Support Wave32 in CK_TILE - Part 1 (#2594)

* Support wave32/wave64 in CK_TILE - Part 1

* remove blocksize in kernel launch

* fix build error

* fix clang format

* fix clang format 2

* fix clang format 3

* fix fmha build error

* fix fmha build 2

* fix fmha build 3

* fix build error 4

* address review comment

* update change log

* replace KernelBlockSize with kBlockSize

* fix CI fail

* fix clang format

* address review comment and rebase code.

* fix universal test fail

---------

Co-authored-by: Lin, Qun <Quentin.Lin+amdeng@amd.com>
Co-authored-by: Thomas Ning <Thomas.Ning@amd.com>
This commit is contained in:
linqunAMD
2025-08-19 01:08:31 +08:00
committed by GitHub
parent 26d3300930
commit 9fcc1ee9fd
113 changed files with 610 additions and 531 deletions

View File

@@ -354,7 +354,7 @@ struct GroupedConvolutionBackwardWeightKernel
using GemmDsLayout = remove_cvref_t<typename EpiloguePipeline::DsLayout>;
static constexpr index_t NumDTensor = GroupedConvTraitsType::NumDTensor;
static constexpr index_t KernelBlockSize = GemmPipeline::BlockSize;
static constexpr index_t kBlockSize = GemmPipeline::BlockSize;
using InDataType = remove_cvref_t<typename GemmPipeline::ADataType>;
using WeiDataType = remove_cvref_t<typename GemmPipeline::BDataType>;
@@ -393,7 +393,7 @@ struct GroupedConvolutionBackwardWeightKernel
TilePartitioner::GridSize(kargs.GemmM, kargs.GemmN), kargs.GemmBatch, kargs.k_batch);
}
CK_TILE_HOST static constexpr auto BlockSize() { return dim3(KernelBlockSize); }
CK_TILE_HOST static constexpr auto BlockSize() { return dim3(kBlockSize); }
CK_TILE_HOST static constexpr GroupedConvBwdWeightKernelArgsSpecialized
MakeKernelArgs(const GroupedConvBwdWeightHostArgs& hostArgs)

View File

@@ -361,7 +361,7 @@ struct GroupedConvolutionForwardKernel
using GemmDsLayout = remove_cvref_t<typename EpiloguePipeline::DsLayout>;
static constexpr index_t NumDTensor = GroupedConvTraitsType::NumDTensor;
static constexpr index_t KernelBlockSize = GemmPipeline::BlockSize;
static constexpr index_t kBlockSize = GemmPipeline::BlockSize;
using InDataType = remove_cvref_t<typename GemmPipeline::ADataType>;
using WeiDataType = remove_cvref_t<typename GemmPipeline::BDataType>;
@@ -398,7 +398,7 @@ struct GroupedConvolutionForwardKernel
TilePartitioner::GridSize(kargs.GemmM, kargs.GemmN), kargs.GemmBatch, kargs.k_batch);
}
CK_TILE_HOST static constexpr auto BlockSize() { return dim3(KernelBlockSize); }
CK_TILE_HOST static constexpr auto BlockSize() { return dim3(kBlockSize); }
CK_TILE_HOST static constexpr GroupedConvFwdKernelArgsSpecialized
MakeKernelArgs(const GroupedConvFwdHostArgs& hostArgs)