Makes sense to me. If you had also benched with 3dMark05, 66.93 probably
would have been the fastest. HL2 just came out two months ago. So take a
guess what benchmark is more relevant? That's right 3dMark05.

DaveL


"Lars-Erik Østerud" .@. wrote in message
news:nhXGd.1619$VR2.87@amstwist00...
OK, did a bit of testing on my Ti-4200 and here are the results:

3dMark 2000 3dMark 2001 3dMark 03 Remark
44.03 12182 11082 1473 1)
45.23 12056 10472 1456
61.77 11467 10013 1596 2)
66.93 11309 9845 1593 3)

First comment must be that it's strange how newer drivers are slower
and slower with 3dMark 2000 and 2001, but faster with 2003 ?????????

What is most important/relevant for games like Quake2, Half-Life,
Half-Life 2, Need for Speed (alle version)? The 2000/2001 or 2003?

1) 44.03 stutters like hell on the end sequence in Half-Life 2
The other drivers all did this very well (45.23 smoothest).
Maybe the others have less stuttering inside HL2 as well?
But this driver is still the fastest in 3dMark 2000/2001.

2) 61.77 rendered the "point sprite" (the horse) in 3dmark 2001
completly wrong. It got transparent and did NOT look good :-(
Still (or maybe because of) it the fastest in 3dMark 2003.

3) 66.93 has a very nice funtion making it possible to set both
quality settings and brightness/gamma/color for each game.
No need for PowerStrip anymore (some things work even better,
like 2xQ AA with HL2, not possible with HL2 or PowerStrip
settings, but when set inside 66.93 it works very very well).
But the 66.93 was also very much slower in 3dMark 2001 :-(

--
Lars-Erik - http://home.chello.no/~larse/ - ICQ # 7297605
Win98se, Asus P4PE, 2.53 GHz, Asus V8420 (Ti4200), SB-Live!