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

Issue 794999 link

Starred by 1 user
Status: Fixed
Owner:
wangxianzhu@chromium.org
Closed: Mar 2018
Cc:
futhark@chromium.org
sullivan@chromium.org
haraken@chromium.org
e...@chromium.org
█ kraynov@chromium.org
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Bug-Regression



Sign in to add a comment

11.1%-11.3% regression in blink_perf.paint at 522380:522413

Project Member Reported by kraynov@chromium.org, Dec 14 2017 
See the link to graphs below.
Project Member

Comment 1 by 42576172...@developer.gserviceaccount.com, Dec 14 2017 

All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=794999

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=c518cc1f69632d5242380ccb96874c6ae97e7c3e601def5776c12699cca5c7d9


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

chromium-rel-win7-gpu-nvidia
win-high-dpi
Project Member

Comment 3 by 42576172...@developer.gserviceaccount.com, Dec 15 2017 

😿 Pinpoint job stopped with an error.
https://pinpoint-dot-chromeperf.appspot.com/job/16cbad8a040000

Comment 6 by sullivan@chromium.org, Jan 23 2018 

Kicking off a few more bisects, since pinpoint did repro.
Project Member

Comment 7 by 42576172...@developer.gserviceaccount.com, Jan 23 2018 

😿 Pinpoint job stopped with an error.
https://pinpoint-dot-chromeperf.appspot.com/job/14e7b50c840000

Comment 8 by sullivan@google.com, Feb 23 2018 

Re-running after the fix for the errors.
Project Member

Comment 10 by 42576172...@developer.gserviceaccount.com, Feb 24 2018

Cc: meade@chromium.org haraken@chromium.org bugsnash@chromium.org ericwilligers@chromium.org
Owner: meade@chromium.org
Status: Assigned (was: Untriaged)
📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/14821790440000

Separate out the four classes contained in CSSPropertyValueSet.h into four different files by meade@chromium.org
https://chromium.googlesource.com/chromium/src/+/951544c9d4f42ebae1262e91a03fe660e09146c5

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

Comment 11 by ericwilligers@chromium.org, Feb 25 2018

Cc: -meade@chromium.org -ericwilligers@chromium.org -bugsnash@chromium.org futhark@chromium.org
Owner: e...@chromium.org
We might need to revert the CL.

Comment 12 by e...@chromium.org, Feb 27 2018 

It was reverted December 18 as r524633 (cedfd986e1aba860a363c54e3367647b0ded5ec2).

Comment 13 by e...@chromium.org, Feb 27 2018

Cc: e...@chromium.org
Owner: ----
Status: Available (was: Assigned)
...after which we saw a recover on the win7 bot. The win10 hidpi one only saw a partial recover.

Comment 14 by sullivan@chromium.org, Mar 16 2018

Cc: sullivan@chromium.org
Owner: wangxianzhu@chromium.org
wangxianzhu: as benchmark owner, do you think there's more that can be done here? The culprit CL was reverted with partial recovery, but to me it looks too noisy to investigate further.

Comment 15 by wangxianzhu@chromium.org, Mar 16 2018

Status: Fixed (was: Available)
The current performance is fine. Also noticed that the reference performance also raised a bit in January.
►

Sign in to add a comment