<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<meta name="Generator" content="Microsoft Exchange Server">
<!-- converted from text --><style><!-- .EmailQuote { margin-left: 1pt; padding-left: 4pt; border-left: #800000 2px solid; } --></style>
</head>
<body>
<div>
<div>
<div></div>
<div>Hi Tobias. </div>
<div>Thanks for taking a look.. Could you repeat the test with yasp<2> because that was the default in the test within the release version. </div>
<div>Thanks </div>
<div>Andreas </div>
<div><br>
</div>
<div><br>
</div>
<div>
<div style="font-size:10px; color:#575757">Sent from Samsung Mobile</div>
</div>
</div>
<br>
<br>
-------- Original message --------<br>
From: Tobias Malkmus <br>
Date:31/07/2014 09:14 (GMT+00:00) <br>
To: dune-fem@dune-project.org <br>
Subject: Re: [dune-fem] Petsc test fails in parallel case <br>
<br>
</div>
<font size="2"><span style="font-size:10pt;">
<div class="PlainText">-----BEGIN PGP SIGNED MESSAGE-----<br>
Hash: SHA1<br>
<br>
Hi Andrea and Andreas<br>
<br>
I tested the l2projection test for petsc (now in<br>
dune/fem/test/l2projection_petsc) and the solver example within the<br>
howto with the newest petsc version ( 3.5 ).<br>
<br>
I can not see any problems in parallel/ serial using ALURGRID_SIMPLEX<br>
as the grid type and openmpi 1.8.1 in the master branch.<br>
<br>
For the release branch the 'old' test fails at another place. I will<br>
look into it. The solver example still works fine.<br>
<br>
As Andreas allready said, petsc tries to catch all errors and reports<br>
them to dune. Dune (fem) reports them as a Petsc error, although it<br>
might be a SegFault or an other system error.<br>
<br>
In you case Andrea, Petsc reports a segmentation fault:<br>
PETSC ERROR: Caught signal number 11 SEGV: Segmentation Violation,<br>
probably memory access out of range<br>
<br>
First try to compile this test with CXXFLAGS="-Wall -g" to enable all<br>
the asserts. Please report the error messages from this test.<br>
<br>
Could you please further check wether the solver example in the howto<br>
shows the same behavior.<br>
<br>
<br>
Best Tobias<br>
<br>
On 07/30/2014 10:18 PM, Andreas Dedner wrote:<br>
> Hi Andrea.<br>
> <br>
> That test has apparently been removed from dune-fem - at least it<br>
> is not there in the master. So it might be old....<br>
> <br>
> I had a short look. I seems to me that if you only compile<br>
> l2projection then it does not use petsc. You need <br>
> PETSCLINEAROPERATOR == 1 and that is defined if you compile<br>
> l2projection_petscmatrix And have usepetsc: true in the parameter<br>
> file.<br>
> <br>
> So I don't really see why you are getting a petsc error with <br>
> l2projextion since it should not be using petsc.... But perhaps I<br>
> missed something. Have you tried the example in the dune-fem-howto<br>
> - I am sure that that worked with the older version of petsc.<br>
> <br>
> I tested it myself and I also get an error. In my case an assert<br>
> is triggered. Are you defining NDEBUG? If so then that might<br>
> explain the problem and I have noticed that sometimes there when is<br>
> a segmentation fault somewhere a petsc error is produced although<br>
> it has nothing to do with petsc So perhaps that is what is happing<br>
