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

Issue 759609 link

Starred by 1 user

Issue metadata

Status: Duplicate
Merged: issue 758921
Owner:
Closed: Sep 2017
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Android
Pri: 2
Type: Bug-Regression



Sign in to add a comment

5.3% regression in system_health.common_mobile at 496243:496345

Project Member Reported by kraynov@chromium.org, Aug 28 2017

Issue description

See the link to graphs below.
 
Project Member

Comment 1 by 42576172...@developer.gserviceaccount.com, Aug 28 2017

All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=759609

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=2d7dca548ebc31a9eaabd76becc41fec2d0152800ccc9b2dc688e63fc97e0f58


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

android-nexus5X
Project Member

Comment 3 by 42576172...@developer.gserviceaccount.com, Aug 28 2017

Cc: mvanouwerkerk@google.com
Owner: mvanouwerkerk@google.com
Status: Assigned (was: Untriaged)

=== Auto-CCing suspected CL author mvanouwerkerk@google.com ===

Hi mvanouwerkerk@google.com, 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 : Michael van Ouwerkerk
  Commit : 9c406aaecb3a05fc6ea931e003fdcf2f3d58560c
  Date   : Tue Aug 22 10:17:36 2017
  Subject: Add entry to test config for NTPCondensedLayout with small logo.

Bisect Details
  Configuration: android_nexus5X_perf_bisect
  Benchmark    : system_health.common_mobile
  Metric       : cpu_time_percentage_avg/browse_chrome/browse_chrome_newtab
  Change       : 3.95% | 0.324456248113 -> 0.337273208943

Revision             Result                      N
chromium@496242      0.324456 +- 0.00467786      6      good
chromium@496255      0.323522 +- 0.00317342      6      good
chromium@496262      0.323887 +- 0.00272454      6      good
chromium@496265      0.32447 +- 0.00276835       6      good
chromium@496267      0.322994 +- 0.0027142       6      good
chromium@496268      0.337655 +- 0.00413848      6      bad       <--
chromium@496294      0.337459 +- 0.00426509      6      bad
chromium@496345      0.337273 +- 0.00215707      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 --story-filter=browse.chrome.newtab system_health.common_mobile

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

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8970006193533525472


For feedback, file a bug with component Speed>Bisection
Components: UI>Browser>NewTabPage
Labels: zine-triaged
Labels: OS-Android
Perf sheriff checking in - is this regression still being worked on?

Comment 7 by fi...@chromium.org, Sep 18 2017

Cc: -mvanouwerkerk@google.com
Owner: mvanouwe...@chromium.org
Cc: erikc...@chromium.org
My CL enabled, primarily, the NTPCondensedLayout feature in the test config. The feature fits more suggestions on the New Tab Page, so some increase in CPU usage makes sense.

Looking at the impacted metric cpu_time_percentage, erikchen seems to be the owner of that. Erik, my feeling is that the increase in CPU usage might be reasonable, we're simply doing more work on the NTP. What do you think?
That seems reasonable to me: Erik, what do you think?
Project Member

Comment 11 by 42576172...@developer.gserviceaccount.com, Sep 21 2017

Mergedinto: 758921
Status: Duplicate (was: Assigned)

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

Suspected Commit
  Author : Brian White
  Commit : 7662ff8cb53121cd4d8916007d430b36efd45895
  Date   : Thu Aug 24 00:03:13 2017
  Subject: Validate histograms before and after each task on Android.

Bisect Details
  Configuration: android_nexus5X_perf_bisect
  Benchmark    : system_health.common_mobile
  Metric       : cpu_time_percentage_avg/browse_chrome/browse_chrome_newtab
  Change       : 6.31% | 0.323989752121 -> 0.344439332736

Revision             Result                      N
chromium@496195      0.32399 +- 0.00318963       6      good
chromium@496562      0.337769 +- 0.00648196      6      good
chromium@496745      0.339192 +- 0.00535137      6      good
chromium@496836      0.341075 +- 0.00452902      6      good
chromium@496859      0.340791 +- 0.0034972       6      good
chromium@496871      0.341085 +- 0.005732        6      good
chromium@496877      0.339005 +- 0.0044246       6      good
chromium@496878      0.477942 +- 0.021016        6      bad       <--
chromium@496879      0.46866 +- 0.0108697        6      bad
chromium@496880      0.460873 +- 0.00536756      6      bad
chromium@496882      0.459225 +- 0.00616776      6      bad
chromium@496928      0.458717 +- 0.00311949      6      bad
chromium@497660      0.344439 +- 0.00139706      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 --story-filter=browse.chrome.newtab system_health.common_mobile

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

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8967840504741171072


For feedback, file a bug with component Speed>Bisection

Sign in to add a comment