kernel                                                 instr   loads     ds     mac   vgpr  th_max     smem
pf_amd_dk128_kf16_vf16                                  4655      17    422    2116    254     256    43008
pf_amd_dk128_kf16_vq4_0                                 4719      18    422    2120    254     256    43008
pf_amd_dk128_kf16_vq8_0                                 4707      21    422    2116    254     256    43008
pf_amd_dk128_kf16_vturbo2                               4703      23    422    2121    254     256    43008
pf_amd_dk128_kf16_vturbo3                               4723      24    422    2120    255     256    43008
pf_amd_dk128_kf16_vturbo4                               4720      24    422    2119    254     256    43008
pf_amd_dk128_kq4_0_vf16                                 4721      18    422    2120    254     256    43008
pf_amd_dk128_kq4_0_vq4_0                                4796      19    422    2126    254     256    43008
pf_amd_dk128_kq4_0_vq8_0                                4788      22    422    2124    254     256    43008
pf_amd_dk128_kq8_0_vf16                                 4706      21    422    2116    254     256    43008
pf_amd_dk128_kq8_0_vq4_0                                4787      23    422    2124    254     256    43008
pf_amd_dk128_kq8_0_vq8_0                                4772      25    422    2120    254     256    43008
pf_amd_dk128_kq8_0_vturbo2                              4769      27    422    2122    254     256    43008
pf_amd_dk128_kq8_0_vturbo3                              4787      28    422    2122    254     256    43008
pf_amd_dk128_kq8_0_vturbo4                              4789      28    422    2122    254     256    43008
pf_amd_dk128_kturbo2_vf16                               4701      23    422    2121    254     256    43008
pf_amd_dk128_kturbo2_vq8_0                              4772      27    422    2122    254     256    43008
pf_amd_dk128_kturbo2_vturbo2                            4724      29    422    2125    254     256    43008
pf_amd_dk128_kturbo2_vturbo3                            4777      30    422    2127    254     256    43008
pf_amd_dk128_kturbo2_vturbo4                            4774      30    422    2126    254     256    43008
pf_amd_dk128_kturbo3_vf16                               4725      24    422    2120    254     256    43008
pf_amd_dk128_kturbo3_vq8_0                              4793      28    422    2122    254     256    43008
pf_amd_dk128_kturbo3_vturbo2                            4780      30    422    2129    254     256    43008
pf_amd_dk128_kturbo3_vturbo3                            4767      31    422    2125    254     256    43008
pf_amd_dk128_kturbo3_vturbo4                            4796      31    422    2126    254     256    43008
pf_amd_dk128_kturbo4_vf16                               4725      24    422    2120    254     256    43008
pf_amd_dk128_kturbo4_vq8_0                              4795      28    422    2122    254     256    43008
pf_amd_dk128_kturbo4_vturbo2                            4780      30    422    2127    254     256    43008
pf_amd_dk128_kturbo4_vturbo3                            4800      31    422    2128    254     256    43008
pf_amd_dk128_kturbo4_vturbo4                            4751      31    422    2124    254     256    43008
pf_amd_dk256_kf16_vf16                                  3815      14    462    1600    255     256    55808
pf_amd_dk256_kf16_vq4_0                                 3878      15    462    1604    255     256    55808
pf_amd_dk256_kf16_vq8_0                                 3867      18    462    1601    255     256    55808
pf_amd_dk256_kf16_vturbo2                               3878      19    462    1601    255     256    55808
pf_amd_dk256_kf16_vturbo3                               3899      20    462    1600    255     256    55808
pf_amd_dk256_kf16_vturbo4                               3894      20    462    1600    255     256    55808
pf_amd_dk256_kq4_0_vf16                                 3880      15    462    1604    255     256    55808
pf_amd_dk256_kq4_0_vq4_0                                3962      16    462    1615    255     256    55808
pf_amd_dk256_kq4_0_vq8_0                                3952      19    462    1612    255     256    55808
pf_amd_dk256_kq8_0_vf16                                 3874      18    462    1601    255     256    55808
pf_amd_dk256_kq8_0_vq4_0                                3955      19    462    1610    255     256    55808
pf_amd_dk256_kq8_0_vq8_0                                3950      22    462    1606    255     256    55808
pf_amd_dk256_kq8_0_vturbo2                              3954      23    462    1606    255     256    55808
pf_amd_dk256_kq8_0_vturbo3                              3977      24    462    1607    255     256    55808
pf_amd_dk256_kq8_0_vturbo4                              3972      24    462    1606    255     256    55808
pf_amd_dk256_kturbo2_vf16                               3882      19    462    1601    255     256    55808
pf_amd_dk256_kturbo2_vq8_0                              3946      23    462    1606    255     256    55808
pf_amd_dk256_kturbo2_vturbo2                            3919      24    462    1601    255     256    55808
pf_amd_dk256_kturbo2_vturbo3                            3979      25    462    1607    255     256    55808
pf_amd_dk256_kturbo2_vturbo4                            3972      25    462    1606    255     256    55808
pf_amd_dk256_kturbo3_vf16                               3905      20    462    1600    255     256    55808
pf_amd_dk256_kturbo3_vq8_0                              3977      24    462    1607    255     256    55808
pf_amd_dk256_kturbo3_vturbo2                            3982      25    462    1607    255     256    55808
pf_amd_dk256_kturbo3_vturbo3                            3962      26    462    1601    255     256    55808
pf_amd_dk256_kturbo3_vturbo4                            3996      26    462    1607    255     256    55808
pf_amd_dk256_kturbo4_vf16                               3896      20    462    1600    255     256    55808
pf_amd_dk256_kturbo4_vq8_0                              3967      24    462    1608    255     256    55808
pf_amd_dk256_kturbo4_vturbo2                            3970      27    462    1606    255     256    55808
pf_amd_dk256_kturbo4_vturbo3                            3988      26    462    1607    255     256    55808
pf_amd_dk256_kturbo4_vturbo4                            3940      26    462    1601    255     256    55808
pf_amd_dk64_kf16_vf16                                   3127      17    246    1075    223     256    26112
pf_amd_dk64_kf16_vq4_0                                  3193      18    246    1078    232     256    26112
