PEXSI::DistSparseMatrix< F > Struct Template Reference

DistSparseMatrix describes a Sparse matrix in the compressed sparse column format (CSC) and distributed with column major partition. More...

#include <sparse_matrix.hpp>

Public Member Functions
LongInt	Nnz ()
	Compute the total number of nonzeros through MPI_Allreduce.

Public Attributes
Int	size
	Matrix dimension.
Int	nnzLocal
	Local number of local nonzeros elements on this processor.
Int	nnz
	Total number of nonzeros elements. More...
IntNumVec	colptrLocal
	Dimension numColLocal + 1, storing the pointers to the nonzero row indices and nonzero values in rowptrLocal and nzvalLocal, respectively. numColLocal is the number of local columns saved on this processor. The indices are 1-based (FORTRAN-convention), i.e. colptrLocal[0] = 1.
IntNumVec	rowindLocal
	Dimension nnzLocal, storing the nonzero row indices. The indices are 1-based (FORTRAN-convention), i.e. the first row index is 1.
NumVec< F >	nzvalLocal
	Dimension nnzLocal, storing the nonzero values.
MPI_Comm	comm
	MPI communicator.

Detailed Description

template<class F>
struct PEXSI::DistSparseMatrix< F >

DistSparseMatrix describes a Sparse matrix in the compressed sparse column format (CSC) and distributed with column major partition.

Note

Since in PEXSI and PPEXSI only symmetric matrix is considered, the compressed sparse row format will also be represented by the compressed sparse column format.

Todo:

Add the parameter of numColLocal

Member Data Documentation

template<class F>

Int PEXSI::DistSparseMatrix< F >::nnz

Total number of nonzeros elements.

FIXME: The datatype should be changed to LongInt in the future.

---

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

• include/sparse_matrix.hpp
• include/sparse_matrix_impl.hpp