kernel                                                 instr  narrow   loads     ds     mac   vgpr  th_max     smem
kernel_mul_mm_bf16_f32                                  1117       0       4     40       4    248     512        0
kernel_mul_mm_f16_f16                                    972       0       3     72     243    208     512        0
kernel_mul_mm_f16_f32                                    989       0       4     72     243    208     512        0
kernel_mul_mm_f32_f32                                   1012       0       6     72     242    212     512        0
kernel_mul_mm_id_bf16_f32                               1167       0       7     53       2    204     512        0
kernel_mul_mm_id_f16_f32                                1019       0       6     83     242    217     512        0
kernel_mul_mm_id_f32_f32                                1034       0       8     83     241    250     512        0
kernel_mul_mm_id_iq1_m_f32                              3276      12      16     83     244    234     512        0
kernel_mul_mm_id_iq1_s_f32                              3235       9      13     83     243    202     512        0
kernel_mul_mm_id_iq2_s_f32                              3268       6      10     83     241    202     512        0
kernel_mul_mm_id_iq2_xs_f32                             2271       5       9     83     244    241     512        0
kernel_mul_mm_id_iq2_xxs_f32                            1760       7      11     83     242    241     512        0
kernel_mul_mm_id_iq3_s_f32                              1774       8      12     83     241    202     512        0
kernel_mul_mm_id_iq3_xxs_f32                            1515      10      14     83     242    241     512        0
kernel_mul_mm_id_iq4_nl_f32                             1150       1      21     83     242    250     512        0
kernel_mul_mm_id_iq4_xs_f32                             1195       1      24     83     242    228     512        0
kernel_mul_mm_id_map0_ne20_1                             155       1       1      2       1    240    1024        0
kernel_mul_mm_id_map0_ne20_16                            277      16      16     32       3    183    1024        0
kernel_mul_mm_id_map0_ne20_2                             170       2       2      4       0    183    1024        0
kernel_mul_mm_id_map0_ne20_4                             184       4       4      8       0    183    1024        0
kernel_mul_mm_id_mxfp4_f32                              1186      17      37     83     241    224     512        0
kernel_mul_mm_id_q1_0_f32                               1119       2       6     83     241    236     512        0
kernel_mul_mm_id_q2_0_f32                               1128       4       8     83     241    229     512        0
kernel_mul_mm_id_q2_K_f32                               1158       4       8     83     257    202     512        0
kernel_mul_mm_id_q3_K_f32                               1240       4       8     83     257    202     512        0
kernel_mul_mm_id_q4_0_f32                               1110       9      13     83     256    202     512        0
kernel_mul_mm_id_q4_1_f32                               1110      10      14     83     256    202     512        0
kernel_mul_mm_id_q4_K_f32                               1144       0       6     83     257    202     512        0
kernel_mul_mm_id_q5_0_f32                               1188       4       8     83     256    202     512        0
kernel_mul_mm_id_q5_1_f32                               1182       4       9     83     256    202     512        0
kernel_mul_mm_id_q5_K_f32                               1244       7      13     83     257    202     512        0
kernel_mul_mm_id_q6_K_f32                               1205       1       5     83     253    202     512        0
kernel_mul_mm_id_q8_0_f32                               1089      16      20     83     241    249     512        0
kernel_mul_mm_iq1_m_f32                                 3237      12      14     72     246    184     512        0
kernel_mul_mm_iq1_s_f32                                 3202       9      11     72     247    192     512        0
kernel_mul_mm_iq2_s_f32                                 3245       6       8     72     242    208     512        0
kernel_mul_mm_iq2_xs_f32                                2238       5       7     72     245    241     512        0
kernel_mul_mm_iq2_xxs_f32                               1738       7       9     72     242    241     512        0
kernel_mul_mm_iq3_s_f32                                 1746       8      10     72     242    208     512        0
kernel_mul_mm_iq3_xxs_f32                               1481      10      12     72     242    241     512        0
kernel_mul_mm_iq4_nl_f32                                1126       1      19     72     243    228     512        0
kernel_mul_mm_iq4_xs_f32                                1171       1      22     72     242    228     512        0
kernel_mul_mm_mxfp4_f32                                 1157      17      36     72     242    248     512        0
