All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=713116

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDgwuvcoQsM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDg_PmJjwoM


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

android-nexus5X
android-webview-nexus5X

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8981888031144260736

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

Hi samans@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 : samans
  Commit : 2040988b1345e6ce738bd03ab7ccd5f4b73a956b
  Date   : Tue Apr 11 23:58:35 2017
  Subject: Use MojoCompositorFrameSink in RendererCompositorFrameSink

Bisect Details
  Configuration: android_nexus5X_perf_bisect
  Benchmark    : thread_times.key_mobile_sites_smooth
  Metric       : thread_renderer_compositor_cpu_time_per_frame/thread_renderer_compositor_cpu_time_per_frame
  Change       : 5.04% | 2.18160364461 -> 2.29165175764

Revision             Result                    N
chromium@463804      2.1816 +- 0.0256929       6      good
chromium@463835      2.18571 +- 0.0438297      6      good
chromium@463843      2.18809 +- 0.0421679      6      good
chromium@463844      2.19756 +- 0.0331425      6      good
chromium@463845      2.27079 +- 0.0402977      6      bad       <--
chromium@463847      2.27732 +- 0.0463822      6      bad
chromium@463851      2.27145 +- 0.0203193      6      bad
chromium@463866      2.28714 +- 0.0324225      6      bad
chromium@463927      2.29165 +- 0.0403673      6      bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests thread_times.key_mobile_sites_smooth

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8981888031144260736

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5272375760781312


| 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!

Based on my understanding, Mojo sends the message directly on the thread that initiated it instead of hopping to IO thread, so I think it's expected that the compositor thread is busier. There's been improvements in IO thread as seen in  crbug.com/713720  which supports this hypothesis although I'm not sure whether it's the browser's IO thread or the renderer's? Mark as WontFix.