Skip to main content
SpringerLink
Log in
SpringerLink
Log in

iPUG: Accelerating Breadth-First Graph Traversals Using Manycore Graphcore IPUs

  • Conference paper
  • First Online:
High Performance Computing (ISC High Performance 2021)

Part of the book series: Lecture Notes in Computer Science ((LNTCS,volume 12728))

Included in the following conference series:

Abstract

The Graphcore Intelligence Processing Unit (IPU) is a newly developed processor type whose architecture does not rely on the traditional caching hierarchies. Developed to meet the need for more and more data-centric applications, such as machine learning, IPUs combine a dedicated portion of SRAM with each of its numerous cores, resulting in high memory bandwidth at the price of capacity. The proximity of processor cores and memory makes the IPU a promising field of experimentation for graph algorithms since it is the unpredictable, irregular memory accesses that lead to performance losses in traditional processors with pre-caching.

This paper aims to test the IPU’s suitability for algorithms with hard-to-predict memory accesses by implementing a breadth-first search (BFS) that complies with the Graph500 specifications. Precisely because of its apparent simplicity, BFS is an established benchmark that is not only subroutine for a variety of more complex graph algorithms, but also allows comparability across a wide range of architectures.

We benchmark our IPU code on a wide range of instances and compare its performance to state-of-the-art CPU and GPU codes. The results indicate that the IPU delivers speedups of up to \(4{\times }\) over the fastest competing result on an NVIDIA V100 GPU, with typical speedups of about \(1.5{\times }\) on most test instances.

This is a preview of subscription content, log in via an institution to check access.

Access this chapter

Log in via an institution
Chapter
USD 29.95
Price excludes VAT (USA)
  • Available as PDF
  • Read on any device
  • Instant download
  • Own it forever
eBook
USD 39.99
Price excludes VAT (USA)
  • Available as EPUB and PDF
  • Read on any device
  • Instant download
  • Own it forever
Softcover Book
USD 54.99
Price excludes VAT (USA)
  • Compact, lightweight edition
  • Dispatched in 3 to 5 business days
  • Free shipping worldwide - see info

Tax calculation will be finalised at checkout

Purchases are for personal use only

Institutional subscriptions

Notes

  1. 1.

    Git commit: 5ee3df5, Online: https://github.com/gunrock/gunrock.

  2. 2.

    Git commit: 426846f, Online: https://github.com/iHeartGraph/Enterprise.

  3. 3.

    https://en.wikichip.org/wiki/amd/epyc/7302p.

  4. 4.

    https://en.wikichip.org/wiki/intel/xeon_gold/6130.

References

  1. Abadi, M., et al.: Tensorflow: a system for large-scale machine learning. In: 12th USENIX Symposium on Operating Systems Design and Implementation (OSDI 2016), pp. 265–283 (2016)

    Google Scholar 

  2. Abu-Khzam, F.N., Collins, R.L., Fellows, M.R., Langston, M.A., Suters, W.H., Symons, C.T.: Kernelization algorithms for the vertex cover problem (2017)

    Google Scholar 

  3. Aho, A.V., Sethi, R., Ullman, J.D.: Compilers, Principles, Techniques, and Tools. Addison-Wesley Pub. Co., Boston (1986)

    MATH  Google Scholar 

  4. Azad, A., Buluç, A.: Distributed-memory algorithms for maximum cardinality matching in bipartite graphs. In: 2016 IEEE International Parallel and Distributed Processing Symposium (IPDPS), pp. 32–42. IEEE (2016)

    Google Scholar 

  5. Bader, D.A., Madduri, K.: Designing multithreaded algorithms for breadth-first search and ST-connectivity on the cray MTA-2. In: 2006 International Conference on Parallel Processing (ICPP 2006), pp. 523–530. IEEE (2006)

    Google Scholar 

  6. Beamer, S., Asanović, K., Patterson, D.: The gap benchmark suite. arXiv preprint arXiv:1508.03619 (2015)

  7. Beamer, S., Asanovic, K., Patterson, D., Beamer, S., Patterson, D.: Searching for a parent instead of fighting over children: a fast breadth-first search implementation for graph500. EECS Department, University of California, Berkeley, Technical report UCB/EECS-2011-117 (2011)

    Google Scholar 

  8. Buluç, A., Beamer, S., Madduri, K., Asanovic, K., Patterson, D.: Distributed-memory breadth-first search on massive graphs. arXiv preprint arXiv:1705.04590 (2017)

