Lecture Notes in Computer Science 4917 CommencedPublicationin1973 FoundingandFormerSeriesEditors: GerhardGoos,JurisHartmanis,andJanvanLeeuwen EditorialBoard DavidHutchison LancasterUniversity,UK TakeoKanade CarnegieMellonUniversity,Pittsburgh,PA,USA JosefKittler UniversityofSurrey,Guildford,UK JonM.Kleinberg CornellUniversity,Ithaca,NY,USA FriedemannMattern ETHZurich,Switzerland JohnC.Mitchell StanfordUniversity,CA,USA MoniNaor WeizmannInstituteofScience,Rehovot,Israel OscarNierstrasz UniversityofBern,Switzerland C.PanduRangan IndianInstituteofTechnology,Madras,India BernhardSteffen UniversityofDortmund,Germany MadhuSudan MassachusettsInstituteofTechnology,MA,USA DemetriTerzopoulos UniversityofCalifornia,LosAngeles,CA,USA DougTygar UniversityofCalifornia,Berkeley,CA,USA MosheY.Vardi RiceUniversity,Houston,TX,USA GerhardWeikum Max-PlanckInstituteofComputerScience,Saarbruecken,Germany Per Stenström Michel Dubois Manolis Katevenis Rajiv Gupta Theo Ungerer (Eds.) High Performance Embedded Architectures and Compilers Third International Conference, HiPEAC 2008 Göteborg, Sweden, January 27-29, 2008 Proceedings 1 3 VolumeEditors PerStenström ChalmersUniversityofTechnology,Dept.ofComputerScienceandEngineering 41296Gothenburg,Sweden E-mail:[email protected] MichelDubois UniversityofSouthernCalifornia,Dept.ofElectricalEngineering 3740McClintockAve.,LosAngeles,CA90089-2562,USA E-mail:[email protected] ManolisKatevenis FORTH-ICS 100PlastiraAve.,VassilikaVouton,70-013Heraklion,Crete,Greece E-mail:[email protected] RajivGupta UniversityofCalifornia,Riverside,Dept.ofComputerScienceandEngineering 408EngineeringBuildingII,Riverside,CA92521,USA E-mail:[email protected] TheoUngerer UniversityofAugsburg,InstituteofComputerScience 86135Augsburg,Germany E-mail:[email protected] LibraryofCongressControlNumber:2007942570 CRSubjectClassification(1998):B.2,C.1,D.3.4,B.5,C.2,D.4 LNCSSublibrary:SL1–TheoreticalComputerScienceandGeneralIssues ISSN 0302-9743 ISBN-10 3-540-77559-5SpringerBerlinHeidelbergNewYork ISBN-13 978-3-540-77559-1SpringerBerlinHeidelbergNewYork Thisworkissubjecttocopyright.Allrightsarereserved,whetherthewholeorpartofthematerialis concerned,specificallytherightsoftranslation,reprinting,re-useofillustrations,recitation,broadcasting, reproductiononmicrofilmsorinanyotherway,andstorageindatabanks.Duplicationofthispublication orpartsthereofispermittedonlyundertheprovisionsoftheGermanCopyrightLawofSeptember9,1965, initscurrentversion,andpermissionforusemustalwaysbeobtainedfromSpringer.Violationsareliable toprosecutionundertheGermanCopyrightLaw. SpringerisapartofSpringerScience+BusinessMedia springer.com ©Springer-VerlagBerlinHeidelberg2008 PrintedinGermany Typesetting:Camera-readybyauthor,dataconversionbyScientificPublishingServices,Chennai,India Printedonacid-freepaper SPIN:12212101 06/3180 543210 Preface The first two editions of the HiPEAC conference series in Barcelona (2005)and Ghent (2007) have really demonstrated that the topics covered by HiPEAC attract a lot of interest. In the 2007 conference, about 200 people attended the conference and its satellite events. The third HiPEAC conference was held in Go¨teborg, the second largest city in Sweden. The offerings of this conference were rich and diverse. We offered attendees a set of four workshops on topics that are all central to the HiPEAC roadmap: multi-cores,compiler optimizations,reconfigurablecomputing,andinterconnec- tion networks. Additionally, a tutorial on the Sunflower Toolsuite was offered. The conference program was as rich as in the last years. It featured many important and timely topics such as multi-core processors, reconfigurable sys- tems, compiler optimization, power-awaretechniques and more. The conference also offered a keynote speech by Mateo Valero – the Eckert-Mauchly Award winner in 2007.Several social events provided opportunities for interaction and exchange of ideas in informal settings such as a tour at the Universeum – a science exhibition center and aquarium – where the banquet took place as well. Thisyearwereceived77submissionsofwhich14paperswereCommitteepa- pers.Papersweresubmitted