MFEM  v4.6.0
Finite element discretization library
Public Member Functions | Protected Types | Protected Member Functions | Protected Attributes | Static Protected Attributes | List of all members
mfem::SparseMatrix Class Reference

Data type sparse matrix. More...

#include <sparsemat.hpp>

Inheritance diagram for mfem::SparseMatrix:
[legend]
Collaboration diagram for mfem::SparseMatrix:
[legend]

Public Member Functions

 SparseMatrix ()
 Create an empty SparseMatrix. More...
 
 SparseMatrix (int nrows, int ncols=-1)
 Create a sparse matrix with flexible sparsity structure using a row-wise linked list (LIL) format. More...
 
 SparseMatrix (int *i, int *j, double *data, int m, int n)
 Create a sparse matrix in CSR format. Ownership of i, j, and data is transferred to the SparseMatrix. More...
 
 SparseMatrix (int *i, int *j, double *data, int m, int n, bool ownij, bool owna, bool issorted)
 Create a sparse matrix in CSR format. Ownership of i, j, and data is optionally transferred to the SparseMatrix. More...
 
 SparseMatrix (int nrows, int ncols, int rowsize)
 Create a sparse matrix in CSR format where each row has space allocated for exactly rowsize entries. More...
 
 SparseMatrix (const SparseMatrix &mat, bool copy_graph=true, MemoryType mt=MemoryType::PRESERVE)
 Copy constructor (deep copy). More...
 
 SparseMatrix (const Vector &v)
 Create a SparseMatrix with diagonal v, i.e. A = Diag(v) More...
 
void OverrideSize (int height_, int width_)
 Sets the height and width of the matrix. More...
 
void UseGPUSparse (bool useGPUSparse_=true)
 Runtime option to use cuSPARSE or hipSPARSE. Only valid when using a CUDA or HIP backend. More...
 
MFEM_DEPRECATED void UseCuSparse (bool useCuSparse_=true)
 Deprecated equivalent of UseGPUSparse(). More...
 
SparseMatrixoperator= (const SparseMatrix &rhs)
 Assignment operator: deep copy. More...
 
void MakeRef (const SparseMatrix &master)
 Clear the contents of the SparseMatrix and make it a reference to master. More...
 
int Size () const
 For backward compatibility, define Size() to be synonym of Height(). More...
 
void Clear ()
 Clear the contents of the SparseMatrix. More...
 
void ClearGPUSparse ()
 Clear the cuSPARSE/hipSPARSE descriptors. This must be called after releasing the device memory of A. More...
 
MFEM_DEPRECATED void ClearCuSparse ()
 Deprecated equivalent of ClearGPUSparse(). More...
 
bool Empty () const
 Check if the SparseMatrix is empty. More...
 
int * GetI ()
 Return the array I. More...
 
const int * GetI () const
 Return the array I, const version. More...
 
int * GetJ ()
 Return the array J. More...
 
const int * GetJ () const
 Return the array J, const version. More...
 
double * GetData ()
 Return the element data, i.e. the array A. More...
 
const double * GetData () const
 Return the element data, i.e. the array A, const version. More...
 
Memory< int > & GetMemoryI ()
 
const Memory< int > & GetMemoryI () const
 
const int * ReadI (bool on_dev=true) const
 
int * WriteI (bool on_dev=true)
 
int * ReadWriteI (bool on_dev=true)
 
const int * HostReadI () const
 
int * HostWriteI ()
 
int * HostReadWriteI ()
 
Memory< int > & GetMemoryJ ()
 
const Memory< int > & GetMemoryJ () const
 
const int * ReadJ (bool on_dev=true) const
 
int * WriteJ (bool on_dev=true)
 
int * ReadWriteJ (bool on_dev=true)
 
const int * HostReadJ () const
 
int * HostWriteJ ()
 
int * HostReadWriteJ ()
 
Memory< double > & GetMemoryData ()
 
const Memory< double > & GetMemoryData () const
 
const double * ReadData (bool on_dev=true) const
 
double * WriteData (bool on_dev=true)
 
double * ReadWriteData (bool on_dev=true)
 
const double * HostReadData () const
 
double * HostWriteData ()
 
double * HostReadWriteData ()
 
int RowSize (const int i) const
 Returns the number of elements in row i. More...
 
int MaxRowSize () const
 Returns the maximum number of elements among all rows. More...
 
int * GetRowColumns (const int row)
 Return a pointer to the column indices in a row. More...
 
const int * GetRowColumns (const int row) const
 Return a pointer to the column indices in a row, const version. More...
 
double * GetRowEntries (const int row)
 Return a pointer to the entries in a row. More...
 
const double * GetRowEntries (const int row) const
 Return a pointer to the entries in a row, const version. More...
 
void SetWidth (int width_=-1)
 Change the width of a SparseMatrix. More...
 
int ActualWidth () const
 Returns the actual Width of the matrix. More...
 
void SortColumnIndices ()
 Sort the column indices corresponding to each row. More...
 
void MoveDiagonalFirst ()
 Move the diagonal entry to the first position in each row, preserving the order of the rest of the columns. More...
 
virtual double & Elem (int i, int j)
 Returns reference to a_{ij}. More...
 
virtual const double & Elem (int i, int j) const
 Returns constant reference to a_{ij}. More...
 
double & operator() (int i, int j)
 Returns reference to A[i][j]. More...
 
const double & operator() (int i, int j) const
 Returns reference to A[i][j]. More...
 
void GetDiag (Vector &d) const
 Returns the Diagonal of A. More...
 
DenseMatrixToDenseMatrix () const
 Produces a DenseMatrix from a SparseMatrix. More...
 
void ToDenseMatrix (DenseMatrix &B) const
 Produces a DenseMatrix from a SparseMatrix. More...
 
virtual MemoryClass GetMemoryClass () const
 Return the MemoryClass preferred by the Operator. More...
 
virtual void Mult (const Vector &x, Vector &y) const
 Matrix vector multiplication. More...
 
virtual void AddMult (const Vector &x, Vector &y, const double a=1.0) const
 y += A * x (default) or y += a * A * x More...
 
virtual void MultTranspose (const Vector &x, Vector &y) const
 Multiply a vector with the transposed matrix. y = At * x. More...
 
virtual void AddMultTranspose (const Vector &x, Vector &y, const double a=1.0) const
 y += At * x (default) or y += a * At * x More...
 
void BuildTranspose () const
 Build and store internally the transpose of this matrix which will be used in the methods AddMultTranspose(), MultTranspose(), and AbsMultTranspose(). More...
 
void ResetTranspose () const
 
void EnsureMultTranspose () const
 Ensures that the matrix is capable of performing MultTranspose(), AddMultTranspose(), and AbsMultTranspose(). More...
 
void PartMult (const Array< int > &rows, const Vector &x, Vector &y) const
 
void PartAddMult (const Array< int > &rows, const Vector &x, Vector &y, const double a=1.0) const
 
void BooleanMult (const Array< int > &x, Array< int > &y) const
 y = A * x, treating all entries as booleans (zero=false, nonzero=true). More...
 
void BooleanMultTranspose (const Array< int > &x, Array< int > &y) const
 y = At * x, treating all entries as booleans (zero=false, nonzero=true). More...
 
void AbsMult (const Vector &x, Vector &y) const
 y = |A| * x, using entry-wise absolute values of matrix A More...
 
void AbsMultTranspose (const Vector &x, Vector &y) const
 y = |At| * x, using entry-wise absolute values of the transpose of matrix A More...
 
double InnerProduct (const Vector &x, const Vector &y) const
 Compute y^t A x. More...
 
void GetRowSums (Vector &x) const
 For all i compute \( x_i = \sum_j A_{ij} \). More...
 
double GetRowNorml1 (int irow) const
 For i = irow compute \( x_i = \sum_j | A_{i, j} | \). More...
 
virtual MatrixInverseInverse () const
 This virtual method is not supported: it always returns NULL. More...
 
void EliminateRow (int row, const double sol, Vector &rhs)
 Eliminates a column from the transpose matrix. More...
 
void EliminateRow (int row, DiagonalPolicy dpolicy=DIAG_ZERO)
 Eliminates a row from the matrix. More...
 
void EliminateCol (int col, DiagonalPolicy dpolicy=DIAG_ZERO)
 Eliminates the column col from the matrix. More...
 
void EliminateCols (const Array< int > &cols, const Vector *x=NULL, Vector *b=NULL)
 Eliminate all columns i for which cols[i] != 0. More...
 
void EliminateCols (const Array< int > &col_marker, SparseMatrix &Ae)
 Similar to EliminateCols + save the eliminated entries into Ae so that (*this) + Ae is equal to the original matrix. More...
 
void EliminateRowCol (int rc, const double sol, Vector &rhs, DiagonalPolicy dpolicy=DIAG_ONE)
 Eliminate row rc and column rc and modify the rhs using sol. More...
 
