Docstoc

Brussels _review_final.pptx - HiPEAC European Network of

Document Sample
Brussels _review_final.pptx - HiPEAC European Network of Powered By Docstoc
					            HiPEAC Review
            Brussels June 25, 2010




HiPEAC:
High-
Performance and
Embedded
Architecture and
Compilation                          Granted: ICT 217068
         Agenda
          09:00-09:30   Private meeting reviewers
          09:30-09:45   Introduction + Agenda
                        Report on the activities of February-May 2010, including
                        -     RC9: Repository of research results
          09:45-10:15   -     RC10: Mini-sabbaticals update
                        -     RC12: Top conferences
                        -     RC13: Mini-sabbaticals: results
          10:15-10:30   Budget overview and planning
Agenda




          10:30-10:45   BREAK
          10:45-11:30   RC1: Future of HiPEAC
          11:30-12:00   RC2: Roadmap update
          12:00-12:15   RC3: A consistent view on the scalability limitations of multicore and
                        manycore systems
          12:15-13:30   LUNCH
                        RC4 : The Project Management should increase HiPEAC’s visibility and the
                        link with Artemis
                        RC5: The project management needs to stimulate much greater and more
          13:30-14:00
                        effective interaction of the research clusters
                        RC6: The Management Report should contain a section giving an overview
                        of related projects and initiatives and their interaction with HiPEAC
          14:00-14:15   RC7: A more systematic approach for analyzing real applications
          14:15-14:45   RC8: The HiPEAC Journal plans
          14:45-15:00   RC11: HiPEAC Industrial Workshop
          15:00-15:30   Reviewers Private Meeting
          15:30-16:00   Feedback
                                                                                                   2
                          Project Status Overview
                             Update since January, 2010
Project Status Overview




                                                           3
                   Change of project manager
WP4: Management




                  Wouter De Raeve                           Jeroen Ongenae
                         Transition period between June 1 and September 1

                                                                             4
WP4: Management
                  Partner change
                Internships
                   Call for topics: Feb 3, 2010
                   Deadline for application: April 14, 2010
                    20 topics, 79 applications
WP1: Mobility



                


                                       Defines topic
                             HiPEAC                                   HiPEAC
                             Company                                  student
                                       Applies to call


                                                     Company chooses candidate



                                                     Approval by SC


                                        Internship




                                                                                 6
                    12 Internships
                   ARM
                    ◦ Improving system energy consumption through optimizing memory usage (Georgia
                      Kouveli)
                    ◦ Power and performance modelling of highly scalable ARM designs (Manuel Comparetti)
                    ◦ Many core exploration on FPGA (Jaume Joven)
WP1: Mobility



                    ◦ Fault-tolerant and Radiation-hardened Processor Design (Massoud Mokhtarpour
                      Ghahroodi)
                    ◦ Scalable Interconnection Network and Cache Coherence using AMBA bus interface
                      (Crispin Gomez Requena)
                    ◦ Optimising and Generating OpenCL code for an Embedded GPU Architecture (Richard
                      Membarth and Dominik Grewe)
                   IBM
                    ◦    Performance tools (Vladimir Čakarević)
                    ◦    Parallel debugging (Lois Orosa Nogueira)
                    ◦    Machine Learning to aid Dynamic Compilation (Zheng Wang)
                    ◦    Split Compilation - synergy between static and dynamic analysis and compilation (Per
                         Larsen)
                   ST
                    ◦ Definition of a L1-data cache prefetch mechanism in a context of "many-core" (J. Lattore)



                                                                                                                7
                Collaboration Grants
                            HiPEAC        Student and host          HiPEAC
                           Institution   prepare a proposal         student
WP1: Mobility



                                                      Ranking by independent reviewers


                                                      Approval by SC

                                           Collaboration
                                               Grant




                 Deadline for application: May 18, 2010
                 34 proposals submitted (27 in 2009)
                 25 granted

                                                                                         8
                Collaboration grants
                Student                        Affiliation                                  Host
                alastair donaldson             University of Oxford                         Albert COHEN
                mohammad reza kakoee           University of Bologna                        Valeria Bertacco
                fabio cancare                  Politecnico di Milano                        Alfonso Jaramillo
                ruben titos-gil                University of Murcia                         Per STENSTROM
                jose ayala                     University Complutenese de Madrid            David ATIENZA
                daniel sanchez                 University of Murcia                         Yiannakis SAZEIDES
WP1: Mobility



                jose maria canales             University of Murcia                         Andy NISBET
                jesús escudero                 University of Castilla-La Mancha             Tor SKEIE
                nehir sonmez                   Barcelona Supercomputing Center              Arvind
                francisco rincon               University Complutense de Madrid             David ATIENZA
                hector pettenghi               INESC-ID                                     Sorin COTOFANA
                juan manuel cebrian gonzalez   University of Murcia                         Stefanos KAXIRAS
                daniel piso-fernández          University of Santiago de Compostela         Kris Gaj
                gokcen kestor                  Barcelona Supercomputing Center              Prof. Michael L. Scott
                daniele ludovici               Delft University of Technology               Davide BERTOZZI
                wei lin guay                   Simula Research Laboratory                   Lars Paul Huse
                mauricio alvarez mesa          Universitat Politècnica de Catalunya (UPC)   Ben JUURLINK
                jordà polo                     Barcelona Supercomputing Center              Malgorzata (Gosia) Steinder
                javier navaridas               The University of the Basque Country         Steve FURBER
                francesco paterna              University of Bologna                        flamand ERIC
                tiago dias                     INESC-ID                                     Sebastian LOPEZ
                alessandro morari              Barcelona Supercomputing Center              Robert WISNIEWSKI
                karthick parashar              INRIA                                        Paolo IENNE
                victor jimenez                 Barcelona Supercomputing Center              Pradip BOSE
                Roberto Goiosa                 Barcelona Supercomputing Center              Michael L. Scott

                                                                                                                          9
                New instrument:
                Conference Grants
                 Stimulate attendance of European students at
                  international conferences which are held in Europe
WP1: Mobility



                 Currently selected conferences:
                    ◦ ISCA 2010, Rennes, André Seznec (GC)
                    ◦ PACT 2010,Vienna, Jens Knoop (PC)
                    ◦ CGO 2011, Chamonix, Olivier Temam (GC) + Mike
                      O’Boyle (PC)
                   ISCA 2010:
                    ◦ Call May 7, deadline for application May 14
                    ◦ 43 applications, 27 accepted
                    ◦ Budget: 14k€

                                                                       10
                            HiPEAC Newsletter
WP3: Spreading Excellence



                             Info 22 - April 2010 (837)   Info 23 – July 2010 (1000)




                                                                                       11
                              Membership
                                  New members (since January 2010)
WP3: Spreading Excellence



                            Member                   Affiliation                       Country
                            Adam Morawiec            ECSI                              France
                            Stefaan Sonck Thiebaut   Opensynergy                       Germany
                            Cyriel Minkenberg        IBM                               Switzerland
                            Mario Nemirovsky         BSC                               Spain
                            Jose Angel Gregorio      University of Cantabria           Spain
                            Valentin Puente Varona   University of Cantabria           Spain
                            Steven Hand              Citrix Systems                    UK
                            Andrew Richards          Codeplay                          UK
                            Christos Bouganis        Imperial College                  UK
                            Mark Zwolinski           University of Southampton         UK
                            Steven Boucque           Alcatel Lucent                    Belgium
                            Zhi hui (Wisdom)         HuaWei Technologies Sweden AB     Sweden
                            Vassos Soteriou          Cyprus University of Technology   Cyprus
                            Sami Khawam              RICAtek                           UK
                            Tom Kean                 Algotronix Ltd                    UK
                            Paolo Palazzari          Ylichron SRL                      Italy
                            Jeremy Bennett           Embecosm                          UK            12
                            ACACES Summer School
                               La Mola, Terrassa, Spain
                                July 11-17, 2010
WP3: Spreading Excellence




                            
                               294 applications, 197 participants
                               Summer school on advanced computer
                                architecture and compilers for high-
                                performance and embedded systems




                                                                       13
                            ACACES Summer School
                               17 industry attendants, 8 companies
                                140 PhD students
WP3: Spreading Excellence




                            
                               59 senior attendants
                               31 countries
                               60 grants (181 requests)




                                                                      14
                             ACACES 2010 courses
                            Instructor         Affiliation              Title
                            Andreas                                     Application-Specific (MP)SoC Architectures for Internet
                                               TU Muenchen
WP3: Spreading Excellence



                            Herkersdorf                                 Networking
                            Michael Scott      University of Rochester Transactional Memory
                                                                        Multicore Programming Models and their Compilation
                            Vivek Sarkar       Rice University
                                                                        Challenges
                            David Brooks       Harvard University       Variation-Aware Processor Design
                                               University of Texas at
                            Derek Chiou                                 Fast and Accurate Computer System Simulators
                                               Austin
                            Scott Mahlke       University of Michigan   Compilation for Multicore Processors
                            Dan Sorin          Duke University          Fault Tolerant Computer Architecture
                            Donatella Sciuto   Politecnico di Milano    FPGA-based reconfigurable computing
                            Steven Hand        Citrix                   System Virtualization
                            Theodore Ts'o      Google                   File Systems and Storage Technologies
                                               Pennsylvania State
                            Mahmut Kandemir                             Embedded Systems: A Software Perspective
                                               University
                            Andrzej Brud and
                                               Chalmers                 How to transform research results into a business
                            Per Stenström


                                                                                                                              15
                            Keynotes
WP3: Spreading Excellence




                            Insup Lee        Jesus Labarta
                            University of    Technical University of
                            Pennsylvania     Catalonia (UPC)

