cTI: a constraint-based termination inference tool for ISO-Prolog

Abstract

We present cTI, the first system for universal left-termination inference of logic programs. Termination inference generalizes termination analysis and checking. Traditionally, a termination analyzer tries to prove that a given class of queries terminates. This class must be provided to the system, for instance by means of user annotations. Moreover, the analysis must be redone every time the class of queries of interest is updated. Termination inference, in contrast, requires neither user annotations nor recomputation. In this approach, terminating classes for all predicates are inferred at once. We describe the architecture of cTI and report an extensive experimental evaluation of the system covering many classical examples from the logic programming termination literature and several Prolog programs of respectable size and complexity.

The Dependency Triple Framework for Termination of Logic Programs

“…12 In addition, one could easily adapt more techniques from 11 Using different reduction pairs for different DT problems resulting from one and the same LP is for instance necessary for programs like the Ackermann function, cf. [15].…”

Nontermination inference of logic programs

“…Looping modes are computed starting from the leaves of the call graph then moving up to its roots. The cTI termination inference tool 2 is detailed in [26,24]. Here is the configuration we used for this experiment: Intel 686, 2.4GHz, 512Mb, Linux 2.4, SICStus Prolog 3.10.1, 24.8 MLips.…”

Determinacy Inference for Logic Programs

“…The table shows that the analysis time is dominated by the analyses that feed the client analysis -the backward analysis. There is no reason why the argument size analysis cannot be improved by replacing a constraint based implementation [2] with one based on a polyhedral library [16]. Moreover, the depth-k analysis would benefit from a more intelligent iteration strategy.…”