CUD@SAT: SAT solving on GPUs

被引:27
作者
Dal Palu, Alessandro [1 ]
Dovier, Agostino [2 ]
Formisano, Andrea [3 ]
Pontelli, Enrico [4 ]
机构
[1] Univ Parma, Dept Math & Comp Sci, I-43100 Parma, Italy
[2] Univ Udine, Dept Math & Comp Sci, I-33100 Udine, Italy
[3] Univ Perugia, Dept Math & Comp Sci, I-06100 Perugia, Italy
[4] New Mexico State Univ, Dept Comp Sci, Las Cruces, NM 88003 USA
基金
美国国家科学基金会;
关键词
Davis-Putnam-Logemann-Loveland procedure; CNF-satisfiability; parallel SAT-solving; general purpose GPU computing;
D O I
10.1080/0952813X.2014.954274
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
The parallel computing power offered by graphic processing units (GPUs) has been recently exploited to support general purpose applications - by exploiting the availability of general API and the single-instruction multiple-thread-style parallelism present in several classes of problems (e.g. numerical simulations and matrix manipulations) - where relatively simple computations need to be applied to all items in large sets of data. This paper investigates the use of GPUs in parallelising a class of search problems, where the combinatorial nature leads to large parallel tasks and relatively less natural symmetries. Specifically, the investigation focuses on the well-known satisfiability testing (SAT) problem and on the use of the NVIDIA compute unified device architecture, one of the most popular platforms for GPU computing. The paper explores ways to identify strong sources of GPU-style parallelism from SAT solving. The paper describes experiments with different design choices and evaluates the results. The outcomes demonstrate the potential for this approach, leading to one order of magnitude of speedup using a simple NVIDIA platform.
引用
收藏
页码:293 / 316
页数:24
相关论文
共 47 条
[1]  
[Anonymous], 2008, NVIDIA Technical Report NVR-2008-004
[2]  
[Anonymous], 2012, NVIDIA CUDA C Programming Guide
[3]  
Apt K., 2009, Principles of constraint programming
[4]  
Baral C., 2010, Knowledge Representation, Reasoning and Declarative Problem Solving
[5]  
Bayardo R., 1997, 14 NAT C ART INT MEN
[6]  
Beckers Sander, 2011, IADIS International Conference of Applied Computing 2011. Proceedings, P435
[7]   Bounded model checking [J].
Biere, Armin .
Frontiers in Artificial Intelligence and Applications, 2009, 185 (01) :457-481
[8]   Parallel propositional satisfiability checking with distributed dynamic learning [J].
Blochinger, W ;
Sinz, C ;
Küchlin, W .
PARALLEL COMPUTING, 2003, 29 (07) :969-994
[9]   A fast parallel SAT-solver - Efficient workload balancing [J].
Bohm, M ;
Speckenmeyer, E .
ANNALS OF MATHEMATICS AND ARTIFICIAL INTELLIGENCE, 1996, 17 (3-4) :381-400
[10]   A GPU Implementation of Large Neighborhood Search for Solving Constraint Optimization Problems [J].
Campeotto, F. ;
Dovier, A. ;
Fioretto, F. ;
Pontelli, E. .
21ST EUROPEAN CONFERENCE ON ARTIFICIAL INTELLIGENCE (ECAI 2014), 2014, 263 :189-+