Hash consing

Abstract: Efficient storage of types within a compiler is necessary to avoid large blowups in space during compilation. Recursive types in particular are important to consider, as naive representations of recursive types may be arbitrarily larger than necessary through unfolding. Hash-consing has been used to efficiently store non-recursive types. Deterministic finite automata techniques have been used to efficiently perform various operations on recursive types. We present a new system for storing recursive types combining hash-consing and deterministic finite automata techniques. The space requirements are linear in the number of distinct types. Both update and lookup operations take polynomial time and linear space and type equality can be checked in constant time once both types are in the system.

Abstract: We report on three different approaches to use hash-consing in programs certified with the Coq system, using binary decision diagrams (BDD) as running example. The use cases include execution inside Coq, or execution of the extracted OCaml code. There are different trade-offs between faithful use of pristine extracted code, and code that is fine-tuned to make use of OCaml programming constructs not available in Coq. We discuss the possible consequences in terms of performances and guarantees.

Abstract: Points-to analysis is a fundamental static analysis, on which many other analyses and optimisations are built. The goal of points-to analysis is to statically approximate the set of abstract objects that a pointer can point to at runtime. Due to the nature of static analysis, points-to analysis introduces much redundancy which can result in duplicate points-to sets and duplicate set union operations, particularly when analysing large programs precisely. To improve performance, there has been extensive effort in mitigating duplication at the algorithmic level through, for example, cycle elimination and variable substitution.

Abstract: Current tabling systems suffer from an increase in space complexity, time complexity or both when dealing with sequences due to the use of data structures for tabled subgoals and answers and the need to copy terms into and from the table area. This symptom can be seen in not only B-Prolog, which uses hash tables, but also systems that use tries such as XSB and YAP. In this paper, we apply hash-consing to tabling structured data in B-Prolog. While hash-consing can reduce the space consumption when sharing is effective, it does not change the time complexity. We enhance hash-consing with two techniques, called input sharing and hash code memoization, for reducing the time complexity by avoiding computing hash codes for certain terms. The improved system is able to eliminate the extra linear factor in the old system for processing sequences, thus significantly enhancing the scalability of applications such as language parsing and bio-sequence analysis applications. We confirm this improvement with experimental results.