Experimenting with SAT Solvers in Vampire
Recently, a new reasoning framework, called AVATAR, integrating first-order theorem proving with SAT solving has been proposed. In this paper, we experimentally analyze the behavior of various SAT solvers within first-order proving. For doing so, we first integrate the Lingeling SAT solver within the first-order theorem prover Vampire and compare the behavior of such an integration with Vampire using a less efficient SAT solver. Interestingly, our experiments on first-order problems show that using the best SAT solvers within AVATAR does not always give best performance. There are some problems that could be solved only by using a less efficient SAT solver than Lingeling. However, the integration of Lingeling with Vampire turned out to be the best when it came to solving most of the hard problems.
Unable to display preview. Download preview PDF.
- 1.HPC Center - West University of Timisoara, http://hpc.uvt.ro/infrastructure/infragrid/
- 3.Biere, A.: Lingeling, plingeling and treengeling entering sat competition 2013. In: SAT Competition 2013, pp. 51–52 (2013)Google Scholar
- 9.Nieuwenhuis, R., Hillenbrand, T., Riazanov, A., Voronkov, A.: On the evaluation of indexing techniques for theorem proving. In: IJCAR, pp. 257–271 (2001)Google Scholar
- 10.Riazanov, A., Voronkov, A.: Splitting without backtracking. In: IJCAI, pp. 611–617 (2001)Google Scholar
- 12.Silva, J.P.M., Lynce, I., Malik, S.: Conflict-driven clause learning sat solvers. In: Handbook of Satisfiability, pp. 131–153 (2009)Google Scholar
- 16.Weidenbach, C.: Combining superposition, sorts and splitting. In: Handbook of Automated Reasoning, pp. 1965–2013 (2001)Google Scholar