Closed Bug 1206721 Opened 6 years ago Closed 6 years ago

22% tps regression on osx on from APZ patch in bug 1191539

Categories

(Core :: Panning and Zooming, defect)

defect
Not set
normal

Tracking

()

RESOLVED INCOMPLETE
Tracking Status
firefox44 --- affected

People

(Reporter: jmaher, Unassigned)

References

Details

(Keywords: perf, regression, Whiteboard: [talos_regression][gfx-noted])

here are the two documented regressions:
http://alertmanager.allizom.org:8080/alerts.html?rev=6e112d30103e1f13d93f3380d83c50485b30b172&showAll=1&testIndex=0&platIndex=0

it is a large regression, but we need to try to address this.
Umm apparently in bug 1191539 comment 41 we said we should let BenWa look at this, which never happened, and this bug got forgotten completely. BenWa, not sure if you want to take a look at this still?
Flags: needinfo?(bgirard)
Whiteboard: [talos_regression] → [talos_regression][gfx-noted]
Right now I'm focusing on bug 1242609 which is related to APZ performance. But it wont improve the TPS score.
Flags: needinfo?(bgirard)
Ok, but what is the next action here? Should we just close this bug, or should I assign it to you to look at eventually?
Flags: needinfo?(bgirard)
I think we should keep it open and unassigned. Anyone who has cycles to work on APZ perf should focus on any user visible regressions first then do any low hanging fruits. When we've fixed these we cycle back to this bug (and similar talos perf delta), assign the bug, see where we stand and fix any remaining differences or explain why they are fundamental in the design.
Flags: needinfo?(bgirard)
I don't think we ever "cycle back" to bugs like these, so there's no point in keeping this open.
Status: NEW → RESOLVED
Closed: 6 years ago
Resolution: --- → INCOMPLETE
You need to log in before you can comment on or make changes to this bug.