mirror of
https://github.com/ROCm/composable_kernel.git
synced 2026-05-21 21:39:15 +00:00
5c7cec1115
* tuning para, * testing on v100 * add fp16 * remove deprecated tensor descriptor * sync with miopen * update build script Co-authored-by: Jing Zhang <jizhan@amd.com>
14 lines
256 B
C++
14 lines
256 B
C++
#ifndef CK_SYNCHRONIZATION_NVIDIA_HPP
|
|
#define CK_SYNCHRONIZATION_NVIDIA_HPP
|
|
|
|
#include "config.hpp"
|
|
|
|
namespace ck {
|
|
|
|
__device__ void block_sync_lds() { __syncthreads(); }
|
|
|
|
__device__ void block_sync_lds_vmem() { __syncthreads(); }
|
|
|
|
} // namespace ck
|
|
#endif
|