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

Issue 797464 link

Starred by 1 user
Status: WontFix
Owner:
█ jamwalla@chromium.org
Last visit > 30 days ago
Closed: Jan 2018
Cc:
danakj@chromium.org
█ briander...@chromium.org
boliu@chromium.org
skyos...@chromium.org
█ jamwalla@chromium.org
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Bug-Regression

Blocking:
issue 687695



Sign in to add a comment

43%-43.3% regression in scheduler.tough_scheduling_cases at 525693:525801

Project Member Reported by briander...@chromium.org, Dec 23 2017 
See the link to graphs below.
Project Member

Comment 1 by 42576172...@developer.gserviceaccount.com, Dec 23 2017 

All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=797464

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=23751ccd5ab777bb2b9aef396403d60249be348abcac702f0eb26ff35360359f


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

chromium-rel-win7-gpu-ati
chromium-rel-win7-gpu-nvidia
Project Member

Comment 3 by 42576172...@developer.gserviceaccount.com, Dec 23 2017 

📍 Couldn't reproduce a difference.
https://pinpoint-dot-chromeperf.appspot.com/job/14c6b5b1040000

Comment 6 by sullivan@chromium.org, Jan 24 2018 

Bisect message is incorrect; it failed to produce values so it's unclear if it could reproduce a difference. Kicking off two more bisects.
Project Member

Comment 7 by 42576172...@developer.gserviceaccount.com, Jan 24 2018

Cc: danakj@chromium.org jamwalla@chromium.org
Owner: jamwalla@chromium.org
Status: Assigned (was: Untriaged)
📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/14fedc1c840000

cc: don't ProcessScheduledActions inside BeginImplFrame
By jamwalla@chromium.org · Thu Dec 21 16:40:07 2017
chromium @ aefa586e7c8a20173fac5a4b8dfa301f8460e852

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

Comment 8 by jamwalla@chromium.org, Jan 24 2018

Cc: boliu@chromium.org
+boliu

It looks like this only regressed on windows, and we don't know how testing input works there.

For reference, the same change regressed smoothness.gpu_rasterization.top_25_smooth on android, but only for synthetic input: crbug/797426

Comment 9 by boliu@chromium.org, Jan 24 2018 

queueing_durations went from ~0.05ms to ~0.07ms, that's like.....still effectively 0

but what queueing duration is it measuring exactly...

Comment 10 by boliu@chromium.org, Jan 24 2018

Blocking: 687695
Cc: skyos...@chromium.org
brian/sami, you two are listed as the test owner, can you explain what queueing_durations is measuring?

Comment 11 by boliu@chromium.org, Jan 25 2018

Status: WontFix (was: Assigned)
I assume this is just that ScheduledActionSendBeginMainFrame happens slightly later than before in the overall BeginImplFrame event. That's expected as  crbug.com/797426  explains, and desired, because now input is proceeded before that.

In which case, wontfix!
Project Member

Comment 12 by 42576172...@developer.gserviceaccount.com, Jan 25 2018

Status: Assigned (was: WontFix)
📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/13ea856c840000

cc: don't ProcessScheduledActions inside BeginImplFrame
By jamwalla@chromium.org · Thu Dec 21 16:40:07 2017
chromium @ aefa586e7c8a20173fac5a4b8dfa301f8460e852

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

Comment 13 by boliu@chromium.org, Jan 25 2018

Status: WontFix (was: Assigned)
►

Sign in to add a comment