Strange bug with GMP-ECM

MatWur-S530113 · 2007-11-18, 21:36

hello,

until yesterday I run GMP-ECM. I installed it step-by-step with MinGW/MSys some weeks ago. I used GMP-ECM v 6.1.2 and GMP v 4.1.4 . Everything worked well, I was able to write batch-files (I'm using WinXP and an Intel E6600 DualCore processor) like this:

ecm -c 300 5e4 < input.file >> output.file

The batch file run 300 curves, results are in output.file and if the cofactor is probable prime ECM stopped. Everything ok and good.

Today I installed GMP-ECM new, again using MinGW/MSys. But I used GMP-ECM v 6.1.3 and GMP v 4.2.2 . I used the same Msys as before and I got a new ecm.exe in a new subdirectory. I started to factor with this new version and some hours everything was ok. But now something strange happened: I run the batch file above, but only 1 curve run, then ECM stops. There is no factor, the input number is composite, I don't see a reason that ECM stops. It should run 300 curves. But it stops after 1 curve.
OK I thought, maybe a little bug in the new version and I went back to the old version. But now there is the same. I only can run 1 curve with the switch -c , then ECM stops. I tried another number to factor, same problem with both versions.

The fault was *not* soon after installing the new version, some hours (about 3) everything was ok. I made a virus check and a torture test with prime95 (2 hours), no problems with the hardware detected.

I don't know where the problem could be. Maybe I should use a new instance of MSys in a new directory? But 3 hours everything was ok, incl. the ECM outputs with the versions of GMP and ECM-GMP. The old version gives ECM 6.1.2/GMP 4.1.4 , the new version output is ECM 6.1.3/GMP 4.2.2 as it should be. Any idea to solve this problem?

Thanks in advance.

greetings

Matthias

Andi47 · 2007-11-18, 22:12

Make sure that the input file ends with a "Carriage Return" character - then it should work.

MatWur-S530113 · 2007-11-19, 00:01

Hello again,

thank you Andi, that it was... I must have deleted the LF/CR by mistake. And I copied the input file to the old version, so there was the bug, too...

ok, problem solved, done, thanks again.

greetings

Matthias

2007-11-18, 21:36	#1
MatWur-S530113 Apr 2007 Spessart/Germany 2·3⁴ Posts	Strange bug with GMP-ECM hello, until yesterday I run GMP-ECM. I installed it step-by-step with MinGW/MSys some weeks ago. I used GMP-ECM v 6.1.2 and GMP v 4.1.4 . Everything worked well, I was able to write batch-files (I'm using WinXP and an Intel E6600 DualCore processor) like this: ecm -c 300 5e4 < input.file >> output.file The batch file run 300 curves, results are in output.file and if the cofactor is probable prime ECM stopped. Everything ok and good. Today I installed GMP-ECM new, again using MinGW/MSys. But I used GMP-ECM v 6.1.3 and GMP v 4.2.2 . I used the same Msys as before and I got a new ecm.exe in a new subdirectory. I started to factor with this new version and some hours everything was ok. But now something strange happened: I run the batch file above, but only 1 curve run, then ECM stops. There is no factor, the input number is composite, I don't see a reason that ECM stops. It should run 300 curves. But it stops after 1 curve. OK I thought, maybe a little bug in the new version and I went back to the old version. But now there is the same. I only can run 1 curve with the switch -c , then ECM stops. I tried another number to factor, same problem with both versions. The fault was not soon after installing the new version, some hours (about 3) everything was ok. I made a virus check and a torture test with prime95 (2 hours), no problems with the hardware detected. I don't know where the problem could be. Maybe I should use a new instance of MSys in a new directory? But 3 hours everything was ok, incl. the ECM outputs with the versions of GMP and ECM-GMP. The old version gives ECM 6.1.2/GMP 4.1.4 , the new version output is ECM 6.1.3/GMP 4.2.2 as it should be. Any idea to solve this problem? Thanks in advance. greetings Matthias

2007-11-18, 22:12	#2
Andi47 Oct 2004 Austria 2·17·73 Posts	Make sure that the input file ends with a "Carriage Return" character - then it should work. Last fiddled with by Andi47 on 2007-11-18 at 22:15 Reason: typo

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Something strange ...	bayanne	Software	6	2016-04-06 04:33
Something really strange	schickel	FactorDB	7	2012-02-02 00:10
A strange applet:	3.14159	Miscellaneous Math	7	2010-06-01 01:29
A strange RSA key	siew	Factoring	70	2010-05-31 22:31
Strange bug	HiddenWarrior	Software	5	2005-08-22 08:34

2007-11-19, 00:01	#3
MatWur-S530113 Apr 2007 Spessart/Germany 2·3⁴ Posts	Hello again, thank you Andi, that it was... I must have deleted the LF/CR by mistake. And I copied the input file to the old version, so there was the bug, too... ok, problem solved, done, thanks again. greetings Matthias