Distributed-memory parallelization of radiative transfer calculation in hypersonic flow

S. Matsuyama; N. Ohnishi; A. Sasoh; Keisuke Sawada

doi:10.1016/B978-044450680-1/50062-0

Distributed-memory parallelization of radiative transfer calculation in hypersonic flow

S. Matsuyama, N. Ohnishi, A. Sasoh, Keisuke Sawada

ENG - Aerospace Engineering

Research output: Chapter in Book/Report/Conference proceeding › Chapter

Abstract

This paper presents a new parallel strategy for radiative transfer calculation coupled with hypersonic flow computations. The cell-interface division parallel strategy is developed based on the distributed-memory parallelization, and applied to the calculation of threedimensional radiative transfer in axisymetric hypersonic flowfieM around the forebody of Fire II vehicle. The developed parallel code realizes fair scalability up to 128 processors. The computational speed achieved by the present parallel strategy using 128 processors of SGI ORIGIN 2000 is approximately 20 Gflops that is 115 times faster than that of a single processor. It is shown that cell-interface division strategy is suitable for large-scale radiation coupling calculations.

Original language	English
Title of host publication	Parallel Computational Fluid Dynamics 2002
Subtitle of host publication	New Frontiers and Multi-Disciplinary Applications
Publisher	Elsevier Inc.
Pages	491-498
Number of pages	8
ISBN (Electronic)	9780080538426
ISBN (Print)	9780444506801
DOIs	https://doi.org/10.1016/B978-044450680-1/50062-0
Publication status	Published - 2003 Apr 25

Keywords

Parallel Algorithms
Supercomputer Applications

ASJC Scopus subject areas

Mathematics(all)

Access to Document

10.1016/B978-044450680-1/50062-0

Cite this

@inbook{1bcec3c7385a49f4865c5caf1b76e8d5,

title = "Distributed-memory parallelization of radiative transfer calculation in hypersonic flow",

abstract = "This paper presents a new parallel strategy for radiative transfer calculation coupled with hypersonic flow computations. The cell-interface division parallel strategy is developed based on the distributed-memory parallelization, and applied to the calculation of threedimensional radiative transfer in axisymetric hypersonic flowfieM around the forebody of Fire II vehicle. The developed parallel code realizes fair scalability up to 128 processors. The computational speed achieved by the present parallel strategy using 128 processors of SGI ORIGIN 2000 is approximately 20 Gflops that is 115 times faster than that of a single processor. It is shown that cell-interface division strategy is suitable for large-scale radiation coupling calculations.",

keywords = "Parallel Algorithms, Supercomputer Applications",

author = "S. Matsuyama and N. Ohnishi and A. Sasoh and Keisuke Sawada",

year = "2003",

month = apr,

day = "25",

doi = "10.1016/B978-044450680-1/50062-0",

language = "English",

isbn = "9780444506801",

pages = "491--498",

booktitle = "Parallel Computational Fluid Dynamics 2002",

publisher = "Elsevier Inc.",

}

TY - CHAP

T1 - Distributed-memory parallelization of radiative transfer calculation in hypersonic flow

AU - Matsuyama, S.

AU - Ohnishi, N.

AU - Sasoh, A.

AU - Sawada, Keisuke

PY - 2003/4/25

Y1 - 2003/4/25

N2 - This paper presents a new parallel strategy for radiative transfer calculation coupled with hypersonic flow computations. The cell-interface division parallel strategy is developed based on the distributed-memory parallelization, and applied to the calculation of threedimensional radiative transfer in axisymetric hypersonic flowfieM around the forebody of Fire II vehicle. The developed parallel code realizes fair scalability up to 128 processors. The computational speed achieved by the present parallel strategy using 128 processors of SGI ORIGIN 2000 is approximately 20 Gflops that is 115 times faster than that of a single processor. It is shown that cell-interface division strategy is suitable for large-scale radiation coupling calculations.

AB - This paper presents a new parallel strategy for radiative transfer calculation coupled with hypersonic flow computations. The cell-interface division parallel strategy is developed based on the distributed-memory parallelization, and applied to the calculation of threedimensional radiative transfer in axisymetric hypersonic flowfieM around the forebody of Fire II vehicle. The developed parallel code realizes fair scalability up to 128 processors. The computational speed achieved by the present parallel strategy using 128 processors of SGI ORIGIN 2000 is approximately 20 Gflops that is 115 times faster than that of a single processor. It is shown that cell-interface division strategy is suitable for large-scale radiation coupling calculations.

KW - Parallel Algorithms

KW - Supercomputer Applications

UR - http://www.scopus.com/inward/record.url?scp=84942235825&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84942235825&partnerID=8YFLogxK

U2 - 10.1016/B978-044450680-1/50062-0

DO - 10.1016/B978-044450680-1/50062-0

M3 - Chapter

AN - SCOPUS:84942235825

SN - 9780444506801

SP - 491

EP - 498

BT - Parallel Computational Fluid Dynamics 2002

PB - Elsevier Inc.

ER -

Distributed-memory parallelization of radiative transfer calculation in hypersonic flow

Abstract

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this