                            Cyber Physical   MareIncognito:
                            Systems:         a vision towards
                            The Next         exascale
                            Computing
                            Revolution
                                                                       16
PUMPS 2010 @ BCW @
BSC




                     17
                                HiPEAC 2011, Heraklion
WP3: Spreading Excellence




                               On schedule
                               86 submissions (last year: 97 submissions)
                               5 workshops, 2 tutorials, probably tutorials on Saturday
                                and workshops on Sunday.
                               Plan for a IEEE Micro special issue workshop
                                                                                           18
                            HiPEAC booth at DATE’10
                               Dresden, Germany
WP3: Spreading Excellence




                               March 09-11, 2010
                               Over 60 booth
                                visitors from
                                academia and
                                industry




                                                      19
                                HiPEAC booth at DAC’10
                               Anaheim, USA
WP3: Spreading Excellence




                               June 4-6, 2010
                               Over 73 booth
                                visitors from
                                academia and
                                industry
     WP3: Spreading Excellence




21
                            Towards hipeac.pl
                               Intend expansion of HiPEAC into new EU
                                member states
WP3: Spreading Excellence




                               E.g. major HiPEAC meeting Oct 2009 in Wroclaw
                               Already several Polish members (e.g. from Poznan,
                                Krakow)
                               Basis for future joint FP7/FP8 projects
                               HiPEAC special session at MIXDES, Wroclaw, June
                                24-26
                               Web portal to be set up soon
                               Contact persons: R. Leupers, Z. Chamski, Świercz

                                                                                22
                    HiPEAC Awards
                    It is necessary to improve the scientific excellence
                    of the network by publishing more papers at
Recommendation 12



                    top conferences. The identification of the
                    best conferences with low numbers of
                    European papers nd the award scheme to
                    motivate more European paper submissions are
                    positive steps. The review panel looks forward to
                    the updated figures of accepted papers at the next
                    review meeting.




                                                                           23
                    HiPEAC Awards
                           Conference              2008     2009   2010
                              POPL                     0     0      0
Recommendation 12


                               PLDI                    1     1      1
                             ASPLOS                    1     2      3
                               ISCA                    2     3      2
                              HPCA                     2     0      4
                              FCCM                     2     6      6
                               DAC                     5     4      4
                             MICRO                     2     6     n/a
                     TOTAL (excl. MICRO)               13   16     20
                             TOTAL                     15   22     n/a

                       Financial awards (excl. Micro):
                        ◦ 2008: 8; 2009: 13; 2010: 9

                       Conference travel grants – 20 k€/year             24
                       Mini-Sabbaticals
                       It is necessary to improve the mini-sabbatical programme in order to achieve 5-10
                       sabbaticals per year (at the moment only 2 mini-sabbatical grants have been
                       awarded) or to draw some conclusions if this is not possible. The mini-sabbaticals
Recommendation 10/13



                       remain a weak point of the project. Only two mini-sabbaticals have been applied for.
                       Results from the sabbatical (Leiden to Sofia) should be presented at the formal
                       review.

                          Sabbaticals done:
                           ◦ 2009:
                                 Prof. Rainer Leupers (Aachen) to ACE
                                 Prof. Enrique Torre (Unizar) to U. Berkeley

                           ◦ 2010:
                                 Prof. Ed Deprettere (Leiden) at Technical University Sofia, March 13 – June 12, 2010
                                 Prof. Stefano Crespi (Milano) at Harvard University, 2-17 June 2010

                          Sabbaticals approved:
                           ◦ 2010:
                                 Prof. Sid Touati (University of Versailles) to Technische Universität München and Leibniz Supercomputing Centre
                                  of the Bavarian Academy of Science, September 2010
                                 Prof. Benjamin Sahelices (Unizar) to University of Illinois at Urbana-Champaign, September 2010
                                 Prof. Jose Miguel-Alonso (University of Cantabria) to University of Manchester, October 2010
                                 [Olivier Temam (INRIA) to China]
                                                                                                                                                    25
                           Mini-Sabbaticals
                       Prof. Ed Deprettere (Univ. Leiden) at Technical University Sofia
                          March 13 – June 12, 2010
Recommendation 10/13



                          Previous contact: the Leiden Embedded Research Center established a
                           new joint education and research embedded systems laboratory at
                           campus of the Technical University Sofia, called the Daedalus
                           Laboratory. (Prof. Marin Marinov and Prof. Angel Popov)
                          The Daedalus laboratory was officially opened on April 7, 2010. It will
                           allow Bulgarian students at The Netherlands to come back home with
                           a job.
                          During the mini-sabbatical stay at Sofia, a 4-week course called
                           “Advanced Design Methods, Techniques, and Tools for Multi-Processor
                           Embedded Systems” has been taught and will be a regular course at
                           Sofia starting 2011.
                          Definition of several Masters Projects for some of the Sofia students
                          Future joint research projects: send a proposal to one of the coming
                           European call for projects.
                                                                                                     26
                       HiPEAC Research results
                       Building a repository of the research results of the HiPEAC
                       community is a very good idea. Progress on this should be monitored
Recommendation 9


                       at the next review.

                       HiPEAC technical reports are not just a repository of research results. It
                        is an easy access tool that allows HiPEAC members to timestamp early
                        ideas and when needed get a feedback on them.
                       Between the pre-review and now the following can be reported:
                        ◦ six new technical reports have been submitted;
                        ◦ Plans to disseminate the technical reports outside the HiPEAC community have been
                          worked out;
                        ◦ The web interface is expanded and is currently under test;
                        ◦ HiPEAC tech reports will be advertised widely on the following mailing lists:
                          micro_publicity, SOCinfo, fpga-list, IEEE-TCSC, hpc-announce, Embedded_news and
                          more;
                        ◦ We will keep monitoring the usage of the time stamping service and promote it within
                          HiPEAC and outside of our network.

                                                                                                            27
                   Tools
                   Tool           Description                                 Institution
                   ArchExplorer   Web infrastructure that aims at facilitating a INRIA,
                                  fair quantitative comparison of architectural UGENT
Recommendation 9


                                  ideas
                   ATMI           Analytical model     of   temperature     in IRISA
                                  microprocessors
                   Collective     Collection of open-source programs with INRIA
                   Benchmark      multiple datasets assembled by the
                                  community to enable realistic benchmarking,
                   Collective     Collaborative repository to share, reuse and INRIA
                   Optimization   reference useful/profitable optimization
                   Database       cases from the community including
                                  compiler optimizations and architecture
                                  configurations to improve code execution
                                  time, code and architecture size, compilation
                                  time, power consumption among others.
                                                                                            28
                   Tool           Description                                     Institution
                   Continuous     Collaborative modular plugin-enabled R&D INRIA
                   Collective     infrastructure to automate program and
                   Compilation    architecture optimizations using empirical,
                   Framework      iterative, statistical and machine learning
Recommendation 9


                                  techniques
                   CPPC           ComPiler for Portable Checkpointing: a University             of
                                  checkpointing tool focused on the insertion of A Coruña
                                  fault tolerance into long-running message-
                                  passing applications
                   cTuning CC     Free, open source compiler collection           INRIA
                   Diablo         Retargetable    link-time   binary      rewriting UGENT
                                  framework
                   GAUT: From C   Academic and open source High-Level Synthesis Université de
                   to RTL         (HLS) tool dedicated to Digital Signal Processing Bretagne-Sud
                                  DSP applications
                   HD-VideoBench Benchmark devoted to High Definition (HD) UPC
                                 digital video processing
                                                                                                29
                   Tool                  Description                                 Institution

                   MAMPS - Multi-      Tool that allows to generate and program TU Eindhoven
                   Application Multi-  multi-processor designs targeting Xilinx
                   Processor Synthesis FPGAs in seconds.
Recommendation 9


                   MICA Microarch.-      Pin tool which allows the user to collect a UGent
                   Independent           number of program characteristics to
                   Characterization of   quantify runtime program behavior.
                   Applications.
                   MILEPOST GCC          Collaborative plugin infrastructure intended University of
                                         to transform popular, stable, production- Edinburgh
                                         quality GCC into a powerful R&D tool.
                   PoCC, the           Source-to-source platform to quickly INRIA
                   Polyhedral Compiler prototype     and     develop    loop nest
                   Collection          optimizations in the polyhedral model
                   SoCLib                System-level virtual prototyping of MPSoC Uni. Pierre &
                                         platforms                                 Marie Curie,
                                                                                   CNRS
                                                                                                   30
         Agenda
          09:00-09:30   Private meeting reviewers
          09:30-09:45   Introduction + Agenda
                        Report on the activities of February-May 2010, including
                        -     RC9: Repository of research results
          09:45-10:15   -     RC10: Mini-sabbaticals update
                        -     RC12: Top conferences
                        -     RC13: Mini-sabbaticals: results
          10:15-10:30   Budget overview and planning
Agenda




          10:30-10:45   BREAK
          10:45-11:30   RC1: Future of HiPEAC
          11:30-12:00   RC2: Roadmap update
          12:00-12:15   RC3: A consistent view on the scalability limitations of multicore and
                        manycore systems
          12:15-13:30   LUNCH
                        RC4 : The Project Management should increase HiPEAC’s visibility and the
                        link with Artemis
                        RC5: The project management needs to stimulate much greater and more
          13:30-14:00
                        effective interaction of the research clusters
                        RC6: The Management Report should contain a section giving an overview
                        of related projects and initiatives and their interaction with HiPEAC
          14:00-14:15   RC7: A more systematic approach for analyzing real applications
          14:15-14:45   RC8: The HiPEAC Journal plans
          14:45-15:00   RC11: HiPEAC Industrial Workshop
          15:00-15:30   Reviewers Private Meeting
          15:30-16:00   Feedback
                                                                                                   31
         Budget & Planning

