Note: The prerequisite of CS 415 / 515 Parallel Programming has been DROPPED. We can give you the needed override for your registration to go through. Contact the undergraduate coordinator uccs at pdx.edu for CS 435, or the instructor for CS 535.
Note: CS 435 is a Programming Intensive course.
The current trend for achieving the highest levels of performance in computing is heterogeneity. Platforms mix general purpose CPUs with special-purpose hardware. This class focuses on heterogeneous approaches that use special-purpose processors and cores to accelerate the execution of a variety of applications. This approach is receiving widespread attention because it can yield a very high rate of operations per watt of power - therefore it is considered a Green approach. Libraries developed for GPUs are in wide use today for machine learning and many scientific applications.
We will start with Why - why multicore, manycore, heterogeneity, GPUs?
Next is What - we will dive into some architectural details - what are the key differences of the accelerated approach? What do they mean for performance? What are the limitations and when should we NOT use the approach?
Much of our time will be spent on How. This class includes hands on programming with NVIDIA Graphics Processing Units (GPUs) and Intel Xeon Phi's. We will learn CUDA C, OpenACC, and programming Intel vector units. Some work will be completed on special-purpose hardware in PPerfLab that will be available to the class.
This course has been redesigned for remote instruction. Zoom will be used for lectures, discussion/problem solving sessions, "lab" sessions, and office hours. I am also exploring some other technologies such as podcasts and will add these if they show promise.
Additional required readings will be from freely available notes, papers and articles.