As the final assignment for the course you have the following options:

1. Submit a project report on a project we have already discussed. This was the original plan and if you have already discussed this option with me please proceed as agreed. 2. Read and summarize the following paper: Self-Optimizing Memory Controllers: A Reinforcement Learning Approach, Ipek et al.

Your summary should include the following:

1. What is the motivation for this work? What problem/need/opportunity it targets? Why is this important? 2. Explain briefly how DRAM memory works at a high level? What is the role of a memory controller? 3. What is the FR-FCFS policy? Provide an example where it works well and another where it is suboptimal. 4. Why is the task of a memory controller hard to solve? 5. What is the method the authors propose? How it relates to reinforcement learning? 6 Explain the hardware implementation? What is/are the key hardware optimizations/compromises made? 7. Summarize key results. Pick one graph and try to explain or speculated as to what the explanation might me behind what the graph reports.