[Haskell-cafe] speeding up fibonacci with memoizing

[ajb at spamcop.net]

G'day all.

Quoting Stefan O'Rear <stefanor at cox.net>:

> Prior art trumps all.  (by a few %)  granted it doesn't do much memoizing
> anymore :)

Ah, butbutbut... of course the Gosper/Salamin one is going to be
faster if you only compute one Fibonacci number per instance.  The
memoed version is optimised for programs that want more than one.

Cheers,
Andrew Bromage