Parallel processing of matrix multiplication in a CPU and GPU heterogeneous environment

被引:0
作者
Ohshima, Satoshi [1 ]
Kise, Kenji [1 ]
Katagiri, Takahiro [1 ]
Yuba, Toshitsugu [1 ]
机构
[1] Univ Electrocommun, Grad Sch Informat Syst, 1-5-1 Chofugaoka, Chofu, Tokyo 182, Japan
来源
HIGH PERFORMANCE COMPUTING FOR COMPUTATIONAL SCIENCE - VECPAR 2006 | 2007年 / 4395卷
关键词
D O I
暂无
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
GPUs for numerical computations are becoming an attractive alternative in research. In this paper, we propose a new parallel processing environment for matrix multiplications by using both CPUs and GPUs. The execution time of matrix multiplications can be decreased to 40.1% by our method, compared with using the fastest of either CPU only case or GPU only case. Our method performs well when matrix sizes are large.
引用
收藏
页码:305 / +
页数:3
相关论文
共 18 条
[1]  
Amada T., 2004, ACM WORKSH GEN PURP, V41, P42
[2]  
[Anonymous], 2004, GPU GEMS PROGRAMMING
[3]  
[Anonymous], P ACM SIGGRAPH EUROG
[4]   Practical experience in the numerical dangers of heterogeneous computing [J].
Blackford, LS ;
Cleary, A ;
Petitet, A ;
Whaley, RC ;
Demmel, J ;
Dhillon, I ;
Ren, H ;
Stanley, K ;
Dongarra, J ;
Hammarling, S .
ACM TRANSACTIONS ON MATHEMATICAL SOFTWARE, 1997, 23 (02) :133-147
[5]  
FATAHALIAN K, 2004, GRAPHICS HARDWARE 20
[6]  
*GPGPU ORG, GEN PURP COMP GPUS G
[7]  
Hall Jesse D., 2003, CACHE BANDWIDTH AWAR
[8]   EXPLOITING FAST MATRIX MULTIPLICATION WITHIN THE LEVEL 3-BLAS [J].
HIGHAM, NJ .
ACM TRANSACTIONS ON MATHEMATICAL SOFTWARE, 1990, 16 (04) :352-368
[9]  
Hillesland K, 2004, P GP2
[10]  
Jiang CH, 2005, PACT 2005: 14TH INTERNATIONAL CONFERENCE ON PARALLEL ARCHITECTURES AND COMPILATION TECHNIQUES, P185