GPU pass through questions

I’m working on upgrading my server soon and the new CPU will have integrated graphics which i’ll only be using for BIOS configuration and troubleshooting if the system doesn’t POST properly or the network stops working, etc. I have a VM running plex server and am thinking of passing the GPU to the VM for the hardware acceleration as this is the only other use I would have for it. I’ve never played around with hardware pass through before so I’m not sure about a couple of things.

I’m assuming the GPU will still work normally during POST, is this correct?

When the GPU is made unavailable to the host system does this happen while the OS is booting or can it be done manually after the system has booted? Ideally I’d like the GPU to work as normal when the system turns on so if there is a problem I can quickly log in and sort it out, if there’s no problems I will use a script to start the VMs.

Yes.

Most people’s setups have the GPU bind to the vfio driver (which will make it stop working untill you pass it to a guest) at boot, so it will stop outputting during the boot process.

However, you can look into single GPU passthrough, where you can create a script to dynamically unbind the GPU from the normal driver so it can be passed through to a VM.

1 Like

Okay that’s great, my VMs are started by a script anyway so that should be perfect, just wasn’t sure if it was something that could be changed after the boot process or not.