Difference Between VirtIO GPU and VirGL GPU

[NextPancake401]

I only ask because the documentation says this

"virtio-gl, often named VirGL is a virtual 3D GPU for use inside VMs that can offload workloads to the host GPU without requiring special (expensive) models and drivers and neither binding the host GPU completely, allowing reuse between multiple guests and or the host.",

so it is my belief that these two are the same thing. So why is there an option for VirGL and VirtIO. Is there a difference or what makes one better than the other?

 

[janssensm]

Since nobody responded yet, my uneducated take .
The virgl option was just recently added in PVE 7.2 [1], with basic support [2].
Virgl is still maturing, and possibly will be replaced by venus [3].
Th pve admin guide is a bit confusing because it doesn´t mention the difference. Also the official QEMU docs seem incomplete.

so it is my belief that these two are the same thing

They are not the same: [4],[5].

To my understanding the virtio: 'VirtIO-GPU' option just implements the standard virtio GPU driver, without opengl host offloading support. Also without needing to install mesa libraries on the host.
And virtio-gl: 'VirGL GPU' implements the virtio gpu driver with on top virgl, adding opengl host offloading, and needing to install mesa libraries as mentioned in the pve admin doc.
VirGL GPU could be faster if your hardware supports opengl, but doesn't necessarily mean better, because virgl needs to mature. So just try them out and see for yourself what fits best.

A technical talk about different gpu virtualisation classes with a video and presentation slides[6].
Presentation start at 8:02:42.

[1] https://pve.proxmox.com/wiki/Roadmap#Proxmox_VE_7.2
[2] https://bugzilla.proxmox.com/show_bug.cgi?id=3849#c11
[3] https://git.proxmox.com/?p=pve-qemu.git;a=commit;h=799cf8c5a32ddb3bbfc742a2868c3b2ee29a4c62
[4] https://git.proxmox.com/?p=qemu-ser...5;hp=a183576e3038f9b06c3a6f23f694693386186591
[5] https://git.proxmox.com/?p=pve-mana...3;hp=a8b93d096697c8dddcd06066c12c9f4381f7017d
[6] https://www.phoronix.com/news/DRM-Native-Contexts-FPS-VM

 

[wireless]

just to be clear to me and maybe help someone else, this Virt* GPU has no relationship with gpu passthrough? I mean, If I want to passthrough the gpu to a vm, I have to use the "adding a pci device" way, or I might select a Virté GPU to have the same effect?

 

[janssensm]

See: https://pve.proxmox.com/pve-docs/pve-admin-guide.html#qm_virtual_machines_settings
Chapter 10.2.8. Display. virtio-gl.

And:
https://pve.proxmox.com/pve-docs/pve-admin-guide.html#qm_pci_passthrough

It's pretty well described there. With most important note:

But, if you pass through a device to a virtual machine, you cannot use that device anymore on the host or in any other VM.

Vir* GPU <> gpu passthrough