mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Factoring (https://www.mersenneforum.org/forumdisplay.php?f=19)
-   -   P-1 factoring attempts at smallest-remaining Mersenne numbers with no known factors (https://www.mersenneforum.org/showthread.php?t=21115)

 UberNumberGeek 2016-03-16 20:23

P-1 factoring attempts at smallest-remaining Mersenne numbers with no known factors

Hello, Geniuses,

I recently ran 4-month long P-1 tests on M1277, M1619, M1753, and M2137.

I would like to ask for your opinion on what I should do next, please.

I still have the 244 MB files generated by the tests, am I able to increase the B1 and B2 for these tests and, essentially, pick up where I left off?

If so, should I do that since I already have so much time invested?

Thank you for your time and any explanations provided!

 VictordeHolland 2016-03-16 21:11

If you want to attempt to factor these low Mersenne numbers, you'll want to run ECM curves with high bounds. ECM with B1=800,000,000 for starters. Prime95 has options to do that. You could also give GMP-ECM a try for the stage 2. Curves wirh higher bounds can find larger factors, but take longer to complete.

 UberNumberGeek 2016-03-16 21:26

[QUOTE=VictordeHolland;429355]If you want to attempt to factor these low Mersenne numbers, you'll want to run ECM curves with high bounds. ECM with B1=800,000,000 for starters. Prime95 has options to do that. You could also give GMP-ECM a try for the stage 2. Curves wirh higher bounds can find larger factors, but take longer to complete.[/QUOTE]

Thank you for replying, Victor!

I am running ECM curves at B1=800,000,000 and B2=80,000,000,000, but on my PC 1 curve takes 2 hours to run, and according to the ECM chart, there are still over 300,000 curves to be run for these numbers at those bounds.

I was hoping that P-1 factoring would find a nice, "smooth" factor and save time. In the 4 months it took my P-1 factoring to complete, I could have only completed a few thousand curves of ECM, still a small percentage.

And, again, I was hoping that, since I have already spent 4 months on P-1 factoring, that I could up the bounds and pick up where I left off and maybe find a decent-sized factor that was bigger than the 65-digit "limit" those bounds imply.

Thank you again!

 Prime95 2016-03-16 21:39

For these small Mersennes it is an absolute must to use prime95 for stage 1 and GMP-ECM for stage 2. There are forum threads that show how to do this.

 UberNumberGeek 2016-03-16 22:01

[QUOTE=Prime95;429360]For these small Mersennes it is an absolute must to use prime95 for stage 1 and GMP-ECM for stage 2. There are forum threads that show how to do this.[/QUOTE]

I am honored that you replied to my thread, thank you.

I am melancholy that I have been doing things the wrong way and have been wasting precious time. :down:

Thank you for letting me know, I am sorry to waste everyone's time. I will try to find how to start using GMP-ECM.

Thank you for everything Prime95 has given and taught me.

 VBCurtis 2016-03-16 22:19

I am running ECM with large bounds (B1 = 4.5e9 B2 = 2e14) on M1277. I have 3 Xeon cores doing so, making slow progress but thus far too lazy to submit manual curve results to George.

1277 has had by far the most work done, so it is least likely to produce a factor per unit of CPU time invested. I'd personally like the help finishing a t65 on it, but the other three get less attention so there's reason to take shots at those too. I second the suggestion to run curves at B1 no lower than 8e8 on these four.

As for getting GMP-ECM to work with P95, P95 needs GmpEcmHook=1 in the settings file (prime.txt or local.txt, I forget). B2 is set no higher than B1, and the results file saves the residue from stage 1. GMP-ECM is then run with -resume <inputfile, B1 set to 1, and B2 set to whatever you like (or, perhaps, not set and GMP-ECM will pick it for you). Lots more detail if you find the other thread.
Good luck!

 ATH 2016-03-16 22:45

I have previously done big P-1 run and 2 big P+1 runs on M1277 using only GMP-ECM for both stage 1 and stage 2:

P-1: B1=10[sup]12[/sup] B2=2.35*10[sup]17[/sup]

P+1: B1=5*10[sup]11[/sup] B2=5.4*10[sup]16[/sup]

P+1: B1=5*10[sup]11[/sup] B2=7.8*10[sup]16[/sup]

I have the save files if anyone want to take them higher.

 GP2 2016-10-18 23:20

I would be interested in P-1 or P+1 save files for small exponents.

