---- Chris Caldwell <

caldwell@...> wrote:

> > I must test numbers in the form k*2^n+1.

> > Is Pfgw faster than prp.exe?

>

> Others will correct me if I am wrong, but I think they use the same

> arithmetic engine now. Of course prp.exe does not prove primality. LLR

> might be the fastest depending on k, but the key thing to do is to

> prescreen well--that is where you can save a great deal of time.

For base 2, I recommend LLR, since it will automatically do a Proth test for k*2^n+1 numbers. You could also use PFGW, but you would need to force a primality test by using the -tm option. The problem with PFGW is that it will not produce a residue for primality tests. LLR produces a residue for all tests. LLR should be just as fast as PFGW for this base.

I suspect that George is keeping PRP up to date with his changes to gwnum, but I think he has slowly been moving those functions to Prime95. As Chris said, PRP cannot do a primality test.

--Mark