Answer:
3.1 ns ; 1.25 ; 3.097
Step-by-step explanation:
Given :
IF, 3 ns;
ID, 2.5 ns;
EX, 2 ns;
MEM, 3 ns;
WB, 1.5 ns.
Use 0.1 ns for the pipelineregisterdelay
maximum time required for MEM = 3 ns
Pipeline register delay = 0.1 ns.
Clock cycled time of the pipelined machine= maximum time required + delay
3ns+0.1 ns = 3.1 ns
2.) for stall after every 4 instruction :
CPI of new machine :
(1 + (1 /4)) = 1 + 0.25 = 1.25
3.)
The speedup of pipelined machine over the single-cycle machine is given by :
Average time per instruction of single cycle ÷ average time per instruction of pipelined
Clock time of original machine = 12ns
Ideal CP1 = 1
CPI of new machine = 1.25
Clock period = 3.1 ns
(12 * 1) / (1.25 * 3.1) = 12 / 3.875
= 3.097
D. Speed up will equal the number of stages in the machine