February 5 – 9, 2024 · Center for Information Services and HPC (ZIH), TU Dresden
Introduction to Computational Fluid Dynamics in High Performance Computing
Trainee
Completed a 30-hour intensive training on CFD within HPC environments, covering numerical methods, turbulence modeling, and MPI parallelization.
I participated in a 30-hour intensive online training course on Computational Fluid Dynamics in High Performance Computing, hosted by the Center for Information Services and HPC (ZIH) at TU Dresden.
The training bridged the gap between fundamental fluid dynamics and advanced computational execution on HPC clusters, featuring hands-on exercises and lectures focused on the following key areas:
Governing Equations & HPC Workflow
- CFD Basics: Introduction to the fundamental equations of fluid dynamics and compressible flows.
- Cluster Operations: Practical workflows for operating on an HPC cluster, including simulation setup and post-processing visualization.
Numerical Methods & Discretization
- Finite Volume Method (FVM): Numerics of Partial Differential Equations, FVM formulation, and exercises resolving Riemann Problems and Flux.
- Finite Element Method (FEM): Fundamentals of FEM applied to incompressible flows.
- High-Order Methods: Implementation of High-Order Finite Volume and Discontinuous Galerkin (DG) Methods, including exercises with High-Order DG in Ateles.
- Lattice Boltzmann Method (LBM): Mesh generation and LBM simulations.
Advanced Physics & Performance
- Turbulence Modeling: Theoretical approaches to turbulence and practical simulations (e.g., subsonic jet).
- Parallelization & Efficiency: Scaling simulations using MPI parallelization, managing Parallel I/O, and conducting rigorous performance assessments.