Zink Is Moving Ahead In 2019 As Mesa-Based OpenGL-Over-Vulkan

[phoronix]

Phoronix: Zink Is Moving Ahead In 2019 As Mesa-Based OpenGL-Over-Vulkan

Remember Zink, that project started a few months back for implementing OpenGL over Vulkan using Mesa/Gallium3D? While there may have not been too much to report on it recently, that side project by Collabora developer Erik Faye-Lund does continue to progress and currently allows for OpenGL 3.0 to be implemented and run over the Intel and Radeon Vulkan drivers...

http://www.phoronix.com/scan.php?page=news_item&px=Zink-Early-2019-State

[boxie]

Oh that is seriously cool!

I wonder if new bring ups (e.g. panfrost) could focus on Vulkan and get a working OpenGL for free* until they decide on bringing up OpenGL themselves.


* I understand that OpenGL Performance over Zink will have a performance hit

[R41N3R]

This is an amazing project and I'm convinced it will help new drivers a lot by focusing on Vulkan only and Virgl will benefit too. I hope this will be merged soon to Mesa mainline. Personally I don't think a performance hit compared to direct OpenGL really matters as there we have now mature drivers for AMD/Intel. But long term running everything on top of just a Vulkan driver sounds great and I'm sure that even for most OpenGL games this would be sufficient! Including DX9 where we can probably use on top of Zink GalliumNine too.

[Britoid]

So I guess this would work on potentially any Vulkan driver as long as you use the Mesa OpenGL driver?

[FireBurn]

This is an amazing project and I'm convinced it will help new drivers a lot by focusing on Vulkan only and Virgl will benefit too. I hope this will be merged soon to Mesa mainline. Personally I don't think a performance hit compared to direct OpenGL really matters as there we have now mature drivers for AMD/Intel. But long term running everything on top of just a Vulkan driver sounds great and I'm sure that even for most OpenGL games this would be sufficient! Including DX9 where we can probably use on top of Zink GalliumNine too.

Gallium Nine isn't OpenGL, there would be nothing for Zink to do here

[aaahaaap]

How is gallium relevant for this? Why doesn't it translate to Vulkan and that's it?

[AsuMagic]

I wonder if this could mean less shit GL drivers on mobile devices where Vulkan is supported.
Assuming their Vulkan drivers aren't complete trash either.

[Speedator]

How is gallium relevant for this? Why doesn't it translate to Vulkan and that's it?

It probably is more easy to just wrap Gallium to Vulkan and not changing all the OpenGL-side to Vulkan like DXVK does for D3D11 and D3D10. That's why R41N3R hopes to easily get GalliumNine on top of Zink, too. Until VK9 will make significant progress, because it doesn't have this extra layer. And D3D10/11 over DXVK probably is faster than OpenGL over Zink for the same reason.

But it sounds like this is just the first step to make progress. They talk about difficulties with NIR extra layer and think about a direct GLSL to SPIR-V compiler. So direct OpenGL to Vulkan is the final goal?

[ptrwis]

Would be cool to see it running on Kazan, software Vulkan implementation

[R41N3R]

Gallium Nine isn't OpenGL, there would be nothing for Zink to do here

Quote from

https://www.collabora.com/news-and-blog/blog/2018/10/31/introducing-zink-opengl-implementation-vulkan/

"3. Enable more integration
Because Zink is implemented as a Gallium driver in Mesa, there's some interesting side-benefits that comes "for free". For instance, projects like Gallium Nine or Clover could in theory work on top of the i965 Vulkan driver through Zink. Please note that this hasn't really been tested, though.

It should also be possible to run Zink on top of a closed-source Vulkan driver, and still get proper window system integration. Not that I promote the idea of using a closed-source Vulkan driver."