An efficient hybrid tridiagonal divide-and-conquer algorithm on distributed memory architectures
文献类型：期刊论文
| 作者 | Li, Shengguo1; Rouet, Francois-Henry2,3; Liu, Jie1,4; Huang, Chun1,4; Gao, Xingyu5; Chi, Xuebin6 |
| 刊名 | Journal of computational and applied mathematics
 |
| 出版日期 | 2018-12-15 |
| 卷号 | 344页码:512-520 |
| ISSN号 | 0377-0427 |
| 关键词 | Scalapack Divide-and-conquer Hss matrix Distributed parallel algorithm |
| DOI | 10.1016/j.cam.2018.05.051 |
| 通讯作者 | Li, shengguo(nudtlsg@nudt.edu.cn) |
| 英文摘要 | In this paper, we propose an efficient divide-and-conquer (dc) algorithm for symmetric tridiagonal matrices based on scalapack and the hierarchically semiseparable (hss) matrices. hss is an important type of rank-structured matrices. the most computationally intensive part of the dc algorithm is computing the eigenvectors via matrix-matrix multiplications (mmm). in our parallel hybrid dc (phdc) algorithm, mmm is accelerated by using hss matrix techniques when the intermediate matrix is large. all the hss computations are performed via the package strumpack. phdc has been tested by using many different matrices. compared with the dc implementation in mkl, phdc can be faster for some matrices with few deflations when using hundreds of processes. however, the gains decrease as the number of processes increases. the comparisons of phdc with elpa (the eigenvalue solvers for petascale applications library) are similar. phdc is usually slower than mkl and elpa when using 300 or more processes on the tianhe-2 supercomputer. (c) 2018 elsevier b.v. all rights reserved. |
| WOS关键词 | POSITIVE-DEFINITE MATRICES ; EIGENVALUE PROBLEM ; SEMISEPARABLE REPRESENTATIONS ; EIGENPROBLEM ; SOLVER ; APPROXIMATION |
| WOS研究方向 | Mathematics |
| WOS类目 | Mathematics, Applied |
| 语种 | 英语 |
| 出版者 | ELSEVIER SCIENCE BV |
| WOS记录号 | WOS:000440394900034 |
| URI标识 | http://www.irgrid.ac.cn/handle/1471x/2374228 |
| 专题 | 计算机网络信息中心 |
| 通讯作者 | Li, Shengguo |
| 作者单位 | 1.Natl Univ Def Technol, Coll Comp, Changsha 410073, Hunan, Peoples R China 2.Livermore Software Technol Corp, Livermore, CA 94550 USA 3.Lawrence Berkeley Natl Lab, Berkeley, CA 94720 USA 4.NUDT, State Key Lab High Performance Comp, Changsha, Hunan, Peoples R China 5.Inst Appl Phys & Computat Math, Beijing 100094, Peoples R China 6.Chinese Acad Sci, Comp Network Informat Ctr, Beijing 100190, Peoples R China |
| 推荐引用方式 GB/T 7714 | Li, Shengguo,Rouet, Francois-Henry,Liu, Jie,et al. An efficient hybrid tridiagonal divide-and-conquer algorithm on distributed memory architectures[J]. Journal of computational and applied mathematics,2018,344:512-520. |
| APA | Li, Shengguo,Rouet, Francois-Henry,Liu, Jie,Huang, Chun,Gao, Xingyu,&Chi, Xuebin.(2018).An efficient hybrid tridiagonal divide-and-conquer algorithm on distributed memory architectures.Journal of computational and applied mathematics,344,512-520. |
| MLA | Li, Shengguo,et al."An efficient hybrid tridiagonal divide-and-conquer algorithm on distributed memory architectures".Journal of computational and applied mathematics 344(2018):512-520. |
入库方式： iSwitch采集
来源：计算机网络信息中心
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。