num::sparse_resolvent

Defined in header <linear/solvers/sparse_resolvent.hpp>

class sparse_resolvent;

The sparsity of \(sI - A\) analyzed once, ready to factor it for any shift.

With SuiteSparse UMFPACK the symbolic analysis is kept and each shift(R, s) does only the numeric factorization. Without it, shift reports that no sparse complex backend is available rather than silently densifying a large matrix. Copies share the analysis.

Members

member	description
explicit sparse_resolvent(const spmat &A, sparse_resolvent_options options = {})	Analyze A's sparsity pattern. A must store its diagonal explicitly.
idx size() const noexcept