Proceedings of the Annual Conference of JSAI
Online ISSN : 2758-7347
36th (2022)
Session ID : 4F3-OS-8b-02
Conference information

A Study on a Framework for Constructing Deterministic Parallel SAT Solvers
*Tsubasa FUKIAGEYuto OBITSUHidetomo NABESHIMAXiao-Nan LU
Author information
CONFERENCE PROCEEDINGS FREE ACCESS

Details
Abstract

The widespread use of multi-core CPUs has led to the development of parallel SAT solvers, but most of them are nondeterministic. In addition, existing deterministic parallel SAT solvers such as ManyGlucose requires high development cost. In this study, we propose a framework to easily develop a deterministic parallel SAT solver, and show that homogeneous parallel SAT solvers based on this framework achieve the comparable performance with specialized one. Next, we attempted to reduce the latency for synchronization in heterogenous parallel SAT solvers. Our strategies achieve the performance improvement but did not lead to significant reductions in latency in the current stage.

Content from these authors
© 2022 The Japanese Society for Artificial Intelligence
Previous article Next article
feedback
Top