Originally Posted by Mad Pistol
The GTX 680 pushes itself over stock on certain programs. It's not the same thing. The GTX 680 is clearly overclocking itself on certain programs to above the stock 1006mhz. This is overclocking.
Nvidia probably measures TDP from the stock clock, not the turbo state. That means power consumption is higher than what they stated in their briefing slides.
This is pretty much exactly how Turbo works on Intel and AMD CPUs, but no one considers the turbo speeds to be an overclock, as they are part of the default spec.
Also my Radeon 69xx cards will throttle themselves to below stock clocks in some programs (crysis, metro 2033, a few distributed computing projects, and any strenuous stress test...which is pretty much the same thing.
As for power consumption, turbo clocks and OCP are usually based on some sort of TDP limiter. TDP does not refer to power consumption, it refers to the ability of the cooling system. It's likely that turbo clocks do not cause the card to overshoot TDP, but take advantage of TDP headroom like pretty much every other turbo implementation.
Originally Posted by Mookster
The point of this is how much more can you get out of the 680? We already know how much more we can get out of the 7970, and it's a lot, and that matters before we can make a decision.