Skip to content

HIP: workaround runtime bug in hipGraph support #11964

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
wants to merge 1 commit into from
Closed

HIP: workaround runtime bug in hipGraph support #11964

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
7 changes: 7 additions & 0 deletions ggml/src/ggml-cuda/ggml-cuda.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2558,7 +2558,14 @@ static void maintain_cuda_graph(ggml_backend_cuda_context * cuda_ctx, std::vecto
for (size_t i = 0; i < cuda_ctx->cuda_graph->num_nodes; i++) {
if(count(ggml_cuda_cpy_fn_ptrs.begin(), ggml_cuda_cpy_fn_ptrs.end(), cuda_ctx->cuda_graph->params[i].func) > 0) {
char ** updated_kernel_arg_ptr = cuda_ctx->cuda_graph->updated_kernel_arg.at(k++);
#if defined(__HIP_PLATFORM_AMD__)
// Workaround for https://github.com/ROCm/clr/issues/138
// The hip runtime fails to copy this and calls delete on it later, so we must
// perform an alloc here.
cuda_ctx->cuda_graph->params[i].kernelParams[1] = new void*(*updated_kernel_arg_ptr);
#else
cuda_ctx->cuda_graph->params[i].kernelParams[1] = updated_kernel_arg_ptr;
#endif // defined(__HIP_PLATFORM_AMD__)
CUDA_CHECK(cudaGraphKernelNodeSetParams(cuda_ctx->cuda_graph->nodes[i], &cuda_ctx->cuda_graph->params[i]));
}
}
Expand Down
Loading