Reconfigurable Massively Parallel Computers

Hungwen Li, Editor
IBM Research Division, Almaden Research Center

Quentin F. Stout, Editor
University of Michigan

Prentice-Hall, Inc., © 1991.

Reconfigurable parallel architectures use reconfigurability of their interconnection network to establish a communication topology optimized for the problem being solved. Processors can make local connectivity decisions based on their own data, rather than have such all decisions dictated by a centralized controller. For some problems, this powerful capability permits far faster solutions than can be achieved by fixed interconnection systems, such as hypercubes or meshes, or even by PRAMs. Reconfiguration also provides significant fault-tolerance ability.

This book surveys most of the existing and proposed reconfigurable massively parallel systems from around the world. It also analyzes many algorithms and applications, particularly in the area of image processing.

Table of Contents

  1. “Reconfigurable Massively Parallel Computers: An Introduction”, by H. Li and Q. F. Stout
  2. “Polymorphic VLSI Arrays with Distributed Control”, by M. Maresca and H. Li
  3. “Implementation and Application of a Gated-Connection Network in Image Understanding”, by D. B. Shu, J. G. Nash, M. M. Eshaghian, and K. Kim
  4. “Reconfiguration in the Low and Intermediate Levels of the Image Understanding Architecture”, by C. C. Weems and D. Rana
  5. “Reconfigurable Arrays Using Local Autonomy”, by T. J. Fountain and I. Y-F. Fung
  6. “Embedding Pyramids into Mesh Arrays”, by M. G. Albanesi, V. Cantoni, U. Cei, M. Ferretti, and M. Mosconi
  7. “Hypercube: a Reconfigurable Mesh”, by C.-T. Ho
  8. “Arrays for Digital Signal Processing Functions: Fault Tolerance and Functional Reconfiguration”, by A. Antola, R. Negrini, M.G. Sami and N. Scarabottolo
  9. “Fault-Tolerant Rectangular Array Processors via Reconfiguration”, by J. S. N. Jean and S. Y. Kung


Related Work

Reconfigurable mesh:
Important early papers are Parallel computations on reconfigurable meshes and Reconfigurable SIMD parallel processors.
Parallel computing:
Overview of my work and my papers.
Here is a somewhat whimsical explanation of parallel computing, and a tutorial.

Quentin's Home Copyright © 2005-2017 Quentin F. Stout.