[petsc-users] question about Create/Duplicate functions

[Matthew Knepley]

On Mon, Dec 26, 2011 at 8:34 AM, Dominik Szczerba <dominik at itis.ethz.ch>wrote:

> On Sun, Dec 25, 2011 at 5:19 PM, Jed Brown <jedbrown at mcs.anl.gov> wrote:
> > On Sun, Dec 25, 2011 at 10:17, Dominik Szczerba <dominik at itis.ethz.ch>
> > wrote:
> >>
> >> If I call a Create (or Duplicate) function on a non-Destroyed object,
> >> will be it destroyed automatically before the new allocation, or is a
> >> memory leak expected?
> >
> >
> > What do you mean? Show the function call you are talking about?
>
> As simple as:
>
> Vec x
> VecCreate...
> VecCreate...
> exit application
>
> I would expect a leak to occur, as Matt confirmed, but was surprised
> that valgrind was not reporting any issues with this code, and that's
> why I asked my question.
> With valgrind option --leak-check=full (option not listed in FAQ) lost
> memory is reported in the first call to VecCreate, but no issue is
> reported due to the lack of VecDestroy. Does Petsc finalize function
> do some type of garbage collection?
>

No, you will see it using -malloc_dump. Valgrind should show you
overwriting the
pointer.

   Matt


> Thanks
> Dominik
>



-- 
What most experimenters take for granted before they begin their
experiments is infinitely more interesting than any results to which their
experiments lead.
-- Norbert Wiener
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20111226/9e81f166/attachment.htm>