            WP1: Mobility

             WP3: Spreading Excellence
Budget




         


            WP2: Partner’s budgets

            WP1+2: Cluster Money


                                         32
         2500000


                                     5.8%
         2000000                     7.7%
                                                 29% underspending
                                     15.4%

         1500000
                                             Excellence
Budget




                                             Mobility
         1000000                             Clusters
                               71%
                                             Partners
                         85%                 Spent
          500000

                   50%

               0
              WP1: Mobility
                                                                    Mobility
                                        700,000
                                        600,000
                                        500,000
                                        400,000
                                                                                               DOW
                                        300,000
                                                                                               Spent
Budget




                                        200,000
                                        100,000
                                                    0
                                                          2008 2009 2010 2011



                                         2008                                          2009                                  2010

              Mobility          Dow     Justified       Remaining     Dow     Budget      Justified    Remaining    Dow     Budget     Remaining

         Collaboration Grants   70000      0             70000       70000    140000      93476,19     46523,81    70000    116523,8   116523,8

         Company Internships      0        0               0         80000     80000          22254     57746      80000    137746      137746

           Mini-sabbaticals     10000      0             10000       20000     30000      7770,84      22229,16    30000    52229,16   44029,16

                Total:          80000      0             80000       170000   250000      123501        126499     180000   306499      298299
                 WP3: Spreading Excellence
                                                    Spreading Excellence
                                        700,000
                                        600,000
                                        500,000
                                        400,000
                                                                                               DOW
                                        300,000
                                                                                               Spent
                                        200,000
Budget




                                        100,000
                                                    0
                                                          2008 2009 2010 2011

                                         2008                                           2009                                     2010
         Spreading excellence   Dow     Justified       Remaining    Dow     Budget            Justified   Remaining    Dow     Budget     Remaining
             Conference         20000      0             20000      20000     40000             20000       20000      20000     40000      20000
         HiPEAC tech reports     0         0               0          0         0                34,7        -34,7       0       -34,7       -34,7
         Industrial workshops    0         0               0          0         0                 0           0          0         0          0
               Journal           0         0               0          0         0                 0           0          0         0          0
             Newsletter         7500    4202,63          3297,37    15000    18297,37          16578,78     1718,59    15000    16718,59   16718,59
              Roadmap            0      1257,98         -1257,98      0      -1257,98          7259,99     -8517,97      0      -8517,97   -8517,97
            Summerschool         0         0               0        100000   100000            105453,7    -5453,72    100000   94546,28      0
            Web seminars         0         0               0          0         0                 0           0          0         0          0
               Website           0         0               0          0         0                 0           0          0         0          0
                Total:          27500   5460,61         22039,39    135000   157039,4          149327,2     7712,2     135000   142712,2   28165,92
              WP2: Partner’s budget
                                                             Partners
                                         5000000

                                         4000000

                                         3000000
                                                                                     DOW
                                         2000000
                                                                                     Spent
                                         1000000
Budget




                                                 0
                                                        2008 2009 2010 2011

                                  2008                                   2009                                   2010
          Partners      Dow     Justified   Remaining      Dow      Budget   Justified   Remaining    Dow       Budget   Remaining
           UGent       142000    54359        87641       209000    296641    218891        77750    209000     286750     286750
          Aachen       85000     89639        -4639       85000     80361     103756       -23395    85000      61605      61605
             BSC       80000     97104       -17104       85000     67896      61685         6211    85000      91211      91211
         Chalmers      80000     60963        19037       85000     104037     96232         7805    85000      92805      92805
          TU Delft     80000     20099        59901       85000     144901     23141       121760    85000      206760     206760
         Edinburgh     80000     11933        68067       85000     153067     55272        97795    85000      182795     182795
           FORTH       80000     23165        56835       85000     141835    156734       -14899    85000      70101      70101
            INRIA      40000     24115        15885       85000     100885     38843        62042    85000      147042     147042
             ARM         0        3646        -3646        5000      1354       1877         -523     5000       4477       4477
             IBM         0        1789        -1789        5000      3211       8091        -4880     5000       120        120
             NXP         0        2299        -2299        5000      2701      589,8       2111,2     5000      7111,2     7111,2
              ST         0          0           0          5000      5000     1542,43      3457,57    5000     8457,57    8457,57
           Total:      667000    389111      277889       824000   1101889   766654,2     335234,8   824000    1159235    1159235

         Grand total   874500   432840,16 441659,84      1304000   1745660   1108429     637231,2    1314000   1951231   1793033
         Partner’s details
         250000



         200000



         150000
Budget




                             DOW2008
                             Spent 2008
         100000
                             DOW 2009"
                             spent 2009

         50000



             0
                WP1+2: Cluster Money
                                                                     Clusters
                                                700000
                                                600000
                                                500000
                                                400000
                                                                                               DOW
                                                300000
                                                                                               Spent
                                                200000
                                                100000
Budget




                                                     0
                                                               2008 2009 2010 2011

                                                                 2008                             2009                              2010
                    Cluster operation                   Dow    Justified Remaining Dow      Budget Justified   Remaining Dow       Budget    Remaining
                         General                       19000   16332,87 2667,13    31000   33667,13 27582,97     6084,16  40000   46084,16    46084,16
                  Adaptive compilation                  9000    1022,63 7977,37    16000   23977,37 2251,29     21726,08 15000    36726,08    35426,08
           Binary translation and virtualization        9000    1255,87 7744,13    16000   23744,13 1129,91     22614,22 15000    37614,22    34598,73
                  Compilation platform                  9000    3480,49 5519,51    16000   21519,51 4779,15     16740,36 15000    31740,36    30085,67
              Design methodology and tools              9000    9193,93 -193,93    16000   15806,07 5309,7      10496,37 15000    25496,37    22296,37
                      Interconnects                     9000    297,19    8702,81  16000   24702,81 6658,32     18044,49 15000    33044,49    32341,59
          Modeling and Simulation Framework             9000    486,23    8513,77  16000   24513,77 3906,36     20607,41 15000    35607,41    35607,41
                 Multi-core architecture                9000    2843,54 6156,46    16000   22156,46 4669,17     17487,29 15000    32487,29    29587,29
                  Programming models                    9000       0       9000    16000     25000    999,05    24000,95 15000    39000,95    19679,95
               Reconfigurable computing                 9000    2924,34 6075,66    16000   22075,66 6923,63     15152,03 15000    30152,03    26794,03
                  Task Force Low Power                    0        0         0       0         0        0           0       0         0           0
               Task Force on Applications                 0        0         0       0         0        0           0       0         0           0
          Task Force on Education and Training            0     431,46    -431,46    0      -431,46     0        -431,46    0      -431,46     -431,46
         Task Force on Reliability and Availability       0        0         0       0         0     4736,63    -4736,63    0     -4736,63    -4736,63
                           Total:                     100000   38268,55 61731,45 175000    236731,5 68946,18    167785,3 175000   342785,3    307333,2
         Agenda
          09:00-09:30   Private meeting reviewers
          09:30-09:45   Introduction + Agenda
                        Report on the activities of February-May 2010, including
                        -     RC9: Repository of research results
          09:45-10:15   -     RC10: Mini-sabbaticals update
                        -     RC12: Top conferences
                        -     RC13: Mini-sabbaticals: results
          10:15-10:30   Budget overview and planning
Agenda




          10:30-10:45   BREAK
          10:45-11:30   RC1: European Advanced Computing Institute
          11:30-12:00   RC2: Roadmap update
          12:00-12:15   RC3: A consistent view on the scalability limitations of multicore and
                        manycore systems
          12:15-13:30   LUNCH
                        RC4 : The Project Management should increase HiPEAC’s visibility and the
                        link with Artemis
                        RC5: The project management needs to stimulate much greater and more
          13:30-14:00
                        effective interaction of the research clusters
                        RC6: The Management Report should contain a section giving an overview
                        of related projects and initiatives and their interaction with HiPEAC
          14:00-14:15   RC7: A more systematic approach for analyzing real applications
          14:15-14:45   RC8: The HiPEAC Journal plans
          14:45-15:00   RC11: HiPEAC Industrial Workshop
          15:00-15:30   Reviewers Private Meeting
          15:30-16:00   Feedback
                                                                                                   39
         Agenda
          09:00-09:30   Private meeting reviewers
          09:30-09:45   Introduction + Agenda
                        Report on the activities of February-May 2010, including
                        -     RC9: Repository of research results
          09:45-10:15   -     RC10: Mini-sabbaticals update
                        -     RC12: Top conferences
                        -     RC13: Mini-sabbaticals: results
          10:15-10:30   Budget overview and planning
Agenda




          10:30-10:45   BREAK
          10:45-11:30   RC1: Future of HiPEAC
          11:30-12:00   RC2: Roadmap update
          12:00-12:15   RC3: A consistent view on the scalability limitations of multicore and
                        manycore systems
          12:15-13:30   LUNCH
                        RC4 : The Project Management should increase HiPEAC’s visibility and the
                        link with Artemis
                        RC5: The project management needs to stimulate much greater and more
          13:30-14:00
                        effective interaction of the research clusters
                        RC6: The Management Report should contain a section giving an overview
                        of related projects and initiatives and their interaction with HiPEAC
          14:00-14:15   RC7: A more systematic approach for analyzing real applications
          14:15-14:45   RC8: The HiPEAC Journal plans
          14:45-15:00   RC11: HiPEAC Industrial Workshop
          15:00-15:30   Reviewers Private Meeting
          15:30-16:00   Feedback
                                                                                                   40
                   European Advanced
                   Computing Institute
                   HiPEAC should investigate the feasibility of setting a
Recommendation 1



                   permanent structure with pan-European outreach to
                   promote research in Computing Systems via a type of
                   "European Advanced Computing Institute". Similar
                   structures in other research areas should be identified
                   and analyzed.The first results of this analysis were presented at
                   the informal review.The development of this thinking should be
                   presented at the next formal review.




