This Article Full Text (PDF) References Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in Web of Science Alert me to new issues of the journal Add to Saved Citations Download to citation manager Request Permissions Request Reprints Add to My Marked Citations Citing Articles Citing Articles via Google Scholar Citing Articles via Scopus Google Scholar Articles by Dahmani, M. Articles by Roy, R. Search for Related Content Social Bookmarking                         What's this?

Scalability Modeling For Deterministic Particle Transport Solvers

Nuclear Engineering Institute École Polytechnique De MontrÉAl P.O.Box 6079, Station Cv, MontrÉal, QuÉBec, Canada H3c 3a7

Robert Roy

Department Of Computer Engineering École Polytechnique De MontrÉal P.O.Box 6079, Station Cv, MontrÉal, QuÉBec, Canada H3c 3a7; Robert.Roy{at}Polymtl.Ca

In this paper, we propose a new parallel solver for the large-scale 3D neutron transport problems used in nuclear reactor simulations. Modern large-memory computers have made possible direct application of transport methods to large-scale computational models. However, many numerical acceleration techniques common to lattice transport codes are not applicable to heterogeneous geometry, especially high-dominance ratio when eigenvalue problems are solved. Consequently, large heterogeneous reactor problems have remained computationally intensive and impractical for routine engineering applications. Based on the characteristics method, this new code solves the transport equation by following neutron tracks. Due to the excessive number of these tracks in the demanding context of 3D large-scale calculations, the parallelization of the solver is the only way to obtain a fast solution. The parallelization is based on distributing a group of tracks, generated by a common ray tracing procedure, on several processors. An analytical model for the communication/ computation ratio is presented in order to predict the specific performance on different kind of architectures. A scalability analysis based on the isoefficiency function is performed on our parallel code when we increase the size of the problem and the number of processors. Tests are done to validate the analytical model by comparing the results to those given by the empirical tests. Results show that our parallel code is scalable and portable.

Key Words: transport theory • method of characteristics • performance model • scalability analysis

International Journal of High Performance Computing Applications, Vol. 20, No. 4, 541-556 (2006)
DOI: 10.1177/1094342006064571


CiteULike    Complore    Connotea    Del.icio.us    Digg    Reddit    Technorati    Twitter    What's this?