mersenneforum.org  

Go Back   mersenneforum.org > Factoring Projects > GMP-ECM

Reply
 
Thread Tools
Old 2014-03-16, 22:32   #1
fivemack
(loop (#_fork))
 
fivemack's Avatar
 
Feb 2006
Cambridge, England

2·29·109 Posts
Default That's odd ... ubuntu-13.10 ecm issue

I'd run about six thousand curves at 3e6 on the C174 cofactor of 5748.1525 to no avail (using GMP-ECM 6.4.2 as distributed with Ubuntu 13.10, on my Avoton box), and thought I ought to do a bit of GPU ECM.

The first job ran (as a screen session) for six hours then stopped and the screen session vanished. Cursing, suspecting some sort of GPU problem (I'd just installed a second GTX580 card), I ran again from a command line in a screen session and it said

Code:
pumpkin@pumpkin:~/ecm-bulk$ ~/cuda-ecm/trunk/ecm -v -gpu -save 5748.1525.1.s1 1e8 1 < 5748.1525.n
GMP-ECM 7.0-dev [configured with GMP 5.1.2, --enable-asm-redc, --enable-gpu, --enable-assert, --enable-openmp] [ECM]
Running on pumpkin
Input number is 240374937214123387734825980441485328571760198890188986168556177043725516179623977715973254104349268626550386607672381746223853794447671062771318068793196116019917029782728411 (174 digits)
Using MODMULN [mulredc:0, sqrredc:2]
Computing batch product (of 144266969 bits) of primes below B1=100000000 took 7117ms
GPU: compiled for a NVIDIA GPU with compute capability 2.0.
GPU: will use device 0: GeForce GTX 580, compute capability 2.0, 16 MPs.
GPU: Selection and initialization of the device took 9ms
Using B1=100000000, B2=1, sigma=3:4227788102-3:4227788613 (512 curves)
dF=0, k=0, d=0, d2=0, i0=0
Expected number of curves to find a factor of n digits:
35      40      45      50      55      60      65      70      75      80
367     1919    11140   71026   492358  3680463 2.9e+07 2.5e+08 2.4e+09 2.7e+10
Computing 512 Step 1 took 848302ms of CPU time / 21208554ms of GPU time
Throughput: 0.024 curves by second (on average 41422.96ms by Step 1)
********** Factor found in step 1: 863037161444255167
Found probable prime factor of 18 digits: 863037161444255167
Probable prime cofactor 278522116952491803716089582283308807148257435762918539361376385659876790193555766353740096713002930457497552258688221497624585776194022548540546514299573733 has 156 digits
So I fear there may be a bug in this particular ubuntu's ecm; the factor is not found with ecm -v -c 10 3e6 on ubuntu-13.10's ecm on my Haswell box either. ubuntu-12.04's ecm finds it immediately.

At least I didn't run GNFS!

Last fiddled with by fivemack on 2014-03-16 at 22:43
fivemack is offline   Reply With Quote
Old 2014-03-16, 23:55   #2
henryzz
Just call me Henry
 
henryzz's Avatar
 
"David"
Sep 2007
Cambridge (GMT/BST)

22·1,433 Posts
Default

Have you tried running exactly the same curve on the problem machines?
henryzz is offline   Reply With Quote
Old 2014-03-19, 17:09   #3
EdH
 
EdH's Avatar
 
"Ed Hall"
Dec 2009
Adirondack Mtns

3,373 Posts
Default

Just to add some more testing, I installed a brand new XUbuntu 13.10 (64-bit) Virtual Machine and grabbed ECM from the repository. I used:
Code:
ecm -v -c 10 3e6 <ecmTestNum
from post #1, and...

nothing!

I then ran 100 curves with identical results:
Code:
xubuntuvm@xubuntuvm-VirtualBox:~/Math$ ecm -v -c 100 3e6 <ecmTestNum
GMP-ECM 6.4.2 [configured with GMP 5.1.2, --enable-asm-redc] [ECM]
Running on xubuntuvm-VirtualBox
Input number is 240374937214123387734825980441485328571760198890188986168556177043725516179623977715973254104349268626550386607672381746223853794447671062771318068793196116019917029782728411 (174 digits)
Using MODMULN [mulredc:0, sqrredc:2]
Using B1=3000000, B2=5706890290, polynomial Dickson(6), sigma=1121040592
dF=16384, k=2, d=158340, d2=11, i0=8
Expected number of curves to find a factor of n digits:
35    40    45    50    55    60    65    70    75    80
324    2351    20272    201449    2247436    2.8e+07    3.9e+08    6e+09    1.1e+11    7.4e+15
Step 1 took 11503ms
Using 21 small primes for NTT
Estimated memory usage: 56M
Initializing tables of differences for F took 11ms
Computing roots of F took 294ms
Building F from its roots took 792ms
Computing 1/F took 459ms
Initializing table of differences for G took 9ms
Computing roots of G took 269ms
Building G from its roots took 747ms
Computing roots of G took 255ms
Building G from its roots took 699ms
Computing G * H took 268ms
Reducing  G * H mod F took 229ms
Computing polyeval(F,G) took 1350ms
Computing product of all F(g_i) took 7ms
Step 2 took 5625ms
Expected time to find a factor of n digits:
35    40    45    50    55    60    65    70    75    80
1.54h    11.18h    4.02d    39.94d    1.22y    15.32y    213.62y    3258y    60442y    4e+09y
Run 2 out of 100:

...

