Skip to main content
SpringerLink
Log in
Book cover

Transactions on Large-Scale Data- and Knowledge-Centered Systems XV pp 1–35Cite as

GPU-Accelerated Database Systems: Survey and Open Challenges

  • Chapter
  • First Online:

Part of the book series: Lecture Notes in Computer Science ((TLDKS,volume 8920))

Abstract

The vast amount of processing power and memory bandwidth provided by modern graphics cards make them an interesting platform for data-intensive applications. Unsurprisingly, the database research community identified GPUs as effective co-processors for data processing several years ago. In the past years, there were many approaches to make use of GPUs at different levels of a database system. In this paper, we explore the design space of GPU-accelerated database management systems. Based on this survey, we present key properties, important trade-offs and typical challenges of GPU-aware database architectures, and identify major open challenges. Additionally, we survey existing GPU-accelerated DBMSs and classify their architectural properties. Then, we summarize typical optimizations implemented in GPU-accelerated DBMSs. Finally, we propose a reference architecture, indicating how GPU acceleration can be integrated in existing DBMSs.

This paper is a substantially extended version of an earlier work [17].

This is a preview of subscription content, 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

Learn about institutional subscriptions

Notes

  1. 1.

    Typically around 2–4 GB on mainstream cards and up to 16 GB on high-end devices.

  2. 2.

    We are aware that this features are included in OpenCL 2.0 but no OpenCL framework supports this features yet.

  3. 3.

    Some potential strategies include keeping the hot set of the data resident on the graphics card, or using the limited graphics card memory as a low-resolution data storage to quickly filter out non-matching data items [47].

  4. 4.

    Note that we deliberately excluded commercial systems such as Jedox [1] or Parstream [2], because they are neither available as open source nor have publications available that provide full architectural details.

  5. 5.

    Source code available at: http://wwwiti.cs.uni-magdeburg.de/iti_db/research/gpu/cogadb/.

  6. 6.

    Source code available at: https://code.google.com/p/gpudb/.

  7. 7.

    Source code available at: http://www.cse.ust.hk/gpuqp/.

  8. 8.

    http://mapd.csail.mit.edu/tweetmap/.

  9. 9.

    Source code available at: http://goo.gl/GHeUv.

  10. 10.

    Source code available at: https://code.google.com/p/omnidb-paralleldbonapu/.

  11. 11.

    Source code available at: https://github.com/bakks/virginian.

  12. 12.

    Note that both systems still apply a block-oriented processing model. This is due to the nature of compilation-based strategies, as discussed in Sect. 3.

  13. 13.

    Since these models need to be able to estimate comparable operator runtimes across different devices, we and others [13] argue that dynamic cost models, which apply techniques from Machine Learning to adapt to the current hardware, are likely required here.

  14. 14.

    We are aware that some in-memory DBMSs can also store data row-oriented, such as HyPer [38]. However, in GDBMSs, row-oriented storage either increases the data volume to be transfered or requires a projection operation before the transfer. A row-oriented layout also makes it difficult to achieve optimal memory access patterns on a GPU.

References

  1. Palo GPU accelerator. White Paper (2010)

    Google Scholar 

  2. Parstream - turning data into knowledge. White Paper, November 2010

    Google Scholar 

  3. Ailamaki, A., DeWitt, D.J., Hill, M.D., Skounakis, M.: Weaving relations for cache performance. In: VLDB, pp. 169–180. Morgan Kaufmann Publishers Inc. (2001)

    Google Scholar 

  4. Andrzejewski, W., Wrembel, R.: GPU-WAH: applying GPUs to compressing bitmap indexes with word aligned hybrid. In: Bringas, P.G., Hameurlain, A., Quirchmayr, G. (eds.) DEXA 2010, Part II. LNCS, vol. 6262, pp. 315–329. Springer, Heidelberg (2010)

    Chapter  Google Scholar 

  5. Augonnet, C., Thibault, S., Namyst, R., Wacrenier, P.-A.: StarPU: a unified platform for task scheduling on heterogeneous multicore architectures. Concurr. Comput. Pract. Exp. 23(2), 187–198 (2011)

