The CDG, UDG, and MEL methods for automatic compile-time parallelization of logic programs for independent and-parallelism

There has been significant interest in parallel execution models for logic programs which exploit Independent And-Parallelism (IAP). In these models, it is necessary to determine which goals are independent and therefore eligible for parallel execution and which goals have to wait for which others during execution. Although this can be done at run-time, it can imply a very heavy overhead. In this paper, we present three algorithms for automatic compiletime parallelization of logic programs using IAP. This is done by converting a clause into a graph-based computational form and then transforming this graph into linear expressions based on &-Prolog, a language for IAP. We also present an algorithm which, given a clause, determines if there is any loss of parallelism due to linearization, for the case in which only unconditional parallelism is desired. Finally, the performance of these annotation algorithms is discussed for some benchmark programs.

Non-strict independent and-parallelism

This paper presents and develops a generalized concept of Non-Strict Independent And Parallelism (NSIAP). NSIAP extends the applicability of Independent And- Parallelism (IAP) by enlarging the class of goals which are eligible for parallel execution. At the same time it maintains IAP's ability to run non-deterministic goals in parallel and to preserve the computational complexity expected in the execution of the program by the programmer. First, a parallel execution framework is defined and some fundamental correctness results, in the sense of equivalence of solutions with the sequential model, are discussed for this framework. The issue of efficiency is then considered. Two new definitions of NSI are given for the cases of puré and impure goals respectively and efficiency results are provided for programs parallelized under these definitions which include treatment of the case of goal failure: not only is reduction of execution time guaranteed (modulo run-time overheads) in the absence of failure but it is also shown that in the worst case of failure no speed-down will occur. In addition to applying to NSI, these results carry over and complete previous results shown in the context of IAP which did not deal with the case of goal failure. Finally, some practical examples of the application of the NSIAP concept to the parallelization of a set of programs are presented and performance results, showing the advantage of using NSI, are given.

&-prolog and its performance: exploiting independent and-parallelism

An Independent And-Parallel Prolog model and implementation, &-Prolog, are described. The description includes a summary of the system's architecture, some details of its execution model (based on the RAP-WAM model), and most importantly, its performance on sequential workstations and shared memory multiprocessors as compared with state-of-the-art Prolog systems. Speedup curves are provided for a collection of benchmark programs which demónstrate significant speed advantages over state-of the art sequential systems.

On the correctness and efficiency of independent and-parallelism in logic programs

This paper presents and proves some fundamental results for independent and-parallelism (IAP). First, the paper treats the issues of correctness and efficiency: after defining strict and non-strict goal independence, it is proved that if strictly independent goals are executed in parallel the solutions obtained are the same as those produced by standard sequential execution. It is also shown that, in the absence of failure, the parallel proof procedure doesn't genérate any additional work (with respect to standard SLDresolution) while the actual execution time is reduced. The same results hold even if non-strictly independent goals are executed in parallel, provided a trivial rewriting of such goals is performed. In addition, and most importantly, treats the issue of compile-time generation of IAP by proposing conditions, to be written at compile-time, to efficiently check strict and non-strict goal independence at run-time and proving the sufficiency of such conditions. It is also shown how simpler conditions can be constructed if some information regarding the binding context of the goals to be executed in parallel is available to the compiler trough either local or program-level analysis. These results therefore provide a formal basis for the automatic compile-time generation of IAP. As a corollary of such results, the paper also proves that negative goals are always non-strictly independent, and that goals which share a first occurrence of an existential variable are never independent.

Constructs and evaluation strategies for intelligent speculative parallelism - armageddon revisited

This report addresses speculative parallelism (the assignment of spare processing resources to tasks which are not known to be strictly required for the successful completion of a computation) at the user and application level. At this level, the execution of a program is seen as a (dynamic) tree —a graph, in general. A solution for a problem is a traversal of this graph from the initial state to a node known to be the answer. Speculative parallelism then represents the assignment of resources to múltiple branches of this graph even if they are not positively known to be on the path to a solution. In highly non-deterministic programs the branching factor can be very high and a naive assignment will very soon use up all the resources. This report presents work assignment strategies other than the usual depth-first and breadth-first. Instead, best-first strategies are used. Since their definition is application-dependent, the application language contains primitives that allow the user (or application programmer) to a) indícate when intelligent OR-parallelism should be used; b) provide the functions that define "best," and c) indícate when to use them. An abstract architecture enables those primitives to perform the search in a "speculative" way, using several processors, synchronizing them, killing the siblings of the path leading to the answer, etc. The user is freed from worrying about these interactions. Several search strategies are proposed and their implementation issues are addressed. "Armageddon," a global pruning method, is introduced, together with both a software and a hardware implementation for it. The concepts exposed are applicable to áreas of Artificial Intelligence such as extensive expert systems, planning, game playing, and in general to large search problems. The proposed strategies, although showing promise, have not been evaluated by simulation or experimentation.

