RawArg Struct Reference

RawArg Struct Reference#

Composable Kernel: ck::tensor_operation::device::DeviceGroupedQueryAttentionForward_Wmma< NumDimG, NumDimM, NumDimL, NumDimK, NumDimN, ADataType, B0DataType, B1DataType, CDataType, Acc0BiasDataType, Acc0DataType, Acc1BiasDataType, Acc1DataType, CShuffleDataType, AElementwiseOperation, B0ElementwiseOperation, AccElementwiseOperation, B1ElementwiseOperation, CElementwiseOperation, GemmSpec, ASpec, B0Spec, B1Spec, CSpec, NumPrefetch, QueryGroupNumber, BlockSize, MPerBlock, LPerBlock, KPerBlock, AK1, BK1, NPerBlock, LTilePerBlock, L1, MPerWmma, LPerWmma, NPerWmma, MRepeat, LRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_K1, ABlockLdsAddExtraM, B0BlockTransferThreadClusterLengths_K0_L_K1, B0BlockTransferThreadClusterArrangeOrder, B0BlockTransferSrcAccessOrder, B0BlockTransferSrcVectorDim, B0BlockTransferSrcScalarPerVector, B0BlockTransferDstScalarPerVector_K1, B0BlockLdsAddExtraL, B1BlockTransferThreadClusterLengths_L0_N_L1, B1BlockTransferThreadClusterArrangeOrder, B1BlockTransferSrcAccessOrder, B1BlockTransferSrcVectorDim, B1BlockTransferSrcScalarPerVector, B1BlockTransferDstScalarPerVector_L1, B1BlockLdsAddExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpec, LoopSched, PipelineVer >::RawArg Struct Reference
Public Member Functions | Public Attributes | List of all members
ck::tensor_operation::device::DeviceGroupedQueryAttentionForward_Wmma< NumDimG, NumDimM, NumDimL, NumDimK, NumDimN, ADataType, B0DataType, B1DataType, CDataType, Acc0BiasDataType, Acc0DataType, Acc1BiasDataType, Acc1DataType, CShuffleDataType, AElementwiseOperation, B0ElementwiseOperation, AccElementwiseOperation, B1ElementwiseOperation, CElementwiseOperation, GemmSpec, ASpec, B0Spec, B1Spec, CSpec, NumPrefetch, QueryGroupNumber, BlockSize, MPerBlock, LPerBlock, KPerBlock, AK1, BK1, NPerBlock, LTilePerBlock, L1, MPerWmma, LPerWmma, NPerWmma, MRepeat, LRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_K1, ABlockLdsAddExtraM, B0BlockTransferThreadClusterLengths_K0_L_K1, B0BlockTransferThreadClusterArrangeOrder, B0BlockTransferSrcAccessOrder, B0BlockTransferSrcVectorDim, B0BlockTransferSrcScalarPerVector, B0BlockTransferDstScalarPerVector_K1, B0BlockLdsAddExtraL, B1BlockTransferThreadClusterLengths_L0_N_L1, B1BlockTransferThreadClusterArrangeOrder, B1BlockTransferSrcAccessOrder, B1BlockTransferSrcVectorDim, B1BlockTransferSrcScalarPerVector, B1BlockTransferDstScalarPerVector_L1, B1BlockLdsAddExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpec, LoopSched, PipelineVer >::RawArg Struct Reference

#include <device_grouped_query_attention_forward_wmma.hpp>

Inheritance diagram for ck::tensor_operation::device::DeviceGroupedQueryAttentionForward_Wmma< NumDimG, NumDimM, NumDimL, NumDimK, NumDimN, ADataType, B0DataType, B1DataType, CDataType, Acc0BiasDataType, Acc0DataType, Acc1BiasDataType, Acc1DataType, CShuffleDataType, AElementwiseOperation, B0ElementwiseOperation, AccElementwiseOperation, B1ElementwiseOperation, CElementwiseOperation, GemmSpec, ASpec, B0Spec, B1Spec, CSpec, NumPrefetch, QueryGroupNumber, BlockSize, MPerBlock, LPerBlock, KPerBlock, AK1, BK1, NPerBlock, LTilePerBlock, L1, MPerWmma, LPerWmma, NPerWmma, MRepeat, LRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_K1, ABlockLdsAddExtraM, B0BlockTransferThreadClusterLengths_K0_L_K1, B0BlockTransferThreadClusterArrangeOrder, B0BlockTransferSrcAccessOrder, B0BlockTransferSrcVectorDim, B0BlockTransferSrcScalarPerVector, B0BlockTransferDstScalarPerVector_K1, B0BlockLdsAddExtraL, B1BlockTransferThreadClusterLengths_L0_N_L1, B1BlockTransferThreadClusterArrangeOrder, B1BlockTransferSrcAccessOrder, B1BlockTransferSrcVectorDim, B1BlockTransferSrcScalarPerVector, B1BlockTransferDstScalarPerVector_L1, B1BlockLdsAddExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpec, LoopSched, PipelineVer >::RawArg:
ck::tensor_operation::device::BaseArgument

Public Member Functions

 RawArg (const ADataType *p_a_grid, const B0DataType *p_b0_grid, const B1DataType *p_b1_grid, CDataType *p_c_grid, index_t M, index_t N, index_t K, index_t O, index_t G0, index_t G1, float alpha, bool input_permute, bool output_permute)
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 B0DataType * p_b0_grid_
const B1DataType * p_b1_grid_
CDataType * p_c_grid_
index_t M_
index_t N_
index_t K_
index_t O_
index_t G0_
index_t G1_
float alpha_
bool input_permute_
bool output_permute_
Public Attributes inherited from ck::tensor_operation::device::BaseArgument
void * p_workspace_ = nullptr

Constructor & Destructor Documentation

◆ RawArg()

template<index_t NumDimG, index_t NumDimM, index_t NumDimL, index_t NumDimK, index_t NumDimN, typename ADataType, typename B0DataType, typename B1DataType, typename CDataType, typename Acc0BiasDataType, typename Acc0DataType, typename Acc1BiasDataType, typename Acc1DataType, typename CShuffleDataType, typename AElementwiseOperation, typename B0ElementwiseOperation, typename AccElementwiseOperation, typename B1ElementwiseOperation, typename CElementwiseOperation, GemmSpecialization GemmSpec, TensorSpecialization ASpec, TensorSpecialization B0Spec, TensorSpecialization B1Spec, TensorSpecialization CSpec, ck::index_t NumPrefetch, ck::index_t QueryGroupNumber, ck::index_t BlockSize, ck::index_t MPerBlock, ck::index_t LPerBlock, ck::index_t KPerBlock, ck::index_t AK1, ck::index_t BK1, ck::index_t NPerBlock, ck::index_t LTilePerBlock, ck::index_t L1, ck::index_t MPerWmma, ck::index_t LPerWmma, ck::index_t NPerWmma, ck::index_t MRepeat, ck::index_t LRepeat, ck::index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, ck::index_t ABlockTransferSrcVectorDim, ck::index_t ABlockTransferSrcScalarPerVector, ck::index_t ABlockTransferDstScalarPerVector_K1, bool ABlockLdsAddExtraM, typename B0BlockTransferThreadClusterLengths_K0_L_K1, typename B0BlockTransferThreadClusterArrangeOrder, typename B0BlockTransferSrcAccessOrder, ck::index_t B0BlockTransferSrcVectorDim, ck::index_t B0BlockTransferSrcScalarPerVector, ck::index_t B0BlockTransferDstScalarPerVector_K1, bool B0BlockLdsAddExtraL, typename B1BlockTransferThreadClusterLengths_L0_N_L1, typename B1BlockTransferThreadClusterArrangeOrder, typename B1BlockTransferSrcAccessOrder, ck::index_t B1BlockTransferSrcVectorDim, ck::index_t B1BlockTransferSrcScalarPerVector, ck::index_t B1BlockTransferDstScalarPerVector_L1, bool B1BlockLdsAddExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, typename CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, index_t CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpecialization MaskingSpec, ck::LoopScheduler LoopSched = make_default_loop_scheduler(), ck::PipelineVersion PipelineVer = ck::PipelineVersion::v1>
ck::tensor_operation::device::DeviceGroupedQueryAttentionForward_Wmma< NumDimG, NumDimM, NumDimL, NumDimK, NumDimN, ADataType, B0DataType, B1DataType, CDataType, Acc0BiasDataType, Acc0DataType, Acc1BiasDataType, Acc1DataType, CShuffleDataType, AElementwiseOperation, B0ElementwiseOperation, AccElementwiseOperation, B1ElementwiseOperation, CElementwiseOperation, GemmSpec, ASpec, B0Spec, B1Spec, CSpec, NumPrefetch, QueryGroupNumber, BlockSize, MPerBlock, LPerBlock, KPerBlock, AK1, BK1, NPerBlock, LTilePerBlock, L1, MPerWmma, LPerWmma, NPerWmma, MRepeat, LRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_K1, ABlockLdsAddExtraM, B0BlockTransferThreadClusterLengths_K0_L_K1, B0BlockTransferThreadClusterArrangeOrder, B0BlockTransferSrcAccessOrder, B0BlockTransferSrcVectorDim, B0BlockTransferSrcScalarPerVector, B0BlockTransferDstScalarPerVector_K1, B0BlockLdsAddExtraL, B1BlockTransferThreadClusterLengths_L0_N_L1, B1BlockTransferThreadClusterArrangeOrder, B1BlockTransferSrcAccessOrder, B1BlockTransferSrcVectorDim, B1BlockTransferSrcScalarPerVector, B1BlockTransferDstScalarPerVector_L1, B1BlockLdsAddExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpec, LoopSched, PipelineVer >::RawArg::RawArg ( const ADataType * p_a_grid,
const B0DataType * p_b0_grid,
const B1DataType * p_b1_grid,
CDataType * p_c_grid,
index_t M,
index_t N,
index_t K,
index_t O,
index_t G0,
index_t G1,
float alpha,
bool input_permute,
bool output_permute )
inline

