The Department of Mechanical Engineering is pleased to offer a multidisciplinary graduate course.
General Description (3 credits): Applications of high-performance computing systems to science and engineering, programing models for computational science and engineering, high-performance numerical algorithms, programming in FORTRAN 90 and High Performance FORTRAN
Prerequisites: Knowledge of Unix, FORTRAN and previous course on numerical methods or equivalent.
Format: Lectures (based on text, class notes, and recent publications), invited speakers, and world wide web reference materials.
Text: Lloyd D. Fosdick, et. al, "An Introduction to High-Performance Scientific Computing," The MIT Press, 1996.
Motivation: In science and engineering, high performance computing
techniques are used to simulate physical events and to process large
quantities of experimental data. The use of simulation in research is now
established as a method of conducting science in cooperation with theory and
experimentation.
Computational science and engineering have been
continually evolving as disciplines over the past two decades. A growing
number of universities have introduced new courses and degree programs in
these areas.
This class is an advanced multidisciplinary graduate
course that provides a broad overview of the state-of-the-art methods and
problems in computational science and engineering. Our numerical laboratory
for this course will use the newly acquired SGI/Cray Origin 2000 system at
the National Supercomputing Center for Energy and the Environment. The
Origin 2000 system is a scalable parallel high-performance computing
platform based on the SGI/MIPS R10000 RISC processor.
Students will learn and employe software tools
including FORTRAN 90, High-Performance FORTRAN, as well as MPI and PVM
environments. The main motivation for the subject matter is derived from
real applications in science and engineering. We attempt to cover the entire
process of problem solving from the point of mathematical modeling to actual
implementation and analysis of results. The course focuses on numerical
methods for solution of partial differential equations, problem solving
techniques, performance measurement, and optimization for scientific and
engineering problems on high-performance computing systems. The scientific
and engineering application areas will be topical, depending on student
interests and disciplines.
[an error occurred while processing this directive]