from22differentnations(about40%fromEurope, 25% from Asia, 30% from North America and 5% from South America), which is a token of the global visibility of the conference. We had the luxury of having a strong Program Committee consisting of 37 experts in allareaswithin the scope of the conference andwe kept allreviewing withintheProgramCommittee.Thus,eachpaperwastypicallyreviewedbyfour ProgramCommittee members. We collected 301 reviews and we were happy to note that each paper was rigorously reviewed before we made the decisions, despite the fact that we shortened the review phase and that reviewing took place during most reviewers’precious vacation time. TheProgramCommitteemeetingwasheldinthecenterofRome,theancient capitalofItaly.DespitealongtripformanymembersoftheProgramCommittee, 16 PC members attended the meeting. For virtually all papers, at least two reviewers were present. The PC meeting was preceded by an e-mail discussion of papers amongthe reviewers.At the PC meeting the papers werediscussed in theorderofaveragescorealsoincludingPCpapers.Whenapaperwasdiscussed where a participating PC member was either a co-author or had conflicts with thatpaper,thatpersonleft the room.We accepted25papersofwhich4 arePC papers, yielding an acceptance rate of 32%. The end result of the whole effort was the high-quality program for the HiPEAC 2008 event. We hope that you learn and get much inspiration from this proceedings volume. VI Preface The planning of a conference starts way ahead of the actual event. If it were not for the unselfish and hard work of a large number of devoted individuals, this conference would not have been as successful as it was. Let us first thank the authors for their great contributions which constitute the core of the con- ference. We were very fortunate to collect a great team to power this event and would like to thank all of them: Mats Brorsson(KTH) for putting together an attractive pre-conference program; Ewa W¨aingelin (Chalmers) for the enor- mous effort she invested in the local arrangement; Per Waborg (Chalmers) for a fantastic job in keeping the costs within budget and running the books; Mike O’Boyle(Edinburgh)fortimelypublicitycampaigns;Jo¨rgMischeandTheoUn- gerer(Augsburg)forthe hardworkinputting togethertheproceedings;Michiel Ronsee(Ghent)foradministeringthesubmissionandreviewsystem;andfinally Sylvie Detournay (Ghent) for administering the web. Thanks to all of you! Finally,wewouldalsoliketomentionthesupportfromtheSixthFramework ProgrammeoftheEuropeanUnion,representedbyprojectofficersMerc`eGriera iFisa andPanagiotisTsarchopoulos,for sponsoringthe eventandfor the travel grants. October 2007 Per Stenstro¨m Michel Dubois Manolis Katevenis Rajiv Gupta Organization Executive Committee General Co-chairs Per Stenstro¨m Chalmers University of Technology, Sweden Michel Dubois University of Southern California, USA ProgramCommittee Co-chairs Manolis Katevenis University of Crete / FORTH, Greece Rajiv Gupta University of California, Riverside, USA Workshop/Tutorials Chair Mats Brorsson KTH, Sweden Local Arrangements Chair Ewa W¨aingelin Chalmers University of Technology, Sweden Finance Chair Per Waborg Chalmers University of Technology, Sweden Publicity Chair Mike O’Boyle University of Edinburgh, UK Publications Chair Theo Ungerer University of Augsburg, Germany Submissions Chair Michiel Ronsse Ghent University, Belgium Web Chair Sylvie Detournay Ghent University, Belgium Program Committee Angelos Bilas FORTH / University of Crete, Greece Mats Brorsson KTH, Sweden Koen De Bosschere University of Ghent, Belgium Jack Davidson University of Virginia, USA Marc Duranton NXP Semiconductors, Netherlands Babak Falsafi Carnegie Mellon University, USA Paolo Faraboschi HP Labs, Spain Kristian Flautner ARM, UK Chris Gniady University of Arizona, USA Wen-mei Hwu University of Illinois, USA Paolo Ienne EPFL, Switzerland Norm Jouppi HP Labs, USA VIII Organization Mahmut Kandemir Pennsylvania State University, USA Stefanos Kaxiras University of Patras, Greece Christos Kozyrakis Stanford University, USA Scott Mahlke University of Michigan, USA Peter Marwedel University of