Dictionary Machines for Cube-Class Networks.

Abstract

A dictionary is a data structure that supports insertion, deletion, and retrieval operations. To maintain a database, a dictionary machine accepts an arbitrary sequence of instructions at a constant rate. This idea presents two new VLSI dictionary machines on networks that emulate the binary cube. One machine runs on a shuffle-exchange network. The other machine runs on a cube-connected cycles network. These two designs demonstrate that general purpose networks can perform dictionary operations. Chapter 1 includes a definition of the dictionary machine problem and an overview of previous work. Chapters 2 and 3 present the bulk of this idea: two dictionary algorithms that run on cube-class architectures. The design in Chapter 2 uses bitonic merge to perform dictionary operations on the SEN. It also presents a novel architecture to implement pipelining. This architecture is an important contribution. The dictionary algorithm of Chapter 3 runs on the CCC and borrows heavily from previous work on tree machines. Chapter 4 places these new designs in perspective with previous work and offers concluding remarks.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Apr 05, 1985
Accession Number
ADA161393

Entities

People

  • Alan M. Schwartz

Organizations

  • University of Illinois Urbana–Champaign

Tags

Communities of Interest

  • Energy and Power Technologies
  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Air Force
  • Algorithms
  • Artificial Intelligence
  • Classification
  • Computations
  • Computer Programming
  • Computer Science
  • Computers
  • Computing System Architectures
  • Contracts
  • Databases
  • Instructions
  • Literature Surveys
  • Parallel Computing
  • Parallel Processing
  • Sequences
  • Trees (Data Structures)

Fields of Study

  • Computer science

Readers

  • Business Analytics
  • Neural Network Machine Learning.
  • Parallel and Distributed Computing.