Re: [Libmesh-users] FW: libmesh problem: libmesh_nullptr was not declared in scope

SourceForge Headquarters 225 Broadway Suite 1600 San Diego, CA 92101 +1 (858) 422-6466

The program I ran wasn't in the libmesh directory when I ran it. I just
compiled it with:

 mpicxx -o HPeigenproblems_PZ_8_2_16_2 HPeigenproblems_PZ_8_2_16_2.C -lz
-Wl,-rpath,/usr/local/lib -L/usr/local/lib -lmesh_opt -I/usr/local/include
-pthread -I/usr/local/include -std=gnu++11 -O2 -felide-constructors
-funroll-loops -fstrict-aliasing -Wdisabled-optimization -fopenmp

So I would think this would mean that libmesh could still have SLEPc.
Additionally I looked in the configure.log file within libmesh-1.0.0-rc1
and it didn't look like SLEPc was missing. The real example files (The ones
that came with libmesh) ran fine.

I ran make -n example-opt in the section of examples on eigenproblems_ex2,
made a copy of eigenproblems_ex2, changed the name of the original
eigenproblems_ex2 to realeigenproblems_ex2, and put my .C file in the new
folder called eigenproblems_ex2 in , deleting the old .C file that was
there (I changed the name of my .C file to eigenproblems_ex2.C). I then
tried running the code that I obtained as output from make -n example-opt
and obtained the following responses:

My input: /opt/libmesh-1.0.0-rc1/examples/eigenproblems/eigenproblems_ex2$
sudo echo "  CXX     " example_opt-eigenproblems_ex2.o;mpicxx
-DHAVE_CONFIG_H -I. -I../../../include  -DNDEBUG
-I/opt/slepc-3.7.0/include -I/opt/slepc-3.7.0/arch-linux2-c-debug/include
-I/opt/petsc-3.7.1/include -I/opt/petsc-3.7.1/arch-linux2-c-debug/include
-I/usr/include/mpich -I/usr/local/include -I../../../include
-I../../../contrib/nanoflann/include -I../../../contrib/fparser
-I../../../contrib/libHilbert/include
-I../../../contrib/nemesis/v5.22/nemesis
-I../../../contrib/exodusii/v5.22/exodus/cbind/include
-I../../../contrib/netcdf/v4/include -I../../../contrib/eigen/eigen
-I../../../contrib/gmv -I../../../contrib/qhull/qhull/src
-I../../../contrib/qhull/qhull/src/libqhull
-I../../../contrib/qhull/qhull/src/libqhullcpp
-I../../../contrib/parmetis/include -I../../../contrib/metis/include
-I../../../contrib/gzstream   -std=gnu++11 -O2 -felide-constructors
-funroll-loops -fstrict-aliasing -Wdisabled-optimization    -fopenmp
-std=gnu++11 -MT example_opt-eigenproblems_ex2.o -MD -MP -MF
.deps/example_opt-eigenproblems_ex2.Tpo -c -o
example_opt-eigenproblems_ex2.o `test -f 'eigenproblems_ex2.C' || echo
'./'`eigenproblems_ex2.C
  CXX      example_opt-eigenproblems_ex2.o
eigenproblems_ex2.C:18:21: fatal error: nr3.h: No such file or directory
     #include "nr3.h"
                     ^
compilation terminated.
hpearce@islands:/opt/libmesh-1.0.0-rc1/examples/eigenproblems/eigenproblems_ex2$
mv -f .deps/example_opt-eigenproblems_ex2.Tpo
.deps/example_opt-eigenproblems_ex2.Po
mv: cannot stat ‘.deps/example_opt-eigenproblems_ex2.Tpo’: No such file or
directory
hpearce@islands:/opt/libmesh-1.0.0-rc1/examples/eigenproblems/eigenproblems_ex2$
rm -f example-opt
rm: cannot remove ‘example-opt’: Permission denied
hpearce@islands:/opt/libmesh-1.0.0-rc1/examples/eigenproblems/eigenproblems_ex2$
echo "  CXXLD   " example-opt;/bin/bash ../../../libtool --silent
--tag=CXX   --mode=link mpicxx -std=gnu++11 -O2 -felide-constructors
-funroll-loops -fstrict-aliasing -Wdisabled-optimization    -fopenmp
-std=gnu++11   -o example-opt example_opt-eigenproblems_ex2.o ../../../
libmesh_opt.la -lz -Wl,-rpath,/opt/slepc-3.7.0/arch-linux2-c-debug/lib
-L/opt/slepc-3.7.0/arch-linux2-c-debug/lib -lslepc
-L/opt/petsc-3.7.1/arch-linux2-c-debug/lib -L/lib/x86_64-linux-gnu -lpetsc
-Wl,-rpath,/opt/petsc-3.7.1/arch-linux2-c-debug/lib -lf2clapack -lf2cblas
-lmpichf90 -lgfortran -lquadmath -lm -lmpichcxx -lstdc++
-Wl,-rpath,/lib/x86_64-linux-gnu -lmpich -lopa -lmpl -lrt -lcr -lpthread
-lgcc_s -ldl -L/usr/lib/x86_64-linux-gnu
-L/usr/lib/gcc/x86_64-linux-gnu/4.8
-Wl,-rpath,/usr/lib/gcc/x86_64-linux-gnu/4.8
-Wl,-rpath,/usr/lib/x86_64-linux-gnu
  CXXLD    example-opt
rm: cannot remove 'example-opt': Permission denied
rm: cannot remove '.libs/example-opt': Permission denied
rm: cannot remove '.libs/lt-example-opt': Permission denied
/usr/bin/ld: cannot open output file .libs/example-opt: Permission denied
collect2: error: ld returned 1 exit status

On Tue, Aug 9, 2016 at 10:52 AM, Mike Marchywka <mar...@ho...>
wrote:

>
>
>
>
> ________________________________
> > From: har...@gm...
