Parallel symbolic computation in ACE

  • Authors:
  • Enrico Pontelli;Gopal Gupta

  • Affiliations:
  • Laboratory for Logic, Databases, and Advanced Programming, Department of Computer Science, New Mexico State University, Box 30001, Las Cruces, NM 88003, USA E-mail: {epontell,gupta&r ...;Laboratory for Logic, Databases, and Advanced Programming, Department of Computer Science, New Mexico State University, Box 30001, Las Cruces, NM 88003, USA E-mail: {epontell,gupta&r ...

  • Venue:
  • Annals of Mathematics and Artificial Intelligence
  • Year:
  • 1997

Quantified Score

Hi-index 0.00

Visualization

Abstract

We present an overview of the ACE system, a sound and complete parallel implementation of Prolog that exploits parallelism transparently (i.e., without any user intervention) from AI programs and symbolic applications coded in Prolog. ACE simultaneously exploits all the major forms of parallelism – Or‐parallelism, Independent And‐parallelism, and Dependent And‐parallelism – found in Prolog programs. These three varieties of parallelism are discussed in detail, along with the problems encountered in their practical exploitation. Our solutions to these problems, incorporated in the ACE system, are presented. The ACE system has been implemented on Sequent Symmetry and Sun Sparc Multiprocessors; performance results from this implementation for several AI programs are presented, which confirm the effectiveness of the choices made.