All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=750148

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=59e6d81a43280ebe19237bb3cc570a1ef7e3e3bf7f0c4b769a30247049fca395


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

android-nexus5X
android-nexus6
android-nexus7v2
android-one
android-webview-nexus5X
linux-release

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8972813341565078416

=== BISECT JOB RESULTS ===
Bisect was unable to run to completion

Error: INFRA_FAILURE

The bisect was able to narrow the range, you can try running with:
  good_revision: 3d467dbf951242c8d10982c79362bdf7bebca1ad
  bad_revision : 620529aba9b0f9fa1ef7704b03af4ae2349e1216

If failures persist contact the team (see below) and report the error.


Bisect Details
  Configuration: android_nexus7_perf_bisect
  Benchmark    : blink_perf.css
  Metric       : Document::rebuildLayoutTree/Document::rebuildLayoutTree

Revision             Result                       N
chromium@489077      0.0407667 +- 0.00313647      6      good
chromium@489123      0.0420228 +- 0.00248855      6      good
chromium@489153      0.0419175 +- 0.00217716      6      good
chromium@489171      0.0418474 +- 0.00193608      6      good
chromium@489174      6.10318 +- 0.104728          6      bad
chromium@489176      6.0818 +- 0.0791164          6      bad
chromium@489180      6.0142 +- 0.102923           6      bad
chromium@489189      5.77054 +- 0.0311886         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.css

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

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8972813341565078416


For feedback, file a bug with component Speed>Bisection

The graphs for this bug improves quite a bit later on, but it's likely because [1] reduces the DOM tree sizes for the relevant tests.

[1] https://chromium.googlesource.com/chromium/src/+/744f09725b57a1d3e0640c300045946756d0774d

I'm not sure I read the graph data correctly, but it seems the regressed graphs actually is an accumulation of Document::RebuildLayoutTree times for all of blink_perf.css. If that's correct, the introduction of a new test[1] in the regression range is probably the "cause".

Leaving to shend@ to mark as wontfix if appropriate.

[1] https://chromium.googlesource.com/chromium/src/+/a65c9b380ae231c5b23ac16573e593260ca575ed

Yep, looking at the range the bisect in #3 narrowed to, it is one of the following:


r489172 "Add bootstrap loading perf tests"
r489173 "Remove dead code"
r489174 "Fix a bad interaction between print preview and the PDF plugin"

The other 2 Cls look very unrelated.

Thanks rune, yeah looks like my patch. Marking as Wont-Fix since we're just adding more tests.

nainar, would this also be the reason why your remove style sharing patch was being blamed for perf regressions?