kernel_mul_mm_q1_0_f32                                  1086       2       4     73     246    184     512        0
kernel_mul_mm_q2_0_f32                                  1090       4       6     72     242    183     512        0
kernel_mul_mm_q2_K_f32                                  1113       4       6     72     261    183     512        0
kernel_mul_mm_q3_K_f32                                  1207       4       6     72     258    184     512        0
kernel_mul_mm_q4_0_f32                                  1074       9      11     72     259    183     512        0
kernel_mul_mm_q4_1_f32                                  1065      10      12     72     259    183     512        0
kernel_mul_mm_q4_K_f32                                  1100       0       4     72     260    183     512        0
kernel_mul_mm_q5_0_f32                                  1168       4       6     72     258    184     512        0
kernel_mul_mm_q5_1_f32                                  1156       4       7     72     260    192     512        0
kernel_mul_mm_q5_K_f32                                  1195       7      11     72     258    184     512        0
kernel_mul_mm_q6_K_f32                                  1168       1       3     72     258    183     512        0
kernel_mul_mm_q8_0_f32                                  1062      16      18     72     244    183     512        0
kernel_mul_mv_bf16_f32                                   444       0       3      5       4    183    1024        0
kernel_mul_mv_bf16_f32_4                                 484       0       5      5       1    208     768        0
kernel_mul_mv_ext_bf16_f32_r1_2                          452       0       9      0       1    248    1024        0
kernel_mul_mv_ext_bf16_f32_r1_3                          519       0      13      0       1    207     768        0
kernel_mul_mv_ext_bf16_f32_r1_4                          594       0      17      0       1    182     512        0
kernel_mul_mv_ext_bf16_f32_r1_5                          698       0      21      0       1    182     768        0
kernel_mul_mv_ext_f16_f32_r1_2                           429       0       9      0       6    185    1024        0
kernel_mul_mv_ext_f16_f32_r1_3                           504       0      13      0       5    185     768        0
kernel_mul_mv_ext_f16_f32_r1_4                           579       0      17      0       3    185     512        0
kernel_mul_mv_ext_f16_f32_r1_5                           660       0      21      0      12    185     768        0
kernel_mul_mv_ext_f32_f32_r1_2                           426       0      12      0       6    185     768        0
kernel_mul_mv_ext_f32_f32_r1_3                           502       0      16      0       3    185     768        0
kernel_mul_mv_ext_f32_f32_r1_4                           573       0      20      0       2    185     512        0
kernel_mul_mv_ext_f32_f32_r1_5                           644       0      24      0       3    185     512        0
kernel_mul_mv_ext_iq4_nl_f32_r1_2                        570       0      26      0       6    228     768        0
kernel_mul_mv_ext_iq4_nl_f32_r1_3                        627       0      28      0       4    228     512        0
kernel_mul_mv_ext_iq4_nl_f32_r1_4                        707       0      32      0       6    228     512        0
kernel_mul_mv_ext_iq4_nl_f32_r1_5                        793       0      36      0       3    228     768        0
kernel_mul_mv_ext_mxfp4_f32_r1_2                         618      20      44      0       3    224     768        0
kernel_mul_mv_ext_mxfp4_f32_r1_3                         674      20      48      0       3    228     512        0
kernel_mul_mv_ext_mxfp4_f32_r1_4                         753      20      52      0       3    229     512        0
kernel_mul_mv_ext_mxfp4_f32_r1_5                         846      20      56      0       3    233     768        0
kernel_mul_mv_ext_q1_0_f32_r1_2                          629       4      12      0       3    184     768        0
kernel_mul_mv_ext_q1_0_f32_r1_3                          695       4      16      0       3    244     512        0
kernel_mul_mv_ext_q1_0_f32_r1_4                          775       4      20      0       3    249     512        0
kernel_mul_mv_ext_q1_0_f32_r1_5                          842       4      24      0       3    192     512        0
kernel_mul_mv_ext_q2_0_f32_r1_2                          516       4      12      0       1    208     768        0
kernel_mul_mv_ext_q2_0_f32_r1_3                          628       4      16      0       1    182     512        0
kernel_mul_mv_ext_q2_0_f32_r1_4                          706       4      20      0       1    208     512        0
kernel_mul_mv_ext_q2_0_f32_r1_5                          778       4      24      0       1    233     512        0
kernel_mul_mv_ext_q2_K_hdr_f32_r1_2_nr0_2                732       8      16      0      35    247     512        0
kernel_mul_mv_ext_q2_K_hdr_f32_r1_3_nr0_2                834       8      20      0      35    236     512        0
kernel_mul_mv_ext_q2_K_hdr_f32_r1_4_nr0_2                966       8      24      0      35    206     512        0
