Originally posted by: Lonyo
I don't see any dual core chips, or anything from AMD.
I'm sorry you feel that way, but if you actually look at the graph instead of reading what it's about then dismissing it out of hand, you would see that as the data set gets larger, the GPU doesn't slow down anywhere near as much as the CPU does. The CPU graphs appear to be some form of a logarithmic curve, whereas the GPUs remain approximately linear.
A second core isn't going to help you there. Even with a second core and
100% efficiency, which almost never happens with SMP, you'd still have the nVidia GPU twice as fast, and the ATi almost 4x as fast. For the large data sets, of course.
An overclocked-to-hell-and-back AMD X2
might be able to keep close to the nVidia GPU, if the sorting algorithm can even be made to be multithreaded...
At any rate, none of this disputes the fact that modern GPUs are hella fast.