Mapping 1D-FFT on an energy efficient 3D FPGA-DRAM architecture

Peter Gadfort, Aravind Dasu, Ali Akoglu

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

The one-dimensional Fast Fourier Transform (1D-FFT) is a fast method of computing the Discrete Fourier Transform (DFT) that is pervasive in signal processing applications. In this paper we present details on the mapping of the 1D-FFT on a power efficient 3D FPGA-DRAM architecture [1], with four data sizes: 256pt, 1024pt, 4096pt and 32768pt to achieve 29 GFLOPs/W with a 65nm technology.

Original languageEnglish (US)
Title of host publication2015 4th Berkeley Symposium on Energy Efficient Electronic Systems, E3S 2015 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781467385688
DOIs
StatePublished - Nov 24 2015
Event4th Berkeley Symposium on Energy Efficient Electronic Systems, E3S 2015 - Berkeley, United States
Duration: Oct 1 2015Oct 2 2015

Publication series

Name2015 4th Berkeley Symposium on Energy Efficient Electronic Systems, E3S 2015 - Proceedings

Other

Other4th Berkeley Symposium on Energy Efficient Electronic Systems, E3S 2015
Country/TerritoryUnited States
CityBerkeley
Period10/1/1510/2/15

Keywords

  • Adders
  • Computer architecture
  • Energy efficiency
  • Field programmable gate arrays
  • Ports (Computers)
  • Random access memory
  • Three-dimensional displays

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'Mapping 1D-FFT on an energy efficient 3D FPGA-DRAM architecture'. Together they form a unique fingerprint.

Cite this