All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=620071

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICg_OXYrQoM


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

chromium-rel-mac-hdd

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

Hi ericrk@chromium.org, the bisect results pointed to your CL below as possibly
causing a regression. Please have a look at this info and see whether
your CL be related.


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


===== SUSPECTED CL(s) =====
Subject : Add timeout to swap fence
Author  : ericrk
Commit description:
  
In some cases (mainly on Mac OS 10.9), it appears that a glFence
may never pass. To work around this issue (and make the code more
robust in similar cases), a 32ms timeout has been added, after which
we will just call glFinish rather than continuing to wait on the fence.

Will add an UMA stat in a follow-up CL to make sure this is very rare
on 10.9 - if this turns out to be at all common, we should probably
modify 10.9 to indicate that it doesn't support glFence.

BUG=618075
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/2064853002
Cr-Commit-Position: refs/heads/master@{#399598}
Commit  : d84042a6b9b061fd319c0068bdd0c786a16fcc85
Date    : Mon Jun 13 23:38:32 2016


===== TESTED REVISIONS =====
Revision         Mean    Std Dev  N  Good?
chromium@399566  3278.0  13.0384  5  good
chromium@399589  3280.0  10.0     5  good
chromium@399595  3276.0  21.9089  5  good
chromium@399597  3260.0  39.37    5  good
chromium@399598  3128.0  17.8885  5  bad    <--
chromium@399600  3128.0  8.3666   5  bad
chromium@399611  3130.0  12.2474  5  bad
chromium@399655  3112.0  8.3666   5  bad

Bisect job ran on: mac_hdd_perf_bisect
Bug ID: 620071

Test Command: src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --upload-results --also-run-disabled-tests oortonline
Test Metric: Total/http___oortonline.gl__run
Relative Change: 5.06%
Score: 99.9

Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/mac_hdd_perf_bisect/builds/580
Job details: https://chromeperf.appspot.com/buildbucket_job_status/9009847546051197376


Not what you expected? We'll investigate and get back to you!
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5838499882729472

| 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!

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


===== SUSPECTED CL(s) =====
Subject : Add timeout to swap fence
Author  : ericrk
Commit description:
  
In some cases (mainly on Mac OS 10.9), it appears that a glFence
may never pass. To work around this issue (and make the code more
robust in similar cases), a 32ms timeout has been added, after which
we will just call glFinish rather than continuing to wait on the fence.

Will add an UMA stat in a follow-up CL to make sure this is very rare
on 10.9 - if this turns out to be at all common, we should probably
modify 10.9 to indicate that it doesn't support glFence.

BUG=618075
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/2064853002
Cr-Commit-Position: refs/heads/master@{#399598}
Commit  : d84042a6b9b061fd319c0068bdd0c786a16fcc85
Date    : Mon Jun 13 23:38:32 2016


===== TESTED REVISIONS =====
Revision         Mean    Std Dev  N  Good?
chromium@399566  3288.0  4.47214  5  good
chromium@399589  3286.0  15.1658  5  good
chromium@399595  3280.0  10.0     5  good
chromium@399597  3284.0  5.47723  5  good
chromium@399598  3136.0  20.7364  5  bad    <--
chromium@399600  3138.0  13.0384  5  bad
chromium@399611  3124.0  19.4936  5  bad
chromium@399655  3100.0  12.2474  5  bad

Bisect job ran on: mac_hdd_perf_bisect
Bug ID: 620071

Test Command: src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --upload-results --also-run-disabled-tests oortonline
Test Metric: Total/http___oortonline.gl__run
Relative Change: 5.72%
Score: 99.9

Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/mac_hdd_perf_bisect/builds/581
Job details: https://chromeperf.appspot.com/buildbucket_job_status/9009847535670581472


Not what you expected? We'll investigate and get back to you!
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5325594888241152

| 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 regression is just un-doing the improvement I checked in in crrev.com/2028303002 (Point ID: 397745 on the graph). While I'd like to keep the improvement, this change is necessary to prevent other (more serious) bugs. I can look into whether a longer timeout/etc... might help, but I think we may have to just swallow this.

Dropping to Pri 3.

Got it, thanks for the info. Given that, let's just close this out.