Skip to content

Fix incorrect MNNVL fabric check #2626

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

Open
nvcastet wants to merge 3 commits into
base: main
Choose a base branch
from
Open

Fix incorrect MNNVL fabric check #2626

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
3e06ef4
Fix incorrect MNNVL fabric check
nvcastet Jan 27, 2026
44a8e03
Merge branch 'main' into fix_mnnvl_check
ptrendx Feb 11, 2026
f1c02f0
Merge branch 'main' into fix_mnnvl_check
ptrendx Feb 12, 2026
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: 3 additions & 2 deletions transformer_engine/common/comm_gemm_overlap/userbuffers/userbuffers-host.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -122,10 +122,11 @@ bool has_mnnvl_fabric(int device_id) {
NVTE_CALL_CHECK_CUDA_NVML(nvmlDeviceGetHandleByIndex_v2, device_id, &local_device);
nvmlGpuFabricInfoV_t fabricInfo = {};
fabricInfo.version = nvmlGpuFabricInfo_v2;
fabricInfo.clusterUuid[0] = '\0';
NVTE_CALL_CHECK_CUDA_NVML(nvmlDeviceGetGpuFabricInfoV, local_device, &fabricInfo);
NVTE_CALL_CHECK_CUDA_NVML(nvmlShutdown);
if (fabricInfo.state >= NVML_GPU_FABRIC_STATE_COMPLETED && fabricInfo.clusterUuid[0] != '\0') {
const unsigned char zero_uuid[NVML_GPU_FABRIC_UUID_LEN] = {0};
if (fabricInfo.state == NVML_GPU_FABRIC_STATE_COMPLETED &&
memcmp(fabricInfo.clusterUuid, zero_uuid, NVML_GPU_FABRIC_UUID_LEN) != 0) {
mnnvl_fabric_support = true;
}
}
Expand Down
Loading