Time and space efficient representation of sparse {0,1}-matrices. More...
#include <zo.h>
Public Member Functions | |
template<class OutVector , class InVector > | |
OutVector & | apply (OutVector &y, const InVector &x) const |
Uses one of the three private utility functions. It calls the generalized utility function _apply if there is no special ordering, _fyapply if there is C_ordering or _fxapply if there is fortran_ordering. | |
template<class OutVector , class InVector > | |
OutVector & | applyTranspose (OutVector &y, const InVector &x) const |
Uses one of the three private utility functions, in the manner described above. Worthy of note is the fact that applyTranspose works by passing the column positions to the _apply functions as if they were rows, and row positions as if they were columns, as if the matrix had been transposed. | |
std::istream & | read (std::istream &is) |
ZeroOne (Field &F, Index *rowP, Index *colP, Index rows, Index cols, Index NNz) | |
template<class OutVector , class InVector > | |
OutVector & | apply (OutVector &y, const InVector &x) const |
Uses one of the three private utility functions. It calls the generalized utility function _apply if there is no special ordering, _fyapply if there is C_ordering or _fxapply if there is fortran_ordering. | |
template<class OutVector , class InVector > | |
OutVector & | applyTranspose (OutVector &y, const InVector &x) const |
Uses one of the three private utility functions, in the manner described above. Worthy of note is the fact that applyTranspose works by passing the column positions to the _apply functions as if they were rows, and row positions as if they were columns, as if the matrix had been transposed. | |
std::istream & | read (std::istream &is) |
Time and space efficient representation of sparse {0,1}-matrices.
A 0-1 matrix is a matrix with all 0's and 1's as entries. We're using a NAG-sparse format. Applies can be performed fast, using only additions. When initalizing this class, you only need to build 2 arrays of equal length: an array of the row indices for the non-zero (1's) entries, and an array of the column indices for the non-zero (1's) entries.
A {0, 1,-1} matrix can be effecively represented as the Dif of two ZeroOne's.
A 0-1 matrix is a matrix with all 0's and 1's as entries. We're using a comp-col or comp-row format. That is we have an array of col indices and an array of pointers indicating where the col indices for each row begins within the col index array. (or vice versa if we have sorted by columns.
Applies can be performed fast, using only additions. When initalizing this class, you only need to build 2 arrays of equal length: an array of the row indices for the non-zero (1's) entries, and an array of the column indices for the non-zero (1's) entries.
A {0, 1,-1} matrix can be effecively represented as the Dif of two ZeroOne's.
ZeroOne | ( | Field & | F, | |
Index * | rowP, | |||
Index * | colP, | |||
Index | rows, | |||
Index | cols, | |||
Index | NNz | |||
) | [inline] |
The real constructor /todo give docs here assuming entries are sorted in lexicographic order by (row,col) pair.
std::istream& read | ( | std::istream & | is | ) | [inline] |
Read the matrix from a stream in the JGD's SMS format
is | Input stream from which to read the matrix |
std::istream& read | ( | std::istream & | is | ) | [inline] |
Read the matrix from a stream in the JGD's SMS format
is | Input stream from which to read the matrix |