All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=778609

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=381805066bf6b35dbc0d25295c819cf3f8f2e9ef0fb34bcd4231c54d44490d18


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

chromium-rel-win8-dual

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8964677053804077280

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

Hi juncai@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 : Jun Cai
  Commit : 012769532c3d31254d7730d4a64613982fde241f
  Date   : Tue Oct 24 04:50:15 2017
  Subject: Fetch source sensors after PlatformSensorFusion object is fully constructed

Bisect Details
  Configuration: win_8_perf_bisect
  Benchmark    : thread_times.tough_scrolling_cases
  Metric       : thread_total_all_cpu_time_per_frame/text_constant_full_page_raster_05000_pixels_per_second
  Change       : 14.12% | 18.3870320561 -> 15.7909894819

Revision             Result                   N
chromium@510942      18.387 +- 0.41704        6      good
chromium@510992      18.6441 +- 0.251249      6      good
chromium@511017      17.5638 +- 0.936728      6      good
chromium@511030      17.7738 +- 1.63148       9      good
chromium@511036      17.8282 +- 0.950551      6      good
chromium@511037      16.4524 +- 1.18629       6      bad       <--
chromium@511038      16.3563 +- 0.406284      6      bad
chromium@511039      16.7432 +- 0.79886       6      bad
chromium@511042      15.791 +- 1.81779        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 --story-filter=text.constant.full.page.raster.05000.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/8964677053804077280


For feedback, file a bug with component Speed>Bisection

The graphs in the link:
https://chromeperf.appspot.com/group_report?bug_id=778609
says "lower is better", and in the above "Bisect Details", it shows the change is:
14.12% | 18.3870320561 -> 15.7909894819
It means the number goes down and if I understand it correctly, it is an improvement.

Hey Simon - the bisect results are really confusing. Is juncai's assessment correct? It seems like it is, which means chromeperf and bisect disagree, but the "lower is better" context is missing from bisect in this case, and the good-->bad transition in bisect output is confusing.

Sorry, was OOO for a week. Bisect doesn't really have a concept of which direction the regression is in, it runs the test at both the start/end (with the assumption start=good, end=bad) and then just looks for where they may have diverged. In this case, not sure why the bisect got different values from the bot but maybe widening the range and running without a story filter might yield better results.