All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=719572

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDgys7Q7woM


Bot(s) for this bug's original alert(s):

chromium-rel-mac12

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8980146084180535456

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8980146077414070384

=== Auto-CCing suspected CL author bsalomon@google.com ===

Hi bsalomon@google.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 : bsalomon
  Commit : 96a53c163a05f4d0f6ce201d685a7f335fb3de99
  Date   : Wed May 03 17:24:33 2017
  Subject: Update Skia GL bindings to ES3

Bisect Details
  Configuration: mac_10_12_perf_bisect
  Benchmark    : blink_perf.canvas
  Metric       : draw-dynamic-canvas-2d-to-hw-accelerated-canvas-2d/draw-dynamic-canvas-2d-to-hw-accelerated-canvas-2d
  Change       : 30.09% | 28105.0657256 -> 19647.811032

Revision             Result                  N
chromium@468953      28105.1 +- 1919.59      6      good
chromium@469003      28599.5 +- 4382.53      6      good
chromium@469016      27732.0 +- 4401.52      6      good
chromium@469017      28478.5 +- 3394.93      6      good
chromium@469018      18268.8 +- 2947.66      6      bad       <--
chromium@469019      19155.6 +- 2166.31      6      bad
chromium@469022      20335.4 +- 3364.87      6      bad
chromium@469028      18807.2 +- 4063.14      6      bad
chromium@469053      19647.8 +- 5484.83      6      bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests blink_perf.canvas

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8980146084180535456

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5290711596924928


| O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq
|  X  | for more information addressing perf regression bugs. For feedback,
| / \ | file a bug with component Speed>Bisection.  Thank you!

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

Suspected Commit
  Author : bsalomon
  Commit : 96a53c163a05f4d0f6ce201d685a7f335fb3de99
  Date   : Wed May 03 17:24:33 2017
  Subject: Update Skia GL bindings to ES3

Bisect Details
  Configuration: mac_10_12_perf_bisect
  Benchmark    : blink_perf.canvas
  Metric       : draw-dynamic-canvas-2d-to-hw-accelerated-canvas-2d/draw-dynamic-canvas-2d-to-hw-accelerated-canvas-2d
  Change       : 30.67% | 28070.6312782 -> 19460.0357482

Revision             Result                  N
chromium@468953      28070.6 +- 2942.74      6      good
chromium@469003      28856.5 +- 4413.19      6      good
chromium@469016      28405.2 +- 4292.37      6      good
chromium@469017      29154.4 +- 2197.37      6      good
chromium@469018      19453.8 +- 1778.85      6      bad       <--
chromium@469019      19077.2 +- 2661.7       6      bad
chromium@469022      19964.2 +- 3522.04      6      bad
chromium@469028      19944.3 +- 2337.21      6      bad
chromium@469053      19460.0 +- 3245.94      6      bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests blink_perf.canvas

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8980146077414070384

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=4960183261855744


| O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq
|  X  | for more information addressing perf regression bugs. For feedback,
| / \ | file a bug with component Speed>Bisection.  Thank you!

I've tried repro'ing this on a couple machines with no luck. I want to keep investigating but given the upcoming branch point I think it makes sense to revert the part of the change that gave Skia access to ES 3.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/8d52fb787ae3f370ba1bf7be4b15ae01287879ac

commit 8d52fb787ae3f370ba1bf7be4b15ae01287879ac
Author: bsalomon <bsalomon@google.com>
Date: Mon May 22 19:58:40 2017

Relimit Skia to ES2 to address performance regression

BUG= 719572 
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.android:android_optional_gpu_tests_rel;master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/2895153004
Cr-Commit-Position: refs/heads/master@{#473666}

[modify] https://crrev.com/8d52fb787ae3f370ba1bf7be4b15ae01287879ac/gpu/skia_bindings/gl_bindings_skia_cmd_buffer.cc

Marking fixed since the revert landed. Please reopen if I'm mistaken.