mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Data (https://www.mersenneforum.org/forumdisplay.php?f=21)
-   -   (Preying for) World Record P-1! (https://www.mersenneforum.org/showthread.php?t=27406)

lisanderke 2021-12-17 08:44

[QUOTE=Prime95;595441]That's not good. The save file knows nothing about the known factors.

Do the save files prior to the 100% reporting find known factors when the B1 bound is increased by 999?[/QUOTE]
Exponent [M]5003[/M] had B1=5x10^9 and B2=4,4x10^15. Re-ran with B1=5x10^9 (ending in 999) for the following result:

[C][Thu Dec 16 00:35:50 2021]
UID: lisander/30.8b5, M5003 completed P-1, B1=5000000000, B2=4419083571242340, Wi4: 3CB31752
[Fri Dec 17 09:35:27 2021]
P-1 found a factor in stage #1, B1=5000000999.
UID: lisander/30.8b5, M5003 has a factor: 3281293532195321113569357982360497098983754047729275907049 (P-1, B1=5000000999)
[/C]
Question: why does my composite factor show up in the results history of this exponent? Shouldn't Primenet already have been aware of this composite factor?

LaurV 2021-12-17 09:12

Is this "new activity" (since I almost disappeared from the forum*) a reason for the "strange" lifting I see in P-1 tops?

I suspected for a while that we have some cheaters there, I mean people jumping hundreds of positions in a week or two, but when you look to their activity, they have very low number of trials, which does not justify the high scores at all, with no possible calculus, unless they did really, and I mean REALLY, deep, DEEP, P-1 (i.e. to very high limits B1 and B2, to get that huge amount of credits, but, BUT, in that case, they were ridiculously unlucky with only 2% (and under) factors found.

So, they either did low-bounds P-1 to find so less factors, which doesn't match with the high credit, or they did high-bounds P-1, which would fit with the credit given, but it doesn't fit with the number of factors.

Or they are cheating (it is easy to cheat the P-1 "no factor" results).

One situation that would still fit, with no cheating, is people doing a lot of P-1 in the very low ranges (which were swapped by TF and lots of ECM already). Is that what's happening here?


---------
*the reason of my "almost disappearance" from the forum I will post later, it is a funny story, not related to math, need time to type...

SethTro 2021-12-17 09:24

[QUOTE=kruoli;595459]petrw1 and masser gave me tons of files some time ago, but mostly from the <2K project. For these, I was planning something similar to your ideas. But I have not put much work into it yet.[/QUOTE]

[QUOTE=ATH;595461]I have quite a few P+1 save files, are those of interest as well?[/QUOTE]

I'll reach out to both of you next week as I make more progress and start adding upload functionality.

[B]Rough website with list of current files I have: [/B] [url]https://pminus1.cloudygo.com/[/url]

ATH 2021-12-17 09:57

[QUOTE=lisanderke;595466]Question: why does my composite factor show up in the results history of this exponent? Shouldn't Primenet already have been aware of this composite factor?[/QUOTE]

Primenet is aware but your local Prime95 is not, you should add the known factors in the worktodo line:

Pminus1=N/A,1,2,5003,-1,5000000000,4400000000000000,"10007,1050631,6087330966532976879,51270087288962073254717551943,68882595903330238838552613955694480987987527"

lisanderke 2021-12-17 09:58

[QUOTE=ATH;595470]Primenet is aware but your local Prime95 is not, you should add the known factors in the worktodo line:

Pminus1=N/A,1,2,5003,-1,5000000000,4400000000000000,"10007,1050631,6087330966532976879,51270087288962073254717551943,68882595903330238838552613955694480987987527"[/QUOTE]
Oops, you misunderstood me. I meant, why does my composite factor show up here [url]https://www.mersenne.org/report_exponent/?exp_lo=5003&full=1[/url]

lisanderke 2021-12-17 10:10

1 Attachment(s)
[QUOTE=LaurV;595467]Is this "new activity" (since I almost disappeared from the forum*) a reason for the "strange" lifting I see in P-1 tops?

I suspected for a while that we have some cheaters there, I mean people jumping hundreds of positions in a week or two, but when you look to their activity, they have very low number of trials, which does not justify the high scores at all, with no possible calculus, unless they did really, and I mean REALLY, deep, DEEP, P-1 (i.e. to very high limits B1 and B2, to get that huge amount of credits, but, BUT, in that case, they were ridiculously unlucky with only 2% (and under) factors found.

