20140921, 16:17  #1 
Tribal Bullet
Oct 2004
3537_{10} Posts 
M991 discussion
Nobody seems to have come forward to coordinate the factorization of M991, and a few forum members see value in completing the job, so I'll put this here as a placeholder.
Here is a thread for the last time this came up. Finishing 2,991 will complete the base 2 Cunningham tables to 1000 bits. It's a big job, but Tom has done slightly smaller jobs singlehanded so it's well within our combined capabilities. I have reserved 2,991 with Sam, attributed to 'mersenneforum'. Just to make things clear, I personally only have the time to keep this thread current. The tasks I can think of are as follows, volunteers should just step forward and reply in this thread.  come up with a GGNFS input file. This will need 2 algebraic and 3 rational large primes and a really hefty factor base. Bob needs to be in charge of his own siever. If it can't handle a job this big, and he can't compile the GGNFS siever, and he can't run outside binaries, then either he has to be responsible for changing one of those three or he can handle all the small specialq.  list out the specialQ ranges up for reservation. I can handle that; since the reservations will have to be manual, I'd recommend biting off a week of work at a time.  find a place to host the relations. If Tom still has ftp space this would be the best option. Otherwise Carlos has offered his cloud storage, and in the worst case I've been meaning to buy a hard drive from this decade so I can do it. We will need on the order of 150GB or so.  find someone to convert Bob's relations to GGNFS format. Serge previously volunteered to do this. Bob will probably have to snailmail a CD or something.  do the postprocessing. I have really nice highmemory machines at work but Tom is best qualified to get this one done, using MPI on his monster box. Okay, over to you... 
20140921, 16:49  #2 
I moo ablest echo power!
May 2013
11011010001_{2} Posts 
I wouldn't feel comfortable doing the coordinating, but I can throw 18 threads worth of computer at this to generate relations.

20140921, 17:52  #3 
"Serge"
Mar 2008
Phi(4,2^7658614+1)/2
24C9_{16} Posts 

20140921, 18:31  #4 
(loop (#_fork))
Feb 2006
Cambridge, England
13·491 Posts 
I don't seem to have ftp space at the moment. Also at present all my sieving CPUs are involved in factoring Fib(1373) ... of course the offer to do the linear algebra stands.
I'll tell you if either of those changes. Last fiddled with by fivemack on 20140921 at 19:48 
20140921, 18:39  #5 
Jun 2014
2^{3}·3·5 Posts 
I'm unsure of the boundaries of ECM vs. NFS, would it be worth taking it to T60/65 etc. or would that be a waste of time?

20140921, 18:40  #6  
Bamboozled!
"πΊππ·π·π"
May 2003
Down not across
10100110110110_{2} Posts 
Quote:
Paul 

20140921, 19:51  #7  
(loop (#_fork))
Feb 2006
Cambridge, England
13×491 Posts 
Quote:


20140921, 20:49  #8 
Tribal Bullet
Oct 2004
DD1_{16} Posts 
Okay, all that's left is to point to the lasieve5 binaries we need, and work out the parameters for them. Sam has confirmed the reservation.

20140921, 20:56  #9 
I moo ablest echo power!
May 2013
5·349 Posts 
Is there a Windows 64bit version of lasieve5? I have lasieve4 64bit.

20140921, 22:50  #10 
"Curtis"
Feb 2005
Riverside, CA
4771_{10} Posts 
I have not yet acquired a windows binary with the 33bit largeprime limit removed, nor lasieve5.
Should this be a 33 or 34 bit project? If 33, do we expect ~1G raw relations? I am interested in contributing 1% or so of the relations, on the order of a desktopmonth of CPU. 
20140922, 00:07  #11 
Tribal Bullet
Oct 2004
3^{3}×131 Posts 
M1061 required lasieve5 because it required specialQ above 2^32. I just assumed M991 was not smaller by enough to merit the smaller siever.

Thread Tools  
Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
M991 press release?  fivemack  Factoring  9  20150704 20:12 
M991 reservations thread  fivemack  Factoring  148  20150619 03:34 
10,375 LA discussion  Raman  Cunningham Tables  27  20081204 21:17 
P1 discussion  AntonVrba  Prime Cullen Prime  5  20070404 04:59 
New .dat discussion  VJS  Prime Sierpinski Project  7  20060725 14:31 