pub fn parallel_sparse_matvec_csr<T>(
y: &mut [T],
rows: usize,
indptr: &[usize],
indices: &[usize],
data: &[T],
x: &[T],
options: Option<ParallelVectorOptions>,
)
Expand description
Parallel sparse matrix-vector multiplication for CSR format
Performs y = A * x where A is a sparse matrix in CSR format, using parallel processing and SIMD acceleration when beneficial.
§Arguments
y
- Output vector (will be overwritten)rows
- Number of rows in the matrixindptr
- Row pointer array (length = rows + 1)indices
- Column indices arraydata
- Non-zero values arrayx
- Input vectoroptions
- Optional configuration (uses default if None)
§Panics
Panics if array dimensions are inconsistent