Anyone that found the old CCC tree style confusing doesn't deserve to be writing articles on a hardware site.

Good article, but I think it should expand beyond graphics. One of the major drawback of Vista, that was never fixed with the SP releases, was poor networking performance. I was using Vista recently but decided to switch back to XP for gaming (use linux for everything else) because of certain issues I was having with Vista, like poor surround sound support and some networking issues. I play on the same CoD4 server almost ever night and just by switching to XP my ping dropped from the 70-80's to the 50's and low 60's. How about a comparison of all 3. I am using the 64bit version of XP (which is essentially Server 2003).

no I didnt miss the point. Just saying I would like to see more comparisons of the two.

If you guys are not going to provide pics showing the image quality differences between competing cards then may it's time to discontinue the use of AdAA and TrAA. Just benchmark the game as it was intended by developers. Like I've said before, everything else in your testbed looks fine IMO.

But lets look at this for a minute. Your testbed shows you using AMD's highest setting possible: Adaptive AA: Quality. While you are using Nvidia's lowest quality possible: Antialiasing, Transparency; Multisampling. Which to me doesn't really make a whole lot of sense here. Again, no IQ comparison pics are provided.

When using this feature from both camps both Quality and Supersampling are the highest quality settings available. What that means is they offer the highest performance hit in favor of offering the best IQ to images with transparent textures. However this is in those games that allows it's use (not all of them do from what I know).

You see what I am talking about here? You are using AMD's highest settings (which provide the highest performance hit) to Nvidia's lowest setting (which provides a minimum performance hit) in games that allows it's use to reduce jaggedness of images with transparent textures. In a review more tailored in providing performance results (among other things) to us the viewers. Which IMO can get confusing if the performance hit noticed in a game is the direct result of it's use. Because there are no IQ comparisons offer to justify why that is.

In ET:QW benches, it says, and I quote: "But in two cases transition to a newer OS causes a serious performance drop for Radeon HD 4890. In 1280x1024 its average readings lower by 21%, and in 1680x1050 – by 9%". But in graphs I don't see it... I see a 1 fps difference, on a 130+ fps, so it's neglible. Have you posted wrong set of graphs, or have you been drinking while writing comments about them?
EDIT: OK, finished reading, nice article overall. There are two more things that should be added to clarify, I presume Windows 7 was also 64bit, but it isn't mentioned in "test system", while Vista says 64bit clearly.Also, at the end when saying that "you should install newest drivers", you forgot to mention one big fix in Win7 with newest Catalyst, and that's the ability to clone/expand desktop to second screen, which didn't work in RC+9.6 driver.And for the end of this comment - too bad you haven't obtained a RTM version of Windows 7 to test. Because despite RC being rock stable, and truly great OS, final version always is - FINAL. So I hope you'll repeat at least part of these benches once you get your hands on RTM

Your results for HAWX for AMD DX10.1 don't make any sence at all. All other sites report up to 30% superiority for AMD solutions over nvidia but in your case AMD loses about 100% to nvidia. Sorry, I don't buy it at all. Perhaps AMD and nvidia results switched places.

Resolution tweaking didn't work in the previous Win 7 versions so i had to install the Vista version for nVidia.
Will have to recheck with the newer versions. all in all, it seems there's no Huge gap for win7. And there's this fact that games like NFS Undercover never really works well with Win7.

Were the default Microsoft drivers used for the Intel northbridge or Intel's latest? Microsoft's built-in are very old (2006) Vista drivers. I bet updating the drivers to the latest off Intel's website would make the Windows 7 numbers look better.

Just installing Intel's setup won't do it on Windows 7. You need to extract the setup using 7Zip and force manual updates to that folder on all the PCI and PCI-E items in Device Manager.

It would be nice to have details on motherboard drivers in any future reviews you have comparing the two operating systems.

Another fake review. 30fps in HAWX, or 35fps in Far Cry 2? at the same 24" resolution and settings but with 8x aa i get more fps with my 4870 at 790MHz for the GPU.Over 65 average fps in Hawx and about 45 in Far Cry 2.

Ohh and stop using those benchmarking tools at last that Nvidia spends it's time optimising their drivers for them, and start running Fraps.

There's neither honesty nor professionality in this review.