Complete and efficient methods for supporting side effects in independent/restricted and-parallelism

It has been shown that it is possible to exploit Independent/Restricted And-parallelism in logic programs while retaining the conventional "don't know" semantics of such programs. In particular, it is possible to parallelize pure Prolog programs while maintaining the semantics of the language. However, when builtin side-effects (such as write or assert) appear in the program, if an identical observable behaviour to that of sequential Prolog implementations is to be preserved, such side-effects have to be properly sequenced. Previously proposed solutions to this problem are either incomplete (lacking, for example, backtracking semantics) or they force sequentialization of significant portions of the execution graph which could otherwise run in parallel. In this paper a series of side-effect synchronization methods are proposed which incur lower overhead and allow more parallelism than those previously proposed. Most importantly, and unlike previous proposals, they have well-defined backward execution behaviour and require only a small modification to a given (And-parallel) Prolog implementation.

Efficient management of backtracking in and-parallelism

A backtracking algorithm for AND-Parallelism and its implementation at the Abstract Machine level are presented: first, a class of AND-Parallelism models based on goal independence is defined, and a generalized version of Restricted AND-Parallelism (RAP) introduced as characteristic of this class. A simple and efficient backtracking algorithm for R A P is then discussed. An implementation scheme is presented for this algorithm which offers minimum overhead, while retaining the performance and storage economy of sequent ial implementations and taking advantage of goal independence to avoid unnecessary backtracking ("restricted intelligent backtracking"). Finally, the implementation of backtracking in sequential and AND-Parallcl systems is explained through a number of examples.

Parallelism and implementation of logic and constraint logic programming

CIAO is an advanced programming environment supporting Logic and Constraint programming. It offers a simple concurrent kernel on top of which declarative and non-declarative extensions are added via librarles. Librarles are available for supporting the ISOProlog standard, several constraint domains, functional and higher order programming, concurrent and distributed programming, internet programming, and others. The source language allows declaring properties of predicates via assertions, including types and modes. Such properties are checked at compile-time or at run-time. The compiler and system architecture are designed to natively support modular global analysis, with the two objectives of proving properties in assertions and performing program optimizations, including transparently exploiting parallelism in programs. The purpose of this paper is to report on recent progress made in the context of the CIAO system, with special emphasis on the capabilities of the compiler, the techniques used for supporting such capabilities, and the results in the áreas of program analysis and transformation already obtained with the system.

Towards high-level execution primitives for and-parallelism: preliminary results

Most implementations of parallel logic programming rely on complex low-level machinery which is arguably difflcult to implement and modify. We explore an alternative approach aimed at taming that complexity by raising core parts of the implementation to the source language level for the particular case of and-parallelism. Therefore, we handle a signiflcant portion of the parallel implementation mechanism at the Prolog level with the help of a comparatively small number of concurrency-related primitives which take care of lower-level tasks such as locking, thread management, stack set management, etc. The approach does not eliminate altogether modiflcations to the abstract machine, but it does greatly simplify them and it also facilitates experimenting with different alternatives. We show how this approach allows implementing both restricted and unrestricted (i.e., non fork-join) parallelism. Preliminary experiments show that the amount of performance sacriflced is reasonable, although granularity control is required in some cases. Also, we observe that the availability of unrestricted parallelism contributes to better observed speedups.

Towards a high-level implementation of flexible parallelism primitives for symbolic languages

An abstract is not available.

