Performance enhancement for Volta Tensor Cores TN layout (#53)

* Fixed performance defect with indirect access to pointer array for Volta TensorCores TN arrangement.

* Updated patch version and changelog.

* Updated patch version and changelog.

* Added link to changelog in readme.

* Fixed markdown link
This commit is contained in:
Andrew Kerr
2019-07-10 10:54:12 -07:00
committed by GitHub
parent eb41735933
commit b5cab177a9
5 changed files with 18 additions and 15 deletions
+1 -4
View File
@@ -183,7 +183,7 @@ TEST(Volta884_f16_s884gemm_128x128x32_tt, short_480x280x224) {
// Contiguous - s884gemm
//
////////////////////////////////////////////////////////////////////////////////////////////////////
#if 0
TEST(Volta884_f16_s884gemm_64x64x32_nt, 64x64x32) {
typedef cutlass::gemm::Volta884GemmTraits<
@@ -218,7 +218,6 @@ TEST(Volta884_f16_s884gemm_64x64x32_nt, 64x64x30_residue) {
run_gemm<GemmTraits>(64, 64, 30);
}
#if 0
////////////////////////////////////////////////////////////////////////////////////////////////////
TEST(Volta884_f16_s884gemm_64x64x32_nt, 64x64x64) {
@@ -874,7 +873,6 @@ TEST(Volta884_f16_s884gemm_128x128x32_nn, 392x264x192) {
run_gemm<GemmTraits>(392, 264, 192);
}
#endif
////////////////////////////////////////////////////////////////////////////////////////////////////
@@ -1281,7 +1279,6 @@ TEST(Volta884_f16_s884gemm_f16_128x256x32_tn, 480x280x224) {
run_gemm<GemmTraits>(480, 280, 224);
}
#endif
////////////////////////////////////////////////////////////////////////////////////////////////////
#endif // if defined(CUTLASS_ENABLE_TENSOR_CORE_MMA)