Abstract

Multidimensional (MD) discrete Fourier transform (DFT) is a key kernel algorithm in many signal processing applications. In this paper we describe an MD-DFT intellectual property (IP) generator and a bandwidth-efficient MD DFT IP for high performance implementations of 2-D and 3-D DFT on field-programmable gate array (FPGA) platforms. The proposed architecture is generated automatically and is based on a decomposition algorithm that takes into account FPGA resources and the characteristics of off-chip memory access, namely, the burst access pattern of the synchronous dynamic RAM (SDRAM). The IP generator has been integrated into an in-house FPGA development platform, AlgoFLEX, for easy verification and fast integration. The corresponding 2-D and 3-D DFT architectures have been ported onto the BEE3 board and their performance measured and analyzed. The results shows that the architecture can maintain the maximum memory bandwidth throughout the whole procedure while avoiding matrix transpose operations used in most other MD DFT implementations. To further enhance the performance, the proposed architecture is being ported onto the newly released Xilinx ML605 board. The simulation results show that 2 K × 2 K images with complex 64-bit precision can be processed in less than 27 ms.

Original languageEnglish (US)
Article number5608525
Pages (from-to)755-764
Number of pages10
JournalIEEE Transactions on Circuits and Systems I: Regular Papers
Volume58
Issue number4
DOIs
StatePublished - 2011

Fingerprint

Intellectual property
Discrete Fourier transforms
Field programmable gate arrays (FPGA)
Bandwidth
Data storage equipment
Random access storage
Signal processing
Decomposition

Keywords

  • Discrete Fourier transform (DFT)
  • dynamic RAM (DRAM)
  • field-programmable gate array (FPGA)
  • multidimensional signal processing

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Cite this

Multidimensional DFT IP generator for FPGA platforms. / Yu, Chi Li; Irick, Kevin; Chakrabarti, Chaitali; Narayanan, Vijaykrishnan.

In: IEEE Transactions on Circuits and Systems I: Regular Papers, Vol. 58, No. 4, 5608525, 2011, p. 755-764.

Research output: Contribution to journalArticle

Yu, Chi Li ; Irick, Kevin ; Chakrabarti, Chaitali ; Narayanan, Vijaykrishnan. / Multidimensional DFT IP generator for FPGA platforms. In: IEEE Transactions on Circuits and Systems I: Regular Papers. 2011 ; Vol. 58, No. 4. pp. 755-764.
@article{bde6e8b4d99f4048bd4443317789563a,
title = "Multidimensional DFT IP generator for FPGA platforms",
abstract = "Multidimensional (MD) discrete Fourier transform (DFT) is a key kernel algorithm in many signal processing applications. In this paper we describe an MD-DFT intellectual property (IP) generator and a bandwidth-efficient MD DFT IP for high performance implementations of 2-D and 3-D DFT on field-programmable gate array (FPGA) platforms. The proposed architecture is generated automatically and is based on a decomposition algorithm that takes into account FPGA resources and the characteristics of off-chip memory access, namely, the burst access pattern of the synchronous dynamic RAM (SDRAM). The IP generator has been integrated into an in-house FPGA development platform, AlgoFLEX, for easy verification and fast integration. The corresponding 2-D and 3-D DFT architectures have been ported onto the BEE3 board and their performance measured and analyzed. The results shows that the architecture can maintain the maximum memory bandwidth throughout the whole procedure while avoiding matrix transpose operations used in most other MD DFT implementations. To further enhance the performance, the proposed architecture is being ported onto the newly released Xilinx ML605 board. The simulation results show that 2 K × 2 K images with complex 64-bit precision can be processed in less than 27 ms.",
keywords = "Discrete Fourier transform (DFT), dynamic RAM (DRAM), field-programmable gate array (FPGA), multidimensional signal processing",
author = "Yu, {Chi Li} and Kevin Irick and Chaitali Chakrabarti and Vijaykrishnan Narayanan",
year = "2011",
doi = "10.1109/TCSI.2010.2078750",
language = "English (US)",
volume = "58",
pages = "755--764",
journal = "IEEE Transactions on Circuits and Systems I: Fundamental Theory and Applications",
issn = "1549-8328",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "4",

}

TY - JOUR

T1 - Multidimensional DFT IP generator for FPGA platforms

AU - Yu, Chi Li

AU - Irick, Kevin

AU - Chakrabarti, Chaitali

AU - Narayanan, Vijaykrishnan

PY - 2011

Y1 - 2011

N2 - Multidimensional (MD) discrete Fourier transform (DFT) is a key kernel algorithm in many signal processing applications. In this paper we describe an MD-DFT intellectual property (IP) generator and a bandwidth-efficient MD DFT IP for high performance implementations of 2-D and 3-D DFT on field-programmable gate array (FPGA) platforms. The proposed architecture is generated automatically and is based on a decomposition algorithm that takes into account FPGA resources and the characteristics of off-chip memory access, namely, the burst access pattern of the synchronous dynamic RAM (SDRAM). The IP generator has been integrated into an in-house FPGA development platform, AlgoFLEX, for easy verification and fast integration. The corresponding 2-D and 3-D DFT architectures have been ported onto the BEE3 board and their performance measured and analyzed. The results shows that the architecture can maintain the maximum memory bandwidth throughout the whole procedure while avoiding matrix transpose operations used in most other MD DFT implementations. To further enhance the performance, the proposed architecture is being ported onto the newly released Xilinx ML605 board. The simulation results show that 2 K × 2 K images with complex 64-bit precision can be processed in less than 27 ms.

AB - Multidimensional (MD) discrete Fourier transform (DFT) is a key kernel algorithm in many signal processing applications. In this paper we describe an MD-DFT intellectual property (IP) generator and a bandwidth-efficient MD DFT IP for high performance implementations of 2-D and 3-D DFT on field-programmable gate array (FPGA) platforms. The proposed architecture is generated automatically and is based on a decomposition algorithm that takes into account FPGA resources and the characteristics of off-chip memory access, namely, the burst access pattern of the synchronous dynamic RAM (SDRAM). The IP generator has been integrated into an in-house FPGA development platform, AlgoFLEX, for easy verification and fast integration. The corresponding 2-D and 3-D DFT architectures have been ported onto the BEE3 board and their performance measured and analyzed. The results shows that the architecture can maintain the maximum memory bandwidth throughout the whole procedure while avoiding matrix transpose operations used in most other MD DFT implementations. To further enhance the performance, the proposed architecture is being ported onto the newly released Xilinx ML605 board. The simulation results show that 2 K × 2 K images with complex 64-bit precision can be processed in less than 27 ms.

KW - Discrete Fourier transform (DFT)

KW - dynamic RAM (DRAM)

KW - field-programmable gate array (FPGA)

KW - multidimensional signal processing

UR - http://www.scopus.com/inward/record.url?scp=79953272919&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79953272919&partnerID=8YFLogxK

U2 - 10.1109/TCSI.2010.2078750

DO - 10.1109/TCSI.2010.2078750

M3 - Article

VL - 58

SP - 755

EP - 764

JO - IEEE Transactions on Circuits and Systems I: Fundamental Theory and Applications

JF - IEEE Transactions on Circuits and Systems I: Fundamental Theory and Applications

SN - 1549-8328

IS - 4

M1 - 5608525

ER -