block_to_ctile_map.hpp File Reference

• include
• ck
• tensor_operation
• gpu
• grid

Classes | Namespaces | Enumerations | Functions

block_to_ctile_map.hpp File Reference

#include "ck/utility/math.hpp"
#include "ck/utility/number.hpp"
#include "ck/utility/tuple.hpp"
#include "ck/tensor_description/tensor_adaptor.hpp"
#include "ck/tensor_description/multi_index_transform_helper.hpp"
#include <limits>
#include <stdlib.h>

Go to the source code of this file.

Classes
struct	ck::BlockToCTileMap_M00_N0_M01< MPerBlock, NPerBlock, CGridDesc_M_N, DeviceCTileIndexCheck >
struct	ck::BlockToCTileMap_M00_N0_M01Adapt< MPerBlock, NPerBlock, void >
struct	ck::BlockToCTileMap_M00_N0_M01Adapt< MPerBlock, NPerBlock, CGridDesc_M_N >
struct	ck::BlockToCTileMap_Grouped_M00_N0_M01Adapt< GroupNum, MPerBlock, NPerBlock >
struct	ck::BlockToCTileMap_N00_M0_N01Adapt< MPerBlock, NPerBlock, void >
struct	ck::BlockToCTileMap_KSplit_M00_N0_M01Adapt< MPerBlock, NPerBlock, CGridDesc_M_N >
struct	ck::BlockToCTileMap_M00_N00_M01_N01< MPerBlock, NPerBlock, CGridDesc_M_N, DeviceCTileIndexCheck >
struct	ck::BlockToCTileMap_KSplit_M00_N00_M01_N01< MPerBlock, NPerBlock, CGridDesc_M_N, DeviceCTileIndexCheck >
struct	ck::OffsettedBlockToCTileMap< UnderlyingBlockToCTileMap >
struct	ck::OffsettedBlockToCTileMap2< UnderlyingBlockToCTileMap >
struct	ck::BlockToCTileMap_3DGrid_KSplit< MPerBlock, NPerBlock >
	Simple tile mapping which creates 3D grid of block of threads. More...
struct	ck::BlockToCTileMap_GemmStreamK< MPerBlock_, NPerBlock_, KPerBlock_, ReductionStrategy_, TileSwizzleSubM_ >
struct	ck::BlockToCTileMap_GemmStreamK_v2< MPerBlock_, NPerBlock_, KPerBlock_, ReductionStrategy_, TileSwizzleSubM_, GroupNum, M01_ >

Namespaces
namespace	ck

Enumerations
enum	ck::StreamKReductionStrategy {   ck::Atomic = 0 ,   ck::Reduction }

Functions
template<typename CTileIdx, typename CTileDim>
__host__ __device__ bool	ck::DefaultValidCTileIndex (const CTileIdx &c_tile_idx, const CTileDim &c_tile_dim)