COMP_ENG 456: Modern Topics in Computer Architecture



COMP_ENG 452 – Advanced Computer Architecture I (or equivalent out-of-order processor architecture course); Useful but not required: COMP_ENG 453 – Parallel Architectures


This course examines fundamental issues and design trade-offs in modern processor architectures. We will discuss some of the constraints that limit the design and programmability of modern processors, and promising techniques to mitigate these constraints. As such, we will draw material from seminal and recent publications in top computer architecture conferences and journals. By exposing the students to state-of-the-art research, the course serves as an entry point to further research in computer architecture. The course will cover a sample of research across a wide spectrum of topics from emerging architectures, including quantum computing, neuromorphic computing, space-time computing, silicon photonics in computer architectures, and advanced techniques in more traditional topics, including memory systems and advanced cache designs, memory consistency models and operational semantics, programmability (e.g., transactional memory, deterministic programming, speculative multithreading), on-chip interconnects, power/thermal management, reliability, and fault tolerance. The exact collection of topics varies across offerings, as the field itself evolves rapidly. The course has a seminar format, and the students are expected to lead multiple presentations throughout the quarter. The course requires the completion of a project in computer architecture. The project component of the course is open-ended, and students are encouraged to draw on their own research interests and prior background for inspiration.

REQUIRED TEXT: None; we'll draw material from seminal and recent publications in top conferences, as well as chapters from the Synthesis Lectures on Computer Architecture by Morgan & Claypool.

COURSE COORDINATOR: Prof. Nikos Hardavellas

COURSE OBJECTIVES: The course aims to offer a firm background for research in computer architecture. The students that successfully complete the course will be exposed to a variety of cutting-edge research topics, be able to read and critique research publications in computer architecture, perform conference-quality paper reviews, perform research presentations, write research reports, and gain familiarity with state-of-the-art tools for research in computer architecture. Along with providing technical knowledge, the course also aims to develop the student's ethos as researchers and research referees, and sharpen the students team-participation skills.

GRADES: Grades are based on class presentations, project, homework assignments (if any), and class participation.

EXAMS: There are no exams in this course.

ABET CONTENT: 100% Engineering