void EliminateRowColMultipleRHS (int rc, const Vector &sol, DenseMatrix &rhs, DiagonalPolicy dpolicy=DIAG_ONE)
 Similar to EliminateRowCol(int, const double, Vector &, DiagonalPolicy), but multiple values for eliminated unknowns are accepted, and accordingly multiple right-hand-sides are used. More...
 
void EliminateRowColDiag (int rc, double value)
 Perform elimination and set the diagonal entry to the given value. More...
 
void EliminateRowCol (int rc, DiagonalPolicy dpolicy=DIAG_ONE)
 Eliminate row rc and column rc. More...
 
void EliminateRowCol (int rc, SparseMatrix &Ae, DiagonalPolicy dpolicy=DIAG_ONE)
 Similar to EliminateRowCol(int, DiagonalPolicy) + save the eliminated entries into Ae so that (*this) + Ae is equal to the original matrix. More...
 
void EliminateBC (const Array< int > &ess_dofs, DiagonalPolicy diag_policy)
 Eliminate essential (Dirichlet) boundary conditions. More...
 
void SetDiagIdentity ()
 If a row contains only one diag entry of zero, set it to 1. More...
 
virtual void EliminateZeroRows (const double threshold=1e-12)
 If a row contains only zeros, set its diagonal to 1. More...
 
void Gauss_Seidel_forw (const Vector &x, Vector &y) const
 Gauss-Seidel forward and backward iterations over a vector x. More...
 
void Gauss_Seidel_back (const Vector &x, Vector &y) const
 
double GetJacobiScaling () const
 Determine appropriate scaling for Jacobi iteration. More...
 
void Jacobi (const Vector &b, const Vector &x0, Vector &x1, double sc, bool use_abs_diag=false) const
 
void DiagScale (const Vector &b, Vector &x, double sc=1.0, bool use_abs_diag=false) const
 x = sc b / A_ii. When use_abs_diag = true, |A_ii| is used. More...
 
void Jacobi2 (const Vector &b, const Vector &x0, Vector &x1, double sc=1.0) const
 
void Jacobi3 (const Vector &b, const Vector &x0, Vector &x1, double sc=1.0) const
 
virtual void Finalize (int skip_zeros=1)
 Finalize the matrix initialization, switching the storage format from LIL to CSR. More...
 
void Finalize (int skip_zeros, bool fix_empty_rows)
 A slightly more general version of the Finalize(int) method. More...
 
bool Finalized () const
 Returns whether or not CSR format has been finalized. More...
 
bool ColumnsAreSorted () const
 Returns whether or not the columns are sorted. More...
 
void Threshold (double tol, bool fix_empty_rows=false)
 Remove entries smaller in absolute value than a given tolerance tol. If fix_empty_rows is true, a zero value is inserted in the diagonal entry (for square matrices only) More...
 
void GetBlocks (Array2D< SparseMatrix *> &blocks) const
 
void GetSubMatrix (const Array< int > &rows, const Array< int > &cols, DenseMatrix &subm) const
 
void SetColPtr (const int row) const
 Initialize the SparseMatrix for fast access to the entries of the given row which becomes the "current row". More...
 
void ClearColPtr () const
 Reset the "current row" set by calling SetColPtr(). This method must be called between any two calls to SetColPtr(). More...
 
double & SearchRow (const int col)
 Perform a fast search for an entry in the "current row". See SetColPtr(). More...
 
void _Add_ (const int col, const double a)
 Add a value to an entry in the "current row". See SetColPtr(). More...
 
void _Set_ (const int col, const double a)
 Set an entry in the "current row". See SetColPtr(). More...
 
double _Get_ (const int col) const
 Read the value of an entry in the "current row". See SetColPtr(). More...
 
double & SearchRow (const int row, const int col)
 
void _Add_ (const int row, const int col, const double a)
 
void _Set_ (const int row, const int col, const double a)
 
void Set (const int i, const int j, const double val)
 
void Add (const int i, const int j, const double val)
 
void SetSubMatrix (const Array< int > &rows, const Array< int > &cols, const DenseMatrix &subm, int skip_zeros=1)
 
void SetSubMatrixTranspose (const Array< int > &rows, const Array< int > &cols, const DenseMatrix &subm, int skip_zeros=1)
 
void AddSubMatrix (const Array< int > &rows, const Array< int > &cols, const DenseMatrix &subm, int skip_zeros=1)
 
bool RowIsEmpty (const int row) const
 
virtual int GetRow (const int row, Array< int > &cols, Vector &srow) const
 Extract all column indices and values from a given row. More...
 
void SetRow (const int row, const Array< int > &cols, const Vector &srow)
 
void AddRow (const int row, const Array< int > &cols, const Vector &srow)
 
void ScaleRow (const int row, const double scale)
 
void ScaleRows (const Vector &sl)
 this = diag(sl) * this; More...
 
void ScaleColumns (const Vector &sr)
 this = this * diag(sr); More...
 
SparseMatrixoperator+= (const SparseMatrix &B)
 Add the sparse matrix 'B' to '*this'. This operation will cause an error if '*this' is finalized and 'B' has larger sparsity pattern. More...
 
void Add (const double a, const SparseMatrix &B)
 Add the sparse matrix 'B' scaled by the scalar 'a' into '*this'. Only entries in the sparsity pattern of '*this' are added. More...
 
SparseMatrixoperator= (double a)
 
SparseMatrixoperator*= (double a)
 
void Print (std::ostream &out=mfem::out, int width_=4) const
 Prints matrix to stream out. More...
 
virtual void PrintMatlab (std::ostream &out=mfem::out) const
 Prints matrix in matlab format. More...
 
void PrintMM (std::ostream &out=mfem::out) const
 Prints matrix in Matrix Market sparse format. More...
 
void PrintCSR (std::ostream &out) const
 Prints matrix to stream out in hypre_CSRMatrix format. More...
 
void PrintCSR2 (std::ostream &out) const
 Prints a sparse matrix to stream out in CSR format. More...
 
void PrintInfo (std::ostream &out) const
 Print various sparse matrix statistics. More...
 
double IsSymmetric () const
 Returns max_{i,j} |(i,j)-(j,i)| for a finalized matrix. More...
 
void Symmetrize ()
 (*this) = 1/2 ((*this) + (*this)^t) More...
 
virtual int NumNonZeroElems () const
 Returns the number of the nonzero elements in the matrix. More...
 
double MaxNorm () const
 
int CountSmallElems (double tol) const
 Count the number of entries with |a_ij| <= tol. More...
 
int CheckFinite () const
 Count the number of entries that are NOT finite, i.e. Inf or Nan. More...
 
void SetGraphOwner (bool ownij)
 Set the graph ownership flag (I and J arrays). More...
 
void SetDataOwner (bool owna)
 Set the data ownership flag (A array). More...
 
bool OwnsGraph () const
 Get the graph ownership flag (I and J arrays). More...
 
bool OwnsData () const
 Get the data ownership flag (A array). More...
 
void LoseData ()
 Lose the ownership of the graph (I, J) and data (A) arrays. More...
 
void Swap (SparseMatrix &other)
 
virtual ~SparseMatrix ()
 Destroys sparse matrix. More...
 
Type GetType () const
 
- Public Member Functions inherited from mfem::AbstractSparseMatrix
 AbstractSparseMatrix (int s=0)
 Creates a square matrix of the given size. More...
 
 AbstractSparseMatrix (int h, int w)
 Creates a matrix of the given height and width. More...
 
virtual ~AbstractSparseMatrix ()
 Destroys AbstractSparseMatrix. More...
 
- Public Member Functions inherited from mfem::Matrix
 Matrix (int s)
 Creates a square matrix of size s. More...
 
 Matrix (int h, int w)
 Creates a matrix of the given height and width. More...
 
bool IsSquare () const
 Returns whether the matrix is a square matrix. More...
 
virtual ~Matrix ()
 Destroys matrix. More...
 
- Public Member Functions inherited from mfem::Operator
void InitTVectors (const Operator *Po, const Operator *Ri, const Operator *Pi, Vector &x, Vector &b, Vector &X, Vector &B) const
 Initializes memory for true vectors of linear system. More...
 
 Operator (int s=0)
 Construct a square Operator with given size s (default 0). More...
 
