mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Software (https://www.mersenneforum.org/forumdisplay.php?f=10)
-   -   A siever for K (b, n, c fixed)? (https://www.mersenneforum.org/showthread.php?t=20347)

pepi37 2015-07-09 15:30

A siever for K (b, n, c fixed)?
 
Is anybody know where to find it?
I know that NewPgen can do fixed k sieve, but I am interesting is any newer ( faster) siever for fixed K?
Thanks

rogue 2015-07-09 16:56

[QUOTE=pepi37;405567]Is anybody know where to find it?
I know that NewPgen can do fixed k sieve, but I am interesting is any newer ( faster) siever for fixed K?
Thanks[/QUOTE]

Fixed or variable b? Fixed or variable n? Fixed or variable c?

Dubslow 2015-07-09 18:32

[QUOTE=rogue;405575]Fixed or variable b? Fixed or variable n? Fixed or variable c?[/QUOTE]

I believe he is referring to Mersenne numbers only, where instead of fixing p and trying a bunch of k (as is standard), fix k and try all p (i.e. n) from (presumably) 1 to 10e9 (PrimeNet's current table boundary). User Tjaoi is known to be sieving Mersennes this way.

I could of course be wrong in my interpretation, but hey.

pepi37 2015-07-09 19:01

[QUOTE=rogue;405575]Fixed or variable b? Fixed or variable n? Fixed or variable c?[/QUOTE]

Fixed b,fixed n, fixed c
So only k is not fixed
Oh I sow error in my first post: so sorry about that

rogue 2015-07-09 19:17

[QUOTE=pepi37;405583]Fixed b,fixed n, fixed c
So only k is not fixed
Oh I sow error in my first post: so sorry about that[/QUOTE]

I believe that you can use fnsievecl. It is fixed k, b, and n with variable c. You can convert k*b^n+c to b^n+c/k by computing the numerical inverse.

pepi37 2015-07-09 21:01

[QUOTE=rogue;405585]I believe that you can use fnsievecl. It is fixed k, b, and n with variable c. You can convert k*b^n+c to b^n+c/k by computing the numerical inverse.[/QUOTE]

So I need sieve program that can made
K=2-200000
b=10
n=1000000
c=1
fncievecl cannot do that

axn 2015-07-10 03:52

[QUOTE=pepi37;405567]Is anybody know where to find it?
I know that NewPgen can do fixed k sieve, but I am interesting is any newer ( faster) siever for fixed K?
Thanks[/QUOTE]

[QUOTE=pepi37;405589]So I need sieve program that can made
K=2-200000
b=10
n=1000000
c=1
fncievecl cannot do that[/QUOTE]

This is what NewPGen is built to do! NewPGen also does fixed-k, but fixed-n is where it really shines.

R. Gerbicz 2015-07-10 04:42

[QUOTE=rogue;405585]I believe that you can use fnsievecl. It is fixed k, b, and n with variable c. You can convert k*b^n+c to b^n+c/k by computing the numerical inverse.[/QUOTE]

That is not working. For different p values you will get very different cc=c/k mod p values, so here cc is not fixed.
(you can still compute these cc values for each prime, and use a program, but that's a painful approach of the problem).


All times are UTC. The time now is 00:41.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.