[MITgcm-support] Chasing a seg fault

[Patrick Heimbach]

Jody,

could you check whether that machine limits stacksize?
I.e. do
    limit
to get info and
    limit stacksize unlimited
to unlimit the stacksize, and try running again.

-Patrick



On Aug 10, 2007, at 3:31 PM, Jody Klymak wrote:

>
> Hello,
>
> I'm trying to get mitgcm running on a cluster of AIX machines at  
> westgrid (IBM AIX) http://www.westgrid.ca/support/programming/ 
> robson.php
>
> I can compile fine if I use the -ignoretime option for genmake2.
>
> I get the segmentation fault after the zeroth mds write has  
> completed, and after the MONITOR output has been written.  gdb  
> indicates that the problem is in the first call to DO_OCEANIC_PHYS,  
> and says:
>
> Program received signal SIGSEGV, Segmentation fault.
> [Switching to Thread 549765433280 (LWP 966)]
> 0x00000000100fd6f0 in do_oceanic_phys (mytime=Cannot access memory  
> at address 0x200026c6070
> ) at do_oceanic_phys.f:424
> 424           SUBROUTINE DO_OCEANIC_PHYS(myTime, myIter, myThid)
>
> My build options file is below....
>
> The exact model setup runs fine on an OSX machine, but under g77,  
> if that makes a difference.
>
> Thanks for any pointers on what I may have done wrong.
>
> Cheers,  Jody
>
> >>> robson<<<
>
> #!/bin/bash
> #
> #  $Header: /u/gcmpack/MITgcm/tools/build_options/sp4+mpi,v 1.3  
> 2005/12/22 01:38:05 ce107 Exp $
> #
>
> #  tested on bluesky, ce107 Dec 2005
> #  using the following invocation:
> #    ../../../tools/genmake2 -mods=../code -of=../../../tools/ 
> build_options/sp4 -make=gmake
>
> S64='$(TOOLSDIR)/set64bitConst.sh'
> MAKEDEPEND=makedepend
> DEFINES='-DWORDLENGTH=4'
>
> INCLUDES=''
> CPP='/usr/bin/cpp -traditional -P'
> CC='xlc'
> FC='xlf95'
> LINK='xlf95'
> LIBS=''
> FFLAGS='-qfixed=132 -q64 -g'
>
> FOPTIM='-q64  -Q -qarch=auto'
> CFLAGS='-q64  -Q -qarch=auto -g'
>
> FC_NAMEMANGLE="#define FC_NAMEMANGLE(X) X ## _"
>
>
>
>
> --
> Jody Klymak
> http://web.uvic.ca/~jklymak/
>
>
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support

---
Patrick Heimbach | heimbach at mit.edu | http://www.mit.edu/~heimbach
MIT | EAPS 54-1518 | 77 Massachusetts Ave | Cambridge MA 02139 USA
FON +1-617-253-5259 | FAX +1-617-253-4464 | SKYPE patrick.heimbach