Limits on the efficiency of event-based algorithms for Monte Carlo neutron transport
Abstract
The traditional form of parallelism in Monte Carlo particle transport simulations, wherein each individual particle history is considered a unit of work, does not lend itself well to data-level parallelism. Event-based algorithms, which were originally used for simulations on vector processors, may offer a path toward better utilizing data-level parallelism in modern computer architectures. In this study, a simple model is developed for estimating the efficiency of the event-based particle transport algorithm under two sets of assumptions. Data collected from simulations of four reactor problems using OpenMC was then used in conjunction with the models to calculate the speedup due to vectorization as a function of the size of the particle bank and the vector width. When each event type is assumed to have constant execution time, the achievable speedup is directly related to the particle bank size. We observed that the bank size generally needs to be at least 20 times greater than vector size to achieve vector efficiency greater than 90%. Lastly, when the execution times for events are allowed to vary, the vector speedup is also limited by differences in execution time for events being carried out in a single event-iteration.
- Authors:
- Publication Date:
- Research Org.:
- Argonne National Lab. (ANL), Argonne, IL (United States)
- Sponsoring Org.:
- USDOE Office of Science (SC), Basic Energy Sciences (BES)
- OSTI Identifier:
- 1371489
- Alternate Identifier(s):
- OSTI ID: 1402080
- Grant/Contract Number:
- AC02-06CH11357
- Resource Type:
- Published Article
- Journal Name:
- Nuclear Engineering and Technology
- Additional Journal Information:
- Journal Name: Nuclear Engineering and Technology Journal Volume: 49 Journal Issue: 6; Journal ID: ISSN 1738-5733
- Publisher:
- Elsevier
- Country of Publication:
- Korea, Republic of
- Language:
- English
- Subject:
- 97 MATHEMATICS AND COMPUTING; Monte Carlo; Neutron transport; Parallelism; Vectorization
Citation Formats
Romano, Paul K., and Siegel, Andrew R. Limits on the efficiency of event-based algorithms for Monte Carlo neutron transport. Korea, Republic of: N. p., 2017.
Web. doi:10.1016/j.net.2017.06.006.
Romano, Paul K., & Siegel, Andrew R. Limits on the efficiency of event-based algorithms for Monte Carlo neutron transport. Korea, Republic of. https://doi.org/10.1016/j.net.2017.06.006
Romano, Paul K., and Siegel, Andrew R. Fri .
"Limits on the efficiency of event-based algorithms for Monte Carlo neutron transport". Korea, Republic of. https://doi.org/10.1016/j.net.2017.06.006.
@article{osti_1371489,
title = {Limits on the efficiency of event-based algorithms for Monte Carlo neutron transport},
author = {Romano, Paul K. and Siegel, Andrew R.},
abstractNote = {The traditional form of parallelism in Monte Carlo particle transport simulations, wherein each individual particle history is considered a unit of work, does not lend itself well to data-level parallelism. Event-based algorithms, which were originally used for simulations on vector processors, may offer a path toward better utilizing data-level parallelism in modern computer architectures. In this study, a simple model is developed for estimating the efficiency of the event-based particle transport algorithm under two sets of assumptions. Data collected from simulations of four reactor problems using OpenMC was then used in conjunction with the models to calculate the speedup due to vectorization as a function of the size of the particle bank and the vector width. When each event type is assumed to have constant execution time, the achievable speedup is directly related to the particle bank size. We observed that the bank size generally needs to be at least 20 times greater than vector size to achieve vector efficiency greater than 90%. Lastly, when the execution times for events are allowed to vary, the vector speedup is also limited by differences in execution time for events being carried out in a single event-iteration.},
doi = {10.1016/j.net.2017.06.006},
journal = {Nuclear Engineering and Technology},
number = 6,
volume = 49,
place = {Korea, Republic of},
year = {Fri Sep 01 00:00:00 EDT 2017},
month = {Fri Sep 01 00:00:00 EDT 2017}
}
https://doi.org/10.1016/j.net.2017.06.006
Web of Science