For a matrix with any rows in a parallel run, The subgraph
containing only vertices associated with matrix rows owned by
the process would have 1 vertex although the actual matrix graph
would have none. This is fixed by checking the number of vertices
of the outer graph and act correctly for zero vertices.
For this we need to mark a matrix without rows built in row_wise mode
as built once the CreateIterator is constructed and correct the number
of vertices for a Subgraph of a MatrixGraph with 0 vertices.

Deprecate SeqPardiso and remove test_pardiso.

Closes #17

See merge request !152

The preconditioner `SeqPardiso` using the sparse direct solver
library Pardiso has not been actively maintained for several
years. It is still based on a very old Pardiso version; finding
the library never went into the CMake build system. Moreover,
Pardiso is not open source. With SuperLU and UMFPack, two working
alternative direct solvers are available.

Drop the support for Pardiso by deprecating `SeqPardiso` and
removing the corresponding test.

call `abort()` instead of `assert(false)`

See merge request !151

remove private `to_real` member function

See merge request !150

`abort()` will always abort the program instead of silently ignoring
the error when built with `-DNDEBUG`

It reimplements `std::real` and is not used anywhere.

typo: alignement → alignment [ci skip]

See merge request !149

add changelogs from release to master

See merge request !148

use `std::make_unique`

See merge request !147

[test] Add check for BCRSMatrix::nonzeroes()

See merge request !93

added empty changelog

See merge request !145

[solvers.hh] Always convert the reduction to double explicitly.

See merge request !138

This already happened for most solvers when setting the reduction in the
result struct.  It is may be needed when the field type is some extended
number type that prohibits implicit conversion to `double`.  This commit
brings the remaining two solvers in line that did not do this yet.

Fix the #include mess in solvers.hh

See merge request !137

[preconditioners] fix type of CGSolver reduction parameter after recent updates

See merge request !132

[bugfix] ensure nonzeroes always return the number of nonzero entries

Closes #28

See merge request !127

[Warnings] Fix MPI 1 deprecation warnings (by switching to MPI 2).

See merge request !125

fix ambiguity of _1 for certain compilers

See merge request !128

[preconditioner] fix ILDL for SIMD types

See merge request !129

the weight is a scalar, but the current code defines it a field_type,
which might be a SIMD type or something else, which does not behave as
a scalar. We fix this by using the SimdScalar traits to extract the
scalar type and use this.

[Warnings] Fix a signed/unsigned warning.

See merge request !126

This fixes warnings of the following form about functions deprecated in MPI 2.
```
/home/joe/Projekte/dune-simd/dune-istl/dune/istl/paamg/test/parallelamgtest.cc: In function ‘int main(int, char**)’:
/home/joe/Projekte/dune-simd/dune-istl/dune/istl/paamg/test/parallelamgtest.cc:217:3: warning: ‘int MPI_Errhandler_create(void (*)(ompi_communicator_t**, int*, ...), ompi_errhandler_t**)’ is deprecated: MPI_Errhandler_create is superseded by MPI_Comm_create_errhandler in MPI-2.0 [-Wdeprecated-declarations]
   MPI_Errhandler_create(MPI_err_handler, &handler);
   ^~~~~~~~~~~~~~~~~~~~~
In file included from /home/joe/Projekte/dune-simd/dune-common/dune/common/parallel/mpitraits.hh:23:0,
                 from /home/joe/Projekte/dune-simd/dune-common/dune/common/parallel/plocalindex.hh:9,
                 from /home/joe/Projekte/dune-simd/dune-istl/dune/istl/paamg/test/anisotropic.hh:8,
                 from /home/joe/Projekte/dune-simd/dune-istl/dune/istl/paamg/test/parallelamgtest.cc:7:
/usr/lib/x86_64-linux-gnu/openmpi/include/mpi.h:1345:20: note: declared here
 OMPI_DECLSPEC  int MPI_Errhandler_create(MPI_Handler_function *function,
                    ^~~~~~~~~~~~~~~~~~~~~
```

bump version to 2.7-git [ci skip]

See merge request !123

Comment out warnings

See merge request !121

These warnings are very irritating. They do not
warn about deprecated or misused feature. They are
issued any time you compile the tests. They don't
provide any useful information unless you look at the
code and have insight in the implementation detains.

To sum up: They don't look like a warning addressed to
the user, but like a reminder on an intended future
implementation addressed to the implementor. Hence
they should not be exposed to the user.

If anyone feels that this is an important user warning
feel free to re-enable it after adding enough information
making the warning actually useful.

[bugfix] Avoid calling std::real for unsupported types

See merge request !120

The condition estimate is only supported for field_types
float and double. Unfortunately, compiling this failed
for any field_type not supported by std::real. Replacing the
dynamic check by a static one helps to avoid the problem.