    Article  Google Scholar 

  6. Bakkum, P., Chakradhar, S.: Efficient data management for GPU databases (2012). http://pbbakkum.com/virginian/paper.pdf

  7. Bakkum, P., Skadron, K.: Accelerating SQL database operations on a GPU with CUDA. In: GPGPU, pp. 94–103. ACM (2010)

    Google Scholar 

  8. Beier, F., Kilias, T., Sattler, K.-U.: GiST scan acceleration using coprocessors. In: DaMoN, pp. 63–69. ACM (2012)

    Google Scholar 

  9. Binnig, C., Hildenbrand, S., Färber, F.: Dictionary-based order-preserving string compression for main memory column stores. In: SIGMOD, pp. 283–296. ACM (2009)

    Google Scholar 

  10. Boncz, P.A., Kersten, M.L., Manegold, S.: Breaking the memory wall in MonetDB. Commun. ACM 51(12), 77–85 (2008)

    Article  Google Scholar 

  11. Boncz, P.A., Zukowski, M., Nes, N.: MonetDB/X100: hyper-pipelining query execution. In: CIDR, pp. 225–237 (2005)

    Google Scholar 

  12. Borkar, S., Chien, A.A.: The future of microprocessors. Commun. ACM 54(5), 67–77 (2011)

    Article  Google Scholar 

  13. Breß, S.: Why it is time for a HyPE: a hybrid query processing engine for efficient GPU coprocessing in dbms. The VLDB PhD Workshop, PVLDB 6(12), 1398–1403 (2013)

    Google Scholar 

  14. Breß, S., Beier, F., Rauhe, H., Sattler, K.-U., Schallehn, E., Saake, G.: Efficient co-processor utilization in database query processing. Inf. Syst. 38(8), 1084–1096 (2013)

    Article  Google Scholar 

  15. Breß, S., Geist, I., Schallehn, E., Mory, M., Saake, G.: A framework for cost based optimization of hybrid CPU/GPU query plans in database systems. Control Cybern. 41(4), 715–742 (2012)

    Google Scholar 

  16. Breß, S., Haberkorn, R., Ladewig, S.: CoGaDB reference manual (2014). http://wwwiti.cs.uni-magdeburg.de/iti_db/research/gpu/cogadb/0.3/doc/refman.pdf

  17. Breß, S., Heimel, M., Siegmund, N., Bellatreche, L., Saake, G.: Exploring the design space of a GPU-aware database architecture. In: Catania, B., et al. (eds.) New Trends in Databases and Information Systems. AISC, vol. 241, pp. 225–234. Springer, Heidelberg (2014)

    Chapter  Google Scholar 

  18. Breß, S., Siegmund, N., Bellatreche, L., Saake, G.: An operator-stream-based scheduling engine for effective GPU coprocessing. In: Catania, B., Guerrini, G., Pokorný, J. (eds.) ADBIS 2013. LNCS, vol. 8133, pp. 288–301. Springer, Heidelberg (2013)

    Chapter  Google Scholar 

  19. Broneske, D., Breß, S., Heimel, M., Saake, G.: Toward hardware-sensitive database operations. In: EDBT, pp. 229–234. OpenProceedings.org (2014)

    Google Scholar 

  20. Dees, J., Sanders, P.: Efficient many-core query execution in main memory column-stores. In: ICDE, pp. 350–361. IEEE (2013)

    Google Scholar 

  21. Diamos, G., Wu, H., Lele, A., Wang, J., Yalamanchili, S.: Efficient relational algebra algorithms and data structures for GPU. Technical report, Center for Experimental Research in Computer Systems (CERS) (2012)

    Google Scholar 

  22. Fang, R., He, B., Lu, M., Yang, K., Govindaraju, N.K., Luo, Q., Sander, P.V.: GPUQP: query co-processing using graphics processors. In: SIGMOD, pp. 1061–1063. ACM (2007)