Member Data Documentation

◆ alpha_

template<index_t NumDimG, index_t NumDimM, index_t NumDimL, index_t NumDimK, index_t NumDimN, typename ADataType, typename B0DataType, typename B1DataType, typename CDataType, typename Acc0BiasDataType, typename Acc0DataType, typename Acc1BiasDataType, typename Acc1DataType, typename CShuffleDataType, typename AElementwiseOperation, typename B0ElementwiseOperation, typename AccElementwiseOperation, typename B1ElementwiseOperation, typename CElementwiseOperation, GemmSpecialization GemmSpec, TensorSpecialization ASpec, TensorSpecialization B0Spec, TensorSpecialization B1Spec, TensorSpecialization CSpec, ck::index_t NumPrefetch, ck::index_t QueryGroupNumber, ck::index_t BlockSize, ck::index_t MPerBlock, ck::index_t LPerBlock, ck::index_t KPerBlock, ck::index_t AK1, ck::index_t BK1, ck::index_t NPerBlock, ck::index_t LTilePerBlock, ck::index_t L1, ck::index_t MPerWmma, ck::index_t LPerWmma, ck::index_t NPerWmma, ck::index_t MRepeat, ck::index_t LRepeat, ck::index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, ck::index_t ABlockTransferSrcVectorDim, ck::index_t ABlockTransferSrcScalarPerVector, ck::index_t ABlockTransferDstScalarPerVector_K1, bool ABlockLdsAddExtraM, typename B0BlockTransferThreadClusterLengths_K0_L_K1, typename B0BlockTransferThreadClusterArrangeOrder, typename B0BlockTransferSrcAccessOrder, ck::index_t B0BlockTransferSrcVectorDim, ck::index_t B0BlockTransferSrcScalarPerVector, ck::index_t B0BlockTransferDstScalarPerVector_K1, bool B0BlockLdsAddExtraL, typename B1BlockTransferThreadClusterLengths_L0_N_L1, typename B1BlockTransferThreadClusterArrangeOrder, typename B1BlockTransferSrcAccessOrder, ck::index_t B1BlockTransferSrcVectorDim, ck::index_t B1BlockTransferSrcScalarPerVector, ck::index_t B1BlockTransferDstScalarPerVector_L1, bool B1BlockLdsAddExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, typename CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, index_t CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpecialization MaskingSpec, ck::LoopScheduler LoopSched = make_default_loop_scheduler(), ck::PipelineVersion PipelineVer = ck::PipelineVersion::v1>
float ck::tensor_operation::device::DeviceGroupedQueryAttentionForward_Wmma< NumDimG, NumDimM, NumDimL, NumDimK, NumDimN, ADataType, B0DataType, B1DataType, CDataType, Acc0BiasDataType, Acc0DataType, Acc1BiasDataType, Acc1DataType, CShuffleDataType, AElementwiseOperation, B0ElementwiseOperation, AccElementwiseOperation, B1ElementwiseOperation, CElementwiseOperation, GemmSpec, ASpec, B0Spec, B1Spec, CSpec, NumPrefetch, QueryGroupNumber, BlockSize, MPerBlock, LPerBlock, KPerBlock, AK1, BK1, NPerBlock, LTilePerBlock, L1, MPerWmma, LPerWmma, NPerWmma, MRepeat, LRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_K1, ABlockLdsAddExtraM, B0BlockTransferThreadClusterLengths_K0_L_K1, B0BlockTransferThreadClusterArrangeOrder, B0BlockTransferSrcAccessOrder, B0BlockTransferSrcVectorDim, B0BlockTransferSrcScalarPerVector, B0BlockTransferDstScalarPerVector_K1, B0BlockLdsAddExtraL, B1BlockTransferThreadClusterLengths_L0_N_L1, B1BlockTransferThreadClusterArrangeOrder, B1BlockTransferSrcAccessOrder, B1BlockTransferSrcVectorDim, B1BlockTransferSrcScalarPerVector, B1BlockTransferDstScalarPerVector_L1, B1BlockLdsAddExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpec, LoopSched, PipelineVer >::RawArg::alpha_