 Operator (int h, int w)
 Construct an Operator with the given height (output size) and width (input size). More...
 
int Height () const
 Get the height (size of output) of the Operator. Synonym with NumRows(). More...
 
int NumRows () const
 Get the number of rows (size of output) of the Operator. Synonym with Height(). More...
 
int Width () const
 Get the width (size of input) of the Operator. Synonym with NumCols(). More...
 
int NumCols () const
 Get the number of columns (size of input) of the Operator. Synonym with Width(). More...
 
virtual void ArrayMult (const Array< const Vector *> &X, Array< Vector *> &Y) const
 Operator application on a matrix: Y=A(X). More...
 
virtual void ArrayMultTranspose (const Array< const Vector *> &X, Array< Vector *> &Y) const
 Action of the transpose operator on a matrix: Y=A^t(X). More...
 
virtual void ArrayAddMult (const Array< const Vector *> &X, Array< Vector *> &Y, const double a=1.0) const
 Operator application on a matrix: Y+=A(X) (default) or Y+=a*A(X). More...
 
virtual void ArrayAddMultTranspose (const Array< const Vector *> &X, Array< Vector *> &Y, const double a=1.0) const
 Operator transpose application on a matrix: Y+=A^t(X) (default) or Y+=a*A^t(X). More...
 
virtual OperatorGetGradient (const Vector &x) const
 Evaluate the gradient operator at the point x. The default behavior in class Operator is to generate an error. More...
 
virtual void AssembleDiagonal (Vector &diag) const
 Computes the diagonal entries into diag. Typically, this operation only makes sense for linear Operators. In some cases, only an approximation of the diagonal is computed. More...
 
virtual const OperatorGetProlongation () const
 Prolongation operator from linear algebra (linear system) vectors, to input vectors for the operator. NULL means identity. More...
 
virtual const OperatorGetRestriction () const
 Restriction operator from input vectors for the operator to linear algebra (linear system) vectors. NULL means identity. More...
 
virtual const OperatorGetOutputProlongation () const
 Prolongation operator from linear algebra (linear system) vectors, to output vectors for the operator. NULL means identity. More...
 
virtual const OperatorGetOutputRestrictionTranspose () const
 Transpose of GetOutputRestriction, directly available in this form to facilitate matrix-free RAP-type operators. More...
 
virtual const OperatorGetOutputRestriction () const
 Restriction operator from output vectors for the operator to linear algebra (linear system) vectors. NULL means identity. More...
 
void FormLinearSystem (const Array< int > &ess_tdof_list, Vector &x, Vector &b, Operator *&A, Vector &X, Vector &B, int copy_interior=0)
 Form a constrained linear system using a matrix-free approach. More...
 
void FormRectangularLinearSystem (const Array< int > &trial_tdof_list, const Array< int > &test_tdof_list, Vector &x, Vector &b, Operator *&A, Vector &X, Vector &B)
 Form a column-constrained linear system using a matrix-free approach. More...
 
virtual void RecoverFEMSolution (const Vector &X, const Vector &b, Vector &x)
 Reconstruct a solution vector x (e.g. a GridFunction) from the solution X of a constrained linear system obtained from Operator::FormLinearSystem() or Operator::FormRectangularLinearSystem(). More...
 
void FormSystemOperator (const Array< int > &ess_tdof_list, Operator *&A)
 Return in A a parallel (on truedofs) version of this square operator. More...
 
void FormRectangularSystemOperator (const Array< int > &trial_tdof_list, const Array< int > &test_tdof_list, Operator *&A)
 Return in A a parallel (on truedofs) version of this rectangular operator (including constraints). More...
 
void FormDiscreteOperator (Operator *&A)
 Return in A a parallel (on truedofs) version of this rectangular operator. More...
 
void PrintMatlab (std::ostream &out, int n, int m=0) const
 Prints operator with input size n and output size m in Matlab format. More...
 
virtual ~Operator ()
 Virtual destructor. More...
 
Type GetType () const
 Return the type ID of the Operator class. More...
 

Protected Types

typedef MemAlloc< RowNode, 1024 > RowNodeAlloc
 

Protected Member Functions

void Destroy ()
 
void SetEmpty ()
 
void InitGPUSparse ()
 
- Protected Member Functions inherited from mfem::Operator
void FormConstrainedSystemOperator (const Array< int > &ess_tdof_list, ConstrainedOperator *&Aout)
 see FormSystemOperator() More...
 
void FormRectangularConstrainedSystemOperator (const Array< int > &trial_tdof_list, const Array< int > &test_tdof_list, RectangularConstrainedOperator *&Aout)
 see FormRectangularSystemOperator() More...
 
OperatorSetupRAP (const Operator *Pi, const Operator *Po)
 Returns RAP Operator of this, using input/output Prolongation matrices Pi corresponds to "P", Po corresponds to "Rt". More...
 

Protected Attributes

RowNode ** Rows
 Array of linked lists, one for every row. This array represents the linked list (LIL) storage format. More...
 
int current_row
 
int * ColPtrJ
 
RowNode ** ColPtrNode
 
SparseMatrixAt
 Transpose of A. Owned. Used to perform MultTranspose() on devices. More...
 
RowNodeAllocNodesMem
 
bool isSorted
 Are the columns sorted already. More...
 
bool useGPUSparse = true
 
bool initBuffers = false
 
cusparseStatus_t status
 
cusparseMatDescr_t descr = 0
 
cusparseSpMatDescr_t matA_descr
 
cusparseDnVecDescr_t vecX_descr
 
cusparseDnVecDescr_t vecY_descr
 
cusparseMatDescr_t matA_descr
 
hipsparseStatus_t status
 
hipsparseMatDescr_t descr = 0
 
hipsparseSpMatDescr_t matA_descr
 
hipsparseDnVecDescr_t vecX_descr
 
hipsparseDnVecDescr_t vecY_descr
 
Arrays used by the CSR storage format.
Memory< int > I
 Array with size (height+1) containing the row offsets. More...
 
Memory< int > J
 Array with size I[height], containing the column indices for all matrix entries, as indexed by the I array. More...
 
Memory< double > A
 Array with size I[height], containing the actual entries of the sparse matrix, as indexed by the I array. More...
 
- Protected Attributes inherited from mfem::Operator
int height
 Dimension of the output / number of rows in the matrix. More...
 
int width
 Dimension of the input / number of columns in the matrix. More...
 

Static Protected Attributes

static int SparseMatrixCount = 0
 
static size_t bufferSize = 0
 
static void * dBuffer = nullptr
 
static cusparseHandle_t handle
 
static hipsparseHandle_t handle
 

Additional Inherited Members

- Public Types inherited from mfem::Operator
enum  DiagonalPolicy { DIAG_ZERO, DIAG_ONE, DIAG_KEEP }
 Defines operator diagonal policy upon elimination of rows and/or columns. More...
 
enum  Type {
  ANY_TYPE, MFEM_SPARSEMAT, Hypre_ParCSR, PETSC_MATAIJ,
  PETSC_MATIS, PETSC_MATSHELL, PETSC_MATNEST, PETSC_MATHYPRE,
  PETSC_MATGENERIC, Complex_Operator, MFEM_ComplexSparseMat, Complex_Hypre_ParCSR,
  Complex_DenseMat, MFEM_Block_Matrix, MFEM_Block_Operator
}
 Enumeration defining IDs for some classes derived from Operator. More...
 

Detailed Description

Data type sparse matrix.

Definition at line 50 of file sparsemat.hpp.

Member Typedef Documentation

◆ RowNodeAlloc

typedef MemAlloc<RowNode, 1024> mfem::SparseMatrix::RowNodeAlloc
protected

Definition at line 83 of file sparsemat.hpp.

Constructor & Destructor Documentation

◆ SparseMatrix() [1/7]

mfem::SparseMatrix::SparseMatrix ( )
inline

Create an empty SparseMatrix.

Definition at line 131 of file sparsemat.hpp.

◆ SparseMatrix() [2/7]

mfem::SparseMatrix::SparseMatrix ( int  nrows,
int  ncols = -1 
)
explicit

Create a sparse matrix with flexible sparsity structure using a row-wise linked list (LIL) format.

New entries are added as needed by methods like AddSubMatrix(), SetSubMatrix(), etc. Calling Finalize() will convert the SparseMatrix to the more compact compressed sparse row (CSR) format.

Definition at line 98 of file sparsemat.cpp.

◆ SparseMatrix() [3/7]

mfem::SparseMatrix::SparseMatrix ( int *  i,
int *  j,
double *  data,
int  m,
int  n 
)

Create a sparse matrix in CSR format. Ownership of i, j, and data is transferred to the SparseMatrix.

Definition at line 124 of file sparsemat.cpp.

◆ SparseMatrix() [4/7]

mfem::SparseMatrix::SparseMatrix ( int *  i,
int *  j,
double *  data,
int  m,
int  n,
bool  ownij,
bool  owna,
bool  issorted 
)

Create a sparse matrix in CSR format. Ownership of i, j, and data is optionally transferred to the SparseMatrix.

If the parameter data is NULL, then the internal A array is allocated by this constructor (initializing it with zeros and taking ownership, regardless of the parameter owna).

Definition at line 143 of file sparsemat.cpp.

◆ SparseMatrix() [5/7]

mfem::SparseMatrix::SparseMatrix ( int  nrows,
int  ncols,
int  rowsize 
)

Create a sparse matrix in CSR format where each row has space allocated for exactly rowsize entries.

