The only feedback would be that it would be nice if patch 8 were broken down somewhat. But, the only suggestions I could think of for possible items to split out were: * intel_bufferobj_buffer: remove flag parameter * set gpu_active_start/end in a lead-up patch
Another question for patch 8. Would gpu_active_start/end be able to also handle the job of gpu_active? (Set them out of range when !gpu_active?) Doesn't seem all that important though. Series (with or without changes mentioned above): Reviewed-by: Jordan Justen <jordan.l.jus...@intel.com> On Tue, 2013-10-08 at 14:00 -0700, Eric Anholt wrote: > Since it sounds like valve won't be able to fix dota2's rendering to use > ARB_mbr soon, here's a series to add just a little bit of tracking that > works around most of the overhead of not using ARB_mbr with their > rendering pattern. 7.69854% +/- 0.909163% (n=3) fps improvement with > default settings. We could also leverage this for some apps that misuse > ARB_mbr in the future. > > This doesn't look like it affects GLB2.7, which has a very special-looking > access pattern to its BO. > > Code is also on the "subdata" branch of my tree. > > _______________________________________________ > mesa-dev mailing list > mesa-dev@lists.freedesktop.org > http://lists.freedesktop.org/mailman/listinfo/mesa-dev _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
