Hi,

At 12:46 AM 11/27/2001 +0100, george de fockert wrote:
>One of my machines does trial factoring.
>Now in the M17900000 range, factoring to 2^66.
>This takes a very long time, is it better to let it double check ?

It is a matter of personal preference.  You can figure out how long
a double-check would take at http://www.mersenne.org/bench.htm

Another option for slow machines is ECM factoring.  See
http://www.mersenne.org/ecm.htm for details (this is not an
automated project)

>Or is it time for a more efficient trial factor implementation (if possible
>?).

If someone wants to volunteer to improve factoring performance that
would be great.  The P4 could use an SSE2 implementation.  CPU-specific
factoring code over 64 bits could be written.

While factoring could probably be improved tens of percent, it has not
been a high priority item for me.  A 10% boost in factoring speed lets
you factor (on average) 10% deeper, which finds about 1 more factor in
every 650 exponents.  That isn't a great throughput improvement for
the GIMPS project.

-- George

_________________________________________________________________________
Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm
Mersenne Prime FAQ      -- http://www.tasam.com/~lrwiman/FAQ-mers

Reply via email to