    Google Scholar 

  23. Fang, W., He, B., Luo, Q.: Database compression on graphics processors. PVLDB 3, 670–680 (2010)

    Google Scholar 

  24. Gaster, B.R., Howes, L., Kaeli, D., Mistry, P., Schaa, D.: Heterogeneous Computing With Opencl. Elsevier Sci. Technol. 1–2 (2012)

    Google Scholar 

  25. Ghodsnia, P.: An in-GPU-memory column-oriented database for processing analytical workloads. In: The VLDB PhD Workshop. VLDB Endowment (2012)

    Google Scholar 

  26. Graefe, G.: Encapsulation of parallelism in the volcano query processing system. In: SIGMOD, pp. 102–111. ACM (1990)

    Google Scholar 

  27. Gregg, C., Hazelwood, K.: Where is the data? why you cannot debate CPU vs. GPU performance without the answer. In: ISPASS, pp. 134–144. IEEE (2011)

    Google Scholar 

  28. He, B., Fang, W., Luo, Q., Govindaraju, N.K., Wang, T.: Mars: a mapreduce framework on graphics processors. In: PACT, pp. 260–269. ACM (2008)

    Google Scholar 

  29. He, B., Lu, M., Yang, K., Fang, R., Govindaraju, N.K., Luo, Q., Sander, P.V.: Relational query co-processing on graphics processors. In: ACM Transactions on Database System, vol. 34. ACM (2009)

    Google Scholar 

  30. He, B., Yang, K., Fang, R., Lu, M., Govindaraju, N., Luo, Q., Sander, P.: Relational joins on graphics processors. In: SIGMOD, pp. 511–524. ACM (2008)

    Google Scholar 

  31. He, B., Yu, J.X.: High-throughput transaction executions on graphics processors. PVLDB 4(5), 314–325 (2011)

    MathSciNet  Google Scholar 

  32. He, J., Lu, M., He, B.: Revisiting co-processing for hash joins on the coupled CPU-GPU architecture. PVLDB 6(10), 889–900 (2013)

    Google Scholar 

  33. Heimel, M., Markl, V.: A first step towards GPU-assisted query optimization. In: ADMS. VLDB Endowment (2012)

    Google Scholar 

  34. Heimel, M., Saecker, M., Pirk, H., Manegold, S., Markl, V.: Hardware-oblivious parallelism for in-memory column-stores. PVLDB 6(9), 709–720 (2013)

    Google Scholar 

  35. Idreos, S., Groffen, F., Nes, N., Manegold, S., Mullender, K.S., Kersten, M.L.: MonetDB: Two decades of research in column-oriented database architectures. IEEE Data Eng. Bull. 35(1), 40–45 (2012)

    Google Scholar 

  36. Ilić, A., Sousa, L.: CHPS: an environment for collaborative execution on heterogeneous desktop systems. Int. J. Netw. Comput. 1(1), 96–113 (2011)

    Google Scholar 

  37. Kaldewey, T., Lohman, G., Mueller, R., Volk, P.: GPU join processing revisited. In: DaMoN, pp. 55–62. ACM (2012)

    Google Scholar 

  38. Kemper, A., Neumann, T.: HyPer: a hybrid OLTP&OLAP main memory database system based on virtual memory snapshots. In: ICDE, pp. 195–206. IEEE (2011)

    Google Scholar 

  39. Kossmann, D.: The state of the art in distributed query processing. ACM Comput. Surv. 32(4), 422–469 (2000)

    Article  Google Scholar 

  40. Manegold, S., Boncz, P., Kersten, M.L.: Generic database cost models for hierarchical memory systems. In: PVLDB, pp. 191–202. VLDB Endowment (2002)

    Google Scholar 

  41. Manegold, S., Boncz, P.A., Kersten, M.L.: Optimizing database architecture for the new bottleneck: memory access. VLDB J. 9(3), 231–246 (2000)