Estrategias Avanzadas de Tabulación y Paralelismo en Programas Lógicos = Advanced Evaluation Strategies for Tabling and Parallelism in Logic Programs

Dentro de los paradigmas de programación en el mundo de la informática tenemos la "Programación Lógica'', cuyo principal exponente es el lenguaje Prolog. Los programas Prolog se componen de un conjunto de predicados, cada uno de ellos definido en base a reglas que aportan un elevado nivel de abstracción y declaratividad al programador. Sin embargo, las formulación con reglas implica, frecuentemente, que un predicado se recompute varias veces para la misma consulta y además, Prolog utiliza un orden fijo para evaluar reglas y objetivos (evaluación SLD) que puede entrar en "bucles infinitos'' cuando ejecuta reglas recursivas declarativamente correctas. Estas limitaciones son atacadas de raiz por la tabulación, que se basa en "recordar'' en una tabla las llamadas realizadas y sus soluciones. Así, en caso de repetir una llamada tendríamos ya disponibles sus soluciones y evitamos la recomputación. También evita "bucles infinitos'' ya que las llamadas que los generan son suspendidas, quedando a la espera de que se computen soluciones para las mismas. La implementación de la tabulación no es sencilla. En particular, necesita de tres operaciones que no pueden ser ejecutadas en tiempo constante simultáneamente. Dichas operaciones son: suspensión de llamadas, relanzamiento de llamadas y {acceso a variables. La primera parte de la tesis compara tres implementaciones de tabulación sobre Ciao, cada una de las cuales penaliza una de estas operaciones. Por tanto, cada solución tiene sus ventajas y sus inconvenientes y se comportan mejor o peor dependiendo del programa ejecutado. La segunda parte de la tesis mejora la funcionalidad de la tabulación para combinarla con restricciones y también para evitar computaciones innecesarias. La programación con restricciones permite la resolución de ecuaciones como medio de programar, mecanismo altamente declarativo. Hemos desarrollado un framework para combinar la tabulación con las restricciones, priorizando objetivos como la flexibilidad, la eficiencia y la generalidad de nuestra solución, obteniendo una sinergia entre ambas técnicas que puede ser aplicada en numerosas aplicaciones. Por otra parte, un aspecto fundamental de la tabulación hace referencia al momento en que se retornan las soluciones de una llamada tabulada. Local evaluation devuelve soluciones cuando todas las soluciones de la llamada tabulada han sido computadas. Por contra, batched evaluation devuelve las soluciones una a una conforme van siendo computadas, por lo que se adapta mejor a problemas donde no nos interesa encontrar todas las soluciones. Sin embargo, su consumo de memoria es exponencialmente peor que el de local evaluation. La tesis presenta swapping evaluation, que devuelve soluciones tan pronto como son computadas pero con un consumo de memoria similar a la de local evaluation. Además, se implementan operadores de poda, once/1, para descartar la búsqueda de soluciones alternativas cuando encontramos la solución deseada. Por último, Prolog adopta con relativa facilidad soluciones para paralelismo gracias a su flexibilidad en el control de la ejecución y a que sus asignaciones son lógicas. La tercera parte de la tesis extiende el paralelismo conjuntivo de Ciao para trabajar con programas no deterministas, lo que presenta dos problemas principales: los objetivos atrapados y la recomputación de objetivos. Las soluciones clásicas para los objetivos atrapados rompían muchos invariantes de la ejecución Prolog, siendo soluciones difíciles de mantener y de extender, que la experiencia nos dice que han caído en desuso. Nosotros proponemos una solución modular (basada en la implementación de swapping evaluation), localizada y que no rompe los invariantes de la ejecución Prolog, pero que mantiene un alto rendimiento de la ejecución paralela. En referencia a la recomputación de objetivos paralelos en presencia de no determinismo hemos adaptado ténicas derivadas de la tabulación para memorizar computaciones de estos objetivos y evitar su recomputación.

A note on data-parallelism and (and-parallel) prolog

Abstract is not available.

A simulation study on parallel backtracking with solution memoing for independent and-parallelism

Goal-level Independent and-parallelism (IAP) is exploited by scheduling for simultaneous execution two or more goals which will not interfere with each other at run time. This can be done safely even if such goals can produce multiple answers. The most successful IAP implementations to date have used recomputation of answers and sequentially ordered backtracking. While in principle simplifying the implementation, recomputation can be very inefficient if the granularity of the parallel goals is large enough and they produce several answers, while sequentially ordered backtracking limits parallelism. And, despite the expected simplification, the implementation of the classic schemes has proved to involve complex engineering, with the consequent difficulty for system maintenance and expansion, and still frequently run into the well-known trapped goal and garbage slot problems. This work presents ideas about an alternative parallel backtracking model for IAP and a simulation studio. The model features parallel out-of-order backtracking and relies on answer memoization to reuse and combine answers. Whenever a parallel goal backtracks, its siblings also perform backtracking, but after storing the bindings generated by previous answers. The bindings are then reinstalled when combining answers. In order not to unnecessarily penalize forward execution, non-speculative and-parallel goals which have not been executed yet take precedence over sibling goals which could be backtracked over. Using a simulator, we show that this approach can bring significant performance advantages over classical approaches.

