fn norm_inf(v: &[DoubleDouble]) -> f64
‖v‖∞ of the leading components; NaN if a component is NaN.
‖v‖∞