A Simple Yet Effective Balanced Edge Partition Model for Parallel Computing

Li, Lingda; Geda, Robel; Hayes, Ari B.; Chen, Yanhao; Chaudhari, Pranav; Zhang, Eddy Z.; Szegedy, Márió

doi:10.1145/3078505.3078520

Cited by 17 publications

(23 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Previously studied vertex reordering algorithms are computationally more complex. The algorithm presented by Li et al [7] has polynomial time complexity in |V |. Gorder [10] takes O( v∈V (deg out (v)) 2 ) steps where…”

Section: E Time Complexitymentioning

confidence: 99%

“…Gonzalez et al proposed vertex cut, a parallel streaming partitioning algorithm that minimizes vertex replication [1]. Li et al [7] and Bourse et al [6] proposed efficient edge-balanced partitioning methods. Bourse et al [6] moreover investigate the interplay between edge balance and vertex balance, which is non-trivial if edge cuts are simultaneously minimized.…”

Section: Related Workmentioning

confidence: 99%

“…Moreover, the partitions should be largely independent to minimize the volume of data communication. It has been demonstrated that partitioning the edge set is more effective than partitioning the vertex set [1], leading to the commonly used heuristic to balance edge counts and minimize vertex replication [1], [6], [7], [2]. These constraints are typically mutually incompatible for scale-free graphs, resulting in a compromise between edge balance and vertex replication [6].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

VEBO: A Vertex- and Edge-Balanced Ordering Heuristic to Load Balance Parallel Graph Processing

Sun,

Vandierendonck,

Nikolopoulos

2018

Preprint

View full text Add to dashboard Cite

Graph partitioning drives graph processing in distributed, disk-based and NUMA-aware systems. A commonly used partitioning goal is to balance the number of edges per partition in conjunction with minimizing the edge or vertex cut. While this type of partitioning is computationally expensive, we observe that such topology-driven partitioning nonetheless results in computational load imbalance.We propose Vertex-and Edge-Balanced Ordering (VEBO): balance the number of edges and the number of unique destinations of those edges. VEBO optimally balances edges and vertices for graphs with a powerlaw degree distribution. Experimental evaluation on three shared-memory graph processing systems (Ligra, Polymer and GraphGrind) shows that VEBO achieves excellent load balance and improves performance by 1.09x over Ligra, 1.41x over Polymer and 1.65x over GraphGrind, compared to their respective partitioning algorithms, averaged across 8 algorithms and 7 graphs.

show abstract

Section: E Time Complexitymentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

VEBO: A Vertex- and Edge-Balanced Ordering Heuristic to Load Balance Parallel Graph Processing

Sun,

Vandierendonck,

Nikolopoulos

2018

Preprint

View full text Add to dashboard Cite

show abstract

“…This effect could be explained by the choice of k-the number of blocks used for partitioning. While we use standard values for (node-based) graph partitioning benchmarks [49], Li et al [32] choose k such that each block contains approximately 10 240 edges. Thus some instances are partitioned into up to 1 692 and 5 952 blocks, which might be too large for current partitioning tools.…”

Section: Scircuit_spmv Cant_spmv and Mc2depi_spmvmentioning

confidence: 99%

“…Similar to (node-based) graph partitioning, the quality of the edge partitioning can have a dramatic effect on parallelization [32]. Noting that edge partitioning can be solved directly with hypergraph partitioners, such as hMETIS [26,27] and PaToH [9], Li et al [32] showed that these techniques give the highest quality partitionings; however, they are also slow. Therefore, a balance of solution quality and speed must be taken into consideration.…”

Section: Introductionmentioning

confidence: 99%

Scalable Edge Partitioning

Schlag¹,

Schulz²,

Seemaier³

et al. 2019

2019 Proceedings of the Twenty-First Workshop on Algorithm Engineering and Experiments (ALENEX)

View full text Add to dashboard Cite

Edge-centric distributed computations have appeared as a recent technique to improve the shortcomings of thinklike-a-vertex algorithms on large scale-free networks. In order to increase parallelism on this model, edge partitioning-partitioning edges into roughly equally sized blocks-has emerged as an alternative to traditional (node-based) graph partitioning. In this work, we give a distributed memory parallel algorithm to compute highquality edge partitions in a scalable way. Our algorithm scales to networks with billions of edges, and runs efficiently on thousands of PEs. Our technique is based on a fast parallelization of split graph construction and a use of advanced node partitioning algorithms. Our extensive experiments show that our algorithm has high quality on large real-world networks and large hyperbolic random graphs-which have a power law degree distribution and are therefore specifically targeted by edge partitioning.

show abstract

Graph Partition

Schulz¹,

Strash²

2018

Encyclopedia of Big Data Technologies

View full text Add to dashboard Cite

A Simple Yet Effective Balanced Edge Partition Model for Parallel Computing

Cited by 17 publications

References 19 publications

VEBO: A Vertex- and Edge-Balanced Ordering Heuristic to Load Balance Parallel Graph Processing

VEBO: A Vertex- and Edge-Balanced Ordering Heuristic to Load Balance Parallel Graph Processing

Scalable Edge Partitioning

Graph Partition

Contact Info

Product

Resources

About