[Webkit-unassigned] [Bug 88482] [Chromium] Compositor should avoid drawing quads when cached textures are available and contents unchanged

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Thu Jun 14 12:39:51 PDT 2012


--- Comment #33 from Adrienne Walker <enne at google.com>  2012-06-14 12:39:51 PST ---
(In reply to comment #31)
> (From update of attachment 147153 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=147153&action=review

> > As I was dumping layer trees debugging this, I found they are for most part pretty flat. Is O(n^2) a realistic possibility here?
> > 
> > I can come up with a separate walk that will propagate the flag, but it will be way more code, and will be much slower in most realistic cases, though technically O(n). I'd prefer not to modify calcDraw at this point as I'm not sure I can do this safely.
> > 
> > What do you think?
> Here's one other possibility: have this getter set the flag for the entire path it just traversed in case of positive result. I.e. instead of just "return true", run up the tree again, this time setting this flag as a side-effect. This will avoid duplicating this work next time our ancestor is queried.

Re: the other possibility.  I would prefer not to have a getter be non-const, even if the result is the same regardless.

It's the negative result case I'm most worried about--not only is it far more likely that transform/opacity is not changed on a given frame, but it's the negative result that requires walking all the way up to the root.

Eh.  Maybe this is fine as-is and we can make it more efficient in the future if need be.

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

More information about the webkit-unassigned mailing list