[petsc-users] memory crash with optimized petsc

[Fande Kong]

When I used valgrind to check memory, I got the following output:

==3055== Conditional jump or move depends on uninitialised value(s)
==3055==    at 0x6CF359E: ??? (in
/usr/lib/x86_64-linux-gnu/libgfortran.so.3.0.0)
==3055==    by 0x6CF2B25: ??? (in
/usr/lib/x86_64-linux-gnu/libgfortran.so.3.0.0)
==3055==    by 0x6C31607: ??? (in
/usr/lib/x86_64-linux-gnu/libgfortran.so.3.0.0)
==3055==    by 0x400F305: call_init.part.0 (dl-init.c:85)
==3055==    by 0x400F3DE: _dl_init (dl-init.c:52)
==3055==    by 0x40016E9: ??? (in /lib/x86_64-linux-gnu/ld-2.15.so)
==3055==    by 0x47: ???
==3055==    by 0x7FEFFFE56: ???
==3055==    by 0x7FEFFFE69: ???
==3055==    by 0x7FEFFFE71: ???
==3055==    by 0x7FEFFFE75: ???
==3055==    by 0x7FEFFFE88: ???
==3055==    by 0x7FEFFFE8B: ???
==3055==    by 0x7FEFFFE95: ???
==3055==    by 0x7FEFFFE9C: ???
==3055==    by 0x7FEFFFEA5: ???
==3055==    by 0x7FEFFFEA9: ???
==3055==    by 0x7FEFFFEB8: ???
==3055==    by 0x7FEFFFEBA: ???
==3055==    by 0x7FEFFFECC: ???
==3055==
==3055== Conditional jump or move depends on uninitialised value(s)
==3055==    at 0x6CF35A9: ??? (in
/usr/lib/x86_64-linux-gnu/libgfortran.so.3.0.0)
==3055==    by 0x6CF2B25: ??? (in
/usr/lib/x86_64-linux-gnu/libgfortran.so.3.0.0)
==3055==    by 0x6C31607: ??? (in
/usr/lib/x86_64-linux-gnu/libgfortran.so.3.0.0)
==3055==    by 0x400F305: call_init.part.0 (dl-init.c:85)
==3055==    by 0x400F3DE: _dl_init (dl-init.c:52)
==3055==    by 0x40016E9: ??? (in /lib/x86_64-linux-gnu/ld-2.15.so)
==3055==    by 0x47: ???
==3055==    by 0x7FEFFFE56: ???
==3055==    by 0x7FEFFFE69: ???
==3055==    by 0x7FEFFFE71: ???
==3055==    by 0x7FEFFFE75: ???
==3055==    by 0x7FEFFFE88: ???
==3055==    by 0x7FEFFFE8B: ???
==3055==    by 0x7FEFFFE95: ???
==3055==    by 0x7FEFFFE9C: ???
==3055==    by 0x7FEFFFEA5: ???
==3055==    by 0x7FEFFFEA9: ???
==3055==    by 0x7FEFFFEB8: ???
==3055==    by 0x7FEFFFEBA: ???
==3055==    by 0x7FEFFFECC: ???
==3055==

How to understand the output?


On Mon, Jun 3, 2013 at 9:00 AM, Jed Brown <jedbrown at mcs.anl.gov> wrote:

> Fande Kong <fande.kong at colorado.edu> writes:
>
> > This was not complete. The complete result is attached.
>
> With petsc-3.4, you can use PetscFunctionBeginUser and -objects_dump to
> eliminate internal allocations from the output.  I'm guessing you missed
> a KSPDestroy.
>
>


-- 
Fande Kong
Department of Computer Science
University of Colorado at Boulder
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20130604/530df756/attachment.html>