> > Date: Tue, 9 Aug 2016 10:35:47 -0700
> > Subject: Re: [Libmesh-users] FW: libmesh problem: libmesh_nullptr was
> > not declared in scope
> > To: mar...@ho...
> > CC: roy...@ic...; lib...@li...
> >
> >
> >
> > Thank you for your help so far. It looks like this problem is close to
> > being finished. Your recommendation of using #include "libmesh/foo.h"
> > worked for compiling the program (I used the full path for
> > libmesh_nullptr.h in the include part and that fixed those errors) but
> > now when I run the executable I get the following error:
> >
> > ERROR: This example requires libMesh to be
> > compiled with SLEPc eigen solvers support!
> >
> > I already have SLEPc though. I would think to fix this problem I would
> > need to either add a compiler flag for SLEPc or add the program I have
> > to the example programs.
> >
> > I would think linking would work just as well as adding to my example
> > programs so I added -L/opt/slepc-3.7.0/sys
> > to the compiling command but this didn't fix the problem.
>
>
> I noticed a couple of things immediately about libmesh that were
> interesting.
> One was the term "exact solution" appearing almost everywhere and
> even being allowed on the command line IIRC in one case- there is
> some hope even if everything runs ok you can check the result. The other
> was the large array of build options and the checks in the examples- I had
> no idea I needed vtk for
> example of what any of the other things brought. To add these, you
> need to rebuild libmesh and reconfigure to enable if
> not found automatically. I ignore all the automake stuff, just run
> ./configure
> with the options set although make clean may help first.
>
> And again go to the examples and run make -n on any target and
> see what you get.
>
> Curious what the name collision is however. I just ran a script to find
> repeated names in my /usr/include and some occur 30 times lol.
> Using directories is probably a good idea.
>
> --- end of comments
>
> >
> >
> > Harry Pearce
> >
> >
> >
> > On Mon, Aug 8, 2016 at 11:20 AM, Mike Marchywka
> > <mar...@ho...<mailto:mar...@ho...>> wrote:
> >
> >
> >
> >
> > ________________________________
> >> From: har...@gm...<mailto:har...@gm...>
> >> Date: Mon, 8 Aug 2016 10:35:42 -0700
> >> Subject: Re: [Libmesh-users] FW: libmesh problem: libmesh_nullptr was
> >> not declared in scope
> >> To: mar...@ho...<mailto:mar...@ho...>;
> > roy...@ic...<mailto:roy...@ic...>
> >> CC:
> > lib...@li...<mailto:libmesh
> -u...@li...>
> >>
> >> Oops I thought I replied to everyone.
> >>
> >>>Did it complain it can't find the header file first?
> >>
> >> It did complain about the libmesh.h header before I added the following
> >> to set my environment:
> >> export PATH=/opt/petsc-3.7.1:/opt/slepc-3.7.0:/opt/libmesh-1.0.
> 0-rc1:$PATH
> >> export PATH=/opt/petsc-3.7.1:/opt/libmesh-1.0.0-rc1:$PATH
> >> export PATH=$PETSC_DIR/arch-linux2-c-debug:$PATH
> >> export CPATH=/opt/libmesh-1.0.0-rc1/include/libmesh/:$CPATH
> >
> > without the libmesh dir on the include it could be a name collision.
> > There is probably some way to get the compiler to tell you which dir
> > it includes from. Also you can get all of include paths, use "find" or
> > something and check for duplicate header names.
