-   Miscellaneous Math (
-   -   Fun Sequence (

Sam Kennedy 2013-02-07 10:36

Fun Sequence
I've started searching this sequence for prime numbers:
[TEX]((n+1)\mathrm{!}\ - \ n\mathrm{!})k + 1[/TEX]

I like this sequence because N-1 is easy to factor so proving primality is straight forward.

Are there any faster primality tests than rabin-miller for numbers like this?

Has anyone searched a sequence similar to this?

axn 2013-02-07 11:16

Your expression simplifies to k.n.n!+1. Why not ditch the extra n and search for k.n!+1? Anyways...

Are you doing any sieving?

I am not aware of anything faster than regular PRP tests (PFGW can do that). PFGW can also prove the primality with a -tm argument.

Sam Kennedy 2013-02-07 11:35

I'm not sure how sieving applies to this, I know it's used in the other prime searching projects but I've never understood how it's actually used.

The only thing I can think of is generating a bunch of the above factorials, trial dividing by small primes below a certain limit, then performing the primality tests on whichever don't divide.

axn 2013-02-07 11:42

Sieving is merely a more efficient mechanism for doing trial factoring across a bunch of candidates, exploiting redundancies. And yes, sieving is applicable for your form, though I am not aware of any existing software that can do this (maybe MultiSieve?). But there are people in this forum who are capable of writing a sieve for this.

Your search space is controlled by 2 parameters, k and n. What are the limits on these? Also, how do you plan to investigate them? Fix a k and search across different n's? Fix an n and search across different k's? Some other way? Depending upon your strategy, the sieving approach also would change.

You could cut down your total search time upto 50% with a good siever.

Sam Kennedy 2013-02-07 11:53

Depending on the complexity of the maths involved, I might be able to write my own sieve.

The way I'm searching is fixing k and incrementing n.

The limit for k is the same as the limit for an unsigned long, n is limited by the amount of memory needed to store the entire term.

Are there any good references which could give me some ideas of coding a sieve?

All times are UTC. The time now is 05:56.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2022, Jelsoft Enterprises Ltd.