Teaching

Tutorials and Courses

Apart from regular teaching activities we are known for our node-level performance engineering tutorials and courses. We are regular lecturers for the PRACE course program at the German Gauss Centre for Supercomputing (GCS). The training course “Node-level Performance Engineering” is available at the GCS sites at LRZ in Garching, HLRS in Stuttgart, and as of late at VSC in Vienna (TU Wien). We give tutorials on node-level performance engineering and hybrid programming at top-ranked conferences. Our full-day tutorial “Node-level Performance Engineering” was accepted for the IEEE/ACM Supercomputing conference every year since 2012. Upon request we also offer our tutorial and course program for interested computing centers, research institutions, and industry.

We regularly teach at the following events:

• One week block course “Parallel Programming of High Performance Systems“  organized once a year in cooperation with Leibniz Supercomputing Centre of the Bavarian Academy of Science and Humanities  (LRZ). The course takes place alternating between Erlangen at RRZE and Garching at LRZ.
• PRACE training course “Node-level Performance Engineering” held at HLRS in Stuttgart, LRZ Garching, and the Vienna Scientific Cluster (VSC) at TU Wien.
• PRACE training course “Advanced Topics in High Performance Computing”  held yearly at LRZ in Garching. Together with staff from LRZ and Intel.
• Full-day tutorial “Node-level Performance Engineering”  at  the IEEE/ACM Supercomputing conference.
• Tutorial “Introduction to Hybrid Programming in HPC” held at HLRS Stuttgart, LRZ in Garching, and TU Wien. Together with Rolf Rabenseifner (HLRS),  Claudia Blaas-Schenner, and Irene Reichl (both VSC, TU Wien).

You can find an overview of our tutorials and courses here.

Regular Lectures

We conduct the summer term lecture “Programming Techniques for Supercomputers” (PTfS), which provides an introduction to supercomputing from the core to the cluster level, including introductions to OpenMP, MPI, and basic performance engineering. Every term we invite students to participate to our seminar “Efficient Numerical Simulation on Multi- and Manycore Processors” (MuCoSim), where they can work on challenging parallelization and performance engineering problems. Finally, the winter-term lecture “Elementary Numerical Mathematics” gives a basic introduction to numerical algorithms.

For an overview of lectures in the current and past terms, see our Lectures and Seminars page.

Theses

Motivated by our various research activities, we always have interesting and challenging projects for master’s and bachelor’s theses. Please contact Prof. Gerhard Wellein if you are interested. Note that master or bachelor candidates should at least have successfully completed the “Programming Techniques for Supercomputers” lecture.

A list of past theses can be found on our theses page.