Re: Fast Fibonacci method


#1

Dear Erlercw,

if you want to implement the Fibonacci calculation to show some
benchmark using the recurrence formula F_{n+2}=F_{n+1}+F_{n},
this isn’t for you.

But I’d bet that the formula 1.3.3 in

http://www.math.upenn.edu/~wilf/gfology.pdf
(http://www.math.upenn.edu/~wilf/gfology.pdf)

is faster (on any machine/computer language ) as it involves
multiplying the
same factors all the time. One could just do one multiplication
using logarithms, actually, and then convert back the result.

Best regards,

Axel


#2

On 27/05/06, removed_email_address@domain.invalid removed_email_address@domain.invalid wrote:

(http://www.math.upenn.edu/~wilf/gfology.pdf)

is faster (on any machine/computer language ) as it involves multiplying
the
same factors all the time. One could just do one multiplication
using logarithms, actually, and then convert back the result.

Best regards,

Axel

Thanks for your suggestion. I tried various methods of implementing
that
formula, taking care to get exact results, but I can’t figure out how to
speed them up enough.