How do I express the speedup of the merge sort algorithm when I parallelize the divide step? I am trying to solve the following exercise:
