Massively parallel, three-dimensional transport solutions for the k-eigenvalue problem

Gregory G. Davidson; Thomas M. Evans; Joshua J. Jarrell; Steven P. Hamilton; Tara M. Pandya; Rachel N. Slaybaugh

doi:10.13182/NSE12-101

Massively parallel, three-dimensional transport solutions for the k-eigenvalue problem

Gregory G. Davidson, Thomas M. Evans, Joshua J. Jarrell, Steven P. Hamilton, Tara M. Pandya, Rachel N. Slaybaugh

Radiation Transport and HPC Methods

Research output: Contribution to journal › Article › peer-review

25 Scopus citations

Abstract

We have implemented a new multilevel parallel decomposition in the Denovo discrete ordinates radiation transport code. In concert with Krylov subspace iterative solvers, the multilevel decomposition allows concurrency over energy in addition to space-angle, enabling scalability beyond the limits imposed by the traditional Koch-Baker-Alcouffe (KBA) space-angle partitioning. Furthermore, a new Arnoldi-based k-eigenvalue solver has been implemented. The added phase-space concurrency combined with the high-performance Krylov and Arnoldi solvers has enabled weak scaling to 0(10⁵) cores on the Titan XK7 supercomputer. The multilevel decomposition provides a mechanism for scaling to exascale computing and beyond.

Original language	English
Pages (from-to)	111-125
Number of pages	15
Journal	Nuclear Science and Engineering
Volume	177
Issue number	2
DOIs	https://doi.org/10.13182/NSE12-101
State	Published - Jun 2014

Access to Document

10.13182/NSE12-101

Cite this

@article{1394cac72d684840994b67b5dae52e96,

title = "Massively parallel, three-dimensional transport solutions for the k-eigenvalue problem",

abstract = "We have implemented a new multilevel parallel decomposition in the Denovo discrete ordinates radiation transport code. In concert with Krylov subspace iterative solvers, the multilevel decomposition allows concurrency over energy in addition to space-angle, enabling scalability beyond the limits imposed by the traditional Koch-Baker-Alcouffe (KBA) space-angle partitioning. Furthermore, a new Arnoldi-based k-eigenvalue solver has been implemented. The added phase-space concurrency combined with the high-performance Krylov and Arnoldi solvers has enabled weak scaling to 0(105) cores on the Titan XK7 supercomputer. The multilevel decomposition provides a mechanism for scaling to exascale computing and beyond.",

author = "Davidson, \{Gregory G.\} and Evans, \{Thomas M.\} and Jarrell, \{Joshua J.\} and Hamilton, \{Steven P.\} and Pandya, \{Tara M.\} and Slaybaugh, \{Rachel N.\}",

year = "2014",

month = jun,

doi = "10.13182/NSE12-101",

language = "English",

volume = "177",

pages = "111--125",

journal = "Nuclear Science and Engineering",

issn = "0029-5639",

publisher = "Taylor and Francis Group",

number = "2",

}

TY - JOUR

T1 - Massively parallel, three-dimensional transport solutions for the k-eigenvalue problem

AU - Davidson, Gregory G.

AU - Evans, Thomas M.

AU - Jarrell, Joshua J.

AU - Hamilton, Steven P.

AU - Pandya, Tara M.

AU - Slaybaugh, Rachel N.

PY - 2014/6

Y1 - 2014/6

N2 - We have implemented a new multilevel parallel decomposition in the Denovo discrete ordinates radiation transport code. In concert with Krylov subspace iterative solvers, the multilevel decomposition allows concurrency over energy in addition to space-angle, enabling scalability beyond the limits imposed by the traditional Koch-Baker-Alcouffe (KBA) space-angle partitioning. Furthermore, a new Arnoldi-based k-eigenvalue solver has been implemented. The added phase-space concurrency combined with the high-performance Krylov and Arnoldi solvers has enabled weak scaling to 0(105) cores on the Titan XK7 supercomputer. The multilevel decomposition provides a mechanism for scaling to exascale computing and beyond.

AB - We have implemented a new multilevel parallel decomposition in the Denovo discrete ordinates radiation transport code. In concert with Krylov subspace iterative solvers, the multilevel decomposition allows concurrency over energy in addition to space-angle, enabling scalability beyond the limits imposed by the traditional Koch-Baker-Alcouffe (KBA) space-angle partitioning. Furthermore, a new Arnoldi-based k-eigenvalue solver has been implemented. The added phase-space concurrency combined with the high-performance Krylov and Arnoldi solvers has enabled weak scaling to 0(105) cores on the Titan XK7 supercomputer. The multilevel decomposition provides a mechanism for scaling to exascale computing and beyond.

UR - https://www.scopus.com/pages/publications/84905506689

U2 - 10.13182/NSE12-101

DO - 10.13182/NSE12-101

M3 - Article

AN - SCOPUS:84905506689

SN - 0029-5639

VL - 177

SP - 111

EP - 125

JO - Nuclear Science and Engineering

JF - Nuclear Science and Engineering

IS - 2

ER -

Massively parallel, three-dimensional transport solutions for the k-eigenvalue problem

Abstract

Access to Document

Other files and links

Fingerprint

Cite this