Abstract
When managing multimedia data such as text, images, videos, etc., it only makes sense to search for similar objects, because it is difficult to imagine that it would be interesting to look for if there is an element in the database exactly the same as another given as an example. Hence, a solution can be modeled through metric spaces. In this scenario, for solving efficient searches, we preprocess the database to build an index; and then, utilizing this index, minimize the number of comparisons required to answer them. However, with very large metric databases this is not enough, it is also necessary to speed up queries using high-performance computing. Then, the GPGPU appears as a profitable alternative. Moreover, there are circumstances in which it is also reasonable to accept quick answers even if they are inexact or approximate.
In this work, we evaluate the trade-off between the answer quality and performance of our GPU implementation of Permutation Index. The implementation is a pure GPU, used to solve in parallel multiple approximate similarity searches on metric databases. Our proposal has two parallelism levels: intra-queries and inter-queries, many queries are solved in parallel (inter-parallelism), and each query is figured out in parallel (intra-parallelism). The experimental results confirm that the GPU Permutation Index is a remarkable recourse to solve approximate similarity searches on large metric databases.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Chávez, E., Navarro, G., Baeza-Yates, R., Marroquín, J.: Searching in metric spaces. ACM Comput. Surv. 33(3), 273–321 (2001)
Chávez, E., Figueroa, K., Navarro, G.: Proximity searching in high dimensional spaces with a proximity preserving order. In: Gelbukh, A., de Albornoz, Á., Terashima-Marín, H. (eds.) MICAI 2005. LNCS (LNAI), vol. 3789, pp. 405–414. Springer, Heidelberg (2005). https://doi.org/10.1007/11579427_41
Ciaccia, P., Patella, M.: Approximate and probabilistic methods. SIGSPATIAL Spec. 2(2), 16–19 (2010). https://doi.org/10.1145/1862413.1862418
Zezula, P., Amato, G., Dohnal, V., Batko, M.: Similarity Search: The Metric Space Approach, ser. Advances in Database Systems, vol. 32. Springer, Boston (2006). https://doi.org/10.1007/0-387-29151-2
Pacheco, P., Malensek, M.: An Introduction to Parallel Programming, ser. An Introduction to Parallel Programming. Elsevier Science (2021). https://books.google.com.ar/books?id=uAfXnQAACAAJ
Robey, R., Zamora, Y.: Parallel and High Performance Computing. Simon and Schuster (2021). https://books.google.com.ar/books?id=jNstEAAAQBAJ
Kirk, D.B., Hwu, W.: Programming Massively Parallel Processors, A Hands on Approach. Elsevier, Morgan Kaufmann (2017). https://doi.org/10.1016/C2015-0-02431-5
Barrientos, R., Millaguir, F., Sánchez, J.L., Arias, E.: GPU-based exhaustive algorithms processing KNN queries. J. Supercomput. 73, 4611–4634 (2017)
Kruliš, M., Osipyan, H., Marchand-Maillet, S.: Employing GPU architectures for permutation-based indexing. Multimedia Tools Appl. 76(05) (2017). https://doi.org/10.1007/s11042-016-3677-7
Li, S., Amenta, N.: Brute-Force k-nearest neighbors search on the GPU. In: Amato, G., Connor, R., Falchi, F., Gennaro, C. (eds.) SISAP 2015. LNCS, vol. 9371, pp. 259–270. Springer, Cham (2015). https://doi.org/10.1007/978-3-319-25087-8_25
Velentzas, P., Vassilakopoulos, M., Corral, A.: In-memory k nearest neighbor GPU-based query processing. In: Proceedings of the 6th International Conference on Geographical Information Systems Theory, Applications and Management - GISTAM, INSTICC, pp. 310–317. SciTePress (2020)
Barrientos, R.J., Gómez, J.I., Tenllado, C., Prieto, M., Zezula, P.: Multi-level clustering on metric spaces using a multi-GPU platform. In: Euro-Par (2013)
Eder dos Santos, R.U.P., Sofia, A.A.O.: Procesamiento de búsquedas por similitud. tecnologías de paralelización e indexación. In: Informe Científico Técnico UNPA, vol. 7, no. 2, pp. 111–138 (2015)
