Anyone doing vGPU on workstations to run CUDA on VMs?

Posted on

Hi Folks,
I'm just a sysadmin, not an AI/ML developer, so I don't know much about CUDA yet. It looks like we're going to have a lot more of our developers beginning to use CUDA, so I wanted to look into isolating those workloads on their workstations to guest VMs if possible.

Most of what I see on Google discusses doing this on Openstack, RHV, or using Nvidia's own "Grid" system and I'm worried it doesn't apply to what I'm trying to do.

I'm hoping if others have already gone down this road and learned lessons or have pointers in mind, that they could post them here.

Thanks!