Talk:Xorg multiseat

From ArchWiki

Hello. Thank you for this page.

  • Could you explain the differences if one uses one video card? Some computers have just one PCI-e slot.
  • We (Fiable.biz) use Mandriva Linux 2009 Spring, and there is no ~/.xinitrc .
  • What is the equivalent of "twm" to launch kde?
  • Where do you launch startx?
--Fiable.biz 11:02, 3 October 2009 (EDT)
  • We can use one video card, but it's a more complicated setup. Xephyr is not necessary but I haven't successfully done this in Linux yet. I do know somebody who sells computers configured this way. I think he uses multiple X-servers but they are separate so they can run graphics acceleration and GL extensions, etc. I'd like to investigate this more and if anybody could contribute it would help.
--Shillshocked 20:31, 13 February 2014 (EDT)

Single card multiseat with Xephyr

The article mentions that is possible to setup multiseat on a single video card with multiple displays using Xephyr. Could someone describe how to do this? I think there would be a lot of people like myself who only have a single videocard with both DVI and HDMI ports. It would be great to be able to set-up the HDMI port as an independent multimedia station connected to a TV.

—This unsigned comment is by Wikijon (talk) 03:28, 20 May 2020‎. Please sign your posts with ~~~~!

I'm also interested in sharing a single GPU among multiple seats. Any pointers on how to do this? Is VirtualGL the way to go, or Xephyr, or is there a simpler / more modern / more flexible way? I'd try to do more research myself, it would already help if one could point me (or everyone here) to the right direction. Thanks in advance, LukeLR (talk) 23:56, 20 December 2020 (UTC)Reply[reply]
The forum is a better place for such topics. Please retry there and if your problem are solved, add your solution here. Thanks. --Fengchao (talk) 13:50, 31 January 2023 (UTC)Reply[reply]

Multiseat with one dedicated GPU and one integrated GPU

There are a few things I'd like to note from my experience when using a multiseat setup with dGPU and iGPU. Let's say you made seat0 with dGPU and iGPU is the [MASTER] device of seat1. Everything on seat1 will be rendered with iGPU, however, I couldn't find a single mention of using DRI_PRIME environment variable. This variable makes the application in that context use dGPU which is in use from seat0. This makes multiseat gaming setups feasible with a single dedicated GPU. While proprietary nvidia drivers don't use that variable, I'm sure you can change the driver with __GLX_VENDOR_LIBRARY_NAME.

Note: Both DRI_PRIME and __GLX_VENDOR_LIBRARY_NAME are only used by OpenGL applications.

Vulkan is not supported on my iGPU so I don't know what the proper way to choose dGPU would be. I couldn't find anything other than vkdevicechooserAUR. Anything that requires vulkan automatically uses my dGPU, as it's the next available device other than software renderer.

Also, would anyone know how to set an environment variable based on which seat is running? preferably something that is user independent, similarly to /etc/environment

Lassebq (talk) 14:42, 11 January 2024 (UTC)Reply[reply]