Argument Struct Reference

• ck
• tensor_operation
• device
• DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle
• Argument

Public Member Functions | Public Attributes | List of all members

ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument Struct Reference

#include <device_grouped_conv_fwd_multiple_d_multiple_r_xdl_cshuffle.hpp>

Inheritance diagram for ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument:

Public Member Functions
	Argument (const void *p_a, const void *p_b, const std::array< const void *, NumDTensor > &p_ds, void *p_e, std::array< void *, NumRTensor > p_rs, const std::array< index_t, NDimSpatial+3 > &a_g_n_c_wis_lengths, const std::array< index_t, NDimSpatial+3 > &a_g_n_c_wis_strides, const std::array< index_t, NDimSpatial+3 > &b_g_k_c_xs_lengths, const std::array< index_t, NDimSpatial+3 > &b_g_k_c_xs_strides, const std::array< std::array< index_t, NDimSpatial+3 >, NumDTensor > &ds_g_n_k_wos_lengths, const std::array< std::array< index_t, NDimSpatial+3 >, NumDTensor > &ds_g_n_k_wos_strides, const std::array< index_t, NDimSpatial+3 > &e_g_n_k_wos_lengths, const std::array< index_t, NDimSpatial+3 > &e_g_n_k_wos_strides, const std::array< index_t, NDimSpatial+2 > &r_g_n_wos_lengths, const std::array< index_t, NDimSpatial+2 > &r_g_n_wos_strides, const std::array< index_t, NDimSpatial > &conv_filter_strides, const std::array< index_t, NDimSpatial > &conv_filter_dilations, const std::array< index_t, NDimSpatial > &input_left_pads, const std::array< index_t, NDimSpatial > &input_right_pads, const AElementwiseOperation &a_element_op, const BElementwiseOperation &b_element_op, const CDEElementwiseOperation &cde_element_op, const QsElementwiseOperation &qs_element_op, const RsElementwiseOperation &rs_element_op)
void	Print () const
Public Member Functions inherited from ck::tensor_operation::device::BaseArgument
	BaseArgument ()=default
	BaseArgument (const BaseArgument &)=default
BaseArgument &	operator= (const BaseArgument &)=default
virtual	~BaseArgument ()

Public Attributes
const ADataType *	p_a_grid_
const BDataType *	p_b_grid_
GridwiseGemm64::DsGridPointer	p_ds_grid_
EDataType *	p_e_grid_
GridwiseGemm64::RsGridPointer	p_rs_grid_
ConvToGemmFwdTransformer	conv_to_gemm_transformer_
AGridDesc_M_K	a_grid_desc_m_k_
BGridDesc_N_K	b_grid_desc_n_k_
EGridDesc_M_N	ds_grid_desc_m_n_
EGridDesc_M_N	e_grid_desc_m_n_
RGridDesc_M	r_grid_desc_m_
AGridDesc_AK0_M_AK1	a_grid_desc_ak0_m_ak1_
BGridDesc_BK0_N_BK1	b_grid_desc_bk0_n_bk1_
Block2ETileMap	block_2_etile_map_
ComputePtrOffsetOfStridedBatch< NumDTensor, NumRTensor >	compute_ptr_offset_of_batch_
AElementwiseOperation	a_element_op_
BElementwiseOperation	b_element_op_
CDEElementwiseOperation	cde_element_op_
QsElementwiseOperation	qs_element_op_
RsElementwiseOperation	rs_element_op_
std::array< index_t, NDimSpatial+3 >	a_g_n_c_wis_lengths_
std::array< index_t, NDimSpatial+3 >	a_g_n_c_wis_strides_
std::array< index_t, NDimSpatial+3 >	b_g_k_c_xs_lengths_
std::array< index_t, NDimSpatial+3 >	b_g_k_c_xs_strides_
std::array< std::array< index_t, NDimSpatial+3 >, NumDTensor >	ds_g_n_k_wos_lengths_
std::array< std::array< index_t, NDimSpatial+3 >, NumDTensor >	ds_g_n_k_wos_strides_
std::array< index_t, NDimSpatial+3 >	e_g_n_k_wos_lengths_
std::array< index_t, NDimSpatial+3 >	e_g_n_k_wos_strides_
std::array< index_t, NDimSpatial >	conv_filter_strides_
std::array< index_t, NDimSpatial >	conv_filter_dilations_
std::array< index_t, NDimSpatial >	input_left_pads_
std::array< index_t, NDimSpatial >	input_right_pads_
Public Attributes inherited from ck::tensor_operation::device::BaseArgument
void *	p_workspace_ = nullptr

