Dynamic Load Balancing with the Parallel Partitioning Tool GridSpiderPar

Dynamic adaptive meshes are often used in high-performance computing. A mesh is locally refined or derefined in spots of interest or where high gradients of an objective function arise. To balance the load of processors, it is required to periodically repartition it. Dynamic load balancing algorithm...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Mathematical models and computer simulations Jg. 14; H. 6; S. 910 - 917
1. Verfasser: Golovchenko, E. N.
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Moscow Pleiades Publishing 01.12.2022
Springer Nature B.V
Schlagworte:
Dynamic loads
Geometric algorithms
Load balancing
Mathematical Modeling and Industrial Mathematics
Mathematics
Mathematics and Statistics
Mesh partitioning
Partitioning
Simulation and Modeling
graph partitioning
mesh partitioning
high-performance computing
ISSN:2070-0482, 2070-0490
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Dynamic adaptive meshes are often used in high-performance computing. A mesh is locally refined or derefined in spots of interest or where high gradients of an objective function arise. To balance the load of processors, it is required to periodically repartition it. Dynamic load balancing algorithms are developed based on the parallel geometric algorithm of mesh partitioning and the parallel incremental algorithm of graph partitioning using the partitioning tool GridSpiderPar. The initial partition of a mesh with a local refinement (6.7 × 10 6 hexahedrons) is compared with the results of repartitioning using the devised algorithms. A comparison of the results shows the advantages of the parallel geometric algorithm on this mesh and the features of using the parallel incremental algorithm.
Bibliographie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:2070-0482
2070-0490
DOI:10.1134/S2070048222060072