Cutlass
CUDA Templates for Linear Algebra Subroutines and Solvers
|
Transposes a tile of 16b elements. Used by HGEMM to construct a K-strided layout in shared memory for multiplicands. More...
Go to the source code of this file.
Classes | |
struct | cutlass::gemm::HgemmSwizzle< GlobalIterator_ > |
Namespaces | |
cutlass | |
cutlass::gemm | |