SetRow() can then be called or the I, J, A arrays can be used directly.

Definition at line 176 of file sparsemat.cpp.

◆ SparseMatrix() [6/7]

mfem::SparseMatrix::SparseMatrix ( const SparseMatrix mat,
bool  copy_graph = true,
MemoryType  mt = MemoryType::PRESERVE 
)

Copy constructor (deep copy).

If mat is finalized and copy_graph is false, the I and J arrays will use a shallow copy (copy the pointers only) without transferring ownership. If mt is MemoryType::PRESERVE the memory type of the resulting SparseMatrix's I, J, and A arrays will be the same as mat, otherwise the type will be mt for those arrays that are deep copied.

Definition at line 199 of file sparsemat.cpp.

◆ SparseMatrix() [7/7]

mfem::SparseMatrix::SparseMatrix ( const Vector v)

Create a SparseMatrix with diagonal v, i.e. A = Diag(v)

Definition at line 268 of file sparsemat.cpp.

◆ ~SparseMatrix()

mfem::SparseMatrix::~SparseMatrix ( )
virtual

Destroys sparse matrix.

Definition at line 4202 of file sparsemat.cpp.

Member Function Documentation

◆ _Add_() [1/2]

void mfem::SparseMatrix::_Add_ ( const int  col,
const double  a 
)
inline

Add a value to an entry in the "current row". See SetColPtr().

Definition at line 585 of file sparsemat.hpp.

◆ _Add_() [2/2]

void mfem::SparseMatrix::_Add_ ( const int  row,
const int  col,
const double  a 
)
inline

Definition at line 594 of file sparsemat.hpp.

◆ _Get_()

double mfem::SparseMatrix::_Get_ ( const int  col) const
inline

Read the value of an entry in the "current row". See SetColPtr().

Definition at line 873 of file sparsemat.hpp.

◆ _Set_() [1/2]

void mfem::SparseMatrix::_Set_ ( const int  col,
const double  a 
)
inline

Set an entry in the "current row". See SetColPtr().

Definition at line 588 of file sparsemat.hpp.

◆ _Set_() [2/2]

void mfem::SparseMatrix::_Set_ ( const int  row,
const int  col,
const double  a 
)
inline

Definition at line 596 of file sparsemat.hpp.

◆ AbsMult()

void mfem::SparseMatrix::AbsMult ( const Vector x,
Vector y 
) const

y = |A| * x, using entry-wise absolute values of matrix A

Definition at line 1079 of file sparsemat.cpp.

◆ AbsMultTranspose()

void mfem::SparseMatrix::AbsMultTranspose ( const Vector x,
Vector y 
) const

y = |At| * x, using entry-wise absolute values of the transpose of matrix A

If the matrix is modified, call ResetTranspose() and optionally EnsureMultTranspose() to make sure this method uses the correct updated transpose.

Definition at line 1128 of file sparsemat.cpp.

◆ ActualWidth()

int mfem::SparseMatrix::ActualWidth ( ) const

Returns the actual Width of the matrix.

This method can be called for matrices finalized or not.

Definition at line 3469 of file sparsemat.cpp.

◆ Add() [1/2]

void mfem::SparseMatrix::Add ( const int  i,
const int  j,
const double  val 
)

Definition at line 2787 of file sparsemat.cpp.

◆ Add() [2/2]

void mfem::SparseMatrix::Add ( const double  a,
const SparseMatrix B 
)

Add the sparse matrix 'B' scaled by the scalar 'a' into '*this'. Only entries in the sparsity pattern of '*this' are added.

Definition at line 3171 of file sparsemat.cpp.

◆ AddMult()

void mfem::SparseMatrix::AddMult ( const Vector x,
Vector y,
const double  a = 1.0 
) const
virtual

y += A * x (default) or y += a * A * x

Implements mfem::AbstractSparseMatrix.

Definition at line 740 of file sparsemat.cpp.

◆ AddMultTranspose()

void mfem::SparseMatrix::AddMultTranspose ( const Vector x,
Vector y,
const double  a = 1.0 
) const
virtual

y += At * x (default) or y += a * At * x

If the matrix is modified, call ResetTranspose() and optionally EnsureMultTranspose() to make sure this method uses the correct updated transpose.

Implements mfem::AbstractSparseMatrix.

Definition at line 915 of file sparsemat.cpp.

◆ AddRow()

void mfem::SparseMatrix::AddRow ( const int  row,
const Array< int > &  cols,
const Vector srow 
)

Definition at line 3022 of file sparsemat.cpp.

◆ AddSubMatrix()

void mfem::SparseMatrix::AddSubMatrix ( const Array< int > &  rows,
const Array< int > &  cols,
const DenseMatrix subm,
int  skip_zeros = 1 
)

Insert the DenseMatrix into this SparseMatrix at the specified rows and columns. If skip_zeros==0 , all entries from the DenseMatrix are added including zeros. If skip_zeros==2 , no zeros are added to the SparseMatrix regardless of their position in the matrix. Otherwise, the default skip_zeros behavior is to omit the zero from the SparseMatrix unless it would break the symmetric structure of the SparseMatrix.

Definition at line 2722 of file sparsemat.cpp.

◆ BooleanMult()

void mfem::SparseMatrix::BooleanMult ( const Array< int > &  x,
Array< int > &  y 
) const

y = A * x, treating all entries as booleans (zero=false, nonzero=true).

The actual values stored in the data array, A, are not used - this means that all entries in the sparsity pattern are considered to be true by this method.

Definition at line 1026 of file sparsemat.cpp.

◆ BooleanMultTranspose()

void mfem::SparseMatrix::BooleanMultTranspose ( const Array< int > &  x,
Array< int > &  y 
) const

y = At * x, treating all entries as booleans (zero=false, nonzero=true).

The actual values stored in the data array, A, are not used - this means that all entries in the sparsity pattern are considered to be true by this method.

Definition at line 1056 of file sparsemat.cpp.

◆ BuildTranspose()

void mfem::SparseMatrix::BuildTranspose ( ) const

Build and store internally the transpose of this matrix which will be used in the methods AddMultTranspose(), MultTranspose(), and AbsMultTranspose().

If this method has been called, the internal transpose matrix will be used to perform the action of the transpose matrix in AddMultTranspose(), MultTranspose(), and AbsMultTranspose().

Warning: any changes in this matrix will invalidate the internal transpose. To rebuild the transpose, call ResetTranspose() followed by (optionally) a call to this method. If the internal transpose is already built, this method has no effect.

When any non-serial-CPU backend is enabled, i.e. the call Device::Allows(~ Backend::CPU_MASK) returns true, the above methods require the internal transpose to be built. If that is not the case (i.e. the internal transpose is not built), these methods will automatically call EnsureMultTranspose(). When using any backend from Backend::CPU_MASK, calling this method is optional.

This method can only be used when the sparse matrix is finalized.

See also
EnsureMultTranspose(), ResetTranspose().

Definition at line 960 of file sparsemat.cpp.

◆ CheckFinite()

int mfem::SparseMatrix::CheckFinite ( ) const

Count the number of entries that are NOT finite, i.e. Inf or Nan.

Definition at line 1664 of file sparsemat.cpp.

◆ Clear()

void mfem::SparseMatrix::Clear ( )
inline

Clear the contents of the SparseMatrix.

Definition at line 209 of file sparsemat.hpp.

◆ ClearColPtr()

void mfem::SparseMatrix::ClearColPtr ( ) const
inline

Reset the "current row" set by calling SetColPtr(). This method must be called between any two calls to SetColPtr().

Definition at line 827 of file sparsemat.hpp.

◆ ClearCuSparse()

MFEM_DEPRECATED void mfem::SparseMatrix::ClearCuSparse ( )
inline

Deprecated equivalent of ClearGPUSparse().

Definition at line 216 of file sparsemat.hpp.

◆ ClearGPUSparse()

void mfem::SparseMatrix::ClearGPUSparse ( )

Clear the cuSPARSE/hipSPARSE descriptors. This must be called after releasing the device memory of A.

Definition at line 81 of file sparsemat.cpp.

◆ ColumnsAreSorted()

bool mfem::SparseMatrix::ColumnsAreSorted ( ) const
inline

Returns whether or not the columns are sorted.

Definition at line 554 of file sparsemat.hpp.

◆ CountSmallElems()

int mfem::SparseMatrix::CountSmallElems ( double  tol) const

Count the number of entries with |a_ij| <= tol.

Definition at line 1636 of file sparsemat.cpp.

◆ Destroy()

void mfem::SparseMatrix::Destroy ( )
protected

Definition at line 3436 of file sparsemat.cpp.

◆ DiagScale()

void mfem::SparseMatrix::DiagScale ( const Vector b,
Vector x,
double  sc = 1.0,
bool  use_abs_diag = false 
) const

x = sc b / A_ii. When use_abs_diag = true, |A_ii| is used.

