Skip to content
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.

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

UCP/TAG: Tag offload bounce buffer optimization #4589

Merged
merged 1 commit into from
Dec 20, 2019
Merged
Changes from all commits
Commits
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
5 changes: 4 additions & 1 deletion src/ucp/tag/offload.c
Original file line number Diff line number Diff line change
Expand Up @@ -256,7 +256,10 @@ ucp_tag_offload_do_post(ucp_request_t *req)

mdi = context->tl_rscs[wiface->rsc_index].md_index;

if (ucs_unlikely(length >= worker->tm.offload.zcopy_thresh)) {
/* Do not use bounce buffer for receives to GPU memory to avoid
* cost of h2d transfers (i.e. cuda_copy from staging to dest memory). */
if ((length >= worker->tm.offload.zcopy_thresh) ||
!UCP_MEM_IS_ACCESSIBLE_FROM_CPU(req->recv.mem_type)) {
if (length > wiface->attr.cap.tag.recv.max_zcopy) {
/* Post maximum allowed length. If sender sends smaller message
* (which is allowed per MPI standard), max recv should fit it.
Expand Down