◆ G0_

template<index_t NumDimG, index_t NumDimM, index_t NumDimL, index_t NumDimK, index_t NumDimN, typename ADataType, typename B0DataType, typename B1DataType, typename CDataType, typename Acc0BiasDataType, typename Acc0DataType, typename Acc1BiasDataType, typename Acc1DataType, typename CShuffleDataType, typename AElementwiseOperation, typename B0ElementwiseOperation, typename AccElementwiseOperation, typename B1ElementwiseOperation, typename CElementwiseOperation, GemmSpecialization GemmSpec, TensorSpecialization ASpec, TensorSpecialization B0Spec, TensorSpecialization B1Spec, TensorSpecialization CSpec, ck::index_t NumPrefetch, ck::index_t QueryGroupNumber, ck::index_t BlockSize, ck::index_t MPerBlock, ck::index_t LPerBlock, ck::index_t KPerBlock, ck::index_t AK1, ck::index_t BK1, ck::index_t NPerBlock, ck::index_t LTilePerBlock, ck::index_t L1, ck::index_t MPerWmma, ck::index_t LPerWmma, ck::index_t NPerWmma, ck::index_t MRepeat, ck::index_t LRepeat, ck::index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, ck::index_t ABlockTransferSrcVectorDim, ck::index_t ABlockTransferSrcScalarPerVector, ck::index_t ABlockTransferDstScalarPerVector_K1, bool ABlockLdsAddExtraM, typename B0BlockTransferThreadClusterLengths_K0_L_K1, typename B0BlockTransferThreadClusterArrangeOrder, typename B0BlockTransferSrcAccessOrder, ck::index_t B0BlockTransferSrcVectorDim, ck::index_t B0BlockTransferSrcScalarPerVector, ck::index_t B0BlockTransferDstScalarPerVector_K1, bool B0BlockLdsAddExtraL, typename B1BlockTransferThreadClusterLengths_L0_N_L1, typename B1BlockTransferThreadClusterArrangeOrder, typename B1BlockTransferSrcAccessOrder, ck::index_t B1BlockTransferSrcVectorDim, ck::index_t B1BlockTransferSrcScalarPerVector, ck::index_t B1BlockTransferDstScalarPerVector_L1, bool B1BlockLdsAddExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, typename CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, index_t CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpecialization MaskingSpec, ck::LoopScheduler LoopSched = make_default_loop_scheduler(), ck::PipelineVersion PipelineVer = ck::PipelineVersion::v1>
index_t ck::tensor_operation::device::DeviceGroupedQueryAttentionForward_Wmma< NumDimG, NumDimM, NumDimL, NumDimK, NumDimN, ADataType, B0DataType, B1DataType, CDataType, Acc0BiasDataType, Acc0DataType, Acc1BiasDataType, Acc1DataType, CShuffleDataType, AElementwiseOperation, B0ElementwiseOperation, AccElementwiseOperation, B1ElementwiseOperation, CElementwiseOperation, GemmSpec, ASpec, B0Spec, B1Spec, CSpec, NumPrefetch, QueryGroupNumber, BlockSize, MPerBlock, LPerBlock, KPerBlock, AK1, BK1, NPerBlock, LTilePerBlock, L1, MPerWmma, LPerWmma, NPerWmma, MRepeat, LRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_K1, ABlockLdsAddExtraM, B0BlockTransferThreadClusterLengths_K0_L_K1, B0BlockTransferThreadClusterArrangeOrder, B0BlockTransferSrcAccessOrder, B0BlockTransferSrcVectorDim, B0BlockTransferSrcScalarPerVector, B0BlockTransferDstScalarPerVector_K1, B0BlockLdsAddExtraL, B1BlockTransferThreadClusterLengths_L0_N_L1, B1BlockTransferThreadClusterArrangeOrder, B1BlockTransferSrcAccessOrder, B1BlockTransferSrcVectorDim, B1BlockTransferSrcScalarPerVector, B1BlockTransferDstScalarPerVector_L1, B1BlockLdsAddExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpec, LoopSched, PipelineVer >::RawArg::G0_

◆ G1_

template<index_t NumDimG, index_t NumDimM, index_t NumDimL, index_t NumDimK, index_t NumDimN, typename ADataType, typename B0DataType, typename B1DataType, typename CDataType, typename Acc0BiasDataType, typename Acc0DataType, typename Acc1BiasDataType, typename Acc1DataType, typename CShuffleDataType, typename AElementwiseOperation, typename B0ElementwiseOperation, typename AccElementwiseOperation, typename B1ElementwiseOperation, typename CElementwiseOperation, GemmSpecialization GemmSpec, TensorSpecialization ASpec, TensorSpecialization B0Spec, TensorSpecialization B1Spec, TensorSpecialization CSpec, ck::index_t NumPrefetch, ck::index_t QueryGroupNumber, ck::index_t BlockSize, ck::index_t MPerBlock, ck::index_t LPerBlock, ck::index_t KPerBlock, ck::index_t AK1, ck::index_t BK1, ck::index_t NPerBlock, ck::index_t LTilePerBlock, ck::index_t L1, ck::index_t MPerWmma, ck::index_t LPerWmma, ck::index_t NPerWmma, ck::index_t MRepeat, ck::index_t LRepeat, ck::index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, ck::index_t ABlockTransferSrcVectorDim, ck::index_t ABlockTransferSrcScalarPerVector, ck::index_t ABlockTransferDstScalarPerVector_K1, bool ABlockLdsAddExtraM, typename B0BlockTransferThreadClusterLengths_K0_L_K1, typename B0BlockTransferThreadClusterArrangeOrder, typename B0BlockTransferSrcAccessOrder, ck::index_t B0BlockTransferSrcVectorDim, ck::index_t B0BlockTransferSrcScalarPerVector, ck::index_t B0BlockTransferDstScalarPerVector_K1, bool B0BlockLdsAddExtraL, typename B1BlockTransferThreadClusterLengths_L0_N_L1, typename B1BlockTransferThreadClusterArrangeOrder, typename B1BlockTransferSrcAccessOrder, ck::index_t B1BlockTransferSrcVectorDim, ck::index_t B1BlockTransferSrcScalarPerVector, ck::index_t B1BlockTransferDstScalarPerVector_L1, bool B1BlockLdsAddExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, typename CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, index_t CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpecialization MaskingSpec, ck::LoopScheduler LoopSched = make_default_loop_scheduler(), ck::PipelineVersion PipelineVer = ck::PipelineVersion::v1>
index_t ck::tensor_operation::device::DeviceGroupedQueryAttentionForward_Wmma< NumDimG, NumDimM, NumDimL, NumDimK, NumDimN, ADataType, B0DataType, B1DataType, CDataType, Acc0BiasDataType, Acc0DataType, Acc1BiasDataType, Acc1DataType, CShuffleDataType, AElementwiseOperation, B0ElementwiseOperation, AccElementwiseOperation, B1ElementwiseOperation, CElementwiseOperation, GemmSpec, ASpec, B0Spec, B1Spec, CSpec, NumPrefetch, QueryGroupNumber, BlockSize, MPerBlock, LPerBlock, KPerBlock, AK1, BK1, NPerBlock, LTilePerBlock, L1, MPerWmma, LPerWmma, NPerWmma, MRepeat, LRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_K1, ABlockLdsAddExtraM, B0BlockTransferThreadClusterLengths_K0_L_K1, B0BlockTransferThreadClusterArrangeOrder, B0BlockTransferSrcAccessOrder, B0BlockTransferSrcVectorDim, B0BlockTransferSrcScalarPerVector, B0BlockTransferDstScalarPerVector_K1, B0BlockLdsAddExtraL, B1BlockTransferThreadClusterLengths_L0_N_L1, B1BlockTransferThreadClusterArrangeOrder, B1BlockTransferSrcAccessOrder, B1BlockTransferSrcVectorDim, B1BlockTransferSrcScalarPerVector, B1BlockTransferDstScalarPerVector_L1, B1BlockLdsAddExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpec, LoopSched, PipelineVer >::RawArg::G1_

