skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Dynamic Load Balancing of Parallel Monte Carlo Transport Calculations

Abstract

The performance of parallel Monte Carlo transport calculations which use both spatial and particle parallelism is increased by dynamically assigning processors to the most worked domains. Since the particle work load varies over the course of the simulation, this algorithm determines each cycle if dynamic load balancing would speed up the calculation. If load balancing is required, a small number of particle communications are initiated in order to achieve load balance. This method has decreased the parallel run time by more than a factor of three for certain criticality calculations.

Authors:
; ;
Publication Date:
Research Org.:
Lawrence Livermore National Lab., Livermore, CA (US)
Sponsoring Org.:
US Department of Energy (US)
OSTI Identifier:
15015938
Report Number(s):
UCRL-CONF-208938
TRN: US0501670
DOE Contract Number:
W-7405-ENG-48
Resource Type:
Conference
Resource Relation:
Conference: Presented at: Monte Carlo 2005, Chattanooga, TN (US), 04/17/2005--04/21/2005; Other Information: PBD: 22 Dec 2004
Country of Publication:
United States
Language:
English
Subject:
22 GENERAL STUDIES OF NUCLEAR REACTORS; ALGORITHMS; COMMUNICATIONS; CRITICALITY; DYNAMIC LOADS; PERFORMANCE; SIMULATION; TRANSPORT; VELOCITY

Citation Formats

O'Brien, M, Taylor, J, and Procassini, R. Dynamic Load Balancing of Parallel Monte Carlo Transport Calculations. United States: N. p., 2004. Web.
O'Brien, M, Taylor, J, & Procassini, R. Dynamic Load Balancing of Parallel Monte Carlo Transport Calculations. United States.
O'Brien, M, Taylor, J, and Procassini, R. Wed . "Dynamic Load Balancing of Parallel Monte Carlo Transport Calculations". United States. doi:. https://www.osti.gov/servlets/purl/15015938.
@article{osti_15015938,
title = {Dynamic Load Balancing of Parallel Monte Carlo Transport Calculations},
author = {O'Brien, M and Taylor, J and Procassini, R},
abstractNote = {The performance of parallel Monte Carlo transport calculations which use both spatial and particle parallelism is increased by dynamically assigning processors to the most worked domains. Since the particle work load varies over the course of the simulation, this algorithm determines each cycle if dynamic load balancing would speed up the calculation. If load balancing is required, a small number of particle communications are initiated in order to achieve load balance. This method has decreased the parallel run time by more than a factor of three for certain criticality calculations.},
doi = {},
journal = {},
number = ,
volume = ,
place = {United States},
year = {Wed Dec 22 00:00:00 EST 2004},
month = {Wed Dec 22 00:00:00 EST 2004}
}

Conference:
Other availability
Please see Document Availability for additional information on obtaining the full-text document. Library patrons may search WorldCat to identify libraries that hold this conference proceeding.

Save / Share:
  • The performance of parallel Monte Carlo transport calculations which use both spatial and particle parallelism is increased by dynamically assigning processors to the most worked domains. Since he particle work load varies over the course of the simulation, this algorithm determines each cycle if dynamic load balancing would speed up the calculation. If load balancing is required, a small number of particle communications are initiated in order to achieve load balance. This method has decreased the parallel run time by more than a factor of three for certain criticality calculations.
  • In order to run computer simulations efficiently on massively parallel computers with hundreds of thousands or millions of processors, care must be taken that the calculation is load balanced across the processors. Examining the workload of every processor leads to an unscalable algorithm, with run time at least as large as O(N), where N is the number of processors. We present a scalable load balancing algorithm, with run time 0(log(N)), that involves iterated processor-pair-wise balancing steps, ultimately leading to a globally balanced workload. We demonstrate scalability of the algorithm up to 2 million processors on the Sequoia supercomputer at Lawrencemore » Livermore National Laboratory. (authors)« less
  • Computational requirements of full scale computational fluid dynamics change as computation progresses on a parallel machine. The change in computational intensity causes workload imbalance of processors, which in turn requires a large amount of data movement at runtime. If parallel CFD is to be successful on a parallel or massively parallel machine, balancing of the runtime load is indispensable. Here a frame work is presented for dynamic load balancing for CFD applications, called Jove. One processor is designated as a decision maker Jove while others are assigned to computational fluid dynamics. Processors running CFD send flags to Jove in amore » predetermined number of iterations to initiate load balancing. Jove starts working on load balancing while other processors continue working with the current data and load distribution. Jove goes through several steps to decide if the new data should be taken, including preliminary evaluate, partition, processor reassignment, cost evaluation, and decision. Jove running on a single SP2 node has been completely implemented. Preliminary experimental results show that the Jove approach to dynamic load balancing can be effective for full scale grid partitioning on the target machine SP2.« less