While a lot of people focus on the floating point and integer processing architectures of various kinds of compute engines, we are spending more and more of our time looking at memory hierarchies and ...