Scalable fast multipole accelerated vortex methods

Handle URI:
http://hdl.handle.net/10754/575821
Title:
Scalable fast multipole accelerated vortex methods
Authors:
Hu, Qi; Gumerov, Nail A.; Yokota, Rio ( 0000-0001-7573-7873 ) ; Barba, Lorena A.; Duraiswami, Ramani
Abstract:
The fast multipole method (FMM) is often used to accelerate the calculation of particle interactions in particle-based methods to simulate incompressible flows. To evaluate the most time-consuming kernels - the Biot-Savart equation and stretching term of the vorticity equation, we mathematically reformulated it so that only two Laplace scalar potentials are used instead of six. This automatically ensuring divergence-free far-field computation. Based on this formulation, we developed a new FMM-based vortex method on heterogeneous architectures, which distributed the work between multicore CPUs and GPUs to best utilize the hardware resources and achieve excellent scalability. The algorithm uses new data structures which can dynamically manage inter-node communication and load balance efficiently, with only a small parallel construction overhead. This algorithm can scale to large-sized clusters showing both strong and weak scalability. Careful error and timing trade-off analysis are also performed for the cutoff functions induced by the vortex particle method. Our implementation can perform one time step of the velocity+stretching calculation for one billion particles on 32 nodes in 55.9 seconds, which yields 49.12 Tflop/s.
KAUST Department:
Extreme Computing Research Center
Publisher:
Institute of Electrical and Electronics Engineers (IEEE)
Journal:
2014 IEEE International Parallel & Distributed Processing Symposium Workshops
Conference/Event name:
28th IEEE International Parallel and Distributed Processing Symposium Workshops, IPDPSW 2014
Issue Date:
May-2014
DOI:
10.1109/IPDPSW.2014.110
Type:
Conference Paper
ISSN:
15302075
ISBN:
9780769552088
Appears in Collections:
Conference Papers; Extreme Computing Research Center; Extreme Computing Research Center

Full metadata record

DC FieldValue Language
dc.contributor.authorHu, Qien
dc.contributor.authorGumerov, Nail A.en
dc.contributor.authorYokota, Rioen
dc.contributor.authorBarba, Lorena A.en
dc.contributor.authorDuraiswami, Ramanien
dc.date.accessioned2015-08-24T09:27:05Zen
dc.date.available2015-08-24T09:27:05Zen
dc.date.issued2014-05en
dc.identifier.isbn9780769552088en
dc.identifier.issn15302075en
dc.identifier.doi10.1109/IPDPSW.2014.110en
dc.identifier.urihttp://hdl.handle.net/10754/575821en
dc.description.abstractThe fast multipole method (FMM) is often used to accelerate the calculation of particle interactions in particle-based methods to simulate incompressible flows. To evaluate the most time-consuming kernels - the Biot-Savart equation and stretching term of the vorticity equation, we mathematically reformulated it so that only two Laplace scalar potentials are used instead of six. This automatically ensuring divergence-free far-field computation. Based on this formulation, we developed a new FMM-based vortex method on heterogeneous architectures, which distributed the work between multicore CPUs and GPUs to best utilize the hardware resources and achieve excellent scalability. The algorithm uses new data structures which can dynamically manage inter-node communication and load balance efficiently, with only a small parallel construction overhead. This algorithm can scale to large-sized clusters showing both strong and weak scalability. Careful error and timing trade-off analysis are also performed for the cutoff functions induced by the vortex particle method. Our implementation can perform one time step of the velocity+stretching calculation for one billion particles on 32 nodes in 55.9 seconds, which yields 49.12 Tflop/s.en
dc.publisherInstitute of Electrical and Electronics Engineers (IEEE)en
dc.subjectFMMen
dc.subjectheterogeneous algorithmen
dc.subjectGPGPUen
dc.subjectvortex methodsen
dc.titleScalable fast multipole accelerated vortex methodsen
dc.typeConference Paperen
dc.contributor.departmentExtreme Computing Research Centeren
dc.identifier.journal2014 IEEE International Parallel & Distributed Processing Symposium Workshopsen
dc.conference.date19 May 2014 through 23 May 2014en
dc.conference.name28th IEEE International Parallel and Distributed Processing Symposium Workshops, IPDPSW 2014en
dc.contributor.institutionUniversity of Maryland, Institute for Advanced Computer Studies, United Statesen
dc.contributor.institutionDepartment of Computer Science, University of MarylandCollege Park, United Statesen
dc.contributor.institutionFantalgo LLCElkridge, MD, United Statesen
dc.contributor.institutionMechanical and Aerospace Engineering, George Washington University, United Statesen
kaust.authorYokota, Rioen
All Items in KAUST are protected by copyright, with all rights reserved, unless otherwise indicated.