DocumentCode :
2460250
Title :
Explicit Spatial Scattering for Load Balancing in Conservatively Synchronized Parallel Discrete Event Simulations
Author :
Thulasidasan, Sunil ; Kasiviswanathan, Shiva Prasad ; Eidenbenz, Stephan ; Romero, Phillip
Author_Institution :
Los Alamos Nat. Lab., Los Alamos, NM, USA
fYear :
2010
fDate :
17-19 May 2010
Firstpage :
1
Lastpage :
8
Abstract :
We re-examine the problem of load balancing in conservatively synchronized parallel, discrete- event simulations executed on high-performance computing clusters, focusing on simulations where computational and messaging load tend to be spatially clustered. Such domains are frequently characterized by the presence of geographic "hot-spots\´\´ - regions that generate significantly more simulation events than others. Examples of such domains include simulation of urban regions, transportation networks and networks where interaction between entities is often constrained by physical proximity. Noting that in conservatively synchronized parallel simulations, the speed of execution of the simulation is determined by the slowest ( i.e most heavily loaded) simulation process, we study different partitioning strategies in achieving equitable processor-load distribution in domains with spatially clustered load. In particular, we study the effectiveness of partitioning via spatial scattering to achieve optimal load balance. In this partitioning technique, nearby entities are explicitly assigned to different processors, thereby scattering the load across the cluster. This is motivated by two observations, namely, (i) since load is spatially clustered, spatial scattering should, intuitively, spread the load across the compute cluster, and (ii) in parallel simulations, equitable distribution of CPU load is a greater determinant of execution speed than message passing overhead. Through large-scale simulation experiments - both of abstracted and real simulation models - on high performance clusters, we observe that scatter partitioning - even with its greatly increased messaging overhead - often significantly outperforms more conventional spatial partitioning techniques that seek to reduce messaging overhead. Further, even if hot-spots change over the course of the simulation, if the underlying feature of spatial clustering is retained, load continues to be balanced with spatial sca- - ttering leading us to the observation that spatial scattering can often obviate the need for dynamic load balancing.
Keywords :
discrete event simulation; parallel processing; resource allocation; workstation clusters; CPU load; conservatively synchronized parallel discrete event simulations; dynamic load balancing; equitable processor load distribution; explicit spatial scattering; geographic hot-spot regions; large-scale simulation; message passing overhead; spatial partitioning techniques; spatial scattering; spatially clustered load; transportation networks; urban region simulation; Character generation; Computational modeling; Concurrent computing; Discrete event simulation; Distributed computing; Large-scale systems; Load management; Message passing; Scattering; Transportation;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Principles of Advanced and Distributed Simulation (PADS), 2010 IEEE Workshop on
Conference_Location :
Atlanta
ISSN :
1087-4097
Print_ISBN :
978-1-4244-7292-5
Electronic_ISBN :
1087-4097
Type :
conf
DOI :
10.1109/PADS.2010.5471664
Filename :
5471664
Link To Document :
بازگشت