11825 frames in 5.0 seconds = 2365.000 FPS
5703 frames in 5.0 seconds = 1140.600 FPS
5998 frames in 5.0 seconds = 1199.600 FPS
6172 frames in 5.0 seconds = 1234.400 FPS
6237 frames in 5.0 seconds = 1247.400 FPS
6189 frames in 5.0 seconds = 1237.800 FPS
6120 frames in 5.0 seconds = 1224.000 FPS
5487 frames in 5.0 seconds = 1097.400 FPS
6158 frames in 5.0 seconds = 1231.600 FPS
5551 frames in 5.0 seconds = 1110.200 FPS
6225 frames in 5.0 seconds = 1245.000 FPS


My specs:

1600x1200, 24-bit color depth, OpenGL 3D acceleration enabled

* Abit NF7 nforce2 mobo (the nforce2 drivers I d/l'ed from nvidia refused to install)
* AMD Athlon 2100+ XP CPU o/c'ed to 2700+
* SB Audigy MP3 sound card
* 1GB Crucial/Micron CAS 2.5 DDR SDRAM (PC2700/333MHz)
* Western Digital 40GB HD ATA100 with UDMA mode fully enabled
* BFG Nvidia GeForce4 ti 4600 VIVO video w/128MB DDR RAM using Nvidia's .4363 video drivers
* SuSE Linux 8.2 Pro out of box without any patches or fixes having been applied

What can I do to improve performance, or am I a nutcase and should accept this? (I'm still freaking over how the first value dropped below 50% by the 2nd result!)

Thanks!

What window manager were you using, what was running, what other screens were under/touching the glxgears one?
It's a very delicate test

Well do you not have to take into account other processes running on your system?

If your system has just started up some cron job to index your web server files, or whatever, then you are going to see variations in your results.

Ideally to do a proper test, you would have to shut down all of the uncessary daemons running on your machine and stop any cron jobs, not to mention other users if you have any of those on your machine.

When you are doing this, you should be running `top' to see what else is running, and remember that by running top you will then affect the result.