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

Issue 732307 link

Starred by 1 user

Issue metadata

Status: Duplicate
Merged: issue 731657
Owner: ----
Closed: Aug 2017
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Bug-Regression



Sign in to add a comment

13.8% regression in blink_perf.css at 477242:477304

Project Member Reported by alexclarke@chromium.org, Jun 12 2017

Issue description

See the link to graphs below.
 
All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=732307

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDglujQpwkM


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

android-nexus5X
Project Member

Comment 3 by 42576172...@developer.gserviceaccount.com, Jun 12 2017

Cc: maxlg@chromium.org
Owner: maxlg@chromium.org

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

Hi maxlg@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 : maxlg
  Commit : 206f88a3c6df64c70511446235ed6ff72a94ce9e
  Date   : Tue Jun 06 15:14:09 2017
  Subject: EQT: Record the max queueing time from navigation start to navigation away

Bisect Details
  Configuration: android_nexus5X_perf_bisect
  Benchmark    : blink_perf.css
  Metric       : CSSPropertySetterGetterMethods/CSSPropertySetterGetterMethods
  Change       : 21.69% | 1222.18667706 -> 957.089824066

Revision             Result                  N
chromium@477241      1222.19 +- 764.076      6       good
chromium@477273      1186.6 +- 730.376       6       good
chromium@477281      1271.23 +- 797.941      6       good
chromium@477285      1057.49 +- 693.659      9       good
chromium@477287      1072.55 +- 922.682      14      good
chromium@477288      1020.48 +- 694.971      14      good
chromium@477289      1028.11 +- 595.71       9       bad       <--
chromium@477304      957.09 +- 10.6885       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 blink_perf.css

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8977001819966360000

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5877702118080512


| 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!

Comment 4 by maxlg@chromium.org, Jun 12 2017

That's weird. My changes doesn't seems to affect CSS setter/getter
Project Member

Comment 7 by 42576172...@developer.gserviceaccount.com, Jun 13 2017


=== BISECT JOB RESULTS ===
Bisect failed for unknown reasons

Please contact the team (see below) and report the error.


Bisect Details
  Configuration: android_nexus5X_perf_bisect
  Benchmark    : blink_perf.css
  Metric       : CSSPropertySetterGetterMethods/CSSPropertySetterGetterMethods


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 blink_perf.css

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8976964214618512656

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5785855853068288


| 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!
Project Member

Comment 8 by 42576172...@developer.gserviceaccount.com, Jun 13 2017


=== BISECT JOB RESULTS ===
Bisect failed for unknown reasons

Please contact the team (see below) and report the error.


Bisect Details
  Configuration: android_nexus5X_perf_bisect
  Benchmark    : blink_perf.css
  Metric       : CSSPropertySetterGetterMethods/CSSPropertySetterGetterMethods


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 blink_perf.css

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8976963589687675168

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5852204507856896


| 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!
Project Member

Comment 11 by 42576172...@developer.gserviceaccount.com, Jun 13 2017


=== BISECT JOB RESULTS ===
NO Perf regression found

Bisect Details
  Configuration: android_nexus5X_perf_bisect
  Benchmark    : blink_perf.css
  Metric       : CSSPropertySetterGetterMethods/CSSPropertySetterGetterMethods

Revision             Result                  N
chromium@477241      1161.03 +- 1157.51      21      good
chromium@477304      1111.01 +- 1097.66      21      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 blink_perf.css

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8976879741246790928

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5785855853068288


| 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!
Project Member

Comment 12 by 42576172...@developer.gserviceaccount.com, Jun 14 2017


=== BISECT JOB RESULTS ===
Perf regression found but unable to narrow commit range

Build failures prevented the bisect from narrowing the range further.


Bisect Details
  Configuration: android_nexus5X_perf_bisect
  Benchmark    : blink_perf.css
  Metric       : CSSPropertySetterGetterMethods/CSSPropertySetterGetterMethods
  Change       : 11.26% | 1586.46876691 -> 1419.92073952

Suspected Commit Range
  2 commits in range
  https://chromium.googlesource.com/chromium/src/+log/ee559102b866a30486a95ffc7b30f9547c007eb1..cfb35dd20e4dbab24f638e389ef17cc94f3a6da2


Revision             Result                  N
chromium@477241      1586.47 +- 644.124      9        good
chromium@477273      1454.86 +- 911.65       14       good
chromium@477288      1653.32 +- 48.9749      6        good
chromium@477293      1660.12 +- 41.0503      9        good
chromium@477295      1598.63 +- 688.617      14       good
chromium@477296      ---                     ---      build failure
chromium@477297      1536.18 +- 71.6309      9        bad
chromium@477304      1419.92 +- 538.671      14       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 blink_perf.css

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8976879729830979104

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5852204507856896


| 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: junov@chromium.org
I'm not sure that I trust the bisect from #12 but the last 'bad' revision was @477297 (https://chromium-review.googlesource.com/523664)

junov@ could you take a look please?  If it's not your patch feel free to just unassign yourself :)

Comment 14 by junov@chromium.org, Jun 19 2017

Owner: ----
No my patch. Code path modified in that CL is in a feature that is not used by the test that regressed.
Project Member

Comment 16 by 42576172...@developer.gserviceaccount.com, Aug 17 2017

Cc: dtapu...@chromium.org
Owner: dtapu...@chromium.org
Status: Assigned (was: Untriaged)

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

Hi dtapuska@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 : Dave Tapuska
  Commit : 75fe739ccdd005ba0faa8eda191c03a94cc269cb
  Date   : Tue Jun 06 15:29:45 2017
  Subject: Allow a passive touchstart listener to add a blocking touchmove listener.

Bisect Details
  Configuration: android_nexus5X_perf_bisect
  Benchmark    : blink_perf.css
  Metric       : CSSPropertySetterGetterMethods/CSSPropertySetterGetterMethods
  Change       : 21.39% | 1717.86600247 -> 1350.43670363

Revision             Result                  N
chromium@477190      1717.87 +- 30.4154      6      good
chromium@477244      1688.26 +- 32.5114      6      good
chromium@477271      1667.03 +- 45.6711      6      good
chromium@477284      1687.51 +- 39.1608      6      good
chromium@477291      1578.1 +- 27.8649       9      good
chromium@477292      1030.21 +- 646.167      6      good
chromium@477293      1656.65 +- 192.917      9      bad       <--
chromium@477294      1599.04 +- 651.81       9      bad
chromium@477297      1384.39 +- 718.694      6      bad
chromium@477402      1348.56 +- 707.841      6      bad
chromium@477613      1350.44 +- 42.134       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 blink_perf.css

More information on addressing performance regressions:
  http://g.co/ChromePerformanceRegressions

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8970992261521368864


For feedback, file a bug with component Speed>Bisection
Components: Blink>CSS
Owner: ----
Status: Untriaged (was: Assigned)
Mine is not related to CSS at all. bisect determined wrong change; the values before and after are the same. 

perhaps the CSS team has an idea since this is a perf test strictly around their code.

Comment 19 by shend@chromium.org, Aug 18 2017

Cc: nainar@chromium.org shend@chromium.org
nainar, could this be related to any of our changes?
I think this is a dup of 731657. A number of devices regressed at around the same time.
Mergedinto: 731657
Status: Duplicate (was: Untriaged)
shend@ - Our work doesn't appear in the regression range. 

ericwilligers@ - To better understand this: fserb@'s patch from the other CL doesn't show up in the blame. But yeah given the trend he describes in the other bug I am inclined to think so too. Duping. Thank you!

Sign in to add a comment