Please use this identifier to cite or link to this item: http://hdl.handle.net/1942/23913
Full metadata record
DC FieldValueLanguage
dc.contributor.authorNEMETH, Balazs-
dc.contributor.authorHABER, Tom-
dc.contributor.authorAshby, Thomas J.-
dc.contributor.authorLAMOTTE, Wim-
dc.date.accessioned2017-06-21T12:21:02Z-
dc.date.available2017-06-21T12:21:02Z-
dc.date.issued2017-
dc.identifier.citationProcedia Computer Science, 108(C), p. 2348-2352 (Art N° 205)-
dc.identifier.issn1877-0509-
dc.identifier.urihttp://hdl.handle.net/1942/23913-
dc.description.abstractTypically, parallel algorithms are developed to leverage the processing power of multiple processors simultaneously speeding up overall execution. At the same time, discrepancy between \{DRAM\} bandwidth and microprocessor speed hinders reaching peak performance. This paper explores how operational intensity improves by performing useful computation during otherwise stalled cycles. While the proposed methodology is applicable to a wide variety of parallel algorithms, and at different scales, the concepts are demonstrated in the machine learning context. Performance improvements are shown for Bayesian logistic regression with a Markov chain Monte Carlo sampler, either with multiple chains or with multiple proposals, on a dense data set two orders of magnitude larger than the last level cache on contemporary systems.-
dc.description.sponsorshipPart of the work presented in this paper was funded by Johnson & Johnson. This project has received funding from the European Union’s Horizon 2020 Research and Innovation programme under Grant Agreement no. 671555.-
dc.language.isoen-
dc.publisherElsevier Science BV-
dc.relation.ispartofseriesProcedia Computer Science-
dc.rights© 2017 The Authors. Published by Elsevier B.V. Peer-review under responsibility of the scientific committee of the International Conference on Computational Science-
dc.subject.otheroperational intensity; MCMC; Bayesian logistic regression; HPC; Big Data-
dc.titleImproving Operational Intensity in Data Bound Markov Chain Monte Carlo-
dc.typeProceedings Paper-
local.bibliographicCitation.authorsKoumoutsakos, Petros-
local.bibliographicCitation.authorsLees, Michael-
local.bibliographicCitation.authorsKrzhizhanovskaya, Valeria-
local.bibliographicCitation.authorsDongarra, Jack-
local.bibliographicCitation.authorsSloot, Peter-
local.bibliographicCitation.conferencedate12-14/06/2017-
local.bibliographicCitation.conferencenameInternational Conference on Computational Science (ICCS 2017)-
local.bibliographicCitation.conferenceplaceZurich, Switzerland-
dc.identifier.epage2352-
dc.identifier.issueC-
dc.identifier.spage2348-
dc.identifier.volume108-
local.bibliographicCitation.jcatC1-
local.publisher.placeAmsterdam, The Netherlands-
dc.relation.references@inproceedings{Amdahl:1967, author = {Amdahl, Gene M.}, title = {Validity of the Single Processor Approach to Achieving Large Scale Computing Capabilities}, booktitle = {Proc. of the April 18-20, 1967, Spring Joint Computer Conf.}, series = {AFIPS '67 (Spring)}, year = {1967}, location = {Atlantic City, New Jersey}, pages = {483--485}, numpages = {3}, url = {http://doi.acm.org/10.1145/1465482.1465560}, acmid = {1465560}, publisher = {ACM}, address = {New York, NY, USA}, } @article{Andrieu2003, title = {{An introduction to MCMC for machine learning}}, year = {2003}, journal = {Machine Learning}, author = {Andrieu, Christophe and De Freitas, Nando and Doucet, Arnaud and Jordan, Michael I.}, number = {1-2}, pages = {5--43}, volume = {50}, isbn = {08856125 (ISSN)}, issn = {08856125}, pmid = {178037200001}, arxivId = {1109.4435v1}, keywords = {MCMC, Markov chain Monte Carlo, Sampling, Stochastic algorithms} } @article{Calderhead2014, title = {{A general construction for parallelizing Metropolis-Hastings algorithms.}}, year = {2014}, journal = {Proc. of the National Academy of Sciences of the United States of America}, author = {Calderhead, Ben}, number = {49}, pages = {17408--13}, volume = {111}, url = {http://www.pnas.org/content/111/49/17408.abstract}, issn = {1091-6490}, pmid = {25422442}, keywords = {Bayesian inference, Hamiltonian dynamics, Markov chain Monte Carlo, parallel computation} } @article{Banterle2014, title = {{Accelerating Metropolis-Hastings algorithms: Delayed acceptance with prefetching}}, year = {2014}, journal = {arXiv}, author = {Banterle, Marco and Grazian, Clara and Robert, Christian P}, pages = {20}, url = {http://arxiv.org/abs/1406.2660}, arxivId = {1406.2660}, keywords = {acceptance probability, big data, higgs boson, jeffreys prior, large scale learning and, likelihood function, mcmc, mixtures of distributions} } @article{Korattikara2013, title = {{Austerity in MCMC Land: Cutting the Metropolis-Hastings Budget}}, year = {2013}, journal = {Proc. of the 30th Int. Conf. on Machine Learning}, author = {Korattikara, Anoop and Chen, Yutian and Welling, Max}, pages = {1--23}, volume = {32}, url = {http://arxiv.org/abs/1304.5299}, isbn = {9781634393973}, arxivId = {1304.5299} } @article{Scott2016, title = {{Bayes and big data: the consensus Monte Carlo algorithm}}, year = {2016}, journal = {Int. Journal of Management Science and Engineering Management}, author = {Scott, Steven L. and Blocker, Alexander W. and Bonassi, Fernando V and Chipman, Hugh A and George, Edward I and McCulloch, Robert E.