Study of the Utilization of Adaptive Computing in Space Applications - PowerPoint PPT Presentation

Loading...

PPT – Study of the Utilization of Adaptive Computing in Space Applications PowerPoint presentation | free to download - id: 582ee3-Y2Q1M



Loading


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation
Title:

Study of the Utilization of Adaptive Computing in Space Applications

Description:

Study of the Utilization of Adaptive Computing in Space Applications Marco Figueiredo ASDP - Adaptive Scientific Data Processing Group website: http://fpga.gsfc.nasa.gov – PowerPoint PPT presentation

Number of Views:748
Avg rating:3.0/5.0
Slides: 17
Provided by: MarcoAFi
Learn more at: http://klabs.org
Category:

less

Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: Study of the Utilization of Adaptive Computing in Space Applications


1
Study of the Utilization of Adaptive Computing in
Space Applications
  • Marco Figueiredo
  • ASDP - Adaptive Scientific Data Processing Group
  • website http//fpga.gsfc.nasa.gov

Microelectronics and Signal Processing Branch -
Code 564 Goddard Space Flight Center
2
Introduction to the ASDP Group
  • ASDP Group
  • Marco Figueiredo (SGT, Inc.)
  • Umesh Patel (SGT, Inc.)
  • Ken Winiecki (Lockheed Martin Corp.)
  • Terry Graessle (Lockheed Martin Corp.)
  • Parminder Ghuman (GSFC)
  • Mission Foster adaptive computing technology
    evolution by applying it to remote sensing
    science applications and providing feedback to
    industry academia

3
Overview
  • The Earth science remote sensing data flow
  • The MODIS reflective calibration prototype
  • The adaptive computing application design process
  • Improvements needed for the cost-effective
    development of adaptive computing applications
  • Conclusions

4
Earth Science Remote SensingData Flow Diagram
5
MODIS Reflective Calibration Algorithm
9.28 times acceleration
6
Application Design Approach
  • Analyze the algorithm and identify computation
    intensive parts that can benefit from execution
    on an adaptive processor.
  • Algorithm Partitioning and Mapping
  • Incremental Design, Test, and Integration

7
Application DesingAlgorithm Partitioning and
Mapping
  • Extract data flow diagram and partition tasks
    between host and adaptive processor
  • Perform a fixed point precision analysis and
    determine data path width
  • Create a detailed hardware data flow diagram
  • Perform FPGA resource estimation
  • Perform design mapping to adaptive processor

8
(No Transcript)
9
(No Transcript)
10
Wildforce Adaptive Processor
11
Mapping Issues
  • Amount of logic that would fit into each device
  • Amount of data to be passed between FPGAs
  • How input parameters would be made available to
    each FPGA
  • How the data would flow in and out of the board

12
Incremental Design, Test, and Integration
  • Design a small portion of the logic in VHDL and
    simulate the design
  • Synthesize, place and route the FPGA design
  • Write host software to download and test the
    design to the board comparing results with
    fixed-point simulation
  • Repeat steps 1 to 3 until design is complete
  • Allows the division of taks among engineers
  • Reduces compilation time and facilitates test

13
Application DevelopmentRequired Skills
  • Algorithm analysis
  • Complex logic design (arithmetic)
  • VHDL design
  • FPGA design
  • Reconfigurable computer architectures
  • Software engineering

14
Improvements Needed
  • Extended library of high performance math
    operators for specific device architectures
  • Transparent design environment that crosses the
    microprocessor (software) and fpga (hardware)
    boundaries
  • JAVA
  • VHDL

15
Improvements Needed (cont.)
  • High level design entry tool for
  • Algorithm capturing
  • Fixed point precision analysis
  • Host/Adaptive processor design partitioning
  • Adaptive processor mapping tool
  • Design documentation
  • Adaptive Computer Engineers

16
Conclusions
  • Adaptive (reconfigurable) computing yields orders
    of magnitude acceleration over microprocessor
    based systems TODAY.
  • The utilization of a spacecraft adaptive computer
    presents the same potentials and limitations
    found on ground applications.

Development cost
Number of applications
About PowerShow.com