Description and Optimization of Abstract Machines in a Dialect of Prolog

Abstract

In order to achieve competitive performance, abstract machines for Prolog and related languages end up being large and intricate, and incorporate sophisticated optimizations, both at the design and at the implementation levels. At the same time, efficiency considerations make it necessary to use low-level languages in their implementation. This makes them laborious to code, optimize, and, especially, maintain and extend. Writing the abstract machine (and ancillary code) in a higher-level language can help tame this inherent complexity. We show how the semantics of most basic components of an efficient virtual machine for Prolog can be described using (a variant of) Prolog. These descriptions are then compiled to C and assembled to build a complete bytecode emulator. Thanks to the high level of the language used and its closeness to Prolog, the abstract machine description can be manipulated using standard Prolog compilation and optimization techniques with relative ease. We also show how, by applying program transformations selectively, we obtain abstract machine implementations whose performance can match and even exceed that of state-of-the-art, highly-tuned, hand-crafted emulators.Comment: 56 pages, 46 figures, 5 tables, To appear in Theory and Practice of Logic Programming (TPLP

Horn clauses as an intermediate representation for program analysis and transformation

“…Albert et al (2012) use a similar approach to cost analysis. Morales et al (2015) explore the use of a logic programming language for the implementation of efficient abstract machines and runtime systems. To this end they use a Prolog variant with certain imperative features (mutable variables) that enables translation into efficient C-style code while still allowing for high-level program transformations, such as partial evaluation of instruction definitions.…”

Fifty Years of Prolog and Beyond

“…Also, the progressively richer information inferred by the PLAI analyzers was applied to enhancing program development, leading to the Ciao assertion language and pre-processor, CiaoPP (Hermenegildo et al, 1999;Puebla et al, 2000a;Hermenegildo et al, 2005), which allowed optionally specifying and checking many properties such as types, modes, determinacy, non-failure, or cost, as well as autodocumentation. A native, optimizing compiler was also developed, and the abstract machine was rewritten in a restricted dialect of Prolog, ImProlog (Morales et al, 2005(Morales et al, , 2016. 1987) SB-Prolog was a Prolog system that, according to the CMU Artificial Intelligence Repository (1995) became available in 1987, and had been started as an exercise to understand the WAM.…”

“…Program analysis and transformation tools, partial evaluators, and automatic parallelizers can be used for Prolog programs, and can be particularly effective for • the perception that it is an "old" language purely declarative parts. For partial evaluation, however, the techniques have still not yet been integrated into Prolog compilers (the discussion by Leuschel and Bruynooghe (2002, Section 7) is mostly still valid today), with the exception of Ciao Prolog's pre-processor (Hermenegildo et al, 2005) and abstract machine generator (Morales et al, 2005(Morales et al, , 2016. However, they have found their way into just-in-time compilers for other languages (Bolz et al, 2011).…”