kernel_mul_mv_ext_q2_K_hdr_f32_r1_5_nr0_2               1102       8      28      0      35    206     512        0
kernel_mul_mv_ext_q3_K_hdr_f32_r1_2_nr0_2                867       8      16      0      35    206     768        0
kernel_mul_mv_ext_q3_K_hdr_f32_r1_3_nr0_2                983      10      22      0      35    206     512        0
kernel_mul_mv_ext_q3_K_hdr_f32_r1_4_nr0_2               1118      10      26      0      35    216     512        0
kernel_mul_mv_ext_q3_K_hdr_f32_r1_5_nr0_2               1238      10      30      0      35    206     512        0
kernel_mul_mv_ext_q4_0_f32_r1_2                          544       4      12      0      13    217     768        0
kernel_mul_mv_ext_q4_0_f32_r1_3                          598       4      16      0      13    182     512        0
kernel_mul_mv_ext_q4_0_f32_r1_4                          679       4      20      0      13    185     512        0
kernel_mul_mv_ext_q4_0_f32_r1_5                          768       4      24      0      19    244     768        0
kernel_mul_mv_ext_q4_1_f32_r1_2                          547       8      16      0      13    245     768        0
kernel_mul_mv_ext_q4_1_f32_r1_3                          603       8      20      0      13    182     512        0
kernel_mul_mv_ext_q4_1_f32_r1_4                          683       8      24      0      13    187     512        0
kernel_mul_mv_ext_q4_1_f32_r1_5                          771       8      28      0      23    243     768        0
kernel_mul_mv_ext_q4_K_raw_f32_r1_2_nr0_2                772       0      12      0       9    206     768        0
kernel_mul_mv_ext_q4_K_raw_f32_r1_3_nr0_2                896       0      16      0      12    206     512        0
kernel_mul_mv_ext_q4_K_raw_f32_r1_4_nr0_2               1042       0      20      0      17    228     512        0
kernel_mul_mv_ext_q4_K_raw_f32_r1_5_nr0_2               1193       0      24      0      21    206     512        0
kernel_mul_mv_ext_q5_0_f32_r1_2                          638      12      20      0      13    228     768        0
kernel_mul_mv_ext_q5_0_f32_r1_3                          680      12      24      0      13    240     512        0
kernel_mul_mv_ext_q5_0_f32_r1_4                          777      12      28      0      13    231     768        0
kernel_mul_mv_ext_q5_0_f32_r1_5                          845      12      32      0      13    252     512        0
kernel_mul_mv_ext_q5_1_f32_r1_2                          613       0      12      0      13    184     512        0
kernel_mul_mv_ext_q5_1_f32_r1_3                          668       0      16      0      13    195     512        0
kernel_mul_mv_ext_q5_1_f32_r1_4                          756       0      20      0      19    251     768        0
kernel_mul_mv_ext_q5_1_f32_r1_5                          836       0      24      0      13    184     512        0
kernel_mul_mv_ext_q5_K_hdr_f32_r1_2_nr0_2                889       0      14      0      35    206     512        0
kernel_mul_mv_ext_q5_K_hdr_f32_r1_3_nr0_2               1016       0      18      0      35    214     512        0
kernel_mul_mv_ext_q5_K_hdr_f32_r1_4_nr0_2               1147       0      22      0      35    206     512        0
kernel_mul_mv_ext_q5_K_hdr_f32_r1_5_nr0_2               1290       0      26      0      35    212     512        0
kernel_mul_mv_ext_q6_K_f32_r1_2_nr0_2                    815       2      10      0      27    206     512        0
kernel_mul_mv_ext_q6_K_f32_r1_3_nr0_2                    936       2      14      0      27    206     512        0
kernel_mul_mv_ext_q6_K_f32_r1_4_nr0_2                   1055       2      18      0      27    234     512        0
kernel_mul_mv_ext_q6_K_f32_r1_5_nr0_2                   1184       2      22      0      27    206     512        0
kernel_mul_mv_ext_q8_0_f32_r1_2                          502      16      24      0       1    182     768        0
kernel_mul_mv_ext_q8_0_f32_r1_3                          555      16      28      0       1    252     512        0
kernel_mul_mv_ext_q8_0_f32_r1_4                          640      16      32      0       1    233     512        0
kernel_mul_mv_ext_q8_0_f32_r1_5                          734      16      36      0       1    209     768        0
kernel_mul_mv_f16_f16_4                                  457       3       5      5       3    182    1024        0
kernel_mul_mv_f16_f32                                    434       0       3      5       3    234    1024        0
kernel_mul_mv_f16_f32_4                                  436       0       5      5       3    220     768        0
kernel_mul_mv_f32_f32                                    435       0      21      5       2    232    1024        0
kernel_mul_mv_f32_f32_4                                  437       0      15      5       1    220     768        0
kernel_mul_mv_id_bf16_f32_4                              610       0       5      5       3    182    1024        0
kernel_mul_mv_id_f16_f32_4                               560       0       5      5       3    188    1024        0
