>>20
lol @ inefficient algorithm
see >>7,10 for better ones.
especially the prime-swing one in >>7. implement that (with a better prime sieve) in assembly and unless you do something incredibly stupid it should be faster than GMP's factorial function.