Wintersemester 2019/20

Scientific computing

Vorlesung

 
Dozent

Prof. Dr. Dominik Göddeke

Zeit und Ort

Vorlesung
Mo.  09:45 - 11:15 PWR 57 - 7.342
Do. 09:45 - 11:15 PWR 57 - 7.122
Erste Vorlesung: Montag, den 14.10.2019

Übung
Do. 11:30 - 13:00 PWR 57 - 7.122

 
Inhalt

We discuss modern topics in the numerical approximation of solutions of partial differential equations. This includes in particular aspects of parallelising discretisation and solution schemes: On the one hand, phones are parallel computers nowadays, but on the other hand, "typical" textbooks and lectures are still sequential and thus, do not match the modern hardware. We will see that exciting new theories are necessary for parallelisation, and that many schemes need to be redesigned from scratch.

List of topics:

  • Abstract Krylov theory and solvers for nonsymmetric problems
  • Crash course OpenMP programming (if requested by participants)
  • Parallel preconditioning
  • Domain decomposition methods: abstract Schwarz theory, hierarchical Schwarz- and substructuring methods
  • Depending on the audience: fault tolerant algorithms, asynchronous algorithms, parallel-in-time schemes
Vorkenntnisse

Some knowledge of Numerics for PDEs, e.g. PDEMAS (Bachelor) or Einführung in die Numerik partieller Differentialgleichungen (Master)

Literatur

will be announced in the lecture

Curricula

M.Sc. Mathematik, M.Sc. SimTech

Leistungspunkte

9 LP

Prüfung

oral exam

Dominik Göddeke
Prof. Dr. rer. nat.

Dominik Göddeke

Institutsleiter und Lehrstuhlinhaber

Zum Seitenanfang