Highly Scalable Parallelism of Integrated Randomized Singular Value Decomposition with Big Data Applications
碩士 === 國立臺灣大學 === 應用數學科學研究所 === 105 === Low-rank approximation plays an important role in big data analysis. Integrated Singular Value Decomposition (iSVD) is an algorithm for computing low-rank approximate singular value decomposition of large size matrices. The iSVD integrates different low-rank S...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Others |
Language: | en_US |
Published: |
2017
|
Online Access: | http://ndltd.ncl.edu.tw/handle/bvks4s |
id |
ndltd-TW-105NTU05507009 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-TW-105NTU055070092019-05-15T23:39:40Z http://ndltd.ncl.edu.tw/handle/bvks4s Highly Scalable Parallelism of Integrated Randomized Singular Value Decomposition with Big Data Applications 利用高度平行之演算法整合多重隨機奇異值分解並應用於巨量資料分析 Mu Yang 楊慕 碩士 國立臺灣大學 應用數學科學研究所 105 Low-rank approximation plays an important role in big data analysis. Integrated Singular Value Decomposition (iSVD) is an algorithm for computing low-rank approximate singular value decomposition of large size matrices. The iSVD integrates different low-rank SVDs obtained by multiple random subspace sketches and achieve higher accuracy and better stability. While iSVD takes higher computational costs due to multiple random sketches and the integration process, these operations can be parallelized to save computational time. We parallelize iSVD for multicore clusters, and modify the algorithms and data structures to increase the scalability and reduce communication. With parallelization, iSVD can find the approximate SVD of matrices with huge size, and achieve near-linear scalability with respect to the matrix size and the number of machines, and gained further 4X faster timing performance on sketching by using GPU. We implement the algorithms in C++, with several techniques for high maintainability, extensibility, and usability. The iSVD is applied on some huge size application using hybrid CPU-GPU supercomputer systems. Weichung Wang 王偉仲 2017 學位論文 ; thesis 88 en_US |
collection |
NDLTD |
language |
en_US |
format |
Others
|
sources |
NDLTD |
description |
碩士 === 國立臺灣大學 === 應用數學科學研究所 === 105 === Low-rank approximation plays an important role in big data analysis. Integrated Singular Value Decomposition (iSVD) is an algorithm for computing low-rank approximate singular value decomposition of large size matrices. The iSVD integrates different low-rank SVDs obtained by multiple random subspace sketches and achieve higher accuracy and better stability. While iSVD takes higher computational costs due to multiple random sketches and the integration process, these operations can be parallelized to save computational time. We parallelize iSVD for multicore clusters, and modify the algorithms and data structures to increase the scalability and reduce communication. With parallelization, iSVD can find the approximate SVD of matrices with huge size, and achieve near-linear scalability with respect to the matrix size and the number of machines, and gained further 4X faster timing performance on sketching by using GPU. We implement the algorithms in C++, with several techniques for high maintainability, extensibility, and usability. The iSVD is applied on some huge size application using hybrid CPU-GPU supercomputer systems.
|
author2 |
Weichung Wang |
author_facet |
Weichung Wang Mu Yang 楊慕 |
author |
Mu Yang 楊慕 |
spellingShingle |
Mu Yang 楊慕 Highly Scalable Parallelism of Integrated Randomized Singular Value Decomposition with Big Data Applications |
author_sort |
Mu Yang |
title |
Highly Scalable Parallelism of Integrated Randomized Singular Value Decomposition with Big Data Applications |
title_short |
Highly Scalable Parallelism of Integrated Randomized Singular Value Decomposition with Big Data Applications |
title_full |
Highly Scalable Parallelism of Integrated Randomized Singular Value Decomposition with Big Data Applications |
title_fullStr |
Highly Scalable Parallelism of Integrated Randomized Singular Value Decomposition with Big Data Applications |
title_full_unstemmed |
Highly Scalable Parallelism of Integrated Randomized Singular Value Decomposition with Big Data Applications |
title_sort |
highly scalable parallelism of integrated randomized singular value decomposition with big data applications |
publishDate |
2017 |
url |
http://ndltd.ncl.edu.tw/handle/bvks4s |
work_keys_str_mv |
AT muyang highlyscalableparallelismofintegratedrandomizedsingularvaluedecompositionwithbigdataapplications AT yángmù highlyscalableparallelismofintegratedrandomizedsingularvaluedecompositionwithbigdataapplications AT muyang lìyònggāodùpíngxíngzhīyǎnsuànfǎzhěnghéduōzhòngsuíjīqíyìzhífēnjiěbìngyīngyòngyújùliàngzīliàofēnxī AT yángmù lìyònggāodùpíngxíngzhīyǎnsuànfǎzhěnghéduōzhòngsuíjīqíyìzhífēnjiěbìngyīngyòngyújùliàngzīliàofēnxī |
_version_ |
1719152366823931904 |