Add tests.

diff --git a/tests/mpi/renumber_z_order_05.cc b/tests/mpi/renumber_z_order_05.cc
new file mode 100644 (file)
index 0000000..27fd2a3
--- /dev/null
+++ b/tests/mpi/renumber_z_order_05.cc
@@ -0,0 +1,119 @@
+// ---------------------------------------------------------------------
+//
+// Copyright (C) 2017 - 2018 by the deal.II authors
+//
+// This file is part of the deal.II library.
+//
+// The deal.II library is free software; you can use it, redistribute
+// it, and/or modify it under the terms of the GNU Lesser General
+// Public License as published by the Free Software Foundation; either
+// version 2.1 of the License, or (at your option) any later version.
+// The full text of the license can be found in the file LICENSE.md at
+// the top level directory of deal.II.
+//
+// ---------------------------------------------------------------------
+
+
+
+// Test that DofRenumbering::hierarchical() gives the same DoF numbers
+// independent of their previous ordering. This was broken before where
+// the hierarchical numbering on each processor re-used the indices
+// the processor previously owned. This worked just fine with the
+// original ordering used by default by DoFHandler, but not if there
+// was another reordering step in between.
+//
+// This test is like the _04 test, but for dim!=spacedim
+
+
+#include <deal.II/base/quadrature_lib.h>
+#include <deal.II/base/tensor.h>
+
+#include <deal.II/distributed/tria.h>
+
+#include <deal.II/dofs/dof_accessor.h>
+#include <deal.II/dofs/dof_handler.h>
+#include <deal.II/dofs/dof_renumbering.h>
+#include <deal.II/dofs/dof_tools.h>
+
+#include <deal.II/fe/fe_q.h>
+#include <deal.II/fe/fe_values.h>
+
+#include <deal.II/grid/grid_generator.h>
+#include <deal.II/grid/grid_out.h>
+#include <deal.II/grid/tria.h>
+#include <deal.II/grid/tria_accessor.h>
+#include <deal.II/grid/tria_iterator.h>
+
+#include <deal.II/lac/trilinos_vector.h>
+
+#include <boost/concept_check.hpp>
+
+#include "../tests.h"
+
+
+void
+test()
+{
+  unsigned int myid = Utilities::MPI::this_mpi_process(MPI_COMM_WORLD);
+
+  parallel::distributed::Triangulation<2, 3> tr(MPI_COMM_WORLD);
+  GridGenerator::subdivided_hyper_rectangle(tr,
+                                            std::vector<unsigned int>{{2, 2}},
+                                            Point<2>(),
+                                            Point<2>(2, 2));
+
+  const FE_Q<2, 3> fe(1);
+
+  for (unsigned int test = 0; test < 2; ++test)
+    {
+      DoFHandler<2, 3> dof_handler(tr);
+      dof_handler.distribute_dofs(fe);
+
+      // in the second test run, revert the global order of DoF
+      // indices. the point is to make sure that processors do not
+      // have strictly increasing, contiguous groups of DoF indices.
