Composable Kernel: ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType > Struct Template Reference

GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType > Struct Template Reference

GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType > Struct Template Reference#

Composable Kernel: ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType > Struct Template Reference

Classes | Public Types | Static Public Member Functions | Static Public Attributes | List of all members

ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType > Struct Template Reference

#include <gridwise_gemm_xdlops_splitk_lds_direct_load.hpp>

Classes
struct	Argument

Public Types
using	ThisThreadBlock = ThisThreadBlock<BlockSize>
using	GridwiseGemmPipe
using	CGridDesc_M_N = remove_cvref_t<decltype(MakeCGridDescriptor_M_N(1, 1, 1))>
using	DefaultBlock2CTileMap = remove_cvref_t<decltype(MakeDefaultBlock2CTileMap())>

Static Public Member Functions
__host__ static __device__ auto	CalculateGridSize (const Argument &karg)
__host__ static __device__ auto	CalculateMPadded (index_t M)
__host__ static __device__ auto	CalculateNPadded (index_t N)
__host__ static __device__ auto	CalculateK0Padded (index_t K, index_t K_Batch=1)
__host__ static __device__ auto	CalculateKPadded (index_t K, index_t K_Batch=1)
__host__ static __device__ auto	MakeAGridDescriptor_KBatch_K0_M_K1 (index_t M, index_t MPad, index_t K, index_t StrideA, index_t KBatch, index_t K0Padded, index_t KPad)
__host__ static __device__ auto	MakeBGridDescriptor_KBatch_K0_N_K1 (index_t K, index_t NPad, index_t N, index_t StrideB, index_t KBatch, index_t K0Padded, index_t KPad)
__host__ static __device__ auto	MakeCGridDescriptor_M_N (index_t M, index_t N, index_t StrideC)
__host__ static __device__ constexpr index_t	GetSharedMemoryNumberOfByte ()
__host__ static __device__ constexpr bool	CheckValidity (const Argument &karg)
__host__ static __device__ auto	GetKPad (index_t K, index_t KBatch)
__host__ static __device__ constexpr bool	CalculateHasMainK0BlockLoop (index_t K0Padded)
template<typename CGridDesc>
__host__ static __device__ constexpr auto	MakeCGridDesc_MBlock_MPerBlock_NBlock_NPerBlock (const CGridDesc &c_m_n_grid_desc)
__host__ static __device__ constexpr auto	GetCBlockDescriptor_MBlock_MPerBlock_NBlock_NPerBlock ()
__host__ static __device__ constexpr auto	MakeDefaultBlock2CTileMap ()
template<bool HasMainKBlockLoop, InMemoryDataOperationEnum CGlobalMemoryDataOperation, typename Block2CTileMap>
static __device__ void	Run (const Argument &karg, void *__restrict__ p_shared_block, const Block2CTileMap &block_2_ctile_map, const AElementwiseOperation a_element_op=AElementwiseOperation{}, const BElementwiseOperation b_element_op=BElementwiseOperation{}, const CElementwiseOperation c_element_op=CElementwiseOperation{})

Static Public Attributes
static constexpr auto	I0 = Number<0>{}
static constexpr auto	I1 = Number<1>{}
static constexpr auto	I2 = Number<2>{}
static constexpr auto	I3 = Number<3>{}
static constexpr auto	I4 = Number<4>{}
static constexpr auto	I5 = Number<5>{}
static constexpr auto	I6 = Number<6>{}
static constexpr auto	I7 = Number<7>{}
static constexpr auto	K1 = Number<K1Value>{}
static constexpr auto	KPerBlock = Number<K1Value * K0PerBlock>{}
static constexpr auto	M01 = 1
static constexpr auto	N01 = 1
static constexpr auto	gemm_padder
static constexpr index_t	MXdlPerWave = MRepeat
static constexpr index_t	NXdlPerWave = NRepeat

Member Typedef Documentation

◆ CGridDesc_M_N

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

using ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::CGridDesc_M_N = remove_cvref_t<decltype(MakeCGridDescriptor_M_N(1, 1, 1))>

◆ DefaultBlock2CTileMap

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

using ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::DefaultBlock2CTileMap = remove_cvref_t<decltype(MakeDefaultBlock2CTileMap())>

◆ GridwiseGemmPipe

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

using ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::GridwiseGemmPipe

Initial value:

remove_cvref_t<

decltype(GridwiseGemmPipeline_Selector<PipelineVer, NumGemmKPrefetchStage, LoopSched>())>

ck::GridwiseGemmPipeline_Selector

constexpr auto GridwiseGemmPipeline_Selector()

Definition gridwise_gemm_pipeline_selector.hpp:31

ck::remove_cvref_t

remove_cv_t< remove_reference_t< T > > remove_cvref_t

Definition type.hpp:297

◆ ThisThreadBlock

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

using ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::ThisThreadBlock = ThisThreadBlock<BlockSize>

Member Function Documentation

◆ CalculateGridSize()

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

__host__ static __device__ auto ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::CalculateGridSize ( const Argument & karg )

inlinestatic

◆ CalculateHasMainK0BlockLoop()

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

__host__ static __device__ constexpr bool ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::CalculateHasMainK0BlockLoop ( index_t K0Padded )

inlinestaticconstexpr

◆ CalculateK0Padded()

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

__host__ static __device__ auto ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::CalculateK0Padded	(	index_t	K,
		index_t	K_Batch = 1 )

inlinestatic

◆ CalculateKPadded()

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

__host__ static __device__ auto ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::CalculateKPadded	(	index_t	K,
		index_t	K_Batch = 1 )

inlinestatic

◆ CalculateMPadded()

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

__host__ static __device__ auto ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::CalculateMPadded ( index_t M )

inlinestatic

◆ CalculateNPadded()

