Zoltan data management services for parallel dynamic applications

Devine, Karen Dragon; Boman, Erik G.; Heaphy, Robert; Hendrickson, Bruce; Vaughan, Courtenay T.

doi:10.1109/5992.988653

Cited by 232 publications

(157 citation statements)

References 21 publications

Supporting

Mentioning

147

Contrasting

Unclassified

Order By: Relevance

“…Also, Sandia National Labs' Zoltan toolkit [25] contains a recently developed parallel hypergraph partitioner [24]. This partitioner is based on the multi-level RB paradigm and currently supports the connectivity-1 cost metric.…”

Section: Hypergraph Partitioning Toolsmentioning

confidence: 99%

Multi-level direct K-way hypergraph partitioning with multiple constraints and fixed vertices

Aykanat

Cambazoğlu

Uçar

2008

Journal of Parallel and Distributed Computing

100

View full text Add to dashboard Cite

K-way hypergraph partitioning has an ever-growing use in parallelization of scientific computing applications. We claim that hypergraph partitioning with multiple constraints and fixed vertices should be implemented using direct K-way refinement, instead of the widely adopted recursive bisection paradigm. Our arguments are based on the fact that recursive-bisection-based partitioning algorithms perform considerably worse when used in the multiple constraint and fixed vertex formulations. We discuss possible reasons for this performance degradation. We describe a careful implementation of a multi-level direct K-way hypergraph partitioning algorithm, which performs better than a well-known recursive-bisection-based partitioning algorithm in hypergraph partitioning with multiple constraints and fixed vertices. We also experimentally show that the proposed algorithm is effective in standard hypergraph partitioning.

show abstract

Section: Hypergraph Partitioning Toolsmentioning

confidence: 99%

Multi-level direct K-way hypergraph partitioning with multiple constraints and fixed vertices

Aykanat

Cambazoğlu

Uçar

2008

Journal of Parallel and Distributed Computing

100

View full text Add to dashboard Cite

show abstract

“…Each algorithm has characteristics and requirements that make it appropriate for certain applications; see [4,35] for examples and [33] for an overview of available methods. The Zoltan Parallel Data Services Toolkit [9,11] provides dynamic load balancing and related capabilities to a wide range of dynamic, unstructured and/or adaptive applications. Using Zoltan, application developers can switch partitioners simply by changing a run-time parameter, facilitating comparisons of the partitioners' effect on the applications.…”

Section: Partitioning and Dynamic Load Balancingmentioning

confidence: 99%

Hierarchical Partitioning and Dynamic Load Balancing for Scientific Computation

Teresco

Faik

Flaherty

2006

Applied Parallel Computing. State of the Art in Scientific Computing

View full text Add to dashboard Cite

Abstract. Cluster and grid computing has made hierarchical and heterogeneous computing systems increasingly common as target environments for large-scale scientific computation. A cluster may consist of a network of multiprocessors. A grid computation may involve communication across slow interfaces. Modern supercomputers are often large clusters with hierarchical network structures. For maximum efficiency, software must adapt to the computing environment. We focus on partitioning and dynamic load balancing, in particular on hierarchical procedures implemented within the Zoltan Toolkit, guided by DRUM, the Dynamic Resource Utilization Model. Here, different balancing procedures are used in different parts of the domain. Preliminary results show that hierarchical partitionings are competitive with the best traditional methods on a small hierarchical cluster.Modern three-dimensional scientific computations must execute in parallel to achieve acceptable performance. Target parallel environments range from clusters of workstations to the largest tightly-coupled supercomputers. Hierarchical and heterogeneous systems are increasingly common as symmetric multiprocessing (SMP) nodes are combined to form the relatively small clusters found in many institutions as well as many of today's most powerful supercomputers. Network hierarchies arise as grid technologies make Internet execution more likely and modern supercomputers are built using hierarchical interconnection networks. MPI implementations may exhibit very different performance characteristics depending on the underlying network and message passing implementation (e.g., [32]). Software efficiency may be improved using optimizations based on system characteristics and domain knowledge. Some have accounted for clusters of SMPs by using a hybrid programming model, with message passing for internode communication and multithreading for intra-node communication (e.g., [1,27]), with varying degress of success, but always with an increased burden on programmers to program both levels of parallelization.Our focus has been on resource-aware partitioning and dynamic load balancing, achieved by adjusting target partition sizes or the choice of a dynamic

show abstract

“…Zoltan [25,9] is a library that contains tools for load balancing and parallel data management. It provides a common interface to graph partitioners like METIS and ParMetis [15,14].…”

Section: Data Distributionmentioning

confidence: 99%

On a parallel multilevel preconditioned Maxwell eigensolver

et al. 2006

View full text Add to dashboard Cite

We report on a parallel implementation of the Jacobi-Davidson algorithm to compute a few eigenvalues and corresponding eigenvectors of a large real symmetric generalized matrix eigenvalue problemThe eigenvalue problem stems from the design of cavities of particle accelerators. It is obtained by the finite element discretization of the time-harmonic Maxwell equation in weak form by a combination of Nédélec (edge) and Lagrange (node) elements.We found the Jacobi-Davidson (JD) method to be a very effective solver provided that a good preconditioner is available for the correction equations that have to be solved in each step of the JD iteration. The preconditioner of our choice is a combination of a hierarchical basis preconditioner and the ML smoothed aggregation AMG preconditioner. It is close to optimal regarding iteration count.The parallel code makes extensive use of the Trilinos software framework. In our examples from accelerator physics we observe satisfactory speedups and efficiencies.

show abstract

Zoltan data management services for parallel dynamic applications

Cited by 232 publications

References 21 publications

Multi-level direct K-way hypergraph partitioning with multiple constraints and fixed vertices

Multi-level direct K-way hypergraph partitioning with multiple constraints and fixed vertices

Hierarchical Partitioning and Dynamic Load Balancing for Scientific Computation

On a parallel multilevel preconditioned Maxwell eigensolver

Contact Info

Product

Resources

About