Reconf 2

Home / Experience / Projects / Reconf 2

Summary

RECONF 2: Design Methodology and Environment for Dynamic RECONFigurable FPGA Funded under 5th FWP (Fifth Framework Programme) (IST–2001–341016)

Overall project summary: The aim of RECONF2 was to allow implementation of adaptive system architectures by developing a complete design environment to take full benefits of dynamic reconfigurable FPGAs (D_FPGAs).

Total project duration: III.2002 – XII.2004

Contact person: Marie Joseph Berard or Philippe Butel, MBDA France

My role

Software researcher at  Work Package 4: Development of the Two Front-End tools:

I was working within the project from II.2003 to XII.2004 (1 year 11 months)

More information

Project objectives
  • Complex SRAM based FPGA are very expensive
  • Detailed analysis of system functional blocks allows to consider:
    • Permanent blocks (Static)
    • Occasional blocks (Dynamic)
  • All the blocks implemented simultaneously into the FPGA result in chip:
    • Over dimension
    • Over consumption
    • Over cost
  • There are different modes of reconfiguration:
    • Global, static
    • Partial, static
    • Partial, Dynamic: Reconfigure some areas of the FPGA while the rest of the design is still running
  • The challenge: Sequentially download the configuration of the dynamic blocks while maintaining, until the switch to the new configuration, the operation of the static blocks
  • RECONF 2 addresses Partial Dynamic Reconfiguration
    • Implementation possible with Xilinx and Atmel technologies
    • Few experiments, mostly research
    • No methodology, no specific tools, no verification available
    • Difficult to implement, relies on modular backend, few usage
Project organization

Reconf_org_big

Project outputs
  • Provide a complete design methodology
    • Relies on existing tools for simulation and synthesis
    • Provide verification methodology for reliable implementation
    • Remains as technology independent as possible
  • Provide Front-end design tools (WP4 outputs)
    • Partitioning of a classic static design
    • Specifying reconfiguration scheduling and others dynamic constraints
    • Provide netlists for dynamic verification through classic static VHDL simulation tools
  • Back-end design tools
  • Use diversified types of application for demonstrating the efficiency of the resulting tools
  • Disseminate, promote and explain the benefits of the resulting tools:
    • Easy applications upgrade
    • Low power consumption

Please head to MIXDES’2005 talk.

close-alt close collapse comment ellipsis expand gallery heart lock menu next pinned previous reply search share star