Re: Surrogate factoring works very well

"rossum" <rossum48@xxxxxxxxxxxx> wrote in message
news:71msv29u87u7nt3obsghra6t3kv2vqadkk@xxxxxxxxxx
The code does compile and run once it is all downloaded. It is very
under-commented and at first glance appears over-complex for what it
does. James has included a certain amount of monitoring/timing code
which can presumably be removed for a production version, though he
has not made it easy to do so.

The big questions then:
Does it actually factor numbers?
What is the asymptotic speed?
What is the realistic speed for 2Kbit numbers?

The first two are the crux of the claims, the third regards it's usability
(see Primes is in P).

The complaint regarding use of isProbablePrime is moot in my opinion, it
works well enough, and since Primes is in P using a more optimized algorithm
for the sizes is acceptable.
Joe


.