Parallel logic programming using an intensional model of computation

This dissertation presents a new eductive model for the execution of logic programs.;This new model provides a mechanism for the parallel execution of logic programs. The model is also capable of executing indexically extended Horn clause languages--including infinitary ones.;This new model is the first suitable for execution in a fine-grain, parallel, dataflow environment and has the following innovative characteristics: support for parallelism at the algorithmic level and at the instruction level; support for all major forms of Prolog parallelism including AND-Parallelism, OR-Parallelism, AND/OR-Parallelism, Stream Parallelism, and Unification Parallelism--or any combination of these. In addition, the model supports all-solution or single-solution evaluation, and permits straightforward exploitation of mixed levels of granularity in the implementation.;This research follows a new approach rather than relying on a coarse-grain, process-oriented execution of the traditional Prolog model; the approach that has been widely explored for the parallel execution of logic programs. This approach is innovative because it uses the eductive model of computation in the evaluation of logic programs. The eductive model of computation is based on the use of a collection of autonomous processing elements for the parallel computation of referentially transparent values. These values result from the evaluation of the Prolog program's logical variable in separate distinct contexts.;The major parts of this work address the development of the new execution model, the definition of a new language that drives the model, an execution architecture for the model, and a compiler that translates from Prolog into the newly defined language.