Reconfigurable Computing Technology

• Hybrid hardware/software data processing systems that implement computationally intensive algorithms in programmable gate array hardware, yielding a 10x to 1000x speed improvement over CPU’s.
• Re-programmable “in-circuit”, allowing configuration changes from mission to mission, day to day or second to second.
• A key element in enabling the Earth Science Vision concepts for future flight and ground systems
  • data driven / event driven hardware configurations
  • real-time on-board processing (direct products, reduced bandwidth)
  • fault tolerance, flexibility, redundancy
• Three elements of our approach:
  • Ground based functional development
  • RC design tools and techniques
  • Flight FPGA part development
• Current efforts:
  • Functional - ALOA, DDF, and NPP
  • Tools & techniques - NC State
  • Flight parts - Atmel, Honeywell, and Sandia
• Key partners, leverage & interests:
  • DARPA, NSA, and DoD
  • Annapolis Microsystems, Xilinx, Honeywell, Actel, and TSI TelSys
  • Northwestern, NC State, Colorado State, Carnegie Mellon, and Brigham Young