Lecture Notes in Computer Science 5336 Commenced Publication in 1973 Founding and Former Series Editors: Gerhard Goos, Juris Hartmanis, and Jan van Leeuwen Editorial Board David Hutchison Lancaster University, UK Takeo Kanade Carnegie Mellon University, Pittsburgh, PA, USA Josef Kittler University of Surrey, Guildford, UK Jon M. Kleinberg Cornell University, Ithaca, NY, USA Alfred Kobsa University of California, Irvine, CA, USA Friedemann Mattern ETH Zurich, Switzerland John C. Mitchell Stanford University, CA, USA Moni Naor Weizmann Institute of Science, Rehovot, Israel Oscar Nierstrasz University of Bern, Switzerland C. Pandu Rangan Indian Institute of Technology, Madras, India Bernhard Steffen University of Dortmund, Germany Madhu Sudan Massachusetts Institute of Technology, MA, USA Demetri Terzopoulos University of California, Los Angeles, CA, USA Doug Tygar University of California, Berkeley, CA, USA Gerhard Weikum Max-Planck Institute of Computer Science, Saarbruecken, Germany
José M. Laginha M. Palma Patrick R. Amestoy Michel Daydé Marta Mattoso João Correia Lopes (Eds.) High Performance Computing for Computational Science - VECPAR 2008 8th International Conference Toulouse,, June 24-27, 2008 Revised Selected Papers 13
Volume Editors José M.Laginha M. Palma João Correia Lopes University of Porto, Faculty of Engineering 4200-465 Porto Portugal E-mail: {jpalma, jlopes}@fe.up.pt Patrick R. Amestoy Michel Daydé University of Toulouse, INP (ENSEEIHT), IRIT 31071 Toulouse CEDEX 7, E-mail: {amestoy, michel.dayde}@enseeiht.fr Marta Mattoso Federal University of Rio de Janeiro 21941-972 Rio de Janeiro RJ, Brazil E-mail: marta@cos.ufrj.br Library of Congress Control Number: Applied for CR Subject Classification (1998): D, F, C.2, G, J.2-3 LNCS Sublibrary: SL 1 Theoretical Computer Science and General Issues ISSN 0302-9743 ISBN-10 3-540-92858-8 Springer Berlin Heidelberg New York ISBN-13 978-3-540-92858-4 Springer Berlin Heidelberg New York This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, re-use of illustrations, recitation, broadcasting, reproduction on microfilms or in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its current version, and permission for use must always be obtained from Springer. Violations are liable to prosecution under the German Copyright Law. springer.com Springer-Verlag Berlin Heidelberg 2008 Printed in Germany Typesetting: Camera-ready by author, data conversion by Scientific Publishing Services, Chennai, India Printed on acid-free paper SPIN: 12591596 06/3180 543210
Preface VECPAR is an international conference series dedicated to the promotion and advancement of all aspects of high-performance computing for computational science, as an industrial technique and academic discipline, extending the frontier of both the state of the art and the state of practice. The audience and participants of VECPAR are researchers in academic departments, government laboratories and industrial organizations. There is now a permanent website for the conference series in http://vecpar.fe.up.pt where the history of the conference is described. The 8th edition of VECPAR was organized in Toulouse (), June 24 27, 2008. It was the third time the conference was celebrated outside Porto after Valencia (Spain) in 2004 and Rio de Janeiro (Brazil) in 2006. The conference programme consisted of 6 invited talks and 53 accepted papers out of 73 contributions that were initially submitted. The major themes are divided into: Large-Scale Simulations and Numerical Algorithms in Computer Science and Engineering (aerospace, earth, environment, finance, geoscience) Computing in Healthcare and Biosciences Multiscale and Multiphysics Problems Cluster Computing and Grid Computing Data-Centric and High-Productivity Computing Cooperative Engineering Problem-Solving Environments Imaging and Graphics Two workshops, in addition to tutorials, were organized before the conference: HPDGrid 2008 International Workshop on High-Performance Data Management in Grid Environments on June 24 and