mersenneforum.org Factoring projects algorithms - What is Aliqueit ?
 Register FAQ Search Today's Posts Mark Forums Read

2016-08-06, 09:36   #12
Romuald

Oct 2015
France

32·7 Posts

Quote:
 Originally Posted by xilman Be careful. YAFU does not play well with GPU-enabled ECM. At least, that's my experience.
What was your software & hardware setup ?

 2016-08-06, 09:43 #13 Dubslow Basketry That Evening!     "Bunslow the Bold" Jun 2011 40
2016-08-06, 11:48   #15
xilman
Bamboozled!

"𒉺𒌌𒇷𒆷𒀭"
May 2003
Down not across

43·271 Posts

Quote:
 Originally Posted by Romuald What was your software & hardware setup ?
Gentoo Linux, so essentially everything built from source. The only exception is the CUDA set-up which Nvidia distributes only as a binary package.

In particular, msieve, ECM-GMP and ggnfs are all built from recent repository code.

See http://www.mersenneforum.org/showthr...667#post437667 for my discovery that YAFU and GPU-enabled ECM do not play well together and Ben's response. I don't know if Ben has fixed it yet; I'm sure he'll let us know if he has.

Paul

Last fiddled with by xilman on 2016-08-06 at 11:50 Reason: Fix clumsy phrasing

 2016-08-06, 11:49 #16 Dubslow Basketry That Evening!     "Bunslow the Bold" Jun 2011 40
2016-08-06, 11:55   #17
xilman
Bamboozled!

"𒉺𒌌𒇷𒆷𒀭"
May 2003
Down not across

43×271 Posts

Quote:
 Originally Posted by Romuald Just hearing from you that factmsieve and all its paraphernalia is obselete.
Some of us find it very far from obselete[sic]. We use it every day for non-trivial factorizations. I'm doing a c168 right now, for instance, under the control of factMsieve.pl

2016-08-06, 21:20   #18
Romuald

Oct 2015
France

32·7 Posts

Quote:
 Originally Posted by Dubslow The setup is harder than the usage.
That is precisely what I mean. By making a full-automated program, why do not include all it needs from beggining?
That question doesn't expect any response.

OK so I'm trying to do all of this:

-GMP and MSIEVE for compiling
-GMP-ECM and GGNFS for executable

But I idn't see any topic giving binaries of GGNFS.
I search tomorrow and in days ahead.

Xilman, what instructions did you follow (if you did) for your factmsieve.pl?

Last fiddled with by Romuald on 2016-08-06 at 21:22 Reason: adding sth

 2016-08-07, 02:09 #19 Dubslow Basketry That Evening!     "Bunslow the Bold" Jun 2011 40
2016-08-07, 06:26   #20
xilman
Bamboozled!

"𒉺𒌌𒇷𒆷𒀭"
May 2003
Down not across

43·271 Posts

Quote:
 Originally Posted by Romuald Xilman, what instructions did you follow (if you did) for your factmsieve.pl?
Use the source Luke.

In the code it's clear how to set up the number of threads, the locations of the sieiving programs, and stuff like that.

Standard usage is trivial. For tidiness sake I tend to run factorizations in a subdirectory of that which holds the msieve code. Put the integer to be factored in a file --- for the sake of example call it foo.n --- and then run ../factMsieve.pl foo.n. It churns for a while as it tries to find a gnfs polynomial. Of course, if you want to run SNFS this is pretty pointless but if you are a complete newbie it does show you the format of a polynomial file, which format you can use for the SNFS poynomial. If you let polynomial finding run to completion factMsieve will continue automagically. If you need to restart factMsieve.pl for whatever reason, including wanting to use a polynomial file found by alternative means, you run ../factMsieve.pl foo.poly

Anything non-trivial --- if I wanted to run in parallel on several machines for instance, or to tweak some of the sieving parameters --- again I'd read the source. It's very rare that I feel the need to do this.