pf_amd_dk64_kf16_vq8_0                                  3184      21    247    1075    251     256    26112
pf_amd_dk64_kq4_0_vf16                                  3194      18    246    1079    232     256    26112
pf_amd_dk64_kq4_0_vq4_0                                 3257      19    246    1087    223     256    26112
pf_amd_dk64_kq4_0_vq8_0                                 3248      22    246    1083    224     256    26112
pf_amd_dk64_kq8_0_vf16                                  3182      21    247    1076    251     256    26112
pf_amd_dk64_kq8_0_vq4_0                                 3246      22    246    1083    224     256    26112
pf_amd_dk64_kq8_0_vq8_0                                 3234      25    246    1079    223     256    26112
pf_amd_dk72_kf16_vf16                                   4259     189    340    1273    255     256    28224
vec_amd_dk128_kf16_vf16                                  885       5     12      43    187     576    34304
vec_amd_dk128_kf16_vf16_w64                              838       5      8      38    212     576    34304
vec_amd_dk128_kf16_vq4_0                                1005       9     12      64    187     640    34304
vec_amd_dk128_kf16_vq4_0_w64                             960       9      8      59    187     640    34304
vec_amd_dk128_kf16_vq8_0                                 970      21     12      48    187     576    34304
vec_amd_dk128_kf16_vq8_0_w64                             922      21      8      44    187     576    34304
vec_amd_dk128_kf16_vturbo2                              1021      25     12      48    232     576    34304
vec_amd_dk128_kf16_vturbo2_w64                           977      25      8      45    232     576    34304
vec_amd_dk128_kf16_vturbo2_wg32                         1050      25     12      46    232     576    34304
vec_amd_dk128_kf16_vturbo3                              1066      29     12      52    248     576    34304
vec_amd_dk128_kf16_vturbo3_w64                          1036      29      8      45    248     512    34304
vec_amd_dk128_kf16_vturbo3_wg32                         1089      29     12      46    248     512    34304
vec_amd_dk128_kf16_vturbo4                              1052      29     12      49    252     576    34304
vec_amd_dk128_kf16_vturbo4_w64                          1013      29      8      46    252     576    34304
vec_amd_dk128_kf16_vturbo4_wg32                         1054      29     12      46    252     576    34304
vec_amd_dk128_kq4_0_vf16                                 890       5     10      41    187     768    25728
vec_amd_dk128_kq4_0_vf16_w64                             953       9      8      61    187     576    34304
vec_amd_dk128_kq4_0_vq4_0                                911       7     10      47    187     768    25728
vec_amd_dk128_kq4_0_vq4_0_w64                            983      13      8      73    187     640    34304
vec_amd_dk128_kq4_0_vq8_0                                918      13     10      39    187     768    25728
vec_amd_dk128_kq4_0_vq8_0_w64                            961      25      8      57    187     576    34304
vec_amd_dk128_kq8_0_vf16                                 951      21     12      48    187     576    34304
vec_amd_dk128_kq8_0_vf16_w64                             923      21      8      46    187     576    34304
vec_amd_dk128_kq8_0_vq4_0                               1013      25     12      61    187     640    34304
vec_amd_dk128_kq8_0_vq4_0_w64                            963      25      8      57    187     576    34304
vec_amd_dk128_kq8_0_vq8_0                                992      37     12      45    187     576    34304
vec_amd_dk128_kq8_0_vq8_0_w64                            944      37      8      41    187     576    34304
vec_amd_dk128_kq8_0_vturbo2                             1056      41     12      46    232     576    34304
vec_amd_dk128_kq8_0_vturbo2_w64                         1006      41      8      43    232     576    34304
vec_amd_dk128_kq8_0_vturbo2_wg32                        1080      41     12      43    232     576    34304
vec_amd_dk128_kq8_0_vturbo3                             1096      45     12      46    248     512    34304
vec_amd_dk128_kq8_0_vturbo3_w64                         1053      45      8      43    248     512    34304
vec_amd_dk128_kq8_0_vturbo3_wg32                        1144      45     12      43    248     512    34304
vec_amd_dk128_kq8_0_vturbo4                             1068      45     12      46    252     576    34304
vec_amd_dk128_kq8_0_vturbo4_w64                         1028      45      8      43    252     576    34304
vec_amd_dk128_kq8_0_vturbo4_wg32                        1119      45     12      43    252     576    34304
vec_amd_dk128_kturbo2_vf16                              1030      25     12      48    232     576    34304
vec_amd_dk128_kturbo2_vf16_w64                           983      25      8      44    232     576    34304
vec_amd_dk128_kturbo2_vf16_wg32                         1033      25     12      46    232     576    34304
vec_amd_dk128_kturbo2_vq8_0                             1053      41     12      46    232     576    34304
vec_amd_dk128_kturbo2_vq8_0_w64                         1015      41      8      42    232     576    34304
vec_amd_dk128_kturbo2_vq8_0_wg32                        1097      41     12      42    232     576    34304
vec_amd_dk128_kturbo2_vturbo2                           1108      45     12      42    240     576    34304
vec_amd_dk128_kturbo2_vturbo2_w64                       1063      45      8      38    232     576    34304
vec_amd_dk128_kturbo2_vturbo2_wg32                      1157      45     12      48    232     576    34304
vec_amd_dk128_kturbo2_vturbo3                           1164      49     12      46    248     576    34304
vec_amd_dk128_kturbo2_vturbo3_w64                       1123      49      8      42    248     512    34304
vec_amd_dk128_kturbo2_vturbo3_wg32                      1210      49     12      42    248     576    34304
vec_amd_dk128_kturbo2_vturbo4                           1135      49     12      46    252     576    34304
vec_amd_dk128_kturbo2_vturbo4_w64                       1097      49      8      42    252     576    34304
vec_amd_dk128_kturbo2_vturbo4_wg32                      1188      49     12      42    252     576    34304
vec_amd_dk128_kturbo3_vf16                              1067      29     12      52    248     576    34304
vec_amd_dk128_kturbo3_vf16_w64                          1039      29      8      44    248     512    34304
vec_amd_dk128_kturbo3_vf16_wg32                         1098      29     12      46    248     512    34304
vec_amd_dk128_kturbo3_vq8_0                             1102      45     12      46    248     512    34304