Sparse Days at CERFACS on June 23 and 24. The most significant contributions are made available in the present book, edited after the conference and after a second review of all orally presented papers at VECPAR 2008 and at the workshops. Note also that the conference gave the opportunity of highlighting French Japanese cooperations in grid computing and computational science with a session devoted to the REDIMPS Project within the VECPAR conference track and two workshops organized on June 23 and 24: PAAP 1 and NEGST 2. Melissa Paes received the Best Student Presentation award for her talk on High-Performance Query Processing (see corresponding paper in Session Data- Centric and High-Productivity Computing). 1 http://www2.lifl.fr/map/paap/ 2 http://www2.lifl.fr/map/negst/
VI Preface The eighth edition of VECPAR took place in Toulouse at INPT/ENSEEIHT: a school of engineers located in the center of Toulouse. The organizational aspects were dealt with by colleagues from the Institut de Recherche en Informatique de Toulouse (IRIT) with the participation of colleagues from CERFACS and the support of the Institut National Polytechnique de Toulouse / ENSEEIHT, Université Paul Sabatier, Pôle de Compétivité Aerospace Valle and Région Midi-Pyrénées. Paper submission and selection were managed via the conference management system, held and operated by the Faculty of Engineering of the University of Porto (FEUP). Websites were maintained both by FEUP and IRIT. Registrations were managed by IRIT. The success of the VECPAR conferences and its long life are a result of the collaboration of many people. As before, given the widespread organization, a large number of collaborators were involved. We mention only some, and through them we thank many others who offered most of their time and commitment to the success of the conference and workshops: Véronique Debats (IRIT) Sabyne Lartigues (IRIT) Sylvie Soler (INPT/ENSEEIHT) For their contribution to the present book, we must thank all the authors for meeting the deadlines and all members of the Scientific Committee who helped us so much in selecting the papers. We also thanks the members of the committees involved in the organization of the workshops held before the conference. October 2008 José M.L.M. Palma Patrick Amestoy Michel Daydé Marta Mattoso João Correia Lopes VECPAR is a series of conferences organized by the Faculty of Engineering of Porto (FEUP) since 1993.
Organization Organizing Committee M. Daydé IRIT/INPT - ENSEEIHT, (Chair) L. Giraud IRIT/INPT - ENSEEIHT, R. Guivarch IRIT/INPT - ENSEEIHT, D. Hagimont IRIT/INPT - ENSEEIHT, J.-P. Jessel IRIT/UPS, J. Correia Lopes FEUP/INESC Porto, Portugal (Web Chair) Steering Committee J. Palma Portugal (Chair) J. Dongarra USA A. Coutinho Brazil J. Fortes USA V. Hernandez Spain O. Marques USA K. Miura Japan Scientific Committee P. Amestoy (Chair) Jacques Bahi Carlos Balsa Portugal Rob Bisseling The Netherlands Vincent Breton Xiao-Chuan Cai USA Mark A. Christon USA Olivier Coulaud José Cardoso e Cunha Portugal Rudnei Cunha Brazil Michel J. Daydé Frédéric Desprez Philippe Devloo Brazil Marcos Donato Ferreira Brazil Jack Dongarra USA Inês de Castro Dutra Portugal Renato Figueiredo USA Wolfgang Gentzsch Germany Alan D. George USA Omar Ghattas USA
VIII Organization Luc Giraud Serge Gratton Ronan Guivarch Daniel Hagimont Abdelkader Hameurlain Bruce Hendrickson Thomas Hérault Vicente Hernandez Vincent Heuveline Jean-Pierre Jessel Christopher R. Johnson Jacko Koster Dieter Kranzmueller Stéphane Lanteri Xiaoye Sherry Li Kuan-Ching Li Rainald Lohner Thomas Ludwig Osni Marques Armando Jorge Padilha José Laginha Palma Dennis Parsons Serge G. Petiton Thierry Priol Heather Ruskin Mitsuhisa Sato Satoshi Sekiguchi António Augusto Sousa Mark A. Stadtherr Domenico Talia Francisco Tirado Miroslav Tuma Patrick Valduriez Xavier Vasseur Roland Wismuller USA Spain Germany USA Norway Austria USA Taiwan USA Germany USA Portugal Portugal USA Ireland Japan Japan Portugal USA Italy Spain Czech Republic Germany Invited Speakers Nicholas Ayache Philippe Bougeault Jack Dongarra Satoshi Matsuoka Thierry Poinsot Michael A. Puso INRIA, European Centre for Medium-Range Weather Forecasts, UK University of Tennessee, USA Tokyo Institute of Technology, Japan CERFACS, Lawrence Livermore National Laboratory, USA
