Adam K Kirchhoff wrote:

I had some time yesterday and thought I'd do a quick comparision of
the DRI drivers and fglrx drivers for three different cards I have,
and I thought others on this list might be interested in the results.
All tests were conducted on a dual 2.8 xeon, with a gig of RAM.  The
cards are a 9600AS with 256 megs of RAM, a 9000Pro with 129 megs of
RAM, and an X700 with 256 megs.  After each test with the DRI
drivers, I rebooted and ran the exact same one with the FireGL
drivers.   The DRI drivers are from Mesa (and DRM) CVS from Thursday.
The FireGL drivers are the latest available at the moment (I don't
remember the version number off the top of my head).  There have been
no game specific changes made in driconf.

Using "timedemo demo1" in doom3 (640x480, "low quality" settings),
this is what I got:

9600 - fgl - 13.4 FPS 9600 - dri - 15 FPS

9000 - fgl - 14 FPS 9000 - dri - 17.6 FPS

X700 - fgl - 13.8 FPS X700 - dri - 14.7 FPS
The fgl numbers are downright awful for the r300 based cards (obviously
more so for the X700), though the dri performance isn't good neither (but honestly you really should expect better performance from the ATI drivers at this point probably). The 9000 scores aren't quite comparable unless you used r_renderer arb on both cards, and you didn't take the rendering errors with dri driver which still are present with hw tcl into account. Though, with Mesa cvs since about last friday, r_renderer r200 actually works on r200 :-) (need to enable software ARB_vertex_program, need to enable 6 texture units, need to disable hw tcl, and for somewhat useful performance also need to use x86 optimized build with MESA_EXPERIMENTAL env var set with a cpu that supports sse...).

Using umark for benchmarking UT2004 (1024x768 with all "low" or "very
 low" display settings)...  First DM-1on1-Albatross:

9600 - fgl - 11.378239 / 35.393394 / 82.763985 fps - Score =
35.407494 9600 - dri - 5.033368 / 8.846323 / 17.930601 fps - Score =
8.850811
Nothing new here, dri performance is still limited by the lack of ARB_vbo or at least decent drawArrays performance avoiding fallbacks.

In retrospect, I should have done ut2004 at 640x480 to get a better comparison of how it stacks up to doom3.
dri numbers would just stay almost completely the same, you can configure ut2k4 however you want but the amount of vertices sent remains pretty much constant.

Roland



-------------------------------------------------------
This SF.Net email is sponsored by xPML, a groundbreaking scripting language
that extends applications into web and mobile media. Attend the live webcast
and join the prime developer group breaking into this new coding territory!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642
--
_______________________________________________
Dri-devel mailing list
Dri-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/dri-devel

Reply via email to