Parallel Implicit Solution of Diffusion-limited Radiation Transport

William D. Gropp, Dinesh K. Kaushik, David E. Keyes, Barry F. Smith

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

We present simulations of diffusion-limited transport in an initi ally cold medium of two different materials subjected to an impulsive radiative load, using a Newton-Krylov-Schwarz solver. The spatial discretization employs Galerkin finite elements with linear piecewise continuous basis functions over simplices in 2D and 3D. Temporal integration is via a solution-adaptive implicit Euler method. The code shows excellent domain-decomposed scalability on the Teragrid, BlueGene, and System X platforms. Comparing implementations for this application with. opintensive residual evaluation, we observe that an analytical Jacobian gives better performance (in terms of the overall execution time to solution) than a Jacobianfree approach.

Original languageEnglish (US)
Title of host publicationDomain Decomposition Methods in Science and Engineering XVI
EditorsOlof Widlund, David Keyes
Pages579-586
Number of pages8
StatePublished - Dec 1 2007

Publication series

NameLecture Notes in Computational Science and Engineering
Volume55
ISSN (Print)1439-7358

Fingerprint

Euler's method
Piecewise continuous
Galerkin
Execution Time
Basis Functions
Scalability
Continuous Function
Discretization
Radiation
Finite Element
Evaluation
Simulation

ASJC Scopus subject areas

  • Modeling and Simulation
  • Engineering(all)
  • Discrete Mathematics and Combinatorics
  • Control and Optimization
  • Computational Mathematics

Cite this

Gropp, W. D., Kaushik, D. K., Keyes, D. E., & Smith, B. F. (2007). Parallel Implicit Solution of Diffusion-limited Radiation Transport. In O. Widlund, & D. Keyes (Eds.), Domain Decomposition Methods in Science and Engineering XVI (pp. 579-586). (Lecture Notes in Computational Science and Engineering; Vol. 55).

Parallel Implicit Solution of Diffusion-limited Radiation Transport. / Gropp, William D.; Kaushik, Dinesh K.; Keyes, David E.; Smith, Barry F.

Domain Decomposition Methods in Science and Engineering XVI. ed. / Olof Widlund; David Keyes. 2007. p. 579-586 (Lecture Notes in Computational Science and Engineering; Vol. 55).

Research output: Chapter in Book/Report/Conference proceedingChapter

Gropp, WD, Kaushik, DK, Keyes, DE & Smith, BF 2007, Parallel Implicit Solution of Diffusion-limited Radiation Transport. in O Widlund & D Keyes (eds), Domain Decomposition Methods in Science and Engineering XVI. Lecture Notes in Computational Science and Engineering, vol. 55, pp. 579-586.
Gropp WD, Kaushik DK, Keyes DE, Smith BF. Parallel Implicit Solution of Diffusion-limited Radiation Transport. In Widlund O, Keyes D, editors, Domain Decomposition Methods in Science and Engineering XVI. 2007. p. 579-586. (Lecture Notes in Computational Science and Engineering).
Gropp, William D. ; Kaushik, Dinesh K. ; Keyes, David E. ; Smith, Barry F. / Parallel Implicit Solution of Diffusion-limited Radiation Transport. Domain Decomposition Methods in Science and Engineering XVI. editor / Olof Widlund ; David Keyes. 2007. pp. 579-586 (Lecture Notes in Computational Science and Engineering).
@inbook{a8fb332e9c3e427ba3f331173df385a5,
title = "Parallel Implicit Solution of Diffusion-limited Radiation Transport",
abstract = "We present simulations of diffusion-limited transport in an initi ally cold medium of two different materials subjected to an impulsive radiative load, using a Newton-Krylov-Schwarz solver. The spatial discretization employs Galerkin finite elements with linear piecewise continuous basis functions over simplices in 2D and 3D. Temporal integration is via a solution-adaptive implicit Euler method. The code shows excellent domain-decomposed scalability on the Teragrid, BlueGene, and System X platforms. Comparing implementations for this application with. opintensive residual evaluation, we observe that an analytical Jacobian gives better performance (in terms of the overall execution time to solution) than a Jacobianfree approach.",
author = "Gropp, {William D.} and Kaushik, {Dinesh K.} and Keyes, {David E.} and Smith, {Barry F.}",
year = "2007",
month = "12",
day = "1",
language = "English (US)",
isbn = "9783540344681",
series = "Lecture Notes in Computational Science and Engineering",
pages = "579--586",
editor = "Olof Widlund and David Keyes",
booktitle = "Domain Decomposition Methods in Science and Engineering XVI",

}

TY - CHAP

T1 - Parallel Implicit Solution of Diffusion-limited Radiation Transport

AU - Gropp, William D.

AU - Kaushik, Dinesh K.

AU - Keyes, David E.

AU - Smith, Barry F.

PY - 2007/12/1

Y1 - 2007/12/1

N2 - We present simulations of diffusion-limited transport in an initi ally cold medium of two different materials subjected to an impulsive radiative load, using a Newton-Krylov-Schwarz solver. The spatial discretization employs Galerkin finite elements with linear piecewise continuous basis functions over simplices in 2D and 3D. Temporal integration is via a solution-adaptive implicit Euler method. The code shows excellent domain-decomposed scalability on the Teragrid, BlueGene, and System X platforms. Comparing implementations for this application with. opintensive residual evaluation, we observe that an analytical Jacobian gives better performance (in terms of the overall execution time to solution) than a Jacobianfree approach.

AB - We present simulations of diffusion-limited transport in an initi ally cold medium of two different materials subjected to an impulsive radiative load, using a Newton-Krylov-Schwarz solver. The spatial discretization employs Galerkin finite elements with linear piecewise continuous basis functions over simplices in 2D and 3D. Temporal integration is via a solution-adaptive implicit Euler method. The code shows excellent domain-decomposed scalability on the Teragrid, BlueGene, and System X platforms. Comparing implementations for this application with. opintensive residual evaluation, we observe that an analytical Jacobian gives better performance (in terms of the overall execution time to solution) than a Jacobianfree approach.

UR - http://www.scopus.com/inward/record.url?scp=84880344334&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84880344334&partnerID=8YFLogxK

M3 - Chapter

AN - SCOPUS:84880344334

SN - 9783540344681

T3 - Lecture Notes in Computational Science and Engineering

SP - 579

EP - 586

BT - Domain Decomposition Methods in Science and Engineering XVI

A2 - Widlund, Olof

A2 - Keyes, David

ER -