Dynamic load balancing for direct-coupled multiphysics simulations. (15th March 2020)
- Record Type:
- Journal Article
- Title:
- Dynamic load balancing for direct-coupled multiphysics simulations. (15th March 2020)
- Main Title:
- Dynamic load balancing for direct-coupled multiphysics simulations
- Authors:
- Niemöller, Ansgar
Schlottke-Lakemper, Michael
Meinke, Matthias
Schröder, Wolfgang - Abstract:
- Highlights: A direct-coupled hybrid multiphysics method is presented for aeroacoustic problems. A dynamic load balancing scheme for coupled multiphysics simulations is presented. The effectiveness to reduce load imbalances for large-scale simulations is shown. The observed performance improvements increase with the degree of parallelism. The approach can balance the workload also in heterogeneous computing environments. Abstract: High parallel efficiency for large-scale coupled multiphysics simulations requires the computational load to be evenly distributed among all compute cores. For complex applications and massively parallel computations, even minor load imbalances can have a severe impact on the overall performance and resource usage. Exemplarily for a volume-coupled multiphysics simulation, a direct-hybrid method is considered, in which a CFD and a CAA simulation are performed concurrently on the same parallel subdomains. For differing load compositions on each subdomain, accurate computational weights for CFD and CAA cells must be known to determine an efficient domain decomposition. Therefore, a dynamic load balancing scheme is presented, which allows to increase the efficiency of complex coupled simulations with non-trivial domain decompositions. A fully-coupled three-dimensional jet simulation with approximately 300 million degrees of freedom demonstrates the effectiveness of the approach to reduce load imbalances. A detailed performance analysis substantiates theHighlights: A direct-coupled hybrid multiphysics method is presented for aeroacoustic problems. A dynamic load balancing scheme for coupled multiphysics simulations is presented. The effectiveness to reduce load imbalances for large-scale simulations is shown. The observed performance improvements increase with the degree of parallelism. The approach can balance the workload also in heterogeneous computing environments. Abstract: High parallel efficiency for large-scale coupled multiphysics simulations requires the computational load to be evenly distributed among all compute cores. For complex applications and massively parallel computations, even minor load imbalances can have a severe impact on the overall performance and resource usage. Exemplarily for a volume-coupled multiphysics simulation, a direct-hybrid method is considered, in which a CFD and a CAA simulation are performed concurrently on the same parallel subdomains. For differing load compositions on each subdomain, accurate computational weights for CFD and CAA cells must be known to determine an efficient domain decomposition. Therefore, a dynamic load balancing scheme is presented, which allows to increase the efficiency of complex coupled simulations with non-trivial domain decompositions. A fully-coupled three-dimensional jet simulation with approximately 300 million degrees of freedom demonstrates the effectiveness of the approach to reduce load imbalances. A detailed performance analysis substantiates the necessity of dynamic load balancing. Furthermore, the results of a strong scaling experiment show the benefit of load balancing to be proportional to the degree of parallelism. In addition, it is shown that the approach allows to attenuate imbalances also for parallel computations on heterogeneous computing hardware. The acoustic field of a chevron nozzle will also be discussed. … (more)
- Is Part Of:
- Computers & fluids. Volume 199(2020)
- Journal:
- Computers & fluids
- Issue:
- Volume 199(2020)
- Issue Display:
- Volume 199, Issue 2020 (2020)
- Year:
- 2020
- Volume:
- 199
- Issue:
- 2020
- Issue Sort Value:
- 2020-0199-2020-0000
- Page Start:
- Page End:
- Publication Date:
- 2020-03-15
- Subjects:
- Dynamic load balancing -- Coupled multiphysics simulation -- Direct-hybrid method -- Parallel efficiency -- Computational aeroacoustics -- Jet noise
Fluid dynamics -- Data processing -- Periodicals
532.050285 - Journal URLs:
- http://www.journals.elsevier.com/computers-and-fluids/ ↗
http://www.elsevier.com/journals ↗ - DOI:
- 10.1016/j.compfluid.2020.104437 ↗
- Languages:
- English
- ISSNs:
- 0045-7930
- Deposit Type:
- Legaldeposit
- View Content:
- Available online (eLD content is only available in our Reading Rooms) ↗
- Physical Locations:
- British Library DSC - 3394.690000
British Library DSC - BLDSS-3PM
British Library HMNTS - ELD Digital store - Ingest File:
- 12755.xml