fun with msieve - Page 2

jasonp · 2007-09-04, 14:29

Quote:

Originally Posted by masser

We have a handful of unfactored candidates in our dat file that our perfect candidates for msieve's QS. I thought that as a fun side project we could individually reserve these candidates and factor them.

Where is the list of composites? I need some SNFS practice, and these look like they would take 2-3 hours each using ggnfs.

masser · 2007-09-04, 16:33

See attached - I have included all of our k*5^n+-1 candidates with n<400.

ltd · 2007-09-04, 16:46

One other question.

Shall we setup an ecmnet server for that?

Lars

masser · 2007-09-04, 18:35

Quote:

Originally Posted by ltd

One other question.

Shall we setup an ecmnet server for that?

Lars

We could if there is interest - however, this factoring work really is not that important for the overall project. Just some fun for people interested in factoring.

Is anyone interested in setting up an ecmnet server for the low-n composites in our dat file?

ValerieVonck · 2007-09-04, 19:30

Quote:

Originally Posted by CedricVonck

Ok I will do 105782*5^138-1 first.

Code:

Number: c102
N=303579451705769573163937715772567363991733593570471099121676691878768394872167846187949180603027343749
  ( 102 digits)
Divisors found:
 r1=387062941725690490886834511486286812101841893 (pp45)
 r2=784315466503416293906930591904861673729193704199342408993 (pp57)
Version: GGNFS-0.77.1-20050930-pentium4
Total time: 27.54 hours.
Scaled time: 13.99 units (timescale=0.508).
Factorization parameters were as follows:
name: c102
n:  303579451705769573163937715772567363991733593570471099121676691878768394872167846187949180603027343749
m:  14867484900659121556
deg: 5
c5: 417912
c4: -2167692490
c3: -542889460261548
c2: 265490848483642090
c1: 1265368479597174727295
c0: -988684725927987579508615
skew: 1635.250
type: gnfs
# adj. I(F,S) = 49.562
# E(F1,F2) = 1.066204e-003
# GGNFS version 0.77.1-20050930-pentium4 polyselect.
# Options were: 
# lcd=1, enumLCD=24, maxS1=59.00000000, seed=1188749814.
# maxskew=2000.0
# These parameters should be manually set:
rlim: 2300000
alim: 2300000
lpbr: 26
lpba: 26
mfbr: 49
mfba: 49
rlambda: 2.6
alambda: 2.6
qintsize: 100000

type: gnfs
Factor base limits: 2300000/2300000
Large primes per side: 3
Large prime bits: 26/26
Max factor residue bits: 49/49
Sieved algebraic special-q in [1150000, 2850001)
Primes: RFBsize:169511, AFBsize:169417, largePrimes:4522881 encountered
Relations: rels:4638652, finalFF:411549
Max relations in full relation-set: 28
Initial matrix: 339004 x 411549 with sparse part having weight 38832499.
Pruned matrix : 287396 x 289155 with weight 25610471.
Total sieving time: 23.61 hours.
Total relation processing time: 0.35 hours.
Matrix solve time: 3.29 hours.
Time per square root: 0.28 hours.
Prototype def-par.txt line would be:
gnfs,101,5,maxs1,maxskew,goodScore,efrac,j0,j1,eStepSize,maxTime,2300000,2300000,26,26,49,49,2.6,2.6,100000
total time: 27.54 hours.
 --------- CPU info (if available) ----------

I will start the next number

ltd · 2007-09-04, 21:53

Quote:

Originally Posted by masser

We could if there is interest - however, this factoring work really is not that important for the overall project. Just some fun for people interested in factoring.

Is anyone interested in setting up an ecmnet server for the low-n composites in our dat file?

If there is interest i can setup the server without any problems.
I have a test server already available.
So i would only have to create the ini file and fire it up again.
It would not run with respawn after a crash but this should
be no problem as the ecmnet server from PSP is now running for months
without any problems.

Lars

ltd · 2007-09-04, 21:54

Quote:

Originally Posted by CedricVonck

Code:

Number: c102
N=303579451705769573163937715772567363991733593570471099121676691878768394872167846187949180603027343749
  ( 102 digits)
Divisors found:
 r1=387062941725690490886834511486286812101841893 (pp45)
 r2=784315466503416293906930591904861673729193704199342408993 (pp57)
Version: GGNFS-0.77.1-20050930-pentium4
Total time: 27.54 hours.
Scaled time: 13.99 units (timescale=0.508).
Factorization parameters were as follows:
name: c102
n:  303579451705769573163937715772567363991733593570471099121676691878768394872167846187949180603027343749
m:  14867484900659121556
deg: 5
c5: 417912
c4: -2167692490
c3: -542889460261548
c2: 265490848483642090
c1: 1265368479597174727295
c0: -988684725927987579508615
skew: 1635.250
type: gnfs
# adj. I(F,S) = 49.562
# E(F1,F2) = 1.066204e-003
# GGNFS version 0.77.1-20050930-pentium4 polyselect.
# Options were: 
# lcd=1, enumLCD=24, maxS1=59.00000000, seed=1188749814.
# maxskew=2000.0
# These parameters should be manually set:
rlim: 2300000
alim: 2300000
lpbr: 26
lpba: 26
mfbr: 49
mfba: 49
rlambda: 2.6
alambda: 2.6
qintsize: 100000

type: gnfs
Factor base limits: 2300000/2300000
Large primes per side: 3
Large prime bits: 26/26
Max factor residue bits: 49/49
Sieved algebraic special-q in [1150000, 2850001)
Primes: RFBsize:169511, AFBsize:169417, largePrimes:4522881 encountered
Relations: rels:4638652, finalFF:411549
Max relations in full relation-set: 28
Initial matrix: 339004 x 411549 with sparse part having weight 38832499.
Pruned matrix : 287396 x 289155 with weight 25610471.
Total sieving time: 23.61 hours.
Total relation processing time: 0.35 hours.
Matrix solve time: 3.29 hours.
Time per square root: 0.28 hours.
Prototype def-par.txt line would be:
gnfs,101,5,maxs1,maxskew,goodScore,efrac,j0,j1,eStepSize,maxTime,2300000,2300000,26,26,49,49,2.6,2.6,100000
total time: 27.54 hours.
 --------- CPU info (if available) ----------

