Chapter

Parallel Computing on Distributed Memory Multiprocessors

Volume 103 of the series NATO ASI Series pp 205-216

Logic Program Execution on Distributed Memory Parallel Computers

  • Mario CannataroAffiliated withCRAI
  • , Giandomenico SpezzanoAffiliated withCRAI
  • , Domenico TaliaAffiliated withCRAI

* Final gross prices may vary according to local VAT.

Get Access

Abstract

The implicit parallelism of logic programs can be exploited by using parallel computers to support their execution. This paper describes the design and the implementation of a logic programming system on a distributed memory parallel architecture in an efficient and scalable way. This implementation is based on the AND/OR Process Model. The AND/OR processes are implemented as Occam processes, and the execution of a logic program is implemented by a tree of parallel processes located on the computing nodes of the parallel computer. A completely decentralized scheduling mechanism is used. The preliminary results show interesting scalability properties of the system.

Keywords

Parallel processing logic programming distributed memory AND/OR parallelism