> >
> > The examples inlcuded "libmesh/foo.h" which is probably a good idea
> > to stick with although maybe someone who has used it more can comment :)
> >
> >
> >>
> >>>if you run "make -n example-opt" and copy the output it should work
> >>
> >> I tried running that command in the directory where my example program
> >> is and it said
> >> "make: *** No rule to make target `example-opt'. Stop."
> >
> > Well, the idea was using the makefile for a working example.
> > Maybe I mistyped the target name but you can look through the
> > examples.
> >
> >>
> >>
> >>
> >>>although without the libmesh dir in the includes I guess you would have
> >>>to add it to the include paths.
> >> libmesh.h is in the includes file.
> >>
> >> I still have the question about how to add my example file to the
> >> configure.ac<http://configure.ac><http://configure.ac> file in the
> > section AC_CONFIG_FILES
> >> since it's required in order to add a new example file according to
> >> https://github.com/libMesh/libmesh/wiki/Adding-a-new-example. If you
> >> don't know how to add the example to the AC_CONFIG_FILES (for example
> >> do I add the directory of the Makefile I want to create in it?) Is
> >> there another potential way to compile my program? I did try editing
> >> the example makefile to include the directory of my project under
> >> SUBDIRS but when I tried to run the makefile the compiler said:
> >
> > What are you using for a makefile now? I'm not sure this is even relevant
> > for what you are doing. Getting your code to compile AFAICT is a
> standalone
> > activity that will just need libmesh include files and libs for linking.
> >
> > --- end of new comments ---
> >
> >>
> >> Making all in eigenproblems/eigenproblems_femviblibtests_ex1
> >> make[1]: Entering directory
> >>
> > `/opt/libmesh-1.0.0-rc1/examples/eigenproblems/
> eigenproblems_femviblibtests_ex1'
> >> make[1]: *** No rule to make target `all'. Stop.
> >> make[1]: Leaving directory
> >>
> > `/opt/libmesh-1.0.0-rc1/examples/eigenproblems/
> eigenproblems_femviblibtests_ex1'
> >> make: *** [all-recursive] Error 1
> >>
> >> On Mon, Aug 8, 2016 at 10:34 AM, Harry Pearce
> >>
> > <har...@gm...<mailto:har...@gm...><mailto:
> har...@gm...<mailto:har...@gm...>>>
> > wrote:
> >>>Did it complain it can't find the header file first?
> >>
> >> It did complain about the libmesh.h header before I added the following
> >> to set my environment:
> >> export PATH=/opt/petsc-3.7.1:/opt/slepc-3.7.0:/opt/libmesh-1.0.
> 0-rc1:$PATH
> >> export PATH=/opt/petsc-3.7.1:/opt/libmesh-1.0.0-rc1:$PATH
> >> export PATH=$PETSC_DIR/arch-linux2-c-debug:$PATH
> >> export CPATH=/opt/libmesh-1.0.0-rc1/include/libmesh/:$CPATH
> >>
> >>>if you run "make -n example-opt" and copy the output it should work
> >>
> >> I tried running that command in the directory where my example program
> >> is and it said
> >> "make: *** No rule to make target `example-opt'. Stop."
> >>
> >>
> >>
> >>>although without the libmesh dir in the includes I guess you would have
> >>>to add it to the include paths.
> >> libmesh.h is in the includes file.
> >>
> >> I still have the question about how to add my example file to the
> >> configure.ac<http://configure.ac><http://configure.ac> file in the
> > section AC_CONFIG_FILES
> >> since it's required in order to add a new example file according to
> >> https://github.com/libMesh/libmesh/wiki/Adding-a-new-example. If you
> >> don't know how to add the example to the AC_CONFIG_FILES (for example
> >> do I add the directory of the Makefile I want to create in it?) Is
> >> there another potential way to compile my program? I did try editing
> >> the example makefile to include the directory of my project under
> >> SUBDIRS but when I tried to run the makefile the compiler said:
> >>
> >> Making all in eigenproblems/eigenproblems_femviblibtests_ex1
> >> make[1]: Entering directory
> >>
> > `/opt/libmesh-1.0.0-rc1/examples/eigenproblems/
> eigenproblems_femviblibtests_ex1'
> >> make[1]: *** No rule to make target `all'. Stop.
> >> make[1]: Leaving directory
> >>
> > `/opt/libmesh-1.0.0-rc1/examples/eigenproblems/
> eigenproblems_femviblibtests_ex1'
> >> make: *** [all-recursive] Error 1
> >>
> >> Harry Pearce
> >>
> >>
> >> On Fri, Aug 5, 2016 at 2:32 PM, Mike Marchywka
> >>
> > <mar...@ho...<mailto:mar...@ho...><mailto:
> mar...@ho...<mailto:mar...@ho...>>>
> > wrote:
> >>
> >>
> >>
> >> Did it complain it can't find the header file first?