vec_amd_dk128_kturbo3_vq8_0_w64                         1063      45      8      42    248     512    34304
vec_amd_dk128_kturbo3_vq8_0_wg32                        1162      45     12      42    248     512    34304
vec_amd_dk128_kturbo3_vturbo2                           1141      49     12      46    248     576    34304
vec_amd_dk128_kturbo3_vturbo2_w64                       1126      49      8      42    248     576    34304
vec_amd_dk128_kturbo3_vturbo2_wg32                      1202      49     12      42    248     576    34304
vec_amd_dk128_kturbo3_vturbo3                           1203      53     12      46    248     576    34304
vec_amd_dk128_kturbo3_vturbo3_w64                       1155      53      8      40    248     512    34304
vec_amd_dk128_kturbo3_vturbo3_wg32                      1270      53     12      51    248     576    34304
vec_amd_dk128_kturbo3_vturbo4                           1185      53     12      46    252     512    34304
vec_amd_dk128_kturbo3_vturbo4_w64                       1150      53      8      42    252     512    34304
vec_amd_dk128_kturbo3_vturbo4_wg32                      1239      53     12      42    252     512    34304
vec_amd_dk128_kturbo4_vf16                              1052      29     12      49    252     640    34304
vec_amd_dk128_kturbo4_vf16_w64                          1019      29      8      46    252     576    34304
vec_amd_dk128_kturbo4_vf16_wg32                         1064      29     12      46    252     576    34304
vec_amd_dk128_kturbo4_vq8_0                             1086      45     12      46    252     576    34304
vec_amd_dk128_kturbo4_vq8_0_w64                         1039      45      8      42    252     576    34304
vec_amd_dk128_kturbo4_vq8_0_wg32                        1113      45     12      42    252     576    34304
vec_amd_dk128_kturbo4_vturbo2                           1136      49     12      46    252     576    34304
vec_amd_dk128_kturbo4_vturbo2_w64                       1090      49      8      42    252     576    34304
vec_amd_dk128_kturbo4_vturbo2_wg32                      1180      49     12      42    252     576    34304
vec_amd_dk128_kturbo4_vturbo3                           1198      53     12      46    252     512    34304
vec_amd_dk128_kturbo4_vturbo3_w64                       1161      53      8      42    252     512    34304
vec_amd_dk128_kturbo4_vturbo3_wg32                      1234      53     12      42    252     512    34304
vec_amd_dk128_kturbo4_vturbo4                           1136      53     12      42    252     576    34304
vec_amd_dk128_kturbo4_vturbo4_w64                       1090      53      8      38    252     576    34304
vec_amd_dk128_kturbo4_vturbo4_wg32                      1201      53     12      51    252     576    34304
vec_amd_dk256_kf16_vf16                                  847       5      8      38    244     576    33536
vec_amd_dk256_kf16_vf16_w64                              848       5      8      38    244     576    33536
vec_amd_dk256_kf16_vf16_w64_wg32                         912       5      8      41    187     576    33536
vec_amd_dk256_kf16_vf16_wg32                             912       5      8      41    187     576    33536
vec_amd_dk256_kf16_vq4_0                                 997       9      8      62    187     576    33536
vec_amd_dk256_kf16_vq4_0_w64                             998       9      8      62    187     576    33536
vec_amd_dk256_kf16_vq4_0_w64_wg32                       1000       9      8      56    187     640    33536
vec_amd_dk256_kf16_vq4_0_wg32                            999       9      8      56    187     640    33536
vec_amd_dk256_kf16_vq8_0                                 954      21      8      44    187     576    33536
vec_amd_dk256_kf16_vq8_0_w64                             955      21      8      44    187     576    33536
vec_amd_dk256_kf16_vq8_0_w64_wg32                        953      21      8      40    187     576    33536
vec_amd_dk256_kf16_vq8_0_wg32                            952      21      8      40    187     576    33536
vec_amd_dk256_kf16_vturbo2                               972      25      8      46    232     576    33536
vec_amd_dk256_kf16_vturbo2_w64                           973      25      8      46    232     576    33536
vec_amd_dk256_kf16_vturbo2_w64_wg32                      988      25      8      42    232     576    33536
vec_amd_dk256_kf16_vturbo2_wg32                          987      25      8      42    232     576    33536
vec_amd_dk256_kf16_vturbo3                              1032      29      8      49    248     576    33536
vec_amd_dk256_kf16_vturbo3_w64                          1033      29      8      49    248     576    33536
vec_amd_dk256_kf16_vturbo3_w64_wg32                     1034      29      8      42    248     576    33536
vec_amd_dk256_kf16_vturbo3_wg32                         1033      29      8      42    248     576    33536
vec_amd_dk256_kf16_vturbo4                              1015      29      8      44    252     576    33536
vec_amd_dk256_kf16_vturbo4_w64                          1016      29      8      44    252     576    33536
vec_amd_dk256_kf16_vturbo4_w64_wg32                     1008      29      8      42    252     576    33536
vec_amd_dk256_kf16_vturbo4_wg32                         1007      29      8      42    252     576    33536
vec_amd_dk256_kq4_0_vf16                                 998       9      8      62    187     576    33536
vec_amd_dk256_kq4_0_vf16_w64                             999       9      8      62    187     576    33536
vec_amd_dk256_kq4_0_vf16_w64_wg32                       1005       9      8      56    187     576    33536
vec_amd_dk256_kq4_0_vf16_wg32                           1004       9      8      56    187     576    33536
vec_amd_dk256_kq4_0_vq4_0                               1052      13      8      73    187     576    33536
vec_amd_dk256_kq4_0_vq4_0_w64                           1052      13      8      73    187     576    33536
vec_amd_dk256_kq4_0_vq4_0_w64_wg32                      1096      13      8      70    200     576    33536
vec_amd_dk256_kq4_0_vq4_0_wg32                          1095      13      8      70    200     576    33536
vec_amd_dk256_kq4_0_vq8_0                               1059      25      8      57    187     512    33536
vec_amd_dk256_kq4_0_vq8_0_w64                           1059      25      8      57    187     512    33536
vec_amd_dk256_kq4_0_vq8_0_w64_wg32                      1094      25      8      54    200     512    33536
vec_amd_dk256_kq4_0_vq8_0_wg32                          1093      25      8      54    200     512    33536
vec_amd_dk256_kq8_0_vf16                                 951      21      8      45    187     576    33536
