"fastfib" in the gf package faster than "fib"



Mine took 1.75 seconds compute pi * e where both were computed using fpprec of 1,000,000.  Typically quantum mechanics time evolutions in 3d are really scary.  I have not even tried it.

Rich




 ------------Original Message------------
From: "Stavros Macrakis" <macrakis at alum.mit.edu>
To: "Richard Hennessy" <rvh2007 at comcast.net>
Cc: "Maxima List" <maxima at math.utexas.edu>
Date: Tue, Jul-8-2008 2:54 PM
Subject: Re: [Maxima] "fastfib" in the gf package faster than "fib"

By the way, the n*log(n) algorithm you're probably thinking of,
Sch?nhage-Strassen (actually O(n*log(n)*log(log(n))), has a crossover
>> 10^10000. (http://en.wikipedia.org/wiki/Sch%C3%B6nhage-Strassen_algorithm).

By the way, on my generic laptop, Maxima currently multiples
10,000-digit numbers in about 1 mS and 10,000,000-digit numbers in
about 4 S -- not so bad for casual calculations.

          -s