• John Richard@lemmy.world
    link
    fedilink
    arrow-up
    39
    arrow-down
    1
    ·
    8 months ago

    Great… now if AMD actually cares about virtualization, maybe they can stop limiting virtual GPUs to their enterprise GPUs. Damn monopolies really don’t want to see consumers have full virtualization support for Windows on Linux.

    • theshatterstone54@feddit.uk
      link
      fedilink
      arrow-up
      10
      arrow-down
      1
      ·
      8 months ago

      I’m sorry, I’m quite inexperienced when it comes to virtualisation. Can you please explain what you mean? Thanks.

      • d3Xt3r@lemmy.nzM
        link
        fedilink
        arrow-up
        40
        ·
        edit-2
        8 months ago

        I think they’re referring to SR-IOV support (Single Root Input/Output Virtualization). It’s a technology that allows a single hardware device, like a GPU, to be shared across multiple virtual machines (VMs) with minimal overhead. In short, it lets you split your GPU into smaller GPUs that you can then distribute to VMs. This, historically, has been the domain of enterprise and industrial applications, but that’s changing. With Linux gaming on the rise, and more tech enthusiasts then ever, more and more people are trying virtualization and more and more consumers feel the need for SR-IOV. Right now, only a handful of expensive, enterprise-tier AMD GPUs have SR-IOV support. I believe it’s the same situation with nVidia, but you can unlock the feature on their consumer GPUs with some third-party tool (AFAIK).

        • Mixel@feddit.de
          link
          fedilink
          arrow-up
          4
          ·
          8 months ago

          Oh yeah I heard about this and saw that mutahar (some ordinary gamers) was doing it once on windows with a 4090. I would love to do that on my GPU and then split it between my host and my VM

    • Alex@lemmy.ml
      link
      fedilink
      arrow-up
      8
      ·
      8 months ago

      They have been working on VirtIO vulkan support as well as native context support for their cards.