I was confused on how different cores at different levels of price and speed achieve these differences with the same architecture. I looked into it and it seems that the idea is that many features for the higher performance and more advanced cores are disabled in the lower end cores. In a way this is the same architecture but different implementations of the same architecture. Implementation vs abstraction.
@colorblue. While it's true there are implementation differences in the SMM across this line, really the fundamental difference among the GeForce chips (the three on the right) is the number of SMMs in each chip. You'd be fine to think of them as the SMM core, replicated different numbers of times.