Why is np.linalg.norm(…, axis=1) slower than writing out the formula for vector norms?

Question

To normalize the rows of a matrix X to unit length, I usually proceed as follows:

X /= np.linalg.norm(X, axis=1, keepdims=True)