Organization IX Additional Reviewers Jose Miguel Alonso Fernando Alvarruiz Gabriel Antoniu Eugenio Cesario Laurent Choy Grzegorz Cieslewski Olivier Coulaud A. Esnard P. Henon Benoit Hudzia Martin Quinson Yvon Jégou G. Latu Kuan-Ching Li Rodrigo F. de Mello Christian Perez Giuseppe Pirrò Jose E. Roman Hung-Hsun Su Christian Tenllado Tien-Hsiung Weng Sponsoring Organizations The Organizing Committee is very grateful to the following organizations for their support: AESE Pôle de Compétivité Aerospace Valley CERFACS European Centre for Research and Advanced Training in ENSEEIHT INPT Institut National Polytechmique de Toulouse IRIT Institut de Recherche en Informatique de Toulouse Région Midi-Pyrénées UP Universidade do Porto, Portugal UPS Université Paul Sabatier
HPDgrid 2008 International Workshop on High-Performance Data Management in Grid Environments Initially developed for the scientific community as a generalization of cluster computing using the Web, grid computing is gaining much interest in other important areas such as enterprise information systems. This makes data management more critical than ever. Compared with cluster computing which deals with homogeneous parallel systems, grids are characterized by high heterogeneity, high autonomy and large-scale distribution of computing and data resources. Managing and transparently accessing large numbers of autonomous, heterogeneous data resources efficiently is an open problem. Furthermore, different grids may have different requirements with respect to autonomy, query expressiveness, efficiency, quality of service, fault-tolerance, security, etc. Thus, different solutions need be investigated, ranging from extensions of distributed and parallel computing techniques to more decentralized, self-adaptive techniques such as peer-to-peer (P2P). The objective of the second edition of this one-day workshop was to bring together researchers and practitioners from the high-performance computing, scientific computing, distributed systems and database communities to discuss the challenges and propose novel solutions in the design and implementation of high-performance data management in Grid environments. The following program was the result of the paper selection, with seven papers presented in two sessions: (1) data-intensive grid applications, and (2) replication in grids, grid services and grid data mining. In addition, we had two keynote sessions: the first one on Grid Data Management: Open Problems and New Issues by Esther Pacitti, and the second one on Data Management in Pervasive Grids by Jean-Marc Pierson. All selected papers were reviewed by three members of the Program Committee. We would like to thank them for the excellent work on coming up with a program that covers many different topics related to grid data management. October 2008 Marta Mattoso Alexandre Lima Esther Pacitti Patrick Valduriez
Organization Workshop General Co-chairs Esther Pacitti Alexandre A.B. Lima INRIA and LINA, Unigranrio, Brazil Executive Chair Patrick Valduriez INRIA and LINA, Program Chair Marta Mattoso COPPE/UFRJ, Brazil Program Committee David Abramson Reza Akbarinia Gabriel Antoniu Rajkumar Buyya Marco A. Casanova Inês de Castro Dutra Alexandre Evsukoff Marcio Faerman Alvaro A.A. Fernandes Harald Kosch Nectarios Koziris Vidal Martins Wagner Meira Jr. Philippe Navaux Jean-Marc Pierson Fabio Porto Gerson Sunye Vladimir Vlassov Monash University, Australia University Nantes, IRISA-INRIA, University of Melbourne, Australia PUC-Rio, Brazil University Porto, Portugal UFRJ, Brazil RNP, Brazil University Manchester, UK University Passau, Germany National Technical University Athens, Greece PUC-Parana, Brazil UFMG, Brazil UFRGS, Brazil IRIT, EPFL, Switzerland University Nantes, KTH, Sweden Invited Speakers Esther Pacitti Jean-Marc Pierson INRIA and LINA, UniversitédeNantes, IRIT, Université Paul Sabatier,
