Re: new factoring algorithm

From: Colin Andrew Percival (cperciva_at_sfu.ca)
Date: 05/30/03 

Date: 30 May 2003 00:07:00 GMT

Bill Unruh <unruh@string.physics.ubc.ca> wrote:
> assymptotically the Number field sieve is faster already than the
> Quadratic sieve. ( exp(L^1/3 lnL ^2/3) rather than exp( L^1/2 lnL^1/2)
> where L is the length of the number if I remeber correctly)

  GNFS is O(exp(C L^(1/3) (log L )^(2/3))) with C slightly above 1.9.

Colin Percival