                                                                                       41
       Situation
                                                                                         FP8
                                           FP7

        HiPEAC1
                                  HiPEAC2                            HiPEAC?

2005    2006    2007      2008    2009     2010     2011     2012    2013       2014   2015




         HRM1                    HRM2




                    Braga           Brussels
                  document         document




            FP7 Call 1:      FP7 Call 4:       FP7 Call 7:      FP7-8 Call ?:
             20+5 M€           25 M€            41+4 M€           ??? M€
                                                                                              42
                   Process
                      We consulted the HiPEAC community on the future of
Recommendation 1


                       HiPEAC.
                      Brainstorm meeting on April 6, 2010 in Brussels Airport.
                       Attendees were: Koen De Bosschere, Marc Duranton, Babak
                       Falsafi, Rainer Leupers, Nacho Navarro, Mike O'Boyle, Yanos
                       Sazeides, Per Stenström, Jürgen Teich, Mateo Valero, Sami
                       Yehia.
                      SC discussion on May 2 and 3, 2010 in Edinburgh.
                      Discussion with project officer Panos Tsarchopoulos on May
                       4, 2010.
                      Conclusions presented at review on June 25, 2010

                                                                                     43
                      What could a network of excellence like HiPEAC do in order to
                       further increase the impact/visibility of our community up to the
                       level of our American peers? What would be our most important action
                       towards that goal?
                       ◦ Let’s not always look across the Atlantic
                             It is today much better than 15 years ago!!
                             Let’s take the lead where we are stronger (embedded)
Recommendation 1


                             Let’s spread our own excellence first
                             Let’s make EU/HiPEAC a stronger brand
                             Let’s make EU/HiPEAC more prestigious
                             If we can’t beat them, let’s join them; Let’s organize events in the US
                       ◦ Let’s change the publication culture to a journal culture. A journal culture has
                         more benefits for non-native speakers. Change the review system.
                       ◦ Let’s teach our undergraduate, graduates and PhD students how to write good
                         papers
                       ◦ Let’s improve our conference. Let’s make HiPEAC no 2 conference in computer
                         architecture. DATE is no 2 in design automation after DAC. Eurosys is on its way
                         to become no 2 in systems. Hence, it is doable.
                       ◦ Let’s create visibility in existing venues
                           Submit to best conferences
                           Special issues of journals – e.g. IEEE Micro
                           Get in touch with key players in the US
                       ◦ Let’s foster excellence
                           There should be enough internal competition for resources
                                                                                                            44
                           Be serious about mobility – also with the US
                      EU funding now happens through instruments like STREPs, IPs, CSAs
                       and NoEs. Are these the right instruments to have maximal impact
                       in our community? Do you know of other instruments that would be
                       better suited? Could the existing instruments be optimized? If so, how?
                       ◦ Current instruments are fine, except for the administrative overhead
                       ◦ More focused funding
Recommendation 1


                       ◦ More small and easy to get grants
                       ◦ Single PI grants (beyond ERC)
                       ◦ Long term investment must be in people, not in a consortium.
                       ◦ More money for FET-open
                       ◦ There is little difference between STREPS and IPs in our community. It would be better to have
                         two types of STREPS and no more IPs.
                          Long term STREPs – coordinated by academics and with industry in an advisory role
                          Shorter term STREPs – coordinated by industry.
                       ◦ New instrument
                          Companies state research challenges
                          Academic groups can propose projects
                          Companies have an advisory role
                          At the end of the project, companies can have funds for 3-6 months of technology transfer
                          Academics should never be forced to adapt their tools to industry in a research project.
                          Impact requires bridging the research-product gap and this is not feasible. I would suggest
                           some support for early development activities, i.e. during the stage that results are being
                           evaluated by industry.There are examples of this happening in some national partners in the
                           EU; Austria is particularly strong. However, perhaps there could be a European dimension to
                           this.                                                                                          45
                      One day, the funding for networks of excellence will stop. How do
                       you see the future of the HiPEAC community after that date? Which
                       activities would you definitely like to see continued, and would you be
                       willing to pay for? Which organizational form would be most appropriate for
                       this new network (non-for profit organization, branch of IEEE/ACM, …)?
                       ◦ No funding, no network
Recommendation 1


                       ◦ Conference could go on
                       ◦ ACACES, Journal
                       ◦ Cluster meetings through CSA/COST
                       ◦ Don’t care, other networks will emerge
                       ◦ Success will critically depend on volunteers




                                                                                                     46
HiPEAC SWOT analysis


     Strengths     Weaknesses




   Opportunities    Threats



                                47
HiPEAC Strengths

       • HiPEAC unites a large community
       • HiPEAC influences the European
         research
    Strengths agenda through the definition
                                  Weaknesses
         of future calls
       • Many new STREPS with HiPEAC
         members have been initiated
       • The HiPEAC conference is doing well
       • The ACACES summer school is doing
         well
       • The roadmap has impact in our
   Opportunities
         community                   Threats



                                               48
HiPEAC Weaknesses

       • The current cluster structure is too static and
         does not reflect the challenges in our domain.
       • It is difficult to determine the effect of all
     Strengths                            Weaknesses
         networking events in terms of actual
         collaborations. Collaborative research should be
         revitalized.
       • HiPEAC is missing appealing grand challenges, like
         Robocup, Darpa Urban Challenge, …
       • In comparison with other communities, we are
         small (especially our industrial base).
       • We have a weak position in the new member
         states.
       • We have not been successful in ERC so far. Even if
   Opportunities                              Threats
         successful, there is no guarantee that an ERC grant
         will tackle the HiPEAC challenges.




                                                               49
HiPEAC Weaknesses

       • Link between academia and industry remains very
         weak in our community.
             • Many academic research results are not
     Strengths                             Weaknesses
               (considered) relevant for the European
               companies.
             • Companies not eager to share their
               tools/methodologies with academia
             • Companies are increasingly focusing on
               short term research. Universities show little
               interest in this kind of research.
             • We lack an industrial board-level network
               (CTO, CEO, …).
             • Solution? More mobility between industry
   Opportunities                               Threats
               and academia? Funding for technology
               transfer? Industry partner program?
               Industrial board? Listen more to industry to
               learn more about their needs?



                                                               50
HiPEAC Weaknesses

       • The lack of open source tools in the computing
         systems domain, lack of affordable CAD tools and
         validation platforms
     Strengths                            Weaknesses
           • Hardware research is not so easy:
                   • Free CAD tools not widely available
                   • FPGA platforms very expensive
                   • CMP and other IC shuttle not so easy
                     (and not for individuals)
           • Solution? creating a “hardware” institute that
              will provide
                   • CAD tools, computer resources
                     (cloud)
                   • Shared FPGA platform for validation
   Opportunities   • Support for chip foundry Threats
                   • Open source cores (a la
                     “opencores.org”)
                   • Paid by commercial use of the results



                                                              51
HiPEAC Opportunities

        • The introduction of massively
          heterogeneous systems for which
          Europe
     Strengthshas lots of expertise. Weaknesses
        • Domain specific languages and
          paradigms
        • Silicon and non-silicon based new
          technologies (molecular, memristive
          devices, etc.)


   Opportunities                  Threats



                                                  52
HiPEAC Threats

       • Computing systems not considered strategic in
         Europe.
       • We are late but not yet too late to create a
     Strengths                             Weaknesses
         HiPEAC industry – China proves it is still possible.
       • The HiPEAC domain is effectively dominated by
         US companies and universities but there is a clear
         shift towards Asia. Competition is tough.
       • SMEs are hard to convince to participate in EU-
         funded project due to administrative overhead.
       • Projects are most useful for an SME if they are
         coordinated by the SME. This is almost unfeasible
         for a small company.
       • Large companies only participate for less strategic
   Opportunities                               Threats
         projects (not in critical path of production)




                                                                53
                       Solutions?
                      We could give more awards to European excellence – and create more prestigious awards
                       programs, not just for academic papers. Startups, fellows, tools, … It is more important to invest in
                       people than in structures.
                       The collaborative research in HiPEAC would benefit from lightweight projects for testing ideas.
Recommendation 1


                   
                       Only high risk/high gain projects.
                      Long term research/basis research needs substantial structural funding, Examples.
                       ◦   UMIC – Aachen – 35 M€ for 5 years for the development of Ultra high speed Mobile Information and
                           Communication.
                       ◦   UPMARC – Lineus grant for 10 years, 6.5 M€.
                       ◦   Priority programs in Germany – 6 M€ per program, running for 3x2 years.
                       ◦   EIT-KIC, ERC, … in Europe; FCRP (Focus Center Research Program, fcrp.src.org), C2S2 in the USA.

                      We should interact more with companies at the board level (CEO, CTO).
                      We could cooperate more actively with ACM Europe.
                      Must increase European presence at International conferences; bring more international
                       conferences to Europe?
                      We could send our best papers to the HiPEAC journal and HiPEAC conference and make them
                       premier publication outlets.
                      We should probably work more closely together with the software objectives, e.g. on the cloud, see
                       http://cordis.europa.eu/fp7/ict/ssai/docs/executivesummary-forweb_en.pdf. Hardware aspects of
                       cloud computing are completely absent.                                                           54
                   Four instruments
Recommendation 1



                           Priority    Innovative
                          programs        Ideas




                          Technology   Fellowship
                            transfer      Fund


                                                    55
                   Priority Programs
                   • A PP is a portfolio of projects working together:
                     • it is focused on clear challenges
Recommendation 1


                     • it allows for medium term planning (up to 6 years, can be stopped
                       after 2 years or 4 years)
                     • it is dynamic (PPs come and go)
                     • it consist of 4-7 (complementary) projects
                     • It is typically proposed by a consortium of companies and
                       academics.

