Returns the first derivative of the error function.
double dxErf(num x) => 2.0 * invSqrtPi * math.exp(-x * x);