research-article

Fast block distributed CUDA implementation of the Hungarian algorithm

Authors:

Paulo A.C. Lopes,

Satyendra Singh Yadav,

Aleksandar Ilic,

Sarat Kumar PatraAuthors Info & Claims

Volume 130, Issue C

Pages 50 - 62

https://doi.org/10.1016/j.jpdc.2019.03.014

Published: 01 August 2019 Publication History

Abstract

The Hungarian algorithm solves the linear assignment problem in polynomial time. A GPU/CUDA implementation of this algorithm is proposed. GPUs are massive parallel machines. In this implementation, the alternating path search phase of the algorithm is distributed by several blocks in a way to minimize global device synchronization. This phase is very important and has a big contribution to the execution time. Other advanced features also implemented are: parallel graph traversal; the parallel detection of multiple alternating paths in a single iteration; a simplified and fast matrix compression that stores the zeros of the slack matrix, resulting in very fast graph traversal; highly optimized reductions for the initial slack matrix calculation and update. This results in a fast implementation for moderate size problems.

Highlights

•

Block Distributed Alternating Path Search with nontrivial proof.

•

Multiple Simultaneous Alternating Paths.

•

Graph Search is directly based on the original Hungarian algorithm resulting in efficient implementation.

•

Optimized reductions for initial dual variables and dual variables update.

•

Chosen O ( n 4 ) Implementation shows better performance than O ( n 3 ) in low range matrices.

References

[1]

Munkres J., Algorithms for the assignment and transportation problems, J. Soc. Ind. Appl. Math. 5 (1) (1957) 32–38.

[2]

Bertsekas D.P., The auction algorithm for assignment and other network flow problems: A tutorial, Interfaces 20 (4) (1990) 133–149.

[3]

Jonker R., Volgenant A., A shortest augmenting path algorithm for dense and sparse linear assignment problems, Computing 38 (4) (1987) 325–340.

Digital Library

[4]

Lawler E.L., Combinatorial Optimization: Networks and Matroids, Dover Books, 1976.

Digital Library

[5]

Date K., Nagi R., GPU-accelerated Hungarian Algorithms for the linear assignment problem, Parallel Comput. 57 (2016) 52–72.

[6]

Balas E., Miller D., Pekny J., Toth P., A parallel shortest augmenting path algorithm for the assignment problem, J. ACM 38 (4) (1991) 985–1004.

[7]

Bertsekas D.P., Castanon D.A., Parallel asynchronous Hungarian methods for the assignment problem, ORSA J. Comput. 5 (3) (1993) 261–274.

[8]

Bertsekas D.P., Castañon D.A., Parallel synchronous and asynchronous implementations of the auction algorithm, Parallel Comput. 17 (6) (1991) 707–732.

[9]

Sathe M., Schenk O., Burkhart H., An auction-based weighted matching implementation on massively parallel architectures, Parallel Comput. 38 (12) (2012) 595–614.

[10]

Zavlanos M.M., Spesivtsev L., Pappas G.J., A distributed auction algorithm for the assignment problem, in: Decision and Control, 2008. CDC 2008. 47th IEEE Conference on, IEEE, 2008, pp. 1212–1217.

[11]

Vasconcelos C.N., Rosenhahn B., Bipartite graph matching computation on GPU, in: International Workshop on Energy Minimization Methods in Computer Vision and Pattern Recognition, Springer, 2009, pp. 42–55.

[12]

Azad A., Halappanavar M., Dobrian F., Pothen A., Computing maximum matching in parallel on bipartite graphs: worth the effort?, in: Proceedings of the 1st Workshop on Irregular Applications: Architectures and Algorithms, ACM, 2011, pp. 11–14.

[13]

Deveci M., Kaya K., Uçar B., Çatalyürek Ü.V., GPU Accelerated maximum cardinality matching algorithms for bipartite graphs, in: European Conference on Parallel Processing, Springer, 2013, pp. 850–861.

[14]

Lawer E., Lenstra J.K., Kan A.R., Shmoys D., The Traveling Salesman Problem: A Guided Tour of Combinatorial Optimization, Wiley and Sons New York, 1985.

[15]

H. Yin, H. Liu, An efficient multiuser loading algorithm for OFDM-based broadband wireless systems, in: Global Telecommunications Conference, 2000. GLOBECOM’00. IEEE, vol. 1, IEEE, 2000, pp. 103–107.

[16]

Pentico D.W., Assignment problems: A golden anniversary survey, European J. Oper. Res. 176 (2) (2007) 774–793.

[17]

Burkard R.E., Dell’Amico M., Martello S., Assignment Problems, Revised Reprint, SIAM, 2009.