Gowanlock, M., Karsin, B.: Accelerating the similarity self-join using the GPU. J. Parallel Distrib. Comput. 133, 06 (2019)
Barrientos, R.J., Riquelme, J.A., Hernández-García, R., Navarro, C.A., Soto-Silva, W.: Fast KNN query processing over a multi-node GPU environment. J. Supercomput., 3045–3071 (2022)
Riquelme, J.A., Barrientos, R.J., Hernández-García, R., Navarro, C.A.: An exhaustive algorithm based on GPU to process a KNN query. In: 2020 39th International Conference of the Chilean Computer Science Society (SCCC), pp. 1–8 (2020)
Lopresti, M., Piccoli, F., Reyes, N.: Goodness of the GPU permutation index: performance and quality results. In: XXVII Congreso Argentino de Ciencias de la Computación, CACIC 2021, pp. 321–332 (2021)
Figueroa, K., Chávez, E., Navarro, G., Paredes, R.: Speeding up spatial approximation search in metric spaces. ACM J. Exp. Algorithmics 14 (2009). Article 3.6
Bustos, B., Navarro, G.: Probabilistic proximity searching algorithms based on compact partitions. J. Discrete Algorithms 2(1), 115–134 (2004). The 9th International Symposium on String Processing and Information Retrieval. https://www.sciencedirect.com/science/article/pii/S1570866703000674
Tokoro, K., Yamaguchi, K., Masuda, S.: Improvements of TLAESA nearest neighbour search algorithm and extension to approximation search. In: Proceedings of the 29th Australasian Computer Science Conference - Volume 48, ser. ACSC 2006, Darlinghurst, Australia, pp. 77–83. Australian Computer Society Inc., Australia (2006). http://dl.acm.org/citation.cfm?id=1151699.1151709
Singh, A., Ferhatosmanoglu, H., Tosun, A.: High dimensional reverse nearest neighbor queries. In: The Twelfth International Conference on Information and Knowledge Management, ser. CIKM 2003, pp. 91–98. ACM, New York (2003). https://doi.org/10.1145/956863.956882
Moreno-Seco, F., Micó, L., Oncina, J.: A modification of the LAESA algorithm for approximated k-NN classification. Pattern Recogn. Lett. 24(1–3), 47–53 (2003). http://www.sciencedirect.com/science/article/pii/S0167865502001873
Fagin, R., Kumar, R., Sivakumar, D.: Comparing top k lists. In: Proceedings of the Fourteenth Annual ACM-SIAM Symposium on Discrete Algorithms, ser. SODA 2003, pp. 28–36. Society for Industrial and Applied Mathematics, Philadelphia (2003). http://dl.acm.org/citation.cfm?id=644108.644113
Baeza-Yates, R.A., Ribeiro-Neto, B.A.: Modern Information Retrieval. Pearson Education Ltd., Harlow (2011)
Cheng, J., Grossman, M.: Professional CUDA C Programming. CreateSpace Independent Publishing Platform (2017). https://books.google.com.ar/books?id=ApBtswEACAAJ
Han, J., Sharma, B.: Learn CUDA Programming: A Beginner’s Guide to GPU Programming and Parallel Computing with CUDA 10.x and C/C++. Packt Publishing (2019). https://books.google.com.ar/books?id=dhWzDwAAQBAJ
NVIDIA: Nvidia CUDA compute unified device architecture, programming guide, in NVIDIA (2020)
Lopresti, M., Miranda, N., Piccoli, F., Reyes, N.: Permutation index and GPU to solve efficiently many queries. In: VI Latin American Symposium on High Performance Computing, HPCLatAm 2013, pp. 101–112 (2013)
Figueroa, K., Reyes, N.: Permutation’s signatures for proximity searching in metric spaces. In: Amato, G., Gennaro, C., Oria, V., Radovanović, M. (eds.) SISAP 2019. LNCS, vol. 11807, pp. 151–159. Springer, Cham (2019). https://doi.org/10.1007/978-3-030-32047-8_14
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2022 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Lopresti, M., Piccoli, F., Reyes, N. (2022). GPU Permutation Index: Good Trade-Off Between Efficiency and Results Quality. In: Pesado, P., Gil, G. (eds) Computer Science – CACIC 2021. CACIC 2021. Communications in Computer and Information Science, vol 1584. Springer, Cham. https://doi.org/10.1007/978-3-031-05903-2_13
Download citation
DOI: https://doi.org/10.1007/978-3-031-05903-2_13
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-05902-5
Online ISBN: 978-3-031-05903-2
eBook Packages: Computer ScienceComputer Science (R0)