115k views
1 vote
Does out-of-order makea machine's performance more or less

sensitive to branch prediction accuracy? Explain.

User Shayan RC
by
6.4k points

1 Answer

6 votes

Answer:

It would be correct to say that out-of-order makes a machine's performance more sensitive to branch prediction accuracy.

Step-by-step explanation:

This can be explained as when a machine is out-of-order, in that state the execution holds importance in prediction accuracy, any increase in these results in rate of prediction near about 25% for the single-issue operating in-order. This is due to the reason that some of the predictions are required for the global pattern history. Most recent outcomes are recorded in the register and for a 4-way machine which is out-of-order, accuracy is very poor as a result of the delay of the branch history for next prediction.

User Mithun Rana
by
7.3k points