apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2026-04-23 00:17:25 +08:00
Code Issues Actions 19 Packages Projects Releases Wiki Activity

router support divided roolout (#6150)

Browse Source
This commit is contained in:
jc
2026-01-22 10:39:39 +08:00
committed by GitHub
parent 9c4db0ac3f
commit 309c7d9764
2 changed files with 142 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files
custom_ops/gpu_ops/swap_cache_layout.cu
+10 -1
View File
@@ -73,6 +73,11 @@ void SwapCacheImpLayout(
copy_kind,
stream);
PADDLE_ENFORCE_EQ(status,
cudaSuccess,
phi::errors::External("cudaMemcpyAsync failed: %s",
cudaGetErrorString(status)));
#ifdef SWAP_DEBUG
cudaStreamSynchronize(stream);
std::cout << "mode:" << mode << ", layer_idx:" << layer_idx
@@ -81,7 +86,11 @@ void SwapCacheImpLayout(
#endif
}
}
cudaStreamSynchronize(stream);
cudaError_t sync_status = cudaStreamSynchronize(stream);
PADDLE_ENFORCE_EQ(sync_status,
cudaSuccess,
phi::errors::External("cudaStreamSynchronize failed: %s",
cudaGetErrorString(sync_status)));
}
void SwapCacheLayout(