/home/docs/checkouts/readthedocs.org/user_builds/advanced-micro-devices-composable-kernel/checkouts/develop/include/ck/tensor_operation/gpu/grid/gridwise_moe_gemm.hpp File Reference

• develop
• include
• ck
• tensor_operation
• gpu
• grid

Classes | Namespaces | Macros | Enumerations | Functions

gridwise_moe_gemm.hpp File Reference

#include "ck/utility/common_header.hpp"
#include "ck/tensor_description/multi_index_transform_helper.hpp"
#include "ck/tensor_description/tensor_descriptor.hpp"
#include "ck/tensor_description/tensor_descriptor_helper.hpp"
#include "ck/tensor_operation/gpu/grid/block_to_ctile_map.hpp"
#include "ck/tensor_operation/gpu/block/blockwise_gemm_pipeline_xdlops_b_preshuffle_selector.hpp"
#include "ck/tensor_operation/gpu/block/thread_group_tensor_slice_transfer_v4r1_gather.hpp"
#include "ck/tensor_operation/gpu/block/thread_group_tensor_slice_transfer_v6r1.hpp"
#include "ck/tensor_operation/gpu/thread/threadwise_tensor_slice_transfer.hpp"
#include "ck/tensor_operation/gpu/element/unary_element_wise_operation.hpp"
#include "ck/tensor_operation/gpu/block/thread_group_tensor_slice_transfer_v7r3_scatter.hpp"

Go to the source code of this file.

Classes
struct	ck::GridwiseMoeGemm< ALayout, BLayout, DsLayout, CLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, CDataType, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1Value, BK1Value, MPerXdl, NPerXdl, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, AThreadTransferSrcResetCoordinateAfterRun, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BThreadTransferSrcResetCoordinateAfterRun, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, CDEShuffleBlockTransferScalarPerVectors, BlkGemmPipeSched, BlkGemmPipelineVer, ActivationOperation, NSwizzle, IsInputGemm, MulRoutedWeight, PerTokenQuant, IndexType, ComputeTypeA, ComputeTypeB, LDSTypeA, LDSTypeB >
struct	ck::GridwiseMoeGemm< ALayout, BLayout, DsLayout, CLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, CDataType, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1Value, BK1Value, MPerXdl, NPerXdl, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, AThreadTransferSrcResetCoordinateAfterRun, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BThreadTransferSrcResetCoordinateAfterRun, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, CDEShuffleBlockTransferScalarPerVectors, BlkGemmPipeSched, BlkGemmPipelineVer, ActivationOperation, NSwizzle, IsInputGemm, MulRoutedWeight, PerTokenQuant, IndexType, ComputeTypeA, ComputeTypeB, LDSTypeA, LDSTypeB >::Problem
struct	ck::GridwiseMoeGemm< ALayout, BLayout, DsLayout, CLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, CDataType, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1Value, BK1Value, MPerXdl, NPerXdl, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, AThreadTransferSrcResetCoordinateAfterRun, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BThreadTransferSrcResetCoordinateAfterRun, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, CDEShuffleBlockTransferScalarPerVectors, BlkGemmPipeSched, BlkGemmPipelineVer, ActivationOperation, NSwizzle, IsInputGemm, MulRoutedWeight, PerTokenQuant, IndexType, ComputeTypeA, ComputeTypeB, LDSTypeA, LDSTypeB >::Argument
struct	ck::GridwiseMoeGemm< ALayout, BLayout, DsLayout, CLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, CDataType, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1Value, BK1Value, MPerXdl, NPerXdl, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, AThreadTransferSrcResetCoordinateAfterRun, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BThreadTransferSrcResetCoordinateAfterRun, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, CDEShuffleBlockTransferScalarPerVectors, BlkGemmPipeSched, BlkGemmPipelineVer, ActivationOperation, NSwizzle, IsInputGemm, MulRoutedWeight, PerTokenQuant, IndexType, ComputeTypeA, ComputeTypeB, LDSTypeA, LDSTypeB >::SplitKBatchOffset

Namespaces
	ck

Macros
#define	DEBUG_LOG   0

Enumerations
enum	ck::Activation {   ck::gelu_and_mul = 0 ,   ck::silu_and_mul = 1 ,   ck::gelu_and_mul = 0 ,   ck::silu_and_mul = 1 ,   ck::gelu_and_mul = 0 ,   ck::silu_and_mul = 1 ,   ck::gelu_and_mul = 0 ,   ck::silu_and_mul = 1 ,   ck::gelu_and_mul = 0 ,   ck::silu_and_mul = 1 }

Functions
template<typename GridwiseGemm , bool HasMainKBlockLoop, InMemoryDataOperationEnum CGlobalMemoryDataOperation, index_t MinimumOccupancy = 1, TailNumber TailNum = TailNumber::Even>
__global__ void	ck::kernel_moe_gemm (typename GridwiseGemm::Argument karg)
template<typename GridwiseGemm , bool HasMainKBlockLoop, InMemoryDataOperationEnum CGlobalMemoryDataOperation, index_t MinimumOccupancy = 1, TailNumber TailNum = TailNumber::Even>
__global__ void	ck::kernel_moe_gemm_2lds (typename GridwiseGemm::Argument karg)

Macro Definition Documentation

◆ DEBUG_LOG

#define DEBUG_LOG   0