◆ input_permute_

template<index_t NumDimG, index_t NumDimM, index_t NumDimL, index_t NumDimK, index_t NumDimN, typename ADataType, typename B0DataType, typename B1DataType, typename CDataType, typename Acc0BiasDataType, typename Acc0DataType, typename Acc1BiasDataType, typename Acc1DataType, typename CShuffleDataType, typename AElementwiseOperation, typename B0ElementwiseOperation, typename AccElementwiseOperation, typename B1ElementwiseOperation, typename CElementwiseOperation, GemmSpecialization GemmSpec, TensorSpecialization ASpec, TensorSpecialization B0Spec, TensorSpecialization B1Spec, TensorSpecialization CSpec, ck::index_t NumPrefetch, ck::index_t QueryGroupNumber, ck::index_t BlockSize, ck::index_t MPerBlock, ck::index_t LPerBlock, ck::index_t KPerBlock, ck::index_t AK1, ck::index_t BK1, ck::index_t NPerBlock, ck::index_t LTilePerBlock, ck::index_t L1, ck::index_t MPerWmma, ck::index_t LPerWmma, ck::index_t NPerWmma, ck::index_t MRepeat, ck::index_t LRepeat, ck::index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, ck::index_t ABlockTransferSrcVectorDim, ck::index_t ABlockTransferSrcScalarPerVector, ck::index_t ABlockTransferDstScalarPerVector_K1, bool ABlockLdsAddExtraM, typename B0BlockTransferThreadClusterLengths_K0_L_K1, typename B0BlockTransferThreadClusterArrangeOrder, typename B0BlockTransferSrcAccessOrder, ck::index_t B0BlockTransferSrcVectorDim, ck::index_t B0BlockTransferSrcScalarPerVector, ck::index_t B0BlockTransferDstScalarPerVector_K1, bool B0BlockLdsAddExtraL, typename B1BlockTransferThreadClusterLengths_L0_N_L1, typename B1BlockTransferThreadClusterArrangeOrder, typename B1BlockTransferSrcAccessOrder, ck::index_t B1BlockTransferSrcVectorDim, ck::index_t B1BlockTransferSrcScalarPerVector, ck::index_t B1BlockTransferDstScalarPerVector_L1, bool B1BlockLdsAddExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, typename CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, index_t CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpecialization MaskingSpec, ck::LoopScheduler LoopSched = make_default_loop_scheduler(), ck::PipelineVersion PipelineVer = ck::PipelineVersion::v1>
bool ck::tensor_operation::device::DeviceGroupedQueryAttentionForward_Wmma< NumDimG, NumDimM, NumDimL, NumDimK, NumDimN, ADataType, B0DataType, B1DataType, CDataType, Acc0BiasDataType, Acc0DataType, Acc1BiasDataType, Acc1DataType, CShuffleDataType, AElementwiseOperation, B0ElementwiseOperation, AccElementwiseOperation, B1ElementwiseOperation, CElementwiseOperation, GemmSpec, ASpec, B0Spec, B1Spec, CSpec, NumPrefetch, QueryGroupNumber, BlockSize, MPerBlock, LPerBlock, KPerBlock, AK1, BK1, NPerBlock, LTilePerBlock, L1, MPerWmma, LPerWmma, NPerWmma, MRepeat, LRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_K1, ABlockLdsAddExtraM, B0BlockTransferThreadClusterLengths_K0_L_K1, B0BlockTransferThreadClusterArrangeOrder, B0BlockTransferSrcAccessOrder, B0BlockTransferSrcVectorDim, B0BlockTransferSrcScalarPerVector, B0BlockTransferDstScalarPerVector_K1, B0BlockLdsAddExtraL, B1BlockTransferThreadClusterLengths_L0_N_L1, B1BlockTransferThreadClusterArrangeOrder, B1BlockTransferSrcAccessOrder, B1BlockTransferSrcVectorDim, B1BlockTransferSrcScalarPerVector, B1BlockTransferDstScalarPerVector_L1, B1BlockLdsAddExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpec, LoopSched, PipelineVer >::RawArg::input_permute_

◆ K_

template<index_t NumDimG, index_t NumDimM, index_t NumDimL, index_t NumDimK, index_t NumDimN, typename ADataType, typename B0DataType, typename B1DataType, typename CDataType, typename Acc0BiasDataType, typename Acc0DataType, typename Acc1BiasDataType, typename Acc1DataType, typename CShuffleDataType, typename AElementwiseOperation, typename B0ElementwiseOperation, typename AccElementwiseOperation, typename B1ElementwiseOperation, typename CElementwiseOperation, GemmSpecialization GemmSpec, TensorSpecialization ASpec, TensorSpecialization B0Spec, TensorSpecialization B1Spec, TensorSpecialization CSpec, ck::index_t NumPrefetch, ck::index_t QueryGroupNumber, ck::index_t BlockSize, ck::index_t MPerBlock, ck::index_t LPerBlock, ck::index_t KPerBlock, ck::index_t AK1, ck::index_t BK1, ck::index_t NPerBlock, ck::index_t LTilePerBlock, ck::index_t L1, ck::index_t MPerWmma, ck::index_t LPerWmma, ck::index_t NPerWmma, ck::index_t MRepeat, ck::index_t LRepeat, ck::index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, ck::index_t ABlockTransferSrcVectorDim, ck::index_t ABlockTransferSrcScalarPerVector, ck::index_t ABlockTransferDstScalarPerVector_K1, bool ABlockLdsAddExtraM, typename B0BlockTransferThreadClusterLengths_K0_L_K1, typename B0BlockTransferThreadClusterArrangeOrder, typename B0BlockTransferSrcAccessOrder, ck::index_t B0BlockTransferSrcVectorDim, ck::index_t B0BlockTransferSrcScalarPerVector, ck::index_t B0BlockTransferDstScalarPerVector_K1, bool B0BlockLdsAddExtraL, typename B1BlockTransferThreadClusterLengths_L0_N_L1, typename B1BlockTransferThreadClusterArrangeOrder, typename B1BlockTransferSrcAccessOrder, ck::index_t B1BlockTransferSrcVectorDim, ck::index_t B1BlockTransferSrcScalarPerVector, ck::index_t B1BlockTransferDstScalarPerVector_L1, bool B1BlockLdsAddExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, typename CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, index_t CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpecialization MaskingSpec, ck::LoopScheduler LoopSched = make_default_loop_scheduler(), ck::PipelineVersion PipelineVer = ck::PipelineVersion::v1>
index_t ck::tensor_operation::device::DeviceGroupedQueryAttentionForward_Wmma< NumDimG, NumDimM, NumDimL, NumDimK, NumDimN, ADataType, B0DataType, B1DataType, CDataType, Acc0BiasDataType, Acc0DataType, Acc1BiasDataType, Acc1DataType, CShuffleDataType, AElementwiseOperation, B0ElementwiseOperation, AccElementwiseOperation, B1ElementwiseOperation, CElementwiseOperation, GemmSpec, ASpec, B0Spec, B1Spec, CSpec, NumPrefetch, QueryGroupNumber, BlockSize, MPerBlock, LPerBlock, KPerBlock, AK1, BK1, NPerBlock, LTilePerBlock, L1, MPerWmma, LPerWmma, NPerWmma, MRepeat, LRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_K1, ABlockLdsAddExtraM, B0BlockTransferThreadClusterLengths_K0_L_K1, B0BlockTransferThreadClusterArrangeOrder, B0BlockTransferSrcAccessOrder, B0BlockTransferSrcVectorDim, B0BlockTransferSrcScalarPerVector, B0BlockTransferDstScalarPerVector_K1, B0BlockLdsAddExtraL, B1BlockTransferThreadClusterLengths_L0_N_L1, B1BlockTransferThreadClusterArrangeOrder, B1BlockTransferSrcAccessOrder, B1BlockTransferSrcVectorDim, B1BlockTransferSrcScalarPerVector, B1BlockTransferDstScalarPerVector_L1, B1BlockLdsAddExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpec, LoopSched, PipelineVer >::RawArg::K_

