In short, a 7950 and 660 are neck and neck. The 660 performs better in some games, the 7950 better in others.
Hitman Absolution: The 7950 wins almost every benchmark here.
http://us.hardware.i...tman-absolution
Aegrus, that's not correct. That was a 7950 Boost, which is significantly faster than stock and costs more than a 660 TI.
In addition, the 7950 spiked higher and ran lower, with less framerate consistency in milliseconds. Lower milliseconds = better.
The 660 TI actually traded blows in that benchmark with a 7950 Boost.
Oooh. . . you seem to be correct. Sorry. D:
Erm, well, I suppose that means the 660 TI is just flat out the better choice, then, considering it has on average the same performance as a 7950 boost, but also costs $50 less and can of course run Phys X effects whereas an AMD card can't.
Keep in mind that, for all these benchmarks, Nvidia's comparable performance for the price is often achieved while it's running phys-x, whereas most users will lose even more fps than the benchmarks suggest with an AMD card, since their CPU (which is often not as good as the i7's that the benchmarks use) will be choked by Phys-x.
Edited by Aegrus, 21 August 2013 - 01:30 AM.