_Consistent_ benchmarking in 3dMark 2001 ?...

[VBboy]

I know I said some things about 3dMark I didn't mean..

I want to benchmark with it, but what display settings do you use?
I am talking about the advanced settings you can set under the "Display Properties" (right-click your desktop). Which MIP-Mapping level do you use, what Anisotropic Filtering sampling do you set, etc? Wouldn't people with the same exact videocard and CPU get different results, based just on their different display settings?

So how do I benchmark my fresh Radeon 9700 Pro against other Radeons 9700 Pro under the SAME conditions?

Thanks..

 

[DeepSix]

Don't apologize, everybody knows 3DM sucks. The main reason it's so popular is because it's free and easy to run, so almost everyone uses it. And because almost everyone does it's tuned into an overall mark to compare one system against another. But just because you get a bad score in 3DM doesn't necessarily mean all your games will run poorly either, and vice-versa.

I use the default 10x7x32 but I run it triple-buffered and turn off all tests but the actual game portion of the mark. Too many times my sys has locked up in AdvPixShader (which is a sign you're clocking the card too high). There are many other tweaks you can make to your vidcard and your PC/CPU that will increase your score but it really doesn't mean anything except you know how to tweak your sys to get a higher 3dM score, it has nothing to do with how well any game will run on it. The real secret to a high 3dM mark, buy the fastest cpu and vidcard you can. Good Luck! (Turn off everything, set for bi/tri filering, turn off FW/SBA and set mipmaps in DX8 to 1.3)

 

[MichaelD]

The universally accepted way to to benchmark using 3DMark2001(SE) is to use the default settings. To ensure you are doing this, open 3DMark. Top-most-righthand button is "Change." Click that. Lower-right-hand-button in that window is "Defaults." Click that. Click OK. Click Benchmark. Done.

 

[VBboy]

Well, there are also the Display Properties settings. How do you make sure that everyone has the same settings there? This is similar to benchmarking the same hard drive in 4 different machines, with the hard drive having very different level of fragmentation on each PC...