32                              1708      30      39     65     201    199     512        0
kernel_mul_mv_id_iq3_xxs_f32                            1458      21      30     74     111    251     576        0
kernel_mul_mv_id_iq4_nl_f32                              737       8      13     33      32    187     768        0
kernel_mul_mv_id_iq4_xs_f32                              739       2      11     33      32    188     768        0
kernel_mul_mv_id_mxfp4_f32                               734       4       9     33      35    187     768        0
kernel_mul_mv_id_q1_0_f32                               1180       8      12      0      24    202     768        0
kernel_mul_mv_id_q2_0_f32                               1718       8      12      0      26    234     576        0
kernel_mul_mv_id_q2_K_f32                               1110      12      30      0     150    253     576        0
kernel_mul_mv_id_q3_K_f32                                792      10      22      0      44    202     576        0
kernel_mul_mv_id_q4_0_f32                                728       0       4      0      62    202     768        0
kernel_mul_mv_id_q4_1_f32                                730       8      12      0      62    202     768        0
kernel_mul_mv_id_q4_K_f32                                730       0      18      0      78    202     768        0
kernel_mul_mv_id_q5_0_f32                                955       8      12      0      63    202     576        0
kernel_mul_mv_id_q5_1_f32                                950       0       8      0      62    202     512        0
kernel_mul_mv_id_q5_K_f32                                728       4      25      0      42    222     640        0
kernel_mul_mv_id_q6_K_f32                                654       1      15      0      39    202    1024        0
kernel_mul_mv_id_q8_0_f32                                494       2       4      5      23    202    1024        0
kernel_mul_mv_iq1_m_f32                                 2859      40      58      0     144    184     768        0
kernel_mul_mv_iq1_s_f32                                 2636      28      36      0     144    185     512        0
kernel_mul_mv_iq2_s_f32                                 3375      32      41    129     137    208     768        0
kernel_mul_mv_iq2_xs_f32                                1988       7      16     74      16    251     576        0
kernel_mul_mv_iq2_xxs_f32                               1560      11      20     74      14    251     640        0
kernel_mul_mv_iq3_s_f32                                 1607      30      39     65     205    208     512        0
kernel_mul_mv_iq3_xxs_f32                               1357      21      30     74     110    251     576        0
kernel_mul_mv_iq4_nl_f32                                 642       8      13     33      35    208    1024        0
kernel_mul_mv_iq4_xs_f32                                 646       2      11     33      36    236     768        0
kernel_mul_mv_mxfp4_f32                                  644       4       9     33      33    187    1024        0
kernel_mul_mv_q1_0_f32                                  1121       8      12      0      20    184     768        0
kernel_mul_mv_q2_0_f32                                  1651       8      12      0      22    183     512        0
kernel_mul_mv_q2_K_f32                                  1035      12      30      0     148    183     576        0
kernel_mul_mv_q3_K_f32                                   716      10      22      0      43    184     576        0
kernel_mul_mv_q4_0_f32                                   648       0       4      0      66    183     768        0
kernel_mul_mv_q4_1_f32                                   659       8      12      0      66    183     640        0
kernel_mul_mv_q4_K_f32                                   656       0      19      0      79    183     768        0
kernel_mul_mv_q4_K_f32_r4                                995       0      30      0     149    206     576        0
kernel_mul_mv_q5_0_f32                                   871       8      12      0      67    184     576        0
kernel_mul_mv_q5_1_f32                                   875       8      16      0      63    184     640        0
kernel_mul_mv_q5_K_f32                                   648       4      25      0      43    184     640        0
kernel_mul_mv_q6_K_f32                                   586       1      15      0      37    183    1024        0
kernel_mul_mv_q8_0_f32                                   422       2       4      5      24    183    1024        0