> >> My includes for libmesh all start libmesh/foo.h
> >>
> >> if you run "make -n example-opt" and copy the output it should work
> >> although without the libmesh dir in the includes I guess you would have
> >> to add it to the include paths.
> >>
> >>
> >> ________________________________
> >>> From:
> > har...@gm...<mailto:har...@gm...><mailto:
> har...@gm...<mailto:har...@gm...>>
> >>> Date: Fri, 5 Aug 2016 14:07:58 -0700
> >>> Subject: Re: [Libmesh-users] FW: libmesh problem: libmesh_nullptr was
> >>> not declared in scope
> >>> To:
> > mar...@ho...<mailto:mar...@ho...><mailto:m
> arc...@ho...<mailto:mar...@ho...>>
> >>> CC:
> > roy...@ic...<mailto:roy...@ic...><mailto:
> roy...@ic...<mailto:roy...@ic...>>;
> >>
> > lib...@li...<mailto:libmesh
> -u...@li...><mailto:lib...@li...
> <mailto:lib...@li...>>
> >>>
> >>>>"I skimmed most of the messages and it did sound like it would be
> > something
> >>>>easy to find by looking at the exact code so I was a bit surprised
> >>> with that last post :)"
> >>>
> >>> Just in case it is something easy to spot I have the code at the bottom
> >>> of the email highlighted. If you don't need the code you can just
> >>> ignore the highlighted part of the email.
> >>>
> >>>>The linking is a bit unorthodox, indeed I think you had some specific
> >>>>recommendations for linking to libmesh that hid most of the details,
> but
> >>>>actually building it and compiling with it were pretty easy. I turned
> >>> on some debug info
> >>>>from the example make script, just copied its invokation of g++ and
> > libtool
> >>>>and that seems to work for some simple case.
> >>>
> >>> This is my first time compiling a libmesh code so I tried to figure out
> >>> what compiler flags to use in the suggested manner
> >>> for simple codes (according to
> >>> http://libmesh.github.io/installation.html) which is libmesh-config
> >>> --cxx` -o HPeigenproblems_PZ_8_2_16_2 HPeigenproblems_PZ_8_2_16_2.C
> >>> `libmesh-config --cxxflags --include --ldflags --libs`.
> >>>
> >>> This is my first time compiling a libmesh code so I thought maybe I
> >>> should use a Makefile.am from a regular libmesh example.
> >>> (My other example files compiled properly) so I was thinking that it
> >>> might work to follow the instructions for adding a new libmesh example
> >>> (https://github.com/libMesh/libmesh/wiki/Adding-a-new-example) which
> >>> invoves changing the Makefile.am and run.sh from another example file
> >>> to run the program.
> >>>
> >>> I started following the directions for adding a new example and am not
> >>> sure what line to put into AC_CONFIG_FILES in
> >>>
> > configure.ac<http://configure.ac><http://configure.ac><http:
> //configure.ac>
> > at the top
> >> level of libmesh. I would
> >>> have thought it would have been the directory of the example file I am
> >>> adding or the Makefile that goes with it but I do not see the
> >>> directories of any other example programs in AC_CONFIG_FILES within the
> >>>
> > configure.ac<http://configure.ac><http://configure.ac><http:
> //configure.ac>
> > program.
> >>>
> >>>
> >>>
> >>>
> >>>
> >>> Harry Pearce
> >>>
> >>> Just in case highlighting doesn't show up on the libmesh users archive,
> >>> the highlighted code is below:
> >>> #include <fstream>
> >>> #include "libmesh.h"
> >>> #include "mesh.h"
> >>> #include "libmesh_config.h"
> >>> //#include "mesh_generation.h"
> >>> // #include "exodusII_io.h"
> >>> #include "eigen_system.h"
> >>> #include "equation_systems.h"
> >>> // #include "fe.h"
> >>> // #include "quadrature_gauss.h"
> >>> // #include "dense_matrix.h"
> >>> // #include "sparse_matrix.h"
> >>> // #include "numeric_vector.h"
> >>> // #include "dof_map.h"
> >>> #include "libmesh_nullptr.h"
> >>> #include <string.h>
> >>> // #include "parallel_algebra.h"
> >>> //#include "parallel.h"
> >>> //#include <auto_ptr.h>
> >>>
> >>>
> >>> using namespace libMesh;
> >>>
> >>> char prefix[80];
> >>>
> >>> void assemble_mass(EquationSystems& es,
> >>>
> >>> const std::string& system_name);
> >>>
> >>> int main (int argc, char** argv)
> >>>
> >>> {
> >>>
> >>> LibMeshInit init (argc, argv);
> >>>
> >>> #ifndef LIBMESH_HAVE_SLEPC
> >>>
> >>> if (libMesh::processor_id() == 0)
> >>>
> >>> std::cerr << "ERROR: This example requires libMesh to be\n"
> >>>
> >>> << "compiled with SLEPc eigen solvers support!"
