[Mesa-dev] V5 Loop unrolling in NIR

[Timothy Arceri]

I forgot to add that Jenkins is saying the following piglit test is
fixed by this series in IVY and HSW. I assume this is due to patch 2

piglit.spec.arb_tessellation_shader.execution.variable-indexing.tes-
both-input-array-vec4-index-rd

On Tue, 2016-10-11 at 09:38 +1100, Timothy Arceri wrote:
> Again thanks to Jason and Connor for their feedback, and to 
> Thomas for getting this work started.
> 
> V5: Has been updated to reflect Jasons feedback and also some
> changes from self review after not looking at this for a few weeks.
> 
> total instructions in shared programs: 8530600 -> 8530365 (-0.00%)
> instructions in affected programs: 46967 -> 46732 (-0.50%)
> helped: 62
> HURT: 92
> 
> total cycles in shared programs: 67279552 -> 67251326 (-0.04%)
> cycles in affected programs: 2548608 -> 2520382 (-1.11%)
> helped: 925
> HURT: 909
> 
> total loops in shared programs: 1868 -> 1214 (-35.01%)
> loops in affected programs: 687 -> 33 (-95.20%)
> helped: 655
> HURT: 17
> 
> LOST:   5
> GAINED: 0
> 
> Most of the HURT instruction count is caused by patch 1. See the
> commit message for more details.

  ^ This should be patch 2.

> 
> The HURT loops are caused by a couple of different things that can be
> addressed in a follow up series. For example handling if statements
> that contain a break but also other instructions.
> 
> _______________________________________________
> mesa-dev mailing list
> mesa-dev at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/mesa-dev