mersenneforum.org Running multiple ecm's
 Register FAQ Search Today's Posts Mark Forums Read

 2016-04-04, 11:33 #1 johnadam74   "John A Burroughs" Jan 2016 Warwick, RI 3·5 Posts Running multiple ecm's Hi, I'm new at running ecm and I need some advice. I have a 6 core windows computer and I'm running 10K curves with a B1=110M using a pre-compiled 7.0 version of ecm. My command is ecm -v -chkpnt C20146980.txt -c 10000 110000000. Should I be running 6 different instances of that program to reduce the running time? Any suggestions that you have would be great. Thanks, John
 2016-04-04, 15:45 #2 ATH Einyen     Dec 2003 Denmark 3,037 Posts Yes, each ecm.exe uses only 1 core, so for maximum throughput you need to run 6 versions in parallel. Just make sure you have enough free RAM to handle up to 6 stage 2 at once. Otherwise you can use the "-maxmem" option to limit the memory used in stage 2.
 2016-04-04, 16:47 #3 firejuggler     Apr 2010 Over the rainbow 32·281 Posts i would reccomend ecm.py, found on this forum, that will help automatise it.
 2016-04-04, 18:35 #4 johnadam74   "John A Burroughs" Jan 2016 Warwick, RI 3×5 Posts Thanks. I'm running ecm.py and I've got six instances running. I appreciate all of you help. Thanks again, John
 2016-04-05, 05:23 #5 schickel     "Frank <^>" Dec 2004 CDP Janesville 2·1,061 Posts The other choice would be to install ECMNET with a server and 6 client instances running. I prefer doing it this way, since I can load more than one number in just in case a factor pops out before I expect it.
 2016-04-05, 05:30 #6 VBCurtis     "Curtis" Feb 2005 Riverside, CA 2·3·19·41 Posts A third way is to use the ecmfactor executable in the new GMP-ECM 7.0, which allows -t 6 as a command line flag. Basically, it does what ecm.py does, but the python script has the nice timing data and ETA so it's better.
 2017-06-30, 22:31 #7 jwaltos     Apr 2012 367 Posts source: [Ecm-discuss] ecmpi "Cyril Bouvier has developped a program named ecmpi to run GMP-ECM on several nodes using MPI. It is very easy to use:... You can download it at https://gite.lirmm.fr/bouvier/ecmpi." I've been running several factorizations at 50e9 for values of several hundreds of digits using vanilla commands, has anyone gone beyond 50e9? Last fiddled with by jwaltos on 2017-06-30 at 22:34 Reason: added question
 2017-07-01, 10:34 #8 ATH Einyen     Dec 2003 Denmark 3,037 Posts You did B1=50e9 ? Did you change MAX_HEIGHT above 32 in batch.c ? According to batch.c the maximum B1 allowed is 50685770166 which is the 2MAX_HEIGHT-1 - 1 = 231-1 'th prime. Can you change it above 32 on Linux? I guess on Windows you can't due to GMP bug? Code: #define MAX_HEIGHT 32 #if ECM_UINT_MAX == 4294967295 /* On a 32-bit machine, with no access to a 64-bit type, the maximum value that can be returned by mpz_sizeinbase(s,2) is = (2^32-1). Multiplying all primes up to the following will result in a product that has (2^32-1) bits. */ #define MAX_B1_BATCH 2977044736UL #elif defined(_WIN32) && __GNU_MP_VERSION <= 6 && !defined(__MPIR_VERSION) /* Due to a limitation in GMP on 64-bit Windows, should also affect 32-bit Windows, sufficient memory cannot be allocated for the batch product s when using primes larger than the following */ #define MAX_B1_BATCH 3124253146UL #else /* nth_prime(2^(MAX_HEIGHT-1))-1 */ #define MAX_B1_BATCH 50685770166ULL #endif Last fiddled with by ATH on 2017-07-01 at 10:35
2017-07-01, 16:45   #9
xilman
Bamboozled!

"𒉺𒌌𒇷𒆷𒀭"
May 2003
Down not across

101001010110112 Posts

Quote:
 Originally Posted by jwaltos source: [Ecm-discuss] ecmpi "Cyril Bouvier has developped a program named ecmpi to run GMP-ECM on several nodes using MPI. It is very easy to use:... You can download it at https://gite.lirmm.fr/bouvier/ecmpi." I've been running several factorizations at 50e9 for values of several hundreds of digits using vanilla commands, has anyone gone beyond 50e9?
Currently having problems building it with a GPU-enabled ECM.

I should contact Cyril.

2017-07-08, 05:29   #10
jwaltos

Apr 2012

367 Posts

Quote:
 Originally Posted by ATH You did B1=50e9 ? Did you change MAX_HEIGHT above 32 in batch.c ? According to batch.c the maximum B1 allowed is 50685770166 which is the 2MAX_HEIGHT-1 - 1 = 231-1 'th prime. Can you change it above 32 on Linux? I guess on Windows you can't due to GMP bug?
Yes B1=50e9 on linux using dev 3022.
I`m running these trials on an old Dell 690 with several days for Stages 1 and 2 to complete with approx. 250-300 GB virtual and 64 GB resident ram utilized. I have not experimented with any of the code on other architectures or operating systems.

 2017-07-08, 12:58 #11 wreck     "Bo Chen" Oct 2005 Wuhan,China 2·83 Posts I'm not sure why you use B1=50e9, but , here is some of my experience. For curiocity / test reason, I ran 1 curve at B1=25e9 (t80) on 10,323- c271, which takes about one week. I use -maxmem 12000 option, processor is i7, operating system is Ubuntu 14.04 LTS 64 bit. Currently, the biggest number factored by gnfs is RSA768, a c232. I guess you should finish t75 first, that means using B1=76e8.

 Similar Threads Thread Thread Starter Forum Replies Last Post numbercruncher Information & Answers 18 2014-04-17 00:17 Neimanator PrimeNet 14 2013-08-10 20:15 hc_grove Software 3 2004-10-10 15:34 Helfire Software 8 2004-01-14 00:09 BillW Software 1 2003-01-21 20:11

All times are UTC. The time now is 22:25.

Mon Mar 1 22:25:09 UTC 2021 up 88 days, 18:36, 0 users, load averages: 2.69, 2.63, 2.52