  9. Buluç, A., Gilbert, J.R.: The combinatorial BLAS: design, implementation, and applications. Int. J. High Perf. Comput. Appl. 25(4), 496–509 (2011)

    Article  Google Scholar 

  10. Buluç, A., Madduri, K.: Parallel breadth-first search on distributed memory systems. In: Proceedings of 2011 International Conference for High Performance Computing, Networking, Storage and Analysis, pp. 1–12 (2011)

    Google Scholar 

  11. Chakrabarti, D., Zhan, Y., Faloutsos, C.: R-MAT: a recursive model for graph mining. In: Proceedings of the 2004 SIAM International Conference on Data Mining, pp. 442–446. SIAM (2004)

    Google Scholar 

  12. Checconi, F., Petrini, F.: Traversing trillions of edges in real time: graph exploration on large-scale parallel machines. In: 2014 IEEE 28th International Parallel and Distributed Processing Symposium, pp. 425–434. IEEE (2014)

    Google Scholar 

  13. Chenglong, Z., Huawei, C., Guobo, W., Qinfen, H., Yang, Z., Xiaochun, Y., Dongrui, F.: Efficient optimization of graph computing on high-throughput computer. J. Comput. Res. Dev. 57(6), 1152 (2020)

    Google Scholar 

  14. Gaihre, A., Wu, Z., Yao, F., Liu, H.: XBFS: exploring runtime optimizations for breadth-first search on GPUs. In: Proceedings of the 28th International Symposium on High-Performance Parallel and Distributed Computing, pp. 121–131 (2019)

    Google Scholar 

  15. Ghosh, R.K., Bhattacharjee, G.: Parallel breadth-first search algorithms for trees and graphs. Int. J. Comput. Math. 15(1–4), 255–268 (1984)

    Article  MathSciNet  MATH  Google Scholar 

  16. Gregor, D., Lumsdaine, A.: Lifting sequential graph algorithms for distributed-memory parallel computation. ACM SIGPLAN Not. 40(10), 423–437 (2005)

    Article  Google Scholar 

  17. Harish, P., Narayanan, P.J.: Accelerating large graph algorithms on the GPU using CUDA. In: Aluru, S., Parashar, M., Badrinath, R., Prasanna, V.K. (eds.) HiPC 2007. LNCS, vol. 4873, pp. 197–208. Springer, Heidelberg (2007). https://doi.org/10.1007/978-3-540-77220-0_21

    Chapter  Google Scholar 

  18. Hennessy, J.L., Patterson, D.A.: A new golden age for computer architecture. Commun. ACM 62(2), 48–60 (2019)

    Article  Google Scholar 

  19. Hong, S., Oguntebi, T., Olukotun, K.: Efficient parallel graph exploration on multi-core CPU and GPU. In: 2011 International Conference on Parallel Architectures and Compilation Techniques, pp. 78–88. IEEE (2011)

    Google Scholar 

  20. Jia, Z., Tillman, B., Maggioni, M., Scarpazza, D.P.: Dissecting the graphcore ipu architecture via microbenchmarking. arXiv preprint arXiv:1912.03413 (2019)

  21. Kaya, K., Langguth, J., Panagiotas, I., Uçar, B.: Karp-Sipser based kernels for bipartite graph matching. In: 2020 Proceedings of the Twenty-Second Workshop on Algorithm Engineering and Experiments (ALENEX), pp. 134–145. SIAM (2020)

    Google Scholar 

  22. Kolodziej, S.P., et al.: The suitesparse matrix collection website interface. J. Open Source Softw. 4(35), 1244 (2019)

    Article  Google Scholar 

  23. Korf, R.E., Schultze, P.: Large-scale parallel breadth-first search. In: AAAI, vol. 5, pp. 1380–1385 (2005)

    Google Scholar 

  24. Langguth, J., Azad, A., Halappanavar, M., Manne, F.: On parallel push-relabel based algorithms for bipartite maximum matching. Parallel Comput. 40(7), 289–308 (2014)

    Article  Google Scholar 

  25. Langguth, J., Cai, X., Sourouri, M.: Memory bandwidth contention: communication vs computation tradeoffs in supercomputers with multicore architectures. In: 2018 IEEE 24th International Conference on Parallel and Distributed Systems (ICPADS), pp. 497–506. IEEE (2018)

