I am doing a theoretical exercise for a class about a theoretical study of parallelizing the mergesort algorithm and the speed-up obtained for different amounts of cores.