> >>>
> >>> << std::endl;
> >>>
> >>> return 0;
> >>>
> >>> #else
> >>>
> >>> #ifdef LIBMESH_DEFAULT_SINGLE_PRECISION
> >>>
> >>> libmesh_example_assert(false, "--disable-singleprecision");
> >>>
> >>> #endif
> >>>
> >>> if (argc < 3)
> >>>
> >>> {
> >>>
> >>> if (libMesh::processor_id() == 0)
> >>>
> >>> std::cerr << "\nUsage: " << argv[0]
> >>>
> >>> << " -n <number of eigen values>"
> >>>
> >>> << std::endl;
> >>>
> >>> libmesh_error();
> >>>
> >>> }
> >>>
> >>>
> >>>
> >>> else
> >>>
> >>> {
> >>>
> >>> std::cout << "Running " << argv[0];
> >>>
> >>>
> >>>
> >>> for (int i=1; i<argc; i++)
> >>>
> >>> std::cout << " " << argv[i];
> >>>
> >>>
> >>>
> >>> std::cout << std::endl << std::endl;
> >>>
> >>> }
> >>>
> >>> const unsigned int nev = std::atoi(argv[2]);
> >>>
> >>> // libmesh_example_assert(3 <= LIBMESH_DIM, "2D support");
> >>>
> >>>
> >>>
> >>> Mesh mesh;
> >>>
> >>> MeshTools::Generation::build_cube (mesh,
> >>>
> >>> 10, 10, 10,
> >>>
> >>> -1., 1.,
> >>>
> >>> -1., 1.,
> >>>
> >>> -1., 1.,
> >>>
> >>> HEX27);
> >>>
> >>> mesh.print_info();
> >>>
> >>>
> >>>
> >>> char xyzfilename[80];
> >>>
> >>> strcpy(prefix,argv[3]);
> >>>
> >>> strcpy(xyzfilename,argv[3]);
> >>>
> >>> strcat(xyzfilename,".xyz");
> >>>
> >>> std::cout << "Writing nodal xyz coordinates to file: " <<
> >>> xyzfilename << std::endl;
> >>>
> >>> std::ofstream xyzfile;
> >>>
> >>> xyzfile.open (xyzfilename);
> >>>
> >>> MeshBase::const_node_iterator node_it = mesh.nodes_begin();
> >>>
> >>>
> >>>
> >>>
> >>> const MeshBase::const_node_iterator node_end = mesh.nodes_end();
> >>>
> >>> xyzfile << "Number of nodes: " << mesh.n_nodes() << "\n";
> >>>
> >>> for (; node_it != node_end; ++node_it)
> >>>
> >>> {
> >>>
> >>> //count++;
> >>>
> >>> const Point& p = **node_it;
> >>>
> >>> const Real px = p(0);
> >>>
> >>> const Real py = p(1);
> >>>
> >>> xyzfile << px << " " << py << " \n";
> >>>
> >>> }
> >>>
> >>> xyzfile.close();
> >>>
> >>> std::cout << "Finished writing nodal xyz coordinates to file: " <<
> >>> xyzfilename << std::endl;
> >>>
> >>>
> >>> EquationSystems equation_systems (mesh);
> >>>
> >>> EigenSystem & eigen_system =
> >>>
> >>> equation_systems.add_system<EigenSystem> ("Eigensystem");
> >>>
> >>>
> >>>
> >>>
> >>> eigen_system.add_variable("u", SECOND);
> >>>
> >>> eigen_system.attach_assemble_function (assemble_mass);
> >>>
> >>>
> >>
> > equation_systems.parameters.se<http://equation_systems.parameters.se><
> http://equation_systems.parameters.se>t<unsigned
> >> int>("eigenpairs") = nev;
> >>>
> >>>
> >>
> > equation_systems.parameters.se<http://equation_systems.parameters.se><
> http://equation_systems.parameters.se>t<unsigned
> >> int>("basis vectors") = nev*3;
> >>>
> >>> eigen_system.eigen_solver->set_eigensolver_type(KRYLOVSCHUR);
> >>>
> >>> // eigen_system.eigen_solver->set_eigensolver_type(LAPACK);
> >>>
> >>> eigen_system.eigen_solver->set_position_of_spectrum(SMALLEST_REAL);
> >>>
> >>>
> >>
> > equation_systems.parameters.se<http://equation_systems.parameters.se><
> http://equation_systems.parameters.se>t<Real>("linear
> >> solver tolerance") =
