Performance testing and analysis of MAGMA library on hybrid architecture (CPU+GPU)

ISCAS OpenIR

	Performance testing and analysis of MAGMA library on hybrid architecture (CPU+GPU)
	Xiao, Xuan-Ji (1); Zhang, Yun-Quan (1); Li, Yu-Cheng (1); Yuan, Liang (1); Xiao, X.-J.(growj@126.com)
	2013
发表期刊	Ruan Jian Xue Bao/Journal of Software
ISSN	10009825
卷号	24 期号:SUPPL.2 页码:118-126
摘要	MAGMA is an open source high performance linear algebra package first developed for next-generation of heterogeneous/ hybrid architectures (CPUs+GPUs) with a dense linear algebra library similar to LAPACK in functionality, data storage, and interface. This paper presents performance testing and analysis of MAGMA. It first studies the matrix decomposition algorithm in MAGMA, then provides some useful suggestions of MAGMA usage and optimization through massive testing and source code analysis, and finally proposes a method for auto-tuning matrix decomposition block algorithms. In this test, the speedup is 1.09 for SGEQRF of square matrix and 1.8 for CGEQRF in terms of tall and skin matrix. © Institute of Software Chinese Academy of Sciences.; MAGMA is an open source high performance linear algebra package first developed for next-generation of heterogeneous/ hybrid architectures (CPUs+GPUs) with a dense linear algebra library similar to LAPACK in functionality, data storage, and interface. This paper presents performance testing and analysis of MAGMA. It first studies the matrix decomposition algorithm in MAGMA, then provides some useful suggestions of MAGMA usage and optimization through massive testing and source code analysis, and finally proposes a method for auto-tuning matrix decomposition block algorithms. In this test, the speedup is 1.09 for SGEQRF of square matrix and 1.8 for CGEQRF in terms of tall and skin matrix. © Institute of Software Chinese Academy of Sciences.
收录类别	EI
部门归属	(1) Laboratory of Parallel Computing, Institute of Software, The Chinese Academy of Sciences, Beijing 100190, China; (2) State Key Laboratory of Computer Science, Institute of Software, The Chinese Academy of Sciences, Beijing 100190, China; (3) University of Chinese Academy of Sciences, Beijing 100190, China
语种	中文
内容类型	期刊论文
URI标识	http://ir.iscas.ac.cn/handle/311060/17045
专题	中国科学院软件研究所
通讯作者	Xiao, X.-J.(growj@126.com)
推荐引用方式 GB/T 7714	Xiao, Xuan-Ji ,Zhang, Yun-Quan ,Li, Yu-Cheng ,et al. Performance testing and analysis of MAGMA library on hybrid architecture (CPU+GPU)[J]. Ruan Jian Xue Bao/Journal of Software,2013,24(SUPPL.2):118-126.
APA	Xiao, Xuan-Ji ,Zhang, Yun-Quan ,Li, Yu-Cheng ,Yuan, Liang ,&Xiao, X.-J..(2013).Performance testing and analysis of MAGMA library on hybrid architecture (CPU+GPU).Ruan Jian Xue Bao/Journal of Software,24(SUPPL.2),118-126.
MLA	Xiao, Xuan-Ji ,et al."Performance testing and analysis of MAGMA library on hybrid architecture (CPU+GPU)".Ruan Jian Xue Bao/Journal of Software 24.SUPPL.2(2013):118-126.