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

Issue 747876 link

Starred by 1 user
Status: Duplicate
Merged: issue 747890
Owner:
█ chongz@chromium.org
Last visit > 30 days ago
Closed: Jul 2017
Cc:
█ chongz@chromium.org
hjd@google.com
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Bug-Regression



Sign in to add a comment

2.1% regression in memory.top_10_mobile at 488447:488584

Project Member Reported by hjd@google.com, Jul 24 2017 
See the link to graphs below.
Project Member

Comment 1 by 42576172...@developer.gserviceaccount.com, Jul 24 2017 

All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=747876

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=3a8bd61a27c4cdebbcabad0135177a77f7d6dd605bbca27397b13026ea8f8a41


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

android-one
Project Member

Comment 3 by sheriffbot@chromium.org, Jul 24 2017

Labels: Hotlist-Google
Project Member

Comment 4 by 42576172...@developer.gserviceaccount.com, Jul 25 2017

Cc: chongz@chromium.org
Owner: chongz@chromium.org

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

Hi chongz@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 : Chong Zhang
  Commit : 990aaa397ddbd256f8184a4168cf6873ef2455a2
  Date   : Thu Jul 20 22:31:29 2017
  Subject: [Android] Use oldest historical event time as hardware timestamp

Bisect Details
  Configuration: android_one_perf_bisect
  Benchmark    : memory.top_10_mobile
  Metric       : memory:chrome:all_processes:reported_by_os:system_memory:java_heap:proportional_resident_size_avg/background/after_http_yandex_ru_touchsearch_text_science
  Change       : 1.48% | 5475181.71429 -> 5556110.22222

Revision             Result                  N
chromium@488446      5475182 +- 204666       14      good
chromium@488447      5771435 +- 55266.0      6       bad       <--
chromium@488448      5770069 +- 47864.0      6       bad
chromium@488449      5593941 +- 91525.4      6       bad
chromium@488451      5569764 +- 108029       9       bad
chromium@488455      5578524 +- 109590       9       bad
chromium@488463      5561913 +- 195641       9       bad
chromium@488481      5570341 +- 228658       14      bad
chromium@488515      5587456 +- 172954       6       bad
chromium@488584      5556110 +- 183633       9       bad

Please refer to the following doc on diagnosing memory regressions:
  https://chromium.googlesource.com/chromium/src/+/master/docs/memory-infra/memory_benchmarks.md

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 memory.top_10_mobile

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

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8973192729772467840


For feedback, file a bug with component Speed>Bisection

Comment 8 by hjd@chromium.org, Jul 25 2017

Owner: ----
hmm, it's not obvious to me how that CL could have caused this regression. I've kicked of more bisects, we'll see if come back to the same place.
Project Member

Comment 9 by 42576172...@developer.gserviceaccount.com, Jul 25 2017

Mergedinto: 747890
Status: Duplicate (was: Untriaged)

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

Suspected Commit
  Author : Anton Vayvod
  Commit : 09fa66e1baaa95d069bbd6620670311ae8aa4dc5
  Date   : Thu Jul 20 23:02:12 2017
  Subject: [RemotePlayback] Add media source info to the RemotePlayback availability url

Bisect Details
  Configuration: android_one_perf_bisect
  Benchmark    : memory.top_10_mobile
  Metric       : memory:chrome:all_processes:reported_by_os:system_memory:java_heap:proportional_resident_size_avg/background/after_https_www_google_co_uk_hl_en_q_science
  Change       : 1.82% | 5424981.33333 -> 5523626.66667

Revision             Result                  N
chromium@488446      5424981 +- 49160.9      6      good
chromium@488464      5431808 +- 87774.8      6      good
chromium@488465      5614421 +- 157406       6      bad       <--
chromium@488466      5574884 +- 225279       9      bad
chromium@488467      5576021 +- 225438       9      bad
chromium@488469      5531819 +- 106007       6      bad
chromium@488473      5531819 +- 87423.7      6      bad
chromium@488481      5533355 +- 92666.8      6      bad
chromium@488515      5521920 +- 95355.8      6      bad
chromium@488584      5523627 +- 93306.8      6      bad

Please refer to the following doc on diagnosing memory regressions:
  https://chromium.googlesource.com/chromium/src/+/master/docs/memory-infra/memory_benchmarks.md

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 memory.top_10_mobile

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

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8973105899936824912


For feedback, file a bug with component Speed>Bisection
Project Member

Comment 10 by 42576172...@developer.gserviceaccount.com, Jul 26 2017

Owner: chongz@chromium.org

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

Hi chongz@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 : Chong Zhang
  Commit : 990aaa397ddbd256f8184a4168cf6873ef2455a2
  Date   : Thu Jul 20 22:31:29 2017
  Subject: [Android] Use oldest historical event time as hardware timestamp

Bisect Details
  Configuration: android_one_perf_bisect
  Benchmark    : memory.top_10_mobile
  Metric       : memory:chrome:all_processes:reported_by_os:system_memory:java_heap:proportional_resident_size_avg/background/after_http_yandex_ru_touchsearch_text_science
  Change       : 1.62% | 5463625.14286 -> 5552241.77778

Revision             Result                  N
chromium@488446      5463625 +- 210345       14      good
chromium@488447      5571072 +- 72128.4      6       bad       <--
chromium@488448      5565212 +- 156694       9       bad
chromium@488449      5556955 +- 286352       14      bad
chromium@488451      5557248 +- 283942       14      bad
chromium@488455      5527780 +- 224860       9       bad
chromium@488464      5554322 +- 292867       14      bad
chromium@488481      5628473 +- 241591       9       bad
chromium@488515      5624832 +- 238836       9       bad
chromium@488584      5552242 +- 182263       9       bad

Please refer to the following doc on diagnosing memory regressions:
  https://chromium.googlesource.com/chromium/src/+/master/docs/memory-infra/memory_benchmarks.md

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 memory.top_10_mobile

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

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8973105871216614496


For feedback, file a bug with component Speed>Bisection
Project Member

Comment 11 by 42576172...@developer.gserviceaccount.com, Jul 26 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_nexus6_perf_bisect
  Benchmark    : memory.top_10_mobile
  Metric       : memory:chrome:all_processes:reported_by_os:system_memory:java_heap:proportional_resident_size_avg/background/after_https_m_facebook_com_rihanna
  Change       : 1.03% | 10374948.5714 -> 10483419.4286

Suspected Commit Range
  2 commits in range
  https://chromium.googlesource.com/chromium/src/+log/a42eff7d37e3a64b5058acbd9bb095277b55fdbf..861fd67c79e141815a7f96929a8df0f129d54cc1


Revision             Result                  N
chromium@488346      10374949 +- 212041      14       good
chromium@488364      10358930 +- 337485      14       good
chromium@488367      10363904 +- 332187      14       good
chromium@488368      ---                     ---      build failure
chromium@488369      10583991 +- 642368      14       bad
chromium@488373      10580992 +- 642588      14       bad
chromium@488381      10629315 +- 751424      21       bad
chromium@488418      10538130 +- 489754      14       bad
chromium@488487      10483419 +- 158819      14       bad

Please refer to the following doc on diagnosing memory regressions:
  https://chromium.googlesource.com/chromium/src/+/master/docs/memory-infra/memory_benchmarks.md

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 memory.top_10_mobile

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

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8973105889619135952


For feedback, file a bug with component Speed>Bisection

Comment 12 by hjd@chromium.org, Jul 26 2017 

Three bisects on basically the same graph and three results :(
The weight of evidence points to issue 747890 however.
►

Sign in to add a comment