vec_amd_dk256_kq8_0_vf16_w64                             952      21      8      45    187     576    33536
vec_amd_dk256_kq8_0_vf16_w64_wg32                        952      21      8      40    187     576    33536
vec_amd_dk256_kq8_0_vf16_wg32                            951      21      8      40    187     576    33536
vec_amd_dk256_kq8_0_vq4_0                               1053      25      8      57    187     576    33536
vec_amd_dk256_kq8_0_vq4_0_w64                           1053      25      8      57    187     576    33536
vec_amd_dk256_kq8_0_vq4_0_w64_wg32                      1089      25      8      54    200     576    33536
vec_amd_dk256_kq8_0_vq4_0_wg32                          1088      25      8      54    200     576    33536
vec_amd_dk256_kq8_0_vq8_0                                973      37      8      41    187     576    33536
vec_amd_dk256_kq8_0_vq8_0_w64                            973      37      8      41    187     576    33536
vec_amd_dk256_kq8_0_vq8_0_w64_wg32                      1019      37      8      38    200     576    33536
vec_amd_dk256_kq8_0_vq8_0_wg32                          1018      37      8      38    200     576    33536
vec_amd_dk256_kq8_0_vturbo2                             1031      41      8      42    232     576    33536
vec_amd_dk256_kq8_0_vturbo2_w64                         1031      41      8      43    232     576    33536
vec_amd_dk256_kq8_0_vturbo2_w64_wg32                    1070      41      8      39    232     576    33536
vec_amd_dk256_kq8_0_vturbo2_wg32                        1069      41      8      39    232     576    33536
vec_amd_dk256_kq8_0_vturbo3                             1068      45      8      42    248     512    33536
vec_amd_dk256_kq8_0_vturbo3_w64                         1068      45      8      43    248     512    33536
vec_amd_dk256_kq8_0_vturbo3_w64_wg32                    1107      45      8      39    248     512    33536
vec_amd_dk256_kq8_0_vturbo3_wg32                        1106      45      8      39    248     512    33536
vec_amd_dk256_kq8_0_vturbo4                             1036      45      8      42    252     576    33536
vec_amd_dk256_kq8_0_vturbo4_w64                         1036      45      8      43    252     576    33536
vec_amd_dk256_kq8_0_vturbo4_w64_wg32                    1090      45      8      40    252     576    33536
vec_amd_dk256_kq8_0_vturbo4_wg32                        1089      45      8      40    252     576    33536
vec_amd_dk256_kturbo2_vf16                               988      25      8      44    232     576    33536
vec_amd_dk256_kturbo2_vf16_w64                           989      25      8      44    232     576    33536
vec_amd_dk256_kturbo2_vf16_w64_wg32                      996      25      8      42    232     576    33536
vec_amd_dk256_kturbo2_vf16_wg32                          995      25      8      42    232     576    33536
vec_amd_dk256_kturbo2_vq8_0                             1031      41      8      42    232     576    33536
vec_amd_dk256_kturbo2_vq8_0_w64                         1032      41      8      42    232     576    33536
vec_amd_dk256_kturbo2_vq8_0_w64_wg32                    1079      41      8      38    232     576    33536
vec_amd_dk256_kturbo2_vq8_0_wg32                        1078      41      8      38    232     576    33536
vec_amd_dk256_kturbo2_vturbo2                           1055      45      8      39    232     576    33536
vec_amd_dk256_kturbo2_vturbo2_w64                       1056      45      8      39    232     576    33536
vec_amd_dk256_kturbo2_vturbo2_w64_wg32                  1119      45      8      41    232     576    33536
vec_amd_dk256_kturbo2_vturbo2_wg32                      1118      45      8      41    232     576    33536
vec_amd_dk256_kturbo2_vturbo3                           1103      49      8      42    248     576    33536
vec_amd_dk256_kturbo2_vturbo3_w64                       1104      49      8      42    248     576    33536
vec_amd_dk256_kturbo2_vturbo3_w64_wg32                  1168      49      8      38    248     576    33536
vec_amd_dk256_kturbo2_vturbo3_wg32                      1167      49      8      38    248     576    33536
vec_amd_dk256_kturbo2_vturbo4                           1081      49      8      42    252     576    33536
vec_amd_dk256_kturbo2_vturbo4_w64                       1082      49      8      42    252     576    33536
vec_amd_dk256_kturbo2_vturbo4_w64_wg32                  1133      49      8      38    252     576    33536
vec_amd_dk256_kturbo2_vturbo4_wg32                      1132      49      8      38    252     576    33536
vec_amd_dk256_kturbo3_vf16                              1037      29      8      44    248     512    33536
vec_amd_dk256_kturbo3_vf16_w64                          1038      29      8      44    248     512    33536
vec_amd_dk256_kturbo3_vf16_w64_wg32                     1035      29      8      42    248     576    33536
vec_amd_dk256_kturbo3_vf16_wg32                         1034      29      8      42    248     576    33536
vec_amd_dk256_kturbo3_vq8_0                             1077      45      8      42    248     512    33536
vec_amd_dk256_kturbo3_vq8_0_w64                         1078      45      8      42    248     512    33536
vec_amd_dk256_kturbo3_vq8_0_w64_wg32                    1125      45      8      38    248     512    33536
vec_amd_dk256_kturbo3_vq8_0_wg32                        1124      45      8      38    248     512    33536
vec_amd_dk256_kturbo3_vturbo2                           1096      49      8      42    248     576    33536
vec_amd_dk256_kturbo3_vturbo2_w64                       1097      49      8      42    248     576    33536
vec_amd_dk256_kturbo3_vturbo2_w64_wg32                  1161      49      8      38    248     576    33536
vec_amd_dk256_kturbo3_vturbo2_wg32                      1160      49      8      38    248     576    33536
vec_amd_dk256_kturbo3_vturbo3                           1146      53      8      40    248     576    33536
vec_amd_dk256_kturbo3_vturbo3_w64                       1147      53      8      40    248     576    33536
vec_amd_dk256_kturbo3_vturbo3_w64_wg32                  1198      53      8      42    248     576    33536
vec_amd_dk256_kturbo3_vturbo3_wg32                      1197      53      8      42    248     576    33536
vec_amd_dk256_kturbo3_vturbo4                           1130      53      8      42    252     512    33536
vec_amd_dk256_kturbo3_vturbo4_w64                       1131      53      8      42    252     512    33536
vec_amd_dk256_kturbo3_vturbo4_w64_wg32                  1176      53      8      38    252     576    33536
vec_amd_dk256_kturbo3_vturbo4_wg32                      1175      53      8      38    252     576    33536
