All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=779213

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=ee2cfcf8a1990e342e50df8e2cec9197a6033e2398feb60e9230f51acc1d8672


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

chromium-rel-win7-dual

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8964547491278617952

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8964547204611248848

I've previously tried bisecting this in  issue 772447  but the issue doesn't repro on the win_perf bisect bots - I've kicked off another bisect just in case. Assigning to Tim as test owner. Do you know what the difference might be between the dual bots and the bisect ones? The regression only seems to apear in win7-dual.

=== BISECT JOB RESULTS ===
NO Perf regression found

Bisect Details
  Configuration: win_perf_bisect
  Benchmark    : thread_times.tough_scrolling_cases
  Metric       : thread_raster_cpu_time_per_frame/text_hover_50000_pixels_per_second

Revision             Result                  N
chromium@506716      8.32853 +- 1.86679      21      good
chromium@506801      8.14748 +- 2.06003      21      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 --story-filter=text.hover.50000.pixels.per.second thread_times.tough_scrolling_cases

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

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8964547491278617952


For feedback, file a bug with component Speed>Bisection

=== BISECT JOB RESULTS ===
Perf regression found but unable to narrow commit range

Build failures prevented the bisect from narrowing the range further.


Bisect Details
  Configuration: win_perf_bisect
  Benchmark    : thread_times.tough_scrolling_cases
  Metric       : thread_raster_cpu_time_per_frame/text_hover_20000_pixels_per_second
  Change       : 16.05% | 2.21107517002 -> 1.94998170474

Suspected Commit Range
  2 commits in range
  https://chromium.googlesource.com/chromium/src/+log/771bc4ee1d609083dd81daa0ee78229b49dfa270..016b6014b2927367bddb752707726bfef36ba9a0


Revision             Result                   N
chromium@506801      2.21108 +- 1.11388       14       good
chromium@506804      1.97148 +- 0.402512      14       good
chromium@506805      ---                      ---      build failure
chromium@506806      1.81286 +- 0.39564       14       bad
chromium@506812      1.86873 +- 0.115318      6        bad
chromium@506825      1.91494 +- 0.514294      14       bad
chromium@506848      1.81151 +- 0.243834      6        bad
chromium@506896      1.94998 +- 0.947576      14       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 --story-filter=text.hover.20000.pixels.per.second thread_times.tough_scrolling_cases

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

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8964547204611248848


For feedback, file a bug with component Speed>Bisection

Presumably win-dual is dual core?

This seems like a problem though, if bisect bots aren't equivalent to the bots doing CI.

+Annie, are we understanding this correctly?

David, this regression is probably your change here, right?

https://chromium.googlesource.com/external/github.com/catapult-project/catapult/+/dbab215523c8d77a7b54fcf3776357b079c36899

Whoops, I got that change confused with https://chromium.googlesource.com/chromium/src/+/771bc4ee1d609083dd81daa0ee78229b49dfa270.

Is there any chance it's the viewport co-ordinate change?

It happened around the same time which is why it got bundled up with the hundredes of other graphs that "regressed" in  issue 772447  but the gesture changes are related to how we deal with page scale. Desktops don't load zoomed out so unless the benchmark does some zooming (this one doesn't) there should be no effect. I've also tried it locally on my machine with my changes reverted and there was no difference so I don't think it's related to my change.

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8964201816990845968

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8964201805388961488

=== BISECT JOB RESULTS ===
NO Perf regression found

Bisect Details
  Configuration: win_perf_bisect
  Benchmark    : thread_times.tough_scrolling_cases
  Metric       : thread_raster_cpu_time_per_frame/text_hover_10000_pixels_per_second

Revision             Result                    N
chromium@506801      0.780638 +- 0.403416      21      good
chromium@506896      0.719103 +- 0.334675      21      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 --story-filter=text.hover.10000.pixels.per.second thread_times.tough_scrolling_cases

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

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8964201805388961488


For feedback, file a bug with component Speed>Bisection

=== BISECT JOB RESULTS ===
NO Perf regression found

Bisect Details
  Configuration: win_perf_bisect
  Benchmark    : thread_times.tough_scrolling_cases
  Metric       : thread_raster_cpu_time_per_frame/text_hover_20000_pixels_per_second

Revision             Result                   N
chromium@506801      2.01341 +- 0.905647      21      good
chromium@506896      2.02235 +- 0.829213      21      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 --story-filter=text.hover.20000.pixels.per.second thread_times.tough_scrolling_cases

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

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8964201816990845968


For feedback, file a bug with component Speed>Bisection

I'm not 100% clear on who ends up triaging these at this point.
I think it's folks on Annie's team?

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/1297efcf040000

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/12fb1a4f040000

Kicking a few more bisects now that pinpoint is available for Windows, to see if we can get a clearer understanding of what's going on.

📍 Couldn't reproduce a difference.
https://pinpoint-dot-chromeperf.appspot.com/job/12fb1a4f040000

Pinpoint shows bimodal values, looks like this went from low to high? Doesn't look like we'll be able to get to the bottom of it.

📍 Couldn't reproduce a difference.
https://pinpoint-dot-chromeperf.appspot.com/job/1297efcf040000