[Mesa-dev] [PATCH 5/5] nir: Lower flrp differently when the alpha value is reused

[Ian Romanick]

On 08/23/2018 06:59 PM, Eric Anholt wrote:
> Ian Romanick <idr at freedesktop.org> writes:
> 
>> From: Ian Romanick <ian.d.romanick at intel.com>
>>
>> For some reason, if I did not move the regular lowering to late
>> optimizations, the new lowering would never trigger.  This also means
>> that the fsub lowering had to be added to late optimizations, and this
>> requires "intel/compiler: Repeat nir_opt_algebraic_late until no more
>> progress".
>>
>> The loops removed by this patch are the same loops added by
>> "intel/compiler: Don't emit flrp for Gen4 or Gen5"
>>
>> I am CC'ing people who are responsible for drivers that set lower_flrp32
>> as this patch will likely affect shader-db results for those drivers.
>>
>> No changes on any Gen6+ platform.
> 
> No change on vc4 in the previous patch, but this patch seems to cause
> flrps to be left in my NIR, so a bunch of traces crash.

Yeah... it will break everyone except Intel.  I moved the lowering to a
algebraic pass that nobody else uses. :(

I have a new series that fixes this issue and is better overall.  I'm
hoping to get that out later today.