Automatic unrestricted independent and-parallelism in logic programs

We present new algorithms which perform automatic parallelization via source-to-source transformations. The objective is to exploit goal-level, unrestricted independent andparallelism. The proposed algorithms use as targets new parallel execution primitives which are simpler and more flexible than the well-known &/2 parallel operator, which makes it possible to generate better parallel expressions by exposing more potential parallelism among the literals of a clause than is possible with &/2. The main differences between the algorithms stem from whether the order of the solutions obtained is preserved or not, and on the use of determinacy information. We briefly describe the environment where the algorithms have been implemented and the runtime platform in which the parallelized programs are executed. We also report on an evaluation of an implementation of our approach. We compare the performance obtained to that of previous annotation algorithms and show that relevant improvements can be obtained.

Algoritmos de distribución de cargas de proceso en computación con membranas

Con el surgir de los problemas irresolubles de forma eficiente en tiempo polinomial en base al dato de entrada, surge la Computación Natural como alternativa a la computación clásica. En esta disciplina se trata de o bien utilizar la naturaleza como base de cómputo o bien, simular su comportamiento para obtener mejores soluciones a los problemas que los encontrados por la computación clásica. Dentro de la computación natural, y como una representación a nivel celular, surge la Computación con Membranas. La primera abstracción de las membranas que se encuentran en las células, da como resultado los P sistemas de transición. Estos sistemas, que podrían ser implementados en medios biológicos o electrónicos, son la base de estudio de esta Tesis. En primer lugar, se estudian las implementaciones que se han realizado, con el fin de centrarse en las implementaciones distribuidas, que son las que pueden aprovechar las características intrínsecas de paralelismo y no determinismo. Tras un correcto estudio del estado actual de las distintas etapas que engloban a la evolución del sistema, se concluye con que las distribuciones que buscan un equilibrio entre las dos etapas (aplicación y comunicación), son las que mejores resultados presentan. Para definir estas distribuciones, es necesario definir completamente el sistema, y cada una de las partes que influyen en su transición. Además de los trabajos de otros investigadores, y junto a ellos, se realizan variaciones a los proxies y arquitecturas de distribución, para tener completamente definidos el comportamiento dinámico de los P sistemas. A partir del conocimiento estático –configuración inicial– del P sistema, se pueden realizar distribuciones de membranas en los procesadores de un clúster para obtener buenos tiempos de evolución, con el fin de que la computación del P sistema sea realizada en el menor tiempo posible. Para realizar estas distribuciones, hay que tener presente las arquitecturas –o forma de conexión– de los procesadores del clúster. La existencia de 4 arquitecturas, hace que el proceso de distribución sea dependiente de la arquitectura a utilizar, y por tanto, aunque con significativas semejanzas, los algoritmos de distribución deben ser realizados también 4 veces. Aunque los propulsores de las arquitecturas han estudiado el tiempo óptimo de cada arquitectura, la inexistencia de distribuciones para estas arquitecturas ha llevado a que en esta Tesis se probaran las 4, hasta que sea posible determinar que en la práctica, ocurre lo mismo que en los estudios teóricos. Para realizar la distribución, no existe ningún algoritmo determinista que consiga una distribución que satisfaga las necesidades de la arquitectura para cualquier P sistema. Por ello, debido a la complejidad de dicho problema, se propone el uso de metaheurísticas de Computación Natural. En primer lugar, se propone utilizar Algoritmos Genéticos, ya que es posible realizar alguna distribución, y basada en la premisa de que con la evolución, los individuos mejoran, con la evolución de dichos algoritmos, las distribuciones también mejorarán obteniéndose tiempos cercanos al óptimo teórico. Para las arquitecturas que preservan la topología arbórea del P sistema, han sido necesarias realizar nuevas representaciones, y nuevos algoritmos de cruzamiento y mutación. A partir de un estudio más detallado de las membranas y las comunicaciones entre procesadores, se ha comprobado que los tiempos totales que se han utilizado para la distribución pueden ser mejorados e individualizados para cada membrana. Así, se han probado los mismos algoritmos, obteniendo otras distribuciones que mejoran los tiempos. De igual forma, se han planteado el uso de Optimización por Enjambres de Partículas y Evolución Gramatical con reescritura de gramáticas (variante de Evolución Gramatical que se presenta en esta Tesis), para resolver el mismo cometido, obteniendo