Abstract multiple specialization and its application to program parallelization.

Program specialization optimizes programs for known valúes of the input. It is often the case that the set of possible input valúes is unknown, or this set is infinite. However, a form of specialization can still be performed in such cases by means of abstract interpretation, specialization then being with respect to abstract valúes (substitutions), rather than concrete ones. We study the múltiple specialization of logic programs based on abstract interpretation. This involves in principie, and based on information from global analysis, generating several versions of a program predicate for different uses of such predicate, optimizing these versions, and, finally, producing a new, "multiply specialized" program. While múltiple specialization has received theoretical attention, little previous evidence exists on its practicality. In this paper we report on the incorporation of múltiple specialization in a parallelizing compiler and quantify its effects. A novel approach to the design and implementation of the specialization system is proposed. The resulting implementation techniques result in identical specializations to those of the best previously proposed techniques but require little or no modification of some existing abstract interpreters. Our results show that, using the proposed techniques, the resulting "abstract múltiple specialization" is indeed a relevant technique in practice. In particular, in the parallelizing compiler application, a good number of run-time tests are eliminated and invariants extracted automatically from loops, resulting generally in lower overheads and in several cases in increased speedups.

Strict and non-strict independent and-parallelism in logic programs: Correctness, efficiency, and compile-time conditions.

This paper presents some fundamental properties of independent and-parallelism and extends its applicability by enlarging the class of goals eligible for parallel execution. A simple model of (independent) and-parallel execution is proposed and issues of correctness and efficiency discussed in the light of this model. Two conditions, "strict" and "non-strict" independence, are defined and then proved sufficient to ensure correctness and efñciency of parallel execution: if goals which meet these conditions are executed in parallel the solutions obtained are the same as those produced by standard sequential execution. Also, in absence of failure, the parallel proof procedure does not genérate any additional work (with respect to standard SLD-resolution) while the actual execution time is reduced. Finally, in case of failure of any of the goals no slow down will occur. For strict independence the results are shown to hold independently of whether the parallel goals execute in the same environment or in sepárate environments. In addition, a formal basis is given for the automatic compile-time generation of independent and-parallelism: compile-time conditions to efficiently check goal independence at run-time are proposed and proved sufficient. Also, rules are given for constructing simpler conditions if information regarding the binding context of the goals to be executed in parallel is available to the compiler.

Abstract interpretation with specialized definitions

The relationship between abstract interpretation and partial evaluation has received considerable attention and (partial) integrations have been proposed starting from both the partial evaluation and abstract interpretation perspectives. In this work we present what we argüe is the first generic algorithm for efñcient and precise integration of abstract interpretation and partial evaluation from an abstract interpretation perspective. Taking as starting point state-of-the-art algorithms for context-sensitive, polyvariant abstract interpretation and (abstract) partial evaluation of logic programs, we present an algorithm which combines the best of both worlds. Key ingredients include the accurate success propagation inherent to abstract interpretation and the powerful program transformations achievable by partial deduction. In our algorithm, the calis which appear in the analysis graph are not analyzed w.r.t. the original definition of the procedure but w.r.t. specialized definitions of these procedures. Such specialized definitions are obtained by applying both unfolding and abstract executability. Also, our framework is parametric w.r.t. different control strategies and abstract domains. Different combinations of these parameters correspond to existing algorithms for program analysis and specialization. Our approach efficiently computes strictly more precise results than those achievable by each of the individual techniques. The algorithm is one of the key components of CiaoPP, the analysis and specialization system of the Ciao compiler.

Experiments in context-sensitive analysis of modular programs

Several models for context-sensitive analysis of modular programs have been proposed, each with different characteristics and representing different trade-offs. The advantage of these context-sensitive analyses is that they provide information which is potentially more accurate than that provided by context-free analyses. Such information can then be applied to validating/debugging the program and/or to specializing the program in order to obtain important performance improvements. Some very preliminary experimental results have also been reported for some of these models which provided initial evidence on their potential. However, further experimentation, which is needed in order to understand the many issues left open and to show that the proposed modes scale and are usable in the context of large, real-life modular programs, was left as future work. The aim of this paper is two-fold. On one hand we provide an empirical comparison of the different models proposed in previous work, as well as experimental data on the different choices left open in those designs. On the other hand we explore the scalability of these models by using larger modular programs as benchmarks. The results have been obtained from a realistic implementation of the models, integrated in a production-quality compiler (CiaoPP/Ciao). Our experimental results shed light on the practical implications of the different design choices and of the models themselves. We also show that contextsensitive analysis of modular programs is indeed feasible in practice, and that in certain critical cases it provides better performance results than those achievable by analyzing the whole program at once, specially in terms of memory consumption and when reanalyzing after making changes to a program, as is often the case during program development.

