Multigroup Monte Carlo on GPUs: Comparison of history- and event-based algorithms

Steven P. Hamilton; Stuart R. Slattery; Thomas M. Evans

doi:10.1016/j.anucene.2017.11.032

Multigroup Monte Carlo on GPUs: Comparison of history- and event-based algorithms

Steven P. Hamilton, Stuart R. Slattery, Thomas M. Evans

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

27 Scopus citations

Abstract

This paper presents an investigation of the performance of different multigroup Monte Carlo transport algorithms on GPUs with a discussion of both history-based and event-based approaches. Several algorithmic improvements are introduced for both approaches. By modifying the history-based algorithm that is traditionally favored in CPU-based MC codes to occasionally filter out dead particles to reduce thread divergence, performance exceeds that of either the pure history-based or event-based approaches. The impacts of several algorithmic choices are discussed, including performance studies on Kepler and Pascal generation NVIDIA GPUs for fixed source and eigenvalue calculations. Single-device performance equivalent to 20–40 CPU cores on the K40 GPU and 60–80 CPU cores on the P100 GPU is achieved. In addition, nearly perfect multi-device parallel weak scaling is demonstrated on more than 16,000 nodes of the Titan supercomputer.

Original language	English
Pages (from-to)	506-518
Number of pages	13
Journal	Annals of Nuclear Energy
Volume	113
DOIs	https://doi.org/10.1016/j.anucene.2017.11.032
State	Published - Mar 2018

Funding

This research was sponsored by the Laboratory Directed Research and Development Program of Oak Ridge National Laboratory , managed by UT-Battelle, LLC, for the U.S. Department of Energy. This research was supported by the Exascale Computing Project (ECP), Project Number: 17-SC-20-SC, a collaborative effort of two DOE organizations—the Office of Science and the National Nuclear Security Administration—responsible for the planning and preparation of a capable exascale ecosystem—including software, applications, hardware, advanced system engineering, and early testbed platforms—to support the nation’s exascale computing imperative. This research used resources of the Oak Ridge Leadership Computing Facility at the Oak Ridge National Laboratory, which is supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC05-00OR22725.

Funders	Funder number
U.S. Department of Energy	17-SC-20-SC
Office of Science
National Nuclear Security Administration
Oak Ridge National Laboratory

Keywords

GPU
Monte Carlo
Radiation transport

Access to Document

10.1016/j.anucene.2017.11.032

Cite this

@article{0225e2fc55ad49acaac8866affb20635,

title = "Multigroup Monte Carlo on GPUs: Comparison of history- and event-based algorithms",

abstract = "This paper presents an investigation of the performance of different multigroup Monte Carlo transport algorithms on GPUs with a discussion of both history-based and event-based approaches. Several algorithmic improvements are introduced for both approaches. By modifying the history-based algorithm that is traditionally favored in CPU-based MC codes to occasionally filter out dead particles to reduce thread divergence, performance exceeds that of either the pure history-based or event-based approaches. The impacts of several algorithmic choices are discussed, including performance studies on Kepler and Pascal generation NVIDIA GPUs for fixed source and eigenvalue calculations. Single-device performance equivalent to 20–40 CPU cores on the K40 GPU and 60–80 CPU cores on the P100 GPU is achieved. In addition, nearly perfect multi-device parallel weak scaling is demonstrated on more than 16,000 nodes of the Titan supercomputer.",

keywords = "GPU, Monte Carlo, Radiation transport",

author = "Hamilton, {Steven P.} and Slattery, {Stuart R.} and Evans, {Thomas M.}",

note = "Publisher Copyright: {\textcopyright} 2017 Elsevier Ltd",

year = "2018",

month = mar,

doi = "10.1016/j.anucene.2017.11.032",

language = "English",

volume = "113",

pages = "506--518",

journal = "Annals of Nuclear Energy",

issn = "0306-4549",

publisher = "Elsevier",

}

TY - JOUR

T1 - Multigroup Monte Carlo on GPUs

T2 - Comparison of history- and event-based algorithms

AU - Hamilton, Steven P.

AU - Slattery, Stuart R.

AU - Evans, Thomas M.

PY - 2018/3

Y1 - 2018/3

N2 - This paper presents an investigation of the performance of different multigroup Monte Carlo transport algorithms on GPUs with a discussion of both history-based and event-based approaches. Several algorithmic improvements are introduced for both approaches. By modifying the history-based algorithm that is traditionally favored in CPU-based MC codes to occasionally filter out dead particles to reduce thread divergence, performance exceeds that of either the pure history-based or event-based approaches. The impacts of several algorithmic choices are discussed, including performance studies on Kepler and Pascal generation NVIDIA GPUs for fixed source and eigenvalue calculations. Single-device performance equivalent to 20–40 CPU cores on the K40 GPU and 60–80 CPU cores on the P100 GPU is achieved. In addition, nearly perfect multi-device parallel weak scaling is demonstrated on more than 16,000 nodes of the Titan supercomputer.

AB - This paper presents an investigation of the performance of different multigroup Monte Carlo transport algorithms on GPUs with a discussion of both history-based and event-based approaches. Several algorithmic improvements are introduced for both approaches. By modifying the history-based algorithm that is traditionally favored in CPU-based MC codes to occasionally filter out dead particles to reduce thread divergence, performance exceeds that of either the pure history-based or event-based approaches. The impacts of several algorithmic choices are discussed, including performance studies on Kepler and Pascal generation NVIDIA GPUs for fixed source and eigenvalue calculations. Single-device performance equivalent to 20–40 CPU cores on the K40 GPU and 60–80 CPU cores on the P100 GPU is achieved. In addition, nearly perfect multi-device parallel weak scaling is demonstrated on more than 16,000 nodes of the Titan supercomputer.

KW - GPU

KW - Monte Carlo

KW - Radiation transport

UR - http://www.scopus.com/inward/record.url?scp=85037543036&partnerID=8YFLogxK

U2 - 10.1016/j.anucene.2017.11.032

DO - 10.1016/j.anucene.2017.11.032

M3 - Article

AN - SCOPUS:85037543036

SN - 0306-4549

VL - 113

SP - 506

EP - 518

JO - Annals of Nuclear Energy

JF - Annals of Nuclear Energy

ER -

Multigroup Monte Carlo on GPUs: Comparison of history- and event-based algorithms

Abstract

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this