Scientific Computing II

What it is about

You will learn the central steps of the simulation pipeline in Scientific Computing, with a focus on iterative solvers and efficient implementation of numerical methods. This is applied to examples from Computational Fluid Dynamics and to particle-based simulations (e.g., Molecular Dynamics). In the end you will be able to select, analyze and implement suitable iterative methods, as well as design and parallelize efficient algorithms for pairwise interactions.

What you will be able to do

• Familiarity with typical problems in algorithm design and implementation of numerical simulations
• Knowledge of the main classes of iterative solvers for large, sparse linear systems
• Assessment of the scope of application and performance characteristics of these methods
• Application and implementation of iterative methods in code
• Recognition of limitations, sources of error and efficiency problems in solver components
• Understanding of the simulation pipeline for particle-based simulations (modeling to visualization)
• Detailed knowledge for Molecular Dynamics and ability to design appropriate simulation software
• Application of efficient algorithms for the calculation of pairwise forces (e.g., Linked Cell, Verlet, Barnes-Hut, Fast Multipole)

What the module consists of

• VorlesungDelivery of lecture content through presentations and talks
• Übungen / TutorienSolving concrete tasks, partially in teams; discussion of selected examples

Teaching method

• VorlesungenIntroduction to theory and concepts; students should independently deepen literature
• Tutorien / ÜbungsgruppenApplying the methods to concrete tasks and working out solutions, sometimes in teams