Computing the mutation score using a randomly selected subset of the mutants seems to work well, though it hasn't been studied a lot. One paper is "Reducing th

Computing the mutation score using a randomly selected subset of the mutants seems to work well, though it hasn't been studied a lot. One paper is "Reducing the cost of mutation testing: An empirical study" by Wong and Mathur.