Question: (e) Suppose two different compilers, compiler A and compiler B, are used to compile a program P thereby generating two versions, PA and PB, of

(e) Suppose two different compilers, compiler A and compiler B, are used to compile a program P thereby generating two versions, PA and PB, of program P. The table below shows the latency and dynamic instruction count of the program P. i. Compute the average overall CPI for program PA and PB if each program executes on a CPU with an operating time period of 1 ns per cycle. ii. Assume the average CPIs computed for PA and PB found in part (i), except that now programs PA and PB execute on two different CPUs. If the execution times on the two CPUs are the same, how much faster is the clock frequency of the CPU executing program PA versus the clock frequency of the CPU executing program PB? iii. A new compiler C is designed. Program P is compiled on compiler C and compiler C generates program version program PC. Program PC comprises 600M instructions and has an average CPI of 1.1 cycles/instr. when executing on the CPU of part (i). What is the speedup of executing program PC on the CPU of part (i) versus executing programs PA and PB on the same CPU
Step by Step Solution
There are 3 Steps involved in it
Get step-by-step solutions from verified subject matter experts
