pub fn csr_matvec<T>( mat: &CompressedSparseData2<T>, vec: &[T], ) -> Result<Vec<T>, SparseErrors>where T: BixverseNumeric, <T as Add>::Output: PartialEq<T>,
Sparse matrix @ dense vector: M @ v.
M @ v
Computes result[i] = sum_j M[i, j] * v[j] by iterating CSR rows.
result[i] = sum_j M[i, j] * v[j]
mat
(m, n)
vec
n
Dense vector of length m. Errors if mat is not CSR.
m