Intel Xe MAX Needs Two Linux Kernels For Now - Meaning You Need To Use A GPU-Accelerated VM

Back in October Intel announced Iris Xe MAX as discrete graphics for laptops. The overall Linux state for Xe MAX hasn't been too clear and we haven't had any hardware access to this Intel laptop discrete graphics hardware to report our own findings, but their developers have now cleared up the situation. The good news is the Xe MAX graphics can be used for a GPU-accelerated Linux virtual machine. The bad news is the Xe MAX support doesn't yet allow for dGPU usage by the host outside of a virtual machine context as it needs "two different [Linux] kernels" for operation in conjunction with the integrated graphics.

Xe MAX launched over a month ago but for now Intel Linux users will want to avoid the hardware unless wanting to resort to a lengthy setup process that involves making use of KVM virtualization to enjoy the discrete graphics while still being able to make use of a laptop's integrated graphics.

Intel's new documentation concerning Iris Xe MAX Graphics for Linux characterize the MAX state as "ongoing", unlike the great Gen12 / Xe-LP Tiger Lake support currently found on Linux as we have shared in our numerous and continuing benchmarks from the Dell XPS 13 9310 with the Core i7 1165G7 Tiger Lake processor.

Here is the interesting summary of the Xe MAX Linux support in that two kernels are currently needed. "To use both the Intel Iris Xe graphics and Intel Iris Xe MAX graphics processors at the same time currently requires two different kernels. To isolate the two required kernel versions, virtualization will be used. The virtual machine (VM) host will have direct control of the display through the Intel Iris Xe graphics processor (via the kernel provided by Ubuntu.) The VM host will also be configured to use PCI passthrough to provide the Intel Iris Xe MAX graphics processor to the VM guest. The VM guest will be running the custom Linux kernel."

Basically the Xe MAX graphics will use PCI pass-through to the guest KVM-based virtual machine and from there the graphics stack can be setup. Outside of the VM context, the integrated Xe Graphics of the laptop will be powering the host desktop / graphics stack short of routing your graphics/compute back through the virtualized stack for executing on Xe MAX.

Intel Xe MAX Graphics are great right now if you want to dedicate a discrete GPU to a virtual machine, but outside of that and the numerous steps involved with the setup, it's not an easy path for those simply wanting a nice Linux laptop with discrete Intel graphics. This also appears to only work with a Linux VM and no mention of Windows virtual machine support, which would be interesting for some with the use-case of having a VM for Windows gaming with a dedicated GPU atop Linux.

Intel's steps for setting up the Xe MAX Graphics on Linux by means of the KVM setup are outlined on dgpu-docs.intel.com. The steps are quite involved so long story short you'll want to wait for the Intel Xe MAX Linux support to further mature and where it works off a single kernel if wanting to avoid the virtualization setup. Again, not to be confused with integrated Xe-LP graphics on the likes of Tiger Lake that work out great on Linux without any fuss -- this is just about Xe MAX being in a primitive state for now on Linux. For close to two decades Intel is normally very punctual with their Linux graphics support being in great shape for time of launch, but their discrete graphics enablement has been slow and thus Xe MAX is an outlier for now but should improve with time.