Quote Originally Posted by DilTech View Post
Quite simple... If the R600 had been able to compete with the 8800GTX, do you really think the next cycle from NVidia would've been the 9800GTX and that ATi would've followed the 2900 with the 3870?

The cycle we were on was 100% increase(in some cases more) for multiple years, as soon as one company couldn't sustain it that cycle dropped the very next round... Doesn't take rocket science to figure out why, as NVidia didn't have a push to keep it up.
No. No and no.

...unless current µArchs would be much more "efficient" than what they are now.

How can AMD make faster cards now? Increase everything. Slap faster RAM and GPU clock. It yields bigger dies, which alone lower the yields.

Basically regardless of R600's hypotethical success, we wouldn't be having any better price/performance ratio, or any better performance/watt ratio, or any better performance/mm² ratio.

The 40 nm process' leakage, transistor density and yields define the speed of the cards.

Actually, if anything, the "slowdown" made game developers to slow down too, which now means better FPS's for us on those games, though possibly at the expense of IQ.

The ONLY way which could have changed how the things are now, is the efficiency(perf/mm²) of the R600 core and it's derivates(Anything until R900(?)). That would've meant that Nvidia would been severe underdog. With current µArchs and TSMC process it's impossible to get any better results really. There are dozens, possibly hundreds chip architects doing everything they can to get better stuff out.