View Single Post
Old 2016-02-22, 11:32   #4
Till
 
Till's Avatar
 
"Tilman Neumann"
Jan 2016
Germany

2×7×31 Posts
Default

Hi Ilya,

you should consider implementing SIQS. It is a big factor faster than the basice quadratic sieve. (some paper reported factor 17, was it Contini?).

MPQS is a good intermediate step because it is already much faster than basic QS but not as complex as SIQS.

Knuth-Schroeppel multipliers give another gain of estimated 50% or so on average.

Then there is a lot of fine-tuning that can be done...

Cheers
Till

Last fiddled with by Till on 2016-02-22 at 11:32 Reason: typo
Till is offline   Reply With Quote