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

Issue 805987 link

Starred by 1 user

Issue metadata

Status: Fixed
Owner:
Closed: Mar 2018
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Bug-Regression



Sign in to add a comment

5.9% regression in rasterize_and_record_micro.top_25 at 531353:531398

Project Member Reported by majidvp@google.com, Jan 25 2018

Issue description

See the link to graphs below.
 
Project Member

Comment 1 by 42576172...@developer.gserviceaccount.com, Jan 25 2018

All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=805987

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=f5f8e255b71d6d2d99f8e468df40ad90963f2574eed1c004045a6d21e10429a9


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

android-nexus5X
Project Member

Comment 3 by 42576172...@developer.gserviceaccount.com, Jan 26 2018

📍 Couldn't reproduce a difference.
https://pinpoint.chromeperf.appspot.com/job/128964dc840000
Components: Internals>GPU>Metrics
Last bisect actually hit a swarming timeout. Trying again on a wider range.
Trying another bisect.
Project Member

Comment 9 by 42576172...@developer.gserviceaccount.com, Mar 16 2018

Cc: x...@chromium.org agrieve@chromium.org dcheng@chromium.org nyquist@chromium.org m...@chromium.org nasko@chromium.org tsepez@chromium.org chfremer@chromium.org g...@chromium.org
Owner: x...@chromium.org
Status: Assigned (was: Untriaged)
📍 Found significant differences after each of 2 commits.
https://pinpoint-dot-chromeperf.appspot.com/job/17e205be440000

Use the actual Chrome OS AFDO profiles on Android by gbiv@chromium.org
https://chromium.googlesource.com/chromium/src/+/1f751e41506c1b1c875ca589681872f725b6d7ab

Move video_capture.mojom out of content. by xjz@chromium.org
https://chromium.googlesource.com/chromium/src/+/a4b93b7a062e8b03bdb1c0a605fe5dabc2bc50b0

Understanding performance regressions:
  http://g.co/ChromePerformanceRegressions

Comment 10 by x...@chromium.org, Mar 16 2018

Owner: g...@chromium.org
My refactoring change is purely namespace change, which sounds unlikely related to this regression. 

gbiv@: Can you PTAL at the other change?

Comment 11 by g...@chromium.org, Mar 16 2018

Status: Started (was: Assigned)
Looking at the longer-term graph, I see a substantial perf improvement when 2e17642e132375121818223a87d8eb02efbf4004, which was supposed to fix perf regressions like this, landed.

Kicked off a pinpoint job in  issue 822868  to verify that 2e17642e132375121818223a87d8eb02efbf4004 gets credit for that perf improvement. If so, we should be good here. :)

Comment 12 by g...@chromium.org, Mar 19 2018

Status: Fixed (was: Started)
So, the pinpoint bots apparently don't like me. :)

Since we saw such a pronounced improvement (and correction from a new profile shortly after), I'm going to go with my original plan of assuming that the improvement was caused by 2e17642e132375121818223a87d8eb02efbf4004.

Sign in to add a comment