Truly stateless, optimal dynamic partial order reduction

Dynamic partial order reduction (DPOR) verifies concurrent programs by exploring all their interleavings up to some equivalence relation, such as the Mazurkiewicz trace equivalence. Doing so involves a complex trade-off between space and time. Existing DPOR algorithms are either exploration-optimal (i.e., explore exactly only interleaving per class) but may use exponential memory in size of program, or maintain polynomial consumption potentially exponentially many redundant interleavings. In this paper, we show that it is possible have best both worlds: one class with linear consumption. Our algorithm, TruSt, formalized Coq, applicable not sequential consistency, also any weak model satisfies few basic assumptions, including TSO, PSO, RC11. addition, TruSt embarrassingly parallelizable: its different exploration options no shared state, can therefore be explored completely parallel. Consequently, outperforms state-of-the-art terms and/or