Title :
Portfolio-based Parallel SAT solving
Author_Institution :
CRIL, Univ. Lille Nord de France, Lille, France
fDate :
June 28 2010-July 2 2010
Abstract :
In this talk we present several contributions to Parallel SAT solving. First, we overview some of the recent work on parallel SAT, then we introduce the portfolio-based parallel framework and discuss its main pros. and cons. Secondly, we address the clause sharing issue, and we describe a novel approach for dynamically controlling the overall exchange between the computing units. Finally, we compare the two main approaches for parallel SAT: the classical divide and conquer and portfolio based parallel models.
Keywords :
Computational modeling; Joints; Learning; Mathematical model; Portfolios;
Conference_Titel :
High Performance Computing and Simulation (HPCS), 2010 International Conference on
Conference_Location :
Caen, France
Print_ISBN :
978-1-4244-6827-0
DOI :
10.1109/HPCS.2010.5547119