Module MAP55640: HPC SEMINARS

Credit weighting (ECTS)

5 credits

Semester/term taught

Second Semester

Module Co-ordinator

Mike Peardon

Intended Learning Outcomes

Students who successfully complete the course should be able to:

Describe the difficulties often encountered in solving large-scale numerical problems on multiple compute cores.
Write software to solve sparse linear systems that will use the MPI library to exploit multiple compute cores.
Describe how efficient N-body simulations can be implemented on a parallel computer.
Describe the fast fourier transform algorithm and its efficient implementation.
Explain the multigrid method as applied to the sparse linear systems often encountered when solving partial differential equations numerically.

Structure and Content

This module consists of independent student research, presentations and written reports on relevant topics in High-Performance Computing. Each student will be asked to carry out an independent literature survey on a topic relevant to high-performance computing and then give a 20- minute presentation to the class summarising their findings. At the end of the semester, a report on the topic will be prepared and submitted. Appropriate research topics will be suggested at the start of each semester. Other topics are also permitted provided they are discussed with and approved by the module co-ordinators prior to the presentation.

Pre-requisite course

MA5633 Numerical Analysis

Assessment Detail

40% continuously assessed programming assignments. 60% two-hour written examinations (April/May)