vec_amd_dk256_kturbo4_vf16                              1010      29      8      46    252     576    33536
vec_amd_dk256_kturbo4_vf16_w64                          1011      29      8      46    252     576    33536
vec_amd_dk256_kturbo4_vf16_w64_wg32                     1018      29      8      42    252     576    33536
vec_amd_dk256_kturbo4_vf16_wg32                         1017      29      8      42    252     576    33536
vec_amd_dk256_kturbo4_vq8_0                             1047      45      8      42    252     576    33536
vec_amd_dk256_kturbo4_vq8_0_w64                         1048      45      8      42    252     576    33536
vec_amd_dk256_kturbo4_vq8_0_w64_wg32                    1086      45      8      38    252     576    33536
vec_amd_dk256_kturbo4_vq8_0_wg32                        1085      45      8      38    252     576    33536
vec_amd_dk256_kturbo4_vturbo2                           1082      49      8      42    252     576    33536
vec_amd_dk256_kturbo4_vturbo2_w64                       1083      49      8      42    252     576    33536
vec_amd_dk256_kturbo4_vturbo2_w64_wg32                  1130      49      8      38    252     576    33536
vec_amd_dk256_kturbo4_vturbo2_wg32                      1129      49      8      38    252     576    33536
vec_amd_dk256_kturbo4_vturbo3                           1122      53      8      42    252     512    33536
vec_amd_dk256_kturbo4_vturbo3_w64                       1123      53      8      42    252     512    33536
vec_amd_dk256_kturbo4_vturbo3_w64_wg32                  1170      53      8      38    252     512    33536
vec_amd_dk256_kturbo4_vturbo3_wg32                      1169      53      8      38    252     512    33536
vec_amd_dk256_kturbo4_vturbo4                           1089      53      8      38    252     576    33536
vec_amd_dk256_kturbo4_vturbo4_w64                       1090      53      8      38    252     576    33536
vec_amd_dk256_kturbo4_vturbo4_w64_wg32                  1137      53      8      43    252     576    33536
vec_amd_dk256_kturbo4_vturbo4_wg32                      1137      53      8      43    252     576    33536
vec_amd_dk384_kf16_vturbo2                               914      19      7      43    232     768    24960
vec_amd_dk384_kf16_vturbo2_w64                           950      19      7      37    232     640    24960
vec_amd_dk384_kf16_vturbo3                               967      22      7      38    248     640    24960
vec_amd_dk384_kf16_vturbo3_w64                           978      22      7      38    248     640    24960
vec_amd_dk384_kf16_vturbo4                               925      23      7      40    252     640    24960
vec_amd_dk384_kf16_vturbo4_w64                           946      22      7      37    252     640    24960
vec_amd_dk384_kq8_0_vturbo2                              949      31      7      35    232     640    24960
vec_amd_dk384_kq8_0_vturbo2_w64                          955      31      7      36    232     640    24960
vec_amd_dk384_kq8_0_vturbo3                              971      34      7      35    248     640    24960
vec_amd_dk384_kq8_0_vturbo3_w64                          988      34      7      36    248     640    24960
vec_amd_dk384_kq8_0_vturbo4                              942      34      7      35    252     640    24960
vec_amd_dk384_kq8_0_vturbo4_w64                          980      34      7      36    252     640    24960
vec_amd_dk384_kturbo2_vf16                               916      19      7      43    232     640    24960
vec_amd_dk384_kturbo2_vf16_w64                           943      19      7      37    232     640    24960
vec_amd_dk384_kturbo2_vq8_0                              948      31      7      35    232     640    24960
vec_amd_dk384_kturbo2_vq8_0_w64                          973      31      7      35    232     640    24960
vec_amd_dk384_kturbo2_vturbo2                            966      34      7      31    232     768    24960
vec_amd_dk384_kturbo2_vturbo2_w64                       1001      34      7      31    232     640    24960
vec_amd_dk384_kturbo2_vturbo3                           1023      37      7      35    248     640    24960
vec_amd_dk384_kturbo2_vturbo3_w64                       1031      37      7      35    248     640    24960
vec_amd_dk384_kturbo2_vturbo4                            977      37      7      35    252     640    24960
vec_amd_dk384_kturbo2_vturbo4_w64                       1017      37      7      35    252     640    24960
vec_amd_dk384_kturbo3_vf16                               967      22      7      38    248     640    24960
vec_amd_dk384_kturbo3_vf16_w64                           988      22      7      37    248     640    24960
vec_amd_dk384_kturbo3_vq8_0                              981      34      7      35    248     640    24960
vec_amd_dk384_kturbo3_vq8_0_w64                          998      34      7      35    248     640    24960
vec_amd_dk384_kturbo3_vturbo2                           1030      37      7      35    248     640    24960
vec_amd_dk384_kturbo3_vturbo2_w64                       1048      37      7      35    248     640    24960
vec_amd_dk384_kturbo3_vturbo3                           1049      40      7      31    248     640    24960
vec_amd_dk384_kturbo3_vturbo3_w64                       1062      40      7      31    248     640    24960
vec_amd_dk384_kturbo3_vturbo4                           1025      40      7      35    252     640    24960
vec_amd_dk384_kturbo3_vturbo4_w64                       1056      40      7      35    252     640    24960
vec_amd_dk384_kturbo4_vf16                               929      23      7      38    252     640    24960
vec_amd_dk384_kturbo4_vf16_w64                           956      22      7      37    252     640    24960
vec_amd_dk384_kturbo4_vq8_0                              949      34      7      35    252     640    24960
vec_amd_dk384_kturbo4_vq8_0_w64                          976      34      7      35    252     640    24960
vec_amd_dk384_kturbo4_vturbo2                            983      37      7      35    252     640    24960
vec_amd_dk384_kturbo4_vturbo2_w64                       1017      37      7      35    252     640    24960
vec_amd_dk384_kturbo4_vturbo3                           1034      40      7      35    252     640    24960
vec_amd_dk384_kturbo4_vturbo3_w64                       1049      40      7      35    252     576    24960
vec_amd_dk384_kturbo4_vturbo4                            974      40      7      31    252     768    24960
vec_amd_dk384_kturbo4_vturbo4_w64                        998      40      7      31    252     640    24960
vec_amd_dk512_kf16_vf16                                  843       5      8      39    230     576    33152