Dortmund, Germany Avi Mendelson Intel, Israel Andreas Moshovos University of Toronto, Canada Mike O’Boyle Edinburgh University, UK Kunle Olukotun Stanford University, USA Yunheung Paek Seoul National Universiry, Korea Santosh Pande Georgia Institute of Technology, USA Yale Patt University of Texas at Austin, USA Alex Ramirez UPC / BSC, Spain Lawrence Rauchwerger Texas A&M University, USA John Regehr University of Utah, USA Andre Seznec INRIA, France Guri Sohi University of Wisconsin, USA Olivier Temam INRIA, France Josep Torrellas University of Illinois, USA Mateo Valero UPC / BSC, Spain David Whalley Florida State University, USA Sudhakar Yalamanchili Georgia Institute of Technology, USA Steering Committee Mateo Valero UPC / BSC, Spain Anant Agarwal MIT, USA Koen De Bosschere Ghent University, Belgium Mike O’Boyle University of Edinburgh, UK Brad Calder University of California, USA Rajiv Gupta University of California, Riverside, USA Wen-mei W. Hwu UIUC, USA Josep Llosa UPC, Spain MargaretMartonosi Princenton University, USA Per Stenstr¨om Chalmers University of Technology, Sweden Olivier Teman INRIA Futurs, France Reviewers Minwook Ahn Mauricio Breternitz Jr. Bruce Childers Mauricio Alvarez Mats Brorsson Doosan Cho Sara Baghsorkhi Felipe Cabarcas Michael Chu Mauro Bianco Francisco Cazorla Romain Cledat Angelos Bilas Luis Ceze Michele Co Organization IX Albert Cohen Chang Joo Lee John Regehr Clark Coleman Jongwon Lee Christopher Rodrigues Adrian Cristal Piotr Lesnicki Shane Ryoo Jack Davidson Yue Li Esther Salami Koen De Bosschere Yongmo Liang Fermin Sanchez Marc Duranton Danny Lynch Oliverio Santana Ayose Falcon Scott Mahlke Andre Seznec Babak Falsafi Peter Marwedel Tim Smith Kevin Fan Mojtaba Mehrara Guri Sohi Paolo Faraboschi Avi Mendelson Jaswanth Sreeram Kristian Flautner Rustam Miftakhutdinov Santhosh Srinath Grigori Fursin Pablo Montesinos Samuel Stone Chris Gniady Miquel Moreto Karin Strauss Daniel Gracia-Perez Andreas Moshovos M. Aater Suleman Brian Greskamp Gilles Mouchard Gabriel Tanase Jason Hiser Veynu Narasiman Olivier Temam Wen-mei Hwu Mike O’Boyle Nathan Thomas Paolo Ienne Kunle Olukotun Abhishek Tiwari Daniel Jimenez Daniel Ortega Josep Torrellas Jose Joao Yunheung Paek Sain-Zee Ueng Norm Jouppi Alex Pajuelo Mateo Valero Mahmut Kandemir Santosh Pande Javier Vera Stefanos Kaxiras Ioannis Papadopoulos David Whalley DaeHo Kim Hyunchul Park Sudhakar Yalamanchili Hokyun Kim Sanghyun Park Sami Yehia Wonsub Kim Yale Patt Jonghee Youn Yongjoo Kim Androniki Pazarloglou Kun Zhang Christos Kozyrakis Maikel Pennings Hongtao Zhong Manjunath Kudlur Antoniu Pop Tushar Kumar Alex Ramirez Christopher Kung Lawrence Rauchwerger Table of Contents Invited Program Supercomputing for the Future, Supercomputing from the Past (Keynote) .................................................. 3 Mateo Valero and Jesu´s Labarta I Multithreaded and Multicore Processors MIPS MT: A Multithreaded RISC Architecture for Embedded Real-Time Processing ............................................ 9 Kevin D. Kissell rMPI: Message Passing on Multicore Processors with On-Chip Interconnect..................................................... 22 James Psota and Anant Agarwal Modeling Multigrain Parallelism on Heterogeneous Multi-core Processors:A Case Study of the Cell BE............................ 38 Filip Blagojevic, Xizhou Feng, Kirk W. Cameron, and Dimitrios S. Nikolopoulos IIa Reconfigurable - ASIP BRAM-LUT Tradeoff on a Polymorphic DES Design ................. 55 Ricardo Chaves, Blagomir Donchev, Georgi Kuzmanov, Leonel Sousa, and Stamatis Vassiliadis Architecture Enhancements for the ADRES Coarse-Grained Reconfigurable Array............................................. 66 Frank Bouwens, Mladen Berekovic, Bjorn De Sutter, and Georgi Gaydadjiev Implementation of an UWB Impulse-Radio Acquisition and Despreading Algorithm on a Low Power ASIP ....................... 82 Jochem Govers, Jos Huisken, Mladen Berekovic, Olivier Rousseaux, Frank Bouwens, Michael de Nil, and Jef Van Meerbergen IIb Compiler Optimizations Fast Bounds Checking Using Debug Register ........................ 99 Tzi-cker Chiueh