                   • A PP could be considered as a cluster with research funds
                   • A PP could also be a JRI
                   • The PP leader manages the priority program, organize PP meetings,
                     workshops, dissemination activities, etc. Other projects (national,
                     European, individual researchers) can join a PP once it is in place. This
                     will further increase the impact of the PP.

                                                                                             56
                   Innovative Ideas
                    The instrument that allows trying out new innovative ideas that do
                     not fit the priority programs.
Recommendation 1


                    This instrument calls for projects with small consortia.
                    Projects can run up to three years.




                                                                                          57
                   Technology transfer
                    Dedicated technology transfer projects, with technology transfer as
                     the prime focus and facilitating it.
Recommendation 1


                    Duration 6-18 months.
                    Consortium of only the key partners of the technology transfer.
                    Easy to get, low overhead, and feasible for any type of company.
                    Might be a particularly attractive instrument for small SMEs.




                                                                                           58
                   Fellowship Fund
                    The best investment for the future is investing in young talent.
Recommendation 1


                    ERC grants do not work that well for our community.
                    Community would benefit from 3 junior and 3 senior ERC style
                     grants working in the core HiPEAC topics.
                    Grants should be prestigious, it would be good to make them
                     substantial (1.5-2.0M€ for 5 years).




                                                                                        59
                   Benefits for the
                   HiPEAC community
                    It creates a framework for structural funding.
Recommendation 1


                      ◦ For young talented researchers: fellowships with substantial funding, and low
                         overhead, running for 5 years. They earn freedom to operate, and the budget
                         could allow them to hire up to 10 persons for 5 years.
                      ◦ For collaborative research: the priority programs running for at least 6 years.
                         This structural funding should allow PP coordinators to effectively tackle
                         concrete challenges.
                    It allows for low overhead collaborative projects.
                      ◦ The focused PP projects with a small number of participants with a clear
                         ambition to work on the challenges set forward in the priority programs, and to
                         collaborate with other projects.
                      ◦ The technology transfer projects with only the partners involved in the
                         technology transfer.
                    It holds to promise for more excellence and visibility in the international publication
                     venues.
                      ◦ By giving substantial resources to young and talented researcher, resulting in
                         many visible publications.
                      ◦ By creating PPs that tackle grand challenges, and thus generating high
                         expectations in the PP and in the public.
                      ◦ By stimulating more companies to start a technology transfer project.
                                                                                                           60
                   Four instruments in FP7
Recommendation 1



                            Priority   Innovative
                           programs       Ideas




                          Technology   Fellowship
                            transfer      Fund


                                                    61
                   Priority Programs
                      Difficult to fit in the existing structure of IP, STREP, CSA, NoE.
Recommendation 1


                      The current closest match to priority programs – content-wise and budget-
                       wise – are the different target outcomes
                        ◦ Parallel and concurrent computing, virtualization, customization, architecture and
                          technology.
                        ◦ Target outcomes could be structured like priority programs, pre-allocating budgets to
                          each outcome to guarantee a good coverage.
                      For bigger priority programs, a completely new research objective could be
                       created. Research objectives have their own consultation meetings, roadmaps
                       documents, and portfolio of projects.
                      Currently, there is no comparable alternative for the longer term funding of
                       priority programs (3x2 years). Successful projects can only submit a follow-up
                       project, and hope it will be accepted. The same holds for networks of excellence.
                      There is no formal structure for networking between the projects belonging to a
                       priority program. Projects are currently not required to collaborate.



                                                                                                                  62
                    Priority program networking
                                       Project 1              Project 2             Project 3
Recommendation 1


                                WP1 WP2 WP3 WP1 WP2 WP3 WP4 WP1 WP2
                    Cluster A     X                                                         X
                    Cluster B                    X             X                     X
                    Cluster C     X                            X                            X
                   Cluster D                            X                    X
                    Cluster E                                         X              X
                    Cluster F            X                     X
                    Cluster G                    X                                   X
                    Cluster H            X                     X                            X

                   Clusters could organize 1 or 2 networking events per year, open to the
                   HiPEAC community.
                                                                                                63
                   Innovative Ideas
                   1. Projects could be submitted in FET open calls.
Recommendation 1


                   2. Future computing system calls could also call for wild
                      and crazy ideas in computing systems to stimulate the
                      preparation of proposals that could be incubators of
                      good ideas for future priority type programs.
                     [The architecture and technology target outcome is the closest match in
                     the next call, but it could be made more explicit in future calls]




                                                                                               64
                    Technology transfer
                   1. As a dedicated STREP project with technology transfer as primary
                      focus. This might however be difficult to justify given the current set
                      of evaluation criteria.
Recommendation 1



                   2. An add-on to existing projects.
                     1.   Technology transfer work package at the end of the STREP or IP project.
                     2.   Budget associated with such work package (e.g. 10% of the total project budget)
                          will only be unlocked based on the approval of a business plan which could be a
                          deliverable by the end of the project and should be approved by the reviewers –
                          possibly extended with a business consultant for that meeting.
                     3.   If the business plan is not approved, the project simply stops, and the money
                          flows back to the commission.
                     4.   The prospect of being able to acquire a dedicated technology transfer budget
                          (probably for a subset of the partners) might revitalize a project in its last phase,
                          and it might even change the focus of the project early on in preparation of the
                          business plan.
                     5.   We believe it should also be possible to engage additional partners in this phase
                          of the project.
                                                                                                                  65
                   Fellowship Fund
                    Fellowships do not fit regular calls (single PI, duration, budget)
Recommendation 1


                    We have to continue trying for ERC




                                                                                          66
         Agenda
          09:00-09:30   Private meeting reviewers
          09:30-09:45   Introduction + Agenda
                        Report on the activities of February-May 2010, including
                        -     RC9: Repository of research results
          09:45-10:15   -     RC10: Mini-sabbaticals update
                        -     RC12: Top conferences
                        -     RC13: Mini-sabbaticals: results
          10:15-10:30   Budget overview and planning
Agenda




          10:30-10:45   BREAK
          10:45-11:30   RC1: Future of HiPEAC
          11:30-12:00   RC2: Roadmap update
          12:00-12:15   RC3: A consistent view on the scalability limitations of multicore and
                        manycore systems
          12:15-13:30   LUNCH
                        RC4 : The Project Management should increase HiPEAC’s visibility and the
                        link with Artemis
                        RC5: The project management needs to stimulate much greater and more
          13:30-14:00
                        effective interaction of the research clusters
                        RC6: The Management Report should contain a section giving an overview
                        of related projects and initiatives and their interaction with HiPEAC
          14:00-14:15   RC7: A more systematic approach for analyzing real applications
          14:15-14:45   RC8: The HiPEAC Journal plans
          14:45-15:00   RC11: HiPEAC Industrial Workshop
          15:00-15:30   Reviewers Private Meeting
          15:30-16:00   Feedback
                                                                                                   67
                   Roadmap
                   It is necessary to focus the Research Roadmap on specific research challenges that
                   are within the scientific and technical competences of HiPEAC. The revision of this
                   Roadmap with the inclusion of timeline and with a fuller presentation of the
Recommendation 2


                   interaction between industry and research should be presented at the formal
                   review.
                    A one day workshop was held in Paris in March 19, 2010 to
                     answer the recommendations.
                    Several actions are started for the next release of the
                     document:
                     ◦ Request to clusters/tasks force leaders to provide a timeline
                       of the research challenges related to their clusters according
                       to the 6 main research challenges
                     ◦ Request the industrial partners and members of the
                       industrial advisory board to give explicit feedback on the
                       current document
                     ◦ The roadmap was presented on several occasions and
                       feedback was collected.
                                                                                                         68
                   Roadmap
                      The Glossary is an important output. It must reflect what new
Recommendation 2


                       qualitative concepts the recent research brought. It should
                       introduce concepts which are at the leading edge of research.
                      Predictability’s definition in the Glossary: It is only one aspect of
                       this concept, related with real-time applications.
                        ◦ Complete rewriting of the glossary.
                      Heterogeneous vs. Homogeneous would be an interesting topic
                       necessary to be synthesized in the Roadmap.
                        ◦ Text about heterogeneous vs. homogeneous was
                          provided by the multi-core cluster and will be integrated
                          in the next release of the document. See also scalabiliy
                          recommendation 3.