XIV Organization Additional Reviewers A.B.M. Russel Alexandre Evsukoff Álvaro A. Fernandes George Tsoukalas Rodrigo Virote Kassick Eduardo Almeida
Table of Contents Parallel and Distributed Computing An Overview of High Performance Computing and Challenges for the Future... 1 Jack Dongarra Parallelization of Sphere-Decoding Methods... 2 Rafael A. Trujillo, Antonio M. Vidal, Víctor M. García, and Alberto González Improving the Performance of a Verified Linear System Solver Using Optimized Libraries and Parallel Computation... 13 Mariana Kolberg, Gerd Bohlender, and Dalcidio Claudio Parallelisation of the CFD Code of a CFD-NWP Coupled System for the Simulation of Atmospheric Flows over Complex Terrain... 27 Fernando A. Castro, Castro M.P. Silva Santos, and José M.L.M. Palma High Performance Computing for Eigenvalue Solver in Density-Matrix Renormalization Group Method: Parallelization of the Hamiltonian Matrix-Vector Multiplication... 39 Susumu Yamada, Masahiko Okumura, and Masahiko Machida Tunable Parallel Experiments in a GridRPC Framework: Application to Linear Solvers... 46 Yves Caniou, Jean-Sébastien Gay, and Pierre Ramet Cluster and Grid Computing The Rise of the Commodity Vectors... 53 Satoshi Matsuoka MOPS A Morphodynamical Prediction System on Cluster Computers... 63 Hartmut Kapitza Implementing a Parallel NetCDF Interface for Seamless Remote I/O Using Multi-dimensional Data... 69 Yuichi Tsujita Vectorized AES Core for High-throughput Secure Environments... 83 Miquel Pericàs, Ricardo Chaves, Georgi N. Gaydadjiev, Stamatis Vassiliadis, and Mateo Valero
XVI Table of Contents A Matrix Inversion Method with YML/OmniRPC on a Large Scale Platform... 95 Maxime Hugues and Serge G. Petiton Can We Connect Existing Production Grids into a World Wide Grid?... 109 Peter Kacsuk, Attila Kertesz, and Tamas Kiss A List Scheduling Algorithm for Scheduling Multi-user Jobs on Clusters... 123 Jorge Barbosa and António P. Monteiro Data Locality Aware Strategy for Two-Phase Collective I/O... 137 Rosa Filgueira, David E. Singh, Juan C. Pichel, Florin Isaila, and Jesús Carretero Problem Solving Environment and Data Centric A Grid-Aware Web Portal with Advanced Service Trading for Linear Algebra Calculations... 150 Hrachya Astsatryan, Vladimir Sahakyan, Yuri Shoukouryan, Michel Daydé, Aurelie Hurault, Marc Pantel, and Eddy Caron Resource Matching in Non-dedicated Multicluster Environments... 160 Josep Lluis Lérida, sc Solsona, sc Giné, Jose Ramon García, and Porfidio Hernández A Parallel Incremental Learning Algorithm for Neural Networks with Fault Tolerance... 174 Jacques M. Bahi, Sylvain Contassot-Vivier, Marc Sauget, and Aurélien Vasseur High-Performance Query Processing of a Real-World OLAP Database with ParGRES... 188 Melissa Paes, Alexandre A.B. Lima, Patrick Valduriez, and Marta Mattoso Improving Search Engines Performance on Multithreading Processors... 201 Carolina Bonacic, Carlos Garcia, Mauricio Marin, Manuel Prieto, Francisco Tirado, and Cesar Vicente Numerical Methods Accomplishments and Challenges in Code Development for Parallel and Multimechanics Simulations... 214 Tony Degroot, Robert Ferencz, Mark Havstad, Neil Hodge, Jerry Lin, Dennis Parsons, Michael Puso, Jerome Solberg, and Edward Zywicz
Table of Contents XVII An Algorithm-by-Blocks for SuperMatrix Band Cholesky Factorization... 228 Gregorio Quintana-Ortí, Enrique S. Quintana-Ortí, Alfredo Remón, and Robert A. van de Geijn An Efficient and Robust Decentralized Algorithm for Detecting the Global Convergence in Asynchronous Iterative Algorithms... 240 Jacques M. Bahi, Sylvain Contassot-Vivier, and Raphaël Couturier A Parallel Implementation of the Trace Minimization Eigensolver... 255 Eloy Romero and Jose E. Roman A Load Balancing Knapsack Algorithm for Parallel Fuzzy c-means Cluster Analysis... 269 Marta V. Modenesi, Alexandre G. Evsukoff, and Myrian C.A. Costa Scalable Parallel 3d FFTs for Electronic Structure Codes... 