[Webkit-unassigned] [Bug 48291] Fix LayoutTests/canvas/philip/tests/2d.composite.uncovered.fill.copy.html

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Wed Jun 22 04:50:49 PDT 2011


https://bugs.webkit.org/show_bug.cgi?id=48291





--- Comment #29 from Mustafizur Rahaman <mustaf.here at gmail.com>  2011-06-22 04:50:48 PST ---
(In reply to comment #26)
> (From update of attachment 97332 [details])
> this looks sensible in the context of what's currently going on in the canvas, but has anyone actually been verifying that we haven't been getting any performance regressions from these changes?  This especially important for "copy" as it's a reasonably common compositing mode (vs. atop, etc).  Have we also verified that this behaviour matches other browsers?

Hi Oliver,

I have tried a CompositeCopy on a 1000*500 canvas [copying to the entire canvas region] on WebKit windows build [Winlauncher]. In each iterations, I am doing fillRect for 100 times & this whole exercise I have done for 10 iterations to get a stable result[loading the page 10 times and noting the result].

With Changes: Total time taken(sum of 10 iterations) = 1139 ms ==> avg=1.13 ms per fillRect

WithOUT Changes: Total time taken(sum of 10 iterations) = 1020 ms ==> avg=1.02 ms per fillRect.

Please let me know your opinion & whether you want any modification in the patch/in the existing logic of this filling transparency elsewhere.

-- 
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