Function step_variance 

fn step_variance<R: Scalar>(
    params: &HullWhiteParameters<R>,
    t0: f64,
    t1: f64,
) -> R

The conditional variance of x over one step from t0 to t1: σ²(1−e^{−2a·dt})/(2a) for a flat volatility, degenerating to σ²·dt as a → 0 (the branch selected on the passive value); under a term, the term’s own V(t0, t1), so a piece boundary inside the step is integrated over it rather than snapped to a slice. The arithmetic runs in the mode.