Reconfigurable Computing

from Wikipedia, the free encyclopedia

The term reconfigurable computing describes the field of reconfigurable computing systems and their applications.

Content

These are computing systems that are structurally programmable, for example on the basis of FPGAs - in contrast to conventional Von Neumann computers , whose instruction streams are programmed as instruction scheduling . The following table shows a classification scheme for computer paradigms published by Nick Tredennicks around 1995 :

Nick Tredennick's computer classification scheme
Computer class Resources / programming source Algorithms / programming source
early historical calculator fixed / none fixed / none
Von Neumann calculator fixed / none variable / software (command streams)
Reconfigurable Computing variable / Configware (configuration) variable / flowware (data streams)

With reconfigurable platforms, a distinction is made between fine-grained reconfigurable, as in FPGAs with only 1-bit wide programmable logic blocks (CLBs), and coarse-grained reconfigurable with word-widths (e.g. 32-bit) reconfigurable data paths (rDPU: reconfigurable data path unit ), which are almost identical to the Are comparable to the main processor of a computer, but do not contain a program counter.

See also

literature

  • Nick Tredennick: The Case for Reconfigurable Computing . In: Microprocessor Report . tape 10 , no. 10 , August 5, 1996, p. 25-27 .
  • Jürgen Teich (Ed.) Et al .: Reconfigurable Computing Systems . Focus issue of the magazine it - Information Technology , Vol. 49, 2007, Issue 3. Oldenbourg Verlag, Munich.
  • Christophe Bobda: Introduction to Reconfigurable Computing: Architectures, algorithms and applications . Springer-Verlag, 2007, ISBN 978-1-4020-6088-5

Web links

Individual evidence

  1. ^ R. Hartenstein: A Decade of Reconfigurable Computing: a Visionary Retrospective . In: Design, Automation and Test in Europe, 2001 . Conference and Exhibition 2001. Proceedings, pp. 642-649.