Encoding floating-point numbers using the SMT theory in ESBMC: An empirical evaluation over the SV-COMP benchmarks

Gadelha, Mikhail Y.R., Cordeiro, Lucas C. and Nicole, Denis A. (2017) Encoding floating-point numbers using the SMT theory in ESBMC: An empirical evaluation over the SV-COMP benchmarks. Cavalheiro, Simone and Fiadeiro, Jose (eds.) In Formal Methods: Foundations and Applications: 20th Brazilian Symposium, SBMF 2017, Recife, Brazil, November 29 — December 1, 2017, Proceedings. vol. 10623, Springer Cham. pp. 91-106 . (doi:10.1007/978-3-319-70848-5_7).

Record type: Conference or Workshop Item (Paper)

Abstract

This paper describes the support for encoding C/C++ programs using the SMT theory of floating-point numbers in ESBMC: an SMT-based context-bounded model checker that provides bit-precise verification of C and C++ programs. In particular, we exploit the availability of two different SMT solvers (MathSAT and Z3) to discharge and check the verification conditions produced by our encoding using the benchmarks from the International Competition on Software Verification (SV-COMP). The experimental results show that our encoding based on MathSAT is able to outperform not only Z3, but also other existing approaches that participated in the most recent edition of SV-COMP.

Text

Encoding:Floating-point numbers using the SMT theory in ESBMC: - Accepted Manuscript

Available under License University of Southampton Accepted Manuscript Licence.

Download (767kB)