Parallel Programming

What it is about

You learn concepts and techniques for parallel programming for distributed and shared memory architectures. The focus is on MPI for distributed systems and OpenMP for shared memory; in addition, dependency analysis, program transformations and newer programming models (e.g. PGAS, CUDA, OpenCL, OpenACC) are covered. In the end you will be able to create programs with MPI and OpenMP, assess their performance and analyze and optimize parallelized applications.

What you will be able to do

• Write MPI programs
• Write OpenMP programs
• Assess performance of different parallelization strategies
• Apply data dependency analysis
• Perform program transformations
• Analyze and tune parallel applications

What the module consists of

• VorlesungIntroduction to models, architectures and concepts of parallel programming
• Übung / zentrale ÜbungssitzungPresentation and discussion of tasks; students solve and submit assignments
• Studentische Kurzpräsentationen (freiwillig)Demonstration of techniques in application areas

Teaching method

• VorlesungConveying the various parallel models and techniques
• Freiwillige KurzpräsentationenPractical examples and deepening by students
• Zentrale Übung mit Aufgaben und EinreichungApplying concepts to larger example programs and checking correctness