radeonhd vs radeon, 2400HD vs 9200SE

[Clemens Eisserer]

>  Either glxgears is a totally bogus test or those system are not
GlxGears measures only very few aspects, so yes, from a benchmark
point of view, its completly useless.
And because if its high throughput of small rendering primitives, its
very CPU dependent.

lg Clemens

2008/4/8, Edgar Toernig <froese at gmx.de>:
> Zoltan Boszormenyi wrote:
>  >
>  > Gene Heskett írta:
>
> > >
>  > > I just swapped out the new 2400HD card for the old 9200SE, rv280 based card,
>  > > and while I saw some sort of a notice go by during boot to a non-x login
>  > > screen, and startx killed the monitor for about 5 seconds, but other than
>  > > that its working MUCH better than when I took it out at least a year ago.
>  > >
>  > > So at a glxgears of 650 frames,
>  >
>  > This doesn't sound right, my Radeon 9200SE produced about 960fps
>  > with glxgears, my current X700 gives around 2100fps.
>
>
> Hmm... I often see these low numbers and don't understand them.
>  configured correctly: I get over 3800fps with my onboard 9100 IGP
>  (24bit color depth, more with 16bit).
>
>  Even while watching TV in parallel (DVB-S, software MPEG decoder,
>  fullscreen XV overlay) glxgears stays well over 3400fps.
>  [Btw, mplayer's GL-renderer with GL_ATI_fragment_shader for YUV
>  to RGB conversion (-vo gl:yuv=5) works well, too.]
>
>  Driver:
>  OpenGL renderer string: Mesa DRI R200 20060602 AGP 4x x86/MMX/SSE2 NO-TCL
>  OpenGL version string: 1.3 Mesa 6.5.2
>
>  Ciao, ET.
>
>
>  _______________________________________________
>  xorg mailing list
>  xorg at lists.freedesktop.org
>  http://lists.freedesktop.org/mailman/listinfo/xorg
>