Generalized Fermat number

From Prime-Wiki
Revision as of 13:22, 6 March 2019 by Karbon (talk | contribs) (cat. changed)
Jump to: navigation, search

There are different kinds of generalized Fermat numbers.

John Cosgrave

John Cosgrave has studied the following numbers:

Numbers of the form: [math]F_{n,r} = \sum_{i=0}^{p-1} \ 2^{i p^{n}} \ = \ 2^{(p-1)p^n}+2^{(p-2)p^n}+...+2^{2p^n}+2^{p^n}+1 \ = \ (2^{p^{n+1}}-1)/(2^{p^n}-1)[/math] where p is the prime of apparition rank r (r(2)=1, r(3)=2, r(5)=3, ...) and n is greater or equal to 0.

[math]F_{0,r}[/math] generates the Mersenne numbers.
[math]F_{n,1}[/math] generates the Fermat numbers.
[math]F_{n,2}[/math] generates the Saouter numbers.

Cosgrave has proven the following properties:

  1. If number [math]\sum_{i=0}^{p-1}\ (2^i)^{m} \ [/math] is prime, then [math]m=p^n[/math].
  2. [math]F_{n,r}[/math] numbers are pairwise relatively prime within a rank and across ranks: [math]gcd(F_{n,i},F_{m,j}) =1 [/math] for all n, m, i and j.
  3. They satisfy a product property like Fermat numbers have. And every [math]F_{n,r}[/math] passes Fermat's test to base 2.

Saouter has proven that [math]F_{n,2}[/math] numbers can be proven prime by using the Pépin's test with k=5.


In 1985, Dubner for the first time built a list of large primes of the form: b2m+1, b ≥ 2 and m ≥ 1.

Björn & Riesel

In 1998, Björn & Riesel for the first time built a list of large primes of the form: a2m+b2m, b > a ≥ 2 and m ≥ 1.

External links


  • Anders Björn and Hans Riesel, Factors of generalized Fermat numbers, Math. Comp. 67 (1998), pp. 441-446