}, number = {2}, pages = {78--88}, volume = {11}, url = {http://www.rob-mcculloch.org/some_papers_and_talks/papers/working/consensus-mc.pdf}, issn = {1750-9653} } @book{Eijkhout2011, title = {{Introduction to High Performance Scientific Computing}}, year = {2011}, author = {Eijkhout, Victor}, pages = {446}, publisher = {lulu.com}, url = {http://www.tacc.utexas.edu/~eijkhout/istc/istc.html}, isbn = {978-1-257-99254-6}, keywords = {parallel computing} } @article{Robert2015, title = {{The Metropolis-Hastings algorithm}}, year = {2015}, author = {Robert, Christian P.}, number = {Mcmc}, pages = {1--15}, url = {http://arxiv.org/abs/1504.01896}, isbn = {9781118445112}, arxivId = {1504.01896}, keywords = {and phrases, bayesian inference, gibbs sampler, hamiltonian monte carlo, hastings algorithm, intractable density, langevin diffusion, markov chains, mcmc meth-, metropolis, ods} } @article{Zhao2015, title = {{Towards cost-effective and high-performance caching middleware for distributed systems}}, year = {2016}, journal = {Int. Journal of Big Data Intell.}, author = {Zhao, Dongfang and Qiao, Kan and Raicu, Ioan}, number = {2}, pages = {92}, volume = {3}, url = {http://www.inderscience.com/link.php?id=77358}, issn = {2053-1389}, keywords = {Heterogeneous Storage, Hy-brid File Systems, Index Terms—Distributed File Systems, SSD, User Level File Systems} } @article{He2014, author="Qiu, Junfei and Wu, Qihui and Ding, Guoru and Xu, Yuhua and Feng, Shuo", title="A survey of machine learning for big data processing", journal="EURASIP Journal on Advances in Signal Processing", year="2016", volume="2016", number="1", pages="67", abstract="There is no doubt that big data are now rapidly expanding in all science and engineering domains. While the potential of these massive data is undoubtedly significant, fully making sense of them requires new ways of thinking and novel learning techniques to address the various challenges. In this paper, we present a literature survey of the latest advances in researches on machine learning for big data processing. First, we review the machine learning techniques and highlight some promising learning methods in recent studies, such as representation learning, deep learning, distributed and parallel learning, transfer learning, active learning, and kernel-based learning. Next, we focus on the analysis and discussions about the challenges and possible solutions of machine learning for big data. Following that, we investigate the close connections of machine learning with signal processing techniques for big data processing. Finally, we outline several open issues and research trends.", issn="1687-6180", doi="10.1186/s13634-016-0355-x", url="http://dx.doi.org/10.1186/s13634-016-0355-x" } @article{McVoy1996, title = {{lmbench: Portable Tools for Performance Analysis}}, year = {1996}, journal = {Proc. of the USENIX Annual Technical Conf.}, author = {McVoy, Larry and Staelin, Carl}, number = {January} } @article{Williams2009, title = {{Roofline: an insightful visual performance model for multicore architectures}}, year = {2009}, journal = {Communications of the ACM}, author = {Williams, Samuel and Waterman, Andrew and Patterson, David}, number = {4}, pages = {65–76}, volume = {52}, url = {www.eecs.berkeley.edu/~waterman/papers/roofline.pdf}, isbn = {0001-0782}, issn = {0001-0782} } @article{Wulf1995, title = {{Hitting the memory wall: Implications of the Obvious}}, year = {1995}, journal = {ACM SIGARCH Computer Architecture News}, author = {Wulf, Wm. a. and McKee, Sally a.}, pages = {20--24}, volume = {23}, isbn = {0163-5964}, issn = {01635964} } @book{P.Murphy2012a, title = {{Machine Learning: A Probabilistic Perspective}}, year = {2012}, booktitle = {Machine Learning: A Probabilistic Perspective}, author = {P. Murphy, Kevin}, url = {https://mitpress.mit.edu/books/machine-learning-0}, isbn = {9780262306164} }-
local.type.refereedRefereed-
local.type.specifiedProceedings Paper-
local.relation.ispartofseriesnr108-
local.bibliographicCitation.artnr205-
local.type.programmeH2020-
local.relation.h2020671555-
dc.identifier.doi10.1016/j.procs.2017.05.024-
dc.identifier.isi000404959000249-
dc.identifier.urlhttp://www.sciencedirect.com/science/article/pii/S1877050917305252-
local.bibliographicCitation.btitleProcedia Computer Science-
item.fulltextWith Fulltext-
item.contributorNEMETH, Balazs-
item.contributorHABER, Tom-
item.contributorAshby, Thomas J.-
item.contributorLAMOTTE, Wim-
item.accessRightsOpen Access-
item.validationecoom 2018-
item.fullcitationNEMETH, Balazs; HABER, Tom; Ashby, Thomas J. & LAMOTTE, Wim (2017) Improving Operational Intensity in Data Bound Markov Chain Monte Carlo. In: Procedia Computer Science, 108(C), p. 2348-2352 (Art N° 205).-
Appears in Collections:Research publications
Files in This Item:
File Description SizeFormat 
1-s2.0-S1877050917305252-main.pdfPublished version389.25 kBAdobe PDFView/Open
Show simple item record

SCOPUSTM   
Citations

1
checked on Sep 5, 2020

WEB OF SCIENCETM
Citations

1
checked on Apr 24, 2024

Page view(s)

88
checked on Sep 5, 2022

Download(s)

110
checked on Sep 5, 2022

Google ScholarTM

Check

Altmetric


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.