Numerische Algorithmen im Hochleistungsrechnen

What it is about

You will learn selected numerical algorithms for high-performance computing (HPC), e.g. solvers for partial differential equations, methods of numerical linear algebra, particle-based simulations and spectral methods. You can assess their range of applications, analyze fundamental performance characteristics and evaluate and use the parallelizability and scalability of the algorithms.

What you will be able to do

• Know important algorithms of parallel and high-performance numerical computation
• Assess where the algorithms can be used in typical applications
• Judge fundamental performance characteristics and scalability
• Decide on parallel verifiability of algorithms and analyze data dependencies
• Understand and exploit parallelism (Vectorization, Threading, MPI) in implementation
• Implement and evaluate parallelization strategies for linear systems and PDEs

What the module consists of

• VorlesungConveying concepts, algorithms and performance models
• Tutorium / ÜbungWorking on concrete tasks, collaborative solution development, partially teamwork

Teaching method

• Vorträge / PräsentationenIntroduction and explanation of the content
• Self-study of literatureDeepening and independent engagement with topics
• Übungsaufgaben und Teamarbeit in TutorienPractical application and discussion of exemplary problems