> here.<br>
> <br>
> The assert btw. is incodimindexset.hh with some invalide index<br>
> being used. That is going to need a bit more investigation I'm<br>
> afraid. Could you please also test the example_solver from the<br>
> dune-fem-howto - I just don't want to fix some test in dune-fem <br>
> which is not used anymore anyway....<br>
> <br>
> Best Andreas<br>
> <br>
> <br>
> On 30/07/14 18:50, Sacconi, Andrea wrote:<br>
>> Hi all,<br>
>> <br>
>> I would like to ask you another question about Petsc and its<br>
>> usage in dune-fem. Recently, I installed the latest release of<br>
>> this library (3.5.0), and thanks to a patch from Tobias Malkmus,<br>
>> it now compiles correctly with dune-fem-1.4.0.<br>
>> <br>
>> I ran the demo test contained in<br>
>> dune/fem/petsc/tests/l2projection, and I am sorry to say that it<br>
>> failed in the parallel case. You can find attached the output of<br>
>> the parallel run with 4 processors.<br>
>> <br>
>> Some additional info:<br>
>> <br>
>> gcc version 4.8.2 OpenMPI 1.6.5 Parmetis: 4.0.3 Petsc 3.5.0) >> I<br>
>> don't know what I'm doing wrong, since I ran the demo tests in<br>
>> Petsc and they all passed without errors. If you need any further<br>
>> information, just let me know.<br>
>> <br>
>> Cheers, Andrea <br>
>> __________________________________________________________<br>
>> <br>
>> Andrea Sacconi PhD student, Applied Mathematics AMMP Section,<br>
>> Department of Mathematics, Imperial College London, London SW7<br>
>> 2AZ, UK a.sacconi11@imperial.ac.uk<br>
>> <br>
>> <br>
>> _______________________________________________ dune-fem mailing<br>
>> list dune-fem@dune-project.org <br>
>> <a href="http://lists.dune-project.org/mailman/listinfo/dune-fem">http://lists.dune-project.org/mailman/listinfo/dune-fem</a><br>
> <br>
> <br>
> <br>
> <br>
> _______________________________________________ dune-fem mailing<br>
> list dune-fem@dune-project.org <br>
> <a href="http://lists.dune-project.org/mailman/listinfo/dune-fem">http://lists.dune-project.org/mailman/listinfo/dune-fem</a><br>
> <br>
<br>
<br>
- -- <br>
Tobias Malkmus <tomalk at mathematik.uni-freiburg.de><br>
<br>
Mathematisches Institut Tel: +49 761 203 5627<br>
Abt. für Angewandte Mathematik Universität Freiburg<br>
Hermann-Herder-Str. 10<br>
79104 Freiburg<br>
<br>
-----BEGIN PGP SIGNATURE-----<br>
Version: GnuPG v2.0.22 (GNU/Linux)<br>
Comment: Using GnuPG with Thunderbird - <a href="http://www.enigmail.net/">http://www.enigmail.net/</a><br>
<br>
iQEcBAEBAgAGBQJT2fq2AAoJEBMY7DHvcWNg9lkIAK7i675PAt9ZDCoDUr6F6gWQ<br>
Rmm5xLnXHjC8uPZ64KFU0CZBPMd1geRoKwBHM6rEQOdjggPrguzEJt2FkcyrPrbS<br>
dSI4oxbjmAVy6Fnn2KbhSv6BddG35qXmSO/ecttOOUwrUUKKrpkKamBmcvShhxc7<br>
GKKTsqe0fi4YZVJp8dkm5K3OihzNGAdtrt81UIDammhL1BXmeXfesnnVCwF/O6FG<br>
YG6ID5nVthmADTeUCAjeU7WPXkepiw8nccVyr81f5PjXlU+GXRNB1RXZ3152tYsZ<br>
d8trlIkX5xhibmow10U1OWYDunB23TFY2xSqm3ADWOEJaokAfLxwxhEFPKN/OTY=<br>
=/MXX<br>
-----END PGP SIGNATURE-----<br>
<br>
_______________________________________________<br>
dune-fem mailing list<br>
dune-fem@dune-project.org<br>
<a href="http://lists.dune-project.org/mailman/listinfo/dune-fem">http://lists.dune-project.org/mailman/listinfo/dune-fem</a><br>
</div>
</span></font>
</body>
</html>