| 
  
  
   | 
 
 
  | 
  
   | 
 
 
  | 
  
   | 
 
 
  
  
  
  
   
     | 
     | 
    
   
     | 
     | 
    
   
     | 
    Contribute to the acceleration of scientific
    discovery, by the use of information technologies | 
    
   
     | 
    Provide high performance supercomputing
    environments for dealing with science�s more challenging problems. | 
    
   
     | 
    Act as a technology transfer agent between
    R&D in information technologies, and computational science | 
    
   
   
   | 
 
 
  
  
  
  
   
     | 
     | 
    
   
     | 
    Capture more physics in the simulation of
    complex systems | 
    
   
     | 
    Complex systems are characterized by multiple
    time and/or length scales | 
    
   
     | 
    Not easy to capture multiple scales in one code | 
    
   
     | 
    Code coupling for multi-physics applications is
    viable alternative in some conditions | 
    
   
     | 
    This leads naturally to computational grids | 
    
   
     | 
    Heterogeneous algorithms map naturally to heterogeneous
    grids. | 
    
   
   
   | 
 
 
  
  
  
  
   
     | 
     | 
    
   
     | 
    Focus on heterogeneous, very high performance
    supercomputing environments. | 
    
   
     | 
    Use of grid technologies to provide a unified
    image and a transparent access to such environments | 
    
   
     | 
    Deploy an application testbed across Europe by
    the integration of partner�s HPC environments | 
    
   
     | 
    Provide a major effort to develop and deploy
    distributed scientific applications (EUROGRID is roughly 2/3 applications
    development, 1/3 technology development) | 
    
   
   
   | 
 
 
  
  
  
  
   
     | 
     | 
     | 
    
   
     | 
    CUSTOM | 
    
   
     | 
    UNICORE | 
    
   
     | 
    COMMODITY | 
    
   
     | 
    MPI (scientific standard, soon interoperable) | 
    
   
     | 
    CORBA - JAVA (internet standard, has significant
    software engineering advantages, but not yet fully adapted to performance
    focused complex applications). | 
    
   
     | 
    � (others) | 
    
   
   
   | 
 
 
  
  
  
  
   
     | 
     | 
     | 
    
   
     | 
    Forschungezentrum Julich (FZJ, project leader) | 
    
   
     | 
    Rechenzentrum Universitat Stuttgart (RUS) | 
    
   
     | 
    Deutscher Wetterdienst Offenbach (DWD) | 
    
   
     | 
    Konnrad-Zuse-Zentrum Berlin (ZIB) | 
    
   
     | 
    Leibniz Rechenzentrum Munchen (LRZ) | 
    
   
     | 
    Rechenzentrum Universitat Karlsrhue (RUKA) | 
    
   
     | 
    Padenborn Center for Parallel Computing (PC2) | 
    
   
     | 
    Technische Universitat Dresden (TUD) | 
    
   
     | 
    Pallas GmbH Bruhl (Pallas) | 
    
   
     | 
    Fecit, subcontractor to Pallas | 
    
   
     | 
     | 
    
   
   
   | 
 
 
  
  
  
  
   
     | 
     | 
    
   
     | 
    UNICORE develops a seamless, secure, intuitive
    software infrastructure to HPC resources | 
    
   
     | 
    Provides consistent batch access to
    heterogeneous remote systems � | 
    
   
     | 
    � with minimal intrusion into the Centers | 
    
   
     | 
    Supports multi-site and multi-systems
    applications for one job | 
    
   
     | 
    Exploits existing and emerging technologies | 
    
   
   
   | 
 
 
  | 
  
   | 
 
 
  
  
  
  
   
     | 
     | 
    
   
     | 
    Resource Modelling (ZIB, Berlin) | 
    
   
     | 
    Data Management Enhancements (RUS, Stuttgart) | 
    
   
     | 
    Extended Job Control (DWD, Offenbach) | 
    
   
     | 
    Application Specific Interfaces (FRZ, Julich) | 
    
   
     | 
    Co-sheduling (PC2, Paderborn) | 
    
   
     | 
    MPI and PACX integration (RUS, Stuttgart) | 
    
   
     | 
    Vampir extensions (TUD, Dresden) | 
    
   
     | 
    Advanced administration (FRZ, Julich) | 
    
   
     | 
     | 
    
   
   
   | 
 
 
  
