Author:
Felker Kyle G.,Siegel Andrew R.,Smith Kord S.,Romano Paul K.,Forget Benoit
Abstract
An algorithm is developed to significantly reduce the on-node footprint of cross section memory in Monte Carlo particle tracking algorithms. The classic method of per-node replication of cross section data is replaced by a memory server model, in which the read-only lookup tables reside on a remote set of disjoint processors. The main particle tracking algorithm is then modified in such a way as to enable efficient use of the remotely stored data in the particle tracking algorithm. Results of a prototype code on a Blue Gene/Q installation reveal that the penalty for remote storage is reasonable in the context of time scales for real-world applications, thus yielding a path forward for a broad range of applications that are memory bound using current techniques.
Reference12 articles.
1. The effect of load imbalances on the performance of Monte Carlo algorithms in LWR analysis
2. Optimizing Memory Constrained Environments in Monte Carlo Nuclear Reactor Simulations
3. Brown F., Recent advances and future prospects for monte carlo, Tech. Rep. LA-UR-10-05634, Los Alamos National Laboratory (2010).
4. Irving D. C., R. M. F. Jr., Kam F., O5R, A general-purpose Monte Carlo neutron transport code, Tech. rep., Oak Ridge National Laboratory (1965).
5. Monte Carlo methods for radiation transport analysis on vector computers
Cited by
3 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献
1. Logically Parallel Communication for Fast MPI+Threads Applications;IEEE Transactions on Parallel and Distributed Systems;2021-12-01
2. RMACXX: An Efficient High-Level C++ Interface over MPI-3 RMA;2021 IEEE/ACM 21st International Symposium on Cluster, Cloud and Internet Computing (CCGrid);2021-05
3. How I learned to stop worrying about user-visible endpoints and love MPI;Proceedings of the 34th ACM International Conference on Supercomputing;2020-06-29