[Ecm-dev] Re: Montgomery coordinated for stage 2

Paul Zimmermann Paul.Zimmermann at loria.fr
Ven 13 Mai 18:56:38 CEST 2005

   From: Alexander Kruppa <alexander.kruppa at mytum.de>
   Date: Fri, 13 May 2005 16:05:53 +0200


   another random thought: why don't we stay in Montgomery coordinates for the 
   root gen of ECM stage 2 if S=1? Having no extgcds would be particularly nice 
   for large base 2 numbers, i.e. Fermat numbers.


Good idea! Do you think it would be a lot of work to implement?
For large numbers (not necessarily Fermat numbers) I'm not sure
the subquadratic stage 2 is optimal. Maybe a more classical standard
continuation would be better.


Plus d'informations sur la liste de diffusion Ecm-discuss