Skip to content
CachyOS

Virtio-Venus

Usecases:

Section titled “Usecases:”
  • Near-Native 3D acceleration

In operating systems that make use of libvulkan, such as Linux, BSD’s and Android, you can achieve near-native performance by translating calls from the VM to vulkan then directly to your GPU and back, without having to pass it to a VM.

Unfortunately, Virtio-Venus is not in VMM, and thus will require you to use qemu-cli. You will need to note the name you gave to your VM and where it is.

Using Virtio-Venus in qemu-cli:

Section titled “Using Virtio-Venus in qemu-cli:”

Once you install whichever VM is eligible, you can open up the terminal and create a qemu-cli prompt for your virtual machine, here’s an example of my CachyOS VM:

Terminal window
qemu-system-x86_64                                                                             \
    -enable-kvm                                                                                \
    -M q35                                                                                     \
    -smp 6                                                                                     \
    -m 12G                                                                                     \
    -cpu host                                                                                  \
    -net nic,model=virtio                                                                      \
    -net user,hostfwd=tcp::2222-:22                                                            \
    -device virtio-sound-pci,audiodev=my_audiodev -audiodev pipewire,id=my_audiodev            \
    -device virtio-vga-gl,hostmem=4G,blob=true,venus=true                                      \
    -vga none                                                                                  \
    -display gtk,gl=on,show-cursor=on                                                          \
    -usb -device usb-tablet                                                                    \
    -object memory-backend-memfd,id=mem1,size=12G                                              \
    -machine memory-backend=mem1                                                               \
    -drive if=pflash,format=raw,readonly=on,file=/usr/share/edk2/x64/OVMF_CODE.secboot.4m.fd   \
    -drive if=pflash,format=raw,file=.config/libvirt/qemu/nvram/archlinux_VARS.fd              \
    -drive file=.local/share/libvirt/images/archlinux.qcow2

To explain:

Section titled “To explain:”
  • -smp 6 means 6 cpu cores
  • -m 12G means 12GB of RAM
  • -net user,hostfwd=tcp::2222-:22 means that port 22 is being forwarded from the host to the vm as port 2222, thus if you wanted an ssh connection to the vm from some device, you’d do:
Terminal window
ssh -p 2222 vmusername@vmipaddress
  • -device virtio-vga-gl,hostmem=4G,blob=true,venus=true this is the venus driver, unfortunately not yet in virt-manager, once it is, this entire section apart from user mode will be superseded.
  • -hostmem=4G means I’m letting the vm have up to 4GB of my VRAM from my GPU.
  • -object memory-backend-memfd,id=mem1,size=12G in this section, the size is the RAM you gave it, thus 12GB in my case.
  • -drive if=pflash,format=raw,readonly=on,file=/usr/share/edk2/x64/OVMF_CODE.secboot.4m.fd this is the firmware file for the vm, this is just the path to it.
  • -drive if=pflash,format=raw,file=.config/libvirt/qemu/nvram/archlinux_VARS.fd this is the nvram of the vm, without this or with this set to read-only, the vm won’t run.
  • -drive file=.local/share/libvirt/images/archlinux.qcow2 this is the path to the .qcow2 file (the storage of the vm).
    (You can see my audio device isn’t ich9 in this case but pipewire, this is irrelevant.)