fnsieve - Page 5

rogue · 2018-01-07, 14:01

Quote:

Originally Posted by Citrix

Is this the fastest program available to trial factor a single k,b,n,+-1 pair? If not, which program is the fastest.

Please be more specific.

Citrix · 2018-01-07, 21:26

I have a set of numbers k*b^n+-1 where k,b,n have no particular pattern. PFGW is extremely slow to trial factor them before PRP testing.

Srsieve is much faster... but I was wondering if there was anything else that could be faster?

rogue · 2018-01-07, 23:43

Quote:

Originally Posted by Citrix

I have a set of numbers k*b^n+-1 where k,b,n have no particular pattern. PFGW is extremely slow to trial factor them before PRP testing.

Srsieve is much faster... but I was wondering if there was anything else that could be faster?

If you have fixed k, b, and c then sr2sieve. If fixed b, n, and c, then fbncsieve. If more than one vary, then you are somewhat out of luck unless there is a pattern.

Citrix · 2018-01-08, 05:49

Quote:

Originally Posted by rogue

If more than one vary, then you are somewhat out of luck unless there is a pattern.

That is what I am trying to do. Trial factor one candidate at a time. What is the best software for this?

ppsieve is the fastest I have found. Is there anything faster?

axn · 2018-01-08, 06:33

Quote:

Originally Posted by Citrix

That is what I am trying to do. Trial factor one candidate at a time. What is the best software for this?

ppsieve is the fastest I have found. Is there anything faster?

You can use NewPGen. Use the "k*b^n+1" form, with kmin = kmax = your k. But doing that one candidate at a time is inefficient.

Hence rogue was suggesting that you combine all your candidates and sieve them in one go -- provide it is one of the forms that he suggested. If you have a range of k values and n values both, it might be doable with ppsieve.

To reiterate, one at a time, bad, all at once, good.

rogue · 2018-01-08, 13:57

Quote:

Originally Posted by Citrix

That is what I am trying to do. Trial factor one candidate at a time. What is the best software for this?

ppsieve is the fastest I have found. Is there anything faster?

That only works for base 2 and is used when you have a range of n and a range of k.

pepi37 · 2018-01-08, 15:07

Can you compile it with all dll-s that program need? It is little problem to start program few times until you find and download dll files program need to start
Thanks

rogue · 2018-01-08, 16:45

Quote:

Originally Posted by pepi37

Can you compile it with all dll-s that program need? It is little problem to start program few times until you find and download dll files program need to start

Possibly. Can you install the mingw64 runtime? You can probably do that before I have time to build an exe with static libraries.

ATH · 2018-01-08, 18:03

You should test the speed. srsieve was faster than newpgen for me when factoring 1*2^n-3 but that was for very few n values, so if it is a large range newpgen might be faster.

ATH · 2018-07-23, 23:14

When I use fnsieve to sieve a c-range of 10^9 it takes like 2min before it says "Beginning sieve at p=3." and even longer if I use a range of 4*10^9. It is on a slow laptop but still takes sometimes on my newest computer as well.

I'm just wondering what it is doing? Is it actually trial factoring with p=3 during that time even if it says beginning sieve afterwards?

axn · 2018-07-24, 02:50

Quote:

Originally Posted by ATH

When I use fnsieve to sieve a c-range of 10^9 it takes like 2min before it says "Beginning sieve at p=3." and even longer if I use a range of 4*10^9. It is on a slow laptop but still takes sometimes on my newest computer as well.

I'm just wondering what it is doing? Is it actually trial factoring with p=3 during that time even if it says beginning sieve afterwards?

Looking for algebraic factors?!

2018-07-23, 23:14	#54
ATH Einyen Dec 2003 Denmark C56₁₆ Posts	When I use fnsieve to sieve a c-range of 10^9 it takes like 2min before it says "Beginning sieve at p=3." and even longer if I use a range of 410^9. It is on a slow laptop but still takes sometimes on my newest computer as well. I'm just wondering what it is doing? Is it actually trial factoring with p=3 during that time even if it says beginning sieve afterwards? Last fiddled with by ATH on 2018-07-23 at 23:14*

2018-01-07, 21:26	#46
Citrix Jun 2003 2·7·113 Posts	I have a set of numbers k*b^n+-1 where k,b,n have no particular pattern. PFGW is extremely slow to trial factor them before PRP testing. Srsieve is much faster... but I was wondering if there was anything else that could be faster?

2018-01-08, 15:07	#51
pepi37 Dec 2011 After milion nines:) 2653₈ Posts	Can you compile it with all dll-s that program need? It is little problem to start program few times until you find and download dll files program need to start Thanks

2018-01-08, 18:03	#53
ATH Einyen Dec 2003 Denmark 2×1,579 Posts	You should test the speed. srsieve was faster than newpgen for me when factoring 1*2^n-3 but that was for very few n values, so if it is a large range newpgen might be faster.