Seamless GPU passthrough on Qubes OS with VirtualGL revisions

Go back to topic: Seamless GPU passthrough on Qubes OS with VirtualGL

  1. v6 anchor; v6 full version
  2. v5 anchor; v5 full version
  3. v4 anchor; v4 full version
  4. v3 anchor; v3 full version
  5. v2 anchor; v2 full version

Revision #6

Edited on
2024-12-09
Edited by user
otter2
Shutdown the template. Create a new VM from it. Set virtualization mode to HVM and kernel to Qube provided. Attach secondary graphics card to it. Boot, and execute **vglrun** with a command to get accelerated graphics. Wrapping a shell also works (**vglrun bash**) to avoid having to type it every time, but there is no obvious way to automate this at startup. Shutdown the template. Create a new VM from it. Set virtualization mode to HVM and kernel to Qube provided. Attach secondary graphics card to it. Boot, and execute **vglrun** with a command to get accelerated graphics. Wrapping a shell also works (**vglrun bash**) to avoid having to type it every time, ~~but there is no obvious way to automate this at startup.~~ From https://wiki.archlinux.org/title/VirtualGL#Running_applications : > **Tip:** `vglrun` is actually just a shell script that (temporarily) sets some environment variables before running the requested application - most importantly it adds the libraries that provide all the VirtualGL functionality to `LD_PRELOAD`. If it better suits your workflow, you could just set those variables yourself instead. The following command lists all environment variables that vglrun would set for your particular set-up: > > > comm -1 -3 <(env | sort) <(vglrun env | grep -v '^\[' | sort)

Revision #5

Edited on
2024-02-22
Edited by user
bamboodabpen
Monero/XMR:87ZPgPnYTdvMnVpdhvoNhvjQ4AvnEPZvQP3XmULGxpjTJCTtpnB9cgmYoRe8Bj6Dh9867CUn67yqQbCU2jwCVeykPjEdMu9We are no longer accepting donations, and have moved on to other projects. We will not be able to adequately respond to further questions. The follow up information provided by users appears to be sufficient for now.

Revision #4

Edited on
2024-02-22
Edited by user
bamboodabpen

Revision #3

Edited on
2023-08-30
Edited by user
bamboodabpen
Device compatibility: Nvidia: Works but lower FPS reported on Nouveau/Mesa driver, proprietary driver is fine. AMD: RadeonSI Mesa driver has issues with EGL on the tested card, but nothing should be an endemic problem with the brand. It is assumed propietary OpenGL will work but it remains untested due to difficulties in patching Ubuntu driver for Debian template. Alternatively, GLX offloading to a dedicated X server works. Conflicts occurred having Nvidia drivers installed with the AMD card, VirtualGL reports no EGL devices while they are installed. A simple fix is possibly preloading the proper libraries. **Device compatibility:** **Nvidia:** Works but lower FPS reported on Nouveau/Mesa driver, proprietary driver is fine. **AMD:** Working with Mesa after disabling MSI in template kernel command line with parameter **pci=nomsi**. Conflicts occurred having Nvidia drivers installed with the AMD card, VirtualGL reports no EGL devices while they are installed. A simple fix is possibly preloading the proper libraries.
Instructions: **Instructions:**
Donations: **Donations:**

Revision #2

Edited on
2023-08-30
Edited by user
bamboodabpen