Course Schedule and Catalog
Course Schedule - Spring Semester 2027
Meeting location information can now be found on student schedules in ESTHER (for students) or on the Course Roster in ESTHER (for faculty and instructors).Additional information available here.
Meeting location information can now be found on student schedules in ESTHER (for students) or on the Course Roster in ESTHER (for faculty and instructors).Additional information available here.
COMP 437 001 (CRN: 26755)
COMPUTER ARCHITECTURE FOR AI
Long Title: COMPUTER ARCHITECTURE FOR ARTIFICIAL INTELLIGENCE
Department: Computer Science
Instructor: Lee, Kyungmi
Meeting: 10:50AM - 12:05PM TR (11-JAN-2027 - 23-APR-2027)
Part of Term: Full Term
Grade Mode: Standard Letter
Course Type: Lecture
Language of Instruction: Taught in English
Method of Instruction: Face to Face
Credit Hours: 3
Course Syllabus:
Course Materials: Rice Campus Store
Restrictions:
Must be enrolled in one of the following Level(s):
Undergraduate Professional
Visiting Undergraduate
Undergraduate
Prerequisites: COMP 222
Section Max Enrollment: 45
Section Enrolled: 0
Total Cross-list Max Enrollment: 90
Total Cross-list Enrolled: 0
Enrollment data as of: 2-NOV-2026 8:35AM
Final Exam: Scheduled Final Exam-OTR Room
Description: Artificial intelligence is driving growing demand for computation, pushing computer architecture in new directions. Graphics processing units have become the dominant hardware platform for AI, while purpose-built accelerators are emerging to deliver greater performance and efficiency. At the same time, rapidly evolving AI workloads, such as large language models and agentic AI, continue to expose new bottlenecks and inspire new hardware designs. This course explores the computer architecture principles behind modern AI hardware and provides the foundation to reason about why today’s AI accelerators are designed the way they are and how future architectures might evolve.
This course will begin with how AI workloads present data parallelism opportunities and how conventional processors exploit data-level parallelism. We then study GPU architecture in depth, examining its parallelism, memory hierarchy, and recent innovations for AI workloads. Next, we will explore domain-specific accelerators, especially those where computations are mapped across spatial arrays of processing elements, and their broad design space. This course will equip students with a working knowledge of architecture principles for AI hardware, and we will be applying these principles to analyze recent advances in the field and conduct hands-on projects. Cross-list: COMP 537. Recommended Prerequisite(s): COMP 326 - Digital Logic Design, Introductory machine learning and/or artificial intelligence course Mutually Exclusive: Cannot register for COMP 437 if student has credit for COMP 537.