Teaching Demo: CPU Caches

Abstract: This teaching demo covers the fundamental elements of CPU caches and their role in the memory hierarchy. It begins by covering general cache design and organization. We will explore how caches operate when a processor executes code and how caches interpret memory addresses. We initially focus on direct-mapped caches and explore their behavior in several different scenarios. Finally, we will motivate and introduce cache associativity as a means of improving system performance.

Bio: Jack Kolb recently finished his PhD at the University of California, Berkeley where he was advised by Randy Katz and David Culler. His research has covered a variety of topics including distributed computing, parallel computing for scientific applications, blockchains, and applications of machine learning to education. Jack has taught courses in computer organization and operating systems and is a recipient of the Berkeley CS department’s outstanding graduate student instructor award.