                                                                                               69
                   Roadmap: timeline
                      Most clusters provided their upcoming challenges with a
Recommendation 2


                       rough timeline of when they will happen.
                      It is organized according to the six main research challenges
                       for inclusion in the document and will be the start of the
                       “roadmap” part of the vision document.
                      "We cannot predict the future, but we can invent it.” (D.M.
                       Cole): the clusters will elaborate on these guidelines to
                       provide inputs for the final roadmap document.
                      “Even Technologists Can't Predict the Future” (WSJ, June 7,
                       2010): of course, we cannot promise that all these will be
                       accurate or even will really happen in the future…


                                                                                       70
                   Roadmap: timeline
                    See next section for the “View on Scalability”
Recommendation 2


                     and associated “Research challenges”
                    Examples for each research challenge:
                       ◦ Design Space Exploration:
                          Efficiency x8 by 2015 and by 64 in 2020
                       ◦ Concurrent programming model:
                          Better specification of concurrency by 2013
                          OS support for heterogeneous cores by 2015
                          Adaptability to available/variable resources at runtime by
                           2015-2020



                                                                                        71
                   Roadmap: timeline
                      Examples for each research challenge (Cont’d):
Recommendation 2


                       ◦ EDA:
                         Full-system simulation with interconnection
                          network by 2015
                         Industrial tools for RC after 2020
                       ◦ Design of optimized components:
                         Performances improved by x8 by 2015 and by x16
                          by 2020
                         New on-chip memory hierarchies (e.g. 3-D
                          stacking) by 2015-2020
                         Different nano-scale technology (stochastic
                          behavior) after 2020
                                                                           72
                   Roadmap: timeline
                      Examples for each research challenge (Cont’d):
Recommendation 2


                       ◦ Self-adapting systems:
                         Hardware support for automatic performance
                          tuning by 2015
                         Flow adaptation and reconfiguration by 2020
                       ◦ Virtualization:
                         Death of the native ISA by 2015
                         Portable performances for Heterogeneous multi-
                          core systems by 2020




                                                                           73
                   Roadmap: industrial
                   interaction
                      The current roadmap was written by participants from
Recommendation 2


                       NXP, Thales, STM, Infineon, Intel, Microsoft, Transitive,
                       HP, CEA.
                      It was distributed and got feedback at least from NXP,
                       Thales, STM, CEA…
                      Other inputs from ARM, IBM,Vector Fabrics…
                      The industrial participants of the ACACES 2009 and
                       2008 were also consulted and provided inputs taken
                       into account in the document.




                                                                                   74
                   Roadmap: industrial
                   interaction
                   Example of industrial highlights:
Recommendation 2


                    HP:
                       ◦ Trends towards Everything as service
                       ◦ Intelligent infrastructure
                       ◦ Energy and sustainability
                       ◦ Embedded sensors, actuators and sensor networks (the CENSE
                         project)
                       ◦ Mobility and user experience
                      NXP:
                       ◦ More than Moore
                       ◦ Embedded heterogeneous systems (including sensors and
                         networking)


                                                                                      75
                   Roadmap: industrial
                   interaction
                   Example of industrial highlights:
Recommendation 2


                    Thales:
                       ◦ Application trends in aerospace, avionics, smart
                         cameras.
                       ◦ More than Moore.
                       ◦ Predictability and time is relevant aspect for safety
                         critical business domain.
                       ◦ Reliability and managing system complexity.


                      …

                                                                                 76
                   Roadmap dissemination
                      The roadmap was presented on different occasions
Recommendation 2


                       ◦   Thales, NXP internally
                       ◦   University of Uppsala (UPMARC)
                       ◦   University of Dublin
                       ◦   Ghent University (for about 200 people)
                       ◦   DAC (compilation and virtualization, by STM)
                       ◦   …
                    The roadmap was distributed at IPDPS, DATE, ISCA,
                     HiPEAC, SASP, DAC, …
                    CACM paper (summary of the vision document) has
                     been submitted

                                                                          77
                   Feedback on the “Vision”
                    Good feedback and interactions
Recommendation 2


                    “New” topics possibly to be added:
                       ◦ CPS: Cyber-Physical Systems
                       ◦ ESA: Embedded Software Automation?
                       ◦ Computing without guarantees (e.g. stochastic, …)
                       ◦ Impact of 3-D stacking: next barrier might be
                         connectivity
                       ◦ …




                                                                             78
         Agenda
          09:00-09:30   Private meeting reviewers
          09:30-09:45   Introduction + Agenda
                        Report on the activities of February-May 2010, including
                        -     RC9: Repository of research results
          09:45-10:15   -     RC10: Mini-sabbaticals update
                        -     RC12: Top conferences
                        -     RC13: Mini-sabbaticals: results
          10:15-10:30   Budget overview and planning
Agenda




          10:30-10:45   BREAK
          10:45-11:30   RC1: Future of HiPEAC
          11:30-12:00   RC2: Roadmap update
          12:00-12:15   RC3: A consistent view on the scalability limitations of multicore and
                        manycore systems
          12:15-13:30   LUNCH
                        RC4 : The Project Management should increase HiPEAC’s visibility and the
                        link with Artemis
                        RC5: The project management needs to stimulate much greater and more
          13:30-14:00
                        effective interaction of the research clusters
                        RC6: The Management Report should contain a section giving an overview
                        of related projects and initiatives and their interaction with HiPEAC
          14:00-14:15   RC7: A more systematic approach for analyzing real applications
          14:15-14:45   RC8: The HiPEAC Journal plans
          14:45-15:00   RC11: HiPEAC Industrial Workshop
          15:00-15:30   Reviewers Private Meeting
          15:30-16:00   Feedback
                                                                                                   79
                   Multi- and manycore systems:
                   Limitations
                   HiPEAC should develop a consistent view on the scalability
Recommendation 3


                   limitations of multicore and especially manycore systems and the
                   specific research directions for overcoming these limitations
                   (shared and distributed memories, message passing, hybrid models
                   etc.). Following discussion at the informal review meeting it is clear that
                   there has been misinterpretation of this recommendation.The action still
                   remains for HiPEAC to present its view on the scalability limitations of multi-
                   core and especially many-core systems and the specific research directions
                   for overcoming these limitations.This view may encompass several fertile
                   research directions. These should be incorporated into the Roadmap in a
                   coherent way together with a clear discussion of the research options.This
                   recommendation should be completed for the formal review.


                                                                                                 80
                   View on Scalability
                   Outlook next 10 years:
Recommendation 3


                    Core count per chip 2X every 2 years
                    Clock frequency per annum: up by 5-10%
                    By 2020, a couple of hundred cores
                    Power budget: effectively constant


                   Implication:
                   To double the compute performance every 2 years,
                   core replication must be combined with
                   specialization leading to heterogeneous multicores

                                                                    81
                   Research Challenges
                   Memory performance will become a major
Recommendation 3


                   bottleneck
                    Latency speed gap is expected to stay.
                    Off-chip bandwidth can become a bottleneck; on-
                     chip memory management is important.
                    Cache coherence needed but must be revisited in
                     the ”on-chip” context 2020.
                    Implicit (through cache coherence) and explicit
                     communication (through message passing) will
                     both be needed for memory locality management.


                                                                   82
         Agenda
          09:00-09:30   Private meeting reviewers
          09:30-09:45   Introduction + Agenda
                        Report on the activities of February-May 2010, including
                        -     RC9: Repository of research results
          09:45-10:15   -     RC10: Mini-sabbaticals update
                        -     RC12: Top conferences
                        -     RC13: Mini-sabbaticals: results
          10:15-10:30   Budget overview and planning
Agenda




          10:30-10:45   BREAK
          10:45-11:30   RC1: Future of HiPEAC
          11:30-12:00   RC2: Roadmap update
          12:00-12:15   RC3: A consistent view on the scalability limitations of multicore and
                        manycore systems
          12:15-13:30   LUNCH
                        RC4 : The Project Management should increase HiPEAC’s visibility and the
                        link with Artemis
                        RC5: The project management needs to stimulate much greater and more
          13:30-14:00
                        effective interaction of the research clusters
                        RC6: The Management Report should contain a section giving an overview
                        of related projects and initiatives and their interaction with HiPEAC
          14:00-14:15   RC7: A more systematic approach for analyzing real applications
          14:15-14:45   RC8: The HiPEAC Journal plans
          14:45-15:00   RC11: HiPEAC Industrial Workshop
          15:00-15:30   Reviewers Private Meeting
          15:30-16:00   Feedback
                                                                                                   83
         Agenda
          09:00-09:30   Private meeting reviewers
          09:30-09:45   Introduction + Agenda
                        Report on the activities of February-May 2010, including
                        -     RC9: Repository of research results
          09:45-10:15   -     RC10: Mini-sabbaticals update
                        -     RC12: Top conferences
                        -     RC13: Mini-sabbaticals: results
          10:15-10:30   Budget overview and planning
Agenda




          10:30-10:45   BREAK
          10:45-11:30   RC1: Future of HiPEAC
          11:30-12:00   RC2: Roadmap update
          12:00-12:15   RC3: A consistent view on the scalability limitations of multicore and
                        manycore systems
          12:15-13:30   LUNCH
                        RC4 : The Project Management should increase HiPEAC’s visibility and the
                        link with Artemis
                        RC5: The project management needs to stimulate much greater and more
          13:30-14:00
                        effective interaction of the research clusters
                        RC6: The Management Report should contain a section giving an overview
                        of related projects and initiatives and their interaction with HiPEAC
          14:00-14:15   RC7: A more systematic approach for analyzing real applications
          14:15-14:45   RC8: The HiPEAC Journal plans
          14:45-15:00   RC11: HiPEAC Industrial Workshop
          15:00-15:30   Reviewers Private Meeting
          15:30-16:00   Feedback
                                                                                                   84
                   Artemis
                   The Project Management should increase HiPEAC’s
Recommendation 4


                   visibility and the link with Artemis, going beyond the
                   current general involvement of becoming a member, but
                   also by implementing effective cooperation and linking
                   directly to each relevant company active in Artemis as was
                   recommended at the first formal review




                                                                                85
                       Artemis Steering Board Members
                   http://cordis.europa.eu/ist/artemis/members.htm