  
  
  
   
     | 
     | 
    
   
     | 
    Computation portal to bio-molecular applications | 
    
   
     | 
    Build interfaces to well known bio-molecular
    applications, simplify access to databases. | 
    
   
     | 
    Integrate interfaces within EUROGRID software. | 
    
   
   
   | 
 
 
  | 
  
   | 
 
 
  
  
  
  
   
     | 
     | 
     | 
    
   
     | 
    Focuses on industrial CAE applications | 
    
   
     | 
    Code coupling and multi-physics optimisations to
    improve system design. | 
    
   
     | 
    ASP - type services : | 
    
   
     | 
    User interfaces to hide the complexity of HPC
    systems to industrial users | 
    
   
     | 
    Supercomputing as an e-business : accurate cost
    prediction of CAE simulations | 
    
   
   
   | 
 
 
  
  
  
  
   
     | 
     | 
    
   
     | 
    Targets : | 
    
   
     | 
    The establishment, by the HPC centres partners
    of EUROGRID, of an application testbed�
    for general purpose HPC distributed applications | 
    
   
     | 
    The installations and tests of EUROGRID software
    releases | 
    
   
     | 
    The development of new relevant GRID
    applications, using existing middlewares, in scientific areas not covered
    by WP1 to WP3. | 
    
   
   
   | 
 
 
  | 
  
   | 
 
 
  
  
  
  
   
     | 
     | 
    
   
     | 
    Interactive monitoring and steering of complex
    simulations (running in batch mode), using JAVA - CORBA technology. | 
    
   
     | 
    Coupling of atmospheric and hydrological models | 
    
   
     | 
    Fluid-structure coupling for space propulsion | 
    
   
     | 
    Direct numerical simulation of turbulent
    combustion. | 
    
   
     | 
    �many others to come in the area of
    electrodynamics, material sciences, quantum chemistry, ... | 
    
   
   
   | 
 
 
  
  
  
  
   
     | 
     | 
     | 
    
   
     | 
    Efficient file transfer (FECIT leading partner) | 
    
   
     | 
    Optimisations of transfer bandwidth and cost | 
    
   
     | 
    Fail-safe and encripted transfer | 
    
   
     | 
    Overlap of transfer and processing | 
    
   
     | 
    Emphasis on quality of service. | 
    
   
     | 
    Resource broker (UoM leading partner) | 
    
   
     | 
    Must handle static and dynamic information to
    match the user�s computational requirements | 
    
   
     | 
    Builds on CSAR past experience. | 
    
   
   
   | 
 
 
  
  
  
  
   
     | 
     | 
     | 
    
   
     | 
    ASP infrastructure (DEBIS leading partner) | 
    
   
     | 
    Supercomputing as an e-business? | 
    
   
     | 
    Build browser-based job submission GUI | 
    
   
     | 
    Build tools for cost-estimation, accounting and
    Web-based billing of services. | 
    
   
     | 
    Interactive access �(Parallab leading partner) | 
    
   
     | 
    Deals with all kinds of issues related to
    simulation steering by visualized output. | 
    
   
     | 
    Integrate these facilities in EUROGRID software. | 
    
   
   
   | 
 
 
  
  
  
  
   
     | 
     | 
     | 
    
   
     | 
    Maintain working versions of EUROGRID software | 
    
   
     | 
    Integrate domain specific extensions (WP1 to
    WP4) and new technology (WP5). | 
    
   
     | 
    EUROGRID 0.0 :��
    UNICORE today | 
    
   
     | 
    EUROGRID 0.5 : | 
    
   
     | 
    Data transfer, application couplings,
    interactive access | 
    
   
     | 
    EUROGRID 1.0 : | 
    
   
     | 
    Resource broker, ASP infrastructure | 
    
   
     | 
    EUROGRID 2 :���
    final version | 
    
   
   
   | 
 
 
  
  
  
  
   
     | 
     | 
    
   
     | 
    Integration of modern grid software technologies
    in European supercomputing infrastructures | 
    
   
     | 
    Major effort in distributed application
    development in science and technology | 
    
   
   
   |