As long as it looks good enough and it runs smoothly for me, I don't see what the big deal is all about.

Where's the PowerVR Series 5 cards? Or the Volari crap cards?

I want someone who can actually produce a faster product with image quality at its best. Not the crap both ATI and Nvidia are pulling. You think with every new product ATI and Nvidia could increase image quality.

Whats the big deal? everyone cheats and if cheating makes a peice of hardware work better on a certain computer game then fine!

isnt it good for the people who buy that non-specific peice of hardware

And another thing...

It's a very mild optimization at the default levels, so of the order of a few percent. But this is a neat algorithm - we encourage you to take the texture slider all the way towards "performance" - we don't think you'll be able to see any image change until the very end, and the performance increases are significant.

Uhh, yeah, every time I've put the "texture preference" slider at anything below "Quality", I've seen dynamic texture compression artifacts (chroma shift particularly) all over the place. It would be better if there was a master switch for texture compression, as well as the "texture preference" slider. Or better yet, more individual settings sliders and better labels for them.

What I'd prefer to see is a benchmarker mode in the drivers that takes away all the sneaky but legit stuff (brilinear filtering, proprietary isotropic filtering algorithms, and so on), so an objective analysis can be made. Gamers can have their speed, and benchmarkers and purists can have honesty from the drivers when they want it. Then the reviewer could put it through a run without the "benchmarker mode", and then report how much of a speed benefit the "optimizations" give, while making sure that the reader knows not to make direct comparisons between "optimized modes" from different manufacturers.

Unforgiveable practices include screwing around with miplevels (a la Quack3), shader substitution for benchmarks (a la 3dMurk), or purposely evading tools used to detect qualities that can't be detected with standard rendering (i.e.: colored miplevels). ATI's description seems to say that colormips demand trilinear while regular rendering works right with their adaptive method; this should've been noted in a page of benchmarker's notes so they know it's not intentional cheating.

As for me, I'm never going to take the bait in April, May, June, October, November, or December. There's too much marketing noise during those times to make an educated decision; besides, the last cycle's gear goes down in price when the new cycle hits.

Oh boy

Even I'm starting to get tired of this.