mersenneforum.org x!^n+/-1 or x#^n+/-1?
 User Name Remember Me? Password
 Register FAQ Search Today's Posts Mark Forums Read

 2020-11-13, 16:01 #1 rogue     "Mark" Apr 2003 Between here and the 181A16 Posts x!^n+/-1 or x#^n+/-1? Is there any interest from anyone here in searching for primes in the form x!^n+/-1 or x#^n+/-1? I would refer to them as "power factorial" or "power primorial". As far as I know these forms have never been searched or n>1. It would be fairly easy to modify fsieve and psieve to support them.
2020-11-13, 16:39   #2
Dr Sardonicus

Feb 2017
Nowhere

2·3·5·139 Posts

Quote:
 Originally Posted by rogue Is there any interest from anyone here in searching for primes in the form x!^n+/-1 or x#^n+/-1? I would refer to them as "power factorial" or "power primorial". As far as I know these forms have never been searched or n>1. It would be fairly easy to modify fsieve and psieve to support them.
Presumably you mean

$\Phi_{n}(x!)\text{, }\Phi_{2n}(x!)\text{,}$

or the same polynomials with the argument x! replaced with x#?

2020-11-13, 17:55   #3
rogue

"Mark"
Apr 2003
Between here and the

2·5·617 Posts

Quote:
 Originally Posted by Dr Sardonicus Presumably you mean $\Phi_{n}(x!)\text{, }\Phi_{2n}(x!)\text{,}$ or the same polynomials with the argument x! replaced with x#?
I'm not fully understanding your notation.

5! = 5*4*3*2*1 so 5!^3 = (5*4*3*2*1)^3
5# = 5*3*2 so 5#^3 = (5*3*2)^3

Here is a list of factorial primes: https://primes.utm.edu/top20/page.php?id=30
Here is a list of primorial primes: https://primes.utm.edu/top20/page.php?id=5.

2020-11-13, 18:02   #4
axn

Jun 2003

12EA16 Posts

Quote:
 Originally Posted by rogue I'm not fully understanding your notation.
x^n+/-1 has trivial algebraic factors. So he was suggesting to look at the nth cyclotomic polynomial instead - i.e. what is left after removing the algebraic factors.

2020-11-13, 18:44   #5
R. Gerbicz

"Robert Gerbicz"
Oct 2005
Hungary

1,429 Posts

Quote:
 Originally Posted by axn x^n+/-1 has trivial algebraic factors. So he was suggesting to look at the nth cyclotomic polynomial instead - i.e. what is left after removing the algebraic factors.
Exactly. There is a not hard pattern for these factors: if additionally you remove all lower exponents factors then for the remaining p prime factors what divides polcyclo(n,x) we have p%n=1. You really need to do this, because for example:
Code:
? factor(polcyclo(20,2))
%24 =
[ 5 1]

[41 1]
Ofcourse after the sieve with these special primes you need to do only a few gcd's on the remaining candidates. [or start the sieve removing these].

Last fiddled with by R. Gerbicz on 2020-11-13 at 18:46 Reason: grammar

 2020-11-13, 19:40 #6 rogue     "Mark" Apr 2003 Between here and the 2×5×617 Posts As for algebraic factors, I had forgotten about those.
 2020-11-13, 20:16 #7 Dylan14     "Dylan" Mar 2017 547 Posts Doing a quick search with pfgw with the -f1 flag, the following numbers are prime or 3-PRP for the form x!^2 +/- 1 up to x = 2000: Code: 0!^2+1 1!^2+1 2!^2+1 3!^2+1 4!^2+1 5!^2+1 9!^2+1 10!^2+1 11!^2+1 13!^2+1 24!^2+1 65!^2+1 76!^2+1 2!^2-1
2020-11-13, 20:59   #8
rogue

"Mark"
Apr 2003
Between here and the

181A16 Posts

Quote:
 Originally Posted by Dylan14 Doing a quick search with pfgw with the -f1 flag, the following numbers are prime or 3-PRP for the form x!^2 +/- 1 up to x = 2000: Code: 0!^2+1 1!^2+1 2!^2+1 3!^2+1 4!^2+1 5!^2+1 9!^2+1 10!^2+1 11!^2+1 13!^2+1 24!^2+1 65!^2+1 76!^2+1 2!^2-1
I wouldn't expect anything else on the -1 side as x^n-1 always factors as (x-1) as a factor. Note that on the +1 side any primes must also be GFNs which means that n will always be a power of 2 for primes.

So from those perspectives, these forms are not that interesting.

 2020-11-14, 11:51 #9 a1call     "Rashid Naimi" Oct 2015 Remote to Here/There 7×283 Posts Remove the +/-1 and add +/-k and divide by k where k | x!^n Things will get interesting Ali candidates will be proveable via N-/+1 method since they will be of form where N has prime factors equal to or less than x. Additionally if both (x!^n+/-k)/k are prime then they are twin primes. Here is non twin, n=1 example: https://www.mersenneforum.org/showpo...8&postcount=19 ETA Unlike non-generalized factorial primes there are plenty of twin-primes in the generated form which are highly unreserved. Last fiddled with by a1call on 2020-11-14 at 12:28
2020-11-14, 13:58   #10
Dr Sardonicus

Feb 2017
Nowhere

2·3·5·139 Posts

Quote:
 Originally Posted by R. Gerbicz Exactly. There is a not hard pattern for these factors: if additionally you remove all lower exponents factors then for the remaining p prime factors what divides polcyclo(n,x) we have p%n=1. You really need to do this, because for example: Code: ? factor(polcyclo(20,2)) %24 = [ 5 1] [41 1] Ofcourse after the sieve with these special primes you need to do only a few gcd's on the remaining candidates. [or start the sieve removing these].
Right, the factor 5 in the above is sometimes called an "intrinsic" prime factor. It shows up because 5 divides polcyclo(4,2). An additional factor of 5 shows up in polcyclo(4*5, 2); another in polcyclo(4*5^2, 2), another in polcyclo(4*5^3, 2) and so on. This is described in detail in The Cyclotomic Polynomials.

There is at most one "intrinsic" prime factor. Let P be the evaluation of the cyclotomic polynomial, and n the exponent. If P has an intrinsic prime factor, it is gcd(P,n).