mersenneforum.org  

Go Back   mersenneforum.org > Great Internet Mersenne Prime Search > Math

Reply
 
Thread Tools
Old 2016-01-23, 16:09   #12
JeppeSN
 
JeppeSN's Avatar
 
"Jeppe"
Jan 2016
Denmark

2·83 Posts
Question

Quote:
Originally Posted by JeppeSN View Post
Can we extend this?

p: factor(s) of 2p-1*(2p-1) + 1
57885161: 7
74207281: ?

/JeppeSN
Update:

42643801: 3593, 7089208037
43112609: 7, 211, 70121, 71647, 1846524311 (was in 2008 post by ATH)
57885161: 7, 22127627
74207281: ?

My trial factoring with the latest perfect number, incremented by one, has not yielded any divisors under $75\cdot 10^9$.

/JeppeSN
JeppeSN is offline   Reply With Quote
Old 2016-01-23, 17:00   #13
ATH
Einyen
 
ATH's Avatar
 
Dec 2003
Denmark

57778 Posts
Default

Yeah for 57885161 there is also a factor: 22127627.

For p=74207281, I trial factored 2p-1*(2p-1) + 1 up to 4.46*1012 with no factor.
ATH is online now   Reply With Quote
Old 2016-01-26, 12:43   #14
JeppeSN
 
JeppeSN's Avatar
 
"Jeppe"
Jan 2016
Denmark

2×83 Posts
Default

Quote:
Originally Posted by ATH View Post
For p=74207281, I trial factored 2p-1*(2p-1) + 1 up to 4.46*1012 with no factor.
Do you know if anyone is going further with trial factoring this beast, $2^{74207280}(2^{74207281}-1)+1$, or even do a primality test of it? It may turn out to be very hard to find a prime factor for this one? /JeppeSN
JeppeSN is offline   Reply With Quote
Old 2016-01-26, 17:20   #15
ATH
Einyen
 
ATH's Avatar
 
Dec 2003
Denmark

37×83 Posts
Default

Yeah the chance of finding a factor is very low, and doing a primality test is impossible. At 44.7 million digits only an LL test would be fast enough but it will not work on a number of this form. There is no software I can think of that can even do a PRP test.

Anyway these numbers are not really that important I think, it was just a question a user asked almost 8 years ago which turned into this tiny trial factoring effort. I have never seen any interest in "perfect numbers + 1" anywhere else.
ATH is online now   Reply With Quote
Old 2016-01-26, 20:34   #16
JeppeSN
 
JeppeSN's Avatar
 
"Jeppe"
Jan 2016
Denmark

2×83 Posts
Default

As R. Gerbicz said in his last post to this thread, an n-1 test should be possible. /JeppeSN
JeppeSN is offline   Reply With Quote
Old 2016-01-27, 06:38   #17
ATH
Einyen
 
ATH's Avatar
 
Dec 2003
Denmark

37·83 Posts
Default

Yes true, but that was for 79502 digits. I do not think programs like PFGW, which can do the N-1 test, can handle 44 million digits or anywhere near this size.

A 44 million digit LL test would take 10-14 days on the very fastest computers / graphic cards with Prime95 / CudaLucas, and an N-1 test would take longer than that. I do not think PFGW can save the test during it and restart, but I'm not certain about that.
ATH is online now   Reply With Quote
Old 2016-01-27, 08:43   #18
Batalov
 
Batalov's Avatar
 
"Serge"
Mar 2008
Phi(4,2^7658614+1)/2

222548 Posts
Default

Quote:
Originally Posted by ATH View Post
Yes true, but that was for 79502 digits. I do not think programs like PFGW, which can do the N-1 test, can handle 44 million digits or anywhere near this size.
PFGW can. The limit is the same (gwnum library is the same) - FFT max size of 32M which is more than enough, but...
Quote:
Originally Posted by ATH View Post
A 44 million digit PRP LL test would take [COLOR=DarkRed]10-14[/COLOR] days on the very fastest computers / graphic cards with Prime95 / CudaLucas, and an N-1 test would take longer than that. I do not think PFGW can save the test during it and restart, but I'm not certain about that.
Because of the form of the number, LL test has no relevance and fast special FFT is inapplicable, so you can safely triple the estimate (and much more than triple if you only use single thread with a vanilla binary). PFGW is single-thread, but you can try to rig Prime95's multithreaded code on zero-padded FFT and write a custom mod step (2^N-2^M+1 is much better than a general number).
Batalov is offline   Reply With Quote
Old 2016-01-27, 12:36   #19
henryzz
Just call me Henry
 
henryzz's Avatar
 
"David"
Sep 2007
Cambridge (GMT/BST)

