[speeder]

I doubt that you can have a much increase in performance as cores increase unless you are calculating numbers with huge amount of bits.

[tbrownaw]

Considering that factorial of 1e6 has about 18e6 bits (and factorial of 1e3 has 8.5e3 bits)? Yes, any factorial that doesn't have a huge amount of bits will be fast enough to calculate that there's not much point to parallelizing it.

[PurplePanda]

as the size of the input to the procedure increases you will indeed be calculating with numbers with a huge amount of bits