+      if (test == 1)
+        {
+          IndexSet locally_owned_dofs = dof_handler.locally_owned_dofs();
+          std::vector<types::global_dof_index> new_numbers(
+            locally_owned_dofs.n_elements());
+          for (auto i : locally_owned_dofs)
+            new_numbers[locally_owned_dofs.index_within_set(i)] =
+              dof_handler.n_dofs() - i - 1;
+          dof_handler.renumber_dofs(new_numbers);
+        }
+
+      // next, let the hierarchical numbering work on it. the
+      // important aspect is that the result of hierarchical
+      // renumbering should not depend on the *previous* numbering. so
+      // we should be getting the same output as in the first run
+      DoFRenumbering::hierarchical(dof_handler);
+
+      // output DoF indices
+      deallog << (test == 0 ? "Without " : "With ")
+              << "prior reordering:" << std::endl;
+      const unsigned int                     dofs_per_cell = fe.dofs_per_cell;
+      std::vector<types::global_dof_index>   local_dof_indices(dofs_per_cell);
+      DoFHandler<2, 3>::active_cell_iterator cell = dof_handler.begin_active(),
+                                             endc = dof_handler.end();
+      for (; cell != endc; ++cell)
+        if (cell->subdomain_id() == tr.locally_owned_subdomain())
+          {
+            deallog << "Cell=" << cell << std::endl;
+            cell->get_dof_indices(local_dof_indices);
+            for (auto i : local_dof_indices)
+              deallog << i << ' ';
+            deallog << std::endl;
+          }
+    }
+}
+
+
+int
+main(int argc, char *argv[])
+{
+  Utilities::MPI::MPI_InitFinalize mpi_initialization(argc, argv, 1);
+  MPILogInitAll                    log;
+
+  test();
+}

diff --git a/tests/mpi/renumber_z_order_05.with_p4est=true.mpirun=4.output b/tests/mpi/renumber_z_order_05.with_p4est=true.mpirun=4.output
new file mode 100644 (file)
index 0000000..e917cb3
--- /dev/null
+++ b/tests/mpi/renumber_z_order_05.with_p4est=true.mpirun=4.output
@@ -0,0 +1,31 @@
+
+DEAL:0::Without prior reordering:
+DEAL:0::Cell=0.0
+DEAL:0::0 1 2 3 
+DEAL:0::With prior reordering:
+DEAL:0::Cell=0.0
+DEAL:0::0 1 2 3 
+
+DEAL:1::Without prior reordering:
+DEAL:1::Cell=0.1
+DEAL:1::1 4 3 5 
+DEAL:1::With prior reordering:
+DEAL:1::Cell=0.1
+DEAL:1::1 4 3 5 
+
+
+DEAL:2::Without prior reordering:
+DEAL:2::Cell=0.2
+DEAL:2::2 3 6 7 
+DEAL:2::With prior reordering:
+DEAL:2::Cell=0.2
+DEAL:2::2 3 6 7 
+
+
+DEAL:3::Without prior reordering:
+DEAL:3::Cell=0.3
+DEAL:3::3 5 7 8 
+DEAL:3::With prior reordering:
+DEAL:3::Cell=0.3
+DEAL:3::3 5 7 8 
+

diff --git a/tests/mpi/renumber_z_order_06.cc b/tests/mpi/renumber_z_order_06.cc
new file mode 100644 (file)
index 0000000..a9e8661
--- /dev/null
+++ b/tests/mpi/renumber_z_order_06.cc
@@ -0,0 +1,123 @@
+// ---------------------------------------------------------------------
+//
+// Copyright (C) 2017 - 2018 by the deal.II authors
+//
+// This file is part of the deal.II library.
+//
+// The deal.II library is free software; you can use it, redistribute
+// it, and/or modify it under the terms of the GNU Lesser General
+// Public License as published by the Free Software Foundation; either
+// version 2.1 of the License, or (at your option) any later version.
+// The full text of the license can be found in the file LICENSE.md at
+// the top level directory of deal.II.
+//
+// ---------------------------------------------------------------------
+
+
+
+// Test that DofRenumbering::hierarchical() gives the same DoF numbers
+// independent of their previous ordering. This was broken before where
+// the hierarchical numbering on each processor re-used the indices
+// the processor previously owned. This worked just fine with the
+// original ordering used by default by DoFHandler, but not if there
+// was another reordering step in between.
+//
+// This test is like the _04 test, but for dim!=spacedim and using an
+// hp::DoFHandler with a single element.
