CUTLASS
CUDA Templates for Linear Algebra Subroutines and Solvers
|
Threadblock swizzling function for split-K GEMMs.
#include <threadblock_swizzle.h>
Public Member Functions | |
CUTLASS_HOST_DEVICE GemmCoord | get_tiled_shape (GemmCoord problem_size, GemmCoord tile_size, int partitions) const |
Returns the shape of the problem in units of logical tiles. More... | |
CUTLASS_HOST_DEVICE dim3 | get_grid_shape (GemmCoord tiled_shape) const |
Computes CUDA grid dimensions given a size in units of logical tiles. More... | |
CUTLASS_DEVICE GemmCoord | get_tile_offset () const |
Obtains the threadblock offset (in units of threadblock-scoped tiles) More... | |
|
inline |
|
inline |
|
inline |