Quote Originally Posted by cegras View Post
I might as well repeat myself: per watt, ATI's architecture is more efficient.
you do realize that an architecture does not determine performance per watt because it is intangible? its more of a battle of process v. architecture. nvidia has a very impressive dx11 arch but it has had trouble on 40nm. nvidia's yields dont allow them to bin gtx480's at the tdp they would like. we could see a silicon spin like B1 or gf100 implemented well on 28nm and it will be an excellent competitor.