Parametric multi-step scheme for GPU-accelerated graph
decomposition into strongly connected components

D 2017

Parametric multi-step scheme for GPU-accelerated graph decomposition into strongly connected components

ALDEGHERI, Stefano, Jiří BARNAT, Nicola BOMBIERI, Federico BUSATO, Milan ČEŠKA et. al.

Basic information

Original name

Parametric multi-step scheme for GPU-accelerated graph decomposition into strongly connected components

Authors

ALDEGHERI, Stefano (380 Italy), Jiří BARNAT (203 Czech Republic, guarantor, belonging to the institution), Nicola BOMBIERI (380 Italy), Federico BUSATO (380 Italy) and Milan ČEŠKA (203 Czech Republic)

Edition

Cham, 22nd International Conference on Parallel and Distributed Computing, Euro-Par 2016, p. 519-531, 13 pp. 2017

Publisher

Springer Verlag

Other information

Language

English

Type of outcome

Stať ve sborníku

Field of Study

10201 Computer sciences, information science, bioinformatics

Country of publisher

Switzerland

Confidentiality degree

není předmětem státního či obchodního tajemství

Publication form

printed version "print"

Impact factor

Impact factor: 0.402 in 2005

RIV identification code

RIV/00216224:14330/17:00100646

Organization unit

Faculty of Informatics

ISBN

978-3-319-58942-8

ISSN

DOI

http://dx.doi.org/10.1007/978-3-319-58943-5_42

UT WoS

000529303100042

Keywords in English

Directed graphs; Distributed computer systems; Problem solving

Abstract

V originále

The problem of decomposing a directed graph into strongly connected components (SCCs) is a fundamental graph problem that is inherently present in many scientific and commercial applications. Clearly, there is a strong need for good high-performance, e.g., GPU-accelerated, algorithms to solve it. Unfortunately, among existing GPU-enabled algorithms to solve the problem, there is none that can be considered the best on every graph, disregarding the graph characteristics. Indeed, the choice of the right and most appropriate algorithm to be used is often left to inexperienced users. In this paper, we introduce a novel parametric multi-step scheme to evaluate existing GPU-accelerated algorithms for SCC decomposition in order to alleviate the burden of the choice and to help the user to identify which combination of existing techniques for SCC decomposition would fit an expected use case the most. We support our scheme with an extensive experimental evaluation that dissects correlations between the internal structure of GPU-based algorithms and their performance on various classes of graphs. The measurements confirm that there is no algorithm that would beat all other algorithms in the decomposition on all of the classes of graphs. Our contribution thus represents an important step towards an ultimate solution of automatically adjusted scheme for the GPU-accelerated SCC decomposition.

Citovat

ALDEGHERI, Stefano, Jiří BARNAT, Nicola BOMBIERI, Federico BUSATO and Milan ČEŠKA. Parametric multi-step scheme for GPU-accelerated graph decomposition into strongly connected components. In 22nd International Conference on Parallel and Distributed Computing, Euro-Par 2016. Cham: Springer Verlag, 2017, p. 519-531. ISBN 978-3-319-58942-8. Available from: https://dx.doi.org/10.1007/978-3-319-58943-5_42.

@inproceedings{1417426,
   author = {Aldegheri, Stefano and Barnat, Jiří and Bombieri, Nicola and Busato, Federico and Češka, Milan},
   address = {Cham},
   booktitle = {22nd International Conference on Parallel and Distributed Computing, Euro-Par 2016},
   doi = {http://dx.doi.org/10.1007/978-3-319-58943-5_42},
   keywords = {Directed graphs; Distributed computer systems; Problem solving},
   howpublished = {tištěná verze "print"},
   language = {eng},
   location = {Cham},
   isbn = {978-3-319-58942-8},
   pages = {519-531},
   publisher = {Springer Verlag},
   title = {Parametric multi-step scheme for GPU-accelerated graph decomposition into strongly connected components},
   year = {2017}
}

TY  - JOUR
ID  - 1417426
AU  - Aldegheri, Stefano - Barnat, Jiří - Bombieri, Nicola - Busato, Federico - Češka, Milan
PY  - 2017
TI  - Parametric multi-step scheme for GPU-accelerated graph decomposition into strongly connected components
PB  - Springer Verlag
CY  - Cham
SN  - 9783319589428
KW  - Directed graphs
KW  - Distributed computer systems
KW  - Problem solving
N2  - The problem of decomposing a directed graph into strongly connected components (SCCs) is a fundamental graph problem that is inherently present in many scientific and commercial applications. Clearly, there is a strong need for good high-performance, e.g., GPU-accelerated, algorithms to solve it. Unfortunately, among existing GPU-enabled algorithms to solve the problem, there is none that can be considered the best on every graph, disregarding the graph characteristics. Indeed, the choice of the right and most appropriate algorithm to be used is often left to inexperienced users. In this paper, we introduce a novel parametric multi-step scheme to evaluate existing GPU-accelerated algorithms for SCC decomposition in order to alleviate the burden of the choice and to help the user to identify which combination of existing techniques for SCC decomposition would fit an expected use case the most. We support our scheme with an extensive experimental evaluation that dissects correlations between the internal structure of GPU-based algorithms and their performance on various classes of graphs. The measurements confirm that there is no algorithm that would beat all other algorithms in the decomposition on all of the classes of graphs. Our contribution thus represents an important step towards an ultimate solution of automatically adjusted scheme for the GPU-accelerated SCC decomposition.
ER  -

ALDEGHERI, Stefano, Jiří BARNAT, Nicola BOMBIERI, Federico BUSATO and Milan ČEŠKA. Parametric multi-step scheme for GPU-accelerated graph decomposition into strongly connected components. In \textit{22nd International Conference on Parallel and Distributed Computing, Euro-Par 2016}. Cham: Springer Verlag, 2017, p.~519-531. ISBN~978-3-319-58942-8. Available from: https://dx.doi.org/10.1007/978-3-319-58943-5\_{}42.

Detailed Information on Publication Record