skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Performance and accuracy of criticality calculations performed using WARP – A framework for continuous energy Monte Carlo neutron transport in general 3D geometries on GPUs

Abstract

In this companion paper to "Algorithmic Choices in WARP - A Framework for Continuous Energy Monte Carlo Neutron Transport in General 3D Geometries on GPUs" (doi:10.1016/j.anucene.2014.10.039), the WARP Monte Carlo neutron transport framework for graphics processing units (GPUs) is benchmarked against production-level central processing unit (CPU) Monte Carlo neutron transport codes for both performance and accuracy. We compare neutron flux spectra, multiplication factors, runtimes, speedup factors, and costs of various GPU and CPU platforms running either WARP, Serpent 2.1.24, or MCNP 6.1. WARP compares well with the results of the production-level codes, and it is shown that on the newest hardware considered, GPU platforms running WARP are between 0.8 to 7.6 times as fast as CPU platforms running production codes. Also, the GPU platforms running WARP were between 15% and 50% as expensive to purchase and between 80% to 90% as expensive to operate as equivalent CPU platforms performing at an equal simulation rate.

Authors:
 [1];  [1];  [1];  [1];  [1]
  1. Univ. of California, Berkeley, CA (United States). Dept. of Nuclear Engineering
Publication Date:
Research Org.:
Univ. of California, Berkeley, CA (United States)
Sponsoring Org.:
USDOE National Nuclear Security Administration (NNSA)
OSTI Identifier:
1344092
Grant/Contract Number:
NA0000979
Resource Type:
Journal Article: Accepted Manuscript
Journal Name:
Annals of Nuclear Energy (Oxford)
Additional Journal Information:
Journal Name: Annals of Nuclear Energy (Oxford); Journal Volume: 103; Journal Issue: C; Journal ID: ISSN 0306-4549
Publisher:
Elsevier
Country of Publication:
United States
Language:
English
Subject:
73 NUCLEAR PHYSICS AND RADIATION PHYSICS; Monte Carlo; Neutron Transport; GPU; CUDA; OptiX; Supercomputing

Citation Formats

Bergmann, Ryan M., Rowland, Kelly L., Radnović, Nikola, Slaybaugh, Rachel N., and Vujić, Jasmina L. Performance and accuracy of criticality calculations performed using WARP – A framework for continuous energy Monte Carlo neutron transport in general 3D geometries on GPUs. United States: N. p., 2017. Web. doi:10.1016/j.anucene.2017.01.027.
Bergmann, Ryan M., Rowland, Kelly L., Radnović, Nikola, Slaybaugh, Rachel N., & Vujić, Jasmina L. Performance and accuracy of criticality calculations performed using WARP – A framework for continuous energy Monte Carlo neutron transport in general 3D geometries on GPUs. United States. doi:10.1016/j.anucene.2017.01.027.
Bergmann, Ryan M., Rowland, Kelly L., Radnović, Nikola, Slaybaugh, Rachel N., and Vujić, Jasmina L. 2017. "Performance and accuracy of criticality calculations performed using WARP – A framework for continuous energy Monte Carlo neutron transport in general 3D geometries on GPUs". United States. doi:10.1016/j.anucene.2017.01.027.
@article{osti_1344092,
title = {Performance and accuracy of criticality calculations performed using WARP – A framework for continuous energy Monte Carlo neutron transport in general 3D geometries on GPUs},
author = {Bergmann, Ryan M. and Rowland, Kelly L. and Radnović, Nikola and Slaybaugh, Rachel N. and Vujić, Jasmina L.},
abstractNote = {In this companion paper to "Algorithmic Choices in WARP - A Framework for Continuous Energy Monte Carlo Neutron Transport in General 3D Geometries on GPUs" (doi:10.1016/j.anucene.2014.10.039), the WARP Monte Carlo neutron transport framework for graphics processing units (GPUs) is benchmarked against production-level central processing unit (CPU) Monte Carlo neutron transport codes for both performance and accuracy. We compare neutron flux spectra, multiplication factors, runtimes, speedup factors, and costs of various GPU and CPU platforms running either WARP, Serpent 2.1.24, or MCNP 6.1. WARP compares well with the results of the production-level codes, and it is shown that on the newest hardware considered, GPU platforms running WARP are between 0.8 to 7.6 times as fast as CPU platforms running production codes. Also, the GPU platforms running WARP were between 15% and 50% as expensive to purchase and between 80% to 90% as expensive to operate as equivalent CPU platforms performing at an equal simulation rate.},
doi = {10.1016/j.anucene.2017.01.027},
journal = {Annals of Nuclear Energy (Oxford)},
number = C,
volume = 103,
place = {United States},
year = 2017,
month = 5
}