                      1. ABB                                         15. ITEA
Recommendation 4


                      2. Airbus                                      16. MEDEA+

                      3. ARM                                         17. Nokia

                      4. BT exact                                    18. Parades

                      5. Comau                                       19. Philips

                      6. Catalonia Technical University              20. Robert Bosch

                      7. ContiTeves                                  21. Siemens

                      8. DaimlerChrysler                             22. ST Microelectronics

                      9. Ericsson                                    23. Symbian

                      10. European Federation of High Tech SMEs      24. Telenor

                      11. Finmeccanica                               25. Thales

                      12. Fraunhofer Institute                       26. Vienna Technical University

                      13. IMEC                                       27. Verimag

                      14. Infineon Technologies
                                                                                                       86
                   Cluster interaction
                   The project management needs to stimulate much greater
Recommendation 5


                   and more effective interaction of the research clusters

                      Actions:
                       ◦   Inter-cluster events in the Edinburgh Innovation Event
                       ◦   Merger of Simulation platform and EDA cluster
                       ◦   Merger of Compilation platform and Adaptive compilation cluster
                       ◦   Brainstorming about new role for clusters/taskforces in combination
                           with FP7 project portfolio




                                                                                                 87
                    Related projects
                    The Management Report should contain a section giving an overview of
                    related projects and initiatives and their interaction with HiPEAC
Recommnendation 6



                                           Project        Members          Coordinator
                                          Partners        HiPEAC            HiPEAC
                      2PARMA                 8               4                  1
                      ADVANCE                9               1                  0
                      ENCORE                 7               7                  1
                      EuroCloud              5               3                  1
                      HEAP                   9               1                  0
                      IOLANES                6               1                  1
                      PEPPHER                10              4                  1
                      PLANETHPC              1               1                  1
                      PRO3D                  9               2                  0
                      Reflect                8               5                  1
                      TERAFLUX               27             25                  1
                      ERA                    9               5                  1
                      EURETILE               6               1                  0
                                            114             60                 9
                                                                                           88
     Recommnendation 6



                   ADVANCE




89
         Agenda
          09:00-09:30   Private meeting reviewers
          09:30-09:45   Introduction + Agenda
                        Report on the activities of February-May 2010, including
                        -     RC9: Repository of research results
          09:45-10:15   -     RC10: Mini-sabbaticals update
                        -     RC12: Top conferences
                        -     RC13: Mini-sabbaticals: results
          10:15-10:30   Budget overview and planning
Agenda




          10:30-10:45   BREAK
          10:45-11:30   RC1: Future of HiPEAC
          11:30-12:00   RC2: Roadmap update
          12:00-12:15   RC3: A consistent view on the scalability limitations of multicore and
                        manycore systems
          12:15-13:30   LUNCH
                        RC4 : The Project Management should increase HiPEAC’s visibility and the
                        link with Artemis
                        RC5: The project management needs to stimulate much greater and more
          13:30-14:00
                        effective interaction of the research clusters
                        RC6: The Management Report should contain a section giving an overview
                        of related projects and initiatives and their interaction with HiPEAC
          14:00-14:15   RC7: A more systematic approach for analyzing real applications
          14:15-14:45   RC8: The HiPEAC Journal plans
          14:45-15:00   RC11: HiPEAC Industrial Workshop
          15:00-15:30   Reviewers Private Meeting
          15:30-16:00   Feedback
                                                                                                   90
                   Analysing real applications
                   A more systematic approach for analyzing real applications should be found.
                   This should involve finding the characteristics of real applications as opposed
Recommendation 7


                   to idealised benchmarks. Quantitative figures for comparing the different
                   platforms should be provided to demonstrate the potential for further
                   research and development. It would be interesting to include in the Roadmap
                   the relationship between the application trends and the benchmarks
                   currently used in the manycore research area. A view about reducing the gap
                   between benchmarks and applications would be useful.The reviewers look
                   forward to seeing HiPEAC’s response on this issue at the formal review.




                                                                                                 91
                   Analysing real applications
                      The members of the TFA are studying advanced complete applications,
                       including their characterization and their evaluation, in the context of
Recommendation 7


                       other FP7 EU projects. These results will be shared with the other
                       members of the Task Force once they become publicly available.
                      The repository has new applications, and they are ported to different
                       Programming Models and evaluated on top of different platforms
                      Success story: the STAP (Space-Time Adaptive Processing Algorithm) ,
                       contributed by Thales, on radar processing
                      Two FP7 Projects, ENCORE and Teraflux, are using the HiPEAC repository
                      Two publications from these projects:
                       ◦ Can Manycores Support the Memory Requirements of Scientific Applications?.
                         Intl. Workshop on Applications for Multi and Many Core Processors (A4MMC),
                         Jun 2010. (best paper award)
                       ◦ Evaluation of Programming Models Supporting Heterogeneity and Local
                         Memories using CellStream, FixedGrid and PBPI: many authors, UPC-
                         BSC,Virginia-TECH, FORTH, Three applications: CellStream, fixedgrid and PBPI,
                         on different PM: SDK, Sequoia, CellGen, StarSs, TPC, and CellMP for the
                         evaluation of the CellBE.
                                                                                                         92
                   Evaluation of Real Applications
                      Real production applications running on the MareNostrum Supercomputer
                      Estimate a critical point where current memory systems won't be able to
Recommendation 7


                       satisfy the demand of the future multicore
                      Bandwidth analysis: L1, L2 and off-chip memory bandwidth




                                                                                                 93
Programming Models Evaluation
    Serial version
    Hand coded
       SDK programming, static scheduling, manual offload
    Based on language constructs
       Sequoia, static sched., manual offload
    Based on program annotations, including hints for data
     transfers
        CellGen, static sched., automatic offload
        StarSs, dynamic sched., manual offload
        TPC, static sched., manual offload
        CellMP, static sched., automatic offload
                    CellStream




                                                                  • Annotations reduce the
           24
                                                                    number of lines
           20                                                       added/removed
           16                                      Sequoia
                                                                  • TPC includes a dispatcher
                                                                  • CellMP separates task,
                                                   StarSs
GBytes/s




                                                   TPC*

                                                                    block, and
           12                                      CellMP*
                                                   SDK

                                                                    copy_in/copy_out
                                                   CellGen-NUMA
           8                                       CellMP*-NUMA

                                                                    annotations
                                                   SDK-NUMA
                                                   Peak
           4
                                                                  • Taking into account NUMA
           0
                1   2     4          8   12   16                    is very important
                        Number of SPUs
                                      Fixedgrid




                            300                                        Serial
                                                                       Sequoia-SIMD
                                                                       CellGen-SIMD
                                                                                         •   CellGen is unable to offload 3
                            250                                        StarSs-OPTK           matrix transpositions, because
                                                                       TPC*-gccSIMD
Execution time in seconds




                                                                       CellMP*-gccSIMD       strided accesses require
                                                                       TPC*-OPTK
                            200
                                                                       CellMP*-OPTK          explicit padding, and padding
                                                                                             makes data incompatible with
                                                                       SDK-SIMD

                            150
                                                                                             SIMD
                            100                                                          •   GCC achieves competitive
                                                                                             vectorization compared to
                            50                                                               hand-coded (OPTK)
                                                                                         •   Taking care of kernel code
                             0
                                  1      2    4               8   12               16        generation is important to
                                             Number of SPUS                                  favor vectorization
                                       PBPI




                                                                                                    • Static task scheduling allows
                            1200                                                   Serial
                                                                                   Sequoia-Scalar

                                                                                                      keeping iterations mapped onto
                            1000                                                   CellGen-Scalar
                                                                                   TPC*-Scalar
Execution time in seconds




                                                                                   CellMP*-Scalar
                            800                                                    Sequoia-SIMD
                                                                                   CellGen-SIMD
                                                                                                      same processors
                            600
                                                                                   StarSs-SIMD
                                                                                   TPC*-SIMD        • Manual SIMDization makes a
                                                                                                      difference in this application
                                                                                   CellMP*-SIMD
                            400                                                    SDK-SIMD


                            200                                                                     • PBPI tasks are small enough in
                              0
                                                                                                      time to have the overhead of
                                   1     2             4            8
                                             Number of SPUs / Serial on PPU
                                                                              12               16
                                                                                                      runtime systems revealed
         Agenda
          09:00-09:30   Private meeting reviewers
          09:30-09:45   Introduction + Agenda
                        Report on the activities of February-May 2010, including
                        -     RC9: Repository of research results
          09:45-10:15   -     RC10: Mini-sabbaticals update
                        -     RC12: Top conferences
                        -     RC13: Mini-sabbaticals: results
          10:15-10:30   Budget overview and planning
Agenda




          10:30-10:45   BREAK
          10:45-11:30   RC1: Future of HiPEAC
          11:30-12:00   RC2: Roadmap update
          12:00-12:15   RC3: A consistent view on the scalability limitations of multicore and
                        manycore systems
          12:15-13:30   LUNCH
                        RC4 : The Project Management should increase HiPEAC’s visibility and the
                        link with Artemis
                        RC5: The project management needs to stimulate much greater and more
          13:30-14:00
                        effective interaction of the research clusters
                        RC6: The Management Report should contain a section giving an overview
                        of related projects and initiatives and their interaction with HiPEAC
          14:00-14:15   RC7: A more systematic approach for analyzing real applications
          14:15-14:45   RC8: The HiPEAC Journal plans
          14:45-15:00   RC11: HiPEAC Industrial Workshop
          15:00-15:30   Reviewers Private Meeting
          15:30-16:00   Feedback
                                                                                                   98
                    The HiPEAC Journal
                   The idea to make the HiPEAC Journal available in visible digital libraries such
Recommendation 8


                   as the ones provided by ACM (ACM DL) or IEEE (IEEExplore) should be
                   progressed. The review panel thinks that this goal is correct and appropriate.
                   Avoidance of overlap with other journals is essential.




