Confidence-Based Work Stealing in Parallel Constraint Programming

Chu, Geoffrey; Schulte, Christian; Stuckey, Peter J.

doi:10.1007/978-3-642-04244-7_20

Cited by 47 publications

(46 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Research on parallelization for CSP solvers includes a broad spectrum of parallel programming models (e.g., OpenMP [5][6][7], Message Passing Interface (MPI ) [8,9]) and a variety of platforms (e.g., single node [5,7], cluster [9,10], and grid [11]), on a scale from a few processors to thousands. In particular, MPI is intended for high-performance parallel computing on platforms without shared memory.…”

Section: Background and Related Workmentioning

confidence: 99%

“…A boolean flag δ i indicates whether a node is closed (both values attempted, δ i = 0, black circle) or open (one value attempted, δ i = 1, white circle) [8]. Although identification of a helpful guiding path is non-trivial (as in [5]), variables with particular properties have proved effective for splitting [7]. Iterative partitioning with clause learning, where search spaces of SAT subproblems may overlap, can also be an effective strategy [11].…”

Section: Background and Related Workmentioning

confidence: 99%

“…Additional parallelization methods include various workload-balancing mechanisms, such as work stealing [5,6,10] and work sharing [9]. The SAT-solver parallelization methods most relevant to Spread are described in [7] and [11].…”

Section: Background and Related Workmentioning

confidence: 99%

See 2 more Smart Citations

A Hybrid Paradigm for Adaptive Parallel Search

Epstein

2012

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. Parallelization offers the opportunity to accelerate search on constraint satisfaction problems. To parallelize a sequential solver under a popular message passing protocol, the new paradigm described here combines portfolio-based methods and search space splitting. To split effectively and to balance processor workload, this paradigm adaptively exploits knowledge acquired during search and allocates additional resources to the most difficult parts of a problem. Extensive experiments in a parallel environment show that this paradigm significantly improves the performance of an underlying sequential solver, outperforms more naive approaches to parallelization, and solves many difficult problems left open after recent solver competitions.

show abstract

Section: Background and Related Workmentioning

confidence: 99%

Section: Background and Related Workmentioning

confidence: 99%

See 1 more Smart Citation

A Hybrid Paradigm for Adaptive Parallel Search

Epstein

2012

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…A different approach is based on work stealing [2,3,4]. The workload is initially distributed to the available workers.…”

Section: Introductionmentioning

confidence: 99%

Self-splitting of Workload in Parallel Computation

Fischetti¹,

Monaci²,

Salvagnin³

2014

Integration of AI and OR Techniques in Constraint Programming

View full text Add to dashboard Cite

Abstract. Parallel computation requires splitting a job among a set of processing units called workers. The computation is generally performed by a set of one or more master workers that split the workload into chunks and distribute them to a set of slave workers. In this setting, communication among workers can be problematic and/or time consuming. Tree search algorithms are particularly suited for being applied in a parallel fashion, as different nodes can be processed by different workers in parallel. In this paper we propose a simple mechanism to convert a sequential tree-search code into a parallel one. In the new paradigm, called SelfSplit, each worker is able to autonomously determine, without any communication with the other workers, the job parts it has to process. Computational results are reported, showing that SelfSplit can achieve an almost linear speedup for hard Constraint Programming applications, even when 64 workers are considered.

show abstract

“…In our experimentation, the winner of the parallel category of the 2013 SAT Competition also achieved a speedup of only about 3 on 32 cores. Constraint programming search appears to be more suitable for parallelization than search for MIP or SAT: different strategies, including a recursive application of search goals [24], work stealing [14], problem decomposition [25], and a dedicated parallel scheme based on limited discrepancy search [23] all exhibit good speedups (sometimes near-linear) of the CP search in certain settings, especially those involving infeasible instances or scenarios where evaluating search tree leaves is costlier than evaluating internal nodes. Yet, recent developments in CP have moved towards more constraint learning during search, for which efficient parallelization becomes increasingly more difficult.…”

Section: Introductionmentioning

confidence: 99%

Parallel Combinatorial Optimization with Decision Diagrams

Bergman

Ciré

Sabharwal

et al. 2014

Integration of AI and OR Techniques in Constraint Programming

View full text Add to dashboard Cite

Abstract. We propose a new approach for parallelizing search for combinatorial optimization that is based on a recursive application of approximate Decision Diagrams. This generic scheme can, in principle, be applied to any combinatorial optimization problem for which a decision diagram representation is available. We consider the maximum independent set problem as a specific case study, and show how a recently proposed sequential branch-and-bound scheme based on approximate decision diagrams can be parallelized efficiently using the X10 parallel programming and execution framework. Experimental results using our parallel solver, DDX10, running on up to 256 compute cores spread across a cluster of machines indicate that parallel decision diagrams scale effectively and consistently. Moreover, on graphs of relatively high density, parallel decision diagrams often outperform state-of-the-art parallel integer programming when both use a single 32-core machine.

show abstract

Confidence-Based Work Stealing in Parallel Constraint Programming

Cited by 47 publications

References 10 publications

A Hybrid Paradigm for Adaptive Parallel Search

A Hybrid Paradigm for Adaptive Parallel Search

Self-splitting of Workload in Parallel Computation

Parallel Combinatorial Optimization with Decision Diagrams

Contact Info

Product

Resources

About