Relating to the homework, this seems computationally very slow presumably because of the number of calculations you need to perform per sample. What are other ways to take into account computational efficiency?