gridwise_elementwise_2d.hpp File Reference#
gridwise_elementwise_2d.hpp File Reference
#include "ck/tensor_description/cluster_descriptor.hpp"#include "ck/utility/data_type.hpp"#include "ck/tensor_operation/gpu/thread/threadwise_tensor_slice_transfer.hpp"#include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"#include "ck/tensor_operation/gpu/block/thread_group_tensor_slice_transfer_v7r2.hpp"#include "ck/tensor_operation/gpu/block/thread_group_tensor_slice_transfer_v4r2.hpp"#include "ck/tensor/static_tensor.hpp"#include "ck/utility/common_header.hpp"Go to the source code of this file.
Namespaces | |
| namespace | ck |
Functions | |
| template<typename GridwiseElementwiseFunctor, typename InGridDescTuple, typename OutGridDescTuple, typename InDataTypePointerTuple, typename OutDataTypePointerTuple, typename Block2TileMap, typename ElementwiseOperation> | |
| __global__ void | ck::kernel_elementwise (const InGridDescTuple in_grid_desc_tuple, const OutGridDescTuple out_grid_desc_tuple, const InDataTypePointerTuple p_in_global_tuple, const OutDataTypePointerTuple p_out_global_tuple, const Block2TileMap block_2_tile_map, const ElementwiseOperation elementwise_op) |
| template<typename GridwiseElementwiseFunctorA, typename GridwiseElementwiseFunctorB, typename InAGridDescTuple, typename InBGridDescTuple, typename OutAGridDescTuple, typename OutBGridDescTuple, typename InADataTypePointerTuple, typename InBDataTypePointerTuple, typename OutADataTypePointerTuple, typename OutBDataTypePointerTuple, typename Block2TileMapA, typename Block2TileMapB, typename ElementwiseOperation> | |
| __global__ void | ck::kernel_elementwise_dual (const InAGridDescTuple in_grid_desc_tuple_a, const InBGridDescTuple in_grid_desc_tuple_b, const OutAGridDescTuple out_grid_desc_tuple_a, const OutBGridDescTuple out_grid_desc_tuple_b, const InADataTypePointerTuple p_in_global_tuple_a, const InBDataTypePointerTuple p_in_global_tuple_b, const OutADataTypePointerTuple p_out_global_tuple_a, const OutBDataTypePointerTuple p_out_global_tuple_b, const Block2TileMapA block_2_tile_map_a, const Block2TileMapB block_2_tile_map_b, const ElementwiseOperation elementwise_op, const index_t a_grid_size) |
| template<typename GridwiseElementwiseFunctorA, typename GridwiseElementwiseFunctorB, typename InAGridDescTuple, typename InBGridDescTuple, typename OutAGridDescTuple, typename OutBGridDescTuple, typename InADataTypePointerTuple, typename InBDataTypePointerTuple, typename OutADataTypePointerTuple, typename OutBDataTypePointerTuple, typename Block2TileMapA, typename Block2TileMapB, typename ElementwiseOperation, index_t NumInputsA, index_t NumInputsB, index_t NumOutputsA, index_t NumOutputsB> | |
| __global__ void | ck::kernel_elementwise_batched_dual (const InAGridDescTuple in_grid_desc_tuple_a, const InBGridDescTuple in_grid_desc_tuple_b, const OutAGridDescTuple out_grid_desc_tuple_a, const OutBGridDescTuple out_grid_desc_tuple_b, const InADataTypePointerTuple p_in_global_tuple_a, const InBDataTypePointerTuple p_in_global_tuple_b, const OutADataTypePointerTuple p_out_global_tuple_a, const OutBDataTypePointerTuple p_out_global_tuple_b, const Block2TileMapA block_2_tile_map_a, const Block2TileMapB block_2_tile_map_b, const ElementwiseOperation elementwise_op, const index_t a_grid_size, const index_t batch_count_a, const index_t batch_count_b, const std::array< index_t, NumInputsA > input_batch_strides_a, const std::array< index_t, NumInputsB > input_batch_strides_b, const std::array< index_t, NumOutputsA > output_batch_strides_a, const std::array< index_t, NumOutputsB > output_batch_strides_b) |
| template<typename GridwiseElementwiseFunctor, typename InGridDescTuple, typename OutGridDescTuple, typename InDataTypePointerTuple, typename OutDataTypePointerTuple, typename Block2TileMap, typename ElementwiseOperation, index_t NumInputs, index_t NumOutputs> | |
| __global__ void | ck::kernel_batched_elementwise (const InGridDescTuple in_grid_desc_tuple, const OutGridDescTuple out_grid_desc_tuple, const InDataTypePointerTuple p_in_global_tuple, const OutDataTypePointerTuple p_out_global_tuple, const Block2TileMap block_2_tile_map, const ElementwiseOperation elementwise_op, const index_t batch_count, const std::array< index_t, NumInputs > input_batch_strides, const std::array< index_t, NumOutputs > output_batch_strides) |