Is it possible to compute this p-value faster ? In the case where R admits a diagonal form, simple linear algebra could help to cut off the computations and answer yes to this question.