An Efficient Architecture for Ultra Long FFTs in FPGAs and ASICs

Abstract

This presentation outlines an architecture for efficient Ultra Long FFTs for use in FPGAs and ASICs. Analysis of accuracy, performance, cost and power consumption are presented. FFTs are at the heart of many real time signal processing applications and Ultra Long FFTs are quite often used for frequency analysis and communications applications. As the processing requirements increase, the use of FPGAs and ASICs become the logical choice for implementing real time FFTs. This presentation describes an efficient framework for implementing the Cooley-Tukey algorithm for Ultra Long FFTs using minimal external memory. Typically for lengths over 16K the memory resources of the FPGA or ASIC are exhausted and external memory is required. The architecture is implemented using two shorter length FFTs (lengths N(sub1) and N(sub2)) to calculate an FFT of length N=N(sub1) x N(sub2). This architecture is optimized for continuous data FFTs, minimizing the external memory requirements and offering flexibility so that it can be used for many different applications.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Sep 30, 2004
Accession Number
ADA433447

Entities

People

  • Tom Dillon

Tags

Communities of Interest

  • Advanced Electronics

DTIC Thesaurus Topics

  • Abstracts
  • Accuracy
  • Algorithms
  • Bandwidth
  • Copyrights
  • Data Sets
  • Energy Consumption
  • Engineering
  • Frequency
  • Image Compression
  • Image Processing
  • Information Operations
  • Lepidoptera
  • Numerical Analysis
  • Sequences
  • Signal Processing
  • Time Signals

Readers

  • Approximation Theory.
  • Computational Modeling and Simulation
  • Integrated Circuit Design and Technology.