[Webkit-unassigned] [Bug 70955] Add CCSmoothedCallback class for smoothing out jittery vsync

Fri Oct 28 02:45:11 PDT 2011

https://bugs.webkit.org/show_bug.cgi?id=70955

--- Comment #8 from Nat Duca <nduca at chromium.org>  2011-10-28 02:45:11 PST ---
(From update of attachment 112712)
View in context: https://bugs.webkit.org/attachment.cgi?id=112712&action=review

I like FakeCCThread, thats pretty cool. Lets peel it off to another patch, regardless. Even if you just shove it up on a webkit bug and assign me owner so we don't lose it, that'd be cool. Long term, I think it belongs with the CCMainThread->CCThread patch... 

I think this algorithm is fascinating. But, there's a lot of code duplication between DelayBasedTimeSource and this, basically because they'e both using a delayed task to build the signal.

Can we rephrase the algorithm as a mode on the CCDelayBasedTimSource, or move the existing algorithm plus this to another subclass?

I'd be fine with setMode(...) and have a few fields on DBTS that aren't used in one mode or the other.

One thing I've been meaning to do is make a variant of the DBTS "achieves60Hz" test that adds noise to the delays. Maybe you can expand on that? I suspect that delay errors are gamma-distributed.

> Source/WebCore/WebCore.gypi:3564
> +            'platform/graphics/chromium/cc/CCSmoothedCallback.cpp',

Lets give this a name that brings it close to CCTimeSource. CCSwapCallbackTimesource?

> Source/WebCore/platform/graphics/chromium/cc/CCSmoothedCallback.cpp:36
> +CCSmoothedCallback::CCSmoothedCallback(double interval, CCThread* thread)

This is based on swapbuffers ack right? /me wonders if there is a clean abstraction so that the graphics context owner (LayerRendererChromium) can create this kind of timesource but then hand it off to the frame rate controller? Its fine if such an abstraction doesn't exist, but this patch in particular pokes at the awkwardness of the plumbing from LayerRendererChromium back to CCFrameRateController.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.