Intel Has More DRM Feature Code Ready For Linux 4.10, GVT Going Mainline

[phoronix]

Phoronix: Intel Has More DRM Feature Code Ready For Linux 4.10, GVT Going Mainline

Intel had already sent in two feature pull requests of new features/changes improvements to their DRM driver for Linux 4.10 (pull requests one and two) while now another feature pull has been submitted and there's also expected to be a last-minute fourth pull request...

http://www.phoronix.com/scan.php?page=news_item&px=Linux-4.10-Intel-DRM-3

[darkbasic]

What's the difference between GVT and Virgil3d?

[sirblackheart]

What's the difference between GVT and Virgil3d?

As far as I can tell GVT is intel specific and allows gpu passtrough which is stated to be out of scope on the virgil3d homepage.

[darkbasic]

allows gpu passtrough which is stated to be out of scope on the virgil3d homepage.

"multiple VMs to the same GPU" doesn't sound like gpu passtrough.

[kobblestown]

"multiple VMs to the same GPU" doesn't sound like gpu passtrough.

It's virtual passthrough So while it still needs GPU-specific support at the VM host side it exposes a regular iGPU to the guest. The guest is using the regular Intel drivers that access the hardware "directly" and is unaware that it's sharing the GPU with other VMs. At least in theory.

I've been waiting for this to become mainstream for a couple of years already. I hope it doesn't take as long as Wayland because I really want to use it.

[sirblackheart]

"multiple VMs to the same GPU" doesn't sound like gpu passtrough.

• Intel® Graphics Virtualization Technology –d (Intel® GVT –d): vDGA: virtual dedicated graphics acceleration (one VM to one physical GPU)
• Intel® Graphics Virtualization Technology –s (Intel® GVT -s): vSGA: virtual shared graphics acceleration (multiple VMs to one physical GPU)
• Intel® Graphics Virtualization Technology –g (Intel® GVT -g): vGPU: virtual graphics processing unit (multiple VMs to one physical GPU)

In response to this challenge, graphics virtualization techniques have evolved to allow a media-optimized workload to run on top of a virtualized environment, such as virtualizing the graphics processor for concurrent use to provide direct, dedicated access to a GPU for a single virtualized workload, or providing a pass-through for direct, shared access to the GPU for a number of virtualized workloads.

Sounds like a kind of pass-trough to me.

[pal666]

What's the difference between GVT and Virgil3d?

gvt is for intel gpus, virgil is for opengl gpus

[darkbasic]

It's virtual passthrough So while it still needs GPU-specific support at the VM host side it exposes a regular iGPU to the guest. The guest is using the regular Intel drivers that access the hardware "directly" and is unaware that it's sharing the GPU with other VMs. At least in theory.

I've been waiting for this to become mainstream for a couple of years already. I hope it doesn't take as long as Wayland because I really want to use it.

Wow, this is awesome. Does it work when sharing the same GPU between the host and the guests?

[andre30correia]

Wow, this is awesome. Does it work when sharing the same GPU between the host and the guests?

yes, one igpu for a lot of vm and host

[anto44]

GVT-g is like a GPU passthrough: original GPU controller and sharing between Host and VM (up to 3 for Haswell, more for newers), about 85% max performance when one machine uses the GPU and no graphics-heavy use for the others.