The performances of R GPU implementations of the GMRES method
Although the performance of commodity computers has improved drastically with the introduction of multicore processors and GPU computing, the standard R distribution is still based on single-threaded model of computation, using only a small fraction of the computational power available now for most...
Main Authors: | , |
---|---|
Format: | Article |
Language: | English |
Published: |
Romanian National Institute of Statistics
2018-03-01
|
Series: | Revista Română de Statistică |
Subjects: | |
Online Access: | http://www.revistadestatistica.ro/wp-content/uploads/2018/03/RRS_1_2018_A09.pdf |
id |
doaj-44e9af5ceee5416db215338df15df4f8 |
---|---|
record_format |
Article |
spelling |
doaj-44e9af5ceee5416db215338df15df4f82020-11-25T02:17:16ZengRomanian National Institute of StatisticsRevista Română de Statistică1018-046X1844-76942018-03-01661121132The performances of R GPU implementations of the GMRES methodBogdan Oancea0Richard Pospisil 1University of BucharestPalacky University of OlomoucAlthough the performance of commodity computers has improved drastically with the introduction of multicore processors and GPU computing, the standard R distribution is still based on single-threaded model of computation, using only a small fraction of the computational power available now for most desktops and laptops. Modern statistical software packages rely on high performance implementations of the linear algebra routines there are at the core of several important leading edge statistical methods. In this paper we present a GPU implementation of the GMRES iterative method for solving linear systems. We compare the performance of this implementation with a pure single threaded version of the CPU. We also investigate the performance of our implementation using different GPU packages available now for R such as gmatrix, gputools or gpuR which are based on CUDA or OpenCL frameworks.http://www.revistadestatistica.ro/wp-content/uploads/2018/03/RRS_1_2018_A09.pdfRGPUstatistical softwareGMRES |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Bogdan Oancea Richard Pospisil |
spellingShingle |
Bogdan Oancea Richard Pospisil The performances of R GPU implementations of the GMRES method Revista Română de Statistică R GPU statistical software GMRES |
author_facet |
Bogdan Oancea Richard Pospisil |
author_sort |
Bogdan Oancea |
title |
The performances of R GPU implementations of the GMRES method |
title_short |
The performances of R GPU implementations of the GMRES method |
title_full |
The performances of R GPU implementations of the GMRES method |
title_fullStr |
The performances of R GPU implementations of the GMRES method |
title_full_unstemmed |
The performances of R GPU implementations of the GMRES method |
title_sort |
performances of r gpu implementations of the gmres method |
publisher |
Romanian National Institute of Statistics |
series |
Revista Română de Statistică |
issn |
1018-046X 1844-7694 |
publishDate |
2018-03-01 |
description |
Although the performance of commodity computers has improved drastically with the introduction of multicore processors and GPU computing, the standard R distribution is still based on single-threaded model of computation, using only a small fraction of the computational power available now for most desktops and laptops. Modern statistical software packages rely on high performance implementations of the linear algebra routines there are at the core of several important leading edge statistical methods. In this paper we present a GPU implementation of the GMRES iterative method for solving linear systems. We compare the performance of this implementation with a pure single threaded version of the CPU. We also investigate the performance of our implementation using different GPU packages available now for R such as gmatrix, gputools or gpuR which are based on CUDA or OpenCL frameworks. |
topic |
R GPU statistical software GMRES |
url |
http://www.revistadestatistica.ro/wp-content/uploads/2018/03/RRS_1_2018_A09.pdf |
work_keys_str_mv |
AT bogdanoancea theperformancesofrgpuimplementationsofthegmresmethod AT richardpospisil theperformancesofrgpuimplementationsofthegmresmethod AT bogdanoancea performancesofrgpuimplementationsofthegmresmethod AT richardpospisil performancesofrgpuimplementationsofthegmresmethod |
_version_ |
1724887331560751104 |