Advanced parallelization strategies using hybrid MPI-CUDA DSMC method for modeling flow through porous media

Revathi Jambunathan, Deborah A. Levin

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

The gridfree octree DSMC approach is parallelized with GPGPUs using CUDA. A linear space filling Morton Z-curve is employed to represent the three dimensional octree structure in a linear array and the advantages of this implementation is discussed. The methodology for a multi-GPU, hybrid MPI-CUDA implementation, involving load balanced domain decomposition and inter-GPU communications is presented. External flow at 1 atm over a fractal-like spherical aggregate is modeled and strong scaling studies for the test case is analysed. The analysis showed that the CHAOS DSMC solver, with 8 GPUs was 92% efficient, while the 16 GPU case was not partitioned equally between all the 16 GPUs. Argon gas flow through a fibrous microstruture modeled with 1.5 million triangular panels was also modeled and showed good scaling. Material permeability of the microstructure was computed to be equal to 253 ×10-12m2.

Original languageEnglish (US)
Title of host publication46th AIAA Thermophysics Conference
PublisherAmerican Institute of Aeronautics and Astronautics Inc, AIAA
ISBN (Print)9781624104350
StatePublished - Jan 1 2016
Event46th AIAA Thermophysics Conference, 2016 - Washington, United States
Duration: Jun 13 2016Jun 17 2016

Publication series

Name46th AIAA Thermophysics Conference

Other

Other46th AIAA Thermophysics Conference, 2016
CountryUnited States
CityWashington
Period6/13/166/17/16

Fingerprint

Porous materials
Fractals
Flow of gases
Argon
Graphics processing unit
Decomposition
Microstructure
Communication

ASJC Scopus subject areas

  • Aerospace Engineering
  • Mechanical Engineering

Cite this

Jambunathan, R., & Levin, D. A. (2016). Advanced parallelization strategies using hybrid MPI-CUDA DSMC method for modeling flow through porous media. In 46th AIAA Thermophysics Conference (46th AIAA Thermophysics Conference). American Institute of Aeronautics and Astronautics Inc, AIAA.

Advanced parallelization strategies using hybrid MPI-CUDA DSMC method for modeling flow through porous media. / Jambunathan, Revathi; Levin, Deborah A.

46th AIAA Thermophysics Conference. American Institute of Aeronautics and Astronautics Inc, AIAA, 2016. (46th AIAA Thermophysics Conference).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Jambunathan, R & Levin, DA 2016, Advanced parallelization strategies using hybrid MPI-CUDA DSMC method for modeling flow through porous media. in 46th AIAA Thermophysics Conference. 46th AIAA Thermophysics Conference, American Institute of Aeronautics and Astronautics Inc, AIAA, 46th AIAA Thermophysics Conference, 2016, Washington, United States, 6/13/16.
Jambunathan R, Levin DA. Advanced parallelization strategies using hybrid MPI-CUDA DSMC method for modeling flow through porous media. In 46th AIAA Thermophysics Conference. American Institute of Aeronautics and Astronautics Inc, AIAA. 2016. (46th AIAA Thermophysics Conference).
Jambunathan, Revathi ; Levin, Deborah A. / Advanced parallelization strategies using hybrid MPI-CUDA DSMC method for modeling flow through porous media. 46th AIAA Thermophysics Conference. American Institute of Aeronautics and Astronautics Inc, AIAA, 2016. (46th AIAA Thermophysics Conference).
@inproceedings{e230c888026b458188afcd5a1837b896,
title = "Advanced parallelization strategies using hybrid MPI-CUDA DSMC method for modeling flow through porous media",
abstract = "The gridfree octree DSMC approach is parallelized with GPGPUs using CUDA. A linear space filling Morton Z-curve is employed to represent the three dimensional octree structure in a linear array and the advantages of this implementation is discussed. The methodology for a multi-GPU, hybrid MPI-CUDA implementation, involving load balanced domain decomposition and inter-GPU communications is presented. External flow at 1 atm over a fractal-like spherical aggregate is modeled and strong scaling studies for the test case is analysed. The analysis showed that the CHAOS DSMC solver, with 8 GPUs was 92{\%} efficient, while the 16 GPU case was not partitioned equally between all the 16 GPUs. Argon gas flow through a fibrous microstruture modeled with 1.5 million triangular panels was also modeled and showed good scaling. Material permeability of the microstructure was computed to be equal to 253 ×10-12m2.",
author = "Revathi Jambunathan and Levin, {Deborah A.}",
year = "2016",
month = "1",
day = "1",
language = "English (US)",
isbn = "9781624104350",
series = "46th AIAA Thermophysics Conference",
publisher = "American Institute of Aeronautics and Astronautics Inc, AIAA",
booktitle = "46th AIAA Thermophysics Conference",

}

TY - GEN

T1 - Advanced parallelization strategies using hybrid MPI-CUDA DSMC method for modeling flow through porous media

AU - Jambunathan, Revathi

AU - Levin, Deborah A.

PY - 2016/1/1

Y1 - 2016/1/1

N2 - The gridfree octree DSMC approach is parallelized with GPGPUs using CUDA. A linear space filling Morton Z-curve is employed to represent the three dimensional octree structure in a linear array and the advantages of this implementation is discussed. The methodology for a multi-GPU, hybrid MPI-CUDA implementation, involving load balanced domain decomposition and inter-GPU communications is presented. External flow at 1 atm over a fractal-like spherical aggregate is modeled and strong scaling studies for the test case is analysed. The analysis showed that the CHAOS DSMC solver, with 8 GPUs was 92% efficient, while the 16 GPU case was not partitioned equally between all the 16 GPUs. Argon gas flow through a fibrous microstruture modeled with 1.5 million triangular panels was also modeled and showed good scaling. Material permeability of the microstructure was computed to be equal to 253 ×10-12m2.

AB - The gridfree octree DSMC approach is parallelized with GPGPUs using CUDA. A linear space filling Morton Z-curve is employed to represent the three dimensional octree structure in a linear array and the advantages of this implementation is discussed. The methodology for a multi-GPU, hybrid MPI-CUDA implementation, involving load balanced domain decomposition and inter-GPU communications is presented. External flow at 1 atm over a fractal-like spherical aggregate is modeled and strong scaling studies for the test case is analysed. The analysis showed that the CHAOS DSMC solver, with 8 GPUs was 92% efficient, while the 16 GPU case was not partitioned equally between all the 16 GPUs. Argon gas flow through a fibrous microstruture modeled with 1.5 million triangular panels was also modeled and showed good scaling. Material permeability of the microstructure was computed to be equal to 253 ×10-12m2.

UR - http://www.scopus.com/inward/record.url?scp=84980351854&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84980351854&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84980351854

SN - 9781624104350

T3 - 46th AIAA Thermophysics Conference

BT - 46th AIAA Thermophysics Conference

PB - American Institute of Aeronautics and Astronautics Inc, AIAA

ER -