Are factors vitrified?

houding · 2014-07-20, 19:08

I think I have read about this somewhere within another post, but I'm not sure.

If a factor for exponent is found (either via TF, P-1 or ECM) and submitted, is this factor also verified? Just to make sure that this was not submitted by a chancer randomly pushing buttons, and that a potential MP might now never be tested?

kracker · 2014-07-20, 19:50

Quote:

Originally Posted by houding

I think I have read about this somewhere within another post, but I'm not sure.

If a factor for exponent is found (either via TF, P-1 or ECM) and submitted, is this factor also verified? Just to make sure that this was not submitted by a chancer randomly pushing buttons, and that a potential MP might now never be tested?

The answer: Try submitting a fake factor.

Mini-Geek · 2014-07-20, 19:56

Yes, it is verified. Verifying a factor is very easy, so PrimeNet does it for all reported factors.

There are still some ways that a Mersenne prime could be missed by GIMPS indefinitely:

People leaving fake LL results to get credit they didn't earn
Bugs in Prime95, etc.
Bugs in PrimeNet

The end problem in the most likely cases would be that GIMPS records incorrect LL residues as double-checked. A future triple-check on independent systems would catch this.

Fortunately, all of these are fairly low-risk in my opinion, since there are some safeguards and checks to prevent each.

TheMawn · 2014-07-21, 01:22

Kracker has it right. It's actually worth trying out. Nothing bad happens; don't worry.

Even though the numbers are very big, there's no reason the math can't be done on them; it just takes longer. Dividing a 300-million digit number by some 40-digit number is just a bit of a longer process than dividing 15 by 3.

Primenet checks factors when they are submitted by manually checking them, i.e. dividing MXXX,XXX,XXX by the alleged factor, which again doesn't take long. The trial factoring software does this millions of times per second when you're looking.

Batalov · 2014-07-21, 01:54

Just yesterday morning I tested if primes are 'vilified' at primes.utm.edu
(by accident, really... lost a leading digit from a number's representation, resulting in six composites in a row: it was a Cunningham chain).

The result was predictable: not only the numbers were quickly discarded, but I also received a yellow card and was blocked from further submissions for a predefined timeout (36 hours). Harsh, but fair!

I don't think primenet will ban you by IP... maybe only after many willfully wrong results in a short period of time. One or two? probably not.

houding · 2014-07-21, 04:49

After many willfully wrong results you will be "vilified"!

Did not notice my typo until now.

Batalov · 2014-07-21, 05:18

(Some thread titles are occasionally tampered with, for a humorous effect.
Most likely, you did not make that typo.)

TheMawn · 2014-07-21, 20:21

Notice that factors are now being vitrified. Better than being experimented on in vitro, I would imagine.

2014-07-20, 19:08	#1
houding "Adolf" Nov 2013 South Africa 111101₂ Posts	Are factors vitrified? I think I have read about this somewhere within another post, but I'm not sure. If a factor for exponent is found (either via TF, P-1 or ECM) and submitted, is this factor also verified? Just to make sure that this was not submitted by a chancer randomly pushing buttons, and that a potential MP might now never be tested?

2014-07-20, 19:56	#3
Mini-Geek Account Deleted "Tim Sorbera" Aug 2006 San Antonio, TX USA 4267₁₀ Posts	Yes, it is verified. Verifying a factor is very easy, so PrimeNet does it for all reported factors. There are still some ways that a Mersenne prime could be missed by GIMPS indefinitely: People leaving fake LL results to get credit they didn't earn Bugs in Prime95, etc. Bugs in PrimeNet The end problem in the most likely cases would be that GIMPS records incorrect LL residues as double-checked. A future triple-check on independent systems would catch this. Fortunately, all of these are fairly low-risk in my opinion, since there are some safeguards and checks to prevent each. Last fiddled with by Mini-Geek on 2014-07-20 at 19:58

2014-07-21, 04:49	#6
houding "Adolf" Nov 2013 South Africa 61 Posts	After many willfully wrong results you will be "vilified"! Did not notice my typo until now. Last fiddled with by houding on 2014-07-21 at 04:50

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Known factors	ATH	PrimeNet	2	2014-09-04 19:45
Big factors	Jeff Gilchrist	Wagstaff PRP Search	10	2013-04-07 11:07
Missing factors at the 'Known Factors' page	MatWur-S530113	PrimeNet	11	2009-01-21 19:08
100 P-1 factors...	dave_0273	Marin's Mersenne-aries	5	2004-12-24 12:54
The factors of 11,199-	Jeff Gilchrist	NFSNET Discussion	2	2004-09-27 23:40

2014-07-21, 01:22	#4
TheMawn May 2013 East. Always East. 11×157 Posts	Kracker has it right. It's actually worth trying out. Nothing bad happens; don't worry. Even though the numbers are very big, there's no reason the math can't be done on them; it just takes longer. Dividing a 300-million digit number by some 40-digit number is just a bit of a longer process than dividing 15 by 3. Primenet checks factors when they are submitted by manually checking them, i.e. dividing MXXX,XXX,XXX by the alleged factor, which again doesn't take long. The trial factoring software does this millions of times per second when you're looking.

2014-07-21, 01:54	#5
Batalov "Serge" Mar 2008 Phi(4,2^7658614+1)/2 9,497 Posts	Just yesterday morning I tested if primes are 'vilified' at primes.utm.edu (by accident, really... lost a leading digit from a number's representation, resulting in six composites in a row: it was a Cunningham chain). The result was predictable: not only the numbers were quickly discarded, but I also received a yellow card and was blocked from further submissions for a predefined timeout (36 hours). Harsh, but fair! I don't think primenet will ban you by IP... maybe only after many willfully wrong results in a short period of time. One or two? probably not.

2014-07-21, 05:18	#7
Batalov "Serge" Mar 2008 Phi(4,2^7658614+1)/2 22431₈ Posts	(Some thread titles are occasionally tampered with, for a humorous effect. Most likely, you did not make that typo.)

2014-07-21, 20:21	#8
TheMawn May 2013 East. Always East. 11·157 Posts	Notice that factors are now being vitrified. Better than being experimented on in vitro, I would imagine.