Optimization of processor allocation for domain decomposed Monte Carlo calculations

•Load imbalance problems can be alleviated through a nonuniform allocation of processors.•Allocations are optimized using runtime diagnostics of a small calibration calculation.•Method cheaply and robustly improves parallel efficiency of radiation transport applications. Load imbalance plagues domai...

Full description

Saved in:
Bibliographic Details
Published in:Parallel computing 2019-09, Vol.87 (C), p.77-86
Main Authors: Ellis, J. Austin, Evans, Thomas M., Hamilton, Steven P., Kelley, C.T., Pandya, Tara M.
Format: Article
Language:English
Subjects:
Domain decomposition
Load imbalance
MATHEMATICS AND COMPUTING
Monte Carlo transport
Noisy optimization
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:•Load imbalance problems can be alleviated through a nonuniform allocation of processors.•Allocations are optimized using runtime diagnostics of a small calibration calculation.•Method cheaply and robustly improves parallel efficiency of radiation transport applications. Load imbalance plagues domain decomposed Monte Carlo calculations when sources are not uniform. Parallel efficiency for domain decomposed Monte Carlo transport calculations improves through a nonuniform allocation of processors over subdomains. We optimize the allocation with runtime diagnostics collected during a calibration step, then complete the full calculation. The diagnostic-based approach is compared to implicit filtering, an optimization algorithm for bound constrained noisy optimization problems. We consider both forward and hybrid radiation transport calculations to measure performance.
ISSN:0167-8191
1872-7336
DOI:10.1016/j.parco.2019.06.001