CART Academic Seminar Courses


CS395T: Readings in Historical Computer Architecture (Burger, Spring 2000)

My goal for this course is to learn about many of the important historical machines and ideas by reading classic, seminal papers. Understanding what has gone before is essential for understanding the evolution of our fast-changing field. An added bonus is that we extend our bag of design tricks.

CS395T: Embedded Architectures and Applications (Keckler, Fall 1999)

In this course, students will examine the architectures and applications of embedded microprocessor based systems. A subset of the topics include digital signal processing, mobile consumer electronic devices, real-time applications, reliability, and power management. This is primarily a reading course in which students will present and discuss papers during class meetings. For more details on this course, please visit the course home page.

CS395T: Applications for Billion-Transistor Architectures (Burger, Spring 1999)

In this research seminar, led by Prof. Doug Burger, students will explore the interaction between key emerging applications and architectural trends. Students will learn about cutting-edge architectures, they will gain a broad view of the emerging application space, and they will study one application each in depth, each characterizing how their application is likely to interact with emerging architectures. For more details on this course, please visit the course home page.

CS395T: Technology Driven Computer Architecture (Keckler, Fall 1998).

An examination of computer architectures for billion transistor chips from the perspective of technological capabilities and constraints. Readings focus both on momentous technological breakthroughs and their implications, including high density solid-state memory, interconnection networks, and notable milestones of integration measured in transistors per chip. The topics of discussion include VLSI scaling, integrated DRAM and processors, on and off-chip communication bandwidth, power consumption, tradeoffs between hardware and software, and novel techniques for using billions of transistors on a single chip.