poissonSolverProblemDevice.h

Go to the documentation of this file.

// ---------------------------------------------------------------------
//
// Copyright (c) 2017-2025  The Regents of the University of Michigan and DFT-FE
// authors.
//
// This file is part of the DFT-FE code.
//
// The DFT-FE code 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 at
// the top level of the DFT-FE distribution.
//
// ---------------------------------------------------------------------
//
 
#if defined(DFTFE_WITH_DEVICE)
#  ifndef poissonSolverProblemDevice_H_
#    define poissonSolverProblemDevice_H_
 
#    include <linearSolverProblemDevice.h>
#    include <constraintMatrixInfo.h>
#    include <constants.h>
#    include <dftUtils.h>
#    include <headers.h>
#    include "FEBasisOperations.h"
#    include "BLASWrapper.h"
 
namespace dftfe
{
  /**
   * @brief poisson solver problem device class template. template parameter FEOrderElectro
   * is the finite element polynomial order. The class should not be used with
   * FLOATING NUCLEAR CHARGES = false or POINT WISE DIRICHLET CONSTRAINT = true
   *
   * @author Gourab Panigrahi
   */
  template <dftfe::uInt FEOrderElectro>
  class poissonSolverProblemDevice : public linearSolverProblemDevice
  {
  public:
    /// Constructor
    poissonSolverProblemDevice(const MPI_Comm &mpi_comm);
 
    /**
     * @brief clears all datamembers and reset to original state.
     *
     *
     */
    void
    clear();
 
    /**
     * @brief reinitialize data structures for total electrostatic potential solve.
     *
     * For Hartree electrostatic potential solve give an empty map to the atoms
     * parameter.
     *
     */
    void
    reinit(
      const std::shared_ptr<
        dftfe::basis::
          FEBasisOperations<double, double, dftfe::utils::MemorySpace::HOST>>
                                              &basisOperationsPtr,
      distributedCPUVec<double>               &x,
      const dealii::AffineConstraints<double> &constraintMatrix,
      const dftfe::uInt                        matrixFreeVectorComponent,
      const dftfe::uInt matrixFreeQuadratureComponentRhsDensity,
      const dftfe::uInt matrixFreeQuadratureComponentAX,
      const std::map<dealii::types::global_dof_index, double> &atoms,
      const std::map<dealii::CellId, std::vector<double>> &smearedChargeValues,
      const dftfe::uInt smearedChargeQuadratureId,
      const dftfe::utils::MemoryStorage<double, dftfe::utils::MemorySpace::HOST>
        &rhoValues,
      const std::shared_ptr<
        dftfe::linearAlgebra::BLASWrapper<dftfe::utils::MemorySpace::DEVICE>>
                        BLASWrapperPtr,
      const bool        isComputeDiagonalA               = true,
      const bool        isComputeMeanValueConstraints    = false,
      const bool        smearedNuclearCharges            = false,
      const bool        isRhoValues                      = true,
      const bool        isGradSmearedChargeRhs           = false,
      const dftfe::uInt smearedChargeGradientComponentId = 0,
      const bool        storeSmearedChargeRhs            = false,
      const bool        reuseSmearedChargeRhs            = false,
      const bool        reinitializeFastConstraints      = false);
 
    /**
     * @brief Compute A matrix multipled by x.
     *
     */
    void
    computeAX(distributedDeviceVec<double> &Ax,
              distributedDeviceVec<double> &x);
 
    /**
     * @brief Compute right hand side vector for the problem Ax = rhs.
     *
     * @param rhs vector for the right hand side values
     */
    void
    computeRhs(distributedCPUVec<double> &rhs);
 
    /**
     * @brief get the reference to x field
     *
     * @return reference to x field. Assumes x field data structure is already initialized
     */
    distributedDeviceVec<double> &
    getX();
 
    /**
     * @brief get the reference to Preconditioner
     *
     * @return reference to Preconditioner
     */
    distributedDeviceVec<double> &
    getPreconditioner();
 
    /**
     * @brief Copies x from Device to Host
     *
     */
    void
    copyXfromDeviceToHost();
 
    /**
     * @brief distribute x to the constrained nodes.
     *
     */
    void
    distributeX();
 
 
    void
    setX();
 
 
  private:
    /**
     * @brief Sets up the matrixfree shapefunction, gradient, jacobian and map for matrixfree computeAX
     *
     */
    void
    setupMatrixFree();
 
    /**
     * @brief Sets up the constraints matrix
     *
     */
    void
    setupConstraints();
 
    /**
     * @brief Compute the diagonal of A.
     *
     */
    void
    computeDiagonalA();
 
    /**
     * @brief Compute mean value constraint which is required in case of fully periodic
     * boundary conditions.
     *
     */
    void
    computeMeanValueConstraint();
 
    /**
     * @brief Mean value constraint distibute
     *
     */
    void
    meanValueConstraintDistribute(distributedDeviceVec<double> &vec) const;
 
    /**
     * @brief Mean value constraint distibute slave to master
     *
     */
    void
    meanValueConstraintDistributeSlaveToMaster(
      distributedDeviceVec<double> &vec) const;
 
    void
    meanValueConstraintDistributeSlaveToMaster(
      distributedCPUVec<double> &vec) const;
 
