New issue
Advanced search Search tips

Issue 685182 link

Starred by 1 user
Status: Duplicate
Merged: issue 701355
Owner:
yigu@chromium.org
Closed: Feb 2017
Cc:
tdres...@chromium.org
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Bug-Regression



Sign in to add a comment

40.2%-43% regression in v8.mobile_infinite_scroll_tbmv2 at 445714:445749

Project Member Reported by tdres...@chromium.org, Jan 25 2017 
See the link to graphs below.
Project Member

Comment 3 by 42576172...@developer.gserviceaccount.com, Jan 25 2017 


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

Bisect Details
  Configuration: android_nexus5X_perf_bisect
  Benchmark    : v8.mobile_infinite_scroll_tbmv2
  Metric       : v8-gc-scavenger_max/v8-gc-scavenger_max

Revision             Result                  N
chromium@445713      25.0139 +- 9.82861      21      good
chromium@445749      25.9085 +- 11.5406      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 v8.mobile_infinite_scroll_tbmv2

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8989486503670261008

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5267381816918016


| 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!
Project Member

Comment 6 by 42576172...@developer.gserviceaccount.com, Jan 26 2017

Cc: yigu@chromium.org
Owner: yigu@chromium.org

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

Hi yigu@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 : yigu
  Commit : c83ddee4e5dcb0fdce873021c2701c9c3a2c547a
  Date   : Tue Jan 24 15:11:45 2017
  Subject: Box shadow should be recorded as main thread scrolling reasons for non-root layers.

Bisect Details
  Configuration: android_nexus5X_perf_bisect
  Benchmark    : smoothness.scrolling_tough_ad_cases
  Metric       : first_gesture_scroll_update_latency/http___androidcentral.com
  Change       : 197.88% | 25.4878833333 -> 75.9223833333

Revision             Result                  N
chromium@445709      25.4879 +- 11.3892      6      good
chromium@445725      21.4587 +- 8.00472      6      good
chromium@445727      19.5939 +- 2.42898      6      good
chromium@445728      71.3214 +- 7.29691      6      bad       <--
chromium@445729      81.2603 +- 17.4779      6      bad
chromium@445733      66.1442 +- 37.5454      6      bad
chromium@445740      75.9224 +- 39.2616      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=http...androidcentral.com smoothness.scrolling_tough_ad_cases

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8989385230698095040

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5041325876969472


| 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!

Comment 8 by yigu@chromium.org, Jan 26 2017

Owner: ----
Not related to the suspected patch.
Project Member

Comment 9 by 42576172...@developer.gserviceaccount.com, Jan 26 2017 


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

Bisect Details
  Configuration: android_nexus5X_perf_bisect
  Benchmark    : v8.mobile_infinite_scroll_tbmv2
  Metric       : v8-gc-incremental-step_max/tumblr

Revision             Result                  N
chromium@445713      5.40629 +- 2.94842      21      good
chromium@445749      5.8139 +- 3.07666       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 v8.mobile_infinite_scroll_tbmv2

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8989394377052894368

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=6752099733340160


| 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!
Project Member

Comment 11 by 42576172...@developer.gserviceaccount.com, Jan 27 2017 


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

Suspected Commit
  Author : yigu
  Commit : c83ddee4e5dcb0fdce873021c2701c9c3a2c547a
  Date   : Tue Jan 24 15:11:45 2017
  Subject: Box shadow should be recorded as main thread scrolling reasons for non-root layers.

Bisect Details
  Configuration: android_nexus5X_perf_bisect
  Benchmark    : smoothness.scrolling_tough_ad_cases
  Metric       : first_gesture_scroll_update_latency/http___androidcentral.com
  Change       : 202.10% | 23.5903833333 -> 71.2664

Revision             Result                  N
chromium@445709      23.5904 +- 7.72074      6      good
chromium@445724      20.2861 +- 5.52626      6      good
chromium@445727      22.6645 +- 10.3777      6      good
chromium@445728      66.9848 +- 20.8413      6      bad       <--
chromium@445729      74.8507 +- 26.6102      6      bad
chromium@445733      67.9794 +- 29.459       6      bad
chromium@445740      71.2664 +- 26.4991      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=http...androidcentral.com smoothness.scrolling_tough_ad_cases

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8989309144605177760

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5706889309978624


| 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!
Project Member

Comment 13 by 42576172...@developer.gserviceaccount.com, Jan 27 2017

Cc: spqc...@chromium.org
Owner: spqc...@chromium.org

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

Hi spqchan@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 : spqchan
  Commit : 6771b80c488c2c1b685ab2f674928ee09654ab15
  Date   : Tue Jan 24 17:22:54 2017
  Subject: [Mac] Security Chip Animation Edge Cases

Bisect Details
  Configuration: android_nexus5X_perf_bisect
  Benchmark    : v8.mobile_infinite_scroll_tbmv2
  Metric       : v8-gc-incremental-step_max/facebook

