1. Assume we use branch history table (1-bit BHT) for the bnez instruction, calculate the miss prediction rate (assume n = 500 in this case).

2. For this code, will the 2-bit predictio do better than the 1-bit prediction? Why or why not?

3. Give an example (in MIPS code or in Java/C++ like code) that demonstrates a 2-bit prediction scheme is better than the 1-bit prediction scheme. (Clearly explain why it is better.)