Constructor & Destructor Documentation

◆ Argument()

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::Argument ( const void * p_a, const void * p_b, const std::array< const void *, NumDTensor > & p_ds, void * p_e, std::array< void *, NumRTensor > p_rs, const std::array< index_t, NDimSpatial+3 > & a_g_n_c_wis_lengths, const std::array< index_t, NDimSpatial+3 > & a_g_n_c_wis_strides, const std::array< index_t, NDimSpatial+3 > & b_g_k_c_xs_lengths, const std::array< index_t, NDimSpatial+3 > & b_g_k_c_xs_strides, const std::array< std::array< index_t, NDimSpatial+3 >, NumDTensor > & ds_g_n_k_wos_lengths, const std::array< std::array< index_t, NDimSpatial+3 >, NumDTensor > & ds_g_n_k_wos_strides, const std::array< index_t, NDimSpatial+3 > & e_g_n_k_wos_lengths, const std::array< index_t, NDimSpatial+3 > & e_g_n_k_wos_strides, const std::array< index_t, NDimSpatial+2 > & r_g_n_wos_lengths, const std::array< index_t, NDimSpatial+2 > & r_g_n_wos_strides, const std::array< index_t, NDimSpatial > & conv_filter_strides, const std::array< index_t, NDimSpatial > & conv_filter_dilations, const std::array< index_t, NDimSpatial > & input_left_pads, const std::array< index_t, NDimSpatial > & input_right_pads, const AElementwiseOperation & a_element_op, const BElementwiseOperation & b_element_op, const CDEElementwiseOperation & cde_element_op, const QsElementwiseOperation & qs_element_op, const RsElementwiseOperation & rs_element_op )

inline

Member Function Documentation

◆ Print()

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

void ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::Print ( ) const

inline

Member Data Documentation

◆ a_element_op_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

AElementwiseOperation ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::a_element_op_

◆ a_g_n_c_wis_lengths_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

std::array<index_t, NDimSpatial + 3> ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::a_g_n_c_wis_lengths_

◆ a_g_n_c_wis_strides_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

std::array<index_t, NDimSpatial + 3> ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::a_g_n_c_wis_strides_

◆ a_grid_desc_ak0_m_ak1_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

AGridDesc_AK0_M_AK1 ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::a_grid_desc_ak0_m_ak1_

◆ a_grid_desc_m_k_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

AGridDesc_M_K ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::a_grid_desc_m_k_

◆ b_element_op_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

BElementwiseOperation ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::b_element_op_

◆ b_g_k_c_xs_lengths_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

std::array<index_t, NDimSpatial + 3> ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::b_g_k_c_xs_lengths_

◆ b_g_k_c_xs_strides_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

std::array<index_t, NDimSpatial + 3> ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::b_g_k_c_xs_strides_

◆ b_grid_desc_bk0_n_bk1_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

BGridDesc_BK0_N_BK1 ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::b_grid_desc_bk0_n_bk1_

◆ b_grid_desc_n_k_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

BGridDesc_N_K ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::b_grid_desc_n_k_

◆ block_2_etile_map_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

Block2ETileMap ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::block_2_etile_map_

◆ cde_element_op_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

CDEElementwiseOperation ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::cde_element_op_

◆ compute_ptr_offset_of_batch_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

ComputePtrOffsetOfStridedBatch<NumDTensor, NumRTensor> ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::compute_ptr_offset_of_batch_

◆ conv_filter_dilations_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

std::array<index_t, NDimSpatial> ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::conv_filter_dilations_

◆ conv_filter_strides_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

std::array<index_t, NDimSpatial> ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::conv_filter_strides_

◆ conv_to_gemm_transformer_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

ConvToGemmFwdTransformer ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::conv_to_gemm_transformer_

◆ ds_g_n_k_wos_lengths_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

std::array<std::array<index_t, NDimSpatial + 3>, NumDTensor> ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::ds_g_n_k_wos_lengths_

◆ ds_g_n_k_wos_strides_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

std::array<std::array<index_t, NDimSpatial + 3>, NumDTensor> ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::ds_g_n_k_wos_strides_

