Don't ya think that ATi could have at least given nVidia a full day of thinking they were going to get the crown back?!? "Go ATi! Go ATi! Go ATi!" ;)

Keep releasing the new hardware guys! (at and nvidia) It'll push the 9700 pros down to a price I can afford :p

it doesnt get faster just because its 0.13! and its cheaper with 0.15 too

ATI has no choice. They've only won once in the battle in the entire history of their products. Does anyone else note how crazy the Inquirer is?

"The RV350 has shown that." [on .13u usefulness in increasing frequency] This leaves out the fact that RV350 has half of the pipes in the R300 design disabled. This should help scalability and make the transition to .13u easier. This logic does not hold true when keeping the entire design intact, apparently. This does not mean that the R300 function set cannot be restructured to operate on .13u at higher frequencies. I believe there are facts and rumors that suggest the original R400 plans have been scrapped and certain "discoveries" were made that will be incorporated into the current R300 series while R400 gets new design implementations. It would seem sensible that if significant additions are made to the R300, that would also be the time to redesign the core for .13u operation. The name I have seen for this tweaked version is R390. It seems logical to expect it on .13u and at a higher core clock than R360.

I'll bought the best card for DooM III & next gen games. So, I'll wait till end of summer & get the fastest card. ATI or nVidia, i don't mind. I only want the fastest.

For reasons of fairness you should keep in mind that whereas the drivers may not be optimized for Doom3, Doom 3 has been built from the ground up to be optimal on Nvidia hardware. The special code paths are evidence of this effort to make the game take advantage of architecture specific features. This, of course, does not rule our Radeon code paths or stronger adherence to the standard ARB path by ATI. It is not in ID's interest to limit their customer base, but coders, notably John Carmack, are limited in their time to bring a product to market and may therefore prefer certain implementations. On the hardware side of things it seems that the R3xx series is successful because it is an open architecture of great brute force. The NV3x series seem more dependent on the code to be optimized, but when it is it can overtake the R3xx in performance. This is somewhat analogous to the Athlon XP versus the Pentium 4, where the Pentium 4 is the less 'forgiving' CPU, but has features that, when implemented, allows it extremely high performance.