Originally posted by JoshSB
I was under the impression that the 8800s ran on a new architecture that sped them up compared to the older architectures (7000 line and 1900 line) Please correct me if I am wrong.
You're quite correct, actually.
The GeForce 8 line of GPUs is very different from the older families. Instead of having specific processing units (vertex/pixel shaders, texture units, etc), each doing one job, the GPUs now feature a large collection of "unified shader processors". Instead of having specific pipelines doing some function, there's now a unified pipeline that can be "programmed" (by the app being run).
Clock speed of the core / memory cannot be used an indicator of performance (and it never has been the case in the past either... pixel pipelines and shader units were usually a better indicator; but now it'll be too difficult to compare pipelines across GPU families).