Assume that we have a program that contains 500 million arithmetic instructions, 300 million load/store instructions, and 100 million branch instructions. We want to run this program on two different computers of the same ISA and assume that each computer operates at 3GHz. On the first computer: - arithmetic instructions have CPI of 1 - load/store operations have CPI of 10, and - branch instructions have CPI of 3. On the other hand, on the second computer: - arithmetic instructions have CPI of 4 - load/store instructions have CPI of 40, and - branch instructions have CPI of 3. a) What would be the execution time of this program on these machines? b) Assume that we added more powerful arithmetic instructions to the given instruction set. On average, by using these more powerful arithmetic instructions, we can reduce the number of arithmetic instructions needed by 25%; however, the clock cycle increases by 10%. Is this a good design choice? Why (show the impact of this change on the performance of both computers)