mersenneforum.org low k/n factoring
 Register FAQ Search Today's Posts Mark Forums Read

 2005-10-04, 04:27 #1 jasong     "Jason Goatcher" Mar 2005 5·701 Posts low k/n factoring When I finish my sieving commitment with Riesel Sieve, I'd like to get into something more hands on. While I can't seem to find the link, I seem to remember there being someone who desired low k/n factorizations. I believe both k and n are below 10,000 on his page. If I'm not totally off my rocker, maybe someone has a link for me?
 2005-10-04, 08:30 #2 geoff     Mar 2003 New Zealand 13·89 Posts This thread might be what you are looking for, the projects mentioned were selected from this page. There are also two mersenneforum projects looking for factors of small PSP candidates and factors of M(p)-2 for Mersenne primes M(p), which are both of the form k*2^n+-c although k or n are larger than 10,000 in some cases.
 2005-10-04, 18:43 #3 VJS     Dec 2004 1001010112 Posts Well there are alot of people who are just factoring for the fun of it k&n both below 10k not sure about that one. However over at SoB there are a few people doing some limited searching for fun... http://www.free-dc.org/forum/showthr...5&pagenumber=4 And here http://www.aooq73.dsl.pipex.com/scores_p.htm If you scroll down a bit to: Smallest unfactored candidates by k... Most everything has been done to at least the 30 digit level. It's all in good fun and doesn't provide much if any benifit to the project. The lowest is the killer currently 24737*2^991+1 it's been done to the 50-digit level I believe. I also personally P-1 it to B1=B2=4.2G and 3 curves of P+1 to B1=10e8 (I believe it was 8 not 9) and B2=10e10. You might want to look at the cunn. tables or oddperfect (Small Comp's)
 2005-10-05, 00:08 #4 jasong     "Jason Goatcher" Mar 2005 5×701 Posts Thanks, guys. Actually, I believe Michael Klasson(sorry if the name is misspelled) is what I'm looking for. One question though: There's something called the doecm_ecm_result_file that I don't know what I'm supposed to do with, if anything at all. Here's the contents(after and before the "-----"): ----- GMP-ECM 5.0.3 [powered by GMP 4.1.2] [ECM] Input number is 86259985035811817016........3547741368129927898751 (233 digits) Using B1=5000000, B2=8316852668, polynomial Dickson(12), sigma=4131865911 Step 1 took 317656ms ----- Should I be doing anything with this?(I'm such a n00b ) Last fiddled with by smh on 2005-10-05 at 07:15 Reason: Trimmed the long composite
2005-10-05, 00:59   #5
fetofs

Aug 2005
Brazil

2×181 Posts

Quote:
 Originally Posted by jasong GMP-ECM 5.0.3 [powered by GMP 4.1.2] [ECM] Input number is 86259985035811817016........3547741368129927898751 (233 digits) Using B1=5000000, B2=8316852668, polynomial Dickson(12), sigma=4131865911 Step 1 took 317656ms
This is the output of GMP-ECM while it's working. It tells you:

a)The program and version (GMP-ECM 5.0.3)
b) The number you're trying to factor (862599...) and the number of digits
c)The value of B1 and B2, along with the polynomial and sigma (specific ECM values)
d)The time it took to do steps 1 and 2.

To make it more clear, this only interests you when it says:

(The number) has a factor: x

BTW: Shouldn't you upgrade to GMP-ECM 6.0.1? Just a thought. :wink

Last fiddled with by smh on 2005-10-05 at 07:15

2005-10-05, 07:18   #6
smh

"Sander"
Oct 2002
52.345322,5.52471

29·41 Posts

Quote:
 Originally Posted by jasong Thanks, guys. Actually, I believe Michael Klasson(sorry if the name is misspelled) is what I'm looking for. One question though: There's something called the doecm_ecm_result_file that I don't know what I'm supposed to do with, if anything at all. Here's the contents(after and before the "-----"):
Actually, doecm is a program is a program that automatically gets a number from the server, runs ECM and submits the results. Something like the ecmclient.

It's been a long time since i used it so i forgot the details.