|2/2015 - 13|
Optimizing the Forward Algorithm for Hidden Markov Model on IBM Roadrunner clustersSOIMAN, S.-I. , RUSU, I. , PENTIUC, S.-G.
|Click to see author's profile in SCOPUS, IEEE Xplore, Web of Science|
|Download PDF (1,006 KB) | Citation | Downloads: 359 | Views: 2,356|
forward algorithm, hidden Markov models, multicore processing, parallel hybrid architectures, parallel programming, performance analysis
parallel(9), models(6), markov(6), hidden(6), cell(6), systems(5), ipdps(5), distributed(5), computing(5), recognition(4)
Blue keywords are present in both the references section and the paper title.
About this article
Date of Publication: 2015-05-31
Volume 15, Issue 2, Year 2015, On page(s): 103 - 108
ISSN: 1582-7445, e-ISSN: 1844-7600
Digital Object Identifier: 10.4316/AECE.2015.02013
Web of Science Accession Number: 000356808900013
SCOPUS ID: 84979846307
In this paper we present a parallel solution of the Forward Algorithm for Hidden Markov Models. The Forward algorithm compute a probability of a hidden state from Markov model at a certain time, this process being recursively. The whole process requires large computational resources for those models with a large number of states and long observation sequences. Our solution in order to reduce the computational time is a multilevel parallelization of Forward algorithm. Two types of cores were used in our implementation, for each level of parallelization, cores that are graved on the same chip of PowerXCell8i processor. This hybrid architecture of processors permitted us to obtain a speedup factor over 40 relative to the sequential algorithm for a model with 24 states and 25 millions of observable symbols. Experimental results showed that the parallel Forward algorithm can evaluate the probability of an observation sequence on a hidden Markov model 40 times faster than the classic one does. Based on the performance obtained, we demonstrate the applicability of this parallel implementation of Forward algorithm in complex problems such as large vocabulary speech recognition.
|References|||||Cited By «-- Click to see who has cited this paper|
| T. F. Oliver, B. Schmidt, Y. Jakop, D. L. Maskell, "High Speed Biological Sequence Analysis With Hidden Markov Models on Reconfigurable Platforms", Information Technology in Biomedicine, IEEE Transactions on 13(5): 740-746. |
[CrossRef] [Web of Science Times Cited 13]
 A. Sand, Pedersen, C. N. S. Pedersen, T. Mailund, A. T. Brask, "HMMlib: A C++ Library for General Hidden Markov Models Exploiting Modern CPUs", 2010 Ninth International Workshop on Parallel and Distributed Methods in Verification, and Second International Workshop on High Performance Computational Systems Biology, IEEE 2010, pp. 126 - 134, 2010.
 J. Nielsen, A. Sand, "Algorithms for a Parallel Implementation of Hidden Markov Models with a Small State Space", in Proc. IPDPS Workshops, IEEE 2011, pp.452-459.
 X. Meng, Y. Ji, "Modern Computational Techniques for the HMMER Sequence Analysis", vol.2013, 13 pages, 2013.
 S. Gorgunoglu, I. M. Orak, A. Cavusoglu, M. Gok, "Examination of Speed Contribution of Parallelization for Several Fingerprint Pre-Processing Algorithms," Advances in Electrical and Computer Engineering, vol. 14, no. 2, pp. 3-8, 2014,
[CrossRef] [Full Text] [Web of Science Times Cited 2]
 L. Yu, Y. Ukidave and D. Kaeli, "GPU-accelerated HMM for Speech Recognition", Workshop - Heterogeneous and Unconventional Cluster Architectures and Applications (HUCAA) September, 2014.
 J. Li, S. Chen, Y. Li, "The fast evaluation of hidden Markov models on GPU," Intelligent Computing and Intelligent Systems, 2009. ICIS 2009. IEEE International Conference on , vol.4, no., pp.426,430, 20-22 Nov. 2009.
[CrossRef] [Web of Science Times Cited 12]
 D. Zhihui, Y. Zhaoming, D.A. Bader, "A tile-based parallel Viterbi algorithm for biological sequence alignment on GPU with CUDA," Parallel & Distributed Processing, Workshops and Phd Forum (IPDPSW), 2010 IEEE International Symposium on , vol., no., pp.1,8, 19-23 April 2010.
 J.P. Walters, V. Balu, S. Kompalli, V. Chaudhary, "Evaluating the use of GPUs in liver image segmentation and HMMER database searches," Parallel & Distributed Processing, 2009. IPDPS 2009. IEEE International Symposium on , vol., no., pp.1,12, 23-29 May 2009.
 W. Lee, J. Kim, I. Lane, "GPU Accelerated Model Combination for Robust Speech Recognition and Keyword Search", GPU Technology Conference, March 2014
 T. Chen, R. Raghavan, J. N. Dale, E. Iwata, "Cell Broadband Engine Architecture and its first implementationA performance view", IBM Journal of Research and Development , vol.51, no.5, pp.559-572, 2007.