◆ M_

template<index_t NumDimG, index_t NumDimM, index_t NumDimL, index_t NumDimK, index_t NumDimN, typename ADataType, typename B0DataType, typename B1DataType, typename CDataType, typename Acc0BiasDataType, typename Acc0DataType, typename Acc1BiasDataType, typename Acc1DataType, typename CShuffleDataType, typename AElementwiseOperation, typename B0ElementwiseOperation, typename AccElementwiseOperation, typename B1ElementwiseOperation, typename CElementwiseOperation, GemmSpecialization GemmSpec, TensorSpecialization ASpec, TensorSpecialization B0Spec, TensorSpecialization B1Spec, TensorSpecialization CSpec, ck::index_t NumPrefetch, ck::index_t QueryGroupNumber, ck::index_t BlockSize, ck::index_t MPerBlock, ck::index_t LPerBlock, ck::index_t KPerBlock, ck::index_t AK1, ck::index_t BK1, ck::index_t NPerBlock, ck::index_t LTilePerBlock, ck::index_t L1, ck::index_t MPerWmma, ck::index_t LPerWmma, ck::index_t NPerWmma, ck::index_t MRepeat, ck::index_t LRepeat, ck::index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, ck::index_t ABlockTransferSrcVectorDim, ck::index_t ABlockTransferSrcScalarPerVector, ck::index_t ABlockTransferDstScalarPerVector_K1, bool ABlockLdsAddExtraM, typename B0BlockTransferThreadClusterLengths_K0_L_K1, typename B0BlockTransferThreadClusterArrangeOrder, typename B0BlockTransferSrcAccessOrder, ck::index_t B0BlockTransferSrcVectorDim, ck::index_t B0BlockTransferSrcScalarPerVector, ck::index_t B0BlockTransferDstScalarPerVector_K1, bool B0BlockLdsAddExtraL, typename B1BlockTransferThreadClusterLengths_L0_N_L1, typename B1BlockTransferThreadClusterArrangeOrder, typename B1BlockTransferSrcAccessOrder, ck::index_t B1BlockTransferSrcVectorDim, ck::index_t B1BlockTransferSrcScalarPerVector, ck::index_t B1BlockTransferDstScalarPerVector_L1, bool B1BlockLdsAddExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, typename CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, index_t CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpecialization MaskingSpec, ck::LoopScheduler LoopSched = make_default_loop_scheduler(), ck::PipelineVersion PipelineVer = ck::PipelineVersion::v1>
index_t ck::tensor_operation::device::DeviceGroupedQueryAttentionForward_Wmma< NumDimG, NumDimM, NumDimL, NumDimK, NumDimN, ADataType, B0DataType, B1DataType, CDataType, Acc0BiasDataType, Acc0DataType, Acc1BiasDataType, Acc1DataType, CShuffleDataType, AElementwiseOperation, B0ElementwiseOperation, AccElementwiseOperation, B1ElementwiseOperation, CElementwiseOperation, GemmSpec, ASpec, B0Spec, B1Spec, CSpec, NumPrefetch, QueryGroupNumber, BlockSize, MPerBlock, LPerBlock, KPerBlock, AK1, BK1, NPerBlock, LTilePerBlock, L1, MPerWmma, LPerWmma, NPerWmma, MRepeat, LRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_K1, ABlockLdsAddExtraM, B0BlockTransferThreadClusterLengths_K0_L_K1, B0BlockTransferThreadClusterArrangeOrder, B0BlockTransferSrcAccessOrder, B0BlockTransferSrcVectorDim, B0BlockTransferSrcScalarPerVector, B0BlockTransferDstScalarPerVector_K1, B0BlockLdsAddExtraL, B1BlockTransferThreadClusterLengths_L0_N_L1, B1BlockTransferThreadClusterArrangeOrder, B1BlockTransferSrcAccessOrder, B1BlockTransferSrcVectorDim, B1BlockTransferSrcScalarPerVector, B1BlockTransferDstScalarPerVector_L1, B1BlockLdsAddExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpec, LoopSched, PipelineVer >::RawArg::M_

◆ N_

template<index_t NumDimG, index_t NumDimM, index_t NumDimL, index_t NumDimK, index_t NumDimN, typename ADataType, typename B0DataType, typename B1DataType, typename CDataType, typename Acc0BiasDataType, typename Acc0DataType, typename Acc1BiasDataType, typename Acc1DataType, typename CShuffleDataType, typename AElementwiseOperation, typename B0ElementwiseOperation, typename AccElementwiseOperation, typename B1ElementwiseOperation, typename CElementwiseOperation, GemmSpecialization GemmSpec, TensorSpecialization ASpec, TensorSpecialization B0Spec, TensorSpecialization B1Spec, TensorSpecialization CSpec, ck::index_t NumPrefetch, ck::index_t QueryGroupNumber, ck::index_t BlockSize, ck::index_t MPerBlock, ck::index_t LPerBlock, ck::index_t KPerBlock, ck::index_t AK1, ck::index_t BK1, ck::index_t NPerBlock, ck::index_t LTilePerBlock, ck::index_t L1, ck::index_t MPerWmma, ck::index_t LPerWmma, ck::index_t NPerWmma, ck::index_t MRepeat, ck::index_t LRepeat, ck::index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, ck::index_t ABlockTransferSrcVectorDim, ck::index_t ABlockTransferSrcScalarPerVector, ck::index_t ABlockTransferDstScalarPerVector_K1, bool ABlockLdsAddExtraM, typename B0BlockTransferThreadClusterLengths_K0_L_K1, typename B0BlockTransferThreadClusterArrangeOrder, typename B0BlockTransferSrcAccessOrder, ck::index_t B0BlockTransferSrcVectorDim, ck::index_t B0BlockTransferSrcScalarPerVector, ck::index_t B0BlockTransferDstScalarPerVector_K1, bool B0BlockLdsAddExtraL, typename B1BlockTransferThreadClusterLengths_L0_N_L1, typename B1BlockTransferThreadClusterArrangeOrder, typename B1BlockTransferSrcAccessOrder, ck::index_t B1BlockTransferSrcVectorDim, ck::index_t B1BlockTransferSrcScalarPerVector, ck::index_t B1BlockTransferDstScalarPerVector_L1, bool B1BlockLdsAddExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, typename CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, index_t CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpecialization MaskingSpec, ck::LoopScheduler LoopSched = make_default_loop_scheduler(), ck::PipelineVersion PipelineVer = ck::PipelineVersion::v1>
index_t ck::tensor_operation::device::DeviceGroupedQueryAttentionForward_Wmma< NumDimG, NumDimM, NumDimL, NumDimK, NumDimN, ADataType, B0DataType, B1DataType, CDataType, Acc0BiasDataType, Acc0DataType, Acc1BiasDataType, Acc1DataType, CShuffleDataType, AElementwiseOperation, B0ElementwiseOperation, AccElementwiseOperation, B1ElementwiseOperation, CElementwiseOperation, GemmSpec, ASpec, B0Spec, B1Spec, CSpec, NumPrefetch, QueryGroupNumber, BlockSize, MPerBlock, LPerBlock, KPerBlock, AK1, BK1, NPerBlock, LTilePerBlock, L1, MPerWmma, LPerWmma, NPerWmma, MRepeat, LRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_K1, ABlockLdsAddExtraM, B0BlockTransferThreadClusterLengths_K0_L_K1, B0BlockTransferThreadClusterArrangeOrder, B0BlockTransferSrcAccessOrder, B0BlockTransferSrcVectorDim, B0BlockTransferSrcScalarPerVector, B0BlockTransferDstScalarPerVector_K1, B0BlockLdsAddExtraL, B1BlockTransferThreadClusterLengths_L0_N_L1, B1BlockTransferThreadClusterArrangeOrder, B1BlockTransferSrcAccessOrder, B1BlockTransferSrcVectorDim, B1BlockTransferSrcScalarPerVector, B1BlockTransferDstScalarPerVector_L1, B1BlockLdsAddExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpec, LoopSched, PipelineVer >::RawArg::N_

