The state of the art for bit-precise reasoning in the context of Satisfiability Modulo Theories (SMT) is a SAT-based technique called bit-blasting where the input formula is first simplified and then translated to an equisatisfiable propositional formula. The main limitation of this technique is scalability, especially in the presence of large bit-widths and arithmetic operators. We introduce an alternative technique, which we call int-blasting, based on a translation to an extension of integer arithmetic rather than propositional logic. We present several translations, discuss their differences, and evaluate them on benchmarks that arise from the verification of rewrite rule candidates for bit-vector solving, as well as benchmarks from SMT-LIB. We also provide preliminary results on 35 benchmarks that arise from smart contract verification. The evaluation shows that this technique is particularly useful for benchmarks with large bit-widths and can solve benchmarks that the state of the art cannot.
|Title of host publication||Verification, Model Checking, and Abstract Interpretation - 23rd International Conference, VMCAI 2022, Proceedings|
|Editors||Bernd Finkbeiner, Thomas Wies|
|Publisher||Springer Science and Business Media Deutschland GmbH|
|Number of pages||23|
|State||Published - 2022|
|Event||23rd International Conference on Verification, Model Checking, and Abstract Interpretation, VMCAI 2022 - Philadelphia, United States|
Duration: 16 Jan 2022 → 18 Jan 2022
|Name||Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)|
|Conference||23rd International Conference on Verification, Model Checking, and Abstract Interpretation, VMCAI 2022|
|Period||16/01/22 → 18/01/22|
Bibliographical noteFunding Information:
This work was supported in part by DARPA (awards N66001-18-C-4012, FA8650-18-2-7854 and FA8650-18-2-7861), ONR (award N68335-17-C-0558), the Stanford Center for Blockchain Research, Certora Inc., and by an NSF Graduate Fellowship (to Makai Mann). A. Irfan—This author’s contributions were made while he was a postdoc at Stanford University.
© 2022, Springer Nature Switzerland AG.