[Webkit-unassigned] [Bug 114329] [Cairo] Take current matrix into pattern matrix to end up at identity scale when cairo_fill runs

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Thu Apr 18 18:35:14 PDT 2013


https://bugs.webkit.org/show_bug.cgi?id=114329





--- Comment #7 from Kyungjin Kim <gen.kim at samsung.com>  2013-04-18 18:33:31 PST ---
(In reply to comment #6)
> So it seems what you are doing here is that instead of creating a scaled pattern matrix, for every render you are creating a temporary image surface with the scaled source? 

Yes

> Is this also slow on non-ARM CPUs? I think a change like this needs a little justification as well. 

I don't think so. But on ARM CPUs this patch helps the performance in some benchmarks. i.e) Browsermark2 2d transform scores 1200 to 2000

> Can it be fixed or optimized in Pixman? 
I doubt it, the function already been neon-optimized as i heard.

> The Skia port used to cache scaled images -- if we *must* scale the image it would probably be better to cache the scaled version.

I can't feel you for caching scaled images part. This patch makes drawPatternToCairoContext look alike Image::drawPattern in skia.

-- 
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