Introduction to Parallel Computing

ENG EK420 / CAS CS420 - Fall Semester

M. Crovella, R. Giles, S. Homer

Parallel and distributed computing systems have emerged as the most powerful computers available, taking over the field of supercomputing. The wide availability of parallel supercomputers is now spurring important scientific progress, as scientists and engineers use parallel computers to explore natural and manmade systems in increasing detail. This course will provide students with an introduction to this field of study and with hands-on experience using state of the art parallel computing technology. The course will introduce fundamental methods for scientific computing in the context of massively parallel computation. Discussions will be organized around important algorithmic concepts and specific applications chosen to illustrate the methods. Various parallel computational models will be evaluated within the framework of specific algorithms. Students will observe, modify and design programs that run on parallel architectures such as the University's SGI Origin2000 and SGI POWER CHALLENGEarray as well as networks of distributed workstations. They will develop competence in both data parallel and message-passing style programming. In addition they will make use of modern computational tools for scientific visualization, interactive debugging, and performance analysis.