Paul

 2016-08-07, 07:16 #21 Romuald     Oct 2015 France 32×7 Posts ... I see. The main difference between your installation and potentially mine is that I would use the python script (i have the bases of python but i am ignorant of PL/I). Actually, I have a functional implementation with factmsieve.py, ggnfs, msieve, just following the thread "Build Instructions for Ubuntu 14.04", but whan I want to factorize, doesn't seem to work: Code: ./factmsieve.py example -> +--------------------------------------------------------------+ -> | Running factmsieve.py, a Python driver for MSIEVE with GGNFS | -> | sieving support. It is Copyright, 2010-2013, Brian Gladman | -> | and is a conversion of factmsieve.pl that is Copyright, 2004 | -> | Chris Monico. This is version 0.77, 27th August 2013. | -> +--------------------------------------------------------------+ -> This is client 1 of 1 -> Running on 6 Cores with 2 hyper-threads per Core -> Working with NAME = example -> Selected default factorization parameters for 100 digit level. -> Selected lattice siever: gnfs-lasieve4I12e -> No parameter change detected, resuming... -> Running setup ... -> Estimated minimum relations needed: 4.095e+06 -> resuming a block for q from 1000000 to 1100000 -> Running lattice siever ... -> entering sieving loop -> making sieve job for q = 1002719 in 1000000 .. 1008333 as file example.job.T0 -> making sieve job for q = 1011359 in 1008333 .. 1016666 as file example.job.T1 -> making sieve job for q = 1019801 in 1016666 .. 1024999 as file example.job.T2 -> making sieve job for q = 1028089 in 1024999 .. 1033332 as file example.job.T3 -> making sieve job for q = 1036531 in 1033332 .. 1041665 as file example.job.T4 -> making sieve job for q = 1044941 in 1041665 .. 1049998 as file example.job.T5 -> making sieve job for q = 1053467 in 1049998 .. 1058331 as file example.job.T6 -> making sieve job for q = 1061171 in 1058331 .. 1066664 as file example.job.T7 -> making sieve job for q = 1069349 in 1066664 .. 1074997 as file example.job.T8 -> making sieve job for q = 1077799 in 1074997 .. 1083330 as file example.job.T9 -> making sieve job for q = 1086331 in 1083330 .. 1091663 as file example.job.T10 -> making sieve job for q = 1094633 in 1091663 .. 1099996 as file example.job.T11 -> Lattice sieving algebraic q from 1000000 to 1100000. -> gnfs-lasieve4I12e -k -o example.out.T0 -v -n0 -a example.job.T0 Traceback (most recent call last): File "./factmsieve.py", line 2017, in run_siever(client_id, num_clients, SV_THREADS, fact_p, lats_p) File "./factmsieve.py", line 1648, in run_siever procs.append(run_exe(lats_p['siever'], args, wait = False)) File "./factmsieve.py", line 350, in run_exe p = subprocess.Popen([ex] + args.split(' '), **al) File "/home/matthias/anaconda3/lib/python3.5/subprocess.py", line 947, in __init__ restore_signals, start_new_session) File "/home/matthias/anaconda3/lib/python3.5/subprocess.py", line 1552, in _execute_child raise child_exception_type(err_msg) subprocess.SubprocessError: Exception occurred in preexec_fn. I'm not specialist... But As far as I can understand, it seems like there's an error with python... and anaconda. I don't see why anaconda is causing trouble, I installed it recently to have maths libraries (sympy...). Last fiddled with by Romuald on 2016-08-07 at 08:03
 2016-08-07, 07:54 #22 Romuald     Oct 2015 France 32×7 Posts Right now, all is ready for YAFU, ecept, of course, ggnfs. So, I've been looking for binaries on the forum, nothing. I try to compile it myself: 'make x86_64' It returns an error occured because of the 'x86_64'. Tried with 'nocona', failed. I realize this is particularly annoying to have this stubborn tool unobtainable, by compiling or downloading ready-to-serve binaries, whereas all other tools GMP, GMP-ECM, Msieve, are OK. Last fiddled with by Romuald on 2016-08-07 at 08:03

 Similar Threads Thread Thread Starter Forum Replies Last Post rogue Factoring 20 2014-11-19 01:08 Raman Hobbies 45 2009-05-11 05:11 ThiloHarich Factoring 0 2007-09-02 20:32 koders333 Factoring 14 2006-01-25 14:08 koders333 Factoring 1 2006-01-19 20:04

All times are UTC. The time now is 21:17.

Mon Jan 30 21:17:54 UTC 2023 up 165 days, 18:46, 0 users, load averages: 0.81, 0.96, 1.02