Run 100 out of 100:
Using MODMULN [mulredc:0, sqrredc:2]
Using B1=3000000, B2=5706890290, polynomial Dickson(6), sigma=1568360326
dF=16384, k=2, d=158340, d2=11, i0=8
Expected number of curves to find a factor of n digits:
35    40    45    50    55    60    65    70    75    80
324    2351    20272    201449    2247436    2.8e+07    3.9e+08    6e+09    1.1e+11    7.4e+15
Step 1 took 11951ms
Using 21 small primes for NTT
Estimated memory usage: 56M
Initializing tables of differences for F took 11ms
Computing roots of F took 301ms
Building F from its roots took 733ms
Computing 1/F took 404ms
Initializing table of differences for G took 8ms
Computing roots of G took 255ms
Building G from its roots took 695ms
Computing roots of G took 250ms
Building G from its roots took 697ms
Computing G * H took 239ms
Reducing  G * H mod F took 209ms
Computing polyeval(F,G) took 1355ms
Computing product of all F(g_i) took 8ms
Step 2 took 5393ms
Expected time to find a factor of n digits:
35    40    45    50    55    60    65    70    75    80
1.56h    11.33h    4.07d    40.44d    1.24y    15.52y    216.33y    3300y    61208y    4e+09y
EdH is offline   Reply With Quote
Old 2014-03-19, 17:19   #4
Xyzzy
 
Xyzzy's Avatar
 
"Mike"
Aug 2002

170278 Posts
Default

Quote:
Just to add some more testing, I installed a brand new XUbuntu 13.10 (64-bit) Virtual Machine and grabbed ECM from the repository.
Try:

Code:
echo '240374937214123387734825980441485328571760198890188986168556177043725516179623977715973254104349268626550386607672381746223853794447671062771318068793196116019917029782728411' | ecm -sigma 2945335458 5e4
Xyzzy is offline   Reply With Quote
Old 2014-03-20, 00:34   #5
EdH
 
EdH's Avatar
 
"Ed Hall"
Dec 2009
Adirondack Mtns

3,373 Posts
Default

Code:
xubuntuvm@xubuntuvm-VirtualBox:~/Math$ echo '240374937214123387734825980441485328571760198890188986168556177043725516179623977715973254104349268626550386607672381746223853794447671062771318068793196116019917029782728411' | ecm -sigma 2945335458 5e4
GMP-ECM 6.4.2 [configured with GMP 5.1.2, --enable-asm-redc] [ECM]
Input number is 240374937214123387734825980441485328571760198890188986168556177043725516179623977715973254104349268626550386607672381746223853794447671062771318068793196116019917029782728411 (174 digits)
Using B1=50000, B2=12746592, polynomial x^2, sigma=2945335458
Step 1 took 209ms
Step 2 took 197ms
xubuntuvm@xubuntuvm-VirtualBox:~/Math$
I tried the above about 25 times and the only variance was the time taken.

From my Fedora host machine for the XUbuntu Virtual one:
Code:
$ echo '240374937214123387734825980441485328571760198890188986168556177043725516179623977715973254104349268626550386607672381746223853794447671062771318068793196116019917029782728411' | ecm -sigma 2945335458 5e4
GMP-ECM 6.4.4 [configured with GMP 5.1.1, --enable-asm-redc] [ECM]
Input number is 240374937214123387734825980441485328571760198890188986168556177043725516179623977715973254104349268626550386607672381746223853794447671062771318068793196116019917029782728411 (174 digits)
Using B1=50000, B2=12746592, polynomial x^2, sigma=2945335458
Step 1 took 169ms
Step 2 took 128ms
********** Factor found in step 2: 863037161444255167
Found probable prime factor of 18 digits: 863037161444255167
Probable prime cofactor 278522116952491803716089582283308807148257435762918539361376385659876790193555766353740096713002930457497552258688221497624585776194022548540546514299573733 has 156 digits

Last fiddled with by EdH on 2014-03-20 at 00:39 Reason: To add in the host machine run.
EdH is offline   Reply With Quote
Old 2014-03-20, 01:59   #6
Xyzzy
 
Xyzzy's Avatar
 
"Mike"
Aug 2002

7,703 Posts
Default

Hopefully we filled out the bug form properly!

