Memory bandwidth optimization of SpMV on GPGPUs

doi:10.1007/s11704-014-4127-1

	Memory bandwidth optimization of SpMV on GPGPUs
	Yan, Chenggang Clarence 1,3; Yu, Hui 1; Xu, Weizhi 2,4; Zhang, Yingping 5; Chen, Bochuan 5; Tian, Zhu 6; Wang, Yuxuan 6; Yin, Jian 6
	2015-06-01
发表期刊	FRONTIERS OF COMPUTER SCIENCE
ISSN	2095-2228
卷号	9 期号:3 页码:431-441
摘要	It is an important task to improve performance for sparse matrix vector multiplication (SpMV), and it is a difficult task because of its irregular memory access. General purpose GPU (GPGPU) provides high computing ability and substantial bandwidth that cannot be fully exploited by SpMV due to its irregularity. In this paper, we propose two novel methods to optimize the memory bandwidth for SpMV on GPGPU. First, a new storage format is proposed to exploit memory bandwidth of GPU architecture more efficiently. The new storage format can ensure that there are as many non-zeros as possible in the format which is suitable to exploit the memory bandwidth of the GPU. Second, we propose a cache blocking method to improve the performance of SpMV on GPU architecture. The sparse matrix is partitioned into sub-blocks that are stored in CSR format.With the blocking method, the corresponding part of vector x can be reused in the GPU cache, so the time to access the global memory for vector x is reduced heavily. Experiments are carried out on three GPU platforms, GeForce 9800 GX2, GeForce GTX 480, and Tesla K40. Experimental results show that both new methods can efficiently improve the utilization of GPU memory bandwidth and the performance of the GPU.
关键词	GPGPU performance tuning SpMV cache blocking memory bandwidth
DOI	10.1007/s11704-014-4127-1
收录类别	SCI
语种	英语
资助项目	China Major ST Project[2013ZX01033001-001-003] ; International S&T Cooperation Project of China[2012DFA11170] ; Tsinghua Indigenous Research Project[20111080997] ; National Natural Science Foundation of China[61472203] ; National Natural Science Foundation of China[61274131] ; NVIDIA Corporation
WOS研究方向	Computer Science
WOS类目	Computer Science, Information Systems ; Computer Science, Software Engineering ; Computer Science, Theory & Methods
WOS记录号	WOS:000354194300009
出版者	HIGHER EDUCATION PRESS
引用统计	被引频次：5[WOS] [WOS记录] [WOS相关记录]
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/9690
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Xu, Weizhi
作者单位	1.Chinese Acad Sci, Inst Comp Technol, Key Lab Intelligent Informat Proc, Beijing 100190, Peoples R China 2.Tsinghua Univ, Inst Microelect, Beijing 100084, Peoples R China 3.Tsinghua Univ, Automat Dept, Beijing 100084, Peoples R China 4.Chinese Acad Sci, Inst Comp Technol, State Key Lab Comp Architecture, Beijing 100190, Peoples R China 5.State Grid Informat & Commun Co Hunan EPC, Changsha 410007, Hunan, Peoples R China 6.Shandong Univ, Dept Comp, Weihai 250101, Peoples R China
推荐引用方式 GB/T 7714	Yan, Chenggang Clarence,Yu, Hui,Xu, Weizhi,et al. Memory bandwidth optimization of SpMV on GPGPUs[J]. FRONTIERS OF COMPUTER SCIENCE,2015,9(3):431-441.
APA	Yan, Chenggang Clarence.,Yu, Hui.,Xu, Weizhi.,Zhang, Yingping.,Chen, Bochuan.,...&Yin, Jian.(2015).Memory bandwidth optimization of SpMV on GPGPUs.FRONTIERS OF COMPUTER SCIENCE,9(3),431-441.
MLA	Yan, Chenggang Clarence,et al."Memory bandwidth optimization of SpMV on GPGPUs".FRONTIERS OF COMPUTER SCIENCE 9.3(2015):431-441.