Quantile regression is a nondifferentiable convex optimization problem. We compare three classical numerical methods, two of them based on linear optimization, and the cutting plane method. We compare them by their required memory and computing time.