3DMark question....HELP (1 Viewer)

[Vader]

OK,

I really hope someone can explain this one, because I am at a loss. I just upgraded from Windows 98SE to Windows XP (I know, I know…), and after installing the video drivers for my new card, did my customary 3D benchmark to see how much I gained. That’s where I am baffled.

Here is my system before:
Windows 98SE
Athlon 1700+ (1.465 Ghz)
GeForce 3 Ti200 (64M, 4x)
512M RAM
3Dmark 2000 score: 9550
3Dmark 2001SE score: 6500

And after:
Windows XP Home Edition
Athlon 3000+ (2.1 Ghz)
GeForce 6600GT (128M, 8x)
1 Gig RAM
3Dmark 2000 score: 13550
3Dmark 2001SE score: 13950

Conventional wisdom would conclude that if 3DMark 2001 is much more demanding than 3Dmark2000, it’s score would increase by a smaller margin. But, in my tests, the 2001SE score increases by over 100%, and the 2000 score by a mere 41%. Shouldn’t the reverse be true? Or, do I have a bottleneck on 3DMark2000? TIA!

 

[MikeH1]

Use 3d mark 05, check guru3d.com to see how you stack up and don't worry about 5 year old tests.

 

[D. Scott MacDonald]

Conventional wisdom goes out the door when you are dealing with benchmarks. While useful, they frequently do not test scenarios that match what you need your system to do. Furthermore, drivers almost always contain hacks to detect a benchmark and then contain tweaks for handling them. Sometimes simply renaming the filename will cause a radically different benchmark result (but not always, as there are other ways to detect the benchmark beside filenames). It may be that your drivers have hacks for 2001, but not for 2000 since OEMs probably abandoned that benchmark due to its age.