Error generating or reading NFS polynomials

[Brownfox]

I'm trying to do a SNFS factorisation (my first) using factMsieve.pl. All goes well during relation gathering until the time comes to check if we have enough relations, when I get the following:

=>"" spairs.out >> poly_c137_29872v2.dat
Found 4537751 relations, need at least 2868670 to proceed.
=> "../msieve.exe" -s poly_c137_29872v2.dat -l ggnfs.log -i poly_c137_29872v2.ini -v -nf poly_c137_29872v2.fb -t 4 -nc1


Msieve v. 1.53 (SVN unknown)
Tue Apr  3 21:31:52 2018
random seeds: c68698f4 24f6390c
factoring 29872666483521855974439785942686916373703548235035544973444687698159082770843176799660815092997681143268940421145038661347691117468625949 (137 digits)
searching for 15-digit factors
commencing number field sieve (137-digit input)
error generating or reading NFS polynomials
elapsed time 00:00:02

It then goes back to getting more relations.

I'm not sure where msieve is trying to get its polynomials from and the documentation isn't very clear.

Can anyone help? File names and format would be wonderful.
Thanks
Steve

[Dubslow]

When you get that error, immediately pause the program and post the output of ls -lt and cat poly_c137_29872v2.ini poly_c137_29872v2.fb here.

[Brownfox]

I'm working on Windows so I've tried to do the nearest equivalent:

C:\Maths\opn\ggnfs_29872>dir
 Volume in drive C is OS
 Volume Serial Number is D4DD-63BD

 Directory of C:\Maths\opn\ggnfs_29872

04/04/2018  07:59    <DIR>          .
04/04/2018  07:59    <DIR>          ..
02/04/2018  17:09                50 .params
04/04/2018  07:59            31,877 ggnfs.log
03/04/2018  20:50               420 msieve.job
03/04/2018  20:07       454,781,541 poly_c137_29872v2 - Copy.dat
04/04/2018  07:54       623,626,892 poly_c137_29872v2.dat
03/04/2018  20:08               137 poly_c137_29872v2.ini
04/04/2018  07:59               420 poly_c137_29872v2.job.T1
02/04/2018  17:17         4,011,760 poly_c137_29872v2.job.T1.afb.0
04/04/2018  07:59               420 poly_c137_29872v2.job.T2
02/04/2018  17:17         4,011,760 poly_c137_29872v2.job.T2.afb.0
04/04/2018  07:59               420 poly_c137_29872v2.job.T3
02/04/2018  17:17         4,011,760 poly_c137_29872v2.job.T3.afb.0
04/04/2018  07:59               420 poly_c137_29872v2.job.T4
02/04/2018  17:17         4,011,760 poly_c137_29872v2.job.T4.afb.0
02/04/2018  15:35               140 poly_c137_29872v2.n
02/04/2018  17:07                 0 poly_c137_29872v2.polsel.51.m.all
02/04/2018  17:07                 0 poly_c137_29872v2.polsel.cand.all
02/04/2018  17:05               139 poly_c137_29872v2.polsel.DESKTOP-9SG7ESK.13348.data
02/04/2018  17:08               616 poly_c137_29872v2.polsel.DESKTOP-9SG7ESK.15812.51.m
02/04/2018  17:08                 0 poly_c137_29872v2.polsel.DESKTOP-9SG7ESK.15812.cand
02/04/2018  17:07               139 poly_c137_29872v2.polsel.DESKTOP-9SG7ESK.15812.data
02/04/2018  17:08             1,107 poly_c137_29872v2.polsel.DESKTOP-9SG7ESK.15812.log
02/04/2018  17:06               139 poly_c137_29872v2.polsel.DESKTOP-9SG7ESK.16376.data
02/04/2018  17:06                 0 poly_c137_29872v2.polsel.DESKTOP-9SG7ESK.16376.log
03/04/2018  19:20               384 poly_c137_29872v2.poly
04/04/2018  07:59           876,702 spairs.add
              26 File(s)  1,095,369,003 bytes
               2 Dir(s)  933,730,783,232 bytes free

C:\Maths\opn\ggnfs_29872>

poly_c137_29872v2.ini:

29872666483521855974439785942686916373703548235035544973444687698159082770843176799660815092997681143268940421145038661347691117468625949

There is no poly_c137_29872v2.fb. My polynomial file poly_c137_29872v2.poly contains:

type: snfs
n: 29872666483521855974439785942686916373703548235035544973444687698159082770843176799660815092997681143268940421145038661347691117468625949
c6: 1
c5: 1
c4: 1
c3: 1
c2: 1
c1: 1
c0: 1
Y1: 1
Y0: -10604081288675176205434139407
skew: 1
rlim: 7400000
alim: 7400000
lpbr: 27
lpba: 27
mfbr: 48
mfba: 48
rlambda: 2.6
alambda: 2.6
q0: 7400001
qintsize: 100000

Thanks for your help
Steve

[Dubslow]

Originally Posted by Brownfox

There is no poly_c137_29872v2.fb.

That sounds like a problem, considering that factmsieve.pl seems to think it exists:

Found 4537751 relations, need at least 2868670 to proceed.
"../msieve.exe" -s poly_c137_29872v2.dat -l ggnfs.log -i poly_c137_29872v2.ini -v -nf poly_c137_29872v2.fb -t 4 -nc1

(Disclaimer: I know ~nothing about factmsieve.pl, but hopefully this can at least point you in the right direction)

Also please use the [code] tags in the future, use the "Go Advanced" editor if you don't want to type it out

[Brownfox]

Thanks for your help. I had thought that this step produced the factor base. I will investigate further.

Sorry about the code tags. I did try to format it but I'm bot used to this forum.

Regards
Steve

[chris2be8]

poly_c137_29872v2.fb should be automatically generated by factMsieve.pl from the .poly file etc. It would contain something like:

N 29872666483521855974439785942686916373703548235035544973444687698159082770843176799660815092997681143268940421145038661347691117468625949
SKEW 1
A6 1
A5 1
A4 1
A3 1
A2 1
A1 1
A0 1
R1 1
R0 -10604081288675176205434139407
FAMAX 7400000
FRMAX 7400000
SALPMAX 134217728
SRLPMAX 134217728

I've spotted one thing that will cause problems, factMsieve.pl will see the poly_c137_29872v2.n file and assume it needs to do polynomial selection, instead of using the poly_c137_29872v2.poly you provided. Try renaming poly_c137_29872v2.n out of the way and re-running.

Chris

[Brownfox]

I needed to restart from scratch and just have a folder with the poly file in. I assumed when the documentation said that the script would look for a poly file first that it would not matter if the .n file was there too. A classic example of nature inventing a better idiot!

All working now, thanks for your help Chris and Dubslow.

[chris2be8]

You can get round that issue by using the factMsieve.pl from http://mersenneforum.org/showthread.php?t=23167&page=2 post 18 (see also posts 21 and 24). It will use msieve for poly selection if necessary, which generates much better polys than GGNFS does.

Chris