View Single Post
Old 05-15-03, 03:56 AM   #82
Registered User
Join Date: Nov 2002
Posts: 104

Originally posted by nVidi0t
... I think 3dmark03 is a ****ty benchmark anyhow. A CPU speed difference of 600-700mhz (which I have tested) makes not a single difference to your 3d Mark score, or miniscule differences to selective tests...
You're right. Higher cpu speed has little impact on 3dmark03. And that's a good thing.

Q: Was 3dmark created to test the performance of the cpu? Or to test the performance of the vidcard?
A: It is meant to test the performance of the vidcard.

These two statements are true:
1. Changing cpu has low impact on 3dmark03 results.
2. Changing vidcards has dramatic effect on 3dmark03 results.

Conclusions: 3dmark03 is an excellent method of measuring vidcards. It is a poor way of measuring cpu's. Therefore, 3dmark succeeds as a vidcard benchmark.

If you want to measure overall system performance, then you need to use a different benchmark. If you want to measure how well a vidcard performs with dx 9 pixel and vertex shaders, 3dmark03 is an excellent tool.

btw, nv40: PS 1.4 is directX 9 spec...if the FX can't handle dx 9 specifications, whose fault is that?
ZoinKs! is offline