Computer architecture is the science and art of selecting and interconnecting hardware components to create a computer that meets functional, performance and cost goals. This course qualitatively and quantitatively examines computer design trade-offs. The topics covered include: advanced processor designs such as superscalar and out-of-order execution, advanced memory systems such as non-blocking caches and multiporting/banking and alternative virtual memory implementations, I/O systems, interconnects, introduction to multiprocessor architectures, performance and power cost metrics, and datacenter architectures.

CSED 311: Compute Architecture (or an equivalent undergraduate course)

(subject to change)
25% Mid-term exam
25% Final exam
40% Project
10% Paper critique

Computer Organization and Design RISC-V Edition, 2nd ed., David Patterson and John Hennessy, 2020

(subject to change)
Week 1: Introduction, backgrounds
Week 2: Memory hierarchy, pipelining
Week 3: Instruction scheduling
Week 4: Tomasulo algorithm, speculative execution
Week 5: Superscalar pipeline, VLIW
Week 6: Multi-threading, vector machine
Week 7: Prefetching, cache coherence
Week 8: Mid-term exam, project progress presentation
Week 9: Virtual memory, memory consistency
Week 10: Synchronization, I/O
Week 11: Data centers, power&energy
Week 12: DRAM, memory controller
Week 13: GPU architecture
Week 14: Domain-specific hardware accelerators
Week 15: Interconnection network
Week 16: Final exam, project final presentation

Lectures will be given in person.