Hybrid CEGAR: combining variable hiding and predicate abstraction
Proceedings of the 2007 IEEE/ACM international conference on Computer-aided design
DiVer: SAT-based model checking platform for verifying large scale systems
TACAS'05 Proceedings of the 11th international conference on Tools and Algorithms for the Construction and Analysis of Systems
SAT-Based verification methods and applications in hardware verification
SFM'06 Proceedings of the 6th international conference on Formal Methods for the Design of Computer, Communication, and Software Systems
GLA: gate-level abstraction revisited
Proceedings of the Conference on Design, Automation and Test in Europe
| Hi-index | 0.00 |
SAT proof analysis techniques have been used recently with BMC in order to generate proof-based abstractions that preserve the correctness of the property up to a certain depth. In this paper, we propose various techniques for handling BMC constraints that are targeted at reducing the size of the abstract model while preserving bounded correctness. We propose lazy constraints, where the idea is to delay propagation of certain implied values without modifying a standard SAT solver. We propose different methods for automatic use of lazy constraints in SAT-based BMC. Their use can be regarded as a heuristic, which typically leads to significant reduction in the size of the abstract model, frequently accompanied by an overall performance improvement. We also use SAT proof analysis to loosen the user-specified environmental constraints, in order to generate smaller abstract models for conservative verification. We briefly describe other SAT-based heuristics for reducing the size of the abstract model. We have implemented our techniques in a prototype verification platform, and demonstrate their effectiveness on several large industry designs.