[CrossRef] [Web of Science Times Cited 107]
 V. Sachdeva, M. Kistler, E. Speight, T.-H. K. Tzeng, "Exploring the viability of the Cell Broadband Engine for bioinformatics applications, " Parallel Computing, vol. 34, no. 11, pp. 616-626, 2008.
[CrossRef] [Web of Science Times Cited 22]
 S.-I. Soiman, I. Rusu, S.-G. Pentiuc, "A parallel accelerated approach of HMM Forward Algorithm for IBM Roadrunner clusters", Proceedings of the 12th Int. Conf. on Development and Appl. Systems, May 2014, pp. 184-188. .
 S.-I. Soiman, I. Rusu, S.-G. Pentiuc, " Multilevel Parallelized Forward Algorithm for Hidden Markov Models on IBM Roadrunner Cluster", Proceedings of the 20th Int. Conf. on Control Systems and Computer Science, May 2015.
 F. Blagojevic, A. Stamatakis, C. D. Antonopoulos, D. S. Nikolopoulos, "RAxML-Cell: Parallel Phylogenetic Tree Inference on the Cell Broadband Engine, " Parallel and Distributed Processing Symposium, IEEE International, pp. 1-10, 2007.
 GRIDNORD Project. High Performance Computing Laboratory of the Faculty of Electrical Engineering and Computer Science, Suceava, Romania, 2012, [Online] Available: Temporary on-line reference link removed - see the PDF document
 A. L. Varbanescu, H. Sips, K.A. Ross, Q. Liu, A. Natsev, J.R. Smith and L.K. Liu, "Evaluating application mapping scenarios on the Cell/B.E, " Concurrency and Computation: Practice and Experience, 21, pp. 85-100, 2009.
[CrossRef] [Web of Science Times Cited 2]
 A. Arevalo, R.M. Matinata, M. Pandian, E. Peri, K. Ruby, F. Thomas, C. Almond: Programming for the Cell Broadband Engine. IBM Redbooks (2008)
 C. A. Tanase, V. G. Gaitan, "Threads Pipelining on the CellBE Systems", Advances in Electrical and Computer Engineering, vol. 13, no. 3, pp. 121-126, 2013.
[CrossRef] [Full Text] [Web of Science Times Cited 3]
 S.-G. Pentiuc, I. Ungurean, "Multilevel Parallelization of Unsupervised Learning Algorithms in Pattern Recognition on a Roadrunner Architecture ", Intelligent Distributed Computing V, vol. 382, pp.71 - 80, 2011.
 I. Ungurean, V.-G. Gaitan, N.-C. Gaitan, "Intensive computing on a large data volume with a short-vector single instruction multiple data processor," Computers & Digital Techniques, IET, vol.8, no.5, pp.219-228, 2014.
[CrossRef] [Web of Science Times Cited 1]
 L. Rabiner, "A tutorial on hidden Markov models and selected applications in speech recognition", Proceedings of IEEE, Vol. 77, pp. 257-285, 1989.
[CrossRef] [Web of Science Times Cited 11340]
Web of Science® Citations for all references: 11,502 TCR
SCOPUS® Citations for all references: 0
Web of Science® Average Citations per reference: 500 ACR
SCOPUS® Average Citations per reference: 0
TCR = Total Citations for References / ACR = Average Citations per Reference
We introduced in 2010 - for the first time in scientific publishing, the term "References Weight", as a quantitative indication of the quality ... Read more
Citations for references updated on 2021-01-11 19:39 in 101 seconds.
Note1: Web of Science® is a registered trademark of Clarivate Analytics.
Note2: SCOPUS® is a registered trademark of Elsevier B.V.
Disclaimer: All queries to the respective databases were made by using the DOI record of every reference (where available). Due to technical problems beyond our control, the information is not always accurate. Please use the CrossRef link to visit the respective publisher site.
Faculty of Electrical Engineering and Computer Science
Stefan cel Mare University of Suceava, Romania
All rights reserved: Advances in Electrical and Computer Engineering is a registered trademark of the Stefan cel Mare University of Suceava. No part of this publication may be reproduced, stored in a retrieval system, photocopied, recorded or archived, without the written permission from the Editor. When authors submit their papers for publication, they agree that the copyright for their article be transferred to the Faculty of Electrical Engineering and Computer Science, Stefan cel Mare University of Suceava, Romania, if and only if the articles are accepted for publication. The copyright covers the exclusive rights to reproduce and distribute the article, including reprints and translations.
Permission for other use: The copyright owner's consent does not extend to copying for general distribution, for promotion, for creating new works, or for resale. Specific written permission must be obtained from the Editor for such copying. Direct linking to files hosted on this website is strictly prohibited.
Disclaimer: Whilst every effort is made by the publishers and editorial board to see that no inaccurate or misleading data, opinions or statements appear in this journal, they wish to make it clear that all information and opinions formulated in the articles, as well as linguistic accuracy, are the sole responsibility of the author.