Definition at line 2620 of file sparsemat.cpp.

◆ Elem() [1/2]

double & mfem::SparseMatrix::Elem ( int  i,
int  j 
)
virtual

Returns reference to a_{ij}.

Implements mfem::Matrix.

Definition at line 601 of file sparsemat.cpp.

◆ Elem() [2/2]

const double & mfem::SparseMatrix::Elem ( int  i,
int  j 
) const
virtual

Returns constant reference to a_{ij}.

Implements mfem::Matrix.

Definition at line 606 of file sparsemat.cpp.

◆ EliminateBC()

void mfem::SparseMatrix::EliminateBC ( const Array< int > &  ess_dofs,
DiagonalPolicy  diag_policy 
)

Eliminate essential (Dirichlet) boundary conditions.

Parameters
[in]ess_dofsindices of the degrees of freedom belonging to the essential boundary conditions.
[in]diag_policypolicy for diagonal entries.

Definition at line 2315 of file sparsemat.cpp.

◆ EliminateCol()

void mfem::SparseMatrix::EliminateCol ( int  col,
DiagonalPolicy  dpolicy = DIAG_ZERO 
)

Eliminates the column col from the matrix.

  • If dpolicy = DIAG_ZERO, all entries in the column will be set to 0.
  • If dpolicy = DIAG_ONE (matrix must be square), the diagonal entry will be set equal to 1 and all other entries in the column to 0.
  • The policy DIAG_KEEP is not supported.

Definition at line 1741 of file sparsemat.cpp.

◆ EliminateCols() [1/2]

void mfem::SparseMatrix::EliminateCols ( const Array< int > &  cols,
const Vector x = NULL,
Vector b = NULL 
)

Eliminate all columns i for which cols[i] != 0.

Elimination of a column means that all entries in the column are set to zero. In addition, if the pointers x and b are not NULL, the eliminated matrix entries are multiplied by the corresponding solution value in *x and subtracted from the r.h.s. vector, *b.

Definition at line 1781 of file sparsemat.cpp.

◆ EliminateCols() [2/2]

void mfem::SparseMatrix::EliminateCols ( const Array< int > &  col_marker,
SparseMatrix Ae 
)

Similar to EliminateCols + save the eliminated entries into Ae so that (*this) + Ae is equal to the original matrix.

Definition at line 1820 of file sparsemat.cpp.

◆ EliminateRow() [1/2]

void mfem::SparseMatrix::EliminateRow ( int  row,
const double  sol,
Vector rhs 
)

Eliminates a column from the transpose matrix.

Definition at line 1693 of file sparsemat.cpp.

◆ EliminateRow() [2/2]

void mfem::SparseMatrix::EliminateRow ( int  row,
DiagonalPolicy  dpolicy = DIAG_ZERO 
)

Eliminates a row from the matrix.

  • If dpolicy = DIAG_ZERO, all the entries in the row will be set to 0.
  • If dpolicy = DIAG_ONE (matrix must be square), the diagonal entry will be set equal to 1 and all other entries in the row to 0.
  • The policy DIAG_KEEP is not supported.

Definition at line 1709 of file sparsemat.cpp.

◆ EliminateRowCol() [1/3]

void mfem::SparseMatrix::EliminateRowCol ( int  rc,
const double  sol,
Vector rhs,
DiagonalPolicy  dpolicy = DIAG_ONE 
)

Eliminate row rc and column rc and modify the rhs using sol.

Eliminates the column rc to the rhs, deletes the row rc and replaces the element (rc,rc) with 1.0; assumes that element (i,rc) is assembled if and only if the element (rc,i) is assembled. By default, elements (rc,rc) are set to 1.0, although this behavior can be adjusted by changing the dpolicy parameter.

Definition at line 1854 of file sparsemat.cpp.

◆ EliminateRowCol() [2/3]

void mfem::SparseMatrix::EliminateRowCol ( int  rc,
DiagonalPolicy  dpolicy = DIAG_ONE 
)

Eliminate row rc and column rc.

Definition at line 2077 of file sparsemat.cpp.

◆ EliminateRowCol() [3/3]

void mfem::SparseMatrix::EliminateRowCol ( int  rc,
SparseMatrix Ae,
DiagonalPolicy  dpolicy = DIAG_ONE 
)

Similar to EliminateRowCol(int, DiagonalPolicy) + save the eliminated entries into Ae so that (*this) + Ae is equal to the original matrix.

Definition at line 2221 of file sparsemat.cpp.

◆ EliminateRowColDiag()

void mfem::SparseMatrix::EliminateRowColDiag ( int  rc,
double  value 
)

Perform elimination and set the diagonal entry to the given value.

Definition at line 2158 of file sparsemat.cpp.

◆ EliminateRowColMultipleRHS()

void mfem::SparseMatrix::EliminateRowColMultipleRHS ( int  rc,
const Vector sol,
DenseMatrix rhs,
DiagonalPolicy  dpolicy = DIAG_ONE 
)

Similar to EliminateRowCol(int, const double, Vector &, DiagonalPolicy), but multiple values for eliminated unknowns are accepted, and accordingly multiple right-hand-sides are used.

Definition at line 1953 of file sparsemat.cpp.

◆ EliminateZeroRows()

void mfem::SparseMatrix::EliminateZeroRows ( const double  threshold = 1e-12)
virtual

If a row contains only zeros, set its diagonal to 1.

Implements mfem::AbstractSparseMatrix.

Definition at line 2369 of file sparsemat.cpp.

◆ Empty()

bool mfem::SparseMatrix::Empty ( ) const
inline

Check if the SparseMatrix is empty.

Definition at line 219 of file sparsemat.hpp.

◆ EnsureMultTranspose()

void mfem::SparseMatrix::EnsureMultTranspose ( ) const

Ensures that the matrix is capable of performing MultTranspose(), AddMultTranspose(), and AbsMultTranspose().

For non-serial-CPU backends (e.g. GPU, OpenMP), multiplying by the transpose requires that the internal transpose matrix be already built. When such a backend is enabled, this function will build the internal transpose matrix, see BuildTranspose().

For the serial CPU backends, the internal transpose is not required, and this function is a no-op. This allows for significant memory savings when the internal transpose matrix is not required.

Definition at line 974 of file sparsemat.cpp.

◆ Finalize() [1/2]

virtual void mfem::SparseMatrix::Finalize ( int  skip_zeros = 1)
inlinevirtual

Finalize the matrix initialization, switching the storage format from LIL to CSR.

This method should be called once, after the matrix has been initialized. Internally, this method converts the matrix from row-wise linked list (LIL) format into CSR (compressed sparse row) format.

Reimplemented from mfem::Matrix.

Definition at line 546 of file sparsemat.hpp.

◆ Finalize() [2/2]

void mfem::SparseMatrix::Finalize ( int  skip_zeros,
bool  fix_empty_rows 
)

A slightly more general version of the Finalize(int) method.

Definition at line 1322 of file sparsemat.cpp.

◆ Finalized()

bool mfem::SparseMatrix::Finalized ( ) const
inline

Returns whether or not CSR format has been finalized.

Definition at line 552 of file sparsemat.hpp.

◆ Gauss_Seidel_back()

void mfem::SparseMatrix::Gauss_Seidel_back ( const Vector x,
Vector y 
) const

Definition at line 2472 of file sparsemat.cpp.

◆ Gauss_Seidel_forw()

void mfem::SparseMatrix::Gauss_Seidel_forw ( const Vector x,
Vector y 
) const

Gauss-Seidel forward and backward iterations over a vector x.

Definition at line 2388 of file sparsemat.cpp.

◆ GetBlocks()

void mfem::SparseMatrix::GetBlocks ( Array2D< SparseMatrix *> &  blocks) const

Split the matrix into M x N blocks of sparse matrices in CSR format. The 'blocks' array is M x N (i.e. M and N are determined by its dimensions) and its entries are overwritten by the new blocks.

Definition at line 1437 of file sparsemat.cpp.

◆ GetData() [1/2]

double* mfem::SparseMatrix::GetData ( )
inline

Return the element data, i.e. the array A.

Definition at line 232 of file sparsemat.hpp.

◆ GetData() [2/2]

const double* mfem::SparseMatrix::GetData ( ) const
inline

Return the element data, i.e. the array A, const version.

Definition at line 234 of file sparsemat.hpp.

◆ GetDiag()

void mfem::SparseMatrix::GetDiag ( Vector d) const

Returns the Diagonal of A.

Definition at line 669 of file sparsemat.cpp.

◆ GetI() [1/2]

int* mfem::SparseMatrix::GetI ( )
inline

Return the array I.

Definition at line 222 of file sparsemat.hpp.

◆ GetI() [2/2]

const int* mfem::SparseMatrix::GetI ( ) const
inline

Return the array I, const version.

Definition at line 224 of file sparsemat.hpp.

◆ GetJ() [1/2]

int* mfem::SparseMatrix::GetJ ( )
inline

Return the array J.

