3

This is required information for some MPI using app that I am working with...

From its Makefile template:

#insert memory per core and uncomment the following line
#PREPROC= -D'MB_PER_CORE=750'

Note that said scientific app runs also on NUMA machines, like Cray, where each core has its own memory. I am asking here what to put in above line on Linux on e.g. 12GB machine with 16 cores.

dhag
  • 15,440
  • 4
  • 54
  • 65
Jakub Narębski
  • 1,228
  • 2
  • 17
  • 30
  • 2
    I don't understand the question. On most architectures, the RAM is shared between all cores. Processes aren't associated with particular cores either. What information are you after? – Gilles 'SO- stop being evil' May 27 '11 at 21:37
  • I have added fragment of Makefile template of said app that talks about "**memory per core**". Your guess is as good as mine in what it mean; I think it might be total memory / number of cores (CPUs). – Jakub Narębski May 28 '11 at 06:13
  • 4
    It depends on what the program does with that parameter. Total memory / number of cores would be a good starting point, but if it actually tries to allocate that much memory per core, you'd probably want to use a smaller value. On the other hand, maybe that parameter is only for use on NUMA machines. You haven't given us enough information for a definitive answer. – cjm May 28 '11 at 08:00
  • Is this scientific app open source? If so, what is its homepage? – Philip Durbin Oct 23 '11 at 00:57
  • Voting to close as there is not enough information provided here. – Stéphane Gimenez Feb 15 '12 at 18:49
  • @PhilipDurbin: It's GENE (Gyrokinetic Electromagnetic Numerical Experiment): http://www.ipp.mpg.de/~fsj/gene/ – Jakub Narębski Mar 09 '12 at 19:47

0 Answers0