0

I am trying to perform a heavy scientific computation on my ubuntu machine. The code is written in Python and is marked as an executable. However, after about one hour the computation is terminated with no error message other than "terminated" as if the system just ran the command "killall python".

Why does ubuntu terminate my computations? There is no infinite loop and the solution converges nicely (but there is one part of the program that needs some optimization).

Thanks!

bmu
  • 35,119
  • 13
  • 91
  • 108
numpy problem
  • 73
  • 1
  • 7

1 Answers1

6

Just from your question as stated it is very hard to pinpoint what the problem might be. If I were you, I'd take a look at the process' memory footprint. What may be happening is that there is some sort of memory leak that grows so huge that eventually the operating system has to kill the process

jmbr
  • 3,298
  • 23
  • 23
  • Thank you for your answer. First of all, I would like to apologize for my late reply. I have been looking deeper into this problem and I have made sure that the finite difference code I used when discovering this problem is not the problem. I have discovered another situation where the same problem occurs. This problem occurs when running [code](convert -density 100 thesis.pdf pages.png)[/code] where thesis.png is the following large pdf file: [link](http://dspace.mit.edu/bitstream/handle/1721.1/60145/680744226.pdf?sequence=1) – numpy problem Jul 22 '12 at 15:15
  • It appears that running a computation heavy enough (with no memory leaks or any problems of that sort) will cause ubuntu to terminate it without giving an error message. I am using ImageMagick 6.6.9-7 on Linux version 3.2.0-26-generic (buildd@batsu) (gcc version 4.6.3 (Ubuntu/Linaro 4.6.3-1ubuntu5) ) If you know how to avoid this problem, please let me know. If not, then I will send a bug report. Thank you – numpy problem Jul 22 '12 at 15:15
  • I'm pretty sure this is a bug. I regularly launch very long computations on Ubuntu and Scientific Linux without any of the problems you report. – jmbr Jul 22 '12 at 17:57