Course purpose
The archived NGSSC overview describes an advanced course in computational science and engineering. It developed practical experience with parallel algorithms and scientific software on high-performance computing systems. Students were expected to apply the techniques to an application connected with their graduate research.
Learning objectives
- Partition computational problems and data to use memory efficiently and reduce communication within and between compute nodes.
- Assess the cost and scalability of parallel algorithms and their implementations.
- Work with parallel libraries for dense linear algebra and software environments used in scientific computing.
- Evaluate high-performance computing methods through a scientific application.
Teaching and course structure
The overview names HPC2N in Umeå and UPPMAX in Uppsala as the collaborating centres, with Mikael Rännar and Maya Neytcheva as the main teachers. Its timetable places preparation in week 47, Umeå lectures in week 48, homework in week 49, Uppsala lectures in week 50 and project work in week 51.
This is a historical course overview. The archived page points readers to the 2012 course; it does not establish the dates or teaching staff of other course editions.