All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=725473

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDgup7vqwoM


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

win-high-dpi

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8978801289506542208

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

Hi sadrul@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 : sadrul
  Commit : 94a4ca46bbe956a4ae033e7e207c584abe1e3645
  Date   : Fri May 19 17:41:00 2017
  Subject: input: Fix running the completion callback for telemetry gesture.

Bisect Details
  Configuration: winx64_high_dpi_perf_bisect
  Benchmark    : scheduler.tough_scheduling_cases
  Metric       : mean_frame_time/mean_frame_time
  Change       : 1.91% | 16.6645729167 -> 16.9832395833

Revision             Result                    N
chromium@473151      16.6646 +- 0.0453957      6      good
chromium@473198      16.6639 +- 0.0349409      6      good
chromium@473221      16.6882 +- 0.0449859      6      good
chromium@473233      16.6721 +- 0.0269089      6      good
chromium@473234      16.6762 +- 0.0407455      6      good
chromium@473235      17.0176 +- 0.0681353      6      bad       <--
chromium@473236      17.02 +- 0.12897          6      bad
chromium@473239      17.0136 +- 0.0840941      6      bad
chromium@473244      16.9832 +- 0.0885172      6      bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=release_x64 --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests scheduler.tough_scheduling_cases

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8978801289506542208

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5281266728960000


| 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!

crrev.com/472595 introduced a bug which would cause the test to terminate a little earlier than it should. That seems to have caused the 'mean frame time' metric to dip (from ~16.8 to ~16.6/16.7). The follow up CL (crrev.com/473235, referenced from the bisect above) fixes that. So the number is expected to go back up to ~16.8. Looks like the numbers went up a little too much though, to ~17.1/17.2, which now seems to have come back down to ~16.8 (since ~crrev.com/474058). So perhaps there was some other regression that has since been fixed?

WontFix?

/cc +dtapuska@, tdresser@

Yeah, WontFix.