How does GPU passthrough work?

R

revvr

New Member
Sep 15, 2023
23
2
3
I'm toying with the idea of spinning up a second node in my network. I have an older PC that I used to use as my main one. It has and AMD 3900X, RTX 2070S, and 64 GB RAM. The 3900X does not have onboard graphics.

I currently run OMV in a VM, which hosts the bulk of my storage and applications (Jellyfin, etc.) and I would like to use the 2070S for transcoding. At the same time, I also run a VM with Ubuntu desktop that I would love to get some acceleration for it so that it feels less laggy. I have other VMs (Home Assistant, Ubuntu Server, etc.) present in the system as well.

My understading is that I will not be able to use the 2070S for both VMs at the same time. Is this correct?

What is the recommended setup in this situation?
 
Last edited:
For those who Google, this seems to be the way! Works great and is a substantial change over what I had on the less powerful machine.

Media that took a while to load while scrubbing is now almost instant.

However, is there a way to check if this is using the GPU for sure to transcode? I can't tell if the better results are from having a significantly faster processor or if the GPU is actually doing the helping.

I have a very big video file, the World Cup final, in 4K at an insane bitrate. The CPU was spiking to 60%-ish.
 
You must log in or register to reply here.
Top Bottom