1

I read this howto and have the following questions:

Is this like the GPU passthrough VFIO (you need two monitors and two different graphic cards, and a motherboard which supports IOMMU feature)?

Or is possible to assign a virtual GPU to guest which can run graphic apps (games or apps needing direct card access) using only one card (of course if GVT-g is supported)?

1

Q1: Is this like the GPU passthrough VFIO (you need two monitors and two different graphic cards, and a motherboard which supports IOMMU feature)?

Looking at this slide from this video on Youtube titled: Intel GVT-g: From Production to Upstream - Zhi Wang, Intel, it looks like you have to have a motherboard that supports IOMMU.

  ss1

Also looking at this slide, it looks like you just have to have a single GPU. Regarding your last question.

Q2: Or is possible to assign a virtual GPU to guest which can run graphic apps (games or apps needing direct card access) using only one card (of course if GVT-g is supported)?

It looks like the guest OSes through QENU will use VFIO MDEV via their GFX drivers internal to them.

  ss2

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.