Graeme Geldenhuys wrote: > 2008/5/17 Graeme Geldenhuys <[EMAIL PROTECTED]>: > >> That is strange. The first complaint I here that it is slow. I ran >> > > OK, I did some more Google'ing. It seems Cairo does have some > performance issues, but I can't remember to which version of Cairo > they used. What I did notice from all my searching is that it is > really unfair to compare Qt vector graphics to anything else! Qt is > simply miles ahead of anybody else in performance. See image. > > http://chaos.troll.no/~zrusin/vectorframeworks.png > http://zrusin.blogspot.com/2006/10/benchmarks.html >
Probably Qt is still faster than cairo but the gap maybe smaller. The test was with an aged cairo version. Recently a lot of optimizations are being done in cairo by taking advantage of using mmx/sse. There's also an GSoC project to improve the polygon rasterizer. Unfortunately, the OpenGL backend is still experimental but there are plans to doing hardware accelerated. A quote from a Mozilla/cairo dev: "Yep, the plan is to move to hardware accelerated rendering using OpenGL and/or DirectX as appropriate. We probably wont’ be using Glitz though, but will develop a different backend for OpenGL/DX as glitz is heavily unmaintained at the moment and there are some things we’d like to do differently. But work on that probably won’t start for another few months at least." Under MACOSX, is already hardware accelerated since uses quartz as backend (http://blog.vlad1.com/2008/03/16/cairo-16-quartz-and-gecko/). One of the reasons of Firefox 3 speed up (http://www.computerworld.com/action/article.do?command=viewArticleBasic&articleId=9068118&intsrc=hm_list) is the use of cairo (http://blog.vlad1.com/2007/12/11/graphics-in-mozilla/) Luiz _______________________________________________ Lazarus mailing list [email protected] http://www.lazarus.freepascal.org/mailman/listinfo/lazarus
