igemm_global_tile.h File Reference

Implements tile iterators to partition the thread block tile into 2D subtiles and efficiently load each. Applies permute transformation to construct 'interleaved K-strided' data layout in which 4-element dot products from the same K index are arranged in consecutive locations within shared memory. More...

#include "cutlass/coord.h"
#include "cutlass/gemm/gemm_global_tile.h"
#include "cutlass/matrix_traits.h"

Go to the source code of this file.

Classes
struct	cutlass::gemm::IgemmGlobalTileTraits< kOperand_, kLayout_, Scalar_, Tile_, Threads_, kAccessSize_ >
struct	cutlass::gemm::IgemmGlobalTileTraits< kOperand_, kLayout_, Scalar_, Tile_, Threads_, kAccessSize_ >::ThreadOffset
	Computes the thread offset in (H, W) based on thread ID. More...
struct	cutlass::gemm::IgemmGlobalIteratorAb< TileTraits_, Index_ >

Namespaces
	cutlass
	cutlass::gemm

Detailed Description

Supports efficient loads from shared memory to target the DP4A instruction.