Probability Convergence in a Multithreaded Counting Application
Abstract
We introduce the PDtree, a data structure for fast counting of any specified combinations of a given set of variables. We describe a static implementation that provides speedup linear in the number of processors on the multithreaded Cray MTA2 machine. Finally, we prove a general convergence property that bounds the nondeterministic deviation of probability estimates relative to a sequential implementation. This convergence result is relevant for any counting application that takes a multithreaded streaming approach.
 Authors:
 Publication Date:
 Research Org.:
 Pacific Northwest National Lab. (PNNL), Richland, WA (United States)
 Sponsoring Org.:
 USDOE
 OSTI Identifier:
 910005
 Report Number(s):
 PNNLSA53295
TRN: US200723%%313
 DOE Contract Number:
 AC0576RL01830
 Resource Type:
 Conference
 Resource Relation:
 Conference: Parallel and Distributed Processing Symposium (IPDPS 2007), 15
 Country of Publication:
 United States
 Language:
 English
 Subject:
 99 GENERAL AND MISCELLANEOUS//MATHEMATICS, COMPUTING, AND INFORMATION SCIENCE; CONVERGENCE; PROBABILITY; P CODES; CRAY COMPUTERS; COMPUTER CALCULATIONS
Citation Formats
Scherrer, Chad, Beagley, Nathaniel, Nieplocha, Jarek, Marquez, Andres, Feo, John, and ChavarríaMiranda, Daniel. Probability Convergence in a Multithreaded Counting Application. United States: N. p., 2007.
Web. doi:10.1109/IPDPS.2007.370688.
Scherrer, Chad, Beagley, Nathaniel, Nieplocha, Jarek, Marquez, Andres, Feo, John, & ChavarríaMiranda, Daniel. Probability Convergence in a Multithreaded Counting Application. United States. doi:10.1109/IPDPS.2007.370688.
Scherrer, Chad, Beagley, Nathaniel, Nieplocha, Jarek, Marquez, Andres, Feo, John, and ChavarríaMiranda, Daniel. Fri .
"Probability Convergence in a Multithreaded Counting Application". United States.
doi:10.1109/IPDPS.2007.370688.
@article{osti_910005,
title = {Probability Convergence in a Multithreaded Counting Application},
author = {Scherrer, Chad and Beagley, Nathaniel and Nieplocha, Jarek and Marquez, Andres and Feo, John and ChavarríaMiranda, Daniel},
abstractNote = {We introduce the PDtree, a data structure for fast counting of any specified combinations of a given set of variables. We describe a static implementation that provides speedup linear in the number of processors on the multithreaded Cray MTA2 machine. Finally, we prove a general convergence property that bounds the nondeterministic deviation of probability estimates relative to a sequential implementation. This convergence result is relevant for any counting application that takes a multithreaded streaming approach.},
doi = {10.1109/IPDPS.2007.370688},
journal = {},
number = ,
volume = ,
place = {United States},
year = {Fri Mar 30 00:00:00 EDT 2007},
month = {Fri Mar 30 00:00:00 EDT 2007}
}
Other availability
Please see Document Availability for additional information on obtaining the fulltext document. Library patrons may search WorldCat to identify libraries that hold this conference proceeding.
Save to My Library
You must Sign In or Create an Account in order to save documents to your library.

Source convergence diagnostics using Boltzmann entropy criterion application to different OECD/NEA criticality benchmarks with the 3D Monte Carlo code Tripoli4
The measurement of the stationarity of Monte Carlo fission source distributions in k{sub eff} calculations plays a central role in the ability to discriminate between fake and 'true' convergence (in the case of a high dominant ratio or in case of loosely coupled systems). Recent theoretical developments have been made in the study of source convergence diagnostics, using Shannon entropy. We will first recall those results, and we will then generalize them using the expression of Boltzmann entropy, highlighting the gain in terms of the various physical problems that we can treat. Finally we will present the results of severalmore » 
Application of convergence acceleration to the reactor kinetic equations: A comparative study
This presentation provides a comparison of two methodologies for the solution of reactor kinetic equations, namely for a standard finite difference and a semianalytical approach. The abovementioned methods are implemented in a convergence acceleration framework to enhance their efficiency and a comparative study is reported to verify whether it is more convenient to use a rudimentary but fast algorithm (finite difference) with respect to the more refined but computationally intense approach of the semianalytical method. Performance on several test cases from the literature are compared. (authors)