Home Explore Help
Sign In
cturan
/
llama.cpp
mirror of https://github.com/cturan/llama.cpp
1
0
Fork 0
Files Issues 0 Wiki
Browse Source

fix : cuda order of synchronization when setting a buffer (ggml/679)

* fix : cuda order of synchronization when setting a buffer

* also sync before memcpy

---------

Co-authored-by: slaren <slarengh@gmail.com>
Erik Scholz 2 years ago
parent
7a9f75c38b
commit
f34432ca1e
1 changed files with 1 additions and 1 deletions
Unified View Show Diff Stats
  1. 1 1
      ggml-cuda.cu

+ 1 - 1
ggml-cuda.cu
View File 

@@ -10184,8 +10184,8 @@ static void ggml_backend_cuda_buffer_set_tensor(ggml_backend_buffer_t buffer, gg
 
 
 
     ggml_cuda_set_device(ctx->device);
 
     ggml_cuda_set_device(ctx->device);
 
     CUDA_CHECK(cudaDeviceSynchronize());
 
     CUDA_CHECK(cudaDeviceSynchronize());
 
-
 
     CUDA_CHECK(cudaMemcpy((char *)tensor->data + offset, data, size, cudaMemcpyHostToDevice));
 
     CUDA_CHECK(cudaMemcpy((char *)tensor->data + offset, data, size, cudaMemcpyHostToDevice));
 
+    CUDA_CHECK(cudaDeviceSynchronize());
 
 }
 
 }
 
 
 
 static void ggml_backend_cuda_buffer_get_tensor(ggml_backend_buffer_t buffer, const ggml_tensor * tensor, void * data, size_t offset, size_t size) {
 
 static void ggml_backend_cuda_buffer_get_tensor(ggml_backend_buffer_t buffer, const ggml_tensor * tensor, void * data, size_t offset, size_t size) {