Re: [Mesa-dev] Reducing GLSL compiler optimization time

2011-03-14 Thread Ian Romanick
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 03/12/2011 02:41 PM, Eric Anholt wrote: > On Sat, 12 Mar 2011 11:30:10 -0800, Kenneth Graunke > wrote: >> On Friday, March 11, 2011 04:06:44 PM Eric Anholt wrote: >>> With the conversion of texenvprogram to producing GLSL IR, all those >>> nice li

Re: [Mesa-dev] Reducing GLSL compiler optimization time

2011-03-14 Thread Ian Romanick
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 03/11/2011 04:06 PM, Eric Anholt wrote: > With the conversion of texenvprogram to producing GLSL IR, all those > nice little optimization passes end up slowing down our > combinatorial-explosion ff tests pretty significantly. This patch > series re

Re: [Mesa-dev] Reducing GLSL compiler optimization time

2011-03-12 Thread Eric Anholt
On Sat, 12 Mar 2011 11:30:10 -0800, Kenneth Graunke wrote: > On Friday, March 11, 2011 04:06:44 PM Eric Anholt wrote: > > With the conversion of texenvprogram to producing GLSL IR, all those > > nice little optimization passes end up slowing down our > > combinatorial-explosion ff tests pretty si

Re: [Mesa-dev] Reducing GLSL compiler optimization time

2011-03-12 Thread Kenneth Graunke
On Friday, March 11, 2011 04:06:44 PM Eric Anholt wrote: > With the conversion of texenvprogram to producing GLSL IR, all those > nice little optimization passes end up slowing down our > combinatorial-explosion ff tests pretty significantly. This patch > series reduces glean texCombine runtime ov

[Mesa-dev] Reducing GLSL compiler optimization time

2011-03-11 Thread Eric Anholt
With the conversion of texenvprogram to producing GLSL IR, all those nice little optimization passes end up slowing down our combinatorial-explosion ff tests pretty significantly. This patch series reduces glean texCombine runtime overall by 20% on my gen6 system. The lowest-hanging fruit at this