Re: faster log factorials

[Brian Gough <bjg at network-theory dot co dot uk>]

James Theiler writes:
 > That is, Lookup is 7x faster than Stirling, which in turn is 3x faster
 > than Lanczos.  To my mind, this is an argument in favor of running up
 > the upper bound on the table for log(n!); further, whenever n exceeds
 > whatever upper bound we choose, we should use Stirling instead of
 > Lanczos for the large n result.

A version with a larger lookup table could go in the new contrib/
directory.  Some people might want to optimise for space instead of
speed.

I think 170 is a good natural limit.  

-- 
Brian