All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=766522

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=a58b372473e810a7bc3112565b038a50e87d1c40281eb953cdb6c2ffcecaa172


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

android-nexus5

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8968031314618311488

=== Auto-CCing suspected CL author reveman@chromium.org ===

Hi reveman@chromium.org, 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 : David Reveman
  Commit : 0d133634f10cc828d8f7d45dbeef628e94d908d1
  Date   : Tue Sep 12 17:05:40 2017
  Subject: cc: Add trilinear filtering support.

Bisect Details
  Configuration: android_nexus5_perf_bisect
  Benchmark    : blink_perf.paint
  Metric       : large-table-background-change-with-zero-width-collapsed-borders/large-table-background-change-with-zero-width-collapsed-borders
  Change       : 7.53% | 926.326166667 -> 996.105583333

Revision             Result                  N
chromium@501224      926.326 +- 17.0509      6      good
chromium@501294      944.382 +- 30.2524      6      good
chromium@501312      967.689 +- 8.00332      6      good
chromium@501317      930.734 +- 16.7083      6      good
chromium@501318      1156.94 +- 15.1444      6      bad       <--
chromium@501319      1156.97 +- 16.3168      6      bad
chromium@501321      1116.27 +- 10.6574      6      bad
chromium@501329      1106.13 +- 13.5542      6      bad
chromium@501364      996.106 +- 17.79        6      bad

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

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

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8968031314618311488


For feedback, file a bug with component Speed>Bisection

This change should not be able to cause a performance regression as it introduces a feature that needs to be explicitly enabled and we have not yet done that on any platform. It's possible that it contains a bug that is causing it to be enable by accident. I'll investigate next week when not traveling.

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8967945961063892192

Lets try another bisect to double check this.

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

Suspected Commit
  Author : David Reveman
  Commit : 0d133634f10cc828d8f7d45dbeef628e94d908d1
  Date   : Tue Sep 12 17:05:40 2017
  Subject: cc: Add trilinear filtering support.

Bisect Details
  Configuration: android_nexus5_perf_bisect
  Benchmark    : blink_perf.paint
  Metric       : large-table-background-change-with-zero-width-collapsed-borders/large-table-background-change-with-zero-width-collapsed-borders
  Change       : 6.60% | 973.340833333 -> 1037.60466667

Revision             Result                  N
chromium@501000      973.341 +- 24.1787      6      good
chromium@501200      920.87 +- 25.7825       6      good
chromium@501300      942.851 +- 9.29668      6      good
chromium@501313      931.01 +- 25.5324       6      good
chromium@501316      935.549 +- 31.5838      6      good
chromium@501317      937.696 +- 18.8053      6      good
chromium@501318      1157.05 +- 13.8349      6      bad       <--
chromium@501319      1152.75 +- 14.7881      6      bad
chromium@501325      1107.49 +- 20.7548      6      bad
chromium@501350      1018.37 +- 14.7935      6      bad
chromium@501400      1037.6 +- 17.3785       6      bad

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

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

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8967945961063892192


For feedback, file a bug with component Speed>Bisection

reveman: Any update here? Adding blink_perf.paint owner wangxianzhu to weigh in on how important it is to look into this regression from September.

If android-nexus5 was the only bot where the regression happened, the regression is not important.

it also doesn't make much sense as this was never used on android. only on chromeos. there's a lot of noise in the graph and it seems like we've recovered and gone back the bad a couple of times since. I don't think there's much that can be done here if we haven't seen a similar regression on other devices.