◆ ds_grid_desc_m_n_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

EGridDesc_M_N ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::ds_grid_desc_m_n_

◆ e_g_n_k_wos_lengths_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

std::array<index_t, NDimSpatial + 3> ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::e_g_n_k_wos_lengths_

◆ e_g_n_k_wos_strides_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

std::array<index_t, NDimSpatial + 3> ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::e_g_n_k_wos_strides_

◆ e_grid_desc_m_n_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

EGridDesc_M_N ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::e_grid_desc_m_n_

◆ input_left_pads_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

std::array<index_t, NDimSpatial> ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::input_left_pads_

◆ input_right_pads_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

std::array<index_t, NDimSpatial> ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::input_right_pads_

◆ p_a_grid_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

const ADataType* ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::p_a_grid_

◆ p_b_grid_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

const BDataType* ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::p_b_grid_

◆ p_ds_grid_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

GridwiseGemm64::DsGridPointer ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::p_ds_grid_

◆ p_e_grid_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

EDataType* ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::p_e_grid_

◆ p_rs_grid_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

GridwiseGemm64::RsGridPointer ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::p_rs_grid_

◆ qs_element_op_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

QsElementwiseOperation ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::qs_element_op_

◆ r_grid_desc_m_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

RGridDesc_M ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::r_grid_desc_m_

◆ rs_element_op_

template<index_t NDimSpatial, typename ALayout, typename BLayout, typename DELayout, typename RLayout, typename ADataType, typename BDataType, typename AccDataType, typename CShuffleDataType, typename DsDataType, typename EDataType, typename ReduceAccDataType, typename RsDataType, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename ThreadReduceOperations, typename RsGlobalMemoryDataOperation, ConvolutionForwardSpecialization ConvForwardSpecialization, GemmSpecialization GemmSpec, index_t NumGemmKPrefetchStage, index_t BlockSize, index_t MPerBlock, index_t NPerBlock, index_t KPerBlock, index_t AK1, index_t BK1, index_t MPerXDL, index_t NPerXDL, index_t MXdlPerWave, index_t NXdlPerWave, typename ABlockTransferThreadClusterLengths_AK0_M_AK1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, index_t ABlockTransferSrcVectorDim, index_t ABlockTransferSrcScalarPerVector, index_t ABlockTransferDstScalarPerVector_AK1, index_t ABlockLdsExtraM, typename BBlockTransferThreadClusterLengths_BK0_N_BK1, typename BBlockTransferThreadClusterArrangeOrder, typename BBlockTransferSrcAccessOrder, index_t BBlockTransferSrcVectorDim, index_t BBlockTransferSrcScalarPerVector, index_t BBlockTransferDstScalarPerVector_BK1, index_t BBlockLdsExtraN, index_t CShuffleMXdlPerWavePerShuffle, index_t CShuffleNXdlPerWavePerShuffle, typename CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, index_t CDEBlockTransferScalarPerVector_NPerBlock, index_t RThreadTransferDstScalarPerVector_MPerBlock, LoopScheduler LoopSched = make_default_loop_scheduler()>

RsElementwiseOperation ck::tensor_operation::device::DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle< NDimSpatial, ALayout, BLayout, DELayout, RLayout, ADataType, BDataType, AccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, ConvForwardSpecialization, GemmSpec, NumGemmKPrefetchStage, BlockSize, MPerBlock, NPerBlock, KPerBlock, AK1, BK1, MPerXDL, NPerXDL, MXdlPerWave, NXdlPerWave, ABlockTransferThreadClusterLengths_AK0_M_AK1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_AK1, ABlockLdsExtraM, BBlockTransferThreadClusterLengths_BK0_N_BK1, BBlockTransferThreadClusterArrangeOrder, BBlockTransferSrcAccessOrder, BBlockTransferSrcVectorDim, BBlockTransferSrcScalarPerVector, BBlockTransferDstScalarPerVector_BK1, BBlockLdsExtraN, CShuffleMXdlPerWavePerShuffle, CShuffleNXdlPerWavePerShuffle, CDRThreadTransferClusterLengths_MPerBlock_NPerBlock, CDEBlockTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument::rs_element_op_

---

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

• device_grouped_conv_fwd_multiple_d_multiple_r_xdl_cshuffle.hpp