All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=738166

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDg7reA8wgM


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

android-webview-nexus5X

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8975418930336352208

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

Hi fmalita@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 : Florin Malita
  Commit : 84a463f08a43176366b7643ed485d42894b61006
  Date   : Wed Jun 28 15:41:40 2017
  Subject: Remove SK_SUPPORT_LEGACY_COLORFILTER_FILTERSPAN

Bisect Details
  Configuration: android_webview_arm64_aosp_perf_bisect
  Benchmark    : smoothness.tough_filters_cases
  Metric       : frame_times/Analog Clock SVG
  Change       : 104.16% | 114.762057607 -> 234.296090454

Revision             Result                  N
chromium@482970      114.762 +- 108.09       9       good
chromium@482986      139.571 +- 89.6305      14      good
chromium@482994      78.1668 +- 12.278       6       good
chromium@482998      96.2079 +- 76.7475      9       good
chromium@482999      108.43 +- 140.251       14      good
chromium@483000      180.117 +- 185.638      9       bad       <--
chromium@483002      234.296 +- 307.076      9       bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=android-webview --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=Analog.Clock.SVG smoothness.tough_filters_cases

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

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8975418930336352208


For feedback, file a bug with component Speed>Bisection

Talking to Mike, we don't expect that change to have such an effect on Android filters (it's CPU only).

Also, the numbers in the bisect are quite noisy.  Looking at the e.g. android_one/Analog_Clock.svg graph, one can notice that there's an extra data point between the baseline and the peak revision linked to this bug.  That intermediate/extra point appears to be the actual regression rev, and it correlates with the regression for other tests (Filter_Terrain.svg being the most obvious): https://chromeperf.appspot.com/report?sid=a07e45223c8d5d7947e4430f8215095059b91581e07414120d6455204a1f7485


So the actual regression range appears to be http://test-results.appspot.com/revision_range?start=482882&end=482941, with the obvious suspect https://chromium-review.googlesource.com/c/551092/.