Hi,
I have not recently used valgrind with mpi, but goooggggling hits
http://www.open-mpi.org/faq/?category=debugging#valgrind_clean
I'm planning to add a valgrind entry to the wiki.
Then we can stash command lines for various 3rd party softwares.
scott
On Wed, Apr 25, 2012 at 11:45:40AM -0400, Jason Swails wrote:
> This is great! Are there a set of rules for any MPIs out there? Those are
> even more annoying, IMO...
>
> On Tue, Apr 24, 2012 at 1:58 AM, Scott Brozell <sbrozell.rci.rutgers.edu>wrote:
> > Valgrind is a great tool for software validation.
> > Unfortunately, Fortran I/O statements cause a bunch of false positive
> > detections of memory management bugs.
> >
> > tools/fortran.supp is a file of Valgrind suppression rules for Fortran.
> > http://valgrind.org/docs/manual/manual-core.html#manual-core.suppress
> >
> > These were generated on x86_64 ifort version 12.1.0
> > for bug 179 in test AmberTools/test/rism1d/spc-kh via
> > ~/bin/valgrind/3.7.0/bin/valgrind -v
> > --suppressions=/amber/tools/fortran.supp --gen-suppressions=all
> > ../../../bin/rism1d spc
> >
> > For that test these rules suppress 670 false positives.
actually 664 due to fortran i/o.
scott
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Wed Apr 25 2012 - 14:00:04 PDT
