[Bug 89580] Implement a NIR -> vec4 pass

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=89580

--- Comment #37 from Jason Ekstrand <jason at jlekstrand.net> ---
I had a chat with Matt on IRC today and He's getting more and more convinced
that making a better register coalescing pass for vec4 is highly impractical. 
Therefore, it looks like anything we can do in NIR to make things nicer for the
backend is probably the way to do it.

I've got two series sitting on the list that both help quite a bit.  One is the
one mentioned in comment #36:

http://lists.freedesktop.org/archives/mesa-dev/2015-September/093737.html

The other I've been talking to Eduardo about already.

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/intel-3d-bugs/attachments/20150911/0cb9cc0d/attachment.html>