Performance properties of large scale parallel systems
Journal of Parallel and Distributed Computing - Special issue on performance of supercomputers
Analyzing scalability of parallel algorithms and architectures
Journal of Parallel and Distributed Computing - Special issue on scalability of parallel algorithms and architectures
Journal of Parallel and Distributed Computing - Special issue on scalability of parallel algorithms and architectures
Using MPI: portable parallel programming with the message-passing interface
Using MPI: portable parallel programming with the message-passing interface
Parallel and Distributed Computing: A Survey of Models, Paradigms and Approaches
Parallel and Distributed Computing: A Survey of Models, Paradigms and Approaches
Isoefficiency: Measuring the Scalability of Parallel Algorithms and Architectures
IEEE Parallel & Distributed Technology: Systems & Technology
Scalability of Parallel Algorithm-Machine Combinations
IEEE Transactions on Parallel and Distributed Systems
Hi-index | 0.00 |
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.