Module serial 

Serial sparse matrix arithmetic routines.

All routines are single-threaded.

Some operations have the prealloc suffix. This means that they expect that the sparsity pattern of the output matrix has already been pre-allocated: that is, the pattern of the result of the operation fits entirely in the output pattern. In the future, there will also be some operations which will be able to dynamically adapt the output pattern to fit the result, but these have yet to be implemented.

Structs

OperationError

A description of the error that occurred during an arithmetic operation.

Enums

OperationErrorKind

The different kinds of operation errors that may occur.

Functions

spadd_csc_prealloc

Sparse matrix addition C <- beta * C + alpha * op(A).

spadd_csr_prealloc

Sparse matrix addition C <- beta * C + alpha * op(A).

spadd_pattern

Sparse matrix addition pattern construction, C <- A + B.

spmm_csc_dense

Sparse-dense matrix-matrix multiplication C <- beta * C + alpha * op(A) * op(B).

spmm_csc_pattern

Sparse matrix multiplication pattern construction, C <- A * B.

spmm_csc_prealloc

Sparse-sparse matrix multiplication, C <- beta * C + alpha * op(A) * op(B).

spmm_csc_prealloc_unchecked

Faster sparse-sparse matrix multiplication, C <- beta * C + alpha * op(A) * op(B). This will not return an error even if the patterns don’t match. Should be used for situations where pattern creation immediately precedes multiplication.

spmm_csr_dense

Sparse-dense matrix-matrix multiplication C <- beta * C + alpha * op(A) * op(B).

spmm_csr_pattern

Sparse matrix multiplication pattern construction, C <- A * B.

spmm_csr_prealloc

Sparse-sparse matrix multiplication, C <- beta * C + alpha * op(A) * op(B).

spmm_csr_prealloc_unchecked

Faster sparse-sparse matrix multiplication, C <- beta * C + alpha * op(A) * op(B). This will not return an error even if the patterns don’t match. Should be used for situations where pattern creation immediately precedes multiplication.

spsolve_csc_lower_triangular

Solve the lower triangular system op(L) X = B.