vec_amd_dk512_kf16_vf16_w64                              845       5      8      39    246     640    33152
vec_amd_dk512_kf16_vf16_w64_wg32                         914       5      8      43    187     640    33152
vec_amd_dk512_kf16_vf16_wg32                             895       5      8      41    187     576    33152
vec_amd_dk512_kf16_vq4_0                                 988       9      8      60    187     576    33152
vec_amd_dk512_kf16_vq4_0_w64                             999       9      8      61    187     576    33152
vec_amd_dk512_kf16_vq4_0_w64_wg32                       1009       9      8      56    187     640    33152
vec_amd_dk512_kf16_vq4_0_wg32                            997       9      8      56    187     576    33152
vec_amd_dk512_kf16_vq8_0                                 928      21      8      45    187     576    33152
vec_amd_dk512_kf16_vq8_0_w64                             951      21      8      43    187     576    33152
vec_amd_dk512_kf16_vq8_0_w64_wg32                        962      21      8      40    187     576    33152
vec_amd_dk512_kf16_vq8_0_wg32                            951      21      8      40    187     576    33152
vec_amd_dk512_kf16_vturbo2                               981      25      8      45    232     576    33152
vec_amd_dk512_kf16_vturbo2_w64                          1012      25      8      45    232     576    33152
vec_amd_dk512_kf16_vturbo2_w64_wg32                     1009      25      8      42    232     576    33152
vec_amd_dk512_kf16_vturbo2_wg32                          992      25      8      42    232     576    33152
vec_amd_dk512_kf16_vturbo3                              1038      29      8      47    248     576    33152
vec_amd_dk512_kf16_vturbo3_w64                          1052      29      8      47    248     576    33152
vec_amd_dk512_kf16_vturbo3_w64_wg32                     1063      29      8      42    248     576    33152
vec_amd_dk512_kf16_vturbo3_wg32                         1036      29      8      42    248     576    33152
vec_amd_dk512_kf16_vturbo4                               972      29      8      44    252     576    33152
vec_amd_dk512_kf16_vturbo4_w64                          1007      29      8      45    252     576    33152
vec_amd_dk512_kf16_vturbo4_w64_wg32                     1012      29      8      42    252     576    33152
vec_amd_dk512_kf16_vturbo4_wg32                          993      29      8      42    252     576    33152
vec_amd_dk512_kq4_0_vf16                                 996       9      8      63    187     576    33152
vec_amd_dk512_kq4_0_vf16_w64                            1009       9      8      60    187     576    33152
vec_amd_dk512_kq4_0_vf16_w64_wg32                       1014       9      8      56    187     576    33152
vec_amd_dk512_kq4_0_vf16_wg32                            995       9      8      56    187     576    33152
vec_amd_dk512_kq4_0_vq4_0                               1050      13      8      73    187     576    33152
vec_amd_dk512_kq4_0_vq4_0_w64                           1062      13      8      73    187     576    33152
vec_amd_dk512_kq4_0_vq4_0_w64_wg32                      1104      13      8      70    200     576    33152
vec_amd_dk512_kq4_0_vq4_0_wg32                          1084      13      8      70    200     576    33152
vec_amd_dk512_kq4_0_vq8_0                               1043      25      8      57    187     512    33152
vec_amd_dk512_kq4_0_vq8_0_w64                           1061      25      8      57    187     512    33152
vec_amd_dk512_kq4_0_vq8_0_w64_wg32                      1103      25      8      54    200     512    33152
vec_amd_dk512_kq4_0_vq8_0_wg32                          1094      25      8      54    200     512    33152
vec_amd_dk512_kq8_0_vf16                                 926      21      8      45    187     576    33152
vec_amd_dk512_kq8_0_vf16_w64                             950      21      8      43    187     576    33152
vec_amd_dk512_kq8_0_vf16_w64_wg32                        961      21      8      40    187     576    33152
vec_amd_dk512_kq8_0_vf16_wg32                            943      21      8      40    187     576    33152
vec_amd_dk512_kq8_0_vq4_0                               1045      25      8      57    187     576    33152
vec_amd_dk512_kq8_0_vq4_0_w64                           1055      25      8      57    187     576    33152
vec_amd_dk512_kq8_0_vq4_0_w64_wg32                      1097      25      8      54    200     576    33152
vec_amd_dk512_kq8_0_vq4_0_wg32                          1088      25      8      54    200     576    33152
vec_amd_dk512_kq8_0_vq8_0                                966      37      8      41    187     576    33152
vec_amd_dk512_kq8_0_vq8_0_w64                            983      37      8      41    187     576    33152
vec_amd_dk512_kq8_0_vq8_0_w64_wg32                      1036      37      8      38    200     576    33152
vec_amd_dk512_kq8_0_vq8_0_wg32                          1001      37      8      38    200     576    33152
vec_amd_dk512_kq8_0_vturbo2                             1021      41      8      42    232     576    33152
vec_amd_dk512_kq8_0_vturbo2_w64                         1037      41      8      43    232     576    33152
vec_amd_dk512_kq8_0_vturbo2_w64_wg32                    1090      41      8      39    232     576    33152
vec_amd_dk512_kq8_0_vturbo2_wg32                        1057      41      8      39    232     576    33152
vec_amd_dk512_kq8_0_vturbo3                             1065      45      8      42    248     512    33152
vec_amd_dk512_kq8_0_vturbo3_w64                         1075      45      8      43    248     512    33152
vec_amd_dk512_kq8_0_vturbo3_w64_wg32                    1136      45      8      39    248     576    33152
vec_amd_dk512_kq8_0_vturbo3_wg32                        1109      45      8      39    248     576    33152
vec_amd_dk512_kq8_0_vturbo4                             1031      45      8      42    252     576    33152
vec_amd_dk512_kq8_0_vturbo4_w64                         1045      45      8      43    252     576    33152
vec_amd_dk512_kq8_0_vturbo4_w64_wg32                    1084      45      8      39    252     576    33152
vec_amd_dk512_kq8_0_vturbo4_wg32                        1065      45      8      39    252     576    33152
vec_amd_dk512_kturbo2_vf16                               984      25      8      47    232     576    33152
vec_amd_dk512_kturbo2_vf16_w64                          1013      25      8      45    232     576    33152
vec_amd_dk512_kturbo2_vf16_w64_wg32                     1009      25      8      42    232     576    33152
vec_amd_dk512_kturbo2_vf16_wg32                          993      25      8      42    232     576    33152
