Module sparse_eigen 

Sparse eigenvalue solvers (Lanczos, Arnoldi). Sparse Eigenvalue Solvers — Lanczos and Arnoldi

Deterministic sparse eigenvalue extraction for symmetric (Lanczos) and general (Arnoldi) matrices. All operations use Kahan summation for order-invariant deterministic reductions.

Structs

ArnoldiResult

Result of an Arnoldi eigenvalue computation for a general (non-symmetric) sparse matrix.

LanczosResult

Result of a Lanczos eigenvalue computation.

Functions

arnoldi_eigs

Compute eigenvalue approximations of a general sparse matrix using the Arnoldi iteration to build an upper Hessenberg matrix, then extract eigenvalues from the Hessenberg form.

lanczos_eigsh

Compute the k largest eigenvalues of a symmetric sparse matrix using the Lanczos algorithm.