The implementation of functional languages on an object-oriented architecture

This thesis concerns the investigation of different implementation strategies for functional languages on a novel platform, the Rekursiv, that provides hardware support for an object- oriented language (Lingo) which was used as the implementation language for this work.<br/><br/>The objective is to evaluate the performance of two different implementation techniques - fixed combinators and supercombinators - for functional languages by interpreting the functional program in two object-oriented styles with different representations for combinators on two different environments:<br/><br/>• A purpose-built object-oriented machine; Harland’s Rekursiv.<br/>• A RISC machine; an IBM RS6000 running a Smalltalk-80 interpreter.<br/><br/>Interpreters are implemented in Smalltalk (for the RS6000) and Lingo (for the Rekursiv) for an applicative language Alcal (a A-calculus language). The interpreters are then used to generate a combinator graph using the fixed SKI set of combinators proposed by Turner, and also Hughes-style program specific supercombinators (Alifting) which are not limited to a fixed set of primitive combinators.<br/><br/>A number of experiments are conducted, based on the implementation of the lazy functional language, Alcal, on the RS6000 in Smalltalk-80 and on the Rekursiv architecture in Lingo. Lingo and Smalltalk allow the creation of objects that contain executable statements, so an object-oriented ‘active graph’ implementation is proposed.<br/><br/>Using these prototypes, the performance of the Rekursiv implementation against the RISC implementation has been evaluated. This includes the proportion of time spent on overhead activities and a statistical performance analysis of the benchmark results to analyse some of the claims made by the designer of the Rekursiv architecture.<br/><br/>Relative to a baseline benchmark written in C, the implementations on the Rekursiv are found to be several times faster than those on the RISC machine. An analysis of variance demonstrates a more subtle interplay between benchmark programs, implementation techniques, representation of combinators and hardware platform.

Name dc:type.qualificationname

PhD

Level dc:type.qualificationlevel

Doctoral Thesis

Year dc:date.issued

1993

Author dc:creator

• Khan, Mohammed

Language dc:language

eng

Identifier

oai:rke.abertay.ac.uk:studenttheses/0189925e-fc0d-4edf-bd11-6e27d0952efb

OAI identifier oai:identifier

oai:rke.abertay.ac.uk:studenttheses/0189925e-fc0d-4edf-bd11-6e27d0952efb