mersenneforum.org F12 factor found?
 Register FAQ Search Today's Posts Mark Forums Read

 2016-11-01, 14:09 #1 johnadam74   "John A Burroughs" Jan 2016 Warwick, RI 52 Posts F12 factor found? Hi, I was running ECM curves on F12 and my prime log states: F12 has a factor: 45477879701734570611058964078361695337745924097 but when I enter that in my factor into my worktodo.txt for future curves, I get the message "45477879701734570611058964078361695337745924097 does not divide F12" Did I find a factor? Thanks, John
 2016-11-01, 14:12 #2 fivemack (loop (#_fork))     Feb 2006 Cambridge, England 193616 Posts You found a factor, but it is a composite factor Code: 114689 * 26017793 * 63766529 * 190274191361 * 1256132134125569 The code is working with F12-with-known-factors-divided-out, so these known factors no longer divide it.
 2016-11-01, 14:23 #3 johnadam74   "John A Burroughs" Jan 2016 Warwick, RI 52 Posts My worktodo.txt contains: ECM2=N/A,1,2,4096,1,999999999,99999999900,10,"114689,26017793,63766529,190274191361,1256132134125569,568630647535356955169033410940867804839360742060818433" I blocked out my reservation code and replaced it with N/A. I will still get composite factors? Thanks, John
2016-11-01, 16:41   #4
GP2

Sep 2003

3·863 Posts

Quote:
 Originally Posted by fivemack The code is working with F12-with-known-factors-divided-out, so these known factors no longer divide it.
If I'm not mistaken, mprime works only on numbers of a specific form, namely: k * b^n ± c

So it doesn't actually divide out the known factors, rather it merely ignores them.

Last fiddled with by GP2 on 2016-11-01 at 16:41

2016-11-01, 16:44   #5
GP2

Sep 2003

3×863 Posts

Quote:
 Originally Posted by johnadam74 I will still get composite factors?
If your ECM2 line contains a double-quoted string with comma-separated known factors at the end, then it should never report any bogus composite factors that are composed of the already-known prime factors.

It's not clear what happened, but maybe at some point you inadvertently added an ECM2 line to your worktodo.txt that was missing the double-quoted string at the end.

2016-11-01, 23:38   #6
Batalov

"Serge"
Mar 2008
Phi(4,2^7658614+1)/2

32×1,117 Posts

Quote:
 Originally Posted by GP2 If I'm not mistaken, mprime works only on numbers of a specific form, namely: k * b^n ± c So it doesn't actually divide out the known factors, rather it merely ignores them.
The code does not "ignore" known factors (a,b,c,a*b,a*c,b*c and a*b*c) "if/when it finds them" - it cannot find them: this is because the code searches for factors of F12/(a*b*c), and is running faster by doing math on F12 and then doing gcd with F12/(a*b*c) at certain steps.

When the code parses an assignment, it goes word by word.
It prepares the FFT structures for F12 and a standalone long number m=F12; then it divides m by every word and skips those that don't divide (and at this point when it reports "X doesn't divide F12" it means X doesn't divide "X doesn't divide F12/product(all factors before X)".

So if you put one factor in the d-quotes twice, the code will complain about the second of them but will work just fine as intended. If you put "(a)", "(product of b*c)" and also "(product of a*c)", one after another in d-quotes, you will get an error "(a*c) doesn't divide F12" but the code will still work as if you provided a, b and c. You can also provide a product of all factors in d-quotes - that will be also fine.

 2016-11-01, 23:57 #7 GP2     Sep 2003 3×863 Posts Ah, OK. On a different note: we've twice had Mersenne primes discovered without anyone noticing for months, because the automatic notification failed. I wonder if there is any notification mechanism if new Fermat factors are found, and are we sure it works? Is it possible that some Fermat factor was found but has been sitting forgotten in a log file somewhere...
 2016-11-02, 01:00 #8 Prime95 P90 years forever!     Aug 2002 Yeehaw, FL 2×4,079 Posts There is no email notifications for Fermat factors. The good news is they aren't hidden. They should get added to the database and displayed for all to see.
2016-11-02, 02:04   #9
GP2

Sep 2003

A1D16 Posts

Quote:
 Originally Posted by Prime95 The good news is they aren't hidden. They should get added to the database and displayed for all to see.
Hmmm, but where can you see them? They don't show up in the Factors Found or Exponent Status reports. And the ECM Progress report only displays the number of curves done.

And assuming you can actually see them somewhere, is anyone viewing that periodically and checking for new ones?

Last fiddled with by GP2 on 2016-11-02 at 02:06

 2016-11-02, 04:37 #10 Prime95 P90 years forever!     Aug 2002 Yeehaw, FL 2×4,079 Posts I guess the only place to see them is to get an ECM assignment. I checked and there are no new ones. Vang's F12 factor was missing.
2016-11-02, 13:54   #11
Xyzzy

Aug 2002

216A16 Posts

Quote:
 Originally Posted by Prime95 Vang's F12 factor was missing.
We would give our left testicle to find another!

 Similar Threads Thread Thread Starter Forum Replies Last Post tha Data 65 2020-08-05 21:11 NBtarheel_33 GPU Computing 11 2012-04-07 21:12 tha Factoring 4 2007-06-18 19:56 jocelynl Software 6 2004-08-07 01:31 Reboot It Data 3 2003-12-03 14:39

All times are UTC. The time now is 14:25.

Tue Jan 31 14:25:45 UTC 2023 up 166 days, 11:54, 0 users, load averages: 0.93, 1.18, 1.21