Thread: Idea for faster sieve View Single Post
2007-06-10, 22:45   #33
geoff

Mar 2003
New Zealand

100100001012 Posts

Quote:
 Originally Posted by Citrix Any progress on this. May be you din't understand my last post or there is some confusion on what I wanted.
No, I just haven't got around to it yet :-). RieselSieve are rapidly approaching the 2^52 limit of the current x86-64 mulmods, so I am working on extending that limit.

BTW I think the existing method I am using for sieving numbers mod 2^y will fail when p > 2^(64-y). For example when sieving GFNs with k=3^16 the limit for the sieve will be 2^59 instead of 2^62.