280 Andrew Canning Linear Algebra Evaluation of Sparse LU Factorization and Triangular Solution on Multicore Platforms... 287 Xiaoye Sherry Li A Parallel Matrix Scaling Algorithm... 301 Patrick R. Amestoy, Iain S. Duff, Daniel Ruiz, and Bora Uçar Design, Tuning and Evaluation of Parallel Multilevel ILU Preconditioners... 314 José I. Aliaga, Matthias Bollhöfer, Alberto F. Martín, and Enrique S. Quintana-Ortí On the I/O Volume in Out-of-Core Multifrontal Methods with a Flexible Allocation Scheme... 328 Emmanuel Agullo, Abdou Guermouche, and Jean-Yves L Excellent Parallel Eigensolvers for a Discretized Radiative Transfer Problem... 336 Paulo B. Vasconcelos, Osni Marques, and Jose E. Roman Computing in Geosciences and Biosciences High Performance Computing and the Progress of Weather and Climate Forecasting... 349 Philippe Bougeault
XVIII Table of Contents Simulation of Seismic Wave Propagation in an Asteroid Based upon an Unstructured MPI Spectral-Element Method: Blocking and Non-blocking Communication Strategies... 350 Roland Martin, Dimitri Komatitsch, Céline Blitz, and Nicolas Le Goff A Simulation of Seismic Wave Propagation at High Resolution in the Inner Core of the Earth on 2166 Processors of MareNostrum... 364 Dimitri Komatitsch, Jesús Labarta, and David Michéa Using a Global Parameter for Gaussian Affinity Matrices in Spectral Clustering... 378 Sandrine Mouysset, Joseph Noailles, and Daniel Ruiz Comparing Some Methods and Preconditioners for Hydropower Plant Flow Simulations... 391 Luiz M. Carvalho, Wagner Fortes, and Luc Giraud Imaging and Graphics Computational Models of the Human Body for Medical Image Analysis... 405 Nicholas Ayache Attaining High Performance in General-Purpose Computations on Current Graphics Processors... 406 Francisco D. Igual, Rafael Mayo, and Enrique S. Quintana-Ortí Optimised Computational Functional Imaging for Arteries... 420 Ramiro Moreno, Ming Chau, Shirod Jeetoo, Franck Nicoud, Frédéric Viart, Anne Salvayre, and Hervé Rousseau Memory Locality Exploitation Strategies for FFT on the CUDA Architecture... 430 Eladio Gutierrez, Sergio Romero, Maria A. Trenas, and Emilio L. Zapata Computing for Aerospace and Engineering Large Eddy Simulation of Combustion on Massively Parallel Machines... 444 Gabriel Staffelbach, Jean Mathieu Senoner, Laurent Gicquel, and Thierry Poinsot Variational Multiscale LES and Hybrid RANS/LES Parallel Simulation of Complex Unsteady Flows... 465 Hilde Ouvrard, Bruno Koobus, Maria-Vittoria Salvetti, Simone Camarri, and Alain Dervieux
Table of Contents XIX Vortex Methods for Massively Parallel Computer Architectures... 479 Philippe Chatelain, Alessandro Curioni, Michael Bergdorf, Diego Rossinelli, Wanda Andreoni, and Petros Koumoutsakos On the Implementation of Boundary Element Engineering Codes on the Cell Broadband Engine... 490 Manoel T.F. Cunha, Jose C.F. Telles, and Alvaro L.G.A. Coutinho High-Performance Data Management in Grid Environments Grid Data Management: Open Problems and New Issues from the P2P Perspective... 505 Esther Pacitti Data Management Concerns in a Pervasive Grid... 506 Jean-Marc Pierson DTR: Distributed Transaction Routing in a Large Scale Network... 521 Idrissa Sarr, Hubert Naacke, and Stéphane Gançarski Distributed Management of Massive Data: An Efficient Fine-Grain Data Access Scheme... 532 Bogdan Nicolae, Gabriel Antoniu, and Luc Bougé BLAST Distributed Execution on Partitioned Databases with Primary Fragments... 544 Daniel Xavier de Sousa, Sergio Lifschitz, and Patrick Valduriez Testing Architectures for Large Scale Systems... 555 Eduardo Cunha de Almeida, Gerson Sunyé, and Patrick Valduriez Data Replication and the Storage Capacity of Data Grids... 567 Silvia Figueira and Tan Trieu Text Mining Grid Services for Multiple Environments... 576 Antonio Anddre Serpa, Valeriana G. Roncero, Myrian C.A. Costa, and Nelson F.F. Ebecken Using Stemming Algorithms on a Grid Environment... 588 Valeriana G. Roncero, Myrian C.A. Costa, and Nelson F.F. Ebecken Author Index... 595