Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add constrained entries to partitioner_coarse in NonNestedMG #15812

Merged
Merged
Show file tree
Hide file tree
Changes from 2 commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
Original file line number Diff line number Diff line change
Expand Up @@ -2224,6 +2224,18 @@ namespace internal
compress_weights(transfer);
}
}


template <int dim, typename Number>
friend void
MGTwoLevelTransferNonNested<dim,
LinearAlgebra::distributed::Vector<Number>>::
reinit(const DoFHandler<dim> & dof_handler_fine,
const DoFHandler<dim> & dof_handler_coarse,
const Mapping<dim> & mapping_fine,
const Mapping<dim> & mapping_coarse,
const AffineConstraints<Number> &constraint_fine,
const AffineConstraints<Number> &constraint_coarse);
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

? I guess this should be in the .h file?

Copy link
Contributor Author

@fdrmrc fdrmrc Jul 31, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The issue is that MGTwoLevelTransferImplementation::create_coarse_partitioner() is defined in .templates.h and is private. Some compilers are also not happy with the present PR, so I moved it to public members and removed the friend declaration. Do you agree with that? I don't know what would be the best option

};


Expand Down Expand Up @@ -3839,12 +3851,9 @@ MGTwoLevelTransferNonNested<dim, LinearAlgebra::distributed::Vector<Number>>::

// create partitioners and internal vectors
{
IndexSet locally_active_dofs =
DoFTools::extract_locally_active_dofs(dof_handler_coarse);
this->partitioner_coarse.reset(
new Utilities::MPI::Partitioner(dof_handler_coarse.locally_owned_dofs(),
locally_active_dofs,
dof_handler_coarse.get_communicator()));
this->partitioner_coarse =
internal::MGTwoLevelTransferImplementation::create_coarse_partitioner(
dof_handler_coarse, constraint_coarse, numbers::invalid_unsigned_int);

this->vec_coarse.reinit(this->partitioner_coarse);
}
Expand Down
209 changes: 209 additions & 0 deletions tests/multigrid-global-coarsening/non_nested_multigrid_04.cc
Original file line number Diff line number Diff line change
@@ -0,0 +1,209 @@
// ---------------------------------------------------------------------
//
// Copyright (C) 2020 - 2023 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.
//
// ---------------------------------------------------------------------


/**
* Check that everything works also for locally-refined grids.
*
*/

#include <deal.II/base/conditional_ostream.h>
#include <deal.II/base/logstream.h>
#include <deal.II/base/mpi.h>
#include <deal.II/base/quadrature_lib.h>

#include <deal.II/distributed/tria.h>

#include <deal.II/dofs/dof_handler.h>

#include <deal.II/fe/fe_dgq.h>
#include <deal.II/fe/fe_q.h>
#include <deal.II/fe/fe_tools.h>
#include <deal.II/fe/mapping_q.h>

#include <deal.II/grid/grid_generator.h>
#include <deal.II/grid/grid_out.h>
#include <deal.II/grid/grid_tools.h>

#include <deal.II/matrix_free/fe_evaluation.h>
#include <deal.II/matrix_free/matrix_free.h>

#include <deal.II/multigrid/mg_constrained_dofs.h>
#include <deal.II/multigrid/mg_transfer_global_coarsening.h>

#include "multigrid_util.h"


template <int dim>
void
create_quadrant(Triangulation<dim> &tria, const unsigned int n_refinements)
{
// Taken from @cite munch2022gc, according to the description in A FLEXIBLE,
// PARALLEL, ADAPTIVE GEOMETRIC MULTIGRID METHOD FOR FEM (Clevenger, Heister,
// Kanschat, Kronbichler): https://arxiv.org/pdf/1904.03317.pdf

GridGenerator::hyper_cube(tria, -1.0, +1.0);

if (n_refinements == 0)
return;

tria.refine_global(1);

for (unsigned int i = 1; i < n_refinements; i++)
{
for (auto cell : tria.active_cell_iterators())
if (cell->is_locally_owned())
{
bool flag = true;
for (int d = 0; d < dim; d++)
if (cell->center()[d] > 0.0)
flag = false;
if (flag)
cell->set_refine_flag();
}
tria.execute_coarsening_and_refinement();
}

AssertDimension(tria.n_global_levels() - 1, n_refinements);
}


template <int dim, typename Number = double>
void
test(const unsigned int n_refinements,
const unsigned int fe_degree_fine,
const bool do_simplex_mesh)
{
using VectorType = LinearAlgebra::distributed::Vector<Number>;

const unsigned int min_level = 0;
const unsigned int max_level = n_refinements;

MGLevelObject<std::shared_ptr<parallel::distributed::Triangulation<dim>>>
triangulations(min_level, max_level);
MGLevelObject<DoFHandler<dim>> dof_handlers(min_level, max_level);
MGLevelObject<AffineConstraints<Number>> constraints(min_level, max_level);
MGLevelObject<std::unique_ptr<Mapping<dim>>> mappings(min_level, max_level);
MGLevelObject<std::shared_ptr<MGTwoLevelTransferNonNested<dim, VectorType>>>
transfers(min_level, max_level);
MGLevelObject<Operator<dim, Number>> operators(min_level, max_level);


// set up levels
for (auto l = min_level; l <= max_level; ++l)
{
triangulations[l] =
std::make_shared<parallel::distributed::Triangulation<dim>>(
MPI_COMM_WORLD);
auto &dof_handler = dof_handlers[l];
auto &constraint = constraints[l];
auto &op = operators[l];

std::unique_ptr<FiniteElement<dim>> fe;
std::unique_ptr<Quadrature<dim>> quad;
std::unique_ptr<Mapping<dim>> mapping;

if (do_simplex_mesh)
{
fe = std::make_unique<FE_SimplexP<dim>>(fe_degree_fine);
quad = std::make_unique<QGaussSimplex<dim>>(fe_degree_fine + 1);
mapping = std::make_unique<MappingFE<dim>>(FE_SimplexP<dim>(1));
}
else
{
fe = std::make_unique<FE_Q<dim>>(fe_degree_fine);
quad = std::make_unique<QGauss<dim>>(fe_degree_fine + 1);
mapping = std::make_unique<MappingQ<dim>>(1);
}

mappings[l] = mapping->clone();

// set up triangulation

create_quadrant(*triangulations[l], 2 * l + 1);

// set up dofhandler
dof_handler.reinit(*triangulations[l]);
dof_handler.distribute_dofs(*fe);

// set up constraints
IndexSet locally_relevant_dofs;
DoFTools::extract_locally_relevant_dofs(dof_handler,
locally_relevant_dofs);
constraint.reinit(locally_relevant_dofs);
VectorTools::interpolate_boundary_values(
*mapping, dof_handler, 0, Functions::ZeroFunction<dim>(), constraint);
DoFTools::make_hanging_node_constraints(dof_handler, constraint);
constraint.close();

// set up operator
op.reinit(*mapping, dof_handler, *quad, constraint);
}

// set up transfer operator
for (unsigned int l = min_level; l < max_level; ++l)
{
transfers[l + 1] =
std::make_shared<MGTwoLevelTransferNonNested<dim, VectorType>>();
transfers[l + 1]->reinit(dof_handlers[l + 1],
dof_handlers[l],
*mappings[l + 1],
*mappings[l],
constraints[l + 1],
constraints[l]);
}

MGTransferGlobalCoarsening<dim, VectorType> transfer(
transfers,
[&](const auto l, auto &vec) { operators[l].initialize_dof_vector(vec); });


GMGParameters mg_data; // TODO

VectorType dst, src;
operators[max_level].initialize_dof_vector(dst);
operators[max_level].initialize_dof_vector(src);

operators[max_level].rhs(src);

ReductionControl solver_control(
mg_data.maxiter, mg_data.abstol, mg_data.reltol, false, false);

mg_solve(solver_control,
dst,
src,
mg_data,
dof_handlers[max_level],
operators[max_level],
operators,
transfer);

deallog << dim << ' ' << fe_degree_fine << ' ' << n_refinements << ' '
<< (do_simplex_mesh ? "tri " : "quad") << ' '
<< solver_control.last_step() << std::endl;
}

int
main(int argc, char **argv)
{
Utilities::MPI::MPI_InitFinalize mpi_initialization(argc, argv, 1);
MPILogInitAll all;

deallog.precision(8);

const unsigned int n_refinements = 3;
const unsigned int degree = 4;
test<2>(n_refinements, degree, false /*quadrilateral*/);
}
Original file line number Diff line number Diff line change
@@ -0,0 +1,2 @@

DEAL:0::2 4 3 quad 3
Original file line number Diff line number Diff line change
@@ -0,0 +1,11 @@

DEAL:0::2 4 3 quad 3

DEAL:1::2 4 3 quad 3


DEAL:2::2 4 3 quad 3


DEAL:3::2 4 3 quad 3