Nvidia's Pascal in 2016... any guesses on pricing?

[tviceman]

50% + performance increases in each tier at about the same price points.
Fastest gpu not exceeding 250 watts.

AMD will be 10% slower in each tier and be priced 50$ lower.
fastest gpu not exceeding 300watts

Happy I wish you were right, but I don't see AMD staying 10-15% behind like they have at 40nm and 28nm. The best air-cooled and cutdown (aka GTX 980 TI) overclocked GM200's destroy the best liquid-cooled Fiji cards (aka Fury X) in performance and consume about the same amount of power when all is said and done. Nvidia left tons of headroom in all the Maxwell chips which make AMD look better out of the box vs. Nvidia than they really are when it comes to FPS. But we all know Nvidia's designs are much leaner; the focused R&D has started to pay dividends with smaller chips, less transistors, higher performance, and lower power usage. AMD's R&D has been slashed time and again over the last few years and is spread out across too many projects to keep up. Nvidia is likely going to double their lead they have now, meaning flagship vs. flagship Nvidia will be 20-25% faster at stock speeds and 40-45% faster OC vs. OC unless AMD hits their design goals out of the park and Nvidia flubs theirs up.

 

[tviceman]

Sorry about that, you are correct there. That being said, 75% faster than GK104 might be stretching it. When the GTX980 came out TPU had it as 50% faster than the GTX770 at 1080, and even in the 950 review from a couple months ago it was 60% faster. There's a few newer games where Kepler really tanks vs Maxwell, but debating why that is is the subject of another extremely long and bitter thread.

Considering the GTX980 is a hair under 50% larger (in transistors) and clocked over 10% faster that GTX770, it would be tough to expect a 20% increase in performance per transistor from uarch alone in the general case. That doesn't preclude specific cases if a game uses some particular new feature that Maxwell doesn't do well, of course.

I'd love it if I were wrong here though.

Lets not compare clock speeds because that is not apples to apples. GTX 980 consumes about the same amount of power at higher clock speeds than GTX 770, which means that it was designed to run at higher clock speeds to achieve it's performance. Power consumption is the best measure of comparison since both cards were designed with a similar TDP and were NOT designed with similar clock speeds.

That said, I am fully aware that Nvidia has been maligned over Kepler performance as of late, but reviews over time show Maxwell pulling ahead before all the crops of DX11 games showed up. From the time GTX 980 came out to early January (less than 4 months), GTX 980 gained another 10% performance over GTX770 with largely the same benchmark suite.

Also keep in mind that Maxwell's architecture has noticeably more headroom than Kepler. Nvidia left plenty of performance on the table and if AMD had actually released a well-rounded product in Hawaii, GTX 980 could have come out of the gate clocked higher.