Revision             Result                  N
chromium@445713      4.57329 +- 1.98352      14      good
chromium@445727      4.98329 +- 3.15566      14      good
chromium@445740      5.12717 +- 1.54231      6       good
chromium@445745      5.279 +- 1.1357         6       good
chromium@445747      5.11067 +- 1.13285      6       good
chromium@445748      5.3005 +- 1.45903       6       good
chromium@445749      4.16514 +- 1.84157      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 v8.mobile_infinite_scroll_tbmv2

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8989367370974426368

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5783081492938752


| 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!
Project Member

Comment 14 by 42576172...@developer.gserviceaccount.com, Jan 27 2017 


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

Bisect Details
  Configuration: android_nexus5X_perf_bisect
  Benchmark    : v8.mobile_infinite_scroll_tbmv2
  Metric       : v8-gc-scavenger_max/v8-gc-scavenger_max

Revision             Result                  N
chromium@445713      22.4377 +- 6.08486      21      good
chromium@445749      22.7638 +- 7.27139      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 v8.mobile_infinite_scroll_tbmv2

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8989304785770738848

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5050410403889152


| 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!

Comment 15 by yigu@chromium.org, Jan 27 2017

Cc: -yigu@chromium.org

Comment 16 by spqc...@chromium.org, Jan 30 2017 

This doesn't make any sense, my CL only affects Mac, not Android
Project Member

Comment 18 by 42576172...@developer.gserviceaccount.com, Jan 31 2017 


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

Suspected Commit
  Author : yigu
  Commit : c83ddee4e5dcb0fdce873021c2701c9c3a2c547a
  Date   : Tue Jan 24 15:11:45 2017
  Subject: Box shadow should be recorded as main thread scrolling reasons for non-root layers.

Bisect Details
  Configuration: android_nexus5X_perf_bisect
  Benchmark    : smoothness.scrolling_tough_ad_cases
  Metric       : first_gesture_scroll_update_latency/http___androidcentral.com
  Change       : 221.40% | 21.07905 -> 67.7472833333

Revision             Result                  N
chromium@445709      21.0791 +- 15.3523      6      good
chromium@445725      23.9831 +- 9.43422      6      good
chromium@445727      21.8826 +- 17.3013      6      good
chromium@445728      72.8348 +- 26.8199      6      bad       <--
chromium@445729      65.0084 +- 21.572       6      bad
chromium@445733      64.6258 +- 17.8249      6      bad
chromium@445740      67.7473 +- 19.3381      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=http...androidcentral.com smoothness.scrolling_tough_ad_cases

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8988993788684317168

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5826680427380736


| 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!
Project Member

Comment 20 by 42576172...@developer.gserviceaccount.com, Feb 3 2017 


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

Suspected Commit
  Author : yigu
  Commit : c83ddee4e5dcb0fdce873021c2701c9c3a2c547a
  Date   : Tue Jan 24 15:11:45 2017
  Subject: Box shadow should be recorded as main thread scrolling reasons for non-root layers.

Bisect Details
  Configuration: android_nexus5X_perf_bisect
  Benchmark    : smoothness.scrolling_tough_ad_cases
  Metric       : first_gesture_scroll_update_latency/http___androidcentral.com
  Change       : 175.43% | 27.4328 -> 75.5572333333

Revision             Result                  N
chromium@445709      27.4328 +- 13.7496      6      good
chromium@445725      25.3761 +- 11.2462      6      good
chromium@445727      21.0616 +- 6.56786      6      good
chromium@445728      73.5683 +- 14.9119      6      bad       <--
chromium@445729      68.1112 +- 24.8464      6      bad
chromium@445733      66.1945 +- 30.6949      6      bad
chromium@445740      75.5572 +- 19.9215      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=http...androidcentral.com smoothness.scrolling_tough_ad_cases

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8988720848136386272

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5820176303390720


| 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!

Comment 21 by tdres...@chromium.org, Feb 6 2017

Cc: -spqc...@chromium.org
Owner: yigu@chromium.org
Yi, the bisect seems pretty insistent that your change caused at least part of this regression.

Do any of the graphs look like they could be related?

Comment 22 by yigu@chromium.org, Feb 7 2017

Status: WontFix (was: Untriaged)
The regression goes away without extra attention.

Comment 23 by yigu@chromium.org, Mar 21 2017

Mergedinto: 701355
Status: Duplicate (was: WontFix)
Project Member

Comment 26 by 42576172...@developer.gserviceaccount.com, Apr 11 2017 


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

Bisect Details
  Configuration: android_nexus5X_perf_bisect
  Benchmark    : v8.mobile_infinite_scroll_tbmv2
  Metric       : v8-gc-scavenger_max/v8-gc-scavenger_max

Revision             Result                  N
chromium@445713      34.2425 +- 8.49832      21      good
chromium@445749      34.4335 +- 11.8011      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 v8.mobile_infinite_scroll_tbmv2

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8982652079528782592

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5050410403889152


| 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