(Solved) : 5 24 Pts Program Executes Following Mixture Instructions Instruction Type Execution Time C Q44118261 . . .
5 (24 pts) A program executes the following mixture of instructions: Instruction Type Execution Time (clock cycles) 10 Fractions of instructions executed 0.3 0.4 0.2 0.1 Loads and stores Integer ALU FP ALU Branch 40 a) What is the average CPI for this program? b) What is the average MIPS rate for this program? (Assume clock period is Ins.) c) You decide to design a new version of this processor with a faster floating-point ALU to improve the processor’s performance when executing this program. What speedup would you expect to see if you could reduce the number of cycles required to execute all floating-point instructions by 5 clock cycles? d) What is the maximum number of clock cycles each floating point operation can take to obtain an overall speedup for this program of 7 percent? e) What is the theoretical limit to the speedup you could obtain by improving the time required to execute floating-point instructions? f) To obtain the biggest performance improvement for the amount of effort expended in redesigning this processor, on which one of the instruction types (load/store, floating point, integer ALU, or branch) should you concentrate your effort? Quantitatively justify your answer. Show transcribed image text 5 (24 pts) A program executes the following mixture of instructions: Instruction Type Execution Time (clock cycles) 10 Fractions of instructions executed 0.3 0.4 0.2 0.1 Loads and stores Integer ALU FP ALU Branch 40 a) What is the average CPI for this program? b) What is the average MIPS rate for this program? (Assume clock period is Ins.) c) You decide to design a new version of this processor with a faster floating-point ALU to improve the processor’s performance when executing this program. What speedup would you expect to see if you could reduce the number of cycles required to execute all floating-point instructions by 5 clock cycles? d) What is the maximum number of clock cycles each floating point operation can take to obtain an overall speedup for this program of 7 percent? e) What is the theoretical limit to the speedup you could obtain by improving the time required to execute floating-point instructions? f) To obtain the biggest performance improvement for the amount of effort expended in redesigning this processor, on which one of the instruction types (load/store, floating point, integer ALU, or branch) should you concentrate your effort? Quantitatively justify your answer.
Expert Answer
Answer to 5 (24 pts) A program executes the following mixture of instructions: Instruction Type Execution Time (clock cycles) 10 F…
OR