                                                                                                     99
                   Background
                   Conferences increasingly face issues
Recommendation 8


                    Not indexed; hurt the domain
                    Reviewing quality poor; history not kept
                    Organization big overhead for community
                    Republication issues
                    Too many small conferences
                    Environmental issues
                   Journals
                    Have problems attracting papers
                    Have very long review cycles
                                                                100
                   A New Publication Model
                   = publication model in other domains of science
Recommendation 8


                     ◦ Journal = publication venue
                     ◦ Conference = meeting venue
                   Plan
                    ◦ Create ACM Transactions on High-Performance and
                      Embedded Computing Systems (TOHPECS)
                    ◦ Give all TOHPECS papers presentation slots in the
                      HiPEAC conference
                    ◦ Starting in 2011 (same for VLDB journal)



                                                                          101
                   Publication model
                                               CFP
Recommendation 8



                   J   F   M   A    M      J         J   A   S   O   N   D




                               Conference presentations
                               in January




                                                                             102
                   Conference
                    Workshops, tutorials, exhibition, keynotes,
                     EU projects, student competition, paper
Recommendation 8



                     presentations are all integrated
                    One registration fee for the whole
                     conference
                    Participants make their own conference
                     program
                    No more PC required. PC-chair makes a
                     program out of the accepted journal
                     papers
                    No more formal proceedings - reprints

                                                               103
                   TOHPECS journal
                      Digital only - TOCS
Recommendation 8


                      Copy-edited
                      Papers of 6000-15000 words
                      3 reviewers
                      Short review cycle (4 weeks for short paper – thanks
                       to review board)
                      Outcome
                       ◦   Accept
                       ◦   Accept with minor revision
                       ◦   Accept with major revision
                       ◦   Revise and submit
                       ◦   Reject
                                                                          104
                   Benefits
                      Shorter review cycles for research papers
                      Higher-quality reviews
Recommendation 8


                      No more hard submission deadlines for papers
                      Better quality control on the final version of the papers
                      No more extreme peaks in the review load for reviewers
                      No more republication issues
                      No more need to extend papers to journal versions
                      All citations go to the same paper
                      Increasing impact factors for the journal
                      No more time consuming physical pc-meetings
                      Easier to handle conflict of interest management
                      Better protection of new ideas
                      The publishing process will be more efficient, cheaper and greener
                      Conferences might attract more delegates
                                                                                        105
         Agenda
          09:00-09:30   Private meeting reviewers
          09:30-09:45   Introduction + Agenda
                        Report on the activities of February-May 2010, including
                        -     RC9: Repository of research results
          09:45-10:15   -     RC10: Mini-sabbaticals update
                        -     RC12: Top conferences
                        -     RC13: Mini-sabbaticals: results
          10:15-10:30   Budget overview and planning
Agenda




          10:30-10:45   BREAK
          10:45-11:30   RC1: Future of HiPEAC
          11:30-12:00   RC2: Roadmap update
          12:00-12:15   RC3: A consistent view on the scalability limitations of multicore and
                        manycore systems
          12:15-13:30   LUNCH
                        RC4 : The Project Management should increase HiPEAC’s visibility and the
                        link with Artemis
                        RC5: The project management needs to stimulate much greater and more
          13:30-14:00
                        effective interaction of the research clusters
                        RC6: The Management Report should contain a section giving an overview
                        of related projects and initiatives and their interaction with HiPEAC
          14:00-14:15   RC7: A more systematic approach for analyzing real applications
          14:15-14:45   RC8: The HiPEAC Journal plans
          14:45-15:00   RC11: HiPEAC Industrial Workshop
          15:00-15:30   Reviewers Private Meeting
          15:30-16:00   Feedback
                                                                                                   106
                     The Industrial Workshop
                    Strategies to increase the appeal of the industrial workshop need to be further
                    developed. It is important to report feedback from industry and how this
                    feedback influenced the project. Progress on this important issue should be
Recommendation 11



                    reported at the formal review.

                       Decided to radical reshape the way the workshop
                        organised
                       Make it SME/spinout orientated
                          ◦ Ask them what they want and do it!
                       Put the industrial programme at the heart of the
                        programme
                       Cross-cutting cluster meetings – challenge orientated
                          ◦ Space off the official programme for clusters and projects
                         Effort put into social programme
                          ◦ Make it welcoming for all

                                                                                                      107
108
                    Activities: 2-way dialogue
                       Keynotes from academics with track records as
                        entrepeneurs
Recommendation 11



                       SMEs 15 minute pitches
                       SME stands
                       Poster session: voted by industials
                       SMEs and industrials in cross-cluster meetings
                       Venture Capitalist and Commercialisation
                        Director
                       FP7 presentations
                       Heterogeneous parallelism and Cloud Computing
                    Social Event
                       In past underestimated the need for social glue
Recommendation 11


                       2 days over 3 increased social contact time
                       Reception on roof garden
                       Atrium meal + poster session
                       Unofficial programme
                        ◦   For younger members
                        ◦   Walk Arthur’s seat
                        ◦   Dungeon tour
                        ◦   Dinner afterwards
                       Made a real difference
                    Edinburgh Innovation Event
                    Day 1: May 3, 2010
                    12:00    Registration
                    13:00    Welcome and Keynote: The High-Tech Research and Entrepreneurship Challenge
Recommendation 11


                             - Mario Nemirovsky (ICREA Research Professor at the Barcelona Supercomputer
                             Center )
                    14:00    Inter Cluster meeting I: "Throughput computing with IA cores" - Toni Juan (Intel)
                    15:00    Break
                    15:30    Inter-cluster Meeting II: Cloud Computing Software and Hardware Challenges
                             Bringing Business to the Cloud - Alex Heneveld (Cloudsoft Corp.)
                             Milliwatt Chips: The Viable Scalability Path for Data Centers - Babak Falsafi (EPFL)
                             mJ per Web-Access: Energy-efficient Next-generation Scalable Computing -
                             Anirban Lahiri (ARM)
                    16:30:   Inter Cluster meeting III: Dusk of general-purpose many-cores, dawn of
                             heterogeneous multi-cores: What are the design and programming issues ?
                    17:30    Drinks Reception
                    19:30    Close of Day
                    Edinburgh Innovation Event
                    Day 2: May 4, 2010
                    9:00     Registration
                    9:30     Colin Adams: Commercialising Informatics Research: The Prospekt Experience
Recommendation 11



                    10:30    Uri Weiser: Industry Innovation? Symmetric and Asymmetric Chip Multi-Core:
                             Applications, processors and scheduling - Initial Thoughts
                    11:30    Morning Coffee
                    12:00    Robert Jelski (Capital-E) Venture Capital's objectives and what they seek in early
                             stage investments
                    13:00:   Lunch
                    14:00    Main Conference: 15-20 minute presentations by
                             ACE, Acumem, Bitwise, CAPS, Cloudsoft, Codeplay,
                             CriticalBlue, Microtech, NemaLabs, Yogitech
                    16:00    Poster Exhibition in Atrium
                    17:30:   Hot Networking buffet & wine to be served
                    19:00    Close of day
                    Edinburgh Innovation Event
                    Day 3: May 5, 2010
                    9:00     Registration
                    9:30     FP 7 Session
Recommendation 11



                             •         2PARMA
                             •         Advance
                             •         Encore
                             •         EuroCloud
                             •         HEAP
                             •         PEPPHER
                             •         PlanetHPC
                             •         PRO3D
                             •         REFLECT
                             •         TERAFLUX
                             •         Description of upcoming Computer Systems Call
                    13:00:   Close of official event + prizes for posters
                    14:00    Parallel Meetings Rooms available
                    16:00    Close of Event
      Recommendation 11   Feedback




114
      Recommendation 11   Feedback




115
      Recommendation 11   Feedback




116
                    What difference did it make?
                    Impact
                         Personally, met several new companies
Recommendation 11


                     
                         ◦ In process of putting together consortia
                        Companies felt they were welcome
                         ◦ and there was great work in EU
                        Attracted many new companies
                         ◦ Informal feedback says they were impressed
                        Helped revitalise members
                         ◦ New ways of engaging
                    One industrial comment
                    “Just a personal note to say that I thoroughly enjoyed the
Recommendation 11


                    HiPEAC event. … You collected some highly talented and
                    curious (in both senses) individuals letting chemistry run
                    its course alongside suitable process and good
                    organisation…. Those that you drew in from the business
                    and broader community immediately recognised the
                    quality of what you were setting out to do and the care
                    your team had taken.”


                    “Creating such camaraderie and a positive open and
                    inclusive culture is important. Thanks for making sure that
                    we are part of it.”
      End of presentation




      &
      A
      Q


119
         Agenda
          09:00-09:30   Private meeting reviewers
          09:30-09:45   Introduction + Agenda
                        Report on the activities of February-May 2010, including
                        -     RC9: Repository of research results
          09:45-10:15   -     RC10: Mini-sabbaticals update
                        -     RC12: Top conferences
                        -     RC13: Mini-sabbaticals: results
          10:15-10:30   Budget overview and planning
Agenda




          10:30-10:45   BREAK
          10:45-11:30   RC1: Future of HiPEAC
          11:30-12:00   RC2: Roadmap update
          12:00-12:15   RC3: A consistent view on the scalability limitations of multicore and
                        manycore systems
          12:15-13:30   LUNCH
                        RC4 : The Project Management should increase HiPEAC’s visibility and the
                        link with Artemis
                        RC5: The project management needs to stimulate much greater and more
          13:30-14:00
                        effective interaction of the research clusters
                        RC6: The Management Report should contain a section giving an overview
                        of related projects and initiatives and their interaction with HiPEAC
          14:00-14:15   RC7: A more systematic approach for analyzing real applications
          14:15-14:45   RC8: The HiPEAC Journal plans
          14:45-15:00   RC11: HiPEAC Industrial Workshop
          15:00-15:30   Reviewers Private Meeting
          15:30-16:00   Feedback
                                                                                                   120

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:30
posted:2/21/2011
language:English
pages:120