> >>> pow(TOLERANCE, 5./3.);
> >>>
> >>>
> >>
> > equation_systems.parameters.se<http://equation_systems.parameters.se><
> http://equation_systems.parameters.se>t<unsigned
> >> int>
> >>>
> >>> ("linear solver maximum iterations") = 1000;
> >>>
> >>> eigen_system.set_eigenproblem_type(GHEP);
> >>>
> >>> equation_systems.init();
> >>>
> >>> equation_systems.print_info();
> >>>
> >>>
> >>>
> >>> eigen_system.solve();
> >>>
> >>> unsigned int nconv = eigen_system.get_n_converged();
> >>>
> >>> std::cout << "Number of converged eigenpairs: " << nconv
> >>>
> >>> << "\n" << std::endl;
> >>>
> >>> if (nconv != 0)
> >>>
> >>> {
> >>>
> >>> eigen_system.get_eigenpair(nconv-1);
> >>>
> >>>
> >>>
> >>> #ifdef LIBMESH_HAVE_EXODUS_API
> >>>
> >>> ExodusII_IO (mesh).write_equation_systems ("out.e", equation_systems);
> >>>
> >>> #endif // #ifdef LIBMESH_HAVE_EXODUS_API
> >>>
> >>> }
> >>>
> >>> else
> >>>
> >>> {
> >>>
> >>> std::cout << "WARNING: Solver did not converge!\n" << nconv <<
> >>> std::endl;
> >>>
> >>> }
> >>>
> >>> #endif // LIBMESH_HAVE_SLEPC
> >>>
> >>> return 0;
> >>>
> >>> }
> >>>
> >>> void assemble_mass(EquationSystems& es,
> >>>
> >>> const std::string& system_name)
> >>>
> >>> {
> >>>
> >>> char potfilename[80];
> >>>
> >>> strcpy(potfilename,prefix);
> >>>
> >>> strcat(potfilename,".pot.out");
> >>>
> >>> std::fstream potfile;
> >>>
> >>> potfile.open(potfilename, std::ios::out);
> >>>
> >>> if(!potfile) {std::cerr << "Error opening PES output file."; return;}
> >>>
> >>> libmesh_assert (system_name == "Eigensystem");
> >>>
> >>> #ifdef LIBMESH_HAVE_SLEPC
> >>>
> >>> const MeshBase& mesh = es.get_mesh();
> >>>
> >>> const unsigned int dim = mesh.mesh_dimension();
> >>>
> >>> EigenSystem & eigen_system = es.get_system<EigenSystem> (system_name);
> >>>
> >>> FEType fe_type = eigen_system.get_dof_map().variable_type(0);
> >>>
> >>> SparseMatrix<Number>& matrix_A = *eigen_system.matrix_A;
> >>>
> >>> SparseMatrix<Number>& matrix_B = *eigen_system.matrix_B;
> >>>
> >>> AutoPtr<FEBase> fe (FEBase::build(dim, fe_type));
> >>>
> >>> QGauss qrule (dim, fe_type.default_quadrature_order());
> >>>
> >>> fe->attach_quadrature_rule (&qrule);
> >>>
> >>> const std::vector<Real>& JxW = fe->get_JxW();
> >>>
> >>> const std::vector<Point>& q_point = fe->get_xyz();
> >>>
> >>> const std::vector<std::vector<Real>>& phi = fe->get_phi();
> >>>
> >>> const std::vector<std::vector<RealGradient>>& dphi = fe->get_dphi();
> >>>
> >>> const DofMap& dof_map = eigen_system.get_dof_map();
> >>>
> >>> DenseMatrix<Number> Me;
> >>>
> >>> DenseMatrix<Number> Ke;
> >>>
> >>> std::vector<unsigned int> dof_indices;
> >>>
> >>> MeshBase::const_element_iterator el =
> >>> mesh.active_local_elements_begin();
> >>>
> >>> const MeshBase::const_element_iterator end_el =
> >>> mesh.active_local_elements_end();
> >>>
> >>>
> >>>
> >>> std::cout << "Writing PES data to file: " << potfilename << std::endl;
> >>>
> >>> double V, gxx, gxy, gyy, gzx, gzy, gzz;
> >>>
> >>> for ( ; el != end_el; ++el)
> >>>
> >>> {
> >>>
> >>> const Elem* elem = *el;
> >>>
> >>> dof_map.dof_indices (elem, dof_indices);
