NIR Experimental Backend For RadeonSI Updated

[phoronix]

Phoronix: NIR Experimental Backend For RadeonSI Updated

Nicolai Hähnle has published his updated massive patch-set for implementing an external NIR back-end in RadeonSI...

http://www.phoronix.com/scan.php?page=news_item&px=RadeonSI-NIR-V2

[FireBurn]

Exciting stuff, hopefully if this goes well Intel might consider switching to Gallium as a lot of the plumbing is already in place for it

[gnarlin]

If Intel switched wouldn't that free up a whole lot of old infrastructure code that could be deleted?

[frosth]

I don't think so. Intel leaves gallium driver years ago.

[frosth]

Edit...left gallium....
and there is no advantages, on one hand Intel classic driver works good and gallium driver will be cost too much effort on the other

[boxie]

Edit...left gallium....
and there is no advantages, on one hand Intel classic driver works good and gallium driver will be cost too much effort on the other

yeah - they have said in the past that there is a whole chunk of work to move over and it is not worth their time.

though - come to think about it, some of those reasons (e.g. 2d acceleration paths for X) are no longer really a cause for holding back.

[darkbasic]

There is no reason to switch to gallium now that you can implement OpenGL on top of Vulkan which is far more standard.

[smitty3268]

There is no reason to switch to gallium now that you can implement OpenGL on top of Vulkan which is far more standard.

Other than the fact that it would take years of development to build such a stack, while Gallium is already done...

Though more to the point, I'm quite certain intel doesn't see any reason to switch away from their current working driver stack to use Gallium either - so it's not going to happen either way.

[darkbasic]

Other than the fact that it would take years of development to build such a stack, while Gallium is already done...

Switching to Gallium would take a couple of years too for Intel.

[duby229]

Other than the fact that it would take years of development to build such a stack, while Gallium is already done...

Though more to the point, I'm quite certain intel doesn't see any reason to switch away from their current working driver stack to use Gallium either - so it's not going to happen either way.

I think I remember initially Gallium was developed with an experimental Intel driver. It was really I think just a driver used to implement the concepts and infrastructure. AMD came on board after it was basically working, advanced it considerably and added their range of hardware support.

So I guess I'm just wondering if Intel as a member of Kronos, did they have some conception of GLnext or Vulkan back in those days? In after thought, they had to have known something else was coming in just a few years.