Skip to content

[UR][CUDA] Fix urUSMContextMemcpyExp synchronization issue #21602

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

Draft
kekaczma wants to merge 1 commit into
base: sycl
Choose a base branch
from
Draft

[UR][CUDA] Fix urUSMContextMemcpyExp synchronization issue #21602

Show file tree
Hide file tree
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
16 changes: 12 additions & 4 deletions unified-runtime/source/adapters/cuda/usm.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -574,10 +574,18 @@ urUSMPoolTrimToExp(ur_context_handle_t hContext, ur_device_handle_t hDevice,
return UR_RESULT_SUCCESS;
}

UR_APIEXPORT ur_result_t UR_APICALL urUSMContextMemcpyExp(ur_context_handle_t,
void *pDst,
const void *pSrc,
size_t Size) {
UR_APIEXPORT ur_result_t UR_APICALL urUSMContextMemcpyExp(
ur_context_handle_t hContext, void *pDst, const void *pSrc, size_t Size) {
// cuMemcpy is synchronous with respect to the host, but it does not
// synchronize with any device streams. We need to synchronize all streams
// in the context before performing the copy to ensure all previous
// operations have completed.
//
// Set the context and synchronize all streams
ScopedContext Active(hContext->getDevices().front());
UR_CHECK_ERROR(cuCtxSynchronize());

// Now perform the synchronous copy
UR_CHECK_ERROR(cuMemcpy((CUdeviceptr)pDst, (CUdeviceptr)pSrc, Size));
return UR_RESULT_SUCCESS;
}
2 changes: 0 additions & 2 deletions unified-runtime/test/conformance/exp_usm_context_memcpy/urUSMContextMemcpyExp.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -81,8 +81,6 @@ struct urUSMContextMemcpyExpTestDevice : urUSMContextMemcpyExpTest {
UUR_INSTANTIATE_DEVICE_TEST_SUITE_MULTI_QUEUE(urUSMContextMemcpyExpTestDevice);

TEST_P(urUSMContextMemcpyExpTestDevice, Success) {
// https://github.com/intel/llvm/issues/19688
UUR_KNOWN_FAILURE_ON(uur::CUDA{});
ASSERT_SUCCESS(
urUSMContextMemcpyExp(context, dst_ptr, src_ptr, allocation_size));
verifyData();
Expand Down
Loading