vibeqc.CPHFOptions

class vibeqc.CPHFOptions(max_iter=100, tol=1e-08, use_preconditioner=True)[source]

Bases: object

Knobs for cphf_solve_rhf().

max_iter

Conjugate-gradient iteration cap. Most cases converge in 20–60 iterations on default basis sets; increase if you see CPHFConvergenceError.

tol

Relative residual tolerance ||r|| < tol·||rhs||. Default 1e-8 — tight enough for analytic Hessian work, loose enough that SCF residual noise doesn’t dominate.

use_preconditioner

When True (default), use \((\varepsilon_a - \varepsilon_i)^{-1}\) as a CG preconditioner. Disable only for diagnostics — convergence is much slower without it.

Parameters:
__init__(max_iter=100, tol=1e-08, use_preconditioner=True)
Parameters:
Return type:

None

Methods

__init__([max_iter, tol, use_preconditioner])

Attributes

max_iter: int = 100
tol: float = 1e-08
use_preconditioner: bool = True