Re: segmentation fault

From: Dennis Shea (shea AT XXXXXX)
Date: Mon Oct 27 2003 - 16:19:57 MST

  • Next message: Constantine Nenkov: "plotting in polar coordinates"

    Or 'more better'

    limit datasize unlimited
    limit stacksize unlimited

    Many of us do this even on NCAR's big data processing machines
    which have 32GB of memory.

    We usually put the above in our .cshrc files. Not sure what the equivalent
    is for MAC X OS.

    Dennis Shea

    >> I am getting a 'segmentation fault' error when running an ncl script on
    >> Mac OS X. The script worked correctly when run on dataproc. The script
    >> reads in an array SST(0,:,:) which is a 180x360 array, regrids this to a
    >> T42 array 'newgrid' using the example on the web pages.
    >> Looking through the ncl-talk previously posted questions and answers, I
    >> guess this error results from a memory allocation problem. What can I do
    >> to get past this error?
    >> An example of the script is attached.
    >> Thank you for your help.
    >I get the same thing for all but the most trivially small programs.
    >The thing to check is your limits:
    >% limit
    >cputime unlimited
    >filesize unlimited
    >datasize 6144 kbytes
    >stacksize 512 kbytes
    >coredumpsize 0 kbytes
    >memoryuse unlimited
    >descriptors 256
    >memorylocked unlimited
    >maxproc 100
    >If yours are the same, I'm sure you are running out of stack. You can
    >increase it:
    >% limit stacksize 8192
    >% unlimit stacksize
    >% unlimit datasize
    >Kate Hedström Arctic Region Supercomputing Center
    >kate AT University of Alaska, Fairbanks
    >ncl-talk mailing list
    >ncl-talk AT

    ncl-talk mailing list
    ncl-talk AT

    This archive was generated by hypermail 2b29 : Mon Oct 27 2003 - 16:16:07 MST