mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Lone Mersenne Hunters (https://www.mersenneforum.org/forumdisplay.php?f=12)
-   -   Found a factor? Post it here. Or forever fold your crease. (https://www.mersenneforum.org/showthread.php?t=13977)

Prime95 2023-02-28 03:06

[QUOTE=Andrew Usher;625793]That is not poaching! Doing work that Primenet expressly makes available...[/QUOTE]

Well, it sort of is [strike]poaching[/strike] playing outside of the rules because Primenet does not make the PRP-CF assignment available for one week.

Andrew Usher 2023-03-01 13:09

Well, you've confirmed that it is supposed to work that way, and I guess I can see why. But if we know there are users that routinely do PRP-CFs as soon as possible, it can't be a surprise when it happens - and if I cared, I'd do what I suggested - submit the factor and PRP result at the same time.

There are ways this could be prevented without compromising the database, but they would seem messy. Less so would be to change a future Prime95 per kruoli's suggestion above, but I would think to avoid complaints the option should be off by default and the maximum exponent configurable.

LordJulius 2023-03-07 05:37

[QUOTE=Prime95;625801]Well, it sort of is [strike]poaching[/strike] playing outside of the rules because Primenet does not make the PRP-CF assignment available for one week.[/QUOTE]

Then [URL="https://www.mersenne.ca/prp.php?show=3"]mersenne.ca[/URL] shouldn't make them available, yes?

kruoli 2023-03-07 11:37

It (mersenne.ca) is no platform to guide you in which work is available, because it cannot reserve anything via mersenne.org (the only site that has policy-making power in this case) and because it is not fully real-time. Since the usual rule is that you need an assignment ID (AID) to do a work unit, this is the only thing you should watch out for. If you are unable to get an AID for some work you want to do, it is either reserved/held back for someone else, unwanted, or in some cases bugged. If you think it is the latter, ask in the "Official Server Problems Thread" and it will be sorted out or deemed a feature, not a bug.

James Heinrich 2023-03-07 15:29

[QUOTE=Prime95;625801]Primenet does not make the PRP-CF assignment available for one week.[/QUOTE]mersenne.ca now should also hide PRP-CF for 7 days after the most recent factor is discovered. Let me know if it's not working as expected.

Rubiksmath 2023-03-08 20:23

Got one:
ECM found a factor in curve #30, stage #2
Sigma=7954931777905094, B1=11000000, B2=52102380330.
M99137 has a factor: 6459953914675791129553263498255728369 (ECM curve 30, B1=11000000, B2=52102380330)
31 to go to get under 1k on 0.0M.

Viliam Furik 2023-03-09 00:50

I was among the ones who requested this testing privilege period for the factor-finder, precisely because those who automatically got the freshly available PRP-CF work have finished before I, or other factor-finders, have even noticed a factor was found. It is true that very few of us actually use this privilege period, but it was established so that those who do want to use it have opportunity to do so.

I could agree with shortening the period to say 5 days, but I believe it has a good purpose and should be honored.

When people forget why the rules have been established, the rules lose their purpose. Just like in times of peace, when people forget about war, we forgot what led to the peace and take it for granted, and then we tend to lose it. We mustn't forget why the rules are the way they are and why we do what we do, whatever it may be.

Si vis pacem, para pactum.

James Heinrich 2023-03-09 01:09

[QUOTE=Viliam Furik;626394]It is true that very few of us actually use this privilege period, but it was established so that those who do want to use it have opportunity to do so.[/QUOTE]It's gradually going to become automatic:[QUOTE=Prime95;626284]30.11 build 1
-- PRP-CF done after new factor is found[/QUOTE]

Rubiksmath 2023-03-09 02:27

For smaller mersennes it looks like factordb performs a PRP-CF almost immediately. No PRP-CF result for the factor of M99137 has been reported to PrimeNet but on factordb when I checked a few hours ago it was aware of the factor and the fact that the cofactor was composite (I had not ran a PRP-CF yet). Not really a big deal since it is outside of GIMPS and it seems no record of the test exists (publicly anyway).

Viliam Furik 2023-03-09 18:42

[QUOTE=James Heinrich;626397]It's gradually going to become automatic:[/QUOTE]

Nice, I didn't notice that. I have not been paying much attention to things here lately. This automatic testing only increases the importance of the privilege period.

I have an idea for resolving these credit conflicts of discovery. Just a prototype thought, so it needs some thinking and discussion before it can be considered as a proposition. When someone is assigned an exponent to test which turns up prime, but someone else has poached that result and got there first - or similarly, someone found a factor and someone else beat them to finding the PRP result - the would-be discoverer and the poacher discoverer could have a shared credit for the discovery (with the unofficial "boo" from the community to the poacher). The person who was assigned the test would have most probably found the prime (or PRP) if they finished the test first (chances of errors and abandoned tests are relatively miniscule, and should be presumed not present similarly to the principle of "innocent until proven guilty"). That is undeniable fact. The poacher has discovered the prime first. Also undeniable fact. And when the are two undeniable facts which don't quite agree, we can either argue both are to be considered false (indicating some contradiction as in proof by contradiction) or both true. Since both follow from the base principles of assignments and discoveries, in my opinion they should both be considered valid. If the poacher didn't work through GIMPS, that would be a different story since then there would be no poaching and the poacher would in fact be a rightful discoverer. But for the cases outside of GIMPS this discussion would be meaningless, so they don't matter much.

What do you think, fellow Mersenne folk?

kruoli 2023-03-09 18:48

Maybe you are misunderstanding the change, it is not that new factors will get distributed more quickly, instead the [I]discoverer's[/I] machine will automagically run the PRP-CF run after a factor was found, so the grace period might not be necessary in this case. So it is much more likely that nobody can poach this, even if they try hard.


All times are UTC. The time now is 22:49.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2023, Jelsoft Enterprises Ltd.