Adaptive dynamic load-balancing with irregular domain decomposition for particle simulations

Begau, Christoph; Sutmann, Godehard

doi:10.1016/j.cpc.2015.01.009

Cited by 26 publications

(16 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Researchers have proposed a number of methods for dynamic load balance. [3][4][5] In these works, static or dynamic load balancing algorithms are implemented, employing centralized or decentralized control. The key point of load balancing is how to control the load distribution adaptively when system parameters and workload characteristics change.…”

Section: Discussionmentioning

confidence: 99%

“…In recent years, remarkable achievements had been achieved in load balancing research, not only in practice but also in theory, such as load balancing strategy, implementation of the elements, and modeling of balance control. Researchers have proposed a number of methods for dynamic load balance . In these works, static or dynamic load balancing algorithms are implemented, employing centralized or decentralized control.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A self‐adaptive load‐dispatching control framework for device data accessing in IoT‐based systems

Wang

Yin

et al. 2017

Int J Communication

View full text Add to dashboard Cite

Summary The Internet of things (IoT) information system plays important roles in disposing of huge volumes of real‐time service requests from heterogeneous devices, targeting for different complex application requirements. Load‐dispatching control (LDC) is a key problem to be solved for devices accessing concurrently in cluster systems. Self‐adaptive LDC optimizes the resource allocation to ensure no overloading node, thus, improving the performance of IoT systems. This paper focuses on adaptive dispatching control problem in IoT information system. First, a device data access platform is proposed for reducing the load imbalance and improving the efficiency of data processing. Then, we propose a processing capability prediction model to evaluate the system performance. On the basis of the model, we present a practical self‐adaptive LDC framework with a self‐adaptive control strategy and a load dispatching method. Finally, a case study is given to verify the framework and the control strategy. Experimental results show that the proposed strategy can meet the requirements of dynamic load balancing with the ability to avoid the load imbalance problem, and the LDC‐based device access platform can process data accessing effectively and ubiquitously.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

A self‐adaptive load‐dispatching control framework for device data accessing in IoT‐based systems

Wang

Yin

et al. 2017

Int J Communication

View full text Add to dashboard Cite

show abstract

“…The third, more complex type of decomposition is the generation of an auxiliary grid in the computational domain, the cells of which, together with the particles in cells, are distributed among the computational processes [13,14]. Load balancing is possible throughout the mobility of grid nodes, the connectivity of which is usually assumed to be fixed.…”

Section: Introductionmentioning

confidence: 99%

Parallel SPH modeling using dynamic domain decomposition and load balancing displacement of Voronoi subdomains

Егорова

Dyachkov

Паршиков

et al. 2019

Computer Physics Communications

View full text Add to dashboard Cite

A highly adaptive load balancing algorithm for parallel simulations using particle methods, such as molecular dynamics and smoothed particle hydrodynamics (SPH), is developed. Our algorithm is based on the dynamic spatial decomposition of simulated material samples between Voronoi subdomains, where each subdomain with all its particles is handled by a single computational process which is typically run on a single CPU core of a multiprocessor computing cluster.The algorithm displaces the positions of neighbor Voronoi subdomains in accordance with the local load imbalance between the corresponding processes. It results in particle transfers from heavy-loaded processes to less-loaded ones. Iteration of the algorithm puts into alignment the processor loads. Convergence to a well-balanced decomposition from imbalanced one is improved by the usage of multi-body terms in the balancing displacements.The high adaptability of the balancing algorithm to simulation conditions is illustrated by SPH modeling of the dynamic behavior of materials under extreme conditions, which are characterized by large pressure and velocity gradients, as a result of which the spatial distribution of particles varies greatly in time. The higher parallel efficiency of our algorithm in such conditions is demonstrated by comparison with the corresponding static decomposition of the computational domain. Our algorithm shows almost perfect strong scalability in tests using from tens to several thousand processes.

show abstract

“…The distributed manager-worker algorithm implements two types of load balancing: (1) adaptive algorithm with non-orthogonal domains, and (2) atom decomposition procedure. A general illustrative example of both routines can be found inFigure 5.On the one hand, the adaptive load balancing algorithm[25] is integrated in the functionality of the framework that gives support to the domain decomposition of the system[24]. Starting from an orthogonal definition of domains, this algorithm moves the vertices of each domain according to a driving force, which is calculated using a system-dependent workload function.…”

mentioning

confidence: 99%

Parallelization comparison and optimization of a scale-bridging framework to model Cottrell atmospheres

Ganesan

Teijeiro

Sutmann

2018

Computational Materials Science

View full text Add to dashboard Cite

Low carbon steels undergo strain aging when heat treated, which causes an increased yield strength that can be observed macroscopically. Such strengthening mechanism is driven by atomistic scale processes, i.e., solute segregation of carbon (C) or nitrogen interstitial atoms. Due to its low solubility, alloying elements can diffuse to defects (e.g., dislocations) and form the so-called Cottrell atmospheres. Consequently, the mobility of defects is strongly reduced because of the interaction with solutes, and higher stresses are needed to unpin them from

show abstract

Adaptive dynamic load-balancing with irregular domain decomposition for particle simulations

Cited by 26 publications

References 30 publications

A self‐adaptive load‐dispatching control framework for device data accessing in IoT‐based systems

A self‐adaptive load‐dispatching control framework for device data accessing in IoT‐based systems

Parallel SPH modeling using dynamic domain decomposition and load balancing displacement of Voronoi subdomains

Parallelization comparison and optimization of a scale-bridging framework to model Cottrell atmospheres

Contact Info

Product

Resources

About