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

Issue 762180 link

Starred by 1 user

Issue metadata

Status: WontFix
Owner:
Closed: Sep 2017
Cc:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Bug-Regression



Sign in to add a comment

9.3%-127.3% regression in thread_times.simple_mobile_sites at 496374:498864

Project Member Reported by briander...@chromium.org, Sep 5 2017

Issue description

See the link to graphs below.
 
All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=762180

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=ec3b1b8505afe140f4de0a7991cc9272e2d29c88f5c9f17d7f9e49877ffe0471


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

android-nexus5
android-nexus5X
android-nexus6
android-nexus7v2
android-webview-nexus5X
android-webview-nexus6
Cc: khushals...@chromium.org
Owner: khushals...@chromium.org
Status: Assigned (was: Untriaged)

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

Hi khushalsagar@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 : Khushal
  Commit : 7875d555052d5faeb68ccc586a3be2d183b2c215
  Date   : Wed Aug 30 19:40:12 2017
  Subject: cc: Only checker images required for pending tree activation.

Bisect Details
  Configuration: android_nexus5X_perf_bisect
  Benchmark    : thread_times.simple_mobile_sites
  Metric       : thread_raster_cpu_time_per_frame/http___www.ebay.co.uk_
  Change       : 104.31% | 0.845887139882 -> 1.72826805036

Revision             Result                     N
chromium@498538      0.845887 +- 0.107155       6      good
chromium@498561      0.837471 +- 0.120501       6      good
chromium@498567      0.805534 +- 0.0565139      6      good
chromium@498570      0.826695 +- 0.08417        6      good
chromium@498571      1.68945 +- 0.106402        6      bad       <--
chromium@498572      1.73302 +- 0.146317        6      bad
chromium@498583      1.70298 +- 0.151567        6      bad
chromium@498628      1.72827 +- 0.141757        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...www.ebay.co.uk. thread_times.simple_mobile_sites

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

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8969257307346037872


For feedback, file a bug with component Speed>Bisection
 Issue 762281  has been merged into this issue.
 Issue 762302  has been merged into this issue.
This is not a regression. The patch that is causing more cpu time on raster workers disabled async decoding of pre-paint images on the image decoding service (which used a thread from the background worker pool). Since these images are now decoded on the raster worker pool, it increases the cpu time on this thread. There should be a decrease in thread_other_cpu_time_per_frame, which captures the thread_time from where this work was happening earlier. I've started a bisect to verify this.

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

Suspected Commit
  Author : Khushal
  Commit : 7875d555052d5faeb68ccc586a3be2d183b2c215
  Date   : Wed Aug 30 19:40:12 2017
  Subject: cc: Only checker images required for pending tree activation.

Bisect Details
  Configuration: android_webview_nexus6_aosp_perf_bisect
  Benchmark    : thread_times.simple_mobile_sites
  Metric       : thread_other_cpu_time_per_frame/http___www.ebay.co.uk_
  Change       : 10.26% | 10.8401414032 -> 9.72802033262

Revision             Result                   N
chromium@498560      10.8401 +- 0.60706       6      good
chromium@498567      10.9856 +- 0.872797      6      good
chromium@498569      11.2229 +- 0.304047      6      good
chromium@498570      10.9748 +- 0.61767       6      good
chromium@498571      10.0856 +- 0.312923      6      bad       <--
chromium@498574      9.92431 +- 0.587743      6      bad
chromium@498588      10.2621 +- 0.54142       9      bad
chromium@498615      9.72802 +- 1.12565       6      bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=android-webview --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=http...www.ebay.co.uk. thread_times.simple_mobile_sites

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

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8969166887667530016


For feedback, file a bug with component Speed>Bisection
Status: WontFix (was: Assigned)
The bisect shows an improvement in other time for the same patch. Marking this as WontFix.

Sign in to add a comment