Cross-listed with ENEE759c. Credit only granted for CMSC818J or ENEE759C.

Must be in the Graduate Program in Computer Science. All other graduate students must request permission.

Explores the design and evaluation of domain-specific architectures (DSAs), or hardware accelerators, for applications such as deep learning, computer vision, scientific computing, and databases. Topics include parallelism, memory-access optimization, data specialization, systolic arrays, near-data processing, hardware-software co-design, and trade-offs among performance, power, area, and generality. The course emphasizes recent research through paper readings and discussions, programming assignments involving DSA implementations or studies, and a research-oriented final project.