So, they either did low-bounds P-1 to find so less factors, which doesn't match with the high credit, or they did high-bounds P-1, which would fit with the credit given, but it doesn't fit with the number of factors.

Or they are cheating (it is easy to cheat the P-1 "no factor" results).

One situation that would still fit, with no cheating, is people doing a lot of P-1 in the very low ranges (which were swapped by TF and lots of ECM already). Is that what's happening here?


---------
*the reason of my "almost disappearance" from the forum I will post later, it is a funny story, not related to math, need time to type...[/QUOTE]
Hi there! A Pre-beta version of Prime95 v30.8 is out ("out" is a big word here, it is meant ONLY for P-1, not any other workload with the current latest build, b5, included.) This version has been pushed out as a pre-beta for P-1 only for the sub two k project specifically. It boasts vast improvements to stage 2 on small(er) exponents. Hence the formerly outrageous bounds you may have seen, see the attachment some of my own testing with 30.8b5 :smile: If you're interested in trying out this pre-beta, I suggest you check out the following thread: [URL]https://www.mersenneforum.org/showthread.php?t=27366[/URL]

lisanderke 2021-12-17 10:24

[QUOTE=SethTro;595468]I'll reach out to both of you next week as I make more progress and start adding upload functionality.

[B]Rough website with list of current files I have: [/B] [URL]https://pminus1.cloudygo.com/[/URL][/QUOTE]
Be sure to keep this thread in the loop! Let us know when some feature/functionality is added/turned on so that we're able to exchange P-1 efforts specifically for this project. Thank you for the work done so far!

lisanderke 2021-12-17 10:27

[QUOTE=Batalov;595456]Just in case (because reading this thread is full of surprises, like "oh, he also doesn't know that", "oh, and that too", but surely the attitude is "what is that qualifying for regional/divisional college competition? ...No, I will go straight for the world record" "...Send me to Beijing, right now!")
-->
this is what you are competing against - [URL]https://members.loria.fr/PZimmermann/records/Pminus1.html[/URL]
Stop thinking in bits, like you will find in [URL="https://mersenneforum.org/showthread.php?t=13977"]many vanity threads[/URL]: "Ah! Look at that factor, it is exceptionally large at 134.47 bits!"
No, [B]66 [/B]digits (or [B]218[/B] bits) is what you want to beat. Good luck![/QUOTE]
I'm afraid I must not have clarified enough that this project would limit itself to Mersenne numbers. With that in mind,

[QUOTE=firejuggler;595439]Also, please refer to [URL]https://www.mersenne.ca/userfactors/pm1/1/bits[/URL] for "big" PM1 results.[/QUOTE]
The 'bit-size' to beat is 173,22. Thus making P-1 efforts done in ranges where ECM-factors of sometimes much higher (say 240) bit levels have been found, quite obsolete. (For my particular search and patience levels).


Though I quite agree with the following statements!
[QUOTE=Prime95;595462]I think top 3 is quite likely (no I didn't do any calculations to back that up!). If we take all Mersennes below 100K to B1=10^9, B2=10^13 (or better) and consider that we get 5 free digits, that's a lot of chances to get lucky.

Even if a record is not found, any new factors are most welcome. One could even lead to a newly fully factored Mersenne.

IMO, that makes the effort worthwhile.[/QUOTE]

firejuggler 2021-12-17 12:21

1 Attachment(s)
I had low hope and I was right, total time, about 27000 seconds (B1 took 11000seconds , B2 15600)

lisanderke 2021-12-17 12:42

[QUOTE=firejuggler;595479]I had low hope and I was right, total time, about 27000 seconds (B1 took 11000seconds , B2 15600)[/QUOTE]
The low hope was probably justified, according to mersenne.ca with T-level 58:

[URL]https://www.mersenne.ca/prob.php?exponent=1811&ecmtlevel=58&b1=40840606334&b2=3.7221820308052E%2B16[/URL]
Probability = [B]0.023413%[/B]

James Heinrich 2021-12-17 14:36

[QUOTE=lisanderke;595466]why does my composite factor show up in the results history of this exponent?
Shouldn't Primenet already have been aware of this composite factor?[/QUOTE]Why wouldn't it show up in the exponent history? Maybe you didn't find any new factors, but we still need to record the space that you searched and what you found in it.


All times are UTC. The time now is 20:23.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2022, Jelsoft Enterprises Ltd.