+
+
+#include <deal.II/base/quadrature_lib.h>
+#include <deal.II/base/tensor.h>
+
+#include <deal.II/distributed/tria.h>
+
+#include <deal.II/dofs/dof_accessor.h>
+#include <deal.II/dofs/dof_renumbering.h>
+#include <deal.II/dofs/dof_tools.h>
+
+#include <deal.II/fe/fe_q.h>
+#include <deal.II/fe/fe_values.h>
+
+#include <deal.II/grid/grid_generator.h>
+#include <deal.II/grid/grid_out.h>
+#include <deal.II/grid/tria.h>
+#include <deal.II/grid/tria_accessor.h>
+#include <deal.II/grid/tria_iterator.h>
+
+#include <deal.II/hp/dof_handler.h>
+
+#include <deal.II/lac/trilinos_vector.h>
+
+#include <boost/concept_check.hpp>
+
+#include "../tests.h"
+
+
+void
+test()
+{
+  unsigned int myid = Utilities::MPI::this_mpi_process(MPI_COMM_WORLD);
+
+  parallel::distributed::Triangulation<2, 3> tr(MPI_COMM_WORLD);
+  GridGenerator::subdivided_hyper_rectangle(tr,
+                                            std::vector<unsigned int>{{2, 2}},
+                                            Point<2>(),
+                                            Point<2>(2, 2));
+
+  hp::FECollection<2, 3> fe;
+  fe.push_back(FE_Q<2, 3>(1));
+
+  for (unsigned int test = 0; test < 2; ++test)
+    {
+      hp::DoFHandler<2, 3> dof_handler(tr);
+      dof_handler.distribute_dofs(fe);
+
+      // in the second test run, revert the global order of DoF
+      // indices. the point is to make sure that processors do not
+      // have strictly increasing, contiguous groups of DoF indices.
+      if (test == 1)
+        {
+          IndexSet locally_owned_dofs = dof_handler.locally_owned_dofs();
+          std::vector<types::global_dof_index> new_numbers(
+            locally_owned_dofs.n_elements());
+          for (auto i : locally_owned_dofs)
+            new_numbers[locally_owned_dofs.index_within_set(i)] =
+              dof_handler.n_dofs() - i - 1;
+          dof_handler.renumber_dofs(new_numbers);
+        }
+
+      // next, let the hierarchical numbering work on it. the
+      // important aspect is that the result of hierarchical
+      // renumbering should not depend on the *previous* numbering. so
+      // we should be getting the same output as in the first run
+      DoFRenumbering::hierarchical(dof_handler);
+
+      // output DoF indices
+      deallog << (test == 0 ? "Without " : "With ")
+              << "prior reordering:" << std::endl;
+      const unsigned int                   dofs_per_cell = fe[0].dofs_per_cell;
+      std::vector<types::global_dof_index> local_dof_indices(dofs_per_cell);
+      hp::DoFHandler<2, 3>::active_cell_iterator cell =
+                                                   dof_handler.begin_active(),
+                                                 endc = dof_handler.end();
+      for (; cell != endc; ++cell)
+        if (cell->subdomain_id() == tr.locally_owned_subdomain())
+          {
+            deallog << "Cell=" << cell << std::endl;
+            cell->get_dof_indices(local_dof_indices);
+            for (auto i : local_dof_indices)
+              deallog << i << ' ';
+            deallog << std::endl;
+          }
+    }
+}
+
+
+int
+main(int argc, char *argv[])
+{
+  Utilities::MPI::MPI_InitFinalize mpi_initialization(argc, argv, 1);
+  MPILogInitAll                    log;
+
+  test();
+}

diff --git a/tests/mpi/renumber_z_order_06.with_p4est=true.mpirun=4.output b/tests/mpi/renumber_z_order_06.with_p4est=true.mpirun=4.output
new file mode 100644 (file)
index 0000000..e917cb3
--- /dev/null
+++ b/tests/mpi/renumber_z_order_06.with_p4est=true.mpirun=4.output
@@ -0,0 +1,31 @@
+
+DEAL:0::Without prior reordering:
+DEAL:0::Cell=0.0
+DEAL:0::0 1 2 3 
+DEAL:0::With prior reordering:
+DEAL:0::Cell=0.0
+DEAL:0::0 1 2 3 
+
+DEAL:1::Without prior reordering:
+DEAL:1::Cell=0.1
+DEAL:1::1 4 3 5 
+DEAL:1::With prior reordering:
+DEAL:1::Cell=0.1
+DEAL:1::1 4 3 5 
+
+
+DEAL:2::Without prior reordering:
+DEAL:2::Cell=0.2
+DEAL:2::2 3 6 7 
+DEAL:2::With prior reordering:
+DEAL:2::Cell=0.2
+DEAL:2::2 3 6 7 
+
+
+DEAL:3::Without prior reordering:
+DEAL:3::Cell=0.3
+DEAL:3::3 5 7 8 
+DEAL:3::With prior reordering:
+DEAL:3::Cell=0.3
+DEAL:3::3 5 7 8 
+