Looks like Extreme Tech has done a retest of the nVidia/ATi benchmark tests from a week or so ago involving the accusations of cheating. It's a pretty long read, but very interesting. Read it here

So, they left AF on the ATI card in build 330, but left it off in build 320....No wonder ATI performace went down w/ the new build:hotbounce

Pretty funny that that goof also helped expose ATI's driver optimizations, though, how funny!

I bet ATI is feeling pretty unlucky these days. After Nvidia downright cheats in 3dM03, it seems as though ATI is taking all the heat for a bonified optimization

That report is just so funny. Who would have thought that benchmarking software would include *countermeasures* to defeat app detection by 3D GPU drivers. The whole benchmarking thing is just looking so rediculous now. I guess both nVidia and ATI both have a lot to play for when it comes to customer perception of which is faster.

This whole fiasco will probably make it better in the long run for the consumer. Defenses are made to be broken, but in the case of nVidia and ATi, they will probably walk the straight and narrow line because of this, because they know they are being watched closely. And we, as consumers will have to keep using 3D Mark, or at least some benchmark as an indicator of what card we buy. We can all say, use real games to benchmark, but that might be subjective to the eye of the beholder, so to speak. And different computer configurations, run differently with the same card. Mine will run Unreal II, with everything maxed out for quality and 1280X960 resolution, and the game runs great. But, I can't get Ghost Recon, which should run easily on my computer, to run evem pm the lowest resolution and settings.

The bottom line is we probably need a yardstick to look at when we buy video cards, unless we are just going to stick to brand loyalty as some do. I think 3D Mark will be a better test now because of all the so called cheating.