NVIDIA with c-series integration in UCS
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
05-07-2014 02:46 AM - edited 03-01-2019 11:39 AM
Hello,
Is it possible/supported to integrate C-240M3 with NVIDIA cards in a UCS Fabric Interconnect environment? I guess UCS Service Profiles don't know about NVIDIA cards, so I'm wondering how will the hypervisor know about its existence.
Thanks a lot,
- Labels:
-
Unified Computing System (UCS)

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
05-07-2014 04:36 AM
Have a look at
http://www.cisco.com/c/dam/en/us/products/collateral/servers-unified-computing/ucs-c-series-rack-servers/nvidia_grid_vgx.pdf
At a high level.
edit: I noticed your question is specific to integration with UCSM. For UCSM it will just be seen as s PCIe device. Just remember to have latest greatest on your C240 (1.5.4+)
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
05-07-2014 04:47 AM
So (provided we have last firmware) you can confirm that the scenario won't pose any problems or concerns? It will simply work as if it was a standalone server?
Thanks

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
05-07-2014 06:03 AM
In essence UCSM should see them as a PCI card it cannot automatically update or apply GPU service profiles to. That's meant as an interim step, AFAIU.
That being said, I would still check this out with your SE, to see how you will be deploying it and what are the limitations, guidelines.
