[prev in list] [next in list] [prev in thread] [next in thread] 

List:       cairo
Subject:    Re: [cairo] Cairo 1.9.2 performance
From:       Carl Worth <cworth () cworth ! org>
Date:       2009-06-17 23:47:29
Message-ID: 1245282449.23131.144.camel () yoom ! home ! cworth ! org
[Download RAW message or body]

[Attachment #2 (multipart/signed)]


On Wed, 2009-06-17 at 18:52 -0400, cu wrote:

> I realize that 1.9.2 is not a release yet - but 10 times loss of
> performance is something that is not easily overcome. (Personally, I was
> hoping for performance improvement - we can use a few milliseconds, but
> now I'd just like to keep what was there before).

Hi cairouser!

Thanks for your report. Obviously, we don't want to release 1.10 with a
big performance problem like this, so we really appreciate ou going
through the work to test this and letting us know the problem.

The next thing that would be extremely helpful would be to submit a
cairo trace of your application (if possible). Some notes on doing that
are here:

	http://cairographics.org/FAQ/#profiling

But it's basically just a matter of running:

	cairo-trace --profile ./my-application

and then posting the resulting trace file and sending a pointer. You
should be aware that the trace file would allow anyone receiving to
replay and potentially display any graphical results created by your
application, so ensure it doesn't draw anything sensitive that you
wouldn't want to share.

Another useful thing to do would be to run git-bisect to identify which
particular commit introduced the performance regression. This is
precisely what some of us would likely do as soon as receiving the
trace.

Thanks, and we look forward to more details,

-Carl


["signature.asc" (application/pgp-signature)]

_______________________________________________
cairo mailing list
cairo@cairographics.org
http://lists.cairographics.org/mailman/listinfo/cairo

[prev in list] [next in list] [prev in thread] [next in thread] 

Configure | About | News | Add a list | Sponsored by KoreLogic