An experimental study of a parallel shortest path algorithm for solving large-scale graph instances

Kamesh Madduri, David A. Bader, W. Berry Jonathan, Joseph R. Crobak

Research output: Chapter in Book/Report/Conference proceedingConference contribution

38 Citations (Scopus)

Abstract

We present an experimental study of the single source shortest path problem with non-negative edge weights (NSSP) on large-scale graphs using the △-stepping parallel algorithm, We report performance results on the Cray MTA-2, a multithreaded parallel computer. The MTA-2 is a high-end shared memory system offering two unique features that aid the efficient parallel implementation of irregular algorithms: the ability to exploit fine-grained parallelism, and low-overhead synchronization primitives. Our implementation exhibits remarkable parallel speedup when compared with competitive sequential algorithms, for low-diameter sparse graphs. For instance, △-stepping on a directed scalefree graph of 100 million vertices and 1 billion edges takes less than ten seconds on 40 processors of the MTA-2, with a relative speedup of close to 30. To our knowledge, these are the first performance results of a shortest path problem on realistic graph instances in the order of billions of vertices and edges.

Original languageEnglish (US)
Title of host publicationProceedings of the 9th Workshop on Algorithm Engineering and Experiments and the 4th Workshop on Analytic Algorithms and Combinatorics
Pages23-35
Number of pages13
StatePublished - Aug 22 2007
Event9th Workshop on Algorithm Engineering and Experiments and the 4th Workshop on Analytic Algorithms and Combinatorics - New Orleans, LA, United States
Duration: Jan 6 2007Jan 6 2007

Publication series

NameProceedings of the 9th Workshop on Algorithm Engineering and Experiments and the 4th Workshop on Analytic Algorithms and Combinatorics

Other

Other9th Workshop on Algorithm Engineering and Experiments and the 4th Workshop on Analytic Algorithms and Combinatorics
CountryUnited States
CityNew Orleans, LA
Period1/6/071/6/07

Fingerprint

Shortest Path Algorithm
Parallel Algorithms
Experimental Study
Shortest Path Problem
Directed graphs
Speedup
Graph in graph theory
Parallel algorithms
Synchronization
Sparse Graphs
Sequential Algorithm
Parallel Computers
Parallel Implementation
Shared Memory
Efficient Implementation
Data storage equipment
Parallelism
Irregular
Non-negative

All Science Journal Classification (ASJC) codes

  • Engineering(all)
  • Applied Mathematics

Cite this

Madduri, K., Bader, D. A., Jonathan, W. B., & Crobak, J. R. (2007). An experimental study of a parallel shortest path algorithm for solving large-scale graph instances. In Proceedings of the 9th Workshop on Algorithm Engineering and Experiments and the 4th Workshop on Analytic Algorithms and Combinatorics (pp. 23-35). (Proceedings of the 9th Workshop on Algorithm Engineering and Experiments and the 4th Workshop on Analytic Algorithms and Combinatorics).
Madduri, Kamesh ; Bader, David A. ; Jonathan, W. Berry ; Crobak, Joseph R. / An experimental study of a parallel shortest path algorithm for solving large-scale graph instances. Proceedings of the 9th Workshop on Algorithm Engineering and Experiments and the 4th Workshop on Analytic Algorithms and Combinatorics. 2007. pp. 23-35 (Proceedings of the 9th Workshop on Algorithm Engineering and Experiments and the 4th Workshop on Analytic Algorithms and Combinatorics).
@inproceedings{3bddbeab896a4878b904ee9d2ba36f80,
title = "An experimental study of a parallel shortest path algorithm for solving large-scale graph instances",
abstract = "We present an experimental study of the single source shortest path problem with non-negative edge weights (NSSP) on large-scale graphs using the △-stepping parallel algorithm, We report performance results on the Cray MTA-2, a multithreaded parallel computer. The MTA-2 is a high-end shared memory system offering two unique features that aid the efficient parallel implementation of irregular algorithms: the ability to exploit fine-grained parallelism, and low-overhead synchronization primitives. Our implementation exhibits remarkable parallel speedup when compared with competitive sequential algorithms, for low-diameter sparse graphs. For instance, △-stepping on a directed scalefree graph of 100 million vertices and 1 billion edges takes less than ten seconds on 40 processors of the MTA-2, with a relative speedup of close to 30. To our knowledge, these are the first performance results of a shortest path problem on realistic graph instances in the order of billions of vertices and edges.",
author = "Kamesh Madduri and Bader, {David A.} and Jonathan, {W. Berry} and Crobak, {Joseph R.}",
year = "2007",
month = "8",
day = "22",
language = "English (US)",
isbn = "0898716284",
series = "Proceedings of the 9th Workshop on Algorithm Engineering and Experiments and the 4th Workshop on Analytic Algorithms and Combinatorics",
pages = "23--35",
booktitle = "Proceedings of the 9th Workshop on Algorithm Engineering and Experiments and the 4th Workshop on Analytic Algorithms and Combinatorics",

