Deconvolution of huge 3-D images: Parallelization strategies on
a multi-GPU system

D 2013

Deconvolution of huge 3-D images: Parallelization strategies on a multi-GPU system

KARAS, Pavel, Michal KUDERJAVÝ and David SVOBODA

Basic information

Original name

Deconvolution of huge 3-D images: Parallelization strategies on a multi-GPU system

Name in Czech

Dekonvoluce velkých 3-D obrazů: Strategie pro paralelizaci na multi-GPU systému

Authors

KARAS, Pavel (203 Czech Republic, guarantor, belonging to the institution), Michal KUDERJAVÝ (203 Czech Republic, belonging to the institution) and David SVOBODA (703 Slovakia, belonging to the institution)

Edition

Neuveden, Algorithms and Architectures for Parallel Processing, p. 279-290, 12 pp. 2013

Publisher

Springer International Publishing

Other information

Language

English

Type of outcome

Stať ve sborníku

Field of Study

10201 Computer sciences, information science, bioinformatics

Country of publisher

Italy

Confidentiality degree

není předmětem státního či obchodního tajemství

Publication form

printed version "print"

References:

URL

Impact factor

Impact factor: 0.402 in 2005

RIV identification code

RIV/00216224:14330/13:00066399

Organization unit

Faculty of Informatics

ISBN

978-3-319-03858-2

ISSN

DOI

http://dx.doi.org/10.1007/978-3-319-03859-9_24

Keywords (in Czech)

dekonvoluce; gpu; multi-gpu; paralelizace; implementace; algoritmus; em-mle; richardson-lucy; ictm; wiener

Keywords in English

deconvolution; gpu; multi-gpu; parallelization; implementation; algorithm; em-mle; richardson-lucy; ictm; wiener

Abstract

V originále

In this paper, we discuss strategies to parallelize selected deconvolution methods on a multi-GPU system. We provide a comparison of several approaches to split the deconvolution into subtasks while keeping the amount of costly data transfers as low as possible, and propose own implementation of three deconvolution methods which achieves up to 65x speedup over the CPU one. In the experimental part, we analyse how the individual stages of the computation contribute to the overall computation time as well as how the multi-GPU implementation scales in various setups. Finally, we identify bottlenecks of the system.

Links

GBP302/12/G157, research and development project

Name: Dynamika a organizace chromosomů během buněčného cyklu a při diferenciaci v normě a patologii

Investor: Czech Science Foundation

MUNI/A/0760/2012, interní kód MU

Name: Rozsáhlé výpočetní systémy: modely, aplikace a verifikace II. (Acronym: FI MAV II.)

Investor: Masaryk University, Category A

Citovat

KARAS, Pavel, Michal KUDERJAVÝ and David SVOBODA. Deconvolution of huge 3-D images: Parallelization strategies on a multi-GPU system. In Kołodziej, Joanna and Martino, Beniamino and Talia, Domenico and Xiong, Kaiqi. Algorithms and Architectures for Parallel Processing. Neuveden: Springer International Publishing, 2013, p. 279-290. ISBN 978-3-319-03858-2. Available from: https://dx.doi.org/10.1007/978-3-319-03859-9_24.

@inproceedings{1124393,
   author = {Karas, Pavel and Kuderjavý, Michal and Svoboda, