1.3-GOPS parallel DSP for high-performance image-processing applications

Publikation: Beitrag in FachzeitschriftArtikelForschungPeer-Review

Autoren

  • Willm Hinrichs
  • Jens Peter Wittenburg
  • Hanno Lieske
  • Helge Kloos
  • Martin Ohmacht
  • Peter Pirsch
Forschungs-netzwerk anzeigen

Details

OriginalspracheEnglisch
Seiten (von - bis)946-952
Seitenumfang7
FachzeitschriftIEEE Journal of Solid-State Circuits
Jahrgang35
Ausgabenummer7
PublikationsstatusVeröffentlicht - Juli 2000

Abstract

A programmable digital signal processor (DSP) for real-time image processing is presented that combines the concepts of single-instruction multiple-data (SIMD) and very long instruction word with a high utilization of parallel resources on instruction and data level. The SIMD approach has been extended with autonomous instruction selection capabilities (ASIMD), which offers to control four parallel datapaths with low area overhead. The memory concept is adapted to image-processing requirements and follows two basic rules: shared data have to be accessed regularly in the shape of a matrix and are stored in the matrix memory. As soon as data are accessed irregularly, they are stored in the private cache memories. The matrix memory allows parallel, conflict-free access from all datapaths in a single clock cycle. The DSP achieves 1.3-GOPS performance at 66 MHz. A first prototype in 0.5-μm CMOS technology has been fabricated.

ASJC Scopus Sachgebiete

Zitieren

1.3-GOPS parallel DSP for high-performance image-processing applications. / Hinrichs, Willm; Wittenburg, Jens Peter; Lieske, Hanno et al.
in: IEEE Journal of Solid-State Circuits, Jahrgang 35, Nr. 7, 07.2000, S. 946-952.

Publikation: Beitrag in FachzeitschriftArtikelForschungPeer-Review

Hinrichs, W, Wittenburg, JP, Lieske, H, Kloos, H, Ohmacht, M & Pirsch, P 2000, '1.3-GOPS parallel DSP for high-performance image-processing applications', IEEE Journal of Solid-State Circuits, Jg. 35, Nr. 7, S. 946-952. https://doi.org/10.1109/4.848202
Hinrichs, W., Wittenburg, J. P., Lieske, H., Kloos, H., Ohmacht, M., & Pirsch, P. (2000). 1.3-GOPS parallel DSP for high-performance image-processing applications. IEEE Journal of Solid-State Circuits, 35(7), 946-952. https://doi.org/10.1109/4.848202
Hinrichs W, Wittenburg JP, Lieske H, Kloos H, Ohmacht M, Pirsch P. 1.3-GOPS parallel DSP for high-performance image-processing applications. IEEE Journal of Solid-State Circuits. 2000 Jul;35(7):946-952. doi: 10.1109/4.848202
Hinrichs, Willm ; Wittenburg, Jens Peter ; Lieske, Hanno et al. / 1.3-GOPS parallel DSP for high-performance image-processing applications. in: IEEE Journal of Solid-State Circuits. 2000 ; Jahrgang 35, Nr. 7. S. 946-952.
Download
@article{9bf88f74d65245a5a54420a971fc6a96,
title = "1.3-GOPS parallel DSP for high-performance image-processing applications",
abstract = "A programmable digital signal processor (DSP) for real-time image processing is presented that combines the concepts of single-instruction multiple-data (SIMD) and very long instruction word with a high utilization of parallel resources on instruction and data level. The SIMD approach has been extended with autonomous instruction selection capabilities (ASIMD), which offers to control four parallel datapaths with low area overhead. The memory concept is adapted to image-processing requirements and follows two basic rules: shared data have to be accessed regularly in the shape of a matrix and are stored in the matrix memory. As soon as data are accessed irregularly, they are stored in the private cache memories. The matrix memory allows parallel, conflict-free access from all datapaths in a single clock cycle. The DSP achieves 1.3-GOPS performance at 66 MHz. A first prototype in 0.5-μm CMOS technology has been fabricated.",
author = "Willm Hinrichs and Wittenburg, {Jens Peter} and Hanno Lieske and Helge Kloos and Martin Ohmacht and Peter Pirsch",
note = "Funding Information: Manuscript received December 20, 1999; revised February 7, 2000. This work was supported by BWB, Germany. The authors are with the Laboratorium f{\"u}r Informationstechnologie, Univer-sit{\"a}t Hannover, Hannover 30167 Germany. Publisher Item Identifier S 0018-9200(00)03879-8.",
year = "2000",
month = jul,
doi = "10.1109/4.848202",
language = "English",
volume = "35",
pages = "946--952",
journal = "IEEE Journal of Solid-State Circuits",
issn = "0018-9200",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "7",

}

Download

TY - JOUR

T1 - 1.3-GOPS parallel DSP for high-performance image-processing applications

AU - Hinrichs, Willm

AU - Wittenburg, Jens Peter

AU - Lieske, Hanno

AU - Kloos, Helge

AU - Ohmacht, Martin

AU - Pirsch, Peter

N1 - Funding Information: Manuscript received December 20, 1999; revised February 7, 2000. This work was supported by BWB, Germany. The authors are with the Laboratorium für Informationstechnologie, Univer-sität Hannover, Hannover 30167 Germany. Publisher Item Identifier S 0018-9200(00)03879-8.

PY - 2000/7

Y1 - 2000/7

N2 - A programmable digital signal processor (DSP) for real-time image processing is presented that combines the concepts of single-instruction multiple-data (SIMD) and very long instruction word with a high utilization of parallel resources on instruction and data level. The SIMD approach has been extended with autonomous instruction selection capabilities (ASIMD), which offers to control four parallel datapaths with low area overhead. The memory concept is adapted to image-processing requirements and follows two basic rules: shared data have to be accessed regularly in the shape of a matrix and are stored in the matrix memory. As soon as data are accessed irregularly, they are stored in the private cache memories. The matrix memory allows parallel, conflict-free access from all datapaths in a single clock cycle. The DSP achieves 1.3-GOPS performance at 66 MHz. A first prototype in 0.5-μm CMOS technology has been fabricated.

AB - A programmable digital signal processor (DSP) for real-time image processing is presented that combines the concepts of single-instruction multiple-data (SIMD) and very long instruction word with a high utilization of parallel resources on instruction and data level. The SIMD approach has been extended with autonomous instruction selection capabilities (ASIMD), which offers to control four parallel datapaths with low area overhead. The memory concept is adapted to image-processing requirements and follows two basic rules: shared data have to be accessed regularly in the shape of a matrix and are stored in the matrix memory. As soon as data are accessed irregularly, they are stored in the private cache memories. The matrix memory allows parallel, conflict-free access from all datapaths in a single clock cycle. The DSP achieves 1.3-GOPS performance at 66 MHz. A first prototype in 0.5-μm CMOS technology has been fabricated.

UR - http://www.scopus.com/inward/record.url?scp=0034225390&partnerID=8YFLogxK

U2 - 10.1109/4.848202

DO - 10.1109/4.848202

M3 - Article

AN - SCOPUS:0034225390

VL - 35

SP - 946

EP - 952

JO - IEEE Journal of Solid-State Circuits

JF - IEEE Journal of Solid-State Circuits

SN - 0018-9200

IS - 7

ER -