3. (6 points) On a pipelined Mrps processor. the berakdown of imwrucrions executed for as Application is ahown below, 10 s of the memory accewse rewilt in ene eysle pipotine mall. a) Assume all branch instructions have one cycle penalty as the instruction following the branch is delayed. What is the CPI of the application? Show all the calculation steps. (3 points) b) Assume in a certain design, if the branch is not taken, there is no performance penalty. If the branch is taken, there is one cycle penalty. Suppose 60% of the branch instructions are taken. What is the CPI of the application? Show all the calculation steps. (3 points)