> >>>
> >>> fe->reinit (elem);
> >>>
> >>> Ke.resize (dof_indices.size(), dof_indices.size());
> >>>
> >>> Me.resize (dof_indices.size(), dof_indices.size());
> >>>
> >>> for (unsigned int qp=0; qp<qrule.n_points(); qp++)
> >>>
> >>> {
> >>>
> >>> const double x = q_point[qp](0);
> >>>
> >>> const double y = q_point[qp](1);
> >>>
> >>> const double z = q_point[qp](2);
> >>>
> >>> /*
> >>>
> >>> Doub potpoint(ndim);
> >>>
> >>> potpoint[0] = x;
> >>>
> >>> potpoint[1] = y;
> >>>
> >>> potpoint[2] = z;
> >>>
> >>> */
> >>>
> >>> V=x*x + y*y + z*z;
> >>>
> >>> gxx=0.10;
> >>>
> >>> gyy=1.20;
> >>>
> >>> gxy=1.10;
> >>>
> >>> gzx=0.003;
> >>>
> >>> gzy=0.40;
> >>>
> >>> gzz=0.50;
> >>>
> >>>
> >>> potfile << x << '\t' << y << '\t' << z << '\t' << V << '\t' << "\n";
> >>>
> >>>
> >>> /*
> >>>
> >>> const Elem* elem = *el;
> >>>
> >>> dof_map.dof_indices (elem, dof_indices);
> >>>
> >>> fe->reinit (elem);
> >>>
> >>> Ke.resize (dof_indices.size(), dof_indices.size());
> >>>
> >>> Me.resize (dof_indices.size(), dof_indices.size());
> >>>
> >>> */
> >>>
> >>> // for (unsigned int qp=0; qp<qrule.n_points(); qp++)
> >>>
> >>> for (unsigned int i=0; i<phi.size(); i++)
> >>>
> >>> for (unsigned int j=0; j<phi.size(); j++)
> >>>
> >>> {
> >>>
> >>>
> >>> Ke(i,j) += JxW[qp]*(
> >>>
> >>> 0.5*( gxx*dphi[i][qp](0)*dphi[j][qp](0)
> >>>
> >>> + gyy*dphi[i][qp](1)*dphi[j][qp](1)
> >>>
> >>> + gzz*dphi[i][qp](2)*dphi[j][qp](2)
> >>>
> >>> + gxy*dphi[i][qp](0)*dphi[j][qp](1)
> >>>
> >>> + gxy*dphi[i][qp](1)*dphi[j][qp](0)
> >>>
> >>> + gzx*dphi[i][qp](0)*dphi[j][qp](2)
> >>>
> >>> + gzx*dphi[i][qp](2)*dphi[j][qp](0)
> >>>
> >>> + gzy*dphi[i][qp](1)*dphi[j][qp](2)
> >>>
> >>> + gzy*dphi[i][qp](2)*dphi[j][qp](1)
> >>>
> >>> // + dgxxdx*dphi[i][qp](0)
> >>>
> >>> // + dgyydy*dphi[i][qp](1)
> >>>
> >>> // + dgzzdz*dphi[i][qp](2)
> >>>
> >>> // + dgxydx*dphi[i][qp](1)
> >>>
> >>> // + dgxydy*dphi[i][qp](0)
> >>>
> >>> // + dgxzdx*dphi[i][qp](2)
> >>>
> >>> // + dgzxdz*dphi[i][qp](0)
> >>>
> >>> // + dgyzdy*dphi[i][qp](2)
> >>>
> >>> // + dgzydz*dphi[i][qp](1)
> >>>
> >>> )*33.71526 + V*phi[i][qp]*phi[j][qp]);
> >>>
> >>>
> >>>
> >>> Me(i,j) += JxW[qp]*phi[i][qp]*phi[j][qp];
> >>>
> >>> }
> >>>
> >>> matrix_A.add_matrix (Ke, dof_indices);
> >>>
> >>> matrix_B.add_matrix (Me, dof_indices);
> >>>
> >>> }
> >>>
> >>> } // end of element loop
> >>>
> >>> std::cout << "Finished writing PES data to file: " << potfilename
> >>> << std::endl;
> >>>
> >>> potfile.close();
> >>>
> >>> std::cout << "Matrix A assembly completed! " << "\n" << std::endl;
> >>>
> >>> std::cout << "Matrix B assembly completed! " << "\n" << std::endl;
> >>>
> >>> std::cout << "Exporting Matrix A in matlab sparse matrix format...
> >>> " << "\n" << std::endl;
> >>>
> >>> char buf1[14];
> >>>
> >>> sprintf (buf1, "A.m");
> >>>
> >>> matrix_A.print_matlab(buf1);
> >>>
> >>> std::cout << "Matrix A Exported! " << "\n" << std::endl;
> >>>
> >>> // matrix_A.clear();
> >>>
> >>> std::cout << "Matrix A Cleared from Memory! " << "\n" << std::endl;
> >>>
> >>> std::cout << "Exporting Matrix B in matlab sparse matrix format...