    /**
     * @brief Mean value constraint set zero
     *
     */
    void
    meanValueConstraintSetZero(distributedDeviceVec<double> &vec) const;
 
    void
    meanValueConstraintSetZero(distributedCPUVec<double> &vec) const;
 
    /// storage for diagonal of the A matrix
    distributedCPUVec<double>    d_diagonalA;
    distributedDeviceVec<double> d_diagonalAdevice;
 
    /// storage for smeared charge rhs in case of total potential solve (doesn't
    /// change every scf)
    distributedCPUVec<double> d_rhsSmearedCharge;
 
    /// pointer to dealii MatrixFree object
    const dealii::MatrixFree<3, double> *d_matrixFreeDataPtr;
 
    /// pointer to the x vector being solved for
    distributedCPUVec<double>   *d_xPtr;
    distributedDeviceVec<double> d_xDevice;
 
    // number of cells local to each mpi task, number of degrees of freedom
    // locally owned and total degrees of freedom including ghost
    dftfe::Int d_nLocalCells, d_xLocalDof, d_xLen;
 
    // shape function value, gradient, jacobian and map for matrixfree
    dftfe::utils::MemoryStorage<double, dftfe::utils::MemorySpace::DEVICE>
      d_shapeFunction, d_jacobianFactor;
    dftfe::utils::MemoryStorage<dftfe::Int, dftfe::utils::MemorySpace::DEVICE>
      d_map;
 
    // Pointers to shape function value, gradient, jacobian and map for
    // matrixfree
    double     *d_shapeFunctionPtr;
    double     *d_jacobianFactorPtr;
    dftfe::Int *d_mapPtr;
 
 
    // constraints
    dftUtils::constraintMatrixInfo<dftfe::utils::MemorySpace::DEVICE>
      d_constraintsTotalPotentialInfo,
      d_inhomogenousConstraintsTotalPotentialInfo;
 
    /// pointer to dealii dealii::AffineConstraints<double> object
    const dealii::AffineConstraints<double> *d_constraintMatrixPtr;
 
    /// matrix free index required to access the DofHandler and
    /// dealii::AffineConstraints<double> objects corresponding to the problem
    dftfe::uInt d_matrixFreeVectorComponent;
 
    /// matrix free quadrature index
    dftfe::uInt d_matrixFreeQuadratureComponentRhsDensity;
 
    /// matrix free quadrature index
    dftfe::uInt d_matrixFreeQuadratureComponentAX;
 
    /// pointer to electron density cell quadrature data
    const dftfe::utils::MemoryStorage<double, dftfe::utils::MemorySpace::HOST>
      *d_rhoValuesPtr;
 
    /// pointer to smeared charge cell quadrature data
    const std::map<dealii::CellId, std::vector<double>>
      *d_smearedChargeValuesPtr;
 
    ///
    dftfe::uInt d_smearedChargeQuadratureId;
 
    /// pointer to map between global dof index in current processor and the
    /// atomic charge on that dof
    const std::map<dealii::types::global_dof_index, double> *d_atomsPtr;
 
    /// shape function gradient integral storage
    std::vector<double> d_cellShapeFunctionGradientIntegralFlattened;
 
    /// storage for mean value constraint vector
    distributedCPUVec<double> d_meanValueConstraintVec;
 
    /// storage for mean value constraint device vector
    distributedDeviceVec<double> d_meanValueConstraintDeviceVec;
 
    /// boolean flag to query if mean value constraint datastructures are
    /// precomputed
    bool d_isMeanValueConstraintComputed;
 
    ///
    bool d_isGradSmearedChargeRhs;
 
    ///
    bool d_isStoreSmearedChargeRhs;
 
    ///
    bool d_isReuseSmearedChargeRhs;
 
    ///
    dftfe::uInt d_smearedChargeGradientComponentId;
 
    /// mean value constraints: mean value constrained node
    dealii::types::global_dof_index d_meanValueConstraintNodeId;
 
    /// mean value constrained node local id
    dealii::types::global_dof_index d_meanValueConstraintNodeIdLocal;
 
    /// mean value constraints: constrained proc id containing the mean value
    /// constrained node
    dftfe::uInt d_meanValueConstraintProcId;
 
    /// duplicate constraints object with flattened maps for faster access
    dftUtils::constraintMatrixInfo<dftfe::utils::MemorySpace::HOST>
      d_constraintsInfo;
    std::shared_ptr<
      dftfe::basis::
        FEBasisOperations<double, double, dftfe::utils::MemorySpace::HOST>>
      d_basisOperationsPtr;
    ///
    std::shared_ptr<
      dftfe::linearAlgebra::BLASWrapper<dftfe::utils::MemorySpace::DEVICE>>
         d_BLASWrapperPtr;
    bool d_isFastConstraintsInitialized;
    bool d_isHomogenousConstraintsInitialized;
 
    const MPI_Comm             mpi_communicator;
    const dftfe::uInt          n_mpi_processes;
    const dftfe::uInt          this_mpi_process;
    dealii::ConditionalOStream pcout;
  };
 
} // namespace dftfe
#  endif // poissonSolverProblemDevice_H_
#endif