Question 4: Pipeline Performance Calculation (1 Point) Consi…
Question 4: Pipeline Performance Calculation (1 Point) Consider a pipeline with the following stages and their respective delays: Stage 1 (Fetch): 200 ps (can be parallelized over 2 units, each unit costs 150 ps) Stage 2 (Decode): 150 ps (cannot be parallelized) Stage 3 (Execute): 300 ps (can be parallelized over 3 units, each unit costs 120 ps) Stage 4 (Memory): 250 ps (cannot be parallelized) Stage 5 (Write Back): 100 ps (can be parallelized over 2 units, one unit costs 70ps, and the other costs 90ps) a) Calculate the proper clock frequency for this pipeline to achieve the highest performance without any timing violations. b) Calculate the total delay to process one instruction without any parallelization, assuming we would like to achieve the highest performance without any timing violations. c) Calculate the new stage delays when parallelization is applied as specified for each of the stages, and determine the new total delay to process one instruction, assuming we would like to achieve the highest performance without any timing violations. d) Compute the speedup achieved through parallelization, assuming the pipeline can be fully loaded. e) Based on the parallelized processor in (c) Assuming we need to run 10 instructions and there is no hazard, compute the total delay to finish all the 10 instructions.