◆ O_

template<index_t NumDimG, index_t NumDimM, index_t NumDimL, index_t NumDimK, index_t NumDimN, typename ADataType, typename B0DataType, typename B1DataType, typename CDataType, typename Acc0BiasDataType, typename Acc0DataType, typename Acc1BiasDataType, typename Acc1DataType, typename CShuffleDataType, typename AElementwiseOperation, typename B0ElementwiseOperation, typename AccElementwiseOperation, typename B1ElementwiseOperation, typename CElementwiseOperation, GemmSpecialization GemmSpec, TensorSpecialization ASpec, TensorSpecialization B0Spec, TensorSpecialization B1Spec, TensorSpecialization CSpec, ck::index_t NumPrefetch, ck::index_t QueryGroupNumber, ck::index_t BlockSize, ck::index_t MPerBlock, ck::index_t LPerBlock, ck::index_t KPerBlock, ck::index_t AK1, ck::index_t BK1, ck::index_t NPerBlock, ck::index_t LTilePerBlock, ck::index_t L1, ck::index_t MPerWmma, ck::index_t LPerWmma, ck::index_t NPerWmma, ck::index_t MRepeat, ck::index_t LRepeat, ck::index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, ck::index_t ABlockTransferSrcVectorDim, ck::index_t ABlockTransferSrcScalarPerVector, ck::index_t ABlockTransferDstScalarPerVector_K1, bool ABlockLdsAddExtraM, typename B0BlockTransferThreadClusterLengths_K0_L_K1, typename B0BlockTransferThreadClusterArrangeOrder, typename B0BlockTransferSrcAccessOrder, ck::index_t B0BlockTransferSrcVectorDim, ck::index_t B0BlockTransferSrcScalarPerVector, ck::index_t B0BlockTransferDstScalarPerVector_K1, bool B0BlockLdsAddExtraL, typename B1BlockTransferThreadClusterLengths_L0_N_L1, typename B1BlockTransferThreadClusterArrangeOrder, typename B1BlockTransferSrcAccessOrder, ck::index_t B1BlockTransferSrcVectorDim, ck::index_t B1BlockTransferSrcScalarPerVector, ck::index_t B1BlockTransferDstScalarPerVector_L1, bool B1BlockLdsAddExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, typename CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, index_t CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpecialization MaskingSpec, ck::LoopScheduler LoopSched = make_default_loop_scheduler(), ck::PipelineVersion PipelineVer = ck::PipelineVersion::v1>
index_t ck::tensor_operation::device::DeviceGroupedQueryAttentionForward_Wmma< NumDimG, NumDimM, NumDimL, NumDimK, NumDimN, ADataType, B0DataType, B1DataType, CDataType, Acc0BiasDataType, Acc0DataType, Acc1BiasDataType, Acc1DataType, CShuffleDataType, AElementwiseOperation, B0ElementwiseOperation, AccElementwiseOperation, B1ElementwiseOperation, CElementwiseOperation, GemmSpec, ASpec, B0Spec, B1Spec, CSpec, NumPrefetch, QueryGroupNumber, BlockSize, MPerBlock, LPerBlock, KPerBlock, AK1, BK1, NPerBlock, LTilePerBlock, L1, MPerWmma, LPerWmma, NPerWmma, MRepeat, LRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_K1, ABlockLdsAddExtraM, B0BlockTransferThreadClusterLengths_K0_L_K1, B0BlockTransferThreadClusterArrangeOrder, B0BlockTransferSrcAccessOrder, B0BlockTransferSrcVectorDim, B0BlockTransferSrcScalarPerVector, B0BlockTransferDstScalarPerVector_K1, B0BlockLdsAddExtraL, B1BlockTransferThreadClusterLengths_L0_N_L1, B1BlockTransferThreadClusterArrangeOrder, B1BlockTransferSrcAccessOrder, B1BlockTransferSrcVectorDim, B1BlockTransferSrcScalarPerVector, B1BlockTransferDstScalarPerVector_L1, B1BlockLdsAddExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpec, LoopSched, PipelineVer >::RawArg::O_

◆ output_permute_

template<index_t NumDimG, index_t NumDimM, index_t NumDimL, index_t NumDimK, index_t NumDimN, typename ADataType, typename B0DataType, typename B1DataType, typename CDataType, typename Acc0BiasDataType, typename Acc0DataType, typename Acc1BiasDataType, typename Acc1DataType, typename CShuffleDataType, typename AElementwiseOperation, typename B0ElementwiseOperation, typename AccElementwiseOperation, typename B1ElementwiseOperation, typename CElementwiseOperation, GemmSpecialization GemmSpec, TensorSpecialization ASpec, TensorSpecialization B0Spec, TensorSpecialization B1Spec, TensorSpecialization CSpec, ck::index_t NumPrefetch, ck::index_t QueryGroupNumber, ck::index_t BlockSize, ck::index_t MPerBlock, ck::index_t LPerBlock, ck::index_t KPerBlock, ck::index_t AK1, ck::index_t BK1, ck::index_t NPerBlock, ck::index_t LTilePerBlock, ck::index_t L1, ck::index_t MPerWmma, ck::index_t LPerWmma, ck::index_t NPerWmma, ck::index_t MRepeat, ck::index_t LRepeat, ck::index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, ck::index_t ABlockTransferSrcVectorDim, ck::index_t ABlockTransferSrcScalarPerVector, ck::index_t ABlockTransferDstScalarPerVector_K1, bool ABlockLdsAddExtraM, typename B0BlockTransferThreadClusterLengths_K0_L_K1, typename B0BlockTransferThreadClusterArrangeOrder, typename B0BlockTransferSrcAccessOrder, ck::index_t B0BlockTransferSrcVectorDim, ck::index_t B0BlockTransferSrcScalarPerVector, ck::index_t B0BlockTransferDstScalarPerVector_K1, bool B0BlockLdsAddExtraL, typename B1BlockTransferThreadClusterLengths_L0_N_L1, typename B1BlockTransferThreadClusterArrangeOrder, typename B1BlockTransferSrcAccessOrder, ck::index_t B1BlockTransferSrcVectorDim, ck::index_t B1BlockTransferSrcScalarPerVector, ck::index_t B1BlockTransferDstScalarPerVector_L1, bool B1BlockLdsAddExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, typename CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, index_t CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpecialization MaskingSpec, ck::LoopScheduler LoopSched = make_default_loop_scheduler(), ck::PipelineVersion PipelineVer = ck::PipelineVersion::v1>
bool ck::tensor_operation::device::DeviceGroupedQueryAttentionForward_Wmma< NumDimG, NumDimM, NumDimL, NumDimK, NumDimN, ADataType, B0DataType, B1DataType, CDataType, Acc0BiasDataType, Acc0DataType, Acc1BiasDataType, Acc1DataType, CShuffleDataType, AElementwiseOperation, B0ElementwiseOperation, AccElementwiseOperation, B1ElementwiseOperation, CElementwiseOperation, GemmSpec, ASpec, B0Spec, B1Spec, CSpec, NumPrefetch, QueryGroupNumber, BlockSize, MPerBlock, LPerBlock, KPerBlock, AK1, BK1, NPerBlock, LTilePerBlock, L1, MPerWmma, LPerWmma, NPerWmma, MRepeat, LRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_K1, ABlockLdsAddExtraM, B0BlockTransferThreadClusterLengths_K0_L_K1, B0BlockTransferThreadClusterArrangeOrder, B0BlockTransferSrcAccessOrder, B0BlockTransferSrcVectorDim, B0BlockTransferSrcScalarPerVector, B0BlockTransferDstScalarPerVector_K1, B0BlockLdsAddExtraL, B1BlockTransferThreadClusterLengths_L0_N_L1, B1BlockTransferThreadClusterArrangeOrder, B1BlockTransferSrcAccessOrder, B1BlockTransferSrcVectorDim, B1BlockTransferSrcScalarPerVector, B1BlockTransferDstScalarPerVector_L1, B1BlockLdsAddExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpec, LoopSched, PipelineVer >::RawArg::output_permute_

