Add multiple Ds instances

This commit is contained in:
Enrico Degregori
2025-08-06 13:00:33 +00:00
parent e6b7d5ed65
commit c71f2f25eb
7 changed files with 171 additions and 21 deletions

View File

@@ -273,8 +273,8 @@ struct DeviceGroupedConvBwdWeightMultipleD_Wmma_CShuffleV3
using CGridDesc_M_N = remove_cvref_t<decltype(ABCGridDescs{}[I2])>;
using GridwiseGemm = GridwiseGemm_wmma_cshuffle_v3<
tensor_layout::gemm::RowMajor,
tensor_layout::gemm::ColumnMajor,
tensor_layout::gemm::RowMajor,
Tuple<>,
tensor_layout::gemm::RowMajor,
Tuple<ADataType>,

View File

@@ -323,8 +323,8 @@ struct DeviceGroupedConvBwdWeightTwoStage_Wmma_CShuffleV3
remove_cvref_t<decltype(GetElementwiseCGridDesc<NDimSpatial>())>;
using GridwiseGemm = GridwiseGemm_wmma_cshuffle_v3<
tensor_layout::gemm::RowMajor,
tensor_layout::gemm::ColumnMajor,
tensor_layout::gemm::RowMajor,
Tuple<>,
tensor_layout::gemm::RowMajor,
Tuple<ADataType>,