[petsc-users] Unexplained memory leaks

[Matthew Knepley]

On Fri, Aug 28, 2020 at 1:32 PM baikadi pranay <pranayreddy865 at gmail.com>
wrote:

> Thank you Mark and Matthew. I will try to use the suppression file to hide
> the leaks.
>
> On a different note, I see a similar error related to MPI when I couple my
> poisson solver with the schrodinger solver (written with SLEPc). I am
> attaching a screenshot of the error. I was wondering if you could comment
> on this issue as well.
>
> Please let me know if you need any further information from me.
>

It looks like you called an MPI function after you called PetscFinalize().
In order to do this, you have to call MPI_Init() before you call
PetscInitialize().

  Thanks,

     Matt


> Thank you.
>
> Best Regards,
> Pranay.
> ᐧ
>
> On Fri, Aug 28, 2020 at 5:48 AM Mark Lohry <mlohry at gmail.com> wrote:
>
>> Looking through my own valgrind I saw OpenMPI mentions on their faq they
>> provide a suppression file to hide their known leaks:
>> https://www.open-mpi.org/faq/?category=debugging#valgrind_clean
>>
>> mpirun -np 2 valgrind --suppressions=$PREFIX/share/openmpi/openmpi-valgrind.supp
>>
>>
>>
>> On Fri, Aug 28, 2020 at 7:41 AM Matthew Knepley <knepley at gmail.com>
>> wrote:
>>
>>> OpenMPI leaks memory at the end. As you can see it is small, so they are
>>> unmotivated to fix that.
>>>
>>>   Thanks,
>>>
>>>      Matt
>>>
>>> On Fri, Aug 28, 2020 at 4:31 AM baikadi pranay <pranayreddy865 at gmail.com>
>>> wrote:
>>>
>>>> Hi,
>>>> I am building a 2D solver for the semiconductor Poisson-Boltzmann
>>>> equation. I detected a memory leak when running the program using valgrind
>>>> but I am unable to solve this issue as there are no signs in the valgrind
>>>> output indicating that the source of the error is in the modules I have
>>>> written. I am attaching you a text file containing the valgrind output.
>>>>
>>>> I have seen that a similar question was asked earlier (found here
>>>> <https://lists.mcs.anl.gov/pipermail/petsc-users/2010-October/007132.html>)
>>>> but I could not find a final solution to that problem. Could you let me
>>>> know the source of the problem?
>>>>
>>>> Please let me know if you need any further information.
>>>>
>>>> Thank you,
>>>> Pranay.
>>>> ᐧ
>>>>
>>>
>>>
>>> --
>>> 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
>>>
>>> https://www.cse.buffalo.edu/~knepley/
>>> <http://www.cse.buffalo.edu/~knepley/>
>>>
>>

-- 
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

https://www.cse.buffalo.edu/~knepley/ <http://www.cse.buffalo.edu/~knepley/>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20200828/41366502/attachment.html>