Comment by sds357

I have a Proxmox machine with dual P40 gpus that pass thru to a vm with ollama running for local inference. Been working great for over 2 years.