public_git/whisper.cpp
1
0
Fork 0
mirror of https://github.com/ggml-org/whisper.cpp.git synced 2026-10-20 17:23:00 +02:00
Code Issues Packages Projects Releases Wiki Activity

ggml-cuda: use per-thread stream for buffer-init padding memset (llama/28782)

Browse Source 
* ggml-cuda: use per-thread stream for buffer-init padding memset

* ci : re-enable test-backend-ops -j for ROCm
This commit is contained in:
Harkirat Gill
2026-10-19 09:47:10 +02:00
committed by Daniel Bevenius
parent d05db0d10b
commit 1ea1089b6b
1 changed files with 2 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ggml/src/ggml-cuda/ggml-cuda.cu
+2 -1
View File
@@ -762,7 +762,8 @@ static enum ggml_status ggml_backend_cuda_buffer_init_tensor(ggml_backend_buffer
if (padded_size > original_size) {
ggml_cuda_set_device(ctx->device);
CUDA_CHECK(cudaMemset((char *)tensor->data + original_size, 0, padded_size - original_size));
CUDA_CHECK(cudaMemsetAsync((char *)tensor->data + original_size, 0, padded_size - original_size, cudaStreamPerThread));
CUDA_CHECK(cudaStreamSynchronize(cudaStreamPerThread));
}
}
return GGML_STATUS_SUCCESS;