For exponents in the low single-digit-thousands, I think these files should be only a few hundred bytes long. The intermediate files (.bu or .bu2) might be very large (hundreds of MB), but the non-bu final savefile should be really tiny for those very small exponents, assuming the P-1 or P+1 test actually ran to completion.

Amazon EC2 cloud has some instances with large amounts of memory, all the way up to x1.16xlarge with 976 GiB and 32 cores, and x1.32xlarge with 1952 GiB and 64 cores. The spot prices are about \$0.70/hour and \$1.50/hour for these, and there are others with less memory for lower cost. I have no idea how long stage 2 might take with GMP-ECM, but if it's days rather than months then something might be feasible.

When you guys did your tests, were you constrained by available memory or by CPU time, or both?

Sadly, openmp does not seem to work properly with P-1 for GMP-ECM, so only one thread could be used per exponent. If this bug could be fixed then perhaps it might go faster.

 GP2 2016-11-16 19:07

[QUOTE=ATH;429378]I have previously done big P-1 run and 2 big P+1 runs on M1277 using only GMP-ECM for both stage 1 and stage 2:

P-1: B1=10[sup]12[/sup] B2=2.35*10[sup]17[/sup]

P+1: B1=5*10[sup]11[/sup] B2=5.4*10[sup]16[/sup]

P+1: B1=5*10[sup]11[/sup] B2=7.8*10[sup]16[/sup]

I have the save files if anyone want to take them higher.[/QUOTE]

I obtained these save files and took them to:

Pā1: B1=1014774826757 (unchanged) B2=4.55e18
P+1: B1=ā510021907147 (unchanged) B2=1.00e18
P+1: B1=ā512993898541 (unchanged) B2=9.15e17

with no result.

 petrw1 2016-11-16 22:21

[QUOTE=Prime95;429360]For these small Mersennes it is an absolute must to use prime95 for stage 1 and GMP-ECM for stage 2. There are forum threads that show how to do this.[/QUOTE]

When you say "...smaller Mersennes..." up to what size?
I'm looking to try to ECM Factor a Prime or 2 for Primes under 20,000.

 petrw1 2016-12-14 18:36

Interested in furthering P-1 for sub 20,000 Exponents.

I put together this list of sub-20,000 Exponents with the lowest B1/B2 done so far.
Granted even these are VERY large but I am thinking of taking these higher.
I have a decent PC with 16GB of RAM running nothing but GIMPS.

Unless the wiser among you advise this a a big waste of time...
I am also going to run LOTS of ECM as well.
My 2017 goal is to first-factor at least 1 sub-20,000 exponent.

I listed the latest B1/B2 on record and who did that test.
If any any you still have the the save files PM me.

