Direct output to a display connected to the server

[CDYT]

Hi,
I'm looking into installing proxmox as the primary OS on my new PC, but I wasn't able to find any way to attach a display to the machine running the server and get the output to that display. As I want to run multiple VMs on that machine, I was wondering if I can just get the output on connected displays instead of using another device to access the VMs. Also if there's a way to project the VMs on assigned displays on startup.
Thank you.

 

[oguz]

hi,

I'm looking into installing proxmox as the primary OS on my new PC, but I wasn't able to find any way to attach a display to the machine running the server and get the output to that display

hmm, are you doing GPU passthrough inside the VMs? if that's the case then you could just connect your GPU(s) to the corresponding monitors.

also another possibility is installing X server on your PVE host [0] and using SPICE [1] to access the VMs locally (should be still good speed and usable because of low latency locally), and you wouldn't need a 2nd device to access the VMs.

hope this helps!

[0]: https://pve.proxmox.com/wiki/Developer_Workstations_with_Proxmox_VE_and_X11
[1]: https://pve.proxmox.com/wiki/SPICE

 

[CDYT]

hi,


hmm, are you doing GPU passthrough inside the VMs? if that's the case then you could just connect your GPU(s) to the corresponding monitors.

also another possibility is installing X server on your PVE host [0] and using SPICE [1] to access the VMs locally (should be still good speed and usable because of low latency locally), and you wouldn't need a 2nd device to access the VMs.

hope this helps!

[0]: https://pve.proxmox.com/wiki/Developer_Workstations_with_Proxmox_VE_and_X11
[1]: https://pve.proxmox.com/wiki/SPICE

Do I need an individual GPU for every VM if I wanna go the GPU passthrough route?

 

[Dunuin]

Do I need an individual GPU for every VM if I wanna go the GPU passthrough route?

Jup. And without GPU passthrough your VMs would be kinda slow anyway (at least if you want to be able to use 3D or watch videos) because no VM could use any GPU accelleration of a GPU in your machine that is not passed through. So without it you are basically back in the 80s where everything had to be rendered in software by the CPU.

There are some ways (SR-IOV, vGPU) to split a single GPU into multiple virtual GPUs that then could be passthed through but that is still very experimental at the moment.