Function

sparse_matrix_variable_block_create_float

Returns a single-precision sparse matrix object that is stored in block-entry format with a variable block size.

Declaration

sparse_matrix_float sparse_matrix_variable_block_create_float(sparse_dimension Mb, sparse_dimension Nb, const sparse_dimension *K, const sparse_dimension *L);

Parameters

Mb

The number of rows in terms of blocks of the matrix. Must be greater than 0.

Nb

The number of columns in terms of blocks of the matrix. Must be greater than 0.

K

Array containing row dimensions of the blocks. The i'th row in terms of blocks will have a dimension K[i]. K is expected to hold Mb elements. All values of K are expected to be greater than 0.

L

Array containing column dimensions of the blocks. The j'th column in terms of blocks will have a dimension L[j]. L is expected to hold Nb elements. All values of L are expected to be greater than 0.

Return Value

A matrix object that is ready for receiving entries. If an error occurs, nil is returned.

Discussion

Create a sparse matrix object that is stored in block-entry format and is ready to receive values from the various block insert routines. Blocks are are of variable dimension where the i,j'th block index has a dimension K[i] * L[j]. Block-entry format means blocks of dense regions will be stored at block indices i,j. For point wise format use non block version of create. See the various insert routines for details on inserting values into this matrix object.

See Also

Matrix Creation and Population

sparse_matrix_block_create_double

Returns a double-precision sparse matrix object that is stored in block-entry format with a fixed block size.

sparse_matrix_block_create_float

Returns a single-precision sparse matrix object that is stored in block-entry format with a fixed block size.

sparse_matrix_variable_block_create_double

Returns a double-precision sparse matrix object that is stored in block-entry format with a variable block size.

sparse_insert_block_double

Inserts a dense block of entries into a double-precision matrix.

sparse_insert_block_float

Inserts a dense block of entries into a single-precision matrix.