ConceptPythonIntermediate
Conjugate gradients and the condition number: why iterations grow as √κ
- ends = np.sort(np.concatenate([eigsh(A, k=1, which="LA", v0=v0, return_eigenvectors=False),
- eigsh(A, k=1, sigma=0, v0=v0, return_eigenvectors=False)]))
- top = eigsh(A, k=1, M=M_ssor, which="LA", v0=v0, return_eigenvectors=False)[0]
- bottom = eigsh(A, k=1, M=M_ssor, sigma=0, v0=v0, return_eigenvectors=False)[0]