CUDA Accelerated LTL Model Checking - Revisited
Petr Bauch, Milan Češka · DROPS (Schloss Dagstuhl – Leibniz Center for Informatics) · 2011
Recently, the massively parallel architecture has been used to significantly accelerate many computation demanding tasks. For example, in [Baier, Kateon, The MIT Press, 2009; Barnat, Brim, Ceska, ICPADS 2009] we have shown how CUDA technology can be employed to accelerate the process of Linear Temporal Logic (LTL) Model Checking. In this paper we redesign the One-Way-Catch-Them-Young (OWCTY) algorithm [Cerna, Pelanek, SPIN'03] in order to devise a new CUDA accelerated OWCTY algorithm that will significantly outperform the original CUDA accelerated algorithm and will be resistant to slowdown caused by improper ordering of the input data representation.