Originally Posted by Linuxhippy
I wrote a micro benchmark which isn't KDE based at all. Its a simply program drawing tons of subpixel-AA strings.
In this benchmark my old Duron outperforms my new Athlon64 by a factor of two - keeping in mind that the Athlon is 4 times as fast, this is a regression that leads to a performance reduction of about 800%!
So I don't speak about bad performance or not implemented accaleration, I speak about a performance regression. Till you don't get this I would be pleased to not read replies any more
Regression? Several versions ago, subpixel AA was actually accelerated. It was disabled since then, because it rendered improperly. You're talking about that QT program that draws fonts that was posted before, right?
Can you do me a favor?
Post your .fonts.conf file in your home directory on both PCs. Also, what font are you using, on both PCs? Can you try arial? Arial has byte code info in it, so it makes for a good test.
There has to be something that is different... A duron beating out an a64 running the same code...it seems impossible.