pub fn eigsh<T, O>(
op: &mut O,
k: usize,
which: Which,
opts: &EigsOptions<T>,
) -> Result<SymEigen<T>, StatlaError>Expand description
k eigenpairs of a symmetric operator (the caller asserts the
symmetry; it is not checked).
Algorithm KS of Chapter 12.6 with the symmetric small solver — the
thick-restart Lanczos method (Algorithm L of the contract): full DGKS
reorthogonalization, Jacobi on the projected arrowhead, Krylov–Schur
truncation with locking, ARPACK’s test ρ ≤ tol·max(ε^{2/3}, |θ|).
Acceptance (L3.4): on symmetric Matrix Market matrices the k
extreme eigenvalues agree with scipy.sparse.linalg.eigsh (ARPACK) to
1e-10 relative where the eigenvalue gap is ≥ 1e-6 ‖A‖ (S9), and
every pair has ‖Ay − θy‖ ≤ (tol + c_K·ncv·u) ‖A‖ (c_K = 8).
§Errors
StatlaError::InvalidArgument (not square; which not symmetric; k = 0; k ≥ n; ncv ∉ (k, n]; tol < 0 or not finite; reorth_eta ∉ (0, 1); v0 = 0), StatlaError::Dimension (v0),
StatlaError::NonFinite (v0, or the operator returned NaN/∞),
StatlaError::Operator, StatlaError::NoConvergence.