Simulating Spatially Explicit Problems on High Performance Architectures

Deelman, Ewa; Szymanski, Boleslaw K.

doi:10.1006/jpdc.2001.1802

Cited by 5 publications

(13 citation statements)

References 49 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…An algorithm based on process migration -a challenging task in itself -for dynamic load balancing for conservative parallel simulations is described in [2]. Another dynamic load balancing algorithm based on spatial partitioning for a simulation employing optimistic synchronization is described in [4] and [5], which involves the movement of spatial data between neighboring processes. A scheme using PVM [1] for parallel discrete event simulations is described in [8].…”

Section: Introductionmentioning

confidence: 99%

Explicit Spatial Scattering for Load Balancing in Conservatively Synchronized Parallel Discrete Event Simulations

Thulasidasan

Kasiviswanathan

Eidenbenz

et al. 2010

2010 IEEE Workshop on Principles of Advanced and Distributed Simulation

View full text Add to dashboard Cite

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 modelswe observe that scatter partitioning, even with its greatly increased messaging overhead, 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 scattering leading us to the observation that spatial scattering can often obviate the need for dynamic load balancing.

show abstract

Section: Introductionmentioning

confidence: 99%

Explicit Spatial Scattering for Load Balancing in Conservatively Synchronized Parallel Discrete Event Simulations

Thulasidasan

Kasiviswanathan

Eidenbenz

et al. 2010

2010 IEEE Workshop on Principles of Advanced and Distributed Simulation

View full text Add to dashboard Cite

show abstract

“…Simulations of Lyme disease are typically performed using spatially explicit models [7,6]. Spatially explicit models provide a convenient conceptual representation for modeling Eco systems as they have an explicit notion of space or environment [7,6,16,17]. The motivation to treat the environment as a separate dimension is that it undergoes its own sequence of changes.…”

Section: Introductionmentioning

confidence: 99%

“…In addition, capacity (or resource) limitations of the workstations often prevent simulation of large models [13]. Different approaches have been proposed to address these bottlenecks [7,6,13]. Amongst the various approaches, the application of parallel simulation methodologies have shown to provide considerable capacity and performance improvements [7,6].…”

Section: Introductionmentioning

confidence: 99%

“…In order to achieve efficient parallel simulations, spatially explicit models are subdivided into smaller, nonoverlapping regions [7,6]. Subdivision of space (say into k subspaces) reduces the average time complexity for simulating interactions between each pair of n entities in the simulation from O(n 2 ) to O(k * [(n/k) 2 ]).…”

Section: Introductionmentioning

confidence: 99%

“…Amongst the various methods, computer-based simulations have gained significant importance primarily because they provide a powerful and intuitive mechanism to explore scenarios that cannot be analyzed using other approaches. Simulations of Lyme disease are typically performed using spatially explicit models [7,6]. Spatially explicit models provide a convenient conceptual representation for modeling Eco systems as they have an explicit notion of space or environment [7,6,16,17].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Accelerating Spatially Explicit Simulations of Spread of Lyme Disease

Rao¹,

Wilsey²

38th Annual Simulation Symposium

View full text Add to dashboard Cite

The factors influencing spread of Lyme disease are often studied using computer-based simulations and spatially explicit models. However, simulating large and complex models is a time consuming task, even when parallel simulation techniques are employed. In an endeavor to accelerate such simulations, an alternative approach involving dynamic (i.e., during simulation) changes to spatial resolution of the model via a novel methodology called Dynamic Component Substitution (DCS) is proposed. Changes to the resolution are performed such that the total number of interactions between the entities in the model is optimized, thereby improving overall performance but introducing minor (< ±1%) deviations in the results. This paper explores the effectiveness and issues involved in applying DCS to accelerate sequential and parallel simulations of spatially explicit Lyme disease models. The paper also presents a brief description of the simulation environment along with empirical results. Our experiments indicate that performance improvements can be obtained using the proposed approach.

show abstract

Switching to High Gear: Opportunities for Grand-Scale Real-Time Parallel Simulations

Perumalla

2009

2009 13th IEEE/ACM International Symposium on Distributed Simulation and Real Time Applications

View full text Add to dashboard Cite

Simulating Spatially Explicit Problems on High Performance Architectures

Cited by 5 publications

References 49 publications

Explicit Spatial Scattering for Load Balancing in Conservatively Synchronized Parallel Discrete Event Simulations

Explicit Spatial Scattering for Load Balancing in Conservatively Synchronized Parallel Discrete Event Simulations

Accelerating Spatially Explicit Simulations of Spread of Lyme Disease

Switching to High Gear: Opportunities for Grand-Scale Real-Time Parallel Simulations

Contact Info

Product

Resources

About