Efficient local unfolding with ancestor stacks for full prolog

The integration of powerful partial evaluation methods into practical compilers for logic programs is still far from reality. This is related both to 1) efficiency issues and to 2) the complications of dealing with practical programs. Regarding efnciency, the most successful unfolding rules used nowadays are based on structural orders applied over (covering) ancestors, i.e., a subsequence of the atoms selected during a derivation. Unfortunately, maintaining the structure of the ancestor relation during unfolding introduces significant overhead. We propose an efficient, practical local unfolding rule based on the notion of covering ancestors which can be used in combination with any structural order and allows a stack-based implementation without losing any opportunities for specialization. Regarding the second issue, we propose assertion-based techniques which allow our approach to deal with real programs that include (Prolog) built-ins and external predicates in a very extensible manner. Finally, we report on our implementation of these techniques in a practical partial evaluator, embedded in a state of the art compiler which uses global analysis extensively (the Ciao compiler and, specifically, its preprocessor CiaoPP). The performance analysis of the resulting system shows that our techniques, in addition to dealing with practical programs, are also significantly more efficient in time and somewhat more efficient in memory than traditional tree-based implementations.

Improving the compilation of prolog to C using moded types and determinism information

We describe the current status of and provide performance results for a prototype compiler of Prolog to C, ciaocc. ciaocc is novel in that it is designed to accept different kinds of high-level information, typically obtained via an automatic analysis of the initial Prolog program and expressed in a standardized language of assertions. This information is used to optimize the resulting C code, which is then processed by an off-the-shelf C compiler. The basic translation process essentially mimics the unfolding of a bytecode emulator with respect to the particular bytecode corresponding to the Prolog program. This is facilitated by a flexible design of the instructions and their lower-level components. This approach allows reusing a sizable amount of the machinery of the bytecode emulator: predicates already written in C, data definitions, memory management routines and áreas, etc., as well as mixing emulated bytecode with native code in a relatively straightforward way. We report on the performance of programs compiled by the current versión of the system, both with and without analysis information.

Efficient negation using abstract interpretation

While negation has been a very active área of research in logic programming, comparatively few papers have been devoted to implementation issues. Furthermore, the negation-related capabilities of current Prolog systems are limited. We recently presented a novel method for incorporating negation in a Prolog compiler which takes a number of existing methods (some modified and improved by us) and uses them in a combined fashion. The method makes use of information provided by a global analysis of the source code. Our previous work focused on the systematic description of the techniques and the reasoning about correctness and completeness of the method, but provided no experimental evidence to evalúate the proposal. In this paper, we report on an implementation, using the Ciao Prolog system preprocessor, and provide experimental data which indicates that the method is not only feasible but also quite promising from the efficiency point of view. In addition, the tests have provided new insight as to how to improve the proposal further. Abstract interpretation techniques are shown to offer important improvements in this application.

A documentation generator for (C)LP systems

We describe lpdoc, a tool which generates documentation manuals automatically from one or more logic program source files, written in Ciao, ISO-Prolog, and other (C)LP languages. It is particularly useful for documenting library modules, for which it automatically generates a rich description of the module interface. However, it can also be used quite successfully to document full applications. A fundamental advantage of using lpdoc is that it helps maintaining a true correspondence between the program and its documentation, and also identifying precisely to what versión of the program a given printed manual corresponds. The quality of the documentation generated can be greatly enhanced by including within the program text assertions (declarations with types, modes, etc. ...) for the predicates in the program, and machine-readable comments. One of the main novelties of lpdoc is that these assertions and comments are written using the Ciao system asseriion language, which is also the language of communication between the compiler and the user and between the components of the compiler. This allows a significant synergy among specification, debugging, documentation, optimization, etc. A simple compatibility library allows conventional (C)LP systems to ignore these assertions and comments and treat normally programs documented in this way. The documentation can be generated interactively from emacs or from the command line, in many formats including texinfo, dvi, ps, pdf, info, ascii, html/css, Unix nroff/man, Windows help, etc., and can include bibliographic citations and images, lpdoc can also genérate "man" pages (Unix man page format), nicely formatted plain ASCII "readme" files, installation scripts useful when the manuals are included in software distributions, brief descriptions in html/css or info formats suitable for inclusión in on-line Índices of manuals, and even complete WWW and info sites containing on-line catalogs of documents and software distributions. The lpdoc manual, all other Ciao system manuals, and parts of this paper are generated by lpdoc.

