Lecture 3 hours; 3 credits.
Prerequisites: Math 316U; C or FORTRAN.
Laboratory work required.
The motivation for and successes of parallel computing.
A taxonomy of commercially available
parallel computers. Strategies for parallel decompositions. Parallel
performance metrics. Parallel algorithms and their relation to corresponding
serial algorithms. Numerous examples from scientific computing, mainly in
linear algebra and differential equations. Implementations using
public-domain network libraries on workstation clusters and parallel
computers.