Definition at line 227 of file sparsemat.hpp.

◆ GetJ() [2/2]

const int* mfem::SparseMatrix::GetJ ( ) const
inline

Return the array J, const version.

Definition at line 229 of file sparsemat.hpp.

◆ GetJacobiScaling()

double mfem::SparseMatrix::GetJacobiScaling ( ) const

Determine appropriate scaling for Jacobi iteration.

Definition at line 2555 of file sparsemat.cpp.

◆ GetMemoryClass()

virtual MemoryClass mfem::SparseMatrix::GetMemoryClass ( ) const
inlinevirtual

Return the MemoryClass preferred by the Operator.

This is the MemoryClass that will be used to access the input and output vectors in the Mult() and MultTranspose() methods.

For example, classes using the mfem::forall macro for implementation can return the value returned by Device::GetMemoryClass().

The default implementation of this method in class Operator returns MemoryClass::HOST.

Reimplemented from mfem::Operator.

Definition at line 342 of file sparsemat.hpp.

◆ GetMemoryData() [1/2]

Memory<double>& mfem::SparseMatrix::GetMemoryData ( )
inline

Definition at line 269 of file sparsemat.hpp.

◆ GetMemoryData() [2/2]

const Memory<double>& mfem::SparseMatrix::GetMemoryData ( ) const
inline

Definition at line 270 of file sparsemat.hpp.

◆ GetMemoryI() [1/2]

Memory<int>& mfem::SparseMatrix::GetMemoryI ( )
inline

Definition at line 237 of file sparsemat.hpp.

◆ GetMemoryI() [2/2]

const Memory<int>& mfem::SparseMatrix::GetMemoryI ( ) const
inline

Definition at line 238 of file sparsemat.hpp.

◆ GetMemoryJ() [1/2]

Memory<int>& mfem::SparseMatrix::GetMemoryJ ( )
inline

Definition at line 253 of file sparsemat.hpp.

◆ GetMemoryJ() [2/2]

const Memory<int>& mfem::SparseMatrix::GetMemoryJ ( ) const
inline

Definition at line 254 of file sparsemat.hpp.

◆ GetRow()

int mfem::SparseMatrix::GetRow ( const int  row,
Array< int > &  cols,
Vector srow 
) const
virtual

Extract all column indices and values from a given row.

If the matrix is finalized (i.e. in CSR format), cols and srow will simply be references to the specific portion of the J and A arrays. As required by the AbstractSparseMatrix interface this method returns:

  • 0, if cols and srow are copies of the values in the matrix, i.e. when the matrix is open.
  • 1, if cols and srow are views of the values in the matrix, i.e. when the matrix is finalized.
    Warning
    This method breaks the const-ness when the matrix is finalized because it gives write access to the J and A arrays.

Implements mfem::AbstractSparseMatrix.

Definition at line 2935 of file sparsemat.cpp.

◆ GetRowColumns() [1/2]

int * mfem::SparseMatrix::GetRowColumns ( const int  row)

Return a pointer to the column indices in a row.

Definition at line 391 of file sparsemat.cpp.

◆ GetRowColumns() [2/2]

const int * mfem::SparseMatrix::GetRowColumns ( const int  row) const

Return a pointer to the column indices in a row, const version.

Definition at line 398 of file sparsemat.cpp.

◆ GetRowEntries() [1/2]

double * mfem::SparseMatrix::GetRowEntries ( const int  row)

Return a pointer to the entries in a row.

Definition at line 405 of file sparsemat.cpp.

◆ GetRowEntries() [2/2]

const double * mfem::SparseMatrix::GetRowEntries ( const int  row) const

Return a pointer to the entries in a row, const version.

Definition at line 412 of file sparsemat.cpp.

◆ GetRowNorml1()

double mfem::SparseMatrix::GetRowNorml1 ( int  irow) const

For i = irow compute \( x_i = \sum_j | A_{i, j} | \).

Definition at line 1237 of file sparsemat.cpp.

◆ GetRowSums()

void mfem::SparseMatrix::GetRowSums ( Vector x) const

For all i compute \( x_i = \sum_j A_{ij} \).

Definition at line 1214 of file sparsemat.cpp.

◆ GetSubMatrix()

void mfem::SparseMatrix::GetSubMatrix ( const Array< int > &  rows,
const Array< int > &  cols,
DenseMatrix subm 
) const

Definition at line 2886 of file sparsemat.cpp.

◆ GetType()

Type mfem::SparseMatrix::GetType ( ) const
inline

Definition at line 708 of file sparsemat.hpp.

◆ HostReadData()

const double* mfem::SparseMatrix::HostReadData ( ) const
inline

Definition at line 277 of file sparsemat.hpp.

◆ HostReadI()

const int* mfem::SparseMatrix::HostReadI ( ) const
inline

Definition at line 245 of file sparsemat.hpp.

◆ HostReadJ()

const int* mfem::SparseMatrix::HostReadJ ( ) const
inline

Definition at line 261 of file sparsemat.hpp.

◆ HostReadWriteData()

double* mfem::SparseMatrix::HostReadWriteData ( )
inline

Definition at line 281 of file sparsemat.hpp.

◆ HostReadWriteI()

int* mfem::SparseMatrix::HostReadWriteI ( )
inline

Definition at line 249 of file sparsemat.hpp.

◆ HostReadWriteJ()

int* mfem::SparseMatrix::HostReadWriteJ ( )
inline

Definition at line 265 of file sparsemat.hpp.

◆ HostWriteData()

double* mfem::SparseMatrix::HostWriteData ( )
inline

Definition at line 279 of file sparsemat.hpp.

◆ HostWriteI()

int* mfem::SparseMatrix::HostWriteI ( )
inline

Definition at line 247 of file sparsemat.hpp.

◆ HostWriteJ()

int* mfem::SparseMatrix::HostWriteJ ( )
inline

Definition at line 263 of file sparsemat.hpp.

◆ InitGPUSparse()

void mfem::SparseMatrix::InitGPUSparse ( )
protected

Definition at line 64 of file sparsemat.cpp.

◆ InnerProduct()

double mfem::SparseMatrix::InnerProduct ( const Vector x,
const Vector y 
) const

Compute y^t A x.

Definition at line 1173 of file sparsemat.cpp.

◆ Inverse()

MatrixInverse * mfem::SparseMatrix::Inverse ( ) const
virtual

This virtual method is not supported: it always returns NULL.

Implements mfem::Matrix.

Definition at line 1688 of file sparsemat.cpp.

◆ IsSymmetric()

double mfem::SparseMatrix::IsSymmetric ( ) const

Returns max_{i,j} |(i,j)-(j,i)| for a finalized matrix.

Definition at line 1527 of file sparsemat.cpp.

◆ Jacobi()

void mfem::SparseMatrix::Jacobi ( const Vector b,
const Vector x0,
Vector x1,
double  sc,
bool  use_abs_diag = false 
) const

One scaled Jacobi iteration for the system A x = b. x1 = x0 + sc D^{-1} (b - A x0) where D is the diag of A. Absolute values of D are used when use_abs_diag = true.

Definition at line 2588 of file sparsemat.cpp.

◆ Jacobi2()

void mfem::SparseMatrix::Jacobi2 ( const Vector b,
const Vector x0,
Vector x1,
double  sc = 1.0 
) const

x1 = x0 + sc D^{-1} (b - A x0) where \( D_{ii} = \sum_j |A_{ij}| \).

Definition at line 2708 of file sparsemat.cpp.

◆ Jacobi3()

void mfem::SparseMatrix::Jacobi3 ( const Vector b,
const Vector x0,
Vector x1,
double  sc = 1.0 
) const

x1 = x0 + sc D^{-1} (b - A x0) where \( D_{ii} = \sum_j A_{ij} \).

Definition at line 2715 of file sparsemat.cpp.

◆ LoseData()

void mfem::SparseMatrix::LoseData ( )
inline

Lose the ownership of the graph (I, J) and data (A) arrays.

Definition at line 701 of file sparsemat.hpp.

◆ MakeRef()

void mfem::SparseMatrix::MakeRef ( const SparseMatrix master)

Clear the contents of the SparseMatrix and make it a reference to master.

After this call, the matrix will point to the same data as master but it will not own its data. The master must be finalized.

Definition at line 313 of file sparsemat.cpp.

◆ MaxNorm()

double mfem::SparseMatrix::MaxNorm ( ) const

Definition at line 1611 of file sparsemat.cpp.

◆ MaxRowSize()

int mfem::SparseMatrix::MaxRowSize ( ) const

Returns the maximum number of elements among all rows.

Definition at line 367 of file sparsemat.cpp.

◆ MoveDiagonalFirst()

void mfem::SparseMatrix::MoveDiagonalFirst ( )

Move the diagonal entry to the first position in each row, preserving the order of the rest of the columns.

