[MODEST] MODEST 4b: developing parallel N-body kernels

[Simon Portegies Zwart]

    MODEST-4b Satellite workshop for MOdeling DEnse STellar systems:
              performing direct N-body calculations on
              massive parallel distributed memory platforms with
              special purpose hardware

       It is a pleasure to invite you for our MODEST-4b (MOdeling DEnse
    STellar systems) satellite workshop on the construction of
    realistic simulation models for star clusters.  This meeting will
    be held in June 7-8 (2004) at the Astronomical Institute 'Anton
    Pannekoek' and the Section Computational Science of the University
    of Amsterdam in Amsterdam, the Netherlands.

        This should be a real working meeting for those actively
    involved in parallel N-body software and hardware development.  We
    therefore have no open registration and participation is by
    invitation.  On the other hand, we do not want to exclude anybody
    from our collaborative effort, which is the core of the MODEST
    philosophy; if you are interested in participating and have not yet
    received an invitation, please contact the undersigned organizers.
    Invited participants are encouraged to bring a graduate student or
    Post Doctoral research associate.

        Our main aim is to bring together a number of experts from the
    fields of stellar dynamics and computational science, though we
    welcome the usual MODEST collaborators to attend this workshop.
    During this satellite meeting we will discuss efficient parallel
    algorithms for performing direct N-body simulations on
    GRAPE-supported distributed and Beowulf platforms.

        The main reason for such a specialized meeting is the recent
    rather widespread availability if GRAPE systems on interconnected
    PC's or Beowulf clusters. This recent development requires the
    parallelization of direct N-body algorithms, which is quite a
    challenge.

       We plan to make this a real `working' workshop, somewhat in the
    style of the Aspen conferences, with few talks, much discussion
    about hands-on aspects of developing an efficient parallel
    algorithm for solving the direct N-body problem with GRAPE.  We
    will not schedule any talks. Instead, we simply encourage all of
    you to bring some relevant presentations and inform the organizers
    in advance of your planned contribution.  The real input of the
    meeting will be your collective expertise in parallelizing N-body
    kernels and the use of GRAPE or other computer hardware.  For the
    output, we envision some budding collaborations leading to the
    development of an efficient direct N-body algorithm for parallel
    platform.

       In keeping with the spirit of an informal work-workshop, there
    will be no registration fee, nor will we have any means to offer
    financial assistance to the participants.  Instead of a banquet, we
    will sample the numerous fine restaurants/bars at walking distance
    to your hotel. We do plan to have a welcoming reception in the
    evening of Sunday June 6, at the East of Eden.

       More information can be found at
    http://www.carol.wins.uva.nl/~spz/act/modest4b or at the usual
    modest website http://www.manybody.org. These pages also contain
    information about registration, hotel reservations, directions for
    travel, a tentative schedule and a list of participants.

    We hope to welcome you at MODEST-4b.

    MODEST greetings,
    Simon Portegies Zwart
          and
    Rainer Spurzem