All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=726787

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDg-o2utAkM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDg-pHu6QgM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDg-snlpQkM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDg-oD7vAkM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDghtitqQoM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDg-ry3pwkM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDg-s2hpwoM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDg-t2lvAkM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDg-qXvsAkM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDg-s36ugoM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDg-pzg8wgM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDg-qPfqwkM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDg-vWb8AgM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDghqrNoQoM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDg-of7tAsM


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

chromium-rel-mac11-pro

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8978509058843313696

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

Hi fsamuel@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 : fsamuel
  Commit : b520968f4ac40fb89ce25329134b388ca42cca12
  Date   : Thu May 25 03:10:42 2017
  Subject: cc::SurfaceDependencyTracker should not crash when a Display goes away

Bisect Details
  Configuration: mac_pro_perf_bisect
  Benchmark    : media.tough_video_cases
  Metric       : idle_wakeups_browser/video.html?src_tulip2.ogg_type_audio
  Change       : 179.58% | 385.333333333 -> 1077.33333333

Revision             Result                  N
chromium@474496      385.333 +- 11.3725      6      good
chromium@474535      378.833 +- 20.6599      6      good
chromium@474538      381.167 +- 29.9138      6      good
chromium@474539      1103.5 +- 67.1379       6      bad       <--
chromium@474540      1056.0 +- 160.468       6      bad
chromium@474545      1108.17 +- 109.183      6      bad
chromium@474555      1086.0 +- 36.5513       6      bad
chromium@474574      1077.33 +- 32.7923      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=video.html.src.tulip2.ogg.type.audio media.tough_video_cases

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8978509058843313696

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5783178821763072


| 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!

 Issue 726798  has been merged into this issue.

 Issue 726894  has been merged into this issue.

Adding Eric Seckler because he landed CLs around the same time that affect the display scheduler. It seems unlikely that my CL would cause performance or memory regressions given surface synchronization is not turned on in Chrome yet. With that said, maybe I missed something and I'm trying to repro locally.

 Issue 726892  has been merged into this issue.

Ohh I just realized what's causing this regression. SurfaceDependencyTracker is always observing the PrimaryBeginFrameSource which always observes the display's BeginFrameSource. This is causing these regressions. I have a CL in flight to fix this. https://codereview.chromium.org/2887453002/

 Issue 726936  has been merged into this issue.

 Issue 726893  has been merged into this issue.