SC²S Colloquium - Feb 14, 2018
| Date: | Feb 14, 2019 |
| Room: | 00.08.053 |
| Time: | 15:00 - 16:00 |
Kyle Davis: Simulating heart valves with preCICE
This is a talk invited by Gerasimos Chourdakis. Kyle Davis started working at the IPVS, University of Stuttgart in December 2018, in the group of Prof. Miriam Mehl. Before that, he acquired a M.Eng. in Biomedical Engineering from the University of Stellenbosch, South Africa. He is now one of the preCICE developers.
Kyle is going to be in Garching between February 13-15 for the preCICE coding days and he is available for further discussions.
Performing multiphysics simulations is a very complex challenge, where easy methods of performing these simulations is not readily available. preCICE aims to reduce the complexity of these simulations for real world applications, for the energy sector to biomedical applications. This talk will cover real problems from users worldwide, as well as future plans for preCICE, with a focus on biomedical applications.
Keywords: preCICE, multi-physics, fluid-structure interaction
Language: English Error creating thumbnail: Unable to save thumbnail to destination
Sascha Sauermann: Performance Modelling for Auto-Tuning of Molecular Dynamics Simulations
This is an IDP submission talk advised by Fabio Gratl.
Computing pairwise short-range particle interactions is a time-consuming and unavoidable part of Molecular Dynamics simulations. For the parallelization of the Linked Cell Algorithm there exist different traversal patterns having varying performance depending on the parameters and state of the simulation.
We use Extra-P to create performance models from benchmarks and use them to apply auto-tuning in ls1-mardyn. Thus we predict the optimal traversal given the current state of the simulation and use it until reevaluation.
For single-parameter models the prediction is quite good but the prediction quality declines fast for multi-parameter models. Additionally, the time required for benchmarks also gets infeasible quickly when adding more and more dimensions to the parameter space because the number of samples required grows exponentially.
We thus model the runtime depending only on the density. Auto-tuning with this simple model on each MPI rank separately already delivers close to optimal performance.
Keywords: Molecular Dynamics, auto-tuning, ls1-mardyn
Language: English Error creating thumbnail: Unable to save thumbnail to destination