True baseline benchmarking results.

This commit is contained in:
Ville Pietilä
2026-02-04 07:55:28 -05:00
parent 20cf6df685
commit 3e6415a8ea

View File

@@ -0,0 +1,5 @@
1 1 1 0 1 0 1 2 32 32 4 4 3 3 200 200 1 1 1 1 1 1 1 1; DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle<256, 64, 64, 32, Default, 16, 16, 2, 2, 4, 4, 4, 1, 1, 1>; 2.85567
1 1 1 0 1 0 1 2 32 32 8 8 3 3 200 200 2 2 1 1 1 1 1 1; DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle_V3_DirectLoad<128, 16, 64, 64, Default, 16, 16, 1, 2, 8, 8, 4, 1, 1, BlkGemmPipelineScheduler: Intrawave, BlkGemmPipelineVersion: v1>; 10.3518
1 1 1 0 1 0 1 2 32 32 8 8 3 3 100 100 1 2 1 1 1 1 1 1; DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle_V3_DirectLoad<256, 256, 32, 64, Default, 32, 32, 2, 1, 8, 8, 8, 1, 1, BlkGemmPipelineScheduler: Intrawave, BlkGemmPipelineVersion: v1>; 15.4196
1 1 1 0 1 0 1 2 1 32 2376 256 3 3 100 100 1 1 1 1 1 1 1 1; DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle<256, 256, 128, 32, Default, 32, 32, 4, 2, 8, 8, 8, 1, 1, 1>; 849.773
1 1 1 0 1 0 1 2 1 32 256 256 3 3 100 100 1 1 1 1 1 1 1 1; DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle<256, 256, 128, 32, OddC, 32, 32, 4, 2, 8, 8, 8, 1, 1, 1>; 990.991
1 1 1 1 0 1 0 1 2 32 32 4 4 3 3 200 200 1 1 1 1 1 1 1 1 DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle<256, 64, 64, 32, Default, 16, 16, 2, 2, 4, 4, 4, 1, 1, 1> 2.85567
2 1 1 1 0 1 0 1 2 32 32 8 8 3 3 200 200 2 2 1 1 1 1 1 1 DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle_V3_DirectLoad<128, 16, 64, 64, Default, 16, 16, 1, 2, 8, 8, 4, 1, 1, BlkGemmPipelineScheduler: Intrawave, BlkGemmPipelineVersion: v1> 10.3518
3 1 1 1 0 1 0 1 2 32 32 8 8 3 3 100 100 1 2 1 1 1 1 1 1 DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle_V3_DirectLoad<256, 256, 32, 64, Default, 32, 32, 2, 1, 8, 8, 8, 1, 1, BlkGemmPipelineScheduler: Intrawave, BlkGemmPipelineVersion: v1> 15.4196
4 1 1 1 0 1 0 1 2 1 32 2376 256 3 3 100 100 1 1 1 1 1 1 1 1 DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle<256, 256, 128, 32, Default, 32, 32, 4, 2, 8, 8, 8, 1, 1, 1> 849.773
5 1 1 1 0 1 0 1 2 1 32 256 256 3 3 100 100 1 1 1 1 1 1 1 1 DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle<256, 256, 128, 32, OddC, 32, 32, 4, 2, 8, 8, 8, 1, 1, 1> 990.991