otro tipo de distribuciones, y pudiendo realizar una comparativa de las arquitecturas. Por último, el uso de estimadores para el tiempo de aplicación y comunicación, y las variaciones en la topología de árbol de membranas que pueden producirse de forma no determinista con la evolución del P sistema, hace que se deba de monitorizar el mismo, y en caso necesario, realizar redistribuciones de membranas en procesadores, para seguir obteniendo tiempos de evolución razonables. Se explica, cómo, cuándo y dónde se deben realizar estas modificaciones y redistribuciones; y cómo es posible realizar este recálculo. Abstract Natural Computing is becoming a useful alternative to classical computational models since it its able to solve, in an efficient way, hard problems in polynomial time. This discipline is based on biological behaviour of living organisms, using nature as a basis of computation or simulating nature behaviour to obtain better solutions to problems solved by the classical computational models. Membrane Computing is a sub discipline of Natural Computing in which only the cellular representation and behaviour of nature is taken into account. Transition P Systems are the first abstract representation of membranes belonging to cells. These systems, which can be implemented in biological organisms or in electronic devices, are the main topic studied in this thesis. Implementations developed in this field so far have been studied, just to focus on distributed implementations. Such distributions are really important since they can exploit the intrinsic parallelism and non-determinism behaviour of living cells, only membranes in this case study. After a detailed survey of the current state of the art of membranes evolution and proposed algorithms, this work concludes that best results are obtained using an equal assignment of communication and rules application inside the Transition P System architecture. In order to define such optimal distribution, it is necessary to fully define the system, and each one of the elements that influence in its transition. Some changes have been made in the work of other authors: load distribution architectures, proxies definition, etc., in order to completely define the dynamic behaviour of the Transition P System. Starting from the static representation –initial configuration– of the Transition P System, distributions of membranes in several physical processors of a cluster is algorithmically done in order to get a better performance of evolution so that the computational complexity of the Transition P System is done in less time as possible. To build these distributions, the cluster architecture –or connection links– must be considered. The existence of 4 architectures, makes that the process of distribution depends on the chosen architecture, and therefore, although with significant similarities, the distribution algorithms must be implemented 4 times. Authors who proposed such architectures have studied the optimal time of each one. The non existence of membrane distributions for these architectures has led us to implement a dynamic distribution for the 4. Simulations performed in this work fix with the theoretical studies. There is not any deterministic algorithm that gets a distribution that meets the needs of the architecture for any Transition P System. Therefore, due to the complexity of the problem, the use of meta-heuristics of Natural Computing is proposed. First, Genetic Algorithm heuristic is proposed since it is possible to make a distribution based on the premise that along with evolution the individuals improve, and with the improvement of these individuals, also distributions enhance, obtaining complexity times close to theoretical optimum time. For architectures that preserve the tree topology of the Transition P System, it has been necessary to make new representations of individuals and new algorithms of crossover and mutation operations. From a more detailed study of the membranes and the communications among processors, it has been proof that the total time used for the distribution can be improved and individualized for each membrane. Thus, the same algorithms have been tested, obtaining other distributions that improve the complexity time. In the same way, using Particle Swarm Optimization and Grammatical Evolution by rewriting grammars (Grammatical Evolution variant presented in this thesis), to solve the same distribution task. New types of distributions have been obtained, and a comparison of such genetic and particle architectures has been done. Finally, the use of estimators for the time of rules application and communication, and variations in tree topology of membranes that can occur in a non-deterministic way with evolution of the Transition P System, has been done to monitor the system, and if necessary, perform a membrane redistribution on processors to obtain reasonable evolution time. How, when and where to make these changes and redistributions, and how it can perform this recalculation, is explained.