https://bugs.launchpad.net/ubuntu/+s...m/+bug/1294929

Xyzzy is offline   Reply With Quote
Old 2014-03-20, 02:56   #7
WraithX
 
WraithX's Avatar
 
Mar 2006

23·59 Posts
Default

Quote:
Originally Posted by Xyzzy View Post
Hopefully we filled out the bug form properly!

https://bugs.launchpad.net/ubuntu/+s...m/+bug/1294929

This is funny, I just sent an email to the "maintainers" as indicated on the following Ubuntu page:
http://packages.ubuntu.com/saucy/math/gmp-ecm

I let them know that this problem was discussed and solved on the ecm-discuss list here:
http://lists.gforge.inria.fr/piperma...ly/004234.html

Hopefully, through either your bug report or my email the official Ubuntu gmp-ecm package will be updated to 6.4.4.
WraithX is offline   Reply With Quote
Old 2014-04-01, 02:23   #8
WraithX
 
WraithX's Avatar
 
Mar 2006

23×59 Posts
Default

Well, the reply from the maintainers list was to file a backport request. You can see the response here and here

However, it looks like the easiest way to fill out a backport request is by someone who uses Ubuntu. You can see details about the process here

One helpful quote from the backport page is:
Quote:
The Ubuntu Backporters team recommends using the requestbackport tool in the ubuntu-dev-tools package to file backport requests. This tool is aware of many conventions for backports and can walk you through most of the verification steps.
Perhaps someone here can go through that process to submit the backport request?
WraithX is offline   Reply With Quote
Old 2014-04-08, 07:25   #9
Xyzzy
 
Xyzzy's Avatar
 
"Mike"
Aug 2002

7,703 Posts
Default

Ubuntu 14.04 (LTS) Beta 2:

Code:
$ echo '240374937214123387734825980441485328571760198890188986168556177043725516179623977715973254104349268626550386607672381746223853794447671062771318068793196116019917029782728411' | ecm -sigma 2945335458 5e4
GMP-ECM 6.4.4 [configured with GMP 5.1.3, --enable-asm-redc] [ECM]
Input number is 240374937214123387734825980441485328571760198890188986168556177043725516179623977715973254104349268626550386607672381746223853794447671062771318068793196116019917029782728411 (174 digits)
Using B1=50000, B2=12746592, polynomial x^2, sigma=2945335458
Step 1 took 136ms
Step 2 took 79ms
********** Factor found in step 2: 863037161444255167
Found probable prime factor of 18 digits: 863037161444255167
Probable prime cofactor 278522116952491803716089582283308807148257435762918539361376385659876790193555766353740096713002930457497552258688221497624585776194022548540546514299573733 has 156 digits
Xyzzy is offline   Reply With Quote
Old 2014-04-18, 21:28   #10
Xyzzy
 
Xyzzy's Avatar
 
"Mike"
Aug 2002

7,703 Posts
Default

Ubuntu 14.04 (LTS):

Code:
$ echo '240374937214123387734825980441485328571760198890188986168556177043725516179623977715973254104349268626550386607672381746223853794447671062771318068793196116019917029782728411' | ecm -sigma 2945335458 5e4
GMP-ECM 6.4.4 [configured with GMP 5.1.3, --enable-asm-redc] [ECM]
Input number is 240374937214123387734825980441485328571760198890188986168556177043725516179623977715973254104349268626550386607672381746223853794447671062771318068793196116019917029782728411 (174 digits)
Using B1=50000, B2=12746592, polynomial x^2, sigma=2945335458
Step 1 took 315ms
Step 2 took 191ms
********** Factor found in step 2: 863037161444255167
Found probable prime factor of 18 digits: 863037161444255167
Probable prime cofactor 278522116952491803716089582283308807148257435762918539361376385659876790193555766353740096713002930457497552258688221497624585776194022548540546514299573733 has 156 digits
Xyzzy is offline   Reply With Quote
Old 2014-04-18, 22:01   #11
henryzz
Just call me Henry
 
henryzz's Avatar
 
"David"
Sep 2007
Cambridge (GMT/BST)

22×1,433 Posts
Default

Is that the same pc? If so that is a huge time difference.
henryzz is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
latex issue wildrabbitt Math 2 2017-10-30 12:47
DC performance issue cuBerBruce Information & Answers 2 2017-07-10 13:43
GWNUM issue ET_ Software 4 2017-06-15 09:52
32/64 bit gmp-ecm issue... WraithX GMP-ECM 15 2016-12-19 17:42
C443 issue em99010pepe No Prime Left Behind 86 2008-12-22 22:54

All times are UTC. The time now is 11:29.

Fri Oct 23 11:29:30 UTC 2020 up 43 days, 8:40, 0 users, load averages: 0.86, 1.27, 1.24

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.