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

__host__ static __device__ auto ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::CalculateNPadded ( index_t N )

inlinestatic

◆ CheckValidity()

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

__host__ static __device__ constexpr bool ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::CheckValidity ( const Argument & karg )

inlinestaticconstexpr

◆ GetCBlockDescriptor_MBlock_MPerBlock_NBlock_NPerBlock()

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

__host__ static __device__ constexpr auto ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::GetCBlockDescriptor_MBlock_MPerBlock_NBlock_NPerBlock ( )

inlinestaticconstexpr

◆ GetKPad()

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

__host__ static __device__ auto ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::GetKPad	(	index_t	K,
		index_t	KBatch )

inlinestatic

◆ GetSharedMemoryNumberOfByte()

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

__host__ static __device__ constexpr index_t ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::GetSharedMemoryNumberOfByte ( )

inlinestaticconstexpr

◆ MakeAGridDescriptor_KBatch_K0_M_K1()

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

__host__ static __device__ auto ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::MakeAGridDescriptor_KBatch_K0_M_K1	(	index_t	M,
		index_t	MPad,
		index_t	K,
		index_t	StrideA,
		index_t	KBatch,
		index_t	K0Padded,
		index_t	KPad )

inlinestatic

◆ MakeBGridDescriptor_KBatch_K0_N_K1()

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

__host__ static __device__ auto ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::MakeBGridDescriptor_KBatch_K0_N_K1	(	index_t	K,
		index_t	NPad,
		index_t	N,
		index_t	StrideB,
		index_t	KBatch,
		index_t	K0Padded,
		index_t	KPad )

inlinestatic

◆ MakeCGridDesc_MBlock_MPerBlock_NBlock_NPerBlock()

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

template<typename CGridDesc>

__host__ static __device__ constexpr auto ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::MakeCGridDesc_MBlock_MPerBlock_NBlock_NPerBlock ( const CGridDesc & c_m_n_grid_desc )

inlinestaticconstexpr

◆ MakeCGridDescriptor_M_N()

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

__host__ static __device__ auto ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::MakeCGridDescriptor_M_N	(	index_t	M,
		index_t	N,
		index_t	StrideC )

inlinestatic

◆ MakeDefaultBlock2CTileMap()

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

__host__ static __device__ constexpr auto ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::MakeDefaultBlock2CTileMap ( )

inlinestaticconstexpr

◆ Run()

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

template<bool HasMainKBlockLoop, InMemoryDataOperationEnum CGlobalMemoryDataOperation, typename Block2CTileMap>

__device__ void ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::Run	(	const Argument &	karg,
		void *__restrict__	p_shared_block,
		const Block2CTileMap &	block_2_ctile_map,
		const AElementwiseOperation	a_element_op = AElementwiseOperation{},
		const BElementwiseOperation	b_element_op = BElementwiseOperation{},
		const CElementwiseOperation	c_element_op = CElementwiseOperation{} )

inlinestatic

Member Data Documentation

◆ gemm_padder

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

auto ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::gemm_padder

staticconstexpr

Initial value:

=
        tensor_operation::device::GemmPadder<GemmSpec, index_t, index_t, index_t>{
            MPerBlock, NPerBlock, K1* K0PerBlock}

◆ I0

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

auto ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::I0 = Number<0>{}

staticconstexpr

◆ I1

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

auto ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::I1 = Number<1>{}

staticconstexpr

◆ I2

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

auto ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::I2 = Number<2>{}

staticconstexpr

◆ I3

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

auto ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::I3 = Number<3>{}

staticconstexpr

◆ I4

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

auto ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::I4 = Number<4>{}

staticconstexpr

◆ I5

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

auto ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::I5 = Number<5>{}

staticconstexpr

◆ I6

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

auto ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::I6 = Number<6>{}

staticconstexpr

◆ I7

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

auto ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::I7 = Number<7>{}

staticconstexpr

◆ K1

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

auto ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::K1 = Number<K1Value>{}

staticconstexpr

◆ KPerBlock

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

auto ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::KPerBlock = Number<K1Value * K0PerBlock>{}

staticconstexpr

◆ M01

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

auto ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::M01 = 1

staticconstexpr

◆ MXdlPerWave

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

index_t ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::MXdlPerWave = MRepeat

staticconstexpr

◆ N01

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

auto ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::N01 = 1

staticconstexpr

◆ NXdlPerWave

template<index_t BlockSize, typename FloatA, typename FloatB, typename FloatAcc, typename FloatC, typename ALayout, typename BLayout, typename CLayout, typename AElementwiseOperation, typename BElementwiseOperation, typename CElementwiseOperation, tensor_operation::device::GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t MPerBlock, index_t NPerBlock, index_t K0PerBlock, index_t MPerXdl, index_t NPerXdl, index_t K1Value, index_t MRepeat, index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, bool ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_K0_N_K1, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, bool BBlockLdsExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, index_t CBlockTransferScalarPerVector_NWaveNPerXDL, typename CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler(), PipelineVersion PipelineVer = PipelineVersion::v4, typename ComputeType = FloatC>

index_t ck::GridwiseGemm_xdlops_splitk_lds_direct_load< BlockSize, FloatA, FloatB, FloatAcc, FloatC, ALayout, BLayout, CLayout, AElementwiseOperation, BElementwiseOperation, CElementwiseOperation, GemmSpec, NumGemmKPrefetchStage, MPerBlock, NPerBlock, K0PerBlock, MPerXdl, NPerXdl, K1Value, MRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_K0_N_K1, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockLdsExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CBlockTransferScalarPerVector_NWaveNPerXDL, CBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, LoopSched, PipelineVer, ComputeType >::NXdlPerWave = NRepeat

staticconstexpr

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

gridwise_gemm_xdlops_splitk_lds_direct_load.hpp