Fun stuff. But why only take n up to around 4 million? Not too long ago, I was able to compute the billionth Fibonacci number on a pretty ordinary machine, usi

Fun stuff. But why only take n up to around 4 million? Not too long ago, I was able to compute the billionth Fibonacci number on a pretty ordinary machine, using a Haskell implementation of the fast doubling method. (The number took 20 minutes to print out in my console window.) By the way, I did a Python implementation, too. And I learned that GHC large-integer multiplication is a lot faster than CPython large-integer multiplication. I wonder why. (No, it has nothing to do with dynamic typing; almost all of the time is going to be spent executing internal integer-multiplication routines written in C.)