CS_CRS< T > Class Template Reference

The compressed row storage sparse matrix data structure. More...

#include <CS_CRS.hpp>

Collaboration diagram for CS_CRS< T >:

[legend]List of all members.

Public Member Functions
	CS_CRS (const ULI number_of_nonzeros, const ULI number_of_rows, const ULI number_of_cols, const T zero)
	Base constructor which only initialises the internal arrays.
	CS_CRS (CRS< T > &toCopy)
	Copy constructor.
	CS_CRS (CS_CRS< T > &toCopy)
	Copy constructor.
	CS_CRS (std::vector< Triplet< T > > input, unsigned long int currow, unsigned long int curcol, T &zero)
	Constructor which transforms a collection of input triplets to CRS format.
	CS_CRS (std::vector< Triplet< T > > input, T &zero)
	Constructor which transforms a collection of input triplets to CRS format.
T &	random_access (ULI i, ULI j)
	Method which provides random matrix access to the stored sparse matrix.
T *	zax (T *x_orig)
	Calculates and returns z=Ax.
	~CS_CRS ()
	Base deconstructor.
Protected Member Functions
bool	find (ULI col_index, ULI search_start, ULI search_end, ULI &ret)
	Helper function which finds a value with a given column index on a given subrange of indeces.
Static Protected Member Functions
int	compareTriplets (const void *left, const void *right)
	Comparation function for 1D columnwise sort.
Protected Attributes
CS_ELEMENT< ULI > *	nnz
	Number of non-zeros.
CS_ELEMENT< ULI > *	nor
	Number of rows.
CS_ELEMENT< ULI > *	noc
	Number of columns.
CS_ARRAY< ULI > *	row_start
	Array keeping track of individual row starting indeces.
CS_ARRAY< T > *	nzs
	Array containing the actual nnz non-zeros.
CS_ARRAY< ULI > *	col_ind
	Array containing the column indeces corresponding to the elements in nzs.
CS_ELEMENT< T > *	zero_element
	Which element is considered as zero.

---

Detailed Description

template<typename T>
class CS_CRS< T >

The compressed row storage sparse matrix data structure.

---

Constructor & Destructor Documentation

template<typename T> CS_CRS< T >::CS_CRS (  const ULI  number_of_nonzeros, const ULI  number_of_rows, const ULI  number_of_cols, const T  zero )  [inline]

Base constructor which only initialises the internal arrays. Note that to gain a valid CRS structure, these arrays have to be filled by some external mechanism (i.e., after calling this constructor, the internal arrays contain garbage, resuling in invalid datastructure). Parameters: number_of_nonzeros  The number of non-zeros of the matrix to be stored. number_of_rows  The number of rows of the matrix to be stored. number_of_cols  The number of columns of the matrix to be stored. zero  Which element is considered to be the zero element.

template<typename T> CS_CRS< T >::CS_CRS (  CS_CRS< T > &  toCopy  )  [inline]

Copy constructor. Parameters: toCopy  reference to the CS_CRS datastructure to copy.

template<typename T> CS_CRS< T >::CS_CRS (  std::vector< Triplet< T > >  input, unsigned long int  currow, unsigned long int  curcol, T &  zero )  [inline]

Constructor which transforms a collection of input triplets to CRS format. The input collection is considered to have at most one triplet with unique pairs of indeces. Unspecified behaviour occurs when this assumption is not met. Parameters: input  The input collection. currow  The number of rows of the input matrix. curcol  The number of columns of the input matrix. zero  Which element is considered zero.

template<typename T> CS_CRS< T >::CS_CRS (  std::vector< Triplet< T > >  input, T &  zero )  [inline]

Constructor which transforms a collection of input triplets to CRS format. The input collection is considered to have at most one triplet with unique pairs of indeces. Unspecified behaviour occurs when this assumption is not met. Parameters: input  The input collection. zero  Which element is considered zero.

---

Member Function Documentation

template<typename T> int CS_CRS< T >::compareTriplets (  const void *  left, const void *  right )  [inline, static, protected]

Comparation function for 1D columnwise sort. Parameters: left  The triplet at the left side of the comparation operator < right  The triplet at the right side of the comparation operator < Returns: -1 iff the column of left is less than that of right, 1 if it is larger, 0 if equal.

template<typename T> bool CS_CRS< T >::find (  ULI  col_index, ULI  search_start, ULI  search_end, ULI &  ret )  [inline, protected]

Helper function which finds a value with a given column index on a given subrange of indeces. Parameters: col_index  The given column index. search_start  The start index of the subrange (inclusive). search_end  The end index of the subrange (exlusive). ret  Reference to the variable where the return *index* is stored. Returns: Whether or not a non-zero value should be returned.

template<typename T> T& CS_CRS< T >::random_access (  ULI  i, ULI  j )  [inline]

Method which provides random matrix access to the stored sparse matrix. Parameters: i  Row index. j  Column index. Returns: Matrix valuei at (i,j).

template<typename T> T* CS_CRS< T >::zax (  T *  x_orig  )  [inline]

Calculates and returns z=Ax. The vectors x should have appropiately set length; this is not enforced. Memory leaks, segmentation faults, the works; one or more of these will occur if dimensions do not make sense. The return array is allocated to a length equal to the number of rows in this function. Cache simulation is only performed on access to elements from A, x or z. Parameters: x_orig  The input (dense) x-vector. Returns: The matrix-vector multiplication Ax, where A corresponds to the currently stored matrix.

---

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

• CS_CRS.hpp

---