We review the design of the concurrent logic language GHC, the basis of the kernel language for the Parallel Inference Machine being developed in the Japanese Fifth Generation Computer Systems project, and the design of the parallel language KL1, the actual kernel language being implemented and used. The key idea in the design of these languages is the separation of concurrency and parallelism. Clarication of concepts of this kind seems to play an important role in bridging the gap between parallel inference systems and knowledge information processing in a coherent manner. In particular, design of a new kernel language has always encouraged us to re-examine and reorganize various existing notions related to programming and to invent new ones.

Abstract not found

Automated problem solving is viewed typically as the allocation of computational resources to solve one or more problems passed to a reasoning system. In response to each problem received, effort is applied in real time to generate a solution and problem solving ends when a solution is rendered. We examine continual computation, reasoning policies that capture a broader conception of problem by considering the proactive allocation of computational resources to potential future challenges. We explore policies for allocating idle time for several settings and present applications that highlight opportunities for harnessing continual computation in real-world tasks. 2001 Elsevier Science B.V. All rights reserved. Keywords: Bounded rationality; Decision-theoretic control; Metareasoning; Deliberation; Compilation; Speculative execution; Value of computation 1.

Speculative execution refers to the execution of parts of a computation before the execution of the conditional operations that decide

Lazy programs are beautiful, but they are slow because they build many thunks. Simple measurements show that most of these thunks are unnecessary: they are in fact always evaluated, or are always cheap. In this paper we describe Optimistic Evaluation --- an evaluation strategy that exploits this observation. Optimistic Evaluation complements compile-time analyses with run-time experiments: it evaluates a thunk speculatively, but has an abortion mechanism to back out if it makes a bad choice. A run-time adaption mechanism records expressions found to be unsuitable for speculative evaluation, and arranges for them to be evaluated more lazily in the future.

Recent research in thread-level speculation (TLS) has proposed several mechanisms for optimistic execution of di#cultto -analyze serial codes in parallel. Though it has been shown that TLS helps to achieve higher levels of parallelism, evaluation of the unique performance potential of TLS, i.e., performance gain that be achieved only through speculation, has not received much attention. In this paper, we evaluate this aspect, by separating the speedup achievable via true TLP (thread-level parallelism) and TLS, for the SPEC CPU2000 benchmark. Further, we dissect the performance potential of each type of speculation --- control speculation, data dependence speculation and data value speculation. To the best of our knowledge, this is the first dissection study of its kind. Assuming an oracle TLS mechanism --- which corresponds to perfect speculation and zero threading overhead --- whereby the execution time of a candidate program region (for speculative execution) can be reduced to zero, our study shows that, at the loop-level, the upper bound on the arithmetic mean and geometric mean speedup achievable via TLS across SPEC CPU2000 is 39.16% (standard deviation = 31.23) and 18.18% respectively.

We examine the task of concurrently computing alternative solutions to a problem. We restrict our interest to the case where only one of the solutions is needed; in this case we need some rule for selecting between the solutions. We use ‘‘fastest first’’, where the first successful alternative is selected. For problems where the required execution time is unpredictable, such as database queries, this method can show substantial execution time performance increases. These increases are dependent on the mean execution time of the alternatives, the fastest execution time, and the overhead involved in concurrent computation. Among the problems with exploring multiple alternatives inparallel are side-effects and combinatorial explosion in the amount of state which must be preserved. These are solved by process management and an application of ‘‘copy-on-write’’ virtual memory management. The side effects resulting from interprocess communication are handled by a specialized message layer which interacts with process management. In order to test the utility of the design, we show how it can be applied to two application areas, distributed execution of recovery blocks and OR-parallelism in Prolog. 1.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xiii 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1 Motivation: The Problems of Parallel Programming . . . . . . . . . . . . . . . . . . . . . 2 1.2 A First Solution: Functional Programming. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.3 A Refined Solution: Speculative Evaluation. . . . . . . . . . . . . . . . . . . . . . . . . . . 10 1.4 A Final Solution: Effective Speculative Evaluation . . . . . . . . . . . . . . . . . . . . . 12 1.5 Objective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 1.6 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 2. Supercombinator Graph Reduction . . . . . . . . . . . . . . . . . . . . . . . . . ...

One way to execute a program faster on a parallel computer is to allocate some processors to speculative computation: work which may prove to be unnecessary. However, this technique, speculative parallelism, requires the programmer to have some control over the physical processors in the machine, and is therefore not usually possible in high-level concurrent languages. This paper investigates the exploitation of speculative parallelism in concurrent logic programming languages. We propose a simple language extension that makes the technique possible, and illustrate its use in two applications: (1) the familiar branch-and-bound search algorithm, and (2) a new dynamic buffer protocol for interprocess communication. The new language feature has been implemented by extending the JAM Parallel Parlog system. The results of our experiments confirm that speculative parallelism can result in greater parallel speedups than is possible with other (divide and conquer) forms of parallelism alone. K...

Abstract. Optimization problems where the evaluation step is computationally intensive are becoming increasingly common in both engineering design and model parameter estimation. We describe a tool, Nimrod/O, that expedites the solution of such problems by performing evaluations concurrently, utilizing a range of platforms from workstations to widely distributed parallel machines. Nimrod/O offers a range of optimization algorithms adapted to take advantage of parallel batches of evaluations. We describe a selection of case studies where Nimrod/O has been successfully applied, showing the parallelism achieved by this approach. 1