Abstract
Parallel logic programming systems are sophisticated examples of symbolic computing systems. They address problems such as dynamic memory allocation, scheduling irregular execution patterns, and managing different types of implicit parallelism. Most parallel logic programming systems have been developed for bus-based shared-memory architectures. The complexity of parallel logic programming systems and the large amount of data they process raises the question of whether logic programming systems can still obtain good performance on scalable architectures, such as distributed shared-memory systems. In this work we use execution-driven simulation to investigate the access patterns and caching behaviour exhibited by a parallel logic programming system, Andorra-I. We show that the system obtains reasonable performance, but that it does not scale well. By studying the behaviour of the major data structures in Andorra-I in detail, we conclude that this result is largely a consequence of the scheduling and work manipulation implementation used in the system. We also show that the Andorra-I's data structures exhibit widely-varying memory access patterns and caching behaviour, which not only depend on the number of processors, but also on the amount and type of parallelism available in the application program. Some of these data structures clearly favour invalidate-based cache coherence protocols, while others favour update-based protocols. Since most of Andorra-I's data structures are common to other parallel logic programming systems, we believe that these systems can greatly benefit from flexible coherence schemes where either the compiler can specify the protocol to be used for each data structure or the protocol can adapt to varying memory access patterns.

Abstract
In this paper we introduce MOMEMTO (MOre MEMory Than Others) a new set of kernel mechanisms that allow users to have full control of the distributed shared memory on a cluster of personal computers. In contrast to many existing software DSM systems, MOMEMTO supports efficiently and flexibly global shared-memory allowing applications to address larger memory space than that available in a single node. MOMEMTO has been implemented in the Linux 2.4 kernel and preliminary performance results show that MOMEMTO has low memory management and communication overheads and that it can indeed perform very well for large memory configurations.

Abstract
Practical Prolog programs usually contain extra-logical features like cuts, side-effects, and database manipulating predicates. In order to exploit implicit parallelism from real applications while preserving sequential Prolog semantics, a parallel logic programming system should necessarily support these features. In this paper we show how Prolog's extra-logical features can be supported in an and-or parallel logic programming system. We show that to support extra-logical features an and-or parallel logic programming system should recompute the solutions to independent goals instead of sharing them. We describe an abstraction called the composition tree for representing and-or parallel execution with recomputation. We introduce the notion of ''local-leftmostness'' in the composition tree and use it for deriving complete and efficient methods for supporting extra-logical predicates in and-or parallel logic programming systems based on the composition tree abstraction.

Abstract
We propose dynamic compilation for Prolog, in the style of Just-In-Time compilers. Our approach adapts to the actual characteristics of the target program by (i) compiling only the parts of the program that are executed frequently, and (ii) adapting to actual call patterns. This allows aggressive optimization of the parts of the program that are really executed, and better informed heuristics to drive these optimizations. Our compiler does need to support all features in the language, only what is deemed important to performance. Complex execution patterns, such as the ones caused by error handling, may be left to the interpreter. On the other hand, compilation is now part of the run-time, and thus incurs run-time overheads. We have implemented dynamic compilation for YAP system. Our initial results suggest that dynamic compilation achieves very substantial performance improvements over the original interpreter, and that it can approach and even out-perform state-of-the-art native code systems. We believe that we have shown that dynamic compilation is worthwhile and fits naturally with Prolog execution.

Abstract
A critical issue in the design of logic programming systems is their memory performance, both in terms of total memory usage and locality in memory accesses. BEAM, as most modern Prolog systems, requires both good emulator design and good memory performance for best performance. We report on a detailed study of the memory management techniques used on our sequential implementation of the EAM. We address questions like how effective are the techniques the BEAM uses to recover and reuse memory space, how garbage collection affects performance and how to classify and unify variables in a EAM environment. We also propose a finer variable allocation scheme to reduce memory overheads that is quite effective at reducing memory pressure, with only a small overhead.

Abstract