Euforia Project - European Community for Iter Applications

The EUFORIA consortium involves the next 14 European institutions:

1.Chalmers University of Technology (Coordinator) from Sweden

2.Max Plank Institute for Plasma Physics (IPP) from Germany

3.Centro Superior de Investigaciones Científicas (CSIC) from Spain

4.Centro de Investigaciones Energéticas, Medio Ambientales y Tecnológicas (CIEMAT) from Spain

5.Forschungszentrum Karlsruhe (FZK) from Germany

6.Finnish IT Center for Science (CSC) from Finland

7.Abo Akademi University (ABO) from Finland

8.University of Edinburgh (UEDIN) from United Kingdom

9.Barcelona Supercomputing Center (BSC) from Spain

10.French Atomic Energy Commission (CEA) from France

11.University Louis Pasteur - Strasbourg I (ULP) from France

12.University of Ljubljana (UOL) from Slovenia

13.Poznan Supercomputing and Networking Center PSNC from Poland

14.Italian National Agency for New Technologies, Energy and the Environment (ENEA) from Italy

Structure of the project

EUFORIA project has been structured into a set of networking, services and research activities according to the following overall organisation:

Networking activities

* NA1 - Management

* NA2 - User Documentation and Training

* NA3 - Dissemination

Service Activities  Infrastructure deployment and Operation

* SA1 - Grid Infrastructure

* SA2 - HPC infrastructure

* SA3 - User Support for HPC and Grid activities

Joint Research Activities

* JRA1 - Adaptation of codes and tools for Grid Infrastructure

* JRA2 - Adaptation of codes and tools for HPC infrastructure

* JRA3 - Physics Integration ?\ Workflow orchestration tools

* JRA4 - Visualization

Network Activities (NA)

Activity NA1: Management

NA1 provides administrative support and management for the consortium as a whole. In particular the global coordination of between different research and service activities is an NA1 responsibility and emphasis has been put on promoting and ensuring communication within research and service activities to guarantee full integration of the projects.

The links to other international projects and initiatives, in Grid or HPC technology as well as the ties to the fusion community, will be investigated. Management will liaise with code owners and users from the fusion community with aim to release additional computational resources for direct physics exploration using the EUFORIA adapted codes. Additional sources of funding will be investigated including funds from the EU, and from the countries of the partners signing the proposal.

As the project progresses in time and the different activities becomes more integrated, emphasis will need to be put on supporting horizontal activities between the different JRA's as well as joint development between Research and Service activities. An efficient coordination will be needed in order to achieve the goals of the project. One of the purposes of the coordination between dissemination and user training activities is that trainees can also act themselves as further dissemination carriers within their communities and the coordination of this is essential for the broader availability of the EUFORIA tools to the fusion user and code developer community.

To harmonize the einfrastructure development with ITER modelling requirements and to maintain in contact with the physics development and modelling activities being performed within the international ITER partners a fusion based coordinating team has been opted for with a strong technology support within the steering group.

NA1 Objectives

- Provide administrative and management services for the consortium as a whole

- Provide Global (horizontal) coordination between and within the Joint Research and Service Activities

- Coordinate the efforts of activity NA3 (Dissemination) and activity NA2 (Knowledge Dissemination)

- Coordinate activities with EGEE, DEISA, ICEAGE and BELIEF using the already available contacts

- Investigate possible liaisons with other EU and International Grid and HPC projects

- Investigate additional sources of funding for the project through contacts with politicians and other decision makers

Successful management is manifested in the performance of the project as a whole and is obtained through consistent monitoring and review of milestones and decision points within the tasks and a continuous assessment of the progress made and the implementation of corrective measures.

Activity NA2: User Documentation and Training

NA2 will provide scientific users with appropriate documentation and training to exploit the HPC and Grid infrastructure of this project. The activity encompasses a training programme, designed to enable users to develop the required knowledge and skills to utilise the infrastructure. In addition, the project will provide a comprehensive set of documentation, including users¡- guides and FAQs.

NA2 Objectives

- Provide documents, FAQs and best practice information to users on the Grid and HPC infrastructure

- Provide a comprehensive training programme for developers and users, including the development of a companion guide

Activity NA3: Dissemination

Dissemination will be essential in identifying, reaching and providing an accurate image of the project and its potential use to the prospective users. This image must be simultaneously appealing by highlighting the benefits that can be obtained from joining the DEISA, EGEE and EUFORIA grid and HPC communities. Dissemination will also play a key role in bringing potential users to attend the knowledge dissemination events that in turn will be fundamental in their induction. It is expected that dissemination will provide a general picture of the project.

Most users, developers and communities will have their first contact with the EUFORIA project through the dissemination activities. The embrace of the project to a broadened user community will depend on the quality and effectiveness of the actions taken to raise the awareness of the project.

NA3 Objectives

- To disseminate the purposes and benefits of Grid Computing and Supercomputing in the Fusion area, and of joining or using the EUFORIA infrastructure

- To define the dissemination methods and message content for reaching each community

- To provide dissemination materials adapted to the target audiences and make them available

- To inform the scientific communities on how to get involved in the project

- To bring directly potential users so that they may become EUFORIA users

- To keep the communities informed of new improvements and functionalities

Service Activities (SA)

Activity SA1: Grid Infrastructure Deployment and Operation

The objective of this activity is to provide an advanced Grid empowered infrastructure for scientific computing targeted to support the fusion activities in the European Research Area. The project will integrate distributed computing resources into a coherent Grid service enabling fusion researchers to embrace new scientific challenges. The Grid Infrastructure Operations activity aims to deploy, integrate, maintain and operate this large distributed Grid computing infrastructure.

