Cairo [was: rawhide report: 20050621 changes]
Jon Nettleton
jon.nettleton at gmail.com
Wed Jun 22 17:44:34 UTC 2005
On Wed, 2005-06-22 at 17:06 +0200, Tarjei Knapstad wrote:
> On Wed, 2005-06-22 at 14:40, Rodd Clarkson wrote:
> > On Wed, 2005-06-22 at 10:26 +0200, Tarjei Knapstad wrote:
> > > On Wed, 2005-06-22 at 02:46, Rodd Clarkson wrote:
> > > > On Tue, 2005-06-21 at 07:02 -0400, Build System wrote:
> >
>
> <snip>
>
> > > accelerated graphics. Those without it shouldn't notice any difference.
> >
> > Don't get me wrong (and I'm not sure if you have).
>
> I think it might be the other way around :) (or we're both confused :))
>
> > What I'm wanting to know is will people without hardware acceleration be
> > worse off. Are non-hw-accelerated users going to end up with a system
> > that runs something like non-hw-accelerated 3D (which really sucks, even
> > with a good processor)? Or will the rendering on non-hw-accelerated
> > systems be quite good, and rendering with hw-accelerated systems will be
> > brilliant?
> >
>
> What I was trying to say in the last line above was that users without
> hardware accelerated graphics shouldn't notice any performance hit
> compared to today. I would be highly surprised if this wasn't a design
> goal in Cairo given that all hardware accelerated graphics currently
> require the use of proprietary drivers (99.9% anyway...) which most of
> the FOSS community seems to bark at.
>
> Normal X window rendering should be on par or better than with current
> vector drawing libraries. I can't imagine that serious performance
> regressions against earlier vector drawing solutions won't be fixed
> before Cairo goes live in gtk+ etc.
>
> In short, I wouldn't worry too much.
>
> --
> Tarjei
>
If anyone who hasn't upgraded to the newest rawhide ( Unfortunately I
jumped the gun before thinking about it ) wants to collect some numbers.
I would suggest grabbing gtkperf
http://gtkperf.sourceforge.net/index.php?page=testing&id=1 and running a
test case, before and after gtk2 with cairo support. I actually think
so far the update feels a bit faster, however some definitive numbers
would be interesting.
Jon
More information about the fedora-devel-list
mailing list