Consider the 5-stage pipeline design we discussed in the lecture, shown in Figure 1, without forwarding paths. Optimize execution time by re-ordering instructions and then insert the minimum number of required NOPs into the instruction stream. The re-ordering must be correct, i.e. not change the output result of the code when applying transformations. Calculate the average CPI of this code, assuming a CPI of one for all instructions. Assume that registers from the register file can be written and read in the same cycle.

Figure : 15 -stage Pipeline without forwarding