pub fn l2_error(computed: &[f64], exact: &[f64], dx: f64) -> f64
Computes the discrete L2 error between computed and exact solutions.