We plan to show graphs based on the speedups of a smart approach to our prefetching techniques compared to simple guesses. Also when prefetching is more useful based on the two protocols that we will be observing.

Our concerning issues is the steep learning curve in order to work effectively. Once we figure out how to manipulate gem5 to accept the pragma prefetching and how to manipulate the coherence protocol on demand, then we should be in good shape to just begin the coding and testing aspect of our project.

Met with Professor Mowry to rework our project.