Fast execution mechanisms of parallel inference engine PIE: PIEpelined goal rewriting and goal multicasting
PIE is a highly parallel inference machine which executes logic programs based on the goal rewriting model. Goal rewriting makes goals completely independent and reduces access conflict, although it accompanies a substantial copying overhead. To cope with this overhead, two new execution mechanisms are proposed and evaluated. These methods are effective especially to raise parallelism quickly and to exploit other kind of parallelism in programs where OR-parallelism is not high.
KeywordsExecution Time Parallel Inference Unification Time Definition Clause Inference Machine
Unable to display preview. Download preview PDF.
- [Mot84]Moto-oka,T., Tanaka,T., Aida,H., Maruyama,T. and Hirata,K., "The Architecture of a Parallel Inference Engine-PIE-", FGCS'84, ICOT, 1984.Google Scholar
- [Got84]Goto,A., Tanaka.H. and Moto-oka,T., "Highly Parallel Inference Engine PIE — Goal Rewriting Model and Machine Architecture —", New Generation Computing, Vol.2, OHMSHA and Springer-Verlag, 1984.Google Scholar
- [Yuh83]Yuhara, M.Aida,H., Tanaka,H. and Moto-oka,T., "Unify Processor and its Reduction Algorithm of the Highly Parallel Inference Engine — PIE (Japanese)", Technical Research Report, EC83-30, IECE of Japan, 1983.Google Scholar
- [Yuh84]Yuhara,M., Koike,H., Tanaka,H. and Moto-oka,T., "A Unify Processor Pilot Machine for PIE", Proc. of the Logic Programming Conference '84, Tokyo, 1984.Google Scholar
- [Tic84]Tick,E. and Warren,D.H.D., "Towards a Pipelined Prolog Processor", International Symposium on Logic Programming, IEEE, 1984.Google Scholar
- [War83]Warren,D.H.D., "An Abstract Prolog Instruction Set", Tech. report 309, A.I.Center, SRI International, 1983.Google Scholar