All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=711255

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDggvDWugoM


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

chromium-rel-mac11

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8982431423355498816

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

Hi trchen@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 : trchen
  Commit : c3593fe02146ff693ce324259b9b292bd43e4910
  Date   : Tue Apr 04 02:12:12 2017
  Subject: [2/5] Disable transformed rasterization if layer is opaque

Bisect Details
  Configuration: mac_10_11_perf_bisect
  Benchmark    : blink_perf.shadow_dom
  Metric       : SlotDistibutedNextPrevious/SlotDistibutedNextPrevious
  Change       : 20.41% | 5138053.63355 -> 4092390.69505

Revision             Result                  N
chromium@461584      5138054 +- 142040       14      good
chromium@461600      5035126 +- 1661627      21      good
chromium@461608      5129054 +- 114389       9       good
chromium@461612      5121467 +- 294935       14      good
chromium@461613      4879653 +- 2260749      21      bad       <--
chromium@461614      5041257 +- 361571       21      bad
chromium@461615      4966264 +- 1311681      14      bad
chromium@461646      4092391 +- 165344       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.shadow_dom

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8982431423355498816

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=4982914762145792


| 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!

Pretty sure the bisect is wrong. It looks like the regression range should be 461616:461646. How can I kick off another bisect?

Hey Chris, Do you think it is due to the revert of your optimization?

    Revert of Make GeometryMapper fully static (patchset #1 id:1 of https://codereview.chromium.org/2793993003/ )
    Review-Url: https://codereview.chromium.org/2798513002
    Cr-Commit-Position: refs/heads/master@{#461620}

I re-landed the patch, so no.

Re comment 4: you can use the telemetry ui to kick off a bisect. Click on the graph and you will see a bisect button.

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8982401169612793232

Bisect kicked off. Hmm, there are a bunch of land/reland happening in the range. Better experiment than guess.

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

Hi lpy@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 : lpy
  Commit : 71bcf50b0c34114c514e431af16527792c55006a
  Date   : Tue Apr 04 03:29:05 2017
  Subject: [Page Load Metrics] Add page load metrics for omnibox.

Bisect Details
  Configuration: mac_10_11_perf_bisect
  Benchmark    : blink_perf.shadow_dom
  Metric       : SlotDistibutedNextPrevious/SlotDistibutedNextPrevious
  Change       : 20.23% | 5048163.90576 -> 4007319.8234

Revision             Result                  N
chromium@461615      5048164 +- 373348       9       good
chromium@461619      4838082 +- 1734850      14      good
chromium@461620      4736717 +- 55799.7      6       good
chromium@461621      3954551 +- 805382       9       bad       <--
chromium@461623      3996993 +- 422267       9       bad
chromium@461631      4013611 +- 275261       6       bad
chromium@461646      4007320 +- 503566       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.shadow_dom

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8982401169612793232

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5041885736861696


| 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!

assigning to hayato, owner of blink_perf.shadow_dom benchmark. This is like a random number generator, are we sure this benchmark is useful?

This looks too unstable benchmark. Let me disable it.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/852fdb990ef3e2999b14608e3f7c0038762f3199

commit 852fdb990ef3e2999b14608e3f7c0038762f3199
Author: hayato <hayato@chromium.org>
Date: Fri Apr 14 06:11:53 2017

Remove unstable SlotDistibutedNextPrevious.html benchmark

The benchmark result looks too unstable, according to the bug report.
It would be better to remove it.

BUG= 711255 

Review-Url: https://codereview.chromium.org/2818923002
Cr-Commit-Position: refs/heads/master@{#464688}

[delete] https://crrev.com/2383850584645fc89acd0332c11a592e0b0fd21c/third_party/WebKit/PerformanceTests/ShadowDOM/SlotDistibutedNextPrevious.html