Communication Dans Un Congrès Année : 2025

Partial-Order Reduction Is Hard

Résumé

The goal of partial-order methods is to accelerate the exploration of concurrent systems by examining only a representative subset of all possible runs. The stateful approach builds a transition system with representative runs, while the stateless method simply enumerates them. The stateless approach may be preferable if the transition system is tree-like; otherwise, the stateful method is more effective. In the last decade, optimality has been a guiding principle for developing stateless partial-order reduction algorithms, and without doubt contributed to big progress in the field. In this paper we ask if we can get a similar principle for the stateful approach. We show that in stateful exploration, a polynomially close to optimal partial-order algorithm cannot exist unless P=NP. The result holds even for acyclic programs with just await instructions. This lower bound result justifies systematic study of heuristics for stateful partial-order reduction. We propose a notion of IFS oracle as a useful abstraction. The oracle can be used to get a very simple optimal stateless algorithm, which can then be adapted to a non-optimal stateful algorithm. While in general the oracle problem is NP-hard, we show a simple case where it can be solved in linear time.

Fichier principal
Vignette du fichier
LIPIcs.CONCUR.2025.22.pdf (1.57 Mo) Télécharger le fichier
Origine Fichiers éditeurs autorisés sur une archive ouverte
Licence

Dates et versions

hal-05399232 , version 1 (04-12-2025)

Licence

Identifiants

Citer

Frédéric Herbreteau, Sarah Larroze-Jardiné, Igor Walukiewicz. Partial-Order Reduction Is Hard. 36th International Conference on Concurrency Theory (CONCUR 2025), Aug 2025, Aarhus, Denmark. pp.22:1--22:20, ⟨10.4230/LIPIcs.CONCUR.2025.22⟩. ⟨hal-05399232⟩
106 Consultations
52 Téléchargements

Altmetric

Partager

  • More