Peter,

David Broadhurst provided similar arguments to show that there is
likely to infinite number of such primes and that their count grows
proportionally to log(log(x)):
http://groups.yahoo.com/group/primeform/message/6938

But he still has some doubts (and I share them) if probabilistic
assumptions he made hold for this problem:
http://groups.yahoo.com/group/primeform/message/6965

But from the point of finding a new prime, these results are
interesting but useless. At least I do not see how they can help in
the search.

Max

On 2/18/06, [EMAIL PROTECTED]
<[EMAIL PROTECTED]> wrote:
> Here are the values of    p, factorization of  n = q^2 - q + 1
> for small p == 1 (mod 12).  They were found through Maple's
> ifactor(n, easy) function.
> The value of n is a probable prime when n = 25 or 901.
>
>      If we pick a large p == 1 (mod 12), then n will be about 4^p.
> We estimate n will be prime with probability
> 1/log(4^p) = (1/log(4))/p.  p will be prime with probability about
> 1/log(p),  Summing
>
>           1/log(4)       1
>           --------   -------
>               p       log(p)
>
> over p = 13, 25, 37, 49, ..., the series diverges,
> so we expect infinitely many cases where both are prime.
> But they can be hard to find, just as 1093 and 3511 are
> the only known primes p for which 2^p == 2 (mod p^2).
>
>
>                                      1, 1
>
>                              13,  (193)  (347587)
>
>                             25,  (1125899806179331)
>
>                   37,  (367)  (379)  (3234619)  (41984721013)
>
>                 49,  (254110928791681607803)  (48337)  (25801)
>
>                                61,  (937) _c34_1
>
>          73,  (43)  (21002580922364503755525539978253220069)  (98773)
>
>                           85,  (223) _c44_1  (222553)
>
>                           97, _c47_1  (865209213469)
>
>
>
>           Peter Montgomery
>
>
> >
> > On Fri, 17 Feb 2006, Max wrote:
> >
> >> Hello!
> >> This problem is somewhat related to Mersenne numbers.
> >> Let p be a prime number and q = 2^p - 1 (so q is a Mersenne number but
> not necessary prime). Can the number q^2 - q + 1 be prime?
> >> Substituting the value for q, we can formulate the same problem as
> follows:
> >> Can the number P(p) = 2^(2*p) - 3*2^p + 3 be prime for prime p? What is
> known:
> >> If p is not limited to primes, then there are many primes P(p). There
> are also many primes P(p) if p is limited to powers of primes. But no
> prime P(p) found for primes p < 150000.
> >> Can anybody prove that P(p) can never be prime for prime p or
> >> find prime P(p) for some prime p (in which case p must be > 150000) ?
> Thanks,
> >> Max
> >> P.S. The origin of this problem (in russian):
> >> http://www.nsu.ru/phorum/read.php?f=29&i=5095&t=5095
> >> _______________________________________________
> >> Prime mailing list
> >> [email protected]
> >> http://hogranch.com/mailman/listinfo/prime
> >
> >
> > Hello Max,
> >
> >   I think your problem could be solved with the technique of
> > "covering congruences". It is easy to prove that for the case
> > p=6k-1 the number P(p) is always divisible by 7. The remaining
> > numbers p=6k+1 can be split into two classes, p=12k+1 and
> > p=12k+7. For the case p=12k+7 it is easy to prove that P(p) is
> > aiways divisible by 13.---I have unfortunately not been able to
> > do more research on the remaining case p=12k+1, because I have
> > been given a new computer, which has not yet had all systems
> > installed. But if you could send me factorizations of the
> > numbers P(p) for p=13, 25, 37, 49, and 61, say, I'll give it a
> > try.
> >
> > Sincerely yours,
> >
> > Hans Riesel
> >
> >
> > _______________________________________________
> > Prime mailing list
> > [email protected]
> > http://hogranch.com/mailman/listinfo/prime
> >
>
>
>
>
>
> _______________________________________________
> Prime mailing list
> [email protected]
> http://hogranch.com/mailman/listinfo/prime
>
_______________________________________________
Prime mailing list
[email protected]
http://hogranch.com/mailman/listinfo/prime

Reply via email to