ScrollInertial metrics regression. |
||
Issue descriptionThere seems to be regressions on ScrollInertial metrics in this range: https://chromium.googlesource.com/chromium/src/+log/69.0.3453.0..69.0.3457.0?pretty=fuller&n=10000 https://uma.googleplex.com/p/chrome/timeline_v2/?sid=5d0560fb8983fc37877227afdfd9d6b7 The 50th percentile graphs show regressions in these metrics: HandledToRendererSwap2_Impl/Main RendererSwapToBrowserNotified2 TimeToHandled2_Impl/Main TimeToScrollUpdateSwapBegin2
,
Jul 4
I don't know enough about these metrics or how fling is handled renderer-side to comment. Probably the same issue with BeginFrames that you saw on Aura platforms? Have you come up with a solution for that issue?
,
Jul 4
>I don't know enough about these metrics or how fling is handled renderer-side to comment. The fling is not handled on the renderer side anymore. If you mean the generated GSU events from the fling controller, they are handled identically to GSU events from normal scrolling. >Probably the same issue with BeginFrames that you saw on Aura platforms? Have you come up with a solution for that issue? I am not sure what you are referring at.. Are you talking about the regression that happened on Aura when we switched to OnAnimiationStep instead of setNeedsBeginFrame? +sadrul@ do you know how using OnAnimate and SetNeedsAnimate to progress fling can cause regressions on RendererSwapToBrowserNotified2 and HandledToRendererSwap2_Impl/Main metrics?
,
Jul 6
> If you mean the generated GSU events... Yes I mean generally I'm not very familiar with how input is handled renderer-side. > Are you talking about the regression that Yes, probably something similar has happened here. If the events arrive after BeginFrame, wouldn't these regressions happen? |
||
►
Sign in to add a comment |
||
Comment 1 by sahel@chromium.org
, Jul 3Status: Assigned (was: Available)