kernel_mul_mv_id_f32_f32_4                               559       0      15      5       3    183    1024        0
kernel_mul_mv_id_iq1_m_f32                              3384      40      58      0       2    240     768        0
kernel_mul_mv_id_iq1_s_f32                              3239      28      36      0       2    202     512        0
kernel_mul_mv_id_iq2_s_f32                              3279      16      25      9       1    209     512        0
kernel_mul_mv_id_iq2_xs_f32                             2323       7      16     46       4    241     768        0
kernel_mul_mv_id_iq2_xxs_f32                            1733      11      20     18       2    241     768        0
kernel_mul_mv_id_iq3_s_f32                              1909      30      39     65      17    232     512        0
kernel_mul_mv_id_iq3_xxs_f32                            1600      21      30     74       2    241     512        0
kernel_mul_mv_id_iq4_nl_f32                              777       8      13     33       4    228    1024        0
kernel_mul_mv_id_iq4_xs_f32                              799       2      11     33       4    228     768        0
kernel_mul_mv_id_mxfp4_f32                               791       4       9     33       3    224    1024        0
kernel_mul_mv_id_q1_0_f32                               1241       8      12      0       9    202     512        0
kernel_mul_mv_id_q2_0_f32                               1783       8      12      0       1    244    1024        0
kernel_mul_mv_id_q2_K_f32                               1167      12      30      0      17    202     768        0
kernel_mul_mv_id_q3_K_f32                                840      10      22      0      11    242     512        0
kernel_mul_mv_id_q4_0_f32                                765       0       4      0       5    233     768        0
kernel_mul_mv_id_q4_1_f32                                774       8      12      0       1    239     512        0
kernel_mul_mv_id_q4_K_f32                                759       0      18      0      13    247     512        0
kernel_mul_mv_id_q5_0_f32                                984       8      12      0       5    202     768        0
kernel_mul_mv_id_q5_1_f32                                993       0       8      0       1    203     768        0
kernel_mul_mv_id_q5_K_f32                                761       4      25      0       2    234     512        0
kernel_mul_mv_id_q6_K_f32                                619      14      22      0       1    238     768        0
kernel_mul_mv_id_q8_0_f32                                535       2       4      5       1    228    1024        0
kernel_mul_mv_iq1_m_f32                                 3257      40      59      0       7    184     768        0
kernel_mul_mv_iq1_s_f32                                 3111      28      38      0       6    231     512        0
kernel_mul_mv_iq2_s_f32                                 3152      16      26      9       2    251     512        0
kernel_mul_mv_iq2_xs_f32                                2187       7      16     46       4    241     768        0
kernel_mul_mv_iq2_xxs_f32                               1625      11      20     18       2    241     768        0
kernel_mul_mv_iq3_s_f32                                 1784      30      39     65      18    220     512        0
kernel_mul_mv_iq3_xxs_f32                               1474      21      30     74       2    241     512        0
kernel_mul_mv_iq4_nl_f32                                 667       8      13     33       4    228    1024        0
kernel_mul_mv_iq4_xs_f32                                 691       2      11     33       4    228     768        0
kernel_mul_mv_mxfp4_f32                                  676       4       9     33       4    224    1024        0
kernel_mul_mv_q1_0_f32                                  1130       8      12      0      11    192     512        0
kernel_mul_mv_q2_0_f32                                  1671       8      12      0       3    183    1024        0
kernel_mul_mv_q2_K_f32                                  1047      12      30      0      21    183     768        0
kernel_mul_mv_q3_K_f32                                   711      10      22      0      16    224     512        0
kernel_mul_mv_q4_0_f32                                   635       0       4      0       5    183     512        0
kernel_mul_mv_q4_1_f32                                   648       8      12      0       1    183     512        0
kernel_mul_mv_q4_K_f32_r4                               1008       0      30      0      27    238     768        0
kernel_mul_mv_q5_0_f32                                   852       8      12      0       7    184     768        0
kernel_mul_mv_q5_1_f32                                   858       8      16      0       1    184     768        0
kernel_mul_mv_q5_K_f32                                   634       4      25      0       7    184     512        0
kernel_mul_mv_q6_K_f32                                   498      14      22      0       4    183     768        0
kernel_mul_mv_q8_0_f32                                   415       2       4      5       1    183    1024        0
