Assume that you have a program that would take 20s to execute if you were running the program on infinitely many cores (if that was possible). Assume further that if you ran the program on 8 cores, you would get a speedup of 4. Assume also that N number of cores result in N times improvement for the part of the program that is possible to parallelize. What would then the speedup be if you executed the program on 20 cores