New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 708640 link

Starred by 2 users

Issue metadata

Status: WontFix
Owner: ----
Closed: Apr 2017
Cc:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Bug-Regression



Sign in to add a comment

12.3% regression in blink_perf.events at 461182:461305

Project Member Reported by pmeenan@chromium.org, Apr 5 2017

Issue description

See the link to graphs below.
 
All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=708640

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDgnP-HtAoM


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

chromium-rel-win7-gpu-intel
Cc: sadrul@chromium.org
Owner: sadrul@chromium.org

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

Hi sadrul@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 : sadrul
  Commit : 74685881f39e33005f2ff9034479bda6da1c1463
  Date   : Sat Apr 01 01:33:43 2017
  Subject: gpu: Fix running callback on IO thread.

Bisect Details
  Configuration: winx64intel_perf_bisect
  Benchmark    : blink_perf.events
  Metric       : EventsDispatchingInShadowTrees/EventsDispatchingInShadowTrees
  Change       : 2.73% | 25898.8559359 -> 25192.0187086

Revision             Result                  N
chromium@461181      25898.9 +- 144.745      6      good
chromium@461243      26073.2 +- 253.485      6      good
chromium@461274      25948.7 +- 147.537      6      good
chromium@461290      25599.2 +- 234.82       6      good
chromium@461298      25640.4 +- 172.719      6      good
chromium@461300      25608.8 +- 135.751      6      good
chromium@461301      25332.6 +- 91.8396      6      bad       <--
chromium@461302      25118.9 +- 210.841      6      bad
chromium@461305      25192.0 +- 495.944      6      bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=release_x64 --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests blink_perf.events

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8983135211159779952

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5862921470476288


| 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!
Owner: ----
That change should not affect the performance. I would say the bisect is wrong (how often is that the case?). Is it possible to re-run?
Way more often than we'd like.  I'll kick off a new bisect.
Status: WontFix (was: Untriaged)
Nevermind, the reference build graph wasn't displayed by default and the reference build moved by the same amount at the same time (infrastructure change, not code issue).  Sorry for the false alarm.

=== BISECT JOB RESULTS ===
Perf regression found with culprit

Suspected Commit
  Author : sadrul
  Commit : 74685881f39e33005f2ff9034479bda6da1c1463
  Date   : Sat Apr 01 01:33:43 2017
  Subject: gpu: Fix running callback on IO thread.

Bisect Details
  Configuration: winx64intel_perf_bisect
  Benchmark    : blink_perf.events
  Metric       : EventsDispatchingInShadowTrees/EventsDispatchingInShadowTrees
  Change       : 2.34% | 26057.3393752 -> 25448.4862113

Revision             Result                  N
chromium@461181      26057.3 +- 203.725      6      good
chromium@461243      25992.4 +- 159.494      6      good
chromium@461274      25984.5 +- 207.539      6      good
chromium@461290      25734.2 +- 205.883      6      good
chromium@461298      25773.5 +- 206.409      6      good
chromium@461300      25701.9 +- 245.804      6      good
chromium@461301      25421.8 +- 127.89       6      bad       <--
chromium@461302      25454.4 +- 140.687      6      bad
chromium@461305      25448.5 +- 281.004      6      bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=release_x64 --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests blink_perf.events

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8983126707146602272

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5893463855333376


| 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!

Sign in to add a comment