All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=638013

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICgjufIuAkM


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

chromium-rel-mac11

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/9004219503562973136

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


=== Bisection aborted ===
The bisect was aborted because The metric values for the initial "good" and "bad" revisions do not represent a clear regression.
Please contact the the team (see below) if you believe this is in error.

=== Warnings ===
The following warnings were raised by the bisect job:

 * Bisect failed to reproduce the regression with enough confidence.

===== TESTED REVISIONS =====
Revision         Mean     Std Dev  N   Good?
chromium@411923  27.7223  1.19814  18  good
chromium@411924  28.3008  1.90539  18  bad

Bisect job ran on: mac_10_11_perf_bisect
Bug ID: 638013

Test Command: src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --upload-results --also-run-disabled-tests power.gpu_rasterization.top_10
Test Metric: energy_consumption_mwh/energy_consumption_mwh
Relative Change: 1.90%
Score: 0

Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/mac_10_11_perf_bisect/builds/832
Job details: https://chromeperf.appspot.com/buildbucket_job_status/9004219503562973136


Not what you expected? We'll investigate and get back to you!
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5802855986888704

| 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!

Charlie, can you look at the graphs in comment 1?

I'm pretty sure this should be ignored. The difference between the energy consumption before and after the ref/ToT builds settle remains pretty constant, and, zooming out, it looks like both the ref build and ToT seem to pretty regularly jump between 26mWh and 30mWh (there's a definite bimodal distribution). I suspect this has something to do with the fact that these tests use the Intel MSRs to report power measurements, which we've had bad experiences with. (This was the impetus for getting BattOr-based tracing working.)