}

Madduri, K, Bader, DA, Jonathan, WB & Crobak, JR 2007, An experimental study of a parallel shortest path algorithm for solving large-scale graph instances. in Proceedings of the 9th Workshop on Algorithm Engineering and Experiments and the 4th Workshop on Analytic Algorithms and Combinatorics. Proceedings of the 9th Workshop on Algorithm Engineering and Experiments and the 4th Workshop on Analytic Algorithms and Combinatorics, pp. 23-35, 9th Workshop on Algorithm Engineering and Experiments and the 4th Workshop on Analytic Algorithms and Combinatorics, New Orleans, LA, United States, 1/6/07.

An experimental study of a parallel shortest path algorithm for solving large-scale graph instances. / Madduri, Kamesh; Bader, David A.; Jonathan, W. Berry; Crobak, Joseph R.

Proceedings of the 9th Workshop on Algorithm Engineering and Experiments and the 4th Workshop on Analytic Algorithms and Combinatorics. 2007. p. 23-35 (Proceedings of the 9th Workshop on Algorithm Engineering and Experiments and the 4th Workshop on Analytic Algorithms and Combinatorics).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - An experimental study of a parallel shortest path algorithm for solving large-scale graph instances

AU - Madduri, Kamesh

AU - Bader, David A.

AU - Jonathan, W. Berry

AU - Crobak, Joseph R.

PY - 2007/8/22

Y1 - 2007/8/22

N2 - We present an experimental study of the single source shortest path problem with non-negative edge weights (NSSP) on large-scale graphs using the △-stepping parallel algorithm, We report performance results on the Cray MTA-2, a multithreaded parallel computer. The MTA-2 is a high-end shared memory system offering two unique features that aid the efficient parallel implementation of irregular algorithms: the ability to exploit fine-grained parallelism, and low-overhead synchronization primitives. Our implementation exhibits remarkable parallel speedup when compared with competitive sequential algorithms, for low-diameter sparse graphs. For instance, △-stepping on a directed scalefree graph of 100 million vertices and 1 billion edges takes less than ten seconds on 40 processors of the MTA-2, with a relative speedup of close to 30. To our knowledge, these are the first performance results of a shortest path problem on realistic graph instances in the order of billions of vertices and edges.

AB - We present an experimental study of the single source shortest path problem with non-negative edge weights (NSSP) on large-scale graphs using the △-stepping parallel algorithm, We report performance results on the Cray MTA-2, a multithreaded parallel computer. The MTA-2 is a high-end shared memory system offering two unique features that aid the efficient parallel implementation of irregular algorithms: the ability to exploit fine-grained parallelism, and low-overhead synchronization primitives. Our implementation exhibits remarkable parallel speedup when compared with competitive sequential algorithms, for low-diameter sparse graphs. For instance, △-stepping on a directed scalefree graph of 100 million vertices and 1 billion edges takes less than ten seconds on 40 processors of the MTA-2, with a relative speedup of close to 30. To our knowledge, these are the first performance results of a shortest path problem on realistic graph instances in the order of billions of vertices and edges.

UR - http://www.scopus.com/inward/record.url?scp=34547953707&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=34547953707&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:34547953707

SN - 0898716284

SN - 9780898716283

T3 - Proceedings of the 9th Workshop on Algorithm Engineering and Experiments and the 4th Workshop on Analytic Algorithms and Combinatorics

SP - 23

EP - 35

BT - Proceedings of the 9th Workshop on Algorithm Engineering and Experiments and the 4th Workshop on Analytic Algorithms and Combinatorics

ER -

Madduri K, Bader DA, Jonathan WB, Crobak JR. An experimental study of a parallel shortest path algorithm for solving large-scale graph instances. In Proceedings of the 9th Workshop on Algorithm Engineering and Experiments and the 4th Workshop on Analytic Algorithms and Combinatorics. 2007. p. 23-35. (Proceedings of the 9th Workshop on Algorithm Engineering and Experiments and the 4th Workshop on Analytic Algorithms and Combinatorics).