[Mesa-dev] [PATCH] nir: Add variants of some of the comparison simplifications.

[Jason Ekstrand]

On Wed, Jan 28, 2015 at 11:20 PM, Eric Anholt <eric at anholt.net> wrote:

> Jason Ekstrand <jason at jlekstrand.net> writes:
>
> > I've got a patch I cooked up today that generalizes this a bit.  I
> haven't
> > sent it yet because it lacks shader-db statistics.  That said, I'm fine
> > with merging this in the mean time.
>
> Not sure what generalization you're working on, but something that let
> me have "any sort of bool" as an expression node, and that "and(flt(a,
> b), flt(c, d))" was recognized as a bool would be nice for some of these
> patterns.  I was wondering if band/bor operations that were only defined
> on bool arguments would be a good idea (even though they'd codegen just
> like normal iand/ior), but then I got to thinking that this sounds like
> just a specific case of generally wanting possible-value categorization
> for SSA nodes.
>

Yeah, that's what it does.  It's an extension to the current system to
allow you to specify that a variable must match a certain type.  I just
sent the series.
--Jason
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/mesa-dev/attachments/20150129/2ed4eea7/attachment.html>