Journal Article:
Free Publicly Available Full Text
This content will become publicly available on May 1, 2018
Publisher's Version of Record

Save / Share:
  • Version 00 (1) Problems to be solved: MVP/GMVP II can solve eigenvalue and fixed-source problems. The multigroup code GMVP can solve forward and adjoint problems for neutron, photon and neutron-photon coupled transport. The continuous-energy code MVP can solve only the forward problems. Both codes can also perform time-dependent calculations. (2) Geometry description: MVP/GMVP employs combinatorial geometry to describe the calculation geometry. It describes spatial regions by the combination of the 3-dimensional objects (BODIes). Currently, the following objects (BODIes) can be used. - BODIes with linear surfaces : half space, parallelepiped, right parallelepiped, wedge, right hexagonal prism - BODIes with quadraticmore » surface and linear surfaces : cylinder, sphere, truncated right cone, truncated elliptic cone, ellipsoid by rotation, general ellipsoid - Arbitrary quadratic surface and torus The rectangular and hexagonal lattice geometry can be used to describe the repeated geometry. Furthermore, the statistical geometry model is available to treat coated fuel particles or pebbles for high temperature reactors. (3) Particle sources: The various forms of energy-, angle-, space- and time-dependent distribution functions can be specified. See Abstract for more detail.« less
  • Validation of the continuous-energy Monte Carlo criticality-safety analysis system, comprising the MVP code and neutron cross sections based on JENDL-3.2, was examined using benchmarks evaluated in the 'International Handbook of Evaluated Criticality Safety Benchmark Experiments'. Eight experiments (116 configurations) for the plutonium solution and plutonium-uranium mixture systems performed at Valduc, Battelle Pacific Northwest Laboratories, and other facilities were selected and used in the studies. The averaged multiplication factors calculated with MVP and MCNP-4B using the same neutron cross-section libraries based on JENDL-3.2 were in good agreement. Based on methods provided in the Japanese nuclear criticality-safety handbook, the estimated criticality lower-limitmore » multiplication factors to be used as a subcriticality criterion for the criticality-safety evaluation of nuclear facilities were obtained. The analysis proved the applicability of the MVP code to the criticality-safety analysis of nuclear fuel facilities, particularly to the analysis of systems fueled with plutonium and in homogeneous and thermal-energy conditions.« less
  • We present an energy banding algorithm for Monte Carlo (MC) neutral particle transport simulations which depend on large cross section lookup tables. In MC codes, read-only cross section data tables are accessed frequently, exhibit poor locality, and are typically too much large to fit in fast memory. Thus, performance is often limited by long latencies to RAM, or by off-node communication latencies when the data footprint is very large and must be decomposed on a distributed memory machine. The proposed energy banding algorithm allows maximal temporal reuse of data in band sizes that can flexibly accommodate different architectural features. Themore » energy banding algorithm is general and has a number of benefits compared to the traditional approach. In the present analysis we explore its potential to achieve improvements in time-to-solution on modern cache-based architectures.« less
  • The purpose of the work described in this paper was to evaluate the adequacy of criticality accident alarm system (CAAS) coverage of several buildings at the Portsmouth Gaseous Diffusion Plant (PORTS) located in Piketon, Ohio. An analysis was performed in which the General Monte Carlo N-Particle Transport Code (MCNP) was used to model several of the buildings at PORTS, simulate criticality accidents, and then tally the neutron fluence at the location of the detectors.