|Bill Allombert on Wed, 03 Dec 2008 04:48:23 +0100|
[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]
On Tue, Dec 02, 2008 at 06:33:17PM +0100, Karim Belabas wrote: > I'd have a look at "ideallog" ... > > As you mentionned, the bottleneck is the computation of discrete logs in > (Z_K / P) for maximal ideals P dividing I. This will be slow if Norm P - 1 > is not smooth ( basic Pohlig-Hellman + Shanks ). Well, a basic index calculus implementation for znlog would beat the current znlog, use less memory and be easy to implement (as long as we do not implement sieving/Lanczos/etc.). Is it worth the trouble ? Cheers, Bill.