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

Issue 700824 link

Starred by 1 user

Issue metadata

Status: Duplicate
Merged: issue 700800
Owner:
Last visit > 30 days ago
Closed: Mar 2017
Cc:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Bug-Regression



Sign in to add a comment

12.7%-86.7% regression in v8.runtimestats.browsing_desktop_turbo at 454487:455034

Project Member Reported by hpayer@google.com, Mar 13 2017

Issue description

See the link to graphs below.
 
Project Member

Comment 3 by 42576172...@developer.gserviceaccount.com, Mar 13 2017


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

Bisect Details
  Configuration: mac_10_11_perf_bisect
  Benchmark    : v8.runtimestats.browsing_desktop_turbo
  Metric       : v8-gc-latency-mark-compactor_sum/browse_media/browse_media_pinterest

Revision             Result                  N
chromium@454739      39.6374 +- 36.0367      21      good
chromium@454985      46.0051 +- 87.7391      21      bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=browse.media.pinterest v8.runtimestats.browsing_desktop_turbo

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8985245342821997888

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=4870222911111168


| 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 5 by 42576172...@developer.gserviceaccount.com, Mar 17 2017

Mergedinto: 700800
Status: Duplicate (was: Assigned)

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

Suspected Commit
  Author : Michael Lippautz
  Commit : 68d6b6c613c8641748a390a10d886859e5afa38b
  Date   : Fri Mar 03 12:10:33 2017
  Subject: [heap] Support for minor MC marking during incremental marking part 2

Bisect Details
  Configuration: winx64nvidia_perf_bisect
  Benchmark    : v8.runtimestats.browsing_desktop_turbo
  Metric       : v8-gc-full-mark-compactor_avg/browse_media/browse_media_imgur
  Change       : 12.15% | 59.0592010979 -> 66.2367457802

Revision                           Result                  N
chromium@454500                    59.0592 +- 2.38698      6      good
chromium@454650                    58.1466 +- 2.45045      6      good
chromium@454725                    59.6048 +- 3.54694      6      good
chromium@454763                    59.2746 +- 1.80981      6      good
chromium@454782                    60.612 +- 5.15761       6      good
chromium@454790                    61.0522 +- 5.726        6      good
chromium@454793                    61.5419 +- 4.29971      6      good
chromium@454794                    59.2686 +- 3.33298      6      good
chromium@454794,v8@de52562d8e      59.9334 +- 2.19374      6      good
chromium@454794,v8@f7bc5ef0bc      57.2323 +- 4.15276      6      good
chromium@454794,v8@14106d2d51      57.8566 +- 2.97988      6      good
chromium@454794,v8@68d6b6c613      67.434 +- 3.30193       6      bad       <--
chromium@454794,v8@1a6487fd65      66.7959 +- 3.64769      6      bad
chromium@454795                    67.6972 +- 6.78226      6      bad
chromium@454800                    68.4396 +- 2.86546      6      bad
chromium@455100                    66.2367 +- 4.40428      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 --story-filter=browse.media.imgur v8.runtimestats.browsing_desktop_turbo

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8984888484507013872

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5833855128305664


| 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 7 by 42576172...@developer.gserviceaccount.com, Mar 23 2017


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

Bisect Details
  Configuration: mac_10_11_perf_bisect
  Benchmark    : v8.runtimestats.browsing_desktop_turbo
  Metric       : v8-gc-latency-mark-compactor_sum/browse_media/browse_media_pinterest

Revision             Result                  N
chromium@454739      35.1406 +- 41.3982      21      good
chromium@454985      42.8957 +- 98.668       21      bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=browse.media.pinterest v8.runtimestats.browsing_desktop_turbo

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8984378653174678864

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=4870222911111168


| 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