PACS Course Structured By Common Mathematical Problems & Topics

The Partners for Advanced Computational Services (PACS) Alliance Training group announces the release of their Parallel Numerical Libraries asynchronous, WebCT course. “This tutorial is unique in a number of ways. Rather than present a list of available libraries, we have structured the course by common mathematical problems and topics,” said Dr. David Ennis, PACS Course Development Leader and Ohio Supercomputer Center (OSC) Systems Developer. “This structure makes it easier for someone with a mathematical problem to find the appropriate library,” said Ennis. “We cover all of the popular libraries as well as the topic-specific ones. For each chapter, we have provided tools to enhance the learning such as problem assignments and solutions, self-test, and discussion space for questions and dialogue.” The on-line tutorial, accessible from includes the following topics: • Parallel Work Distribution • Vector & Matrix Operations • Simultaneous Sets of Linear Equations • Statistical Fitting Operations • Sparse Matrices • Fourier Transforms • Eigenvalue / Eigenvector Problems • Random Number Generation • Numerical Integration • Partial Differential Equations (Appendix A) “As with our previous asynchronous courses, we are proud of the quality of the work that went into this course,” said Leslie Southern, PACS Training Lead and OSC HPC Director. “Based upon the feedback we have gotten so far, we expect a large number of people to take this course and share it with their colleagues. Our first asynchronous course, Introduction to MPI, has had over 8600 log-ins and we hope that this course reaches as many or more.” The Parallel Numerical Libraries on-line tutorial was developed by PACS Alliance partners National Center for Supercomputer Applications (NCSA), Boston University, University of Kentucky, University of New Mexico, and OSC.