Parallel parsing of ambiguous languages on hypercube architectures
In this paper we describe a generalization of existing LALR techniques to allow the parallel parsing of ambiguous languages on hypercube architectures. In particular we are interested in the parsing of natural languages in parallel. The technique that is used is a “nondeterministic” version of a standard LALR parser, where when a shift/reduce or reduce/reduce conflict is reached the parser branches and pursues them both independently. We have developed a parallel parser generator RACC, which is developed with the use of YACC under Unix. By specifying a YACC like specification file and running it through RACC one obtains a parallel parser that will run on the nodes of the NCUBE/10 hypercube.
Unable to display preview. Download preview PDF.
- 1.A. H. Aho, R. Sethi and J. Ullman: Compilers: Principles, Techniques, and Tools. Addison-Wesley Publishing Company, Reading, Massachusetts, 1986.Google Scholar
- 2.Y. Matsumoto, A Parallel Parsing System for Natural Language Analysis. Third International Conference on Logic Programming, pp. 396–409, London, July 14–18, 1986.Google Scholar
- 4.M. Tomita: Efficient Parsing for Natural Language: A Fast Algorithm for Practical Systems. Kluwer Academic Publishers, Boston, 1986.Google Scholar
- 5.R. Trehan, and P. Wilk: A Parallel Chart Parser for the Committed Choice Non-Deterministic (CCND) Logic Languages. Artificial Intelligence Applications Institute, University of Edinburgh, Edinburgh, United Kingdom, 1988.Google Scholar