[Wien] Segmentation Fault in lapw0

L. D. Marks L-marks at northwestern.edu
Tue Jul 6 21:20:46 CEST 2004


I noticed some strange behavior with lapw0 on large systems with ifort &
Redhat (not -static). One thing you have to be careful about it
libguide.o. At least with mkl6.1 the version in ifort is different from
the one in mkl, and incompatible. They may have fixed this with mkl7.0, I
don't know. You may want to set LD_LIBRARY_PATH carefully (in .bashrc as
well as ulimit); I use

LD_LIBRARY_PATH=/opt/intel/mkl61/lib/32:/opt/intel_fc_80/lib:/opt/intel_cc_80/lib
export LD_LIBRARY_PATH

Remember that the order matters!

On Tue, 6 Jul 2004, Sam Leitch wrote:

> Thanks for the suggestions, but those don't appear to be the problem.
>
> I used the commandline interface of w2web to check the stack size. Both
> 'ulimit' and 'ulimit -s' return unlimited every time I try it.
> I tried run_lapw from a standard bash shell with 'ulimit -s unlimited'
> and it resulted in a segmentation fault as well.
>
> I had compiled lapw0 -O0 before, but I tried recompiling everything to
> be sure. Same problem.
>
> I'm pretty sure it's a compiler bug with large arrays. There's some
> posts on the intel forums of people who've had the same problem. They
> saw "Segmentation Fault" errors, not "Out of Memory" errors when
> creating large arrays, which looks more like a compiler error than a
> coding error.
>
> I'm going to change paths and fight to get IFC 7.1 working with libc 2.3
> instead.
>
> I'm still open for suggestion though.
>
> Torsten Andersen wrote:
>
> > Dear Sam,
> >
> > the problem is probably with your shell settings.
> >
> > in bash, you need to put (in the file where you defined WIENROOT -
> > .profile or .bash_profile, probably): ulimit -s unlimited
> >
> > in tcsh, you need to put (in .cshrc or .tcshrc - wherever WIENROOT is
> > set): limit stacksize unlimited
> >
> > If this doesn't help (log out and log in again for the settings to
> > take proper effect), you can try to recompile with "-O0"
> > (minus-the.letter.O-zero).
> >
> > Best regards,
> > Torsten Andersen.
>
>
>
> --
> Sam Leitch
> * * * * * * * * * * * * * * *
> Graduate Student
> Department of Physics and Engineering Physics
> University of Saskatchewan
> * * * * * * * * * * * * * * *
>
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>

-----------------------------------------------
Laurence Marks
Department of Materials Science and Engineering
MSE Rm 2036 Cook Hall
2225 N Campus Drive
Northwestern University
Evanston, IL 60201, USA
Tel: (847) 491-3996 Fax: (847) 491-7820
mailto:L-marks at northwestern.edu
http://www.numis.northwestern.edu
-----------------------------------------------




More information about the Wien mailing list