SA1 Objectives

- Deploy, maintain and operate the central services necessary to ensure the integration of the computing resources into a production Grid infrastructure capable of supporting serial and parallel applications

- Coordinate the deployment, maintenance and operation of the Grid resources provided by the participating sites ensuring a robust, secure and dependable service

- Provide support for Virtual Organizations and resource providers thus helping end users, VO managers and site managers to achieve their goals and contribute to a successful usage of the infrastructure.

The context of this project relies on the consolidation of a Grid based research space across Europe. The infrastructure will be compatible with EGEE middleware services contributing to the standardization of the Grid middleware and consolidation of the eInfrastructures in Europe. For this purpose the project will follow the recommendations of international bodies such as eIRG (e-Infrastructure Reflection Group) and GGF (Global Grid Forum).

Activity SA2: HPC Infrastructure

The code developers within JRA2 are focused on optimizing codes from within the plasma fusion community. In order for the optimization work to function properly the code developers and scientists must have access to suitable computers for testing, debugging, and running. In addition, feasibility tests and real production runs from scientists will require supercomputer access.

The main target of SA2 is therefore to secure access for the code developers and scientists to HPC facilities and customer services at a range of supercomputer centres (That is, CSC in Helsinki, BSC in Barcelona and EPCC (UEDIN) in Edinburgh).

SA2 Objectives

- To provide [HPC] infrastructure support for the code optimization effort within JRA2

- To provide run time access to supercomputers for feasibility tests for the HPC codes within the code platform

Activity SA3: User Support for HPC and Grid Activities

SA3 will provide scientific users with the appropriate support mechanisms to utilise the HPC and Grid infrastructure of this project in a rapid and efficient manner. There will be one single interface for users, providing a central point of contact while exploiting the pool of expertise located across partner institutions. This will provide user registration; resource management and accounting; and a single helpdesk facility for all users. This activity underpins many of the other activities, such as JRA1, JRA2, JRA3 and JRA4, providing a first port of call for users to ensure rapid uptake of the applications, technologies and tools in all these activities.

SA3 Objectives

- To provide, through a single user interface, user administration, resource management, accounting, reporting and a central helpdesk for users of the service.

- To provide support for a broad range of applications in porting and initial optimization of their applications on the EUFORIA infrastructure.

Joint Research Activities (JRA)

Activity JRA1: Adaptation of codes for Grid infrastructure

In JRA1, a set of pilot codes used in Plasma Core and Edge Transport modelling will be adapted to the grid computing architecture. The main objective is to explore the possibilities of grid architecture for fusion research, starting with a subset of codes selected from a range of edge and core turbulence and transport code, which is the most appropriated for this purpose. A set of sequential codes will be adapted as a first step and two mixed workflow codes will be installed to the grid in the second step. Special attention will be paid to the connection of serial and parallel workflows for the utility that this kind of mixed computing strategy could have for Fusion in the future. The lessons learned both in porting codes and in data management will be taken as valuable experience for future Fusion simulations and experiments.

JRA1 Objectives

- To port a set of sequential codes to the grid environment together with at least a few mixed sequential?\parallel codes

- To get experience in running the codes and managing the data produced by them within a work flow environment

Activity JRA2: Adaptation of code for HPC infrastructure

In JRA2, a set of central computer codes in plasma physics and magnetic confinement fusion will be adapted for the parallel supercomputer systems available in the participating HPC centres. The aim is to make the programs run faster and to enable them to use a larger number of processors in order to solve substantially larger problem instances, thereby promoting the modelling capabilities in fusion research towards full-scale modelling. The computer codes will be designed to run as efficiently as possible on current tera-scale, as well as next generation peta-scale, HPC systems, and adapted to operate as components within a workflow environment.

JRA2 Objectives

- To improve the parallel scalability of the computer codes

- To optimise their performance on modern processor architectures

- To adapt the codes so that they can be integrated into a single coupled framework through a workflow orchestration tool

Activity JRA3: Workflow Orchestration

The coupling of different computational modules and codes requires a large element of coordination and structured data management and resource scheduling to be performed efficiently. A workflow orchestration tool greatly facilitates the integration process and provides a structure or framework for performing additional tasks such as computational steering and interactive monitoring or control. It also provides a possible entry point for standardized data access and user built post-processing tools.

JRA3 Objectives

- To schedule jobs on the GRID and HPC infrastructures together with jobs running on others computing facilities form a workflow orchestration tool that dynamically can link and launch the different computationally modules installed

- To develop a single user interface to support user administration, user queries, resource management, accounting and reporting. This interface will fulfil the requirements of the SA3 users that will provide the related service

Activity JRA4: Visualization

JRA4 is aimed at providing a set of unified visualization tools for the codes integrated in the platform. Some of these tools will be used for post-processing of the data generated by the codes, others will be integrated into the Kepler Workflow to allow visualization and monitoring of the results during the execution of the workflow. The visualization tools developed in JR4 will be based on existing open source software like Python, numpy and matplotlib on the one hand and VTK on VisIt on the other hand.

JRA4 Objectives

- To provide a set of unified visualization tools available within the Kepler workflow to visualization data remotely available on a computational Grid

- To implement more powerful visualization tools that can be used to post-process the computed data

- To develop a new lossy wavelet based data compression tool in order to handle the very large amount of data generated by some of the codes

- To develop specific visualization tools for a 4D or 5D distribution function