The versions of Fibonacci compute the same result. Which ones run faster?

The * order* of a computation, written *O(f(n))*
(sometimes called
``big O'' notation) describes the time required to perform the
computation for a problem of size *n*, as *n* becomes large.

The iterative ` (fibonacci n)` and recursive ` (fibon n)`
are *O(n)*, that is, the time they
require is linear in the size of the input. We assume that the
operation ` +` takes constant time, which is not strictly true in
this case.

What is the order of the recursive function ` (fib n)`?