MHPC Master Thesis

Parallel SUNDIALS Wrappers for Deal.II

SUNDIALS is an Open Source library that allows one to solve ODE and DAE in a very efficient way. We have implemented some time ago a SUNDIALS Wrapper for deal.II, and we would like to make this wrapper works also in parallel, exploiting either PETSc or Trilinos backends for deal.II Linear Algebra.

Exploiting blender capabilities for scientific computing applications

Blender is an opensource software ( to render and model three-dimensional worlds. It contains a fully programmable python interface, and it implements several algorithms to reconstruct the three-dimensional motion (tracking) of high contrast points (markers) in movie frames. This capability is very useful in several scientific applications, and will be explored in detail for locomotion problems, as well as sports related motion capturing.

Introducing "programming best practices" into existing scientific HPC programs

Existing scientific programs are often poorly and/or badly documented, and very rarely adhere to best practices in scientific computing, like unit testing, agile programming, and so on. In this MHPC master thesis, the student will be presented with an example of such an HPC program which is currently being developed, and will cooperate with the authors of the program to transform it into a fully tested, documented and sustainable HPC application which follows programming best practices.

Subscribe to RSS - MHPC Master Thesis