Interesting look at how to incorporate the cost of random access into complexity estimates.

In short: if you try to squeeze too much L1 cache onto your CPU it will eventually collapse into a black hole, and that would make it awkward to get the results of the computation back to the user.

The Myth of RAM - Part 1
The Myth of RAM - Part 2
The Myth of RAM - Part 3
The Myth of RAM - Part 4