[CODE]7127 63 60,000,000,000 4,800,000,000,000 Mikr
7621 63 60,000,000,000 4,800,000,000,000 Mikr
8291 62 60,000,000,000 4,800,000,000,000 Mikr
10399 63 10,000,000,000 100,000,000,000 bcp19
10463 63 10,000,000,000 500,000,000,000 bcp19
10771 63 10,000,000,000 500,000,000,000 bcp19
10831 63 10,000,000,000 100,000,000,000 bcp19
11027 63 10,000,000,000 100,000,000,000 bcp19
11159 63 10,000,000,000 100,000,000,000 bcp19
11251 63 10,000,000,000 100,000,000,000 bcp19
11351 63 10,000,000,000 100,000,000,000 bcp19
11423 63 10,000,000,000 100,000,000,000 bcp19
11467 63 10,000,000,000 100,000,000,000 bcp19
11483 63 10,000,000,000 100,000,000,000 bcp19
11489 63 10,000,000,000 100,000,000,000 bcp19
11657 63 10,000,000,000 100,000,000,000 bcp19
11821 63 10,000,000,000 100,000,000,000 bcp19
11839 63 10,000,000,000 100,000,000,000 bcp19
12149 63 10,000,000,000 100,000,000,000 bcp19
12161 63 10,000,000,000 100,000,000,000 bcp19
12269 63 10,000,000,000 100,000,000,000 bcp19
12281 63 10,000,000,000 100,000,000,000 bcp19
12377 63 10,000,000,000 100,000,000,000 bcp19
12517 63 10,000,000,000 100,000,000,000 bcp19
12637 63 10,000,000,000 100,000,000,000 bcp19
13007 63 10,000,000,000 500,000,000,000 bcp19
13217 63 10,000,000,000 100,000,000,000 bcp19
13219 63 10,000,000,000 100,000,000,000 bcp19
13327 63 10,000,000,000 100,000,000,000 bcp19
13523 63 10,000,000,000 100,000,000,000 bcp19
13597 63 10,000,000,000 100,000,000,000 bcp19
13691 63 10,000,000,000 100,000,000,000 bcp19
14153 63 10,000,000,000 100,000,000,000 bcp19
14173 63 10,000,000,000 100,000,000,000 bcp19
14243 63 10,000,000,000 100,000,000,000 bcp19
14447 63 10,000,000,000 100,000,000,000 bcp19
14489 63 10,000,000,000 100,000,000,000 bcp19
14557 63 10,000,000,000 100,000,000,000 bcp19
14723 63 10,000,000,000 100,000,000,000 bcp19
14867 63 10,000,000,000 100,000,000,000 bcp19
14951 63 10,000,000,000 9,887,122,214,540,710 bcp19
15017 63 10,000,000,000 100,000,000,000 bcp19
15077 63 10,000,000,000 100,000,000,000 bcp19
15259 63 10,000,000,000 100,000,000,000 bcp19
15349 63 10,000,000,000 100,000,000,000 bcp19
15451 63 10,000,000,000 100,000,000,000 bcp19
15497 63 10,000,000,000 100,000,000,000 bcp19
15559 63 1,446,830,000 43,643,783,079 Jayder
15643 63 10,000,000,000 100,000,000,000 c10ck3r
15649 63 1,446,830,000 43,643,783,079 Jayder
16057 63 1,446,830,000 43,643,783,079 Jayder
16061 63 1,851,000,000 43,643,783,079 Jayder
16253 63 1,446,830,000 43,643,783,079 Jayder
16349 63 10,000,000,000 100,000,000,000 c10ck3r
16369 63 10,000,000,000 100,000,000,000 c10ck3r
16381 63 10,000,000,000 100,000,000,000 c10ck3r
16649 63 984,343,260 49,217,163,000 Jocelyn Larouche
16673 63 982,360,150 49,118,007,500 Jocelyn Larouche
16843 63 20,000,000,000 200,000,000,000 blahpy
17053 63 10,000,000,000 100,000,000,000 blahpy
17077 63 949,978,502 47,498,925,100 Jocelyn Larouche
17203 63 1,935,060,000 35,360,118,869 Jayder
17239 63 20,000,000,000 100,000,000,000 David Campeau
17359 63 2,000,000,000 130,000,000,000 David Campeau
17393 63 1,172,220,000 35,360,118,869 Jayder
17471 63 2,840,000,000 2,840,000,000 Never Odd Or Even
17761 63 1,172,220,000 35,360,118,869 Jayder
17827 63 3,000,000,000 3,000,000,000 Sergiosi
18119 63 2,000,000,000 130,000,000,000 David Campeau
18149 63 1,172,220,000 35,360,118,869 Jayder
18341 63 1,172,220,000 35,360,118,869 Jayder
18397 63 1,172,220,000 35,360,118,869 Jayder
18413 63 1,172,220,000 35,360,118,869 Jayder
18439 63 1,172,220,000 35,360,118,869 Jayder
18457 63 1,172,220,000 35,360,118,869 Jayder
18539 63 1,172,220,000 35,360,118,869 Jayder
18553 63 1,172,220,000 35,360,118,869 Jayder
18583 63 1,172,220,000 35,360,118,869 Jayder
18587 63 1,172,220,000 35,360,118,869 Jayder
19013 63 24,250,000,000 50,000,000,000 Sergiosi
19157 63 1,172,220,000 35,360,118,869 Jayder
19219 63 1,172,220,000 35,360,118,869 Jayder
19373 63 1,172,220,000 35,360,118,869 Jayder
19423 63 1,172,220,000 35,360,118,869 Jayder
19433 63 1,172,220,000 35,360,118,869 Jayder
19483 63 1,172,220,000 35,360,118,869 Jayder
19501 63 1,172,220,000 35,360,118,869 Jayder
19507 63 10,000,000,000 100,000,000,000 bcp19
19531 63 10,000,000,000 100,000,000,000 bcp19
19583 63 10,000,000,000 100,000,000,000 bcp19
19709 63 10,000,000,000 100,000,000,000 bcp19
19753 63 10,000,000,000 100,000,000,000 bcp19
19813 63 10,000,000,000 100,000,000,000 bcp19
19853 63 10,000,000,000 100,000,000,000 bcp19[/CODE]

All times are UTC. The time now is 15:14.