    Google Scholar 

  26. Langguth, J., Patwary, M.M.A., Manne, F.: Parallel algorithms for bipartite matching problems on distributed memory computers. Parallel Comput. 37(12), 820–845 (2011)

    Article  MATH  Google Scholar 

  27. Liu, H., Huang, H.H.: Enterprise: breadth-first graph traversal on GPUs. In: Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis, pp. 1–12 (2015)

    Google Scholar 

  28. Murphy, R.C., Wheeler, K.B., Barrett, B.W., Ang, J.A.: Introducing the graph 500. Cray Users Group (CUG) 19, 45–74 (2010)

    Google Scholar 

  29. Seshadhri, C., Pinar, A., Kolda, T.G.: An in-depth analysis of stochastic Kronecker graphs. J. ACM (JACM) 60(2), 1–32 (2013)

    Article  MathSciNet  MATH  Google Scholar 

  30. Valiant, L.G.: A bridging model for parallel computation. Commun. ACM 33(8), 103–111 (1990)

    Article  Google Scholar 

  31. Wang, Y., Davidson, A., Pan, Y., Wu, Y., Riffel, A., Owens, J.D.: Gunrock: a high-performance graph processing library on the GPU. In: Proceedings of the 21st ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming, pp. 1–12 (2016)

    Google Scholar 

  32. Yang, C., Buluc, A., Owens, J.D.: GraphBLAST: a high-performance linear algebra-based graph framework on the GPU (2020)

    Google Scholar 

  33. Yasui, Y., Fujisawa, K., Goto, K.: NUMA-optimized parallel breadth-first search on multicore single-node system. In: 2013 IEEE International Conference on Big Data, pp. 394–402. IEEE (2013)

    Google Scholar 

  34. Yoo, A., Chow, E., Henderson, K., McLendon, W., Hendrickson, B., Catalyurek, U.: A scalable distributed parallel breadth-first search algorithm on BlueGene/L. In: SC 2005: Proceedings of the 2005 ACM/IEEE Conference on Supercomputing, p. 25. IEEE, November 2005. https://doi.org/10.1109/SC.2005.4

Download references

Author information

Authors and Affiliations

  1. Simula Research Laboratory, Fornebu, Norway

    Luk Burchard, Johannes Moe, Konstantin Pogorelov & Johannes Langguth

  2. University of Oslo, Oslo, Norway

    Johannes Moe

  3. Technical University Berlin, Berlin, Germany

    Luk Burchard & Daniel Thilo Schroeder

  4. Simula Metropolitan Center for Digital Engineering, Oslo, Norway

    Daniel Thilo Schroeder

  5. BI Norwegian Business School, Oslo, Norway

    Johannes Langguth

Authors
  1. Luk Burchard
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Johannes Moe
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Daniel Thilo Schroeder
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Konstantin Pogorelov
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Johannes Langguth
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Luk Burchard .

Editor information

Editors and Affiliations

  1. Hewlett Packard Enterprise, Seattle, WA, USA

    Bradford L. Chamberlain

  2. University of Amsterdam, Amsterdam, The Netherlands

    Ana-Lucia Varbanescu

  3. Extreme Computing Research Center, Thuwal Jeddah, Saudi Arabia

    Hatem Ltaief

  4. The University of Tennessee, Knoxville, Knoxville, TN, USA

    Piotr Luszczek

Rights and permissions

Reprints and permissions

Copyright information

© 2021 Springer Nature Switzerland AG

About this paper

Check for updates. Verify currency and authenticity via CrossMark

Cite this paper

Burchard, L., Moe, J., Schroeder, D.T., Pogorelov, K., Langguth, J. (2021). iPUG: Accelerating Breadth-First Graph Traversals Using Manycore Graphcore IPUs. In: Chamberlain, B.L., Varbanescu, AL., Ltaief, H., Luszczek, P. (eds) High Performance Computing. ISC High Performance 2021. Lecture Notes in Computer Science(), vol 12728. Springer, Cham. https://doi.org/10.1007/978-3-030-78713-4_16

Download citation

  • DOI: https://doi.org/10.1007/978-3-030-78713-4_16

  • Published:

  • Publisher Name: Springer, Cham

  • Print ISBN: 978-3-030-78712-7

  • Online ISBN: 978-3-030-78713-4

  • eBook Packages: Computer ScienceComputer Science (R0)

Publish with us

Policies and ethics

Search

Navigation