vec_amd_dk512_kturbo2_vq8_0                             1028      41      8      42    232     576    33152
vec_amd_dk512_kturbo2_vq8_0_w64                         1046      41      8      42    232     576    33152
vec_amd_dk512_kturbo2_vq8_0_w64_wg32                    1100      41      8      38    232     576    33152
vec_amd_dk512_kturbo2_vq8_0_wg32                        1059      41      8      38    232     576    33152
vec_amd_dk512_kturbo2_vturbo2                           1075      45      8      38    232     576    33152
vec_amd_dk512_kturbo2_vturbo2_w64                       1082      45      8      38    232     576    33152
vec_amd_dk512_kturbo2_vturbo2_w64_wg32                  1152      45      8      42    232     576    33152
vec_amd_dk512_kturbo2_vturbo2_wg32                      1130      45      8      44    232     576    33152
vec_amd_dk512_kturbo2_vturbo3                           1116      49      8      42    248     576    33152
vec_amd_dk512_kturbo2_vturbo3_w64                       1142      49      8      42    248     576    33152
vec_amd_dk512_kturbo2_vturbo3_w64_wg32                  1189      49      8      38    248     576    33152
vec_amd_dk512_kturbo2_vturbo3_wg32                      1155      49      8      38    248     576    33152
vec_amd_dk512_kturbo2_vturbo4                           1065      49      8      42    252     576    33152
vec_amd_dk512_kturbo2_vturbo4_w64                       1090      49      8      42    252     576    33152
vec_amd_dk512_kturbo2_vturbo4_w64_wg32                  1137      49      8      38    252     576    33152
vec_amd_dk512_kturbo2_vturbo4_wg32                      1126      49      8      38    252     576    33152
vec_amd_dk512_kturbo3_vf16                              1035      29      8      45    248     576    33152
vec_amd_dk512_kturbo3_vf16_w64                          1049      29      8      43    248     576    33152
vec_amd_dk512_kturbo3_vf16_w64_wg32                     1064      29      8      42    248     576    33152
vec_amd_dk512_kturbo3_vf16_wg32                         1045      29      8      42    248     576    33152
vec_amd_dk512_kturbo3_vq8_0                             1072      45      8      42    248     576    33152
vec_amd_dk512_kturbo3_vq8_0_w64                         1093      45      8      42    248     576    33152
vec_amd_dk512_kturbo3_vq8_0_w64_wg32                    1147      45      8      38    248     576    33152
vec_amd_dk512_kturbo3_vq8_0_wg32                        1118      45      8      38    248     576    33152
vec_amd_dk512_kturbo3_vturbo2                           1125      49      8      42    248     576    33152
vec_amd_dk512_kturbo3_vturbo2_w64                       1143      49      8      42    248     576    33152
vec_amd_dk512_kturbo3_vturbo2_w64_wg32                  1189      49      8      38    248     576    33152
vec_amd_dk512_kturbo3_vturbo2_wg32                      1171      49      8      38    248     576    33152
vec_amd_dk512_kturbo3_vturbo3                           1154      53      8      40    248     576    33152
vec_amd_dk512_kturbo3_vturbo3_w64                       1174      53      8      40    248     576    33152
vec_amd_dk512_kturbo3_vturbo3_w64_wg32                  1242      53      8      45    248     576    33152
vec_amd_dk512_kturbo3_vturbo3_wg32                      1220      53      8      46    248     576    33152
vec_amd_dk512_kturbo3_vturbo4                           1120      53      8      42    252     576    33152
vec_amd_dk512_kturbo3_vturbo4_w64                       1152      53      8      42    252     576    33152
vec_amd_dk512_kturbo3_vturbo4_w64_wg32                  1191      53      8      38    252     576    33152
vec_amd_dk512_kturbo3_vturbo4_wg32                      1164      53      8      38    252     576    33152
vec_amd_dk512_kturbo4_vf16                               970      29      8      45    252     576    33152
vec_amd_dk512_kturbo4_vf16_w64                          1014      29      8      44    252     576    33152
vec_amd_dk512_kturbo4_vf16_w64_wg32                     1015      29      8      42    252     576    33152
vec_amd_dk512_kturbo4_vf16_wg32                          999      29      8      42    252     576    33152
vec_amd_dk512_kturbo4_vq8_0                             1028      45      8      42    252     576    33152
vec_amd_dk512_kturbo4_vq8_0_w64                         1053      45      8      42    252     576    33152
vec_amd_dk512_kturbo4_vq8_0_w64_wg32                    1098      45      8      38    252     576    33152
vec_amd_dk512_kturbo4_vq8_0_wg32                        1074      45      8      38    252     576    33152
vec_amd_dk512_kturbo4_vturbo2                           1070      49      8      42    252     576    33152
vec_amd_dk512_kturbo4_vturbo2_w64                       1088      49      8      42    252     576    33152
vec_amd_dk512_kturbo4_vturbo2_w64_wg32                  1141      49      8      38    252     576    33152
vec_amd_dk512_kturbo4_vturbo2_wg32                      1132      49      8      38    252     576    33152
vec_amd_dk512_kturbo4_vturbo3                           1116      53      8      42    252     576    33152
vec_amd_dk512_kturbo4_vturbo3_w64                       1144      53      8      42    252     576    33152
vec_amd_dk512_kturbo4_vturbo3_w64_wg32                  1190      53      8      38    252     576    33152
vec_amd_dk512_kturbo4_vturbo3_wg32                      1169      53      8      38    252     576    33152
vec_amd_dk512_kturbo4_vturbo4                           1054      53      8      39    252     576    33152
vec_amd_dk512_kturbo4_vturbo4_w64                       1067      53      8      39    252     576    33152
vec_amd_dk512_kturbo4_vturbo4_w64_wg32                  1133      53      8      44    252     576    33152
vec_amd_dk512_kturbo4_vturbo4_wg32                      1105      53      8      43    252     576    33152
vec_amd_dk576_dv512_kf16_vf16                            880       6      8      43    187     640    33152
vec_amd_dk576_dv512_kf16_vf16_w64                        776       4      6      29    187    1024    16576
vec_amd_dk576_dv512_kf16_vf16_w64_wg32                   812       4      6      26    200    1024    16576
vec_amd_dk576_dv512_kf16_vf16_wg32                       902       6      8      40    200     640    33152
vec_amd_dk576_dv512_kq8_0_vq8_0                         1030      42      8      42    230     640    33152
vec_amd_dk576_dv512_kq8_0_vq8_0_w64                      874      24      6      28    187    1024    16576
vec_amd_dk576_dv512_kq8_0_vq8_0_w64_wg32                 935      24      6      32    187    1024    16576
