Lazy abstraction with interpolants, also known as the Impact algorithm, is en vogue as a state-of-the-art software model-checking technique for sequential programs. However, a direct extension of the Impact algorithm to concurrent programs is bound to be inefficient as it has to explore all thread interleavings, which leads to control-state explosion. To this end, we present a new algorithm that combines a new, symbolic form of partial-order reduction with Impact. Our algorithm carries out the dependence analysis on-the-fly while constructing the abstraction and is thus able to deal precisely with dynamic dependencies arising from accesses to tables or pointers -- a setting where classical static partial-order reduction techniques struggle. We have implemented the algorithm in a prototype tool that analyses concurrent C program with POSIX threads and evaluated it on a number of benchmark programs. To our knowledge, this is the first application of an Impact-like algorithm to concurrent programs.
Proceedings of FMCAD 13, 2013. 8 pages.
PDF
© 2013 IEEE Computer Society Press.