CMSC 22240-1: Computer Architecture for Scientists (and other non-engineers) Winter 2021, TuTh 1120-1240 , Location: TBD
Professor: Prof. Andrew A. Chien
Book: Computer Architecture for Scientists (Cambridge University Press, in print Fall 2021)
The course provides an understanding of the key scientific ideas that underpin the extraordinary capabilities of today’s computers, including speed (gigahertz), illusion of sequential order (relativity), dynamic locality (warping space), parallelism, keeping it cheap - and low-energy (e-field scaling), and of course their ability as universal information processing engines. These scientific principles provide a model for software architects, engineers, data scientists, and other computation users to reason about the performance of their programs on laptops, accelerators (GPUs), servers, and the cloud. Using the scientific principles as underpinnings, we create principled performance models for dynamic locality (caches), scaleup (parallelism), and scaleout (cloud parallelism). This approach arms users of computers with a high-level performance model for computing. Finally, the course provides a longer-term understanding of computer capabilities, performance, and limits to the wealth of computer scientists practicing data science, software development, or machine learning. Students may not receive credit for both CMSC 22200 and 22240. This course is an optional prerequisite for CMSC 23000 (parallel to others)
Students will use a draft version of upcoming book for readings,
discuss lectures. They will learn basics of computer
architecture, perform case studies in performance and
scalability, and also pursue connections of computer
performance scaling principles to other scientific principles.
They will do exercises and labs to exercise the performance models derived
from these scientific principles; learning to reason about
performance and scalability in computer architecture and
systems. And further learn about the implications for
computing systems (software and applications), that enable
them to reason about energy/power and performance in systems
ranging from smartphones to cloud datacenters.
Syllabus |