Quote Originally Posted by Solus Corvus View Post
It is clear that you had to pick the obvious outlier to make your point. All the other cards maintain their relative positioning between the game and benchmark, with only the magnitude between them different because of the wildly different scenes being rendered. Also the Nvidia AND ATI drivers are different.
WTF are you talking about? The relative performance is off by more than 50% in some cases and by more than 20% in most cases.
Don't you think that a difference of 50% in performance would have a "slight" impact on your graphics card purchase decision?

There was also NEVER a Catalyst release that improved performance of Shattered Horrizon, but only one fix for a graphics bug.

So how in hell is this not a prime example that 3DMark is a horrible thing to use to compare performance?