An implementation of flow calculus for complexity analysis (tool paper)
Résumé
Abstract. We present a tool to automatically perform the data-size analysis of imperative programs written in C. This tool, called pymwp, is inspired by a classical work on complexity analysis [10], and allows to certify that the size of the values computed by a program will be bounded by a polynomial in the program’s inputs. Strategies to provide meaningful feedback on non-polynomial programs and to “tame” the non-determinism of the original analysis were implemented following recent progresses [3], but required particular care to accommodate the
growing complexity of the analysis.
The Python source code is intensively documented, and our numerous example files encompass the original examples as well as multiple test cases. A pip package should make it easy to install pymwp on any plat-form, but an on-line demo is also available for convenience.
Origine | Fichiers produits par l'(les) auteur(s) |
---|