Hardware:SandyBridge the score of running "glmark2 --off-screen": mesa 10.2: 1412 mesa 10.3 or master: 979 Through bisect, I found the bad commit is: commit 2f28a0dc23165123cf1e8b5942acad37878edd8a Author: Kristian Høgsberg <krh@bitplanet.net> Date: Mon Jul 7 16:44:58 2014 -0700 i965: Implement fast color clears using meta operations I did the following test to confirm this regression is introduced by commit 2f28a0d 1) git checkout 2f28a0d git revert 2f28a0d the score is 1412 2) git checkout 2f28a0d git am 14c1a2a9 the score is 1052
I think this was fixed by this commit: commit f95733ddb7fff0af923fce3a07ebef78fa3139a4 Author: Kenneth Graunke <kenneth@whitecape.org> Date: Fri Jan 9 20:46:17 2015 -0800 i965: Respect the no_8 flag on Gen6, not just Gen7+. When doing repclears, we only want to use the SIMD16 program, not the SIMD8 one. Kristian added this to the Gen7+ code, but apparently we missed it in the Gen6 code. This patch copies that code over. Approximately doubles the performance in a clear microbenchmark from mesa-demos (clearspd -width 500 -height 500 +color) on Sandybridge. Cc: "10.4 10.3" <mesa-stable@lists.freedesktop.org> Signed-off-by: Kenneth Graunke <kenneth@whitecape.org> Reviewed-by: Kristian Høgsberg <krh@bitplanet.net> Reviewed-by: Matt Turner <mattst88@gmail.com> References: https://code.google.com/p/chrome-os-partner/issues/detail?id=34681
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.