Parallel 3D method of characteristics in MPACT
Abstract
A new parallel 3D MOC kernel has been developed and implemented in MPACT which makes use of the modular ray tracing technique to reduce computational requirements and to facilitate parallel decomposition. The parallel model makes use of both distributed and shared memory parallelism which are implemented with the MPI and OpenMP standards, respectively. The kernel is capable of parallel decomposition of problems in space, angle, and by characteristic rays up to 0(104) processors. Initial verification of the parallel 3D MOC kernel was performed using the Takeda 3D transport benchmark problems. The eigenvalues computed by MPACT are within the statistical uncertainty of the benchmark reference and agree well with the averages of other participants. The MPACT k{sub eff} differs from the benchmark results for rodded and unrodded cases by 11 and 40 pcm, respectively. The calculations were performed for various numbers of processors and parallel decompositions up to 15625 processors; all producing the same result at convergence. The parallel efficiency of the worst case was 60%, while very good efficiency (>95%) was observed for cases using 500 processors. The overall run time for the 500 processor case was 231 seconds and 19 seconds for the case with 15625 processors. Ongoing workmore »
 Authors:
 Department of Nuclear Engineering and Radiological Sciences, University of Michigan, 2200 Bonisteel, Ann Arbor, MI 48109 (United States)
 School of Nuclear Science and Technology, Xi'an Jiaotong University, No. 28 Xianning West Road, Xi'an, Shaanxi 710049 (China)
 Publication Date:
 Research Org.:
 American Nuclear Society, 555 North Kensington Avenue, La Grange Park, IL 60526 (United States)
 OSTI Identifier:
 22212693
 Resource Type:
 Conference
 Resource Relation:
 Conference: M and C 2013: 2013 International Conference on Mathematics and Computational Methods Applied to Nuclear Science and Engineering, Sun Valley, ID (United States), 59 May 2013; Other Information: Country of input: France; 17 refs.; Related Information: In: Proceedings of the 2013 International Conference on Mathematics and Computational Methods Applied to Nuclear Science and Engineering  M and C 2013 3016 p.
 Country of Publication:
 United States
 Language:
 English
 Subject:
 97 MATHEMATICAL METHODS AND COMPUTING; BENCHMARKS; COMPUTER CALCULATIONS; COMPUTER CODES; CONVERGENCE; EIGENVALUES; NEUTRON TRANSPORT; PARALLEL PROCESSING; PERFORMANCE; THREEDIMENSIONAL CALCULATIONS
Citation Formats
Kochunas, B., Dovvnar, T. J., and Liu, Z.. Parallel 3D method of characteristics in MPACT. United States: N. p., 2013.
Web.
Kochunas, B., Dovvnar, T. J., & Liu, Z.. Parallel 3D method of characteristics in MPACT. United States.
Kochunas, B., Dovvnar, T. J., and Liu, Z.. Mon .
"Parallel 3D method of characteristics in MPACT". United States.
doi:.
@article{osti_22212693,
title = {Parallel 3D method of characteristics in MPACT},
author = {Kochunas, B. and Dovvnar, T. J. and Liu, Z.},
abstractNote = {A new parallel 3D MOC kernel has been developed and implemented in MPACT which makes use of the modular ray tracing technique to reduce computational requirements and to facilitate parallel decomposition. The parallel model makes use of both distributed and shared memory parallelism which are implemented with the MPI and OpenMP standards, respectively. The kernel is capable of parallel decomposition of problems in space, angle, and by characteristic rays up to 0(104) processors. Initial verification of the parallel 3D MOC kernel was performed using the Takeda 3D transport benchmark problems. The eigenvalues computed by MPACT are within the statistical uncertainty of the benchmark reference and agree well with the averages of other participants. The MPACT k{sub eff} differs from the benchmark results for rodded and unrodded cases by 11 and 40 pcm, respectively. The calculations were performed for various numbers of processors and parallel decompositions up to 15625 processors; all producing the same result at convergence. The parallel efficiency of the worst case was 60%, while very good efficiency (>95%) was observed for cases using 500 processors. The overall run time for the 500 processor case was 231 seconds and 19 seconds for the case with 15625 processors. Ongoing work is focused on developing theoretical performance models and the implementation of acceleration techniques to minimize the number of iterations to converge. (authors)},
doi = {},
journal = {},
number = ,
volume = ,
place = {United States},
year = {Mon Jul 01 00:00:00 EDT 2013},
month = {Mon Jul 01 00:00:00 EDT 2013}
}

MPACT (Michigan Parallel Characteristics Transport Code) is a new reactor analysis tool. It is being developed by students and research staff at the University of Michigan to be used for an advanced pinresolved transport capability within VERA (Virtual Environment for Reactor Analysis). VERA is the enduser reactor simulation tool being produced by the Consortium for the Advanced Simulation of Light Water Reactors (CASL). The MPACT development project is itself unique for the way it is changing how students do research to achieve the instructional and research goals of an academic institution, while providing immediate value to industry. The MPACT codemore »

Assessment of the 2D MOC solver in MPACT: Michigan parallel characteristics transport code
MPACT (Michigan Parallel Characteristics Transport Code) is a new reactor analysis tool being developed by researchers at the University of Michigan as an advanced pinresolved transport capability within VERA (Virtual Environment for Reactor Analysis). VERA is the enduser reactor simulation tool being developed by the Consortium for the Advanced Simulation of Light Water Reactors (CASL). The MPACT development project is itself unique for the way it is changing how students perform research to achieve the instructional and research goals of an academic institution, while providing immediate value to the industry. One of the major computational pieces in MPACT is themore » 
A parallel iterative method for solving 3D elliptic partial differential equations
In this paper, we will consider the parallel solution of the Dirichlet problem for a secondorder uniformly elliptic equation in two and three dimensions. Specifically, we shall consider the problem, L(u) = f in {Omega}, u = g on {Gamma}, where L(u) = {minus} summation ({partial derivative}/{partial derivative}x{sub i}(a{sub i}({partial derivative}u/{partial derivative}x{sub i}))) from i = 1 to 3 with a{sub i} positive, bounded, and piecewise smooth on bounded {Omega} with boundary {Gamma}. For sake of exposition, we will assume the equation is 2dimensional, however, the extension to 3dimensions of the numerical methods to be described will be obvious. 
High order spatial expansion for the method of characteristics applied to 3D geometries
The method of characteristics is an efficient and flexible technique to solve the neutron transport equation and has been extensively used in twodimensional calculations because it permits to deal with complex geometries. However, because of a very fast increase in storage requirements and number of floating operations, its direct application to threedimensional routine transport calculations it is not still possible. In this work we introduce and analyze several modifications aimed to reduce memory requirements and to diminish the computing burden. We explore highorder spatial approximation, the use of intermediary trajectorydependent flux expansions and the possibility of dynamic trajectory reconstruction frommore »