Factoring humongous Cunningham numbers - Page 113

Dubslow · 2015-04-18, 14:39

Quote:

Originally Posted by LaurV

Ha! They are not!

hehe, yafu decided to store them in number of bits. Good to know.

Code:

n: 15245685683654194070528451784367735927134564102584982532265254072253746820313011591722538299058918175786099377638450142276578178842074686064451121444334808941
skew: 76967649.34
c0: -388219652152137289945728452849156092151552
c1: 31897168474946617482234360164118068
c2: -8309850439619272290936949
c3: -13444173881982964058
c4: -21258070854
c5: 720
Y0: -7330995270849431836665563364183
Y1: 132282895415155891
rlim: 36800000
alim: 36800000
lpbr: 30
lpba: 30
mfbr: 60
mfba: 60
rlambda: 2.6
alambda: 2.6

In the Msieve format (nfs.fb) they are indeed stored as the 2^(bits) rather than just bits.

LaurV · 2015-04-24, 11:45

It started filtering today at 10:20 AM after ~92408000 relations (ninety two millions), and it crashed copiously at 13:45. The reason: no disk space. The all data set is over 17 gigs, and it digested the allowed space on the SSD. Now it is 18:30 and i just arrived home, I will resume it soon. Meantime, beside of ~13-14 gig of relations (nfs.dat) the "matrix" phase crash left on disk:
~2 gigs of nfs.dat.mat, plus
~500 megs of nfs.dat.chk and nfs.dat.bak.chk (together both) plus
~200 megs of nfs.dat.cyc plus
~100 megs of other files together (.p from the poly, plus .hc, .log, .blahblah)

Do I have to follow a "special" procedure to salvage about 3 hours of matrix work? (is it possible?)
If not, I will resume in the "normal way". I made 9 gigs free on this SSD by deleting some movies.

LaurV · 2015-04-24, 13:02

Well, that's it. We resumed.

Code:

nfs: commencing msieve linear algebra
linear algebra completed 168181 of 5753361 dimensions (2.9%, ETA 27h52m)

The "relation tricks" file.
nfs.zip

xilman · 2015-04-25, 08:36

Fourteen factors have been reported over the last month; they are given in the table below. There are now 189 composites left to be factored in the tables on the web.

Code:

9-5	233	C209	856416019019857159138205934516172807990172385292801651863415546005310716474553681.	P128			J Becker		SNFS	2015-03-22
8+5	248	C163	386429205030022897416086016547599971116379975203748491519630715356616273.	P91				J Becker		SNFS	2015-03-24
11+2	214	C214	6528652865996133234378474930552485352555278144523274695163032424968305577742813.	P135			R Silverman		SNFS	2015-03-27
12+11	218	C221	12524238801470181304926531708940532716851886750204629426266238601510995893017.	P145				J Becker		SNFS	2015-04-02
8+7	248	C191	164940544724737116328660018230563660268416243899834086210478713146657.	P122					J Becker		SNFS	2015-04-04
11+3	214	C206	377426956761375702050512156243087464836728791472863287841928362974764923026294732803472309.	P116		R Silverman		SNFS	2015-04-06
8+3	248	C161	5375486127801069000240107169355386392836389726125206201208234009156022185760689.	P83			J Becker		SNFS	2015-04-06
4+3	373	C184	5011732725095091402415465843599219922819625373740880078457035129494921062877155333530559.	P96		J Becker		SNFS	2015-04-09
9+8	236	C171	2624228962249296027594339625733235450567020244082934640556433.	P111						J Becker		SNFS	2015-04-12
5+3	323	C199	858589742488823489249628863164220324270930596610040604565527781561146607.	P127				J Becker		SNFS	2015-04-14
5-3	323	C164	58155176825209089298729578828208843536270139597234322070872395686779.	P97					J Becker		SNFS	2015-04-17
11+8	214	C208	8208116181867158885324930022594195094090492600442416789410841933446936029134585603408777.	P120		R Silverman		SNFS	2015-04-21
10+9	226	C190	80722215428434057205303209313546914367270528193004681.	P137							J Becker		SNFS	2015-04-21
8+5	251	C191	2568416537772921426695619743837540857133868356645501369605700609061156523676291.	P113			J Becker		SNFS	2015-04-24

LaurV · 2015-04-26, 05:49

Ugly spilt ...(as opposite to "nice split" for (almost-)brilliant numbers)

Code:

prp91 = 4128357108091388288089124279269514018391194766611965045335651312012217582958814315306234767
prp67 = 3692918341238784274586962597797592376098995727814752526742535257923

Someone who need the logs, PM.

By an error/idiotic_thinking/force_of_habit, the batch command when I resumed last time, was "factor(blah-blah)" instead of "nfs(blah-blah)", so after the factorization finished last night, then the C161 started and wasted about 12 hours doing ECM.

. Nothing found, of course.

Now I stopped it, and I restarted with the right command. As expected, it found the right form and poly, as BB said. Let's sieve it....

jyb · 2015-04-26, 07:52

Quote:

Originally Posted by LaurV

Ugly spilt ...(as opposite to "nice split" for (almost-)brilliant numbers)

Code:

prp91 = 4128357108091388288089124279269514018391194766611965045335651312012217582958814315306234767
prp67 = 3692918341238784274586962597797592376098995727814752526742535257923

Someone who need the logs, PM.

By an error/idiotic_thinking/force_of_habit, the batch command when I resumed last time, was "factor(blah-blah)" instead of "nfs(blah-blah)", so after the factorization finished last night, then the C161 started and wasted about 12 hours doing ECM.

. Nothing found, of course.

Now I stopped it, and I restarted with the right command. As expected, it found the right form and poly, as BB said. Let's sieve it....