◆ p_a_grid_

template<index_t NumDimG, index_t NumDimM, index_t NumDimL, index_t NumDimK, index_t NumDimN, typename ADataType, typename B0DataType, typename B1DataType, typename CDataType, typename Acc0BiasDataType, typename Acc0DataType, typename Acc1BiasDataType, typename Acc1DataType, typename CShuffleDataType, typename AElementwiseOperation, typename B0ElementwiseOperation, typename AccElementwiseOperation, typename B1ElementwiseOperation, typename CElementwiseOperation, GemmSpecialization GemmSpec, TensorSpecialization ASpec, TensorSpecialization B0Spec, TensorSpecialization B1Spec, TensorSpecialization CSpec, ck::index_t NumPrefetch, ck::index_t QueryGroupNumber, ck::index_t BlockSize, ck::index_t MPerBlock, ck::index_t LPerBlock, ck::index_t KPerBlock, ck::index_t AK1, ck::index_t BK1, ck::index_t NPerBlock, ck::index_t LTilePerBlock, ck::index_t L1, ck::index_t MPerWmma, ck::index_t LPerWmma, ck::index_t NPerWmma, ck::index_t MRepeat, ck::index_t LRepeat, ck::index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, ck::index_t ABlockTransferSrcVectorDim, ck::index_t ABlockTransferSrcScalarPerVector, ck::index_t ABlockTransferDstScalarPerVector_K1, bool ABlockLdsAddExtraM, typename B0BlockTransferThreadClusterLengths_K0_L_K1, typename B0BlockTransferThreadClusterArrangeOrder, typename B0BlockTransferSrcAccessOrder, ck::index_t B0BlockTransferSrcVectorDim, ck::index_t B0BlockTransferSrcScalarPerVector, ck::index_t B0BlockTransferDstScalarPerVector_K1, bool B0BlockLdsAddExtraL, typename B1BlockTransferThreadClusterLengths_L0_N_L1, typename B1BlockTransferThreadClusterArrangeOrder, typename B1BlockTransferSrcAccessOrder, ck::index_t B1BlockTransferSrcVectorDim, ck::index_t B1BlockTransferSrcScalarPerVector, ck::index_t B1BlockTransferDstScalarPerVector_L1, bool B1BlockLdsAddExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, typename CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, index_t CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpecialization MaskingSpec, ck::LoopScheduler LoopSched = make_default_loop_scheduler(), ck::PipelineVersion PipelineVer = ck::PipelineVersion::v1>
const ADataType* ck::tensor_operation::device::DeviceGroupedQueryAttentionForward_Wmma< NumDimG, NumDimM, NumDimL, NumDimK, NumDimN, ADataType, B0DataType, B1DataType, CDataType, Acc0BiasDataType, Acc0DataType, Acc1BiasDataType, Acc1DataType, CShuffleDataType, AElementwiseOperation, B0ElementwiseOperation, AccElementwiseOperation, B1ElementwiseOperation, CElementwiseOperation, GemmSpec, ASpec, B0Spec, B1Spec, CSpec, NumPrefetch, QueryGroupNumber, BlockSize, MPerBlock, LPerBlock, KPerBlock, AK1, BK1, NPerBlock, LTilePerBlock, L1, MPerWmma, LPerWmma, NPerWmma, MRepeat, LRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_K1, ABlockLdsAddExtraM, B0BlockTransferThreadClusterLengths_K0_L_K1, B0BlockTransferThreadClusterArrangeOrder, B0BlockTransferSrcAccessOrder, B0BlockTransferSrcVectorDim, B0BlockTransferSrcScalarPerVector, B0BlockTransferDstScalarPerVector_K1, B0BlockLdsAddExtraL, B1BlockTransferThreadClusterLengths_L0_N_L1, B1BlockTransferThreadClusterArrangeOrder, B1BlockTransferSrcAccessOrder, B1BlockTransferSrcVectorDim, B1BlockTransferSrcScalarPerVector, B1BlockTransferDstScalarPerVector_L1, B1BlockLdsAddExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpec, LoopSched, PipelineVer >::RawArg::p_a_grid_

◆ p_b0_grid_

template<index_t NumDimG, index_t NumDimM, index_t NumDimL, index_t NumDimK, index_t NumDimN, typename ADataType, typename B0DataType, typename B1DataType, typename CDataType, typename Acc0BiasDataType, typename Acc0DataType, typename Acc1BiasDataType, typename Acc1DataType, typename CShuffleDataType, typename AElementwiseOperation, typename B0ElementwiseOperation, typename AccElementwiseOperation, typename B1ElementwiseOperation, typename CElementwiseOperation, GemmSpecialization GemmSpec, TensorSpecialization ASpec, TensorSpecialization B0Spec, TensorSpecialization B1Spec, TensorSpecialization CSpec, ck::index_t NumPrefetch, ck::index_t QueryGroupNumber, ck::index_t BlockSize, ck::index_t MPerBlock, ck::index_t LPerBlock, ck::index_t KPerBlock, ck::index_t AK1, ck::index_t BK1, ck::index_t NPerBlock, ck::index_t LTilePerBlock, ck::index_t L1, ck::index_t MPerWmma, ck::index_t LPerWmma, ck::index_t NPerWmma, ck::index_t MRepeat, ck::index_t LRepeat, ck::index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, ck::index_t ABlockTransferSrcVectorDim, ck::index_t ABlockTransferSrcScalarPerVector, ck::index_t ABlockTransferDstScalarPerVector_K1, bool ABlockLdsAddExtraM, typename B0BlockTransferThreadClusterLengths_K0_L_K1, typename B0BlockTransferThreadClusterArrangeOrder, typename B0BlockTransferSrcAccessOrder, ck::index_t B0BlockTransferSrcVectorDim, ck::index_t B0BlockTransferSrcScalarPerVector, ck::index_t B0BlockTransferDstScalarPerVector_K1, bool B0BlockLdsAddExtraL, typename B1BlockTransferThreadClusterLengths_L0_N_L1, typename B1BlockTransferThreadClusterArrangeOrder, typename B1BlockTransferSrcAccessOrder, ck::index_t B1BlockTransferSrcVectorDim, ck::index_t B1BlockTransferSrcScalarPerVector, ck::index_t B1BlockTransferDstScalarPerVector_L1, bool B1BlockLdsAddExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, typename CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, index_t CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpecialization MaskingSpec, ck::LoopScheduler LoopSched = make_default_loop_scheduler(), ck::PipelineVersion PipelineVer = ck::PipelineVersion::v1>
const B0DataType* ck::tensor_operation::device::DeviceGroupedQueryAttentionForward_Wmma< NumDimG, NumDimM, NumDimL, NumDimK, NumDimN, ADataType, B0DataType, B1DataType, CDataType, Acc0BiasDataType, Acc0DataType, Acc1BiasDataType, Acc1DataType, CShuffleDataType, AElementwiseOperation, B0ElementwiseOperation, AccElementwiseOperation, B1ElementwiseOperation, CElementwiseOperation, GemmSpec, ASpec, B0Spec, B1Spec, CSpec, NumPrefetch, QueryGroupNumber, BlockSize, MPerBlock, LPerBlock, KPerBlock, AK1, BK1, NPerBlock, LTilePerBlock, L1, MPerWmma, LPerWmma, NPerWmma, MRepeat, LRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_K1, ABlockLdsAddExtraM, B0BlockTransferThreadClusterLengths_K0_L_K1, B0BlockTransferThreadClusterArrangeOrder, B0BlockTransferSrcAccessOrder, B0BlockTransferSrcVectorDim, B0BlockTransferSrcScalarPerVector, B0BlockTransferDstScalarPerVector_K1, B0BlockLdsAddExtraL, B1BlockTransferThreadClusterLengths_L0_N_L1, B1BlockTransferThreadClusterArrangeOrder, B1BlockTransferSrcAccessOrder, B1BlockTransferSrcVectorDim, B1BlockTransferSrcScalarPerVector, B1BlockTransferDstScalarPerVector_L1, B1BlockLdsAddExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpec, LoopSched, PipelineVer >::RawArg::p_b0_grid_

