Department of Information Technology

Parallel Algorithms for Scientific Computing

The course focuses on parallel algorithms for scientific computing problems (and not on parallel programming). We will cover parallel algorithms for a range of problem classes in Scientific Computing, from linear algebra (with a focus on sparse problems related to solving PDE), iterative solvers, multigrid, and particle interaction problems (Barnes-Hut, multipole etc).

Passing the course nominally gives 5 hp credits, but by completing an optional project it can also give 7,5 hp credits. The examination will consist of a written exam and written reports on the two assignments. The examination for the optional project consists of a written report.

Teachers: Sverker Holmgren, Maya Neytcheva, and Dimitar Lukarski


Basic sparse linear algebra

  • Sparse formats and representation (DL). April 11, 13-15, room 2145. slides
  • Sparse matrix-vector multiplication (SH). May 8, 13-15, room 2344. slides

Sparse linear systems

Parallel preconditioners - convergence acceleration

Methods for particle-like problems (pairwise interactions)

Assignment 1: a written report should be delivered to Maya Neytcheva not later than by June 12, 2013.

Assignment 2: Article review

Updated  2015-10-01 23:58:35 by Kurt Otto.