> >>> " << "\n" << std::endl;
> >>>
> >>> char buf2[14];
> >>>
> >>> sprintf (buf2, "B.m");
> >>>
> >>> matrix_B.print_matlab(buf2);
> >>>
> >>> std::cout << "Matrix B Exported! " << "\n" << std::endl;
> >>>
> >>> // matrix_B.clear();
> >>>
> >>> std::cout << "Matrix B Cleared from Memory! " << "\n" << std::endl;
> >>>
> >>> #endif // LIBMESH_HAVE_SLEPC
> >>>
> >>> /**
> >>>
> >>> * All done!
> >>>
> >>> */
> >>>
> >>> return;
> >>>
> >>> On Thu, Aug 4, 2016 at 5:41 PM, Mike Marchywka
> >>>
> >>
> > <mar...@ho...<mailto:mar...@ho...><mailto:
> mar...@ho...<mailto:mar...@ho...>><mailto:
> mar...@ho...<mailto:mar...@ho...><mailto:m
> arc...@ho...<mailto:mar...@ho...>>>>
> >> wrote:
> >>>
> >>>
> >>>
> >>>
> >>> ----------------------------------------
> >>>> Date: Thu, 4 Aug 2016 19:11:03 -0500
> >>>> From:
> >>
> > roy...@ic...<mailto:roy...@ic...><mailto:
> roy...@ic...<mailto:roy...@ic...>><mailto:royst
> gn...@ic...<mailto:roy...@ic...><mailto:
> roy...@ic...<mailto:roy...@ic...>>>
> >>>> To:
> >>
> > mar...@ho...<mailto:mar...@ho...><mailto:m
> arc...@ho...<mailto:mar...@ho...>><mailto:m
> arc...@ho...<mailto:mar...@ho...><mailto:ma
> rc...@ho...<mailto:mar...@ho...>>>
> >>>> CC:
> >>>
> >>
> > lib...@li...<mailto:libmesh
> -u...@li...><mailto:lib...@li...
> <mailto:lib...@li...>><mailto:libmesh-users@lists.
> sourceforge.net<mailto:lib...@li...><mailto:
> lib...@li...<mailto:libmesh
> -u...@li...>>>
> >>>> Subject: Re: [Libmesh-users] FW: libmesh problem: libmesh_nullptr was
> >>> not declared in scope
> >>>>
> >>>>
> >>>>
> >>>> On Thu, 4 Aug 2016, Mike Marchywka wrote:
> >>>>
> >>>>>> I have not followed the exchange but sometimes I end up including
> >>> headers within a namespace or
> >>>>>> something. Is libmesh being included within something? Not sure
> >>> what that does with some
> >>>>>> macros and others not lol.
> >>>>
> >>>> That was my first guess, too! That's actually why I was asking about
> >>>> the config macros - if he was in the nullptr-workaround case, then
> >>>> accidentally including a libMesh header within a namespace would have
> >>>> exactly the effect seen: the class we define wouldn't exist in the
> >>>> global namespace and the compiler wouldn't recognize it.
> >>>
> >>> I skimmed most of the messages and it did sound like it would be
> something
> >>> easy to find by looking at the exact code so I was a bit surprised with
> >>> that last post :)
> >>>
> >>> For problems in unfamiliar code grep can be amazingly
> >>> useful even if you expect 1000's of hits they usually end up being easy
> >>> to sort out.
> >>>
> >>> The linking is a bit unorthodox, indeed I think you had some specific
> >>> recommendations for linking to libmesh that hid most of the details,
> but
> >>> actually building it and compiling with it were pretty easy. I turned
> >>> on some debug info
> >>> from the example make script, just copied its invokation of g++ and
> > libtool
> >>> and that seems to work for some simple case.
> >>>
> >>>
> >>>
> >>>
> >>>> ---
> >>>> Roy
> >>>
> >>>
> >>
> > ------------------------------------------------------------
> ------------------
> >>> _______________________________________________
> >>> Libmesh-users mailing list
> >>>
> >>
> > Lib...@li...<mailto:Libmesh
> -u...@li...><mailto:Lib...@li...
> <mailto:Lib...@li...>><mailto:Libmesh-users@lists.
> sourceforge.net<mailto:Lib...@li...><mailto:
> Lib...@li...<mailto:Libmesh
> -u...@li...>>>
> >>> https://lists.sourceforge.net/lists/listinfo/libmesh-users
> >>>
> >>
> >>
> >>
> >
> >
>