For the following pattern of branch outcomes: (T for branch taken and NT for branch not taken)
T, T, T, T, NT, NT, NT, T
a) What is the accuracy of always-taken and always-not-taken predictors?
b) What is the accuracy of a one-bit predictor, assuming the predictor starts off predicting not-taken? What if the pattern repeats forever?
c) What is the accuracy of a two-bit predictor, assuming the predictor starts off in the bottom left state? What if the pattern repeats forever?