    Article  MATH  Google Scholar 

  42. Manegold, S., Kersten, M.L., Boncz, P.: Database architecture evolution: mammals flourished long before dinosaurs became extinct. PVLDB 2(2), 1648–1653 (2009)

    Google Scholar 

  43. Mostak, T.: An overview of MapD (massively parallel database). White Paper, Massachusetts Institute of Technology, April 2013. http://geops.csail.mit.edu/docs/mapd_overview.pdf

  44. Neumann, T.: Efficiently compiling efficient query plans for modern hardware. PVLDB 4(9), 539–550 (2011)

    Google Scholar 

  45. NVIDIA. NVIDIA CUDA C programming guide, pp. 31–36, 40, 213–216, Version 6.0. (2014). http://docs.nvidia.com/cuda/pdf/CUDA_C_Programming_Guide.pdf. Accessed 21 April 2014

  46. Owens, J.D., Luebke, D., Govindaraju, N., Harris, M., Krger, J., Lefohn, A.E., Purcell, T.J.: A survey of general-purpose computation on graphics hardware. Comput. Graph. Forum 26(1), 80–113 (2007)

    Article  Google Scholar 

  47. Pirk, H.: Efficient cross-device query processing. In: The VLDB PhD Workshop. VLDB Endowment (2012)

    Google Scholar 

  48. Pirk, H., Manegold, S., Kersten, M.: Accelerating foreign-key joins using asymmetric memory channels. In: ADMS, pp. 585–597. VLDB Endowment (2011)

    Google Scholar 

  49. Pirk, H., Manegold, S., Kersten, M.: Waste not... efficient co-processing of relational data. In: ICDE. IEEE (2014)

    Google Scholar 

  50. Przymus, P., Kaczmarski, K.: Dynamic compression strategy for time series database using GPU. In: Catania, B., et al. (eds.) New Trends in Databases and Information Systems. AISC, vol. 241, pp. 235–244. Springer, Heidelberg (2014)

    Chapter  Google Scholar 

  51. Przymus, P., Kaczmarski, K., Stencel, K.: A bi-objective optimization framework for heterogeneous CPU/GPU query plans. In: CS&P, pp. 342–354. CEUR-WS (2013)

    Google Scholar 

  52. Rabl, T., Poess, M., Jacobsen, H.-A., O’Neil, P., O’Neil, E.: Variations of the star schema benchmark to test the effects of data skew on query performance. In: ICPE, pp. 361–372. ACM (2013)

    Google Scholar 

  53. Rauhe, H., Dees, J., Sattler, K.-U., Faerber, F.: Multi-level parallel query execution framework for CPU and GPU. In: Catania, B., Guerrini, G., Pokorný, J. (eds.) ADBIS 2013. LNCS, vol. 8133, pp. 330–343. Springer, Heidelberg (2013)

    Chapter  Google Scholar 

  54. Răducanu, B., Boncz, P., Zukowski, M.: Micro adaptivity in vectorwise. In: SIGMOD, pp. 1231–1242. ACM (2013)

    Google Scholar 

  55. Saecker, M., Markl, V.: Big data analytics on modern hardware architectures: a technology survey. In: Aufaure, M.-A., Zimányi, E. (eds.) eBISS 2012. LNBIP, vol. 138, pp. 125–149. Springer, Heidelberg (2013)

    Chapter  Google Scholar 

  56. Sanders, J., Kandrot, E.: CUDA by Example: An Introduction to General-Purpose GPU Programming, 1st edn. Addison-Wesley Professional, Upper Saddle River (2010)

    Google Scholar 

  57. Schäler, M., Grebhahn, A., Schröter, R., Schulze, S., Köppen, V., Saake, G.: QuEval: beyond high-dimensional indexing à la carte. PVLDB 6(14), 1654–1665 (2013)

    Google Scholar 

  58. Selinger, P.G., Astrahan, M.M., Chamberlin, D.D., Lorie, R.A., Price, T.G.: Access path selection in a relational database management system. In: SIGMOD, pp. 23–34. ACM (1979)