Definition at line 577 of file sparsemat.cpp.

◆ Mult()

void mfem::SparseMatrix::Mult ( const Vector x,
Vector y 
) const
virtual

Matrix vector multiplication.

Implements mfem::AbstractSparseMatrix.

Definition at line 733 of file sparsemat.cpp.

◆ MultTranspose()

void mfem::SparseMatrix::MultTranspose ( const Vector x,
Vector y 
) const
virtual

Multiply a vector with the transposed matrix. y = At * x.

If the matrix is modified, call ResetTranspose() and optionally EnsureMultTranspose() to make sure this method uses the correct updated transpose.

Implements mfem::AbstractSparseMatrix.

Definition at line 908 of file sparsemat.cpp.

◆ NumNonZeroElems()

int mfem::SparseMatrix::NumNonZeroElems ( ) const
virtual

Returns the number of the nonzero elements in the matrix.

Implements mfem::AbstractSparseMatrix.

Definition at line 1588 of file sparsemat.cpp.

◆ operator()() [1/2]

double & mfem::SparseMatrix::operator() ( int  i,
int  j 
)

Returns reference to A[i][j].

Definition at line 611 of file sparsemat.cpp.

◆ operator()() [2/2]

const double & mfem::SparseMatrix::operator() ( int  i,
int  j 
) const

Returns reference to A[i][j].

Definition at line 634 of file sparsemat.cpp.

◆ operator*=()

SparseMatrix & mfem::SparseMatrix::operator*= ( double  a)

Definition at line 3220 of file sparsemat.cpp.

◆ operator+=()

SparseMatrix & mfem::SparseMatrix::operator+= ( const SparseMatrix B)

Add the sparse matrix 'B' to '*this'. This operation will cause an error if '*this' is finalized and 'B' has larger sparsity pattern.

Definition at line 3141 of file sparsemat.cpp.

◆ operator=() [1/2]

SparseMatrix & mfem::SparseMatrix::operator= ( const SparseMatrix rhs)

Assignment operator: deep copy.

Definition at line 303 of file sparsemat.cpp.

◆ operator=() [2/2]

SparseMatrix & mfem::SparseMatrix::operator= ( double  a)

Definition at line 3194 of file sparsemat.cpp.

◆ OverrideSize()

void mfem::SparseMatrix::OverrideSize ( int  height_,
int  width_ 
)

Sets the height and width of the matrix.

Warning
This does not modify in any way the underlying CSR or LIL representation of the matrix.

This function should generally be called when manually constructing the CSR I, J, and A arrays in conjunction with the SparseMatrix::SparseMatrix() constructor.

Definition at line 297 of file sparsemat.cpp.

◆ OwnsData()

bool mfem::SparseMatrix::OwnsData ( ) const
inline

Get the data ownership flag (A array).

Definition at line 698 of file sparsemat.hpp.

◆ OwnsGraph()

bool mfem::SparseMatrix::OwnsGraph ( ) const
inline

Get the graph ownership flag (I and J arrays).

Definition at line 695 of file sparsemat.hpp.

◆ PartAddMult()

void mfem::SparseMatrix::PartAddMult ( const Array< int > &  rows,
const Vector x,
Vector y,
const double  a = 1.0 
) const

Definition at line 1008 of file sparsemat.cpp.

◆ PartMult()

void mfem::SparseMatrix::PartMult ( const Array< int > &  rows,
const Vector x,
Vector y 
) const

Definition at line 982 of file sparsemat.cpp.

◆ Print()

void mfem::SparseMatrix::Print ( std::ostream &  out = mfem::out,
int  width_ = 4 
) const
virtual

Prints matrix to stream out.

Reimplemented from mfem::Matrix.

Definition at line 3244 of file sparsemat.cpp.

◆ PrintCSR()

void mfem::SparseMatrix::PrintCSR ( std::ostream &  out) const

Prints matrix to stream out in hypre_CSRMatrix format.

Definition at line 3336 of file sparsemat.cpp.

◆ PrintCSR2()

void mfem::SparseMatrix::PrintCSR2 ( std::ostream &  out) const

Prints a sparse matrix to stream out in CSR format.

Definition at line 3360 of file sparsemat.cpp.

◆ PrintInfo()

void mfem::SparseMatrix::PrintInfo ( std::ostream &  out) const

Print various sparse matrix statistics.

Definition at line 3385 of file sparsemat.cpp.

◆ PrintMatlab()

void mfem::SparseMatrix::PrintMatlab ( std::ostream &  out = mfem::out) const
virtual

Prints matrix in matlab format.

Reimplemented from mfem::Operator.

Definition at line 3292 of file sparsemat.cpp.

◆ PrintMM()

void mfem::SparseMatrix::PrintMM ( std::ostream &  out = mfem::out) const

Prints matrix in Matrix Market sparse format.

Definition at line 3314 of file sparsemat.cpp.

◆ ReadData()

const double* mfem::SparseMatrix::ReadData ( bool  on_dev = true) const
inline

Definition at line 271 of file sparsemat.hpp.

◆ ReadI()

const int* mfem::SparseMatrix::ReadI ( bool  on_dev = true) const
inline

Definition at line 239 of file sparsemat.hpp.

◆ ReadJ()

const int* mfem::SparseMatrix::ReadJ ( bool  on_dev = true) const
inline

Definition at line 255 of file sparsemat.hpp.

◆ ReadWriteData()

double* mfem::SparseMatrix::ReadWriteData ( bool  on_dev = true)
inline

Definition at line 275 of file sparsemat.hpp.

◆ ReadWriteI()

int* mfem::SparseMatrix::ReadWriteI ( bool  on_dev = true)
inline

Definition at line 243 of file sparsemat.hpp.

◆ ReadWriteJ()

int* mfem::SparseMatrix::ReadWriteJ ( bool  on_dev = true)
inline

Definition at line 259 of file sparsemat.hpp.

◆ ResetTranspose()

void mfem::SparseMatrix::ResetTranspose ( ) const

Reset (destroy) the internal transpose matrix. See BuildTranspose() for more details.

Definition at line 968 of file sparsemat.cpp.

◆ RowIsEmpty()

bool mfem::SparseMatrix::RowIsEmpty ( const int  row) const

Definition at line 2914 of file sparsemat.cpp.

◆ RowSize()

int mfem::SparseMatrix::RowSize ( const int  i) const

Returns the number of elements in row i.

Definition at line 344 of file sparsemat.cpp.

◆ ScaleColumns()

void mfem::SparseMatrix::ScaleColumns ( const Vector sr)

this = this * diag(sr);

Definition at line 3113 of file sparsemat.cpp.

◆ ScaleRow()

void mfem::SparseMatrix::ScaleRow ( const int  row,
const double  scale 
)

Definition at line 3054 of file sparsemat.cpp.

◆ ScaleRows()

void mfem::SparseMatrix::ScaleRows ( const Vector sl)

this = diag(sl) * this;

Definition at line 3082 of file sparsemat.cpp.

◆ SearchRow() [1/2]

double & mfem::SparseMatrix::SearchRow ( const int  col)
inline

Perform a fast search for an entry in the "current row". See SetColPtr().

If the matrix is not finalized and the entry is not found in the SparseMatrix, it will be added to the sparsity pattern initialized with zero. If the matrix is finalized and the entry is not found, an error will be generated.

Definition at line 846 of file sparsemat.hpp.

◆ SearchRow() [2/2]

double & mfem::SparseMatrix::SearchRow ( const int  row,
const int  col 
)
inline

Definition at line 887 of file sparsemat.hpp.

◆ Set()

void mfem::SparseMatrix::Set ( const int  i,
const int  j,
const double  val 
)

Definition at line 2768 of file sparsemat.cpp.

◆ SetColPtr()

void mfem::SparseMatrix::SetColPtr ( const int  row) const
inline

Initialize the SparseMatrix for fast access to the entries of the given row which becomes the "current row".

Fast access to the entries of the "current row" can be performed using the methods: SearchRow(const int), Add(const int, const double), Set(const int, const double), and Get(const int).

Definition at line 792 of file sparsemat.hpp.

◆ SetDataOwner()

void mfem::SparseMatrix::SetDataOwner ( bool  owna)
inline

Set the data ownership flag (A array).

Definition at line 692 of file sparsemat.hpp.

◆ SetDiagIdentity()

void mfem::SparseMatrix::SetDiagIdentity ( )

If a row contains only one diag entry of zero, set it to 1.

Definition at line 2358 of file sparsemat.cpp.

◆ SetEmpty()

void mfem::SparseMatrix::SetEmpty ( )
protected

Definition at line 325 of file sparsemat.cpp.

◆ SetGraphOwner()

void mfem::SparseMatrix::SetGraphOwner ( bool  ownij)
inline

Set the graph ownership flag (I and J arrays).

Definition at line 688 of file sparsemat.hpp.