Abstract specialization and its application to program parallelization

Program specialization optimizes programs for known valúes of the input. It is often the case that the set of possible input valúes is unknown, or this set is infinite. However, a form of specialization can still be performed in such cases by means of abstract interpretation, specialization then being with respect to abstract valúes (substitutions), rather than concrete ones. This paper reports on the application of abstract múltiple specialization to automatic program parallelization in the &-Prolog compiler. Abstract executability, the main concept underlying abstract specialization, is formalized, the design of the specialization system presented, and a non-trivial example of specialization in automatic parallelization is given.

Towards independent and-parallelism in CLP

In this paper we propose a complete scheme for automatic exploitation of independent and-parallelism in CLP programs. We first discuss the new problems involved because of the different properties of the independence notions applicable to CLP. We then show how independence can be derived from a number of standard analysis domains for CLP. Finally, we perform a preliminary evaluation of the efficiency, accuracy, and effectiveness of the approach by implementing a parallehzing compiler for CLP based on the proposed ideas and applying it on a number of CLP benchmarks.

Implementation of multiple specialization in logic programs

We study the múltiple specialization of logic programs based on abstract interpretation. This involves in general generating several versions of a program predícate for different uses of such predícate, making use of information obtained from global analysis performed by an abstract interpreter, and finally producing a new, "multiply specialized" program. While the topic of múltiple specialization of logic programs has received considerable theoretical attention, it has never been actually incorporated in a compiler and its effects quantified. We perform such a study in the context of a parallelizing compiler and show that it is indeed a relevant technique in practice. Also, we propose an implementation technique which has the same power as the strongest of the previously proposed techniques but requires little or no modification of an existing abstract interpreter.

Effectiveness of global analysis in strict independence-based automatic program parallelization

This paper presents a study of the effectiveness of global analysis in the parallelization of logic programs using strict independence. A number of well-known approximation domains are selected and tlieir usefulness for the application in hand is explained. Also, methods for using the information provided by such domains to improve parallelization are proposed. Local and global analyses are built using these domains and such analyses are embedded in a complete parallelizing compiler. Then, the performance of the domains (and the system in general) is assessed for this application through a number of experiments. We argüe that the results offer significant insight into the characteristics of these domains, the demands of the application, and the tradeoffs involved.

A comparative study of methods for automatic compile-time parallelization of logic programs

This paper presents a study of the effectiveness of three different algorithms for the parallelization of logic programs based on compile-time detection of independence among goals. The algorithms are embedded in a complete parallelizing compiler, which incorporates different abstract interpretation-based program analyses. The complete system shows the task of automatic program parallelization to be practical. The trade-offs involved in using each of the algorithms in this task are studied experimentally, weaknesses of these identified, and possible improvements discussed.

An automatic translation scheme from prolog to the andorra kernel language

The Andorra family of languages (which includes the Andorra Kernel Language -AKL) is aimed, in principie, at simultaneously supporting the programming styles of Prolog and committed choice languages. On the other hand, AKL requires a somewhat detailed specification of control by the user. This could be avoided by programming in Prolog to run on AKL. However, Prolog programs cannot be executed directly on AKL. This is due to a number of factors, from more or less trivial syntactic differences to more involved issues such as the treatment of cut and making the exploitation of certain types of parallelism possible. This paper provides basic guidelines for constructing an automatic compiler of Prolog programs into AKL, which can bridge those differences. In addition to supporting Prolog, our style of translation achieves independent and-parallel execution where possible, which is relevant since this type of parallel execution preserves, through the translation, the user-perceived "complexity" of the original Prolog program.

A technique for recursive invariance detection and selective program specialization

This paper presents a technique for achieving a class of optimizations related to the reduction of checks within cycles. The technique uses both Program Transformation and Abstract Interpretation. After a ñrst pass of an abstract interpreter which detects simple invariants, program transformation is used to build a hypothetical situation that simpliñes some predicates that should be executed within the cycle. This transformation implements the heuristic hypothesis that once conditional tests hold they may continué doing so recursively. Specialized versions of predicates are generated to detect and exploit those cases in which the invariance may hold. Abstract interpretation is then used again to verify the truth of such hypotheses and conñrm the proposed simpliñcation. This allows optimizations that go beyond those possible with only one pass of the abstract interpreter over the original program, as is normally the case. It also allows selective program specialization using a standard abstract interpreter not speciñcally designed for this purpose, thus simplifying the design of this already complex module of the compiler. In the paper, a class of programs amenable to such optimization is presented, along with some examples and an evaluation of the proposed techniques in some application áreas such as floundering detection and reducing run-time tests in automatic logic program parallelization. The analysis of the examples presented has been performed automatically by an implementation of the technique using existing abstract interpretation and program transformation tools.