[Mesa-dev] [RFC 31/31] nir: Add a bool to float32 lowering pass

Ian Romanick idr at freedesktop.org
Tue Oct 23 00:23:52 UTC 2018


On 10/22/2018 05:09 PM, Alyssa Rosenzweig wrote:
>> Right.  Sorry.  I forgot who's working on what chip these days.  Maybe Ilia
>> can CC the right person.
> 
> No worries at all! It may well have been me (I had touched the b2f code;
> our hardware lacks a dedicated instruction and needs `iand 1.0` to do
> the conversion). I'm not sure which hardware lacks native integers these
> days.

It does have an instruction.  That instruction is just 'iand 1.0'. :)
This is what we do in i965.  I mentioned this in my review of patch 14.
If you want to set me straight, that's probably the better patch to
argue it out. :)

> _______________________________________________
> mesa-dev mailing list
> mesa-dev at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/mesa-dev


More information about the mesa-dev mailing list