All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=629505

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICg6ua3pwoM


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

android-nexus5

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/9006699472949281232

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/9005128036179801328

Yeah, this is just Ganesh going on/off. Merging.

===== BISECT JOB RESULTS =====
Status: completed


===== TESTED REVISIONS =====
Revision         Mean     Std Dev    N  Good?
chromium@405974  1.87494  0.0134182  5  good
chromium@405979  1.89793  0.0519207  5  good
chromium@405980  1.90041  0.051722   5  good
chromium@405981  2.21929  0.029338   5  bad
chromium@405983  2.24847  0.0176203  5  bad

Bisect job ran on: android_nexus6_perf_bisect
Bug ID: 629505

Test Command: src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --upload-results --also-run-disabled-tests thread_times.simple_mobile_sites
Test Metric: thread_browser_cpu_time_per_frame/thread_browser_cpu_time_per_frame
Relative Change: 19.92%
Score: 0

Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/android_nexus6_perf_bisect/builds/2433
Job details: https://chromeperf.appspot.com/buildbucket_job_status/9005128036179801328


Not what you expected? We'll investigate and get back to you!
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5214889107259392

| 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 Tests>AutoBisect.  Thank you!

This is likely not GPU raster related - Looking at the various bisects (including the one posted to the bug), there seems to be a bit of noise - my guess is that 405981 caused this change: https://chromium.googlesource.com/external/github.com/catapult-project/catapult.git/+/d2b666834a289f8d109c5383fb16e7a9cba66145

This was a catapult roll which modified how scroll amounts were computed. It seems entirely possible that this would impact power metrics.

Additionally, GPU raster was landed and reverted a few times in this timeframe, and we don't see significant movement at the other points.

Adding ulan@ to see if it makes sense that his change would have impacted the cpu time on thread_times.simple_mobile_sites, which uses scrolling. If this seems reasonable, please feel free to close this issue out.

It is possible that my CL changes scroll amount for this benchmark.