◆ SetRow()

void mfem::SparseMatrix::SetRow ( const int  row,
const Array< int > &  cols,
const Vector srow 
)

Definition at line 2975 of file sparsemat.cpp.

◆ SetSubMatrix()

void mfem::SparseMatrix::SetSubMatrix ( const Array< int > &  rows,
const Array< int > &  cols,
const DenseMatrix subm,
int  skip_zeros = 1 
)

Definition at line 2806 of file sparsemat.cpp.

◆ SetSubMatrixTranspose()

void mfem::SparseMatrix::SetSubMatrixTranspose ( const Array< int > &  rows,
const Array< int > &  cols,
const DenseMatrix subm,
int  skip_zeros = 1 
)

Definition at line 2845 of file sparsemat.cpp.

◆ SetWidth()

void mfem::SparseMatrix::SetWidth ( int  width_ = -1)

Change the width of a SparseMatrix.

If width_ = -1 (DEFAULT), this routine will set the new width to the actual Width of the matrix awidth = max(J) + 1. Values 0 <= width_ < awidth are not allowed (error check in Debug Mode only)

This method can be called for matrices finalized or not.

Definition at line 419 of file sparsemat.cpp.

◆ Size()

int mfem::SparseMatrix::Size ( ) const
inline

For backward compatibility, define Size() to be synonym of Height().

Definition at line 206 of file sparsemat.hpp.

◆ SortColumnIndices()

void mfem::SparseMatrix::SortColumnIndices ( )

Sort the column indices corresponding to each row.

Definition at line 457 of file sparsemat.cpp.

◆ Swap()

void mfem::SparseMatrix::Swap ( SparseMatrix other)

Definition at line 4182 of file sparsemat.cpp.

◆ Symmetrize()

void mfem::SparseMatrix::Symmetrize ( )

(*this) = 1/2 ((*this) + (*this)^t)

Definition at line 1569 of file sparsemat.cpp.

◆ Threshold()

void mfem::SparseMatrix::Threshold ( double  tol,
bool  fix_empty_rows = false 
)

Remove entries smaller in absolute value than a given tolerance tol. If fix_empty_rows is true, a zero value is inserted in the diagonal entry (for square matrices only)

Definition at line 1261 of file sparsemat.cpp.

◆ ToDenseMatrix() [1/2]

DenseMatrix * mfem::SparseMatrix::ToDenseMatrix ( ) const

Produces a DenseMatrix from a SparseMatrix.

Definition at line 703 of file sparsemat.cpp.

◆ ToDenseMatrix() [2/2]

void mfem::SparseMatrix::ToDenseMatrix ( DenseMatrix B) const

Produces a DenseMatrix from a SparseMatrix.

Definition at line 716 of file sparsemat.cpp.

◆ UseCuSparse()

MFEM_DEPRECATED void mfem::SparseMatrix::UseCuSparse ( bool  useCuSparse_ = true)
inline

Deprecated equivalent of UseGPUSparse().

Definition at line 194 of file sparsemat.hpp.

◆ UseGPUSparse()

void mfem::SparseMatrix::UseGPUSparse ( bool  useGPUSparse_ = true)
inline

Runtime option to use cuSPARSE or hipSPARSE. Only valid when using a CUDA or HIP backend.

Note
This option is enabled by default, so typically one would use this method to disable the use of cuSPARSE/hipSPARSE.

Definition at line 191 of file sparsemat.hpp.

◆ WriteData()

double* mfem::SparseMatrix::WriteData ( bool  on_dev = true)
inline

Definition at line 273 of file sparsemat.hpp.

◆ WriteI()

int* mfem::SparseMatrix::WriteI ( bool  on_dev = true)
inline

Definition at line 241 of file sparsemat.hpp.

◆ WriteJ()

int* mfem::SparseMatrix::WriteJ ( bool  on_dev = true)
inline

Definition at line 257 of file sparsemat.hpp.

Member Data Documentation

◆ A

Memory<double> mfem::SparseMatrix::A
protected

Array with size I[height], containing the actual entries of the sparse matrix, as indexed by the I array.

Definition at line 68 of file sparsemat.hpp.

◆ At

SparseMatrix* mfem::SparseMatrix::At
mutableprotected

Transpose of A. Owned. Used to perform MultTranspose() on devices.

Definition at line 80 of file sparsemat.hpp.

◆ bufferSize

size_t mfem::SparseMatrix::bufferSize = 0
staticprotected

Definition at line 101 of file sparsemat.hpp.

◆ ColPtrJ

int* mfem::SparseMatrix::ColPtrJ
mutableprotected

Definition at line 76 of file sparsemat.hpp.

◆ ColPtrNode

RowNode** mfem::SparseMatrix::ColPtrNode
mutableprotected

Definition at line 77 of file sparsemat.hpp.

◆ current_row

int mfem::SparseMatrix::current_row
mutableprotected

Definition at line 75 of file sparsemat.hpp.

◆ dBuffer

void * mfem::SparseMatrix::dBuffer = nullptr
staticprotected

Definition at line 102 of file sparsemat.hpp.

◆ descr [1/2]

cusparseMatDescr_t mfem::SparseMatrix::descr = 0
protected

Definition at line 108 of file sparsemat.hpp.

◆ descr [2/2]

hipsparseMatDescr_t mfem::SparseMatrix::descr = 0
protected

Definition at line 121 of file sparsemat.hpp.

◆ handle [1/2]

cusparseHandle_t mfem::SparseMatrix::handle
staticprotected

Definition at line 107 of file sparsemat.hpp.

◆ handle [2/2]

hipsparseHandle_t mfem::SparseMatrix::handle
staticprotected

Definition at line 120 of file sparsemat.hpp.

◆ I

Memory<int> mfem::SparseMatrix::I
protected

Array with size (height+1) containing the row offsets.

The data for row r, 0 <= r < height, is at offsets j, I[r] <= j < I[r+1]. The offsets, j, are indices in the J and A arrays. The first entry in this array is always zero, I[0] = 0, and the last entry, I[height], gives the total number of entries stored (at a minimum, all nonzeros must be represented) in the sparse matrix.

Definition at line 62 of file sparsemat.hpp.

◆ initBuffers

bool mfem::SparseMatrix::initBuffers = false
mutableprotected

Definition at line 103 of file sparsemat.hpp.

◆ isSorted

bool mfem::SparseMatrix::isSorted
protected

Are the columns sorted already.

Definition at line 88 of file sparsemat.hpp.

◆ J

Memory<int> mfem::SparseMatrix::J
protected

Array with size I[height], containing the column indices for all matrix entries, as indexed by the I array.

Definition at line 65 of file sparsemat.hpp.

◆ matA_descr [1/3]

cusparseSpMatDescr_t mfem::SparseMatrix::matA_descr
mutableprotected

Definition at line 111 of file sparsemat.hpp.

◆ matA_descr [2/3]

cusparseMatDescr_t mfem::SparseMatrix::matA_descr
mutableprotected

Definition at line 115 of file sparsemat.hpp.

◆ matA_descr [3/3]

hipsparseSpMatDescr_t mfem::SparseMatrix::matA_descr
mutableprotected

Definition at line 123 of file sparsemat.hpp.

◆ NodesMem

RowNodeAlloc* mfem::SparseMatrix::NodesMem
protected

Definition at line 84 of file sparsemat.hpp.

◆ Rows

RowNode** mfem::SparseMatrix::Rows
protected

Array of linked lists, one for every row. This array represents the linked list (LIL) storage format.

Definition at line 73 of file sparsemat.hpp.

◆ SparseMatrixCount

int mfem::SparseMatrix::SparseMatrixCount = 0
staticprotected

Definition at line 100 of file sparsemat.hpp.

◆ status [1/2]

cusparseStatus_t mfem::SparseMatrix::status
protected

Definition at line 106 of file sparsemat.hpp.

◆ status [2/2]

hipsparseStatus_t mfem::SparseMatrix::status
protected

Definition at line 119 of file sparsemat.hpp.

◆ useGPUSparse

bool mfem::SparseMatrix::useGPUSparse = true
protected

Definition at line 93 of file sparsemat.hpp.

◆ vecX_descr [1/2]

cusparseDnVecDescr_t mfem::SparseMatrix::vecX_descr
mutableprotected

Definition at line 112 of file sparsemat.hpp.

◆ vecX_descr [2/2]

hipsparseDnVecDescr_t mfem::SparseMatrix::vecX_descr
mutableprotected

Definition at line 124 of file sparsemat.hpp.

◆ vecY_descr [1/2]

cusparseDnVecDescr_t mfem::SparseMatrix::vecY_descr
mutableprotected

Definition at line 113 of file sparsemat.hpp.

◆ vecY_descr [2/2]

hipsparseDnVecDescr_t mfem::SparseMatrix::vecY_descr
mutableprotected

Definition at line 125 of file sparsemat.hpp.


The documentation for this class was generated from the following files: