mersenneforum.org Sieving freakishly big MMs (was "World record" phone number?)
 Register FAQ Search Today's Posts Mark Forums Read

2013-09-03, 18:17   #144
Batalov

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

2×37×131 Posts

Here's some LLR data that would help you guys to scroll some of these columns a screenful down (first four columns for p<3,030,000)
Attached Files
 some_MMp_llr_res.txt (64.5 KB, 111 views)

 2013-09-03, 21:12 #145 Batalov     "Serge" Mar 2008 Phi(4,2^7658614+1)/2 2×37×131 Posts P.S. Because I actually ran (some months ago) a lot of ad hoc tests on smaller MMp (and I also have the divisibility test binary built from a modified Prime95), I know that the most optimal representation for a candidate factor appears to be k*2p+1-(2k-1), or in (k,b,n,c) form: k,2,p+1,-(2k-1) (for P95's worktodo.txt or for an LLR input file). I.e. extra two goes into the exponent, to keep "k" low. P.P.S. I wanted to mention this before but didn't. Here it goes now: It is easy to demonstrate that if a 2kMp+1 is a factor of MMp (with conventional, i.e. rather small, values of k), then it doesn't need a proof of primality - it has no room to have smaller factors of 2k1Mp+1 form => it is prime. The converse, of course, is not true; rather we'd expect to need to collect a lot of primes of form 2kMp+1 to finally hit a divider of MMp.
2013-09-04, 08:52   #146
ET_
Banned

"Luigi"
Aug 2002
Team Italia

2·41·59 Posts

Quote:
 Originally Posted by aketilander Excellent Batalov, I had not seen that! Thank you! Luigi, if we get the permission, it may be nice including this information at the Double Mersenne page also?
I will ask Will about it, but I can't promise to update immediately after: I'm losing my "real-life"TM job and am quite nervous about it.

Luigi

2013-09-04, 09:03   #147
ET_
Banned

"Luigi"
Aug 2002
Team Italia

2·41·59 Posts

Quote:
 Originally Posted by Batalov P.S. Because I actually ran (some months ago) a lot of ad hoc tests on smaller MMp (and I also have the divisibility test binary built from a modified Prime95), I know that the most optimal representation for a candidate factor appears to be k*2p+1-(2k-1), or in (k,b,n,c) form: k,2,p+1,-(2k-1) (for P95's worktodo.txt or for an LLR input file). I.e. extra two goes into the exponent, to keep "k" low. P.P.S. I wanted to mention this before but didn't. Here it goes now: It is easy to demonstrate that if a 2kMp+1 is a factor of MMp (with conventional, i.e. rather small, values of k), then it doesn't need a proof of primality - it has no room to have smaller factors of 2k1Mp+1 form => it is prime. The converse, of course, is not true; rather we'd expect to need to collect a lot of primes of form 2kMp+1 to finally hit a divider of MMp.
Thank you Serge, I will update the table later today.

Luigi

2013-09-04, 13:17   #148
aketilander

"Åke Tilander"
Apr 2011
Sandviken, Sweden

2×283 Posts

Quote:
 Originally Posted by ET_ I'm losing my "real-life"TM job and am quite nervous about it.
I am sorry to hear that! I hope everything will be OK.

2013-09-04, 16:07   #149
LaurV
Romulan Interpreter

"name field"
Jun 2011
Thailand

2×3×5×7×47 Posts

Quote:
 Originally Posted by ET_ I'm losing my "real-life"TM job and am quite nervous about it.
Don't be. We use to say: Every kick in the butt means a step ahead. You will find better, I am sure.

2013-09-04, 18:11   #150
LaurV
Romulan Interpreter

"name field"
Jun 2011
Thailand

2·3·5·7·47 Posts

Quote:
 Originally Posted by ET_ Please send me your results with a couple of lines that explain what to update,
Ok, I checked all the log files I have, and updated this (very old!) table, here and there. Only the third sheet is updated, and from it, only the columns G, H are important. That is for your history page. The values in the blue rows were all tested from scratch (i.e. double-checking the already-known results) and few of them were extended to higher limits. The pink lines were already too high to be tested from scratch and therefore I only re-tested a bit of the "tail" (for example, only the last 156M were tested for MM#13, as explained in the yellow cells).

There are some question marks for MM#29-MM#31 (in columns H-I), that is because the values were written in the table, but I did not find any log file with the "extensions". I will re-check them anyhow.

Also, for MM#34, at the time when I did the tests, k=91 was last tested, and I checked 92 and 93, they did not divide, so the next k would be 112 (see a former post in this thread) but when I checked your tables today, MM#34 (same as #35-#38) seems to be bloody-much advanced.
Attached Files
 mmpp.zip (29.6 KB, 109 views)

Last fiddled with by LaurV on 2013-09-04 at 18:13

2013-09-04, 23:25   #151
Uncwilly
6809 > 6502

"""""""""""""""""""
Aug 2003
101×103 Posts

280916 Posts

Quote:
 Originally Posted by aketilander I am sorry to hear that! I hope everything will be OK.
Ditto

 2013-09-05, 10:50 #152 aketilander     "Åke Tilander" Apr 2011 Sandviken, Sweden 2·283 Posts If I want to TF one specific MMp with one specific possible factor 2*k*Mp+1 which program is best to use? (I am now thinking of medium to large sized MMp:s).
2013-09-05, 13:36   #153
ATH
Einyen

Dec 2003
Denmark

CB716 Posts

Quote:
 Originally Posted by aketilander If I want to TF one specific MMp with one specific possible factor 2*k*Mp+1 which program is best to use? (I am now thinking of medium to large sized MMp:s).
See post #122 - #131 in this thread.

Last fiddled with by ATH on 2013-09-05 at 13:37

2013-09-05, 13:46   #154
ET_
Banned

"Luigi"
Aug 2002
Team Italia

2×41×59 Posts

Quote:
 Originally Posted by LaurV Ok, I checked all the log files I have, and updated this (very old!) table, here and there. Only the third sheet is updated, and from it, only the columns G, H are important. That is for your history page. The values in the blue rows were all tested from scratch (i.e. double-checking the already-known results) and few of them were extended to higher limits. The pink lines were already too high to be tested from scratch and therefore I only re-tested a bit of the "tail" (for example, only the last 156M were tested for MM#13, as explained in the yellow cells). There are some question marks for MM#29-MM#31 (in columns H-I), that is because the values were written in the table, but I did not find any log file with the "extensions". I will re-check them anyhow. Also, for MM#34, at the time when I did the tests, k=91 was last tested, and I checked 92 and 93, they did not divide, so the next k would be 112 (see a former post in this thread) but when I checked your tables today, MM#34 (same as #35-#38) seems to be bloody-much advanced.
Thank you for the updates, LaurV!

I am going to ask Tony Forbes if there are news in the ranges he is working on, update the table on Doublemersennes, and finaally apply your updates.

Before any modificaton is shown, I'd like that no one steps over pre-owned ranges.

Thanks to everyone for the infos!

Luigi

Last fiddled with by ET_ on 2013-09-05 at 14:22 Reason: email sent to Tony and Will.

 Similar Threads Thread Thread Starter Forum Replies Last Post CRGreathouse Number Theory Discussion Group 51 2018-12-16 21:55 LaurV Hobbies 74 2018-07-11 19:33 Batalov Computer Science & Computational Number Theory 40 2013-03-16 09:19 outlnder Soap Box 20 2005-02-03 09:30 nitai1999 Software 7 2004-08-26 18:12

All times are UTC. The time now is 21:00.

Mon Jan 24 21:00:01 UTC 2022 up 185 days, 15:29, 1 user, load averages: 1.32, 1.29, 1.37