Complete Factorization??? - Page 2

dave_0273 · 2005-04-15, 01:50

Well, it seems that you really want to give this at least a go. So, how about this for starters.

Why not just use "factoroverride" (in undoc.txt) to factor the exponent to a higher bit limit than what primenet allows??

Why not use the ECM that comes with prime95 using the bounds/curves shown at http://www.loria.fr/~zimmerma/records/ecm/params.html??

(For those that read this with actual factoring knowledge, yes, I know that I am sending him on a wild goose chase, but isn't that what he is asking for?? Maybe this will keep him quiet for awhile. And maybe after a couple of months he might even come back and say "You know, it is taking too long")

geoff · 2005-04-15, 03:51

There is a project in this thread looking for factors of 2^n-3 where 2^n-1 is a Mersenne prime. Some of these numbers are huge and are being worked on with Prime95 ECM, though most of the effort is going into the smaller ones. Results so far are recorded here.

If you are testing a Mersenne number then a P-1 test is quite efficient, you should probably do that before starting ECM. Keep the save files from the P-1 test because it can be continued to higher bounds later. Look in the Prime95 readme.txt for how to use the 'Pminus1' and 'ECM' keywords in worktodo.ini to do these tests.

VJS · 2005-04-15, 15:21

Just remember to only factor the composite or whats left over after removing (dividing) the 62-bit factor out, also make sure you have found the smallest factor already, try a few different programs.

The ecm applet probably won't work for a number that large.

Run another P-1 on the (co-factor) with a very high B1 bound through stage1 B1=B2, keeping the save file. Then try the highest B2 bounds your memory will allow.

P+1 is also an option then use ecm (IMHO)

I and a couple people are trying to factor an n=991 with no success thus far. You can imagine that your number will have at least one factor harder than n=991.

Regardless to each his own, I'm just fairly confident that you will give up within a month or two. You may be better off waiting until there are better programs for 64-bit computers with raided memory (yes this exists... but it's years off, sort of an extension of dual channel).

Congrats on the large factor and keep up the good work, but generally one factor for each n is good enough. Regardless I wish you luck and if it interests you all the better.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
6^383+1 by GNFS is complete. Thanks!	fivemack	Factoring	50	2008-03-24 19:57
f14 complete	masser	Sierpinski/Riesel Base 5	2	2006-04-23 16:05
Factoring -1.#J% complete	Peter Nelson	Software	4	2005-04-06 00:17
61.5 thru 62m complete to 2^60	nitro	Lone Mersenne Hunters	0	2003-12-07 13:50
60-69M complete through 58 bits	nitro	Lone Mersenne Hunters	2	2003-07-19 02:06

2005-04-15, 01:50	#12
dave_0273 Oct 2003 Australia, Brisbane 726₈ Posts	Well, it seems that you really want to give this at least a go. So, how about this for starters. Why not just use "factoroverride" (in undoc.txt) to factor the exponent to a higher bit limit than what primenet allows?? Why not use the ECM that comes with prime95 using the bounds/curves shown at http://www.loria.fr/~zimmerma/records/ecm/params.html?? (For those that read this with actual factoring knowledge, yes, I know that I am sending him on a wild goose chase, but isn't that what he is asking for?? Maybe this will keep him quiet for awhile. And maybe after a couple of months he might even come back and say "You know, it is taking too long")

2005-04-15, 03:51	#13
geoff Mar 2003 New Zealand 1157₁₀ Posts	There is a project in this thread looking for factors of 2^n-3 where 2^n-1 is a Mersenne prime. Some of these numbers are huge and are being worked on with Prime95 ECM, though most of the effort is going into the smaller ones. Results so far are recorded here. If you are testing a Mersenne number then a P-1 test is quite efficient, you should probably do that before starting ECM. Keep the save files from the P-1 test because it can be continued to higher bounds later. Look in the Prime95 readme.txt for how to use the 'Pminus1' and 'ECM' keywords in worktodo.ini to do these tests.

2005-04-15, 15:21	#14
VJS Dec 2004 13·23 Posts	Just remember to only factor the composite or whats left over after removing (dividing) the 62-bit factor out, also make sure you have found the smallest factor already, try a few different programs. The ecm applet probably won't work for a number that large. Run another P-1 on the (co-factor) with a very high B1 bound through stage1 B1=B2, keeping the save file. Then try the highest B2 bounds your memory will allow. P+1 is also an option then use ecm (IMHO) I and a couple people are trying to factor an n=991 with no success thus far. You can imagine that your number will have at least one factor harder than n=991. Regardless to each his own, I'm just fairly confident that you will give up within a month or two. You may be better off waiting until there are better programs for 64-bit computers with raided memory (yes this exists... but it's years off, sort of an extension of dual channel). Congrats on the large factor and keep up the good work, but generally one factor for each n is good enough. Regardless I wish you luck and if it interests you all the better.