A new course on superscalar and VLIW computer architectures for real-time image and video computing

A one-quarter electrical and computer engineering course on advanced microprocessors and real-time image/video computing algorithms is described in this paper. It is primarily aimed at graduate students who have some background in computer architecture, signal and image processing, and experience in low-level programming of microprocessors. The course is designed to teach students not only the advanced microprocessors with instruction-level parallelism, e.g., superscalar and very long instruction word (VLIW) computer architectures, but also how to analyze and efficiently map various image and video computing algorithms on these microprocessors. Emphasis is placed on laboratory and individual project work requiring an in-depth study, implementation, testing and verification, demanding intensive participation from both the students and instructor. It is continually updated and improved to reflect new technologies in this dynamically changing area and to incorporate student feedback. This course has been offered for four quarters at the University of Washington