A comparison between the graph reduction model and an original parallel evaluation scheme for functional languages

Describes a comparison between two models for the parallel evaluation of functional languages: the well-known `graph reduction model´ and the model we have defined, which is called P³. In the graph reduction model, an application of a function to an argument is represented by a graph in which a function is bound to its argument by an intermediate application node. Our model is based on two distinct representations for functions and arguments, and no intermediate application nodes are handled. In this paper, we show that intermediate application nodes are the cause of many inefficiency problems of the graph reduction model, and we show how our model palliates these disadvantages