It's been a concern to me, probably since the HD69xx reviews, that OC3D's performance graphs don't make the sense that I'd personally expect with comparisons of cards vS card vS oc'd cards.
Either that or cards since I've been seeing the variations are self governing themselves via their respective bioses ? I can't imagine that to be the case.
I wonder if you guys could either internally review the reviews or have some kind of 1-off face-off between those cards typical in todays review comparison.
In simple terms, you bench a card at stock, then bench it overclocked, it's hard to imagine how on earth it could bench with a lesser performance. (just for example) For sure, if you overclock a card and something within the OS kicks in (like windows just loves to do) it can certainly hamper performance. I've noticed particularly with failed overclocks, Windows 7 runs a program called something like "ApSafetyBlahBlah.exe" that you have to check is not running before you bench again cos it does take up cycles. I think it's Windows trying to help the user not have the same problem again that might've caused the system failure.
Just a suggestion, not having a poke.
Tis a decent card btw, used to love Gigabyte's productions. But instinct would dictate buying the standard card and overclocking it - as mentioned above, it's a hard thing to suggest looking at some results.