I will start the next number

Have you reported it to the sievesubmission already?

fivemack · 2007-09-04, 22:52

Doing these numbers by gnfs is deeply inefficient: you have all the time spent in finding a suitable polynomial, and then you have a gnfs-polynomial with coefficients like -988684725927987579508615 rather than an SNFS polynomial with coefficients like 105782.

snfs for 48394*5^143-1, using poly file

Code:

n: 434012402448225262913811004503886088045932777057101467253834000195600317795197042869403958320617675781249
type: snfs
skew: 1
c4: 48394
c0: -5
Y0: 14551915228366851806640625
Y1: -1

where Y0=5^36 and the number factored is c4*Y0^4+c0, IE 48394*5^144-5, is really very quick: three thousand relations per special-Q, more than 500 relations found per second, about a million relations needed so it takes half an hour.

Code:

Number: 48394*5^143-1
N=434012402448225262913811004503886088045932777057101467253834000195600317795197042869403958320617675781249
  ( 105 digits)
SNFS difficulty: 105 digits.
Divisors found:
 r1=170900861476382989478928598031890143841721828653 (pp48)
 r2=2539556551669003708992032638434970040504709222935330416933 (pp58)
Version: GGNFS-0.77.1-20060513-nocona
Total time: 0.49 hours.
Scaled time: 0.91 units (timescale=1.848).
Factorization parameters were as follows:
n: 434012402448225262913811004503886088045932777057101467253834000195600317795197042869403958320617675781249
type: snfs
skew: 1
c4: 48394
c0: -5
Y0: 14551915228366851806640625
Y1: -1
Factor base limits: 450000/500000
Large primes per side: 3
Large prime bits: 25/25
Max factor residue bits: 44/44
Sieved algebraic special-q in [250000, 330001)
Primes: RFBsize:37706, AFBsize:41659, largePrimes:1015171 encountered
Relations: rels:957167, finalFF:103567
Max relations in full relation-set: 28
Initial matrix: 79431 x 103567 with sparse part having weight 4024611.
Pruned matrix : 64086 x 64547 with weight 1919670.
Total sieving time: 0.47 hours.
Total relation processing time: 0.01 hours.
Matrix solve time: 0.01 hours.
Time per square root: 0.00 hours.
Prototype def-par.txt line would be:
snfs,105,4,0,0,0,0,0,0,0,0,450000,500000,25,25,44,44,2.2,2.2,20000
total time: 0.49 hours.
 --------- CPU info (if available) ----------
[   19.192740] Intel(R) Core(TM)2 CPU          6600  @ 2.40GHz stepping 06
[   18.662922] Memory: 3967336k/4915200k available (2217k kernel code, 151032k reserved, 1162k data, 304k init)
[   18.740386] Calibrating delay using timer specific routine.. 4798.78 BogoMIPS (lpj=9597562)
[   19.192310] Calibrating delay using timer specific routine.. 4795.53 BogoMIPS (lpj=9591075)

jasonp · 2007-09-05, 04:00

Quote:

Originally Posted by masser

See attached - I have included all of our k*5^n+-1 candidates with n<400.

326834*5^146 - 1 =
1285754320076493157639468411457383 (pp34) *
284963915819868309143429201029121716737714765791684584045633868060341827703 (pp75)

(done in 2.5 hours with GGNFS, I picked a factor base size that was too large and needed 2 million relations)

ValerieVonck · 2007-09-05, 04:37

@ltd : sievesubmission?

@fivemack: can you give me good poly for 139606*5^138+1?

Thank you

ltd · 2007-09-05, 05:09

http://sr5.psp-project.de/sieveimport.php

You only have to bring your result to the format like this:

12112762513867651499183 | 24032*5^265+1

Only one factor needs to be reported.

2007-09-05, 04:37	#21
ValerieVonck Mar 2004 Belgium 29² Posts	@ltd : sievesubmission? @fivemack: can you give me good poly for 1396065^138+1? Thank you Last fiddled with by ValerieVonck on 2007-09-05 at 04:38*

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
msieve on KNL	frmky	Msieve	3	2016-11-06 11:45
Msieve on a Mac (Help)	pxp	Msieve	1	2013-02-28 14:56
Using msieve with c	burrobert	Msieve	9	2012-10-26 22:46
msieve help	em99010pepe	Msieve	23	2009-09-27 16:13
Msieve 1.10	RedGolpe	Msieve	6	2006-09-07 12:56

2007-09-04, 16:46	#14
ltd Apr 2003 2²×193 Posts	One other question. Shall we setup an ecmnet server for that? Lars

2007-09-05, 05:09	#22
ltd Apr 2003 2²×193 Posts	http://sr5.psp-project.de/sieveimport.php You only have to bring your result to the format like this: 12112762513867651499183 \| 24032*5^265+1 Only one factor needs to be reported.