TiledCUDA is a highly efficient kernel template library designed to elevate CUDA C’s level of abstraction for processing tiles.
158
stars
10
forks
source link
feat(unittest): Add unittest to ensure the correctness of swizzled layout. #98
Closed
haruhi55 closed 4 months ago
This PR should be merged after this one https://github.com/TiledTensor/TiledCUDA/pull/97