Answer Happy

a.)
Consider the following repeating pattern (e.g., in a loop)
of branch outcomes: (T means the branch is taken. N
means the branch is not taken).
T, T, N, T, N, T
For a one-bit predictor that initially predicts T, what is its
accuracy of predictions, assuming the pattern repeats forever?
Select one:
A.
2/6
B.
3/6
C.
4/6
D.
5/6
E.
1/6
b.)
Consider the following repeating pattern (e.g., in a loop)
of branch outcomes: (T means the branch is taken. N
means the branch is not taken).
T, T, N, T, N, T
For a two-bit predictor that initially predicts weakly taken T,
what would be its accuracy of predictions if the pattern repeats
forever?
Select one:
A.
2/6
B.
3/6
C.
4/6
D.
5/6
E.
6/6
c)
We know that the memory stall cycles due to cache misses can be
a significant factor in a computer's performance. Assume a base CPI
= 1.5 for a perfect cache that never misses. Consider a one-level
L1 cache that has average combined miss ratio per instruction of 3%
for the program P, and the main memory access time of
100 cycles. What is the total effective CPI for the computer with
this cache?
Select one:
A.
1.5
B.
2.5
C.
3.5
D.
4.5
E.
5.5
d.)
If we double the clock rate of processor in previous question
without changing the memory system, which of following statement is
true regarding the CPI and speedup of new processor compared to the
previous one?
Select one:
A.
CPI increases, speedup < 2
B.
CPI decreases, speedup = 2
C.
CPI decreases, speedup < 2
D.
CPI increases, speedup = 2
E.
None of the above
e.)
In order to get better performance, we may add an L2 cache
between the L1 cache and main memory in question 34. Assuming the
L2 cache has access time of 10 cycles and reduces the global miss
rate per instruction to main memory down to 0.5%, what is the total
effective CPI now?
Select one:
A.
1.5
B.
2.0
C.
2.9
D.
2.3
E.
3.5