3×1,951 Posts
Default

Quote:
Originally Posted by Batalov View Post
PFGW can. The limit is the same (gwnum library is the same) - FFT max size of 32M which is more than enough, but...

Because of the form of the number, LL test has no relevance and fast special FFT is inapplicable, so you can safely triple the estimate (and much more than triple if you only use single thread with a vanilla binary). PFGW is single-thread, but you can try to rig Prime95's multithreaded code on zero-padded FFT and write a custom mod step (2^N-2^M+1 is much better than a general number).
Based upon my fiddling with PFGW the FFT length would be probably be 4x as large as the mersenne prime. It would also be general reduction as well.
Possible if someone wants to spend several months on it I think.

Last fiddled with by henryzz on 2016-01-27 at 12:46
henryzz is offline   Reply With Quote
Old 2016-01-27, 14:04   #20
JeppeSN
 
JeppeSN's Avatar
 
"Jeppe"
Jan 2016
Denmark

2×83 Posts
Exclamation

On my Windows system, it crashes.

I try with:

Code:
 
.\pfgw64.exe -q"2^74207280*(2^74207281-1)+1" -t -f0 -h"jeppe_M74207281.txt"
where -t should enforce a deterministic N-1 test, -f0 should skip factoring, and the jeppe_M74207281.txt is a local help file containing just one line with 2^74207281-1 on it. The help file is for making PFGW quickly realize it has the full factorization of N-1 (our even perfect number). Apparently, PFGW does not read news sites on the web telling about the recent discovery of M74207281, so I will have to help it.

pfgw32.exe with the same parameters crashes as well, on my machine here.

Using the binaries from .7z file here: http://sourceforge.net/projects/openpfgw/

Can anyone else get to a point where the N-1 test actually progresses?

/JeppeSN
JeppeSN is offline   Reply With Quote
Old 2016-01-27, 15:40   #21
paulunderwood
 
paulunderwood's Avatar
 
Sep 2002
Database er0rr

E2116 Posts
Default

It is much quicker to run a default 3-PRP test, rather than a fully blown N-1 test. If that goes fine, I am sure the authors will fix up PFGW for you

Here, I would run on linux: ./pfgw64 -q"2^74207280*(2^74207281-1)+1" -f0

The chance it will be prime is next to zero, but you will know!
paulunderwood is offline   Reply With Quote
Old 2016-01-27, 16:52   #22
JeppeSN
 
JeppeSN's Avatar
 
"Jeppe"
Jan 2016
Denmark

2·83 Posts
Default

Maybe it is smarter to sieve even further before attempting a PRP or primality test. After all, what we want to add to ATH's table above is the smallest prime factor, and it is too optimistic to hope the primality test will result in N being its own smallest prime factor. /JeppeSN
JeppeSN is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Is this new formula for Perfect Numbers useful? mahbel Miscellaneous Math 20 2017-03-01 22:41
How to generate base10 representation of Mersenne-prime perfect numbers? James Heinrich Miscellaneous Math 10 2012-03-08 07:20
Odd Perfect Numbers davar55 Miscellaneous Math 16 2011-01-29 01:53
Perfect Numbers MajUSAFRet Math 3 2003-12-13 03:55
Odd Perfect Numbers Zeta-Flux Math 1 2003-05-28 19:41

All times are UTC. The time now is 13:51.

Mon Apr 12 13:51:15 UTC 2021 up 4 days, 8:32, 1 user, load averages: 2.36, 1.73, 1.73

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

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.