ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-04-20 14:59:17 +00:00
Code Issues Packages Projects Releases Wiki Activity

Add an option to change the number of warm-up cycles and iterations. (#1124)

Browse Source 
* allow setting the number of warmup cycles and iterations for profiler

* fix the gemm_splitk and grouped_gemm examples
This commit is contained in:
Illia Silin
2024-01-09 09:43:08 -08:00
committed by GitHub
parent e699dbd8a3
commit 886d9eeb99
9 changed files with 119 additions and 33 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
include/ck/host_utility/kernel_launch.hpp
View File

@@ -30,7 +30,7 @@ float launch_and_time_kernel(const StreamConfig& stream_config,
block_dim.y,
block_dim.z);
printf("Warm up 1 time\n");
printf("Warm up %d times\n", stream_config.cold_niters_);
#endif
// warm up
for(int i = 0; i < stream_config.cold_niters_; ++i)
@@ -103,14 +103,17 @@ float launch_and_time_kernel_with_preprocess(const StreamConfig& stream_config,
block_dim.y,
block_dim.z);
printf("Warm up 1 time\n");
printf("Warm up %d times\n", stream_config.cold_niters_);
#endif
// warm up
preprocess();
kernel<<<grid_dim, block_dim, lds_byte, stream_config.stream_id_>>>(args...);
hip_check_error(hipGetLastError());
for(int i = 0; i < stream_config.cold_niters_; ++i)
{
kernel<<<grid_dim, block_dim, lds_byte, stream_config.stream_id_>>>(args...);
hip_check_error(hipGetLastError());
}
const int nrepeat = 10;
const int nrepeat = stream_config.nrepeat_;
#if DEBUG_LOG
printf("Start running %d times...\n", nrepeat);
#endif