device_gemm_multiple_d_multiple_r_xdl_cshuffle.hpp File Reference

device_gemm_multiple_d_multiple_r_xdl_cshuffle.hpp File Reference#

Composable Kernel: device_gemm_multiple_d_multiple_r_xdl_cshuffle.hpp File Reference
Classes | Namespaces | Functions
device_gemm_multiple_d_multiple_r_xdl_cshuffle.hpp File Reference
#include <iostream>
#include <sstream>
#include "ck/utility/common_header.hpp"
#include "ck/tensor_description/tensor_descriptor.hpp"
#include "ck/tensor_description/tensor_descriptor_helper.hpp"
#include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
#include "ck/tensor_operation/gpu/device/device_gemm_multiple_d_multiple_r.hpp"
#include "ck/tensor_operation/gpu/device/gemm_specialization.hpp"
#include "ck/tensor_operation/gpu/device/matrix_padder.hpp"
#include "ck/tensor_operation/gpu/grid/gridwise_gemm_multiple_d_multiple_r_xdl_cshuffle.hpp"
#include "ck/host_utility/device_prop.hpp"
#include "ck/host_utility/kernel_launch.hpp"

Go to the source code of this file.

Classes

struct  ck::tensor_operation::device::DeviceGemmMultipleDMultipleR_Xdl_CShuffle< ALayout, BLayout, DELayout, ADataType, BDataType, GemmAccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, 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, CDEReduceThreadTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >
struct  ck::tensor_operation::device::DeviceGemmMultipleDMultipleR_Xdl_CShuffle< ALayout, BLayout, DELayout, ADataType, BDataType, GemmAccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, 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, CDEReduceThreadTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Argument
struct  ck::tensor_operation::device::DeviceGemmMultipleDMultipleR_Xdl_CShuffle< ALayout, BLayout, DELayout, ADataType, BDataType, GemmAccDataType, CShuffleDataType, DsDataType, EDataType, ReduceAccDataType, RsDataType, AElementwiseOperation, BElementwiseOperation, CDEElementwiseOperation, QsElementwiseOperation, RsElementwiseOperation, ThreadReduceOperations, RsGlobalMemoryDataOperation, 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, CDEReduceThreadTransferScalarPerVector_NPerBlock, RThreadTransferDstScalarPerVector_MPerBlock, LoopSched >::Invoker

Namespaces

namespace  ck
namespace  ck::tensor_operation
namespace  ck::tensor_operation::device

Functions

template<typename GridwiseGemm, typename FloatAB, typename FloatDsPointer, typename FloatE, typename FloatRsPointer, typename AElementwiseOperation, typename BElementwiseOperation, typename CDEElementwiseOperation, typename QsElementwiseOperation, typename RsElementwiseOperation, typename AGridDesc_AK0_M_AK1, typename BGridDesc_BK0_N_BK1, typename DsGridDescriptor_MBlock_MPerBlock_NBlock_NPerBlock, typename EGridDescriptor_MBlock_MPerBlock_NBlock_NPerBlock, typename RsGridDescriptor_MBlock_MPerBlock, typename Block2ETileMap, bool HasMainKBlockLoop>
__global__ void ck::kernel_gemm_multiple_d_multiple_r_xdl_cshuffle (const FloatAB *__restrict__ p_a_grid, const FloatAB *__restrict__ p_b_grid, FloatDsPointer p_ds_grid, FloatE *__restrict__ p_e_grid, FloatRsPointer p_rs_grid, 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, const AGridDesc_AK0_M_AK1 a_grid_desc_ak0_m_ak1, const BGridDesc_BK0_N_BK1 b_grid_desc_bk0_n_bk1, const DsGridDescriptor_MBlock_MPerBlock_NBlock_NPerBlock ds_grid_desc_mblock_mperblock_nblock_nperblock, const EGridDescriptor_MBlock_MPerBlock_NBlock_NPerBlock e_grid_desc_mblock_mperblock_nblock_nperblock, const RsGridDescriptor_MBlock_MPerBlock rs_grid_desc_mblock_mperblock, const Block2ETileMap block_2_etile_map)