Read Chapter 6. The material covers performance bottlenecks (the"weakest" links in the design of a system, in terms of speed), multilevel memories (using caching to avoid penalties of using slower hardware), and scheduling (running tasks in an order that reduces their total completion time or some other performance metric).