Question

Consider a processor and a program that would have an IPC of 1 with a perfect 1-cycle L1 cache (All accesses are hit). Assume that each additional cycle for cache/memory access causes program execution time to increase by one cycle and 50% of instructions are loads/stores and the L1-Instruction cache hit rate is 100%. Assume the following MPK MIs(Miss Per Kilo Memory Instruction) and latencies for the following caches:

• L1:32 KB: 1-cycle: 80 MPKMI
• L2: 256 KB: 10-cycle: 50 MPKMI
• L3:2 MB: 30-cycle: 20 MPKMI
• L4:32 MB: 100-cycle: 5 MPKMI.
• Memory: 250-cycles

Estimate the CPI for the following cache configurations:

a. L1-L2-L3-L4
b. L1-L2-L3
c. L2-L3-L4
d. L1-L2-L4

Answer 1

Check the explanation

Step-by-step explanation:

check the below presentation to see the answer to all the questions above

1. L1-L2-L3-L4

1000+ 80 X 10 + 50 X 30+20 X 100+5 X 250 = 6550

2. L1-L2-L3

1000 + 80 X 10 + 50 X 30 + 20 X 250 = 8300

3. L2-L3-L4

1000 X 10 + 50 X 30 + 20 X 100 + 5 X 250 = 14750

4. L1-L2-L4

1000 + 80 X 10 + 50 X 100 + 5 X 250 = 8050