Or then again, maybe let's not. If the C161 in question is 8+3,248 (which is actually a Homogeneous Cunningham, unlike the one you just finished and keep talking about in this thread), then it has already been factored. It's always worth a look in the factordb before you start a job.

Might I suggest 11-5,223 C162 instead? It's the smallest remaining composite by actual number of digits. Or 7+3,268 C184, which is the smallest remaining unreserved composite by SNFS difficulty? If you do decide to factor either of these, please reserve it.

LaurV · 2015-04-26, 08:29

Quote:

Originally Posted by jyb

it has already been factored

Well... indeed. It wasn't me. Thanks for telling me, I don't follow this project, and it didn't occur to me that I could check the fdb

This concludes my contribution for this project for now. Happy to move my cores to my other things.

xilman · 2015-04-26, 09:53

Quote:

Originally Posted by LaurV

Ugly spilt ...(as opposite to "nice split" for (almost-)brilliant numbers)

Code:

prp91 = 4128357108091388288089124279269514018391194766611965045335651312012217582958814315306234767
prp67 = 3692918341238784274586962597797592376098995727814752526742535257923

Someone who need the logs, PM.

By an error/idiotic_thinking/force_of_habit, the batch command when I resumed last time, was "factor(blah-blah)" instead of "nfs(blah-blah)", so after the factorization finished last night, then the C161 started and wasted about 12 hours doing ECM.

. Nothing found, of course.

Now I stopped it, and I restarted with the right command. As expected, it found the right form and poly, as BB said. Let's sieve it....

Thanks for completing this one. As mentioned earlier it is a GCW number. In particular, it is W(862) aka 862 . 2⁸⁶²-1.

I'll PM you for details.

Paul

R.D. Silverman · 2015-04-27, 13:57

Quote:

Originally Posted by xilman

Thanks for completing this one. As mentioned earlier it is a GCW number. In particular, it is W(862) aka 862 . 2⁸⁶²-1.

I'll PM you for details.

Paul

Why is it being discussed in this sub-group?

xilman · 2015-04-27, 14:45

Quote:

Originally Posted by R.D. Silverman

Why is it being discussed in this sub-group?

Ancient history. LaurV offered to run GNFS on a HCN or similar. I posted the smallest such composite then remaining and also a GCW candidate. He was told that all the HCN candidates were better done with SNFS so he ran the GCW for me. If you want more detail, go read back issues...

Paul

Looks like post #1206 is the one

jyb · 2015-05-15, 23:56

From another thread:

Quote:

Originally Posted by Batalov

I am tempted to try the 6th hole 2,1100+ as an octic.

Serge, in your experience how much of a penalty does one pay when using octic polynomials? As an experiment, I'm looking at 10+9,570L. Ordinarily this would be a quartic with difficulty ~228, but as an octic its difficulty drops to ~152. Such a huge advantage seems worth it, but when I try sieving I get a fairly crappy yield. Not as bad as the quartic would be, but the composite has only 121 digits, so it would probably go faster with GNFS.

Can you share some of the ways you massaged the various parameters to give a good result?

Thanks.

2015-04-24, 13:02	#1235
LaurV Romulan Interpreter Jun 2011 Thailand 22677₈ Posts	Well, that's it. We resumed. Code: nfs: commencing msieve linear algebra linear algebra completed 168181 of 5753361 dimensions (2.9%, ETA 27h52m) The "relation tricks" file. nfs.zip

2015-04-26, 05:49	#1237
LaurV Romulan Interpreter Jun 2011 Thailand 3×3,221 Posts	Ugly spilt ...(as opposite to "nice split" for (almost-)brilliant numbers) Code: prp91 = 4128357108091388288089124279269514018391194766611965045335651312012217582958814315306234767 prp67 = 3692918341238784274586962597797592376098995727814752526742535257923 Someone who need the logs, PM. By an error/idiotic_thinking/force_of_habit, the batch command when I resumed last time, was "factor(blah-blah)" instead of "nfs(blah-blah)", so after the factorization finished last night, then the C161 started and wasted about 12 hours doing ECM. . Nothing found, of course. Now I stopped it, and I restarted with the right command. As expected, it found the right form and poly, as BB said. Let's sieve it....

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
New phi for homogeneous Cunningham numbers	wpolly	Factoring	26	2016-07-29 04:34
Mathematics of Cunningham Numbers (3rd ed., 2002, A.M.S.)	Xyzzy	Cunningham Tables	42	2014-04-02 18:31
Don't know how to work on Cunningham numbers.	jasong	GMP-ECM	6	2006-06-30 08:51
Doing Cunningham numbers but messed up.	jasong	Factoring	1	2006-04-03 17:18
Need help factoring Cunningham numbers	jasong	Factoring	27	2006-03-21 02:47

2015-04-24, 11:45	#1234
LaurV Romulan Interpreter Jun 2011 Thailand 3·3,221 Posts	It started filtering today at 10:20 AM after ~92408000 relations (ninety two millions), and it crashed copiously at 13:45. The reason: no disk space. The all data set is over 17 gigs, and it digested the allowed space on the SSD. Now it is 18:30 and i just arrived home, I will resume it soon. Meantime, beside of ~13-14 gig of relations (nfs.dat) the "matrix" phase crash left on disk: ~2 gigs of nfs.dat.mat, plus ~500 megs of nfs.dat.chk and nfs.dat.bak.chk (together both) plus ~200 megs of nfs.dat.cyc plus ~100 megs of other files together (.p from the poly, plus .hc, .log, .blahblah) Do I have to follow a "special" procedure to salvage about 3 hours of matrix work? (is it possible?) If not, I will resume in the "normal way". I made 9 gigs free on this SSD by deleting some movies.