Rettinger C (2018)
Publication Type: Conference contribution, Abstract of a poster
Publication year: 2018
Particulate flow simulations of geometrically fully resolved particles enable accurate predictions from first principles but their high computational cost usually limits the admissible system sizes significantly.
Here, we present two distinct approaches to reduce these costs: adaptive mesh refinement (AMR) and dynamic load balancing.
In AMR, the computational grid features a fine resolution only in certain, temporally and spatially changing regions of interest, e.g. around the particles, whereas it is coarsened elsewhere.
Load balancing, on the other hand, aims to distribute the computational load evenly among the available resources to improve their utilization which effectively reduces the time to solution.
APA:
Rettinger, C. (2018). Adaptive Mesh Refinement and Load Balancing Techniques for Particulate Flow Simulations. Poster presentation at CoSaS - International Symposium on Computational Science at Scale, Erlangen, DE.
MLA:
Rettinger, Christoph. "Adaptive Mesh Refinement and Load Balancing Techniques for Particulate Flow Simulations." Presented at CoSaS - International Symposium on Computational Science at Scale, Erlangen 2018.
BibTeX: Download