PARALLEL PROCESSING OF NATURAL LANGUAGE PARSERS
Marcel P. van Lohuizen · 2000
In this paper we describe an implementation of a parallel parser for natural language. We first indicate why these applications are typically hard to parallelize. We then present a work stealing approach for symmetric shared-memory multiprocessors that minimizes overhead by dynamically controlling the granularity of work distribution. In addition, we will explain how we deal with concurrent access of the centralized chart.