Distributed implementations of programming languages with implicit parallelism hold out the prospect that the parallel programs are immediately scalable. This paper presents some of the results of our part of Esprit 415, in which we considered the implementation of lazy functional programming languages on distributed architectures. A compiler and abstract machine were designed to achieve this goal. The abstract parallel machine was formally specified, using Miranda 1 . Each instruction of the abstract machine was then implemented as a macro in the Transputer Assembler. Although macro expansion of the code results in non-optimal code generation, use of the Miranda specification makes it possible to validate the compiler before the Transputer code is generated. The hardware currently available consists of five T800--25's, each board having 16M bytes of memory. Benchmark timings using this hardware are given. In spite of the straight forward code-generation, the resulting system compar...

syntax, and syntactic and semantic domains of a flow graph Figure 9. Semantic equations Def and Exp of a flow graph The first argument to the functions Def and Exp specifies a set of nodes that represent a flow graph, from which the element(s) of current interest are selected by pattern matching.

Lazy evaluation of functional programs incurs time and memory overheads, and restricts parallelism compared with programs that are evaluated strictly. A number of analysis techniques, such as abstract interpretation and projection analysis, have been developed to find out information that can alleviate these overheads. This paper formalises an evaluation model, the evaluation transformer model of reduction, which can use information from these analysis techniques, and proves that the resulting reduction strategies produce the same answers as those obtained using lazy evaluation.

The evaluation transformer model of reduction generalises lazy evaluation in two ways: it can start the evaluation of expressions before their first use, and it can evaluate expressions further than weak head normal form. Moreover, the amount of evaluation required of an argument to a function may depend on the amount of evaluation required of the function application. It is a suitable candidate model for implementing lazy functional languages on parallel machines. In this paper we explore the implementation of lazy functional languages on parallel machines, both shared and distributed memory architectures, using the evaluation transformer model of reduction. We will see that the same code can be produced for both styles of architecture, and the definition of the instruction set is virtually the same for each style. The essential difference is that a distributed memory architecture has one extra node type for non-local pointers, and instructions which involve the value of such nodes need their definitions extended to cover this new type of node. To make our presentation accessible, we base our description on a variant of the well-knon G-machine, a machine for executing lazy functional programs.

The research presented in this thesis is about the design and implementation of Naira, a parallel, parallelising compiler for a rich, purely functional programming language. The source language of the compiler is a subset of Haskell 1.2. The front end of Naira is written entirely in the Haskell subset being compiled. Naira has been successfully parallelised and it is the largest successfully parallelised Haskell program having achieved good absolute speedups on a network of SUN workstations. Having the same basic structure as other production compilers of functional languages, Naira's parallelisation technology should carry forward to other functional language compilers. The back end of Naira is written in C and generates parallel code in the C language which is envisioned to be run on distributed-memory machines. The code generator is based on a novel compilation scheme specified using a restricted form of Milner's ß-calculus which achieves asynchronous communication. We present the f...

. We show that compiler optimisations based on strictness analysis can be expressed formally in the functional framework using continuations. This formal presentation has two benefits: it allows us to give a rigorous correctness proof of the optimised compiler; and it exposes the various optimisations made possible by a strictness analysis. 1 Introduction Realistic compilers for imperative or functional languages include a number of optimisations based on non-trivial global analyses. Proving the correctness of such optimising compilers can be done in three steps: 1. proving the correctness of the original (unoptimised) compiler; 2. proving the correctness of the analysis; and 3. proving the correctness of the modifications of the simple-minded compiler to exploit the results of the analysis. A substantial amount of work has been devoted to steps (1) and (2) but there have been surprisingly few attempts at tackling step (3). In this paper we show how to carry out this third step in the...

Various nonstandard interpretations of functional programs have been proposed in which the basic nonstandard values are projections. We show that every stable function is completely determined by an appropriate abstract value in the backward analysis, and that every continuous function is completely determined by an appropriate value in the forward analysis. 1 Introduction Strictness analysis by abstract (or non-standard) interpretation is an important part of several compilers for lazy functional languages [1, 2], and a wide variety of strictness analysis techniques have been proposed. The term forward is used to describe abstract interpretations in which the goal is to discover information about an entire expression given information at its leaves, while backward describes interpretations in which information flows in the other direction: the goal is to determine information at the leaves of an expression, given information about the entire expression. The development of backward ...

Interpretation versus Demand Propagation Wadler [1987] uses abstract interpretation over a four-point domain for reasoning about strictness on lists. The four points correspond to undefined list (represented by value 0), infinite lists and lists with some tail undefined (value 1), lists with at least one head undefined (value 2), and all lists (value 3). Burn's work [Burn 1987] on evaluation transformers also uses abstract interpretation on the above-mentioned domain for strictness analysis. He defines four evaluators that correspond to the four points mentioned above in the sense that the ith evaluator will fail to produce an answer when given a list with the abstract value i \Gamma 1.

A system of annotated types is proposed as a means of describing and inferring static information, such as strictness and constancy, about functional programs. An abstract semantics is given in terms of projections. A close connection between annotated type assignment and projection analysis is demonstrated.

. Great hopes in the exploitation of the (implicit) parallelism inherent in functional programs have driven a number of projects. General frustration resulted wherever implementations on distributed memory machines were attempted. The grain size of potentially parallel tasks is too small to amortize the enormous costs of the necessary communication. The management of the parallelism is expensive. Decision making in the network operating system suffers from the fact, that most parameters such as the number of potentially parallel tasks and their execution times are dynamic. This article tries to summarise the state of the art in the parallel implementation of functional languages, to give reasons why attempts have failed to show performance, and argues why this will probably remain so. Similar arguments will hold for other declarative languages and mechanisms with implicit fine grain parallelism such as logic and constraint languages. 1 Introduction Parallel architectures are a big pro...