Abstract
Coupling a database and a parallel-programming framework reduces the I/O overhead between them. However, there will be serious issues such as memory bandwidth limitations, load imbalances, and race conditions. Existing frameworks such as MapReduce do not resolve these problems because they adopt flat parallelization, i.e., partitioning a task without regard to its structure. In this paper, we propose a recursive divide-and-conquer-based method for spatial databases which supports high-throughput machine learning. Our approach uses a tree-based task structure, which improves the reference locality, and load balancing is realized by setting the grain size of tasks dynamically. Race conditions are also avoided. We applied our method to the task of learning a hierarchical Poisson mixture model. The results show that our approach achieves strong scalability and robustness against load-imbalanced datasets.
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
Similar content being viewed by others
References
Stauffer, C., Grimson, W.E.L.: Adaptive background mixture models for real-time tracking. In: IEEE Computer Society Conference on Computer Vision and Pattern Recognition, Jun 1999
Miura, K., Noguchi, H., Kawaguchi, H., Yoshimoto, M.: A low memory bandwidth gaussian mixture model(GMM) processor for 20,000-word real-time speech recognition FPGA system. In: International Conference on ICECE Technology, Dec 2008
Gupta, K., Owens, J.D.: Three-layer optimizations for fast GMM computations on GPU-like parallel processors. In: IEEE Workshop on Automatic Speech Recognition & Understanding, Dec 2009
Pereira, S.S., Lopez-Valcarce, R., Pages-Zamora, A.: A diffusion-based EM algorithm for distributed estimation in unreliable sensor networks. IEEE Signal Process. Lett. 20(6), 595–598 (2013)
Kinoshita, A., Takasu, A., Adachi, J.: Traffic incident detection using probabilistic topic model. In: Proceedings of the Workshops of the EDBT/ICDT 2014 Joint Conference, Mar 2014
Kwedlo, W.: A parallel EM algorithm for Gaussian mixture models implemented on a NUMA system using OpenMP. In: 2014 22nd Euromicro International Conference on Parallel, Distributed and Network-Based Processing (PDP), Feb 2014
Dean, J., Ghemawat, S.: MapReduce: simplified data processing on large clusters. In: Proceedings of the 6th Conference on Symposium on Opearting Systems Design & Implementation, vol. 6, Dec 2004
Zaharia, M., Chowdhury, M., Franklin, M.J., Shenker, S., Stoica, I.: Spark: cluster computing with working sets. In: Proceedings of the 2nd USENIX Conference on Hot Topics in Cloud Computing, Jun 2010
Power, R., Li, J.: Piccolo: building fast, distributed programs with partitioned tables. In: Proceedings of the 9th USENIX Conference on Operating Systems Design and Implementation, Oct 2010
Low, Y., Bickson, D., Gonzalez, J., Guestrin, C., Kyrola, A., Hellerstein, J.M.: Distributed GraphLab: a framework for machine learning and data mining in the cloud. In: Proceedings of the VLDB Endowment, Apr 2012
Zaharia, M., Chowdhury, M., Das, T., Dave, A., Ma, J., MacCauley, M., Franklin, M.J., Shenker, S., Stoica, I.: Resilient distributed datasets: a fault-tolerant abstraction for in-memory cluster computing. In: Proceedings of the 9th USENIX conference on Networked Systems Design and Implementation, Apr 2012
Yang, R., Xiong, T., Chen, T., Huang, Z., Feng, S.: DISTRIM: parallel GMM learning on multicore cluster. In: 2012 IEEE International Conference on Computer Science and Automation Engineering (CSAE), May 2012
Mohr, E., Kranz, D.A., Halstead Jr., R.H.: Lazy task creation: a technique for increasing the granularity of parallel programs. In: Proceedings of the 1990 ACM Conference on LISP and Functional Programming, May 1990
Blumofe, R.D., Joerg, C.F., Kuszmaul, B.C., Leiserson, C.E., Randall, K.H., Zhou, Y.: Cilk: an efficient multithreaded runtime system. In: Proceedings of the Fifth ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming, Aug 1995
Nakashima, J., Nakatani, S., Taura, K.: Design and implementation of a customizable work stealing scheduler. In: 3rd International Workshop on Runtime and Operating Systems for Supercomputers, Jun 2013
Guttman, A.: R-trees: a dynamic index structure for spatial searching. In: Proceedings of the 1984 ACM SIGMOD International Conference on Management of Data, Jun 1984
Acknowledgment
This work was supported by the CPS-IIP Project (http://www.cps.nii.ac.jp) in the research promotion program for national challenges “Research and development for the realization of next-generation IT platforms” of the Ministry of Education, Culture, Sports, Science and Technology, Japan. The environment on which we conducted our experiment was provided by Assistant Prof. Hajime Imura at the Meme Media Laboratory, Hokkaido University.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2015 Springer International Publishing Switzerland
About this paper
Cite this paper
Kawakatsu, T., Kinoshita, A., Takasu, A., Adachi, J. (2015). Highly Efficient Parallel Framework: A Divide-and-Conquer Approach. In: Chen, Q., Hameurlain, A., Toumani, F., Wagner, R., Decker, H. (eds) Database and Expert Systems Applications. Globe DEXA 2015 2015. Lecture Notes in Computer Science(), vol 9262. Springer, Cham. https://doi.org/10.1007/978-3-319-22852-5_15
Download citation
DOI: https://doi.org/10.1007/978-3-319-22852-5_15
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-22851-8
Online ISBN: 978-3-319-22852-5
eBook Packages: Computer ScienceComputer Science (R0)