◆ p_b1_grid_

template<index_t NumDimG, index_t NumDimM, index_t NumDimL, index_t NumDimK, index_t NumDimN, typename ADataType, typename B0DataType, typename B1DataType, typename CDataType, typename Acc0BiasDataType, typename Acc0DataType, typename Acc1BiasDataType, typename Acc1DataType, typename CShuffleDataType, typename AElementwiseOperation, typename B0ElementwiseOperation, typename AccElementwiseOperation, typename B1ElementwiseOperation, typename CElementwiseOperation, GemmSpecialization GemmSpec, TensorSpecialization ASpec, TensorSpecialization B0Spec, TensorSpecialization B1Spec, TensorSpecialization CSpec, ck::index_t NumPrefetch, ck::index_t QueryGroupNumber, ck::index_t BlockSize, ck::index_t MPerBlock, ck::index_t LPerBlock, ck::index_t KPerBlock, ck::index_t AK1, ck::index_t BK1, ck::index_t NPerBlock, ck::index_t LTilePerBlock, ck::index_t L1, ck::index_t MPerWmma, ck::index_t LPerWmma, ck::index_t NPerWmma, ck::index_t MRepeat, ck::index_t LRepeat, ck::index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, ck::index_t ABlockTransferSrcVectorDim, ck::index_t ABlockTransferSrcScalarPerVector, ck::index_t ABlockTransferDstScalarPerVector_K1, bool ABlockLdsAddExtraM, typename B0BlockTransferThreadClusterLengths_K0_L_K1, typename B0BlockTransferThreadClusterArrangeOrder, typename B0BlockTransferSrcAccessOrder, ck::index_t B0BlockTransferSrcVectorDim, ck::index_t B0BlockTransferSrcScalarPerVector, ck::index_t B0BlockTransferDstScalarPerVector_K1, bool B0BlockLdsAddExtraL, typename B1BlockTransferThreadClusterLengths_L0_N_L1, typename B1BlockTransferThreadClusterArrangeOrder, typename B1BlockTransferSrcAccessOrder, ck::index_t B1BlockTransferSrcVectorDim, ck::index_t B1BlockTransferSrcScalarPerVector, ck::index_t B1BlockTransferDstScalarPerVector_L1, bool B1BlockLdsAddExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, typename CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, index_t CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpecialization MaskingSpec, ck::LoopScheduler LoopSched = make_default_loop_scheduler(), ck::PipelineVersion PipelineVer = ck::PipelineVersion::v1>
const B1DataType* ck::tensor_operation::device::DeviceGroupedQueryAttentionForward_Wmma< NumDimG, NumDimM, NumDimL, NumDimK, NumDimN, ADataType, B0DataType, B1DataType, CDataType, Acc0BiasDataType, Acc0DataType, Acc1BiasDataType, Acc1DataType, CShuffleDataType, AElementwiseOperation, B0ElementwiseOperation, AccElementwiseOperation, B1ElementwiseOperation, CElementwiseOperation, GemmSpec, ASpec, B0Spec, B1Spec, CSpec, NumPrefetch, QueryGroupNumber, BlockSize, MPerBlock, LPerBlock, KPerBlock, AK1, BK1, NPerBlock, LTilePerBlock, L1, MPerWmma, LPerWmma, NPerWmma, MRepeat, LRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_K1, ABlockLdsAddExtraM, B0BlockTransferThreadClusterLengths_K0_L_K1, B0BlockTransferThreadClusterArrangeOrder, B0BlockTransferSrcAccessOrder, B0BlockTransferSrcVectorDim, B0BlockTransferSrcScalarPerVector, B0BlockTransferDstScalarPerVector_K1, B0BlockLdsAddExtraL, B1BlockTransferThreadClusterLengths_L0_N_L1, B1BlockTransferThreadClusterArrangeOrder, B1BlockTransferSrcAccessOrder, B1BlockTransferSrcVectorDim, B1BlockTransferSrcScalarPerVector, B1BlockTransferDstScalarPerVector_L1, B1BlockLdsAddExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpec, LoopSched, PipelineVer >::RawArg::p_b1_grid_

◆ p_c_grid_

template<index_t NumDimG, index_t NumDimM, index_t NumDimL, index_t NumDimK, index_t NumDimN, typename ADataType, typename B0DataType, typename B1DataType, typename CDataType, typename Acc0BiasDataType, typename Acc0DataType, typename Acc1BiasDataType, typename Acc1DataType, typename CShuffleDataType, typename AElementwiseOperation, typename B0ElementwiseOperation, typename AccElementwiseOperation, typename B1ElementwiseOperation, typename CElementwiseOperation, GemmSpecialization GemmSpec, TensorSpecialization ASpec, TensorSpecialization B0Spec, TensorSpecialization B1Spec, TensorSpecialization CSpec, ck::index_t NumPrefetch, ck::index_t QueryGroupNumber, ck::index_t BlockSize, ck::index_t MPerBlock, ck::index_t LPerBlock, ck::index_t KPerBlock, ck::index_t AK1, ck::index_t BK1, ck::index_t NPerBlock, ck::index_t LTilePerBlock, ck::index_t L1, ck::index_t MPerWmma, ck::index_t LPerWmma, ck::index_t NPerWmma, ck::index_t MRepeat, ck::index_t LRepeat, ck::index_t NRepeat, typename ABlockTransferThreadClusterLengths_K0_M_K1, typename ABlockTransferThreadClusterArrangeOrder, typename ABlockTransferSrcAccessOrder, ck::index_t ABlockTransferSrcVectorDim, ck::index_t ABlockTransferSrcScalarPerVector, ck::index_t ABlockTransferDstScalarPerVector_K1, bool ABlockLdsAddExtraM, typename B0BlockTransferThreadClusterLengths_K0_L_K1, typename B0BlockTransferThreadClusterArrangeOrder, typename B0BlockTransferSrcAccessOrder, ck::index_t B0BlockTransferSrcVectorDim, ck::index_t B0BlockTransferSrcScalarPerVector, ck::index_t B0BlockTransferDstScalarPerVector_K1, bool B0BlockLdsAddExtraL, typename B1BlockTransferThreadClusterLengths_L0_N_L1, typename B1BlockTransferThreadClusterArrangeOrder, typename B1BlockTransferSrcAccessOrder, ck::index_t B1BlockTransferSrcVectorDim, ck::index_t B1BlockTransferSrcScalarPerVector, ck::index_t B1BlockTransferDstScalarPerVector_L1, bool B1BlockLdsAddExtraN, index_t CShuffleMRepeatPerShuffle, index_t CShuffleNRepeatPerShuffle, typename CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, index_t CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpecialization MaskingSpec, ck::LoopScheduler LoopSched = make_default_loop_scheduler(), ck::PipelineVersion PipelineVer = ck::PipelineVersion::v1>
CDataType* ck::tensor_operation::device::DeviceGroupedQueryAttentionForward_Wmma< NumDimG, NumDimM, NumDimL, NumDimK, NumDimN, ADataType, B0DataType, B1DataType, CDataType, Acc0BiasDataType, Acc0DataType, Acc1BiasDataType, Acc1DataType, CShuffleDataType, AElementwiseOperation, B0ElementwiseOperation, AccElementwiseOperation, B1ElementwiseOperation, CElementwiseOperation, GemmSpec, ASpec, B0Spec, B1Spec, CSpec, NumPrefetch, QueryGroupNumber, BlockSize, MPerBlock, LPerBlock, KPerBlock, AK1, BK1, NPerBlock, LTilePerBlock, L1, MPerWmma, LPerWmma, NPerWmma, MRepeat, LRepeat, NRepeat, ABlockTransferThreadClusterLengths_K0_M_K1, ABlockTransferThreadClusterArrangeOrder, ABlockTransferSrcAccessOrder, ABlockTransferSrcVectorDim, ABlockTransferSrcScalarPerVector, ABlockTransferDstScalarPerVector_K1, ABlockLdsAddExtraM, B0BlockTransferThreadClusterLengths_K0_L_K1, B0BlockTransferThreadClusterArrangeOrder, B0BlockTransferSrcAccessOrder, B0BlockTransferSrcVectorDim, B0BlockTransferSrcScalarPerVector, B0BlockTransferDstScalarPerVector_K1, B0BlockLdsAddExtraL, B1BlockTransferThreadClusterLengths_L0_N_L1, B1BlockTransferThreadClusterArrangeOrder, B1BlockTransferSrcAccessOrder, B1BlockTransferSrcVectorDim, B1BlockTransferSrcScalarPerVector, B1BlockTransferDstScalarPerVector_L1, B1BlockLdsAddExtraN, CShuffleMRepeatPerShuffle, CShuffleNRepeatPerShuffle, CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock, CShuffleBlockTransferScalarPerVector_NPerBlock, MaskingSpec, LoopSched, PipelineVer >::RawArg::p_c_grid_
The documentation for this struct was generated from the following file: