“All stages are perfectly balanced” means ALL STAGES TAKE SAME TIME (say, t time)
“No cycle time overhead of pipelining” means there is no buffer
Given , k=6
for Non-pipelined processor, tn=6t and CPI=1
for Pipelined Processor, tp= MAX(t,t,t,t,t,t)=t
FINDING CPI:
say, we have n-instructions
25% of n-instructions incur 2 stalls each
CPI= (n + 0.25*n*2)/n = 1.5
Speedup= slow process/ fast process
= 1 instruction execution time for non-pipeline/ 1 instruction execution time for pipeline
= (CPI for non-pipeline)* tn / (CPI for pipeline)*tp
= 1*6t/1.5*t
=4