Composable Kernel: ck::tensor_operation::device::DeviceGemmMultipleDLayernorm_Wmma_CShuffleV3< ALayout, BLayout, DsLayout, HLayout, ADataType, BDataType, DsDataType, HDataType, AccDataType, CShuffleDataType, EMeanVarDataType, GammaDataType, BetaDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, HElementwiseOperation, GemmSpec, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerWmma, NPerWmma, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CDEShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, CDEShuffleBlockTransferScalarPerVector, LayernormThreadClusterSize_M_N, LayernormThreadSliceSize_M, BlkGemmPipeSched, BlkGemmPipelineVer, ComputeTypeA, ComputeTypeB, PermuteA, PermuteB >::Invoker Struct Reference

Invoker Struct Reference

Invoker Struct Reference#

Composable Kernel: ck::tensor_operation::device::DeviceGemmMultipleDLayernorm_Wmma_CShuffleV3< ALayout, BLayout, DsLayout, HLayout, ADataType, BDataType, DsDataType, HDataType, AccDataType, CShuffleDataType, EMeanVarDataType, GammaDataType, BetaDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, HElementwiseOperation, GemmSpec, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerWmma, NPerWmma, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CDEShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, CDEShuffleBlockTransferScalarPerVector, LayernormThreadClusterSize_M_N, LayernormThreadSliceSize_M, BlkGemmPipeSched, BlkGemmPipelineVer, ComputeTypeA, ComputeTypeB, PermuteA, PermuteB >::Invoker Struct Reference

Public Member Functions | List of all members

ck::tensor_operation::device::DeviceGemmMultipleDLayernorm_Wmma_CShuffleV3< ALayout, BLayout, DsLayout, HLayout, ADataType, BDataType, DsDataType, HDataType, AccDataType, CShuffleDataType, EMeanVarDataType, GammaDataType, BetaDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, HElementwiseOperation, GemmSpec, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerWmma, NPerWmma, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CDEShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, CDEShuffleBlockTransferScalarPerVector, LayernormThreadClusterSize_M_N, LayernormThreadSliceSize_M, BlkGemmPipeSched, BlkGemmPipelineVer, ComputeTypeA, ComputeTypeB, PermuteA, PermuteB >::Invoker Struct Reference

#include <device_gemm_multiple_d_layernorm_wmma_cshuffle_v3.hpp>

Inheritance diagram for ck::tensor_operation::device::DeviceGemmMultipleDLayernorm_Wmma_CShuffleV3< ALayout, BLayout, DsLayout, HLayout, ADataType, BDataType, DsDataType, HDataType, AccDataType, CShuffleDataType, EMeanVarDataType, GammaDataType, BetaDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, HElementwiseOperation, GemmSpec, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerWmma, NPerWmma, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CDEShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, CDEShuffleBlockTransferScalarPerVector, LayernormThreadClusterSize_M_N, LayernormThreadSliceSize_M, BlkGemmPipeSched, BlkGemmPipelineVer, ComputeTypeA, ComputeTypeB, PermuteA, PermuteB >::Invoker:

Public Member Functions
float	Run (const Argument &arg, const StreamConfig &stream_config=StreamConfig{})
float	Run (const BaseArgument *p_arg, const StreamConfig &stream_config=StreamConfig{}) override
Public Member Functions inherited from ck::tensor_operation::device::BaseInvoker
	BaseInvoker ()=default
	BaseInvoker (const BaseInvoker &)=default
BaseInvoker &	operator= (const BaseInvoker &)=default
virtual	~BaseInvoker ()

Member Function Documentation

◆ Run() [1/2]

template<typename ALayout, typename BLayout, typename DsLayout, typename HLayout, typename ADataType, typename BDataType, typename DsDataType, typename HDataType, typename AccDataType, typename CShuffleDataType, typename EMeanVarDataType, typename GammaDataType, typename BetaDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename HElementwiseOperation, GemmSpecialization GemmSpec, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerWmma, index_t NPerWmma, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, typename CDEShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, index_t CDEShuffleBlockTransferScalarPerVector, typename LayernormThreadClusterSize_M_N, index_t LayernormThreadSliceSize_M, BlockGemmPipelineScheduler BlkGemmPipeSched = BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion BlkGemmPipelineVer = BlockGemmPipelineVersion::v1, typename ComputeTypeA = HDataType, typename ComputeTypeB = ComputeTypeA, bool PermuteA = false, bool PermuteB = false>

float ck::tensor_operation::device::DeviceGemmMultipleDLayernorm_Wmma_CShuffleV3< ALayout, BLayout, DsLayout, HLayout, ADataType, BDataType, DsDataType, HDataType, AccDataType, CShuffleDataType, EMeanVarDataType, GammaDataType, BetaDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, HElementwiseOperation, GemmSpec, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerWmma, NPerWmma, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CDEShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, CDEShuffleBlockTransferScalarPerVector, LayernormThreadClusterSize_M_N, LayernormThreadSliceSize_M, BlkGemmPipeSched, BlkGemmPipelineVer, ComputeTypeA, ComputeTypeB, PermuteA, PermuteB >::Invoker::Run	(	const Argument &	arg,
		const StreamConfig &	stream_config = StreamConfig{} )

inline

◆ Run() [2/2]

template<typename ALayout, typename BLayout, typename DsLayout, typename HLayout, typename ADataType, typename BDataType, typename DsDataType, typename HDataType, typename AccDataType, typename CShuffleDataType, typename EMeanVarDataType, typename GammaDataType, typename BetaDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename HElementwiseOperation, GemmSpecialization GemmSpec, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerWmma, index_t NPerWmma, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, typename CDEShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, index_t CDEShuffleBlockTransferScalarPerVector, typename LayernormThreadClusterSize_M_N, index_t LayernormThreadSliceSize_M, BlockGemmPipelineScheduler BlkGemmPipeSched = BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion BlkGemmPipelineVer = BlockGemmPipelineVersion::v1, typename ComputeTypeA = HDataType, typename ComputeTypeB = ComputeTypeA, bool PermuteA = false, bool PermuteB = false>

float ck::tensor_operation::device::DeviceGemmMultipleDLayernorm_Wmma_CShuffleV3< ALayout, BLayout, DsLayout, HLayout, ADataType, BDataType, DsDataType, HDataType, AccDataType, CShuffleDataType, EMeanVarDataType, GammaDataType, BetaDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, HElementwiseOperation, GemmSpec, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerWmma, NPerWmma, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CDEShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, CDEShuffleBlockTransferScalarPerVector, LayernormThreadClusterSize_M_N, LayernormThreadSliceSize_M, BlkGemmPipeSched, BlkGemmPipelineVer, ComputeTypeA, ComputeTypeB, PermuteA, PermuteB >::Invoker::Run	(	const BaseArgument *	p_arg,
		const StreamConfig &	stream_config = StreamConfig{} )

inlineoverridevirtual

Reimplemented from ck::tensor_operation::device::BaseInvoker.

The documentation for this struct was generated from the following file:

device_gemm_multiple_d_layernorm_wmma_cshuffle_v3.hpp