Abstract
Single-source shortest path (SSSP) is an important graph search algorithm for data-intensive applications which finds the minimum distance from a source vertex to any other vertex in a given graph. Although having been extensively studied for both single- and multi-node scenarios, SSSP search still brings severe challenge to communication when processing large graphs that consist of billions of vertices involving hundreds of computing nodes. To address this problem, in this paper we propose XSP, a fast SSSP search method based on communication-computation collaboration, which optimizes the communication of parallel SSSP in two aspects. First, we design a group-based scalable batching mechanism which effectively reduces the inter-machine communication overhead. Second, we propose a CCO (Communication-Computation Overlapping) method which realizes non-blocking execution of communication and computation. We have implemented XSP and extensive evaluation results show that the performance of XSP is significantly higher than that of the state-of-the-art parallel SSSP methods.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Eason, G., Noble, B., Sneddon, I.N.: On certain integrals of Lipschitz-Hankel type involving products of Bessel functions. Phil. Trans. Roy. Soc. London A247, 529–551 (1955)
Ueno, K., Suzumura, T.: Highly scalable graph search for the Graph500 benchmark. In: Proceedings of the 21st International Symposium on High-Performance Parallel and Distributed Computing, New York, NY, USA (2012)
Fuentes, P., Bosque, J.L., Beivide, R., Valero, M., Minkenberg, C.: Characterizing the communication demands of the Graph500 benchmark on a commodity cluster. In: Proceeding of the 2014 IEEE/ACM International Symposium on Big Data Computing, London, UK (2014)
Suzumura, T., Ueno, K., Sato, H., Fujisawa, K., Matsuoka, S.: Performance characteristics of Graph500 on large-scale distributed environment. In: Proceeding of the 2011 IEEE International Symposium on Workload Characterization (IISWC), Austin, TX, USA (2011)
Nakao, M., Ueno, K., Fujisawa, K., Kodama, Y., Satoh, M.: Performance evaluation of supercomputer Fugaku using breadth-first search benchmark in Graph500. In: Proceeding of the 2020 IEEE International Conference on Cluster Computing (CLUSTER), Kobe, Japan (2020)
Ueno, K., Suzumura, T.: Book 2D Partitioning Based Graph Search for the Graph500 Benchmark, pp. 1925–1931. IEEE (2012)
Beamer, S., Asanovic, K., Patterson, D.: Book Direction-optimizing Breadth-First Search, pp. 1–10. IEEE (2012)
Wang, R., et al.: Brief introduction of TianHe exascale prototype system. Tsinghua Sci. Technol. 26(3), 361–369 (2021)
Graph500 https://www.graph500.org/
Li, S., Huang, P.-C., Jacob, B.: Exascale Interconnect Topology Characterization and Parameter Exploration, pp. 810–819. IEEE (2018)
Zhu, Y., Taylor, M., Baden, S.B., Cheng, C.-K.: Advancing Supercomputer Performance Through Interconnection Topology Synthesis, pp. 555–558. IEEE (2008)
Blocksome, M., et al.: Design and Implementation of a One-Sided Communication Interface for the IBM eServer Blue Gene, p. 54. IEEE (2006)
Faraj, A., Kumar, S., Smith, B., Mamidala, A., Gunnels, J.: MPI Collective Communications on The Blue Gene/P Supercomputer Algorithms and Optimizations, pp. 63–72. IEEE (2009)
Yu, H., Chung, I.-H., Moreira, J.: Topology Mapping for Blue Gene/L Supercomputer, p. 52. IEEE (2006)
Kumar, S., et al.: PAMI A Parallel Active Message Interface for the Blue Gene/Q Supercomputer, pp. 763–773. IEEE (2012)
Shah, G., et al.: Performance and Experience with LAPI-a New High-Performance Communication Library for the IBM RS/6000 SP, pp. 260–266. IEEE (1998)
Shida, N., Sumimoto, S., Uno, A.: MPI library and low-level communication on the K computer. Fujitsu Sci. Tech. J. 48(3), 324–330 (2012)
Li, M., Lu, X., Potluri, S., Hamidouche, K., Tomko, J.J.K., Panda, D.K.: Scalable Graph500 design with MPI-3 RMA, pp. 230–238. IEEE (2014)
Dijkstra, E.W.: A note on two problems in connection with graphs. Numer. Math. 1(1), 269–271 (1959)
Nikas, K., Anastopoulos, N., Goumas, G., Koziris, N.: Employing Transactional Memory and Helper Threads to Speedup Dijkstras Algorithm, pp. 388–395. IEEE (2009)
Bellman, R.: On a routing problem. Q. Appl. Math. 16, 88–90 (1958)
Plimpton, S.J., Devine, K.D.: MAPREDUCE in MPI for large-scale graph algorithms. Parallel Comput. 37(9), 610–632 (2011)
Zhang, Y., et al.: TopoX topology refactorization for minimizing network communication in graph computations. IEEE/ACM Trans. Networking 28(6), 2768–2782 (2020)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2022 IFIP International Federation for Information Processing
About this paper
Cite this paper
Gan, X., Tan, W., Jia, M., Liu, J., Zhang, Y. (2022). XSP: Fast SSSP Based on Communication-Computation Collaboration. In: Cérin, C., Qian, D., Gaudiot, JL., Tan, G., Zuckerman, S. (eds) Network and Parallel Computing. NPC 2021. Lecture Notes in Computer Science(), vol 13152. Springer, Cham. https://doi.org/10.1007/978-3-030-93571-9_5
Download citation
DOI: https://doi.org/10.1007/978-3-030-93571-9_5
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-93570-2
Online ISBN: 978-3-030-93571-9
eBook Packages: Computer ScienceComputer Science (R0)
