An Out-of-Core Branch and Bound Method for Solving the 0-1 Knapsack Problem on a GPU

被引:7
作者
Shen, Jingcheng [1 ]
Shigeoka, Kentaro [2 ]
Ino, Fumihiko [1 ]
Hagihara, Kenichi [1 ]
机构
[1] Osaka Univ, Grad Sch Informat Sci & Technol, 1-5 Yamadaoka, Suita, Osaka 5650871, Japan
[2] Hitachi High Technol Corp, Minato Ku, 14-24-1 Nishishinjuku, Tokyo 1058717, Japan
来源
ALGORITHMS AND ARCHITECTURES FOR PARALLEL PROCESSING, ICA3PP 2017 | 2017年 / 10393卷
基金
日本学术振兴会;
关键词
Out-of-core computation; Branch and bound; Knapsack; GPU; ALGORITHMS;
D O I
10.1007/978-3-319-65482-9_17
中图分类号
TP3 [计算技术、计算机技术];
学科分类号
0812 ;
摘要
In this paper, we propose an out-of-core branch and bound (B&B) method for solving the 0-1 knapsack problem on a graphics processing unit (GPU). Given a large problem that produces many subproblems, the proposed method dynamically swaps subproblems out to CPU memory. We adopt two strategies to realize this swapping-out procedure with minimum amount of CPU-GPU data transfer. The first strategy is a GPU-based stream compaction strategy that reduces the sparseness of arrays. The second strategy is a double buffering strategy that hides the data transfer overhead by overlapping data transfer with GPU-based B&B operations. Experimental results show that the proposed method can store 33.7 times more subproblems than the previous method, solving twice more instances on the GPU. As for the stream compaction strategy, an input-output separated scheme runs 13.1% faster than an input-output unified scheme.
引用
收藏
页码:254 / 267
页数:14
相关论文
共 19 条
[1]  
[Anonymous], 2017, CUDA TOOLKIT DOCUMEN
[2]  
Bell N., 2011, THRUST PRODUCTIVITY
[3]  
Boukedjar A., 2012, Proceedings of the 2012 20th Euromicro International Conference on Parallel, Distributed and Network-Based Processing (PDP 2012), P392, DOI 10.1109/PDP.2012.23
[4]   Solving knapsack problems on GPU [J].
Boyer, V. ;
El Baz, D. ;
Elkihel, M. .
COMPUTERS & OPERATIONS RESEARCH, 2012, 39 (01) :42-47
[5]  
Carneiro T., 2011, 2011 Proceedings of 23rd International Symposium on Computer Architecture and High Performance Computing (SBAC-PAD 2011), P41, DOI 10.1109/SBAC-PAD.2011.20
[6]   DISCRETE-VARIABLE EXTREMUM PROBLEMS [J].
DANTZIG, GB .
OPERATIONS RESEARCH, 1957, 5 (02) :266-277
[7]  
Eckstein J., 2001, PROC INHERENTLY PARA, P219, DOI DOI 10.1016/S1570-579X(01)80014-8
[8]   Master–Worker: An Enabling Framework for Applications on the Computational Grid [J].
Jean-Pierre Goux ;
Sanjeev Kulkarni ;
Michael Yoder ;
Jeff Linderoth .
Cluster Computing, 2001, 4 (1) :63-70
[9]   Sequence Homology Search Using Fine Grained Cycle Sharing of Idle GPUs [J].
Ino, Fumihiko ;
Munekawa, Yuma ;
Hagihara, Kenichi .
IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS, 2012, 23 (04) :751-759
[10]   GPU Implementation of the Branch and Bound method for knapsack problems [J].
Lalami, Mohamed Esseghir ;
El-Baz, Didier .
2012 IEEE 26TH INTERNATIONAL PARALLEL AND DISTRIBUTED PROCESSING SYMPOSIUM WORKSHOPS & PHD FORUM (IPDPSW), 2012, :1769-1777