ANGLE roll:

https://chromium.googlesource.com/angle/angle.git/+log/cbcb96f..6d94f06

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8964797611158768336

=== Auto-CCing suspected CL author jiajia.qin@intel.com ===

Hi jiajia.qin@intel.com, the bisect results pointed to your CL, please take a look at the
results.


=== BISECT JOB RESULTS ===
Perf regression found with culprit

Suspected Commit
  Author : Qin Jiajia
  Commit : 7478ce7f4e074436d52c78bb704e47861495c4ab
  Date   : Fri Oct 20 15:34:40 2017
  Subject: D3D11: Minor Optimization to DrawCallNeedsTranslation

Bisect Details
  Configuration: winx64nvidia_perf_bisect
  Benchmark    : angle_perftests
  Metric       : DrawElementsPerf_gl_index_buffer_changed/score
  Change       : 49.28% | 5213.66666667 -> 2644.33333333

Revision                              Result                  N
chromium@510710                       5213.67 +- 280.309      6      good
chromium@510716                       5076.5 +- 339.811       6      good
chromium@510716,angle@111a99e82d      5165.0 +- 234.068       6      good
chromium@510716,angle@7478ce7f4e      2645.5 +- 61.5264       6      bad       <--
chromium@510716,angle@fb9051a588      2642.83 +- 61.146       6      bad
chromium@510717                       2626.0 +- 63.3561       6      bad
chromium@510718                       2634.0 +- 62.9285       6      bad
chromium@510719                       2646.33 +- 58.6629      6      bad
chromium@510721                       2644.33 +- 59.7104      6      bad

To Run This Test
  .\src\out\Release_x64\angle_perftests.exe --test-launcher-print-test-stdio=always --test-launcher-jobs=1

More information on addressing performance regressions:
  http://g.co/ChromePerformanceRegressions

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8964797611158768336


For feedback, file a bug with component Speed>Bisection

Jiajia can you try to reproduce the regression locally? I'm not sure how your change could affect the GL renderer.

I don't think it's caused by my patch. My change is only for d3d11. It won't affect the GL renderer. I compared the score of DrawElementsPerf_gl_index_buffer_changed between the current code and the current code + revert my patch. There is no big differ.

OK, thanks for checking. Indeed a huge gap.