Application of General-Purpose Computing on Graphics Processing Units for Acceleration of Basic Linear Algebra Operations and Principal Components Analysis Method

被引:4
作者
Majchrowicz, Michal [1 ]
Kapusta, Pawel [1 ]
Was, Lukasz [2 ]
Wiak, Slawomir [2 ]
机构
[1] Lodz Univ Technol, Inst Appl Comp Sci, Lodz, Poland
[2] Lodz Univ Technol, Inst Mech & Informat Syst, Lodz, Poland
来源
MAN-MACHINE INTERACTIONS 3 | 2014年 / 242卷
关键词
PCA; GPU; linear algebra operations;
D O I
10.1007/978-3-319-02309-0_57
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Nowadays speed of performed calculations can have significant impact not only on industrial application of developed solutions but also on conducted research, it is not uncommon for scientist to perform computations that may take hours or even days to complete. In this paper two different kinds of computations were analyzed: principal components analysis method and basic linear algebra operations. Authors present in this paper two different methods performing calculations using graphic cards. First one, used for PCA algorithm, uses single device, whereas second one, used in Landweber algorithm, uses multiple devices for conducting calculations.
引用
收藏
页码:519 / 527
页数:9
相关论文
共 9 条
[1]   Parallel GPU Implementation of Iterative PCA Algorithms [J].
Andrecut, M. .
JOURNAL OF COMPUTATIONAL BIOLOGY, 2009, 16 (11) :1593-1599
[2]  
[Anonymous], NEURAL NETWORKS TERM
[3]  
[Anonymous], 2016, Programming massively parallel processors: a hands-on approach
[4]  
[Anonymous], 2000, Principles of multivariate analysis
[5]  
Jolliffe Ian T., 1986, PRINCIPAL COMPONENT
[6]   Real-time PCA calculation for spectral imaging (using SIMD and GP-GPU) [J].
Josth, Radovan ;
Antikainen, Jukka ;
Havel, Jiri ;
Herout, Adam ;
Zemcik, Pavel ;
Hauta-Kasari, Markku .
JOURNAL OF REAL-TIME IMAGE PROCESSING, 2012, 7 (02) :95-103
[7]  
Kapusta P., 2010, P 17 INT C INF TECHN, P1
[8]  
Lionheart WR B., 2001, PROC 2 WORLD C IND P, P4
[9]  
Yang W., 2003, MEASUREMENT SCI TECH, V14