vec_amd_dk576_dv512_kq8_0_vq8_0_wg32                    1096      42      8      45    187     640    33152
vec_amd_dk640_kf16_vturbo2                              1051      31      9      53    232     512    41344
vec_amd_dk640_kf16_vturbo2_w64                          1075      31      9      51    232     512    41344
vec_amd_dk640_kf16_vturbo3                              1115      36      9      50    248     512    41344
vec_amd_dk640_kf16_vturbo3_w64                          1149      36      9      55    248     576    41344
vec_amd_dk640_kf16_vturbo4                              1054      36      9      50    252     512    41344
vec_amd_dk640_kf16_vturbo4_w64                          1066      37      9      51    252     512    41344
vec_amd_dk640_kq8_0_vturbo2                             1108      51      9      49    232     512    41344
vec_amd_dk640_kq8_0_vturbo2_w64                         1122      51      9      50    232     512    41344
vec_amd_dk640_kq8_0_vturbo3                             1189      56      9      49    248     640    41344
vec_amd_dk640_kq8_0_vturbo3_w64                         1192      56      9      50    248     576    41344
vec_amd_dk640_kq8_0_vturbo4                             1108      56      9      49    252     512    41344
vec_amd_dk640_kq8_0_vturbo4_w64                         1120      56      9      50    252     512    41344
vec_amd_dk640_kturbo2_vf16                              1057      31      9      54    232     512    41344
vec_amd_dk640_kturbo2_vf16_w64                          1066      31      9      50    232     512    41344
vec_amd_dk640_kturbo2_vq8_0                             1101      51      9      49    232     512    41344
vec_amd_dk640_kturbo2_vq8_0_w64                         1133      51      9      49    232     512    41344
vec_amd_dk640_kturbo2_vturbo2                           1161      56      9      48    232     512    41344
vec_amd_dk640_kturbo2_vturbo2_w64                       1180      56      9      48    232     512    41344
vec_amd_dk640_kturbo2_vturbo3                           1225      61      9      49    248     512    41344
vec_amd_dk640_kturbo2_vturbo3_w64                       1260      61      9      49    248     640    41344
vec_amd_dk640_kturbo2_vturbo4                           1166      61      9      49    252     512    41344
vec_amd_dk640_kturbo2_vturbo4_w64                       1183      61      9      49    252     512    41344
vec_amd_dk640_kturbo3_vf16                              1102      36      9      51    248     512    41344
vec_amd_dk640_kturbo3_vf16_w64                          1149      36      9      51    248     448    41344
vec_amd_dk640_kturbo3_vq8_0                             1165      56      9      49    248     512    41344
vec_amd_dk640_kturbo3_vq8_0_w64                         1191      56      9      49    248     576    41344
vec_amd_dk640_kturbo3_vturbo2                           1225      61      9      49    248     512    41344
vec_amd_dk640_kturbo3_vturbo2_w64                       1241      61      9      49    248     512    41344
vec_amd_dk640_kturbo3_vturbo3                           1265      66      9      48    248     512    41344
vec_amd_dk640_kturbo3_vturbo3_w64                       1324      66      9      47    248     640    41344
vec_amd_dk640_kturbo3_vturbo4                           1222      66      9      49    252     512    41344
vec_amd_dk640_kturbo3_vturbo4_w64                       1247      66      9      49    252     512    41344
vec_amd_dk640_kturbo4_vf16                              1068      36      9      51    252     512    41344
vec_amd_dk640_kturbo4_vf16_w64                          1071      37      9      51    252     448    41344
vec_amd_dk640_kturbo4_vq8_0                             1112      56      9      49    252     512    41344
vec_amd_dk640_kturbo4_vq8_0_w64                         1130      56      9      49    252     576    41344
vec_amd_dk640_kturbo4_vturbo2                           1170      61      9      49    252     512    41344
vec_amd_dk640_kturbo4_vturbo2_w64                       1182      61      9      49    252     512    41344
vec_amd_dk640_kturbo4_vturbo3                           1229      66      9      49    252     512    41344
vec_amd_dk640_kturbo4_vturbo3_w64                       1256      66      9      49    252     576    41344
vec_amd_dk640_kturbo4_vturbo4                           1135      66      9      45    252     512    41344
vec_amd_dk640_kturbo4_vturbo4_w64                       1153      66      9      47    252     512    41344
vec_amd_dk64_kf16_vf16                                   865       3     16      35    200    1024    35840
vec_amd_dk64_kf16_vf16_w64                               796       3     10      27    200    1024    35840
vec_amd_dk64_kf16_vq4_0                                  884       5     13      42    187    1024    26880
vec_amd_dk64_kf16_vq4_0_w64                              852       5     10      36    200    1024    35840
vec_amd_dk64_kf16_vq8_0                                  898      11     16      36    200    1024    35840
vec_amd_dk64_kf16_vq8_0_w64                              838      11     10      28    216    1024    35840
vec_amd_dk64_kq4_0_vf16                                  890       5     13      42    187     768    26880
vec_amd_dk64_kq4_0_vf16_w64                              850       5     10      36    200     768    35840
vec_amd_dk64_kq4_0_vq4_0                                 942       7     13      52    187    1024    26880
vec_amd_dk64_kq4_0_vq4_0_w64                             914       9     10      49    187    1024    35840
vec_amd_dk64_kq4_0_vq8_0                                 946      13     13      44    187     768    26880
vec_amd_dk64_kq4_0_vq8_0_w64                             907      13     10      40    187     768    35840
vec_amd_dk64_kq8_0_vf16                                  896      11     16      36    200    1024    35840
vec_amd_dk64_kq8_0_vf16_w64                              836      11     10      28    216     768    35840
vec_amd_dk64_kq8_0_vq4_0                                 935      13     13      44    187    1024    26880
vec_amd_dk64_kq8_0_vq4_0_w64                             906      13     10      40    187    1024    35840
vec_amd_dk64_kq8_0_vq8_0                                 948      19     16      39    192    1024    35840
vec_amd_dk64_kq8_0_vq8_0_w64                             879      19     10      35    187    1024    35840
vec_amd_dk72_kf16_vf16                                   821       3     10      28    201    1024    19968
vec_amd_dk72_kf16_vf16_w64                               775       3      6      23    201    1024    19968