[18]

Biggs N., Lloyd E.K., Wilson R.J., Graph Theory, 1736–1936, Oxford University Press, 1976.

[19]

Nvidia, Compute Unified Device Architecture (CUDA) Programming Guide, 2007.

[20]

P.A.C. Lopes, S.S. Yadav, A. Ilic, S.K. Patra, HungarianGPU on GitHub, http://github.com/paclopes/HungarianGPU.

[21]

Harris M., et al., Optimizing parallel reduction in CUDA, NVIDIA Dev. Technol. 2 (4) (2007).

Cited By

Kawtikwar SNagi R(2024)HyLACJournal of Parallel and Distributed Computing10.1016/j.jpdc.2024.104838187:COnline publication date: 1-May-2024
https://dl.acm.org/doi/10.1016/j.jpdc.2024.104838
Hijma PHeldens SSclocco Avan Werkhoven BBal H(2023)Optimization Techniques for GPU ProgrammingACM Computing Surveys10.1145/357063855:11(1-81)Online publication date: 16-Mar-2023
https://dl.acm.org/doi/10.1145/3570638
Wubben JHernández DCecilia JImberón BCalafate CCano JManzoni PToh C(2023)Assignment and Take-Off Approaches for Large-Scale Autonomous UAV SwarmsIEEE Transactions on Intelligent Transportation Systems10.1109/TITS.2023.324276524:5(4836-4847)Online publication date: 1-May-2023
https://dl.acm.org/doi/10.1109/TITS.2023.3242765

Index Terms

Fast block distributed CUDA implementation of the Hungarian algorithm

Index terms have been assigned to the content through auto-classification.

Recommendations

GPU-accelerated Hungarian algorithms for the Linear Assignment Problem

Linear Assignment is one of the most fundamental problems in operations research.A creative parallelization of a Hungarian-like algorithm on GPU cluster.Efficient parallelization of the augmenting path search step.Large problems with 1.6 billion ...
HyLAC: Hybrid linear assignment solver in CUDA
Abstract
The Linear Assignment Problem (LAP) is a fundamental combinatorial optimization problem with a wide range of applications. Over the years, significant progress has been made in developing efficient algorithms to solve the LAP, particularly in the ...
Highlights
- The fastest Linear Assignment Problem (LAP) solver that uses GPUs.
- Improved implementation of classical and tree variants of Hungarian algorithm.
- Solves a stream of small LAPs 22.59× faster than the existing solution.
A CUDA implementation of the Continuous Space Language Model

The training phase of the Continuous Space Language Model (CSLM) was implemented in the NVIDIA hardware/software architecture Compute Unified Device Architecture (CUDA). A detailed explanation of the CSLM algorithm is provided. Implementation was ...

Comments

Information & Contributors

Information

Published In

cover image Journal of Parallel and Distributed Computing

Journal of Parallel and Distributed Computing Volume 130, Issue C

Aug 2019

209 pages

ISSN:0743-7315

Issue’s Table of Contents

Elsevier Inc.

Publisher

Academic Press, Inc.

United States

Publication History

Published: 01 August 2019

Author Tags

Qualifiers

Research-article

Contributors

Other Metrics

View Article Metrics

Bibliometrics & Citations

Bibliometrics

Article Metrics

3
Total Citations
View Citations
0
Total Downloads

Downloads (Last 12 months)0
Downloads (Last 6 weeks)0

Reflects downloads up to 04 Oct 2024

Other Metrics

View Author Metrics

Citations

Cited By

Kawtikwar SNagi R(2024)HyLACJournal of Parallel and Distributed Computing10.1016/j.jpdc.2024.104838187:COnline publication date: 1-May-2024
https://dl.acm.org/doi/10.1016/j.jpdc.2024.104838
Hijma PHeldens SSclocco Avan Werkhoven BBal H(2023)Optimization Techniques for GPU ProgrammingACM Computing Surveys10.1145/357063855:11(1-81)Online publication date: 16-Mar-2023
https://dl.acm.org/doi/10.1145/3570638
Wubben JHernández DCecilia JImberón BCalafate CCano JManzoni PToh C(2023)Assignment and Take-Off Approaches for Large-Scale Autonomous UAV SwarmsIEEE Transactions on Intelligent Transportation Systems10.1109/TITS.2023.324276524:5(4836-4847)Online publication date: 1-May-2023
https://dl.acm.org/doi/10.1109/TITS.2023.3242765

View Options

View options

Get Access

Login options

Check if you have access through your login credentials or your institution to get full access on this article.

Full Access

Get this Publication

Media

Figures

Other

Tables

View Issue’s Table of Contents