    Google Scholar 

  59. Tsirogiannis, D., Harizopoulos, S., Shah, M.A.: Analyzing the energy efficiency of a database server. In: SIGMOD, pp. 231–242. ACM (2010)

    Google Scholar 

  60. Viglas, S.D.: Just-in-time compilation for SQL query processing. PVLDB 6(11), 1190–1191 (2013)

    Google Scholar 

  61. Wu, H., Diamos, G., Cadambi, S., Yalamanchili, S.: Kernel weaver: automatically fusing database primitives for efficient GPU computation. In: MICRO, pp. 107–118. IEEE (2012)

    Google Scholar 

  62. Yuan, Y., Lee, R., Zhang, X.: The yin and yang of processing data warehousing queries on GPU devices. PVLDB 6(10), 817–828 (2013)

    Google Scholar 

  63. Zhang, S., He, J., He, B., OmniDB, M.L.: Towards portable and efficient query processing on parallel CPU/GPU architectures. PVLDB 6(12), 1374–1377 (2013)

    Google Scholar 

  64. Zhong, J., He, B.: Medusa: simplified graph processing on gpus. IEEE Trans. Parallel Distrib. Syst. 99, 1–14 (2013)

    Google Scholar 

  65. Zhong, J., He, B.: Parallel graph processing on graphics processors made easy. PVLDB 6(12), 1270–1273 (2013)

    Google Scholar 

Download references

Acknowledgements

We thank Tobias Lauer from Jedox AG and the anonymous reviewers of the GPUs in Databases Workshop for their helpful feedback on the workshop version of this paper [17]. We thank Jens Teubner from TU Dortmund University, Michael Saecker from ParStream GmbH, and the anonymous reviewers of the TLDKS journal for their helpful comments on the journal version of this paper.

Author information

Authors and Affiliations

  1. University of Magdeburg, Magdeburg, Germany

    Sebastian Breß & Gunter Saake

  2. Technische Universität Berlin, Berlin, Germany

    Max Heimel

  3. University of Passau, Passau, Germany

    Norbert Siegmund

  4. LIAS/ISAE-ENSMA, Futuroscope, Poitiers, France

    Ladjel Bellatreche

Authors
  1. Sebastian Breß
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Max Heimel
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Norbert Siegmund
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Ladjel Bellatreche
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Gunter Saake
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Sebastian Breß .

Editor information

Editors and Affiliations

  1. IRIT, Paul Sabatier University, Toulouse, France

    Abdelkader Hameurlain

  2. FAW, University of Linz, Linz, Austria

    Josef Küng

  3. FAW, University of Linz, Linz, Austria

    Roland Wagner

  4. University of Genoa, Genoa, Italy

    Barbara Catania

  5. University of Genoa, Genoa, Italy

    Giovanna Guerrini

  6. LIPADE, Paris Descartes University, Paris, France

    Themis Palpanas

  7. Charles University, Prague, Czech Republic

    Jaroslav Pokorný

  8. Aristotle University of Thessalonik, Thessaloniki, Greece

    Athena Vakali

Rights and permissions

Reprints and permissions

Copyright information

© 2014 Springer-Verlag Berlin Heidelberg

About this chapter

Cite this chapter

Breß, S., Heimel, M., Siegmund, N., Bellatreche, L., Saake, G. (2014). GPU-Accelerated Database Systems: Survey and Open Challenges. In: Hameurlain, A., et al. Transactions on Large-Scale Data- and Knowledge-Centered Systems XV. Lecture Notes in Computer Science(), vol 8920. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-662-45761-0_1

Download citation

  • DOI: https://doi.org/10.1007/978-3-662-45761-0_1

  • Published:

  • Publisher Name: Springer, Berlin, Heidelberg

  • Print ISBN: 978-3-662-45760-3

  • Online ISBN: 978-3-662-45761-0

  • eBook Packages: Computer ScienceComputer Science (R0)

Publish with us

Policies and ethics

Search

Navigation