seqsolverbackend.hh

Go to the documentation of this file.

// -*- mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4 -*-
// vi: set et ts=4 sw=4 sts=4:
//
// SPDX-FileCopyrightInfo: Copyright © DuMux Project contributors, see AUTHORS.md in root folder
// SPDX-License-Identifier: GPL-3.0-or-later
//
#ifndef DUMUX_SEQ_SOLVER_BACKEND_HH
#define DUMUX_SEQ_SOLVER_BACKEND_HH
 
#include <type_traits>
#include <tuple>
#include <utility>
 
#include <dune/istl/preconditioners.hh>
#include <dune/istl/solvers.hh>
#include <dune/istl/io.hh>
#include <dune/common/indices.hh>
#include <dune/common/hybridutilities.hh>
 
#include <dumux/common/parameters.hh>
#include <dumux/common/typetraits/matrix.hh>
#include <dumux/common/typetraits/utility.hh>
#include <dumux/linear/solver.hh>
#include <dumux/linear/preconditioners.hh>
#include <dumux/linear/linearsolverparameters.hh>
#include <dumux/linear/parallelmatrixadapter.hh>
 
namespace Dumux {
 
class IterativePreconditionedSolverImpl
{
public:
 
    // solve with generic parameter tree
    template<class Preconditioner, class Solver, class Matrix, class Vector>
    static bool solveWithParamTree(const Matrix& A, Vector& x, const Vector& b,
                                   const Dune::ParameterTree& params)
    {
        // make a linear operator from a matrix
        using MatrixAdapter = Dune::MatrixAdapter<Matrix, Vector, Vector>;
        const auto linearOperator = std::make_shared<MatrixAdapter>(A);
 
        auto precond = std::make_shared<Preconditioner>(linearOperator, params.sub("preconditioner"));
        Solver solver(linearOperator, precond, params);
 
        Vector bTmp(b);
 
        Dune::InverseOperatorResult result;
        solver.apply(x, bTmp, result);
 
        return result.converged;
    }
};
 
template<class M>
constexpr std::size_t preconditionerBlockLevel() noexcept
{
    return isMultiTypeBlockMatrix<M>::value ? 2 : 1;
}
 
class ExplicitDiagonalSolver : public LinearSolver
{
public:
    using LinearSolver::LinearSolver;
 
    template<class Matrix, class Vector>
    bool solve(const Matrix& A, Vector& x, const Vector& b)
    {
        Vector rhs(b);
        constexpr auto precondBlockLevel = preconditionerBlockLevel<Matrix>();
        Dune::SeqJac<Matrix, Vector, Vector, precondBlockLevel> jac(A, 1, 1.0);
        jac.pre(x, rhs);
        jac.apply(x, rhs);
        jac.post(x);
        return true;
    }
 
    std::string name() const
    {
        return "Explicit diagonal matrix solver";
    }
};
 
// \{
 
template <class LinearSolverTraits>
class UzawaBiCGSTABBackend : public LinearSolver
{
public:
    using LinearSolver::LinearSolver;
 
    template<class Matrix, class Vector>
    bool solve(const Matrix& A, Vector& x, const Vector& b)
    {
        using Preconditioner = SeqUzawa<Matrix, Vector, Vector>;
        using Solver = Dune::BiCGSTABSolver<Vector>;
        static const auto solverParams = LinearSolverParameters<LinearSolverTraits>::createParameterTree(this->paramGroup());
        return IterativePreconditionedSolverImpl::template solveWithParamTree<Preconditioner, Solver>(A, x, b, solverParams);
    }
 
    std::string name() const
    {
        return "Uzawa preconditioned BiCGSTAB solver";
    }
};
 
template<class M, class X, class Y, int blockLevel = 2>
class BlockDiagILU0Preconditioner : public Dune::Preconditioner<X, Y>
{
    template<std::size_t i>
    using DiagBlockType = std::decay_t<decltype(std::declval<M>()[Dune::index_constant<i>{}][Dune::index_constant<i>{}])>;
 
    template<std::size_t i>
    using VecBlockType = std::decay_t<decltype(std::declval<X>()[Dune::index_constant<i>{}])>;
 
    template<std::size_t i>
    using BlockILU = Dune::SeqILU<DiagBlockType<i>, VecBlockType<i>, VecBlockType<i>, blockLevel-1>;
 
    using ILUTuple = typename makeFromIndexedType<std::tuple, BlockILU, std::make_index_sequence<M::N()> >::type;
 
public:
    using matrix_type = typename std::decay_t<M>;
    using domain_type = X;
    using range_type = Y;
    using field_type = typename X::field_type;
 
    BlockDiagILU0Preconditioner(const M& m, double w = 1.0)
    : BlockDiagILU0Preconditioner(m, w, std::make_index_sequence<M::N()>{})
    {
        static_assert(blockLevel >= 2, "Only makes sense for MultiTypeBlockMatrix!");
    }
 
    void pre (X& v, Y& d) final {}
 
    void apply (X& v, const Y& d) final
    {
        using namespace Dune::Hybrid;
        forEach(integralRange(Dune::Hybrid::size(ilu_)), [&](const auto i)
        {
            std::get<decltype(i)::value>(ilu_).apply(v[i], d[i]);
        });
    }
 
    void post (X&) final {}
 
    Dune::SolverCategory::Category category() const final
    {
        return Dune::SolverCategory::sequential;
    }
 
private:
    template<std::size_t... Is>
    BlockDiagILU0Preconditioner (const M& m, double w, std::index_sequence<Is...> is)
    : ilu_(std::make_tuple(BlockILU<Is>(m[Dune::index_constant<Is>{}][Dune::index_constant<Is>{}], w)...))
    {}
 
    ILUTuple ilu_;
};
 
 
class BlockDiagILU0BiCGSTABSolver : public LinearSolver
{
 
public:
    using LinearSolver::LinearSolver;
 
    template<class Matrix, class Vector>
    bool solve(const Matrix& M, Vector& x, const Vector& b)
    {
        BlockDiagILU0Preconditioner<Matrix, Vector, Vector> preconditioner(M);
        Dumux::ParallelMultiTypeMatrixAdapter<Matrix, Vector, Vector> op(M);
        Dune::BiCGSTABSolver<Vector> solver(op, preconditioner, this->residReduction(),
                                            this->maxIter(), this->verbosity());
        auto bTmp(b);
        solver.apply(x, bTmp, result_);
 
        return result_.converged;
    }
 
    const Dune::InverseOperatorResult& result() const
    {
      return result_;
    }
 
    std::string name() const
    { return "block-diagonal ILU0-preconditioned BiCGSTAB solver"; }
 
private:
    Dune::InverseOperatorResult result_;
};
 
class BlockDiagILU0RestartedGMResSolver : public LinearSolver
{
 
public:
    using LinearSolver::LinearSolver;
 
    template<int precondBlockLevel = 2, class Matrix, class Vector>
    bool solve(const Matrix& M, Vector& x, const Vector& b)
    {
        BlockDiagILU0Preconditioner<Matrix, Vector, Vector> preconditioner(M);
        Dumux::ParallelMultiTypeMatrixAdapter<Matrix, Vector, Vector> op(M);
        static const int restartGMRes = getParamFromGroup<int>(this->paramGroup(), "LinearSolver.GMResRestart");
        Dune::RestartedGMResSolver<Vector> solver(op, preconditioner, this->residReduction(), restartGMRes,
                                                  this->maxIter(), this->verbosity());
        auto bTmp(b);
        solver.apply(x, bTmp, result_);
 
        return result_.converged;
    }
 
    const Dune::InverseOperatorResult& result() const
    {
      return result_;
    }
 
    std::string name() const
    { return "block-diagonal ILU0-preconditioned restarted GMRes solver"; }
 
private:
    Dune::InverseOperatorResult result_;
};
 
template<class M, class X, class Y, int blockLevel = 2>
class BlockDiagAMGPreconditioner : public Dune::Preconditioner<X, Y>
{
    template<std::size_t i>
    using DiagBlockType = std::decay_t<decltype(std::declval<M>()[Dune::index_constant<i>{}][Dune::index_constant<i>{}])>;
 
    template<std::size_t i>
    using VecBlockType = std::decay_t<decltype(std::declval<X>()[Dune::index_constant<i>{}])>;
 
    template<std::size_t i>
    using Smoother = Dune::SeqSSOR<DiagBlockType<i>, VecBlockType<i>, VecBlockType<i>>;
 
    template<std::size_t i>
    using LinearOperator = Dune::MatrixAdapter<DiagBlockType<i>, VecBlockType<i>, VecBlockType<i>>;
 
    template<std::size_t i>
    using ScalarProduct = Dune::SeqScalarProduct<VecBlockType<i>>;
 
    template<std::size_t i>
    using BlockAMG = Dune::Amg::AMG<LinearOperator<i>, VecBlockType<i>, Smoother<i>>;
 
    using AMGTuple = typename makeFromIndexedType<std::tuple, BlockAMG, std::make_index_sequence<M::N()> >::type;
 
public:
    using matrix_type = typename std::decay_t<M>;
    using domain_type = X;
    using range_type = Y;
    using field_type = typename X::field_type;
 
    template<class LOP, class Criterion, class SmootherArgs>
    BlockDiagAMGPreconditioner(const LOP& lop, const Criterion& c, const SmootherArgs& sa)
    : BlockDiagAMGPreconditioner(lop, c, sa, std::make_index_sequence<M::N()>{})
    {
        static_assert(blockLevel >= 2, "Only makes sense for MultiTypeBlockMatrix!");
    }
 
    void pre (X& v, Y& d) final
    {
        using namespace Dune::Hybrid;
        forEach(integralRange(Dune::Hybrid::size(amg_)), [&](const auto i)
        {
            std::get<decltype(i)::value>(amg_).pre(v[i], d[i]);
        });
    }
 
    void apply (X& v, const Y& d) final
    {
        using namespace Dune::Hybrid;
        forEach(integralRange(Dune::Hybrid::size(amg_)), [&](const auto i)
        {
            std::get<decltype(i)::value>(amg_).apply(v[i], d[i]);
        });
    }
 
    void post (X& v) final
    {
        using namespace Dune::Hybrid;
        forEach(integralRange(Dune::Hybrid::size(amg_)), [&](const auto i)
        {
            std::get<decltype(i)::value>(amg_).post(v[i]);
        });
    }
 
    Dune::SolverCategory::Category category() const final
    {
        return Dune::SolverCategory::sequential;
    }
 
private:
    template<class LOP, class Criterion, class SmootherArgs, std::size_t... Is>
    BlockDiagAMGPreconditioner (const LOP& lop, const Criterion& c, const SmootherArgs& sa, std::index_sequence<Is...> is)
    : amg_(std::make_tuple(BlockAMG<Is>(*std::get<Is>(lop), *std::get<Is>(c), *std::get<Is>(sa))...))
    {}
 
    AMGTuple amg_;
};
 
class BlockDiagAMGBiCGSTABSolver : public LinearSolver
{
    template<class M, std::size_t i>
    using DiagBlockType = std::decay_t<decltype(std::declval<M>()[Dune::index_constant<i>{}][Dune::index_constant<i>{}])>;
 
    template<class X, std::size_t i>
    using VecBlockType = std::decay_t<decltype(std::declval<X>()[Dune::index_constant<i>{}])>;
 
    template<class M, class X, std::size_t i>
    using Smoother = Dune::SeqSSOR<DiagBlockType<M, i>, VecBlockType<X, i>, VecBlockType<X, i>>;
 
    template<class M, class X, std::size_t i>
    using SmootherArgs = typename Dune::Amg::SmootherTraits<Smoother<M, X, i>>::Arguments;
 
    template<class M, std::size_t i>
    using Criterion = Dune::Amg::CoarsenCriterion<Dune::Amg::SymmetricCriterion<DiagBlockType<M, i>, Dune::Amg::FirstDiagonal>>;
 
    template<class M, class X, std::size_t i>
    using LinearOperator = Dune::MatrixAdapter<DiagBlockType<M, i>, VecBlockType<X, i>, VecBlockType<X, i>>;
 
public:
    using LinearSolver::LinearSolver;
 
    // Solve saddle-point problem using a Schur complement based preconditioner
    template<class Matrix, class Vector>
    bool solve(const Matrix& m, Vector& x, const Vector& b)
    {
        Dune::Amg::Parameters params(15, 2000, 1.2, 1.6, Dune::Amg::atOnceAccu);
        params.setDefaultValuesIsotropic(3);
        params.setDebugLevel(this->verbosity());
 
        auto criterion = makeCriterion_<Criterion, Matrix>(params, std::make_index_sequence<Matrix::N()>{});
        auto smootherArgs = makeSmootherArgs_<SmootherArgs, Matrix, Vector>(std::make_index_sequence<Matrix::N()>{});
 
        using namespace Dune::Hybrid;
        forEach(std::make_index_sequence<Matrix::N()>{}, [&](const auto i)
        {
            auto& args = std::get<decltype(i)::value>(smootherArgs);
            args->iterations = 1;
            args->relaxationFactor = 1;
        });
 
        auto linearOperator = makeLinearOperator_<LinearOperator, Matrix, Vector>(m, std::make_index_sequence<Matrix::N()>{});
 
        BlockDiagAMGPreconditioner<Matrix, Vector, Vector> preconditioner(linearOperator, criterion, smootherArgs);
 
        Dumux::ParallelMultiTypeMatrixAdapter<Matrix, Vector, Vector> op(m);
        Dune::BiCGSTABSolver<Vector> solver(op, preconditioner, this->residReduction(),
                                            this->maxIter(), this->verbosity());
        auto bTmp(b);
        solver.apply(x, bTmp, result_);
 
        return result_.converged;
    }
 
    const Dune::InverseOperatorResult& result() const
    {
      return result_;
    }
 
    std::string name() const
    { return "block-diagonal AMG-preconditioned BiCGSTAB solver"; }
 
private:
    template<template<class M, std::size_t i> class Criterion, class Matrix, class Params, std::size_t... Is>
    auto makeCriterion_(const Params& p, std::index_sequence<Is...>)
    {
        return std::make_tuple(std::make_shared<Criterion<Matrix, Is>>(p)...);
    }
 
    template<template<class M, class X, std::size_t i> class SmootherArgs, class Matrix, class Vector, std::size_t... Is>
    auto makeSmootherArgs_(std::index_sequence<Is...>)
    {
        return std::make_tuple(std::make_shared<SmootherArgs<Matrix, Vector, Is>>()...);
    }
 
    template<template<class M, class X, std::size_t i> class LinearOperator, class Matrix, class Vector, std::size_t... Is>
    auto makeLinearOperator_(const Matrix& m, std::index_sequence<Is...>)
    {
        return std::make_tuple(std::make_shared<LinearOperator<Matrix, Vector, Is>>(m[Dune::index_constant<Is>{}][Dune::index_constant<Is>{}])...);
    }
 
    Dune::InverseOperatorResult result_;
};
 
// \}
 
} // end namespace Dumux
 
#endif