1 Release 260 Tags

RSS Feed

• v12.81 983f41a1f0

  Compare

  fix(nvidia): detect NVSwitch fabric via driver, not host-PCIe

  mchus released this 2026-07-29 18:58:19 +03:00

  On HGX B200 (Kaytus KR9288-X3) NVSwitch is managed via HMC/ERoT and never
  appears on host lspci as 10de:0680, so the old check always returned
  "not present" and silently skipped fabricmanager — leaving the fabric
  untrained and CUDA/NCCL/dcgmi failing with "system not yet initialized"
  on otherwise-healthy hardware. bee-check-nvswitch now checks, in order,
  nvidia-smi's GPU Fabric GUID, /proc/driver/nvidia-nvswitch/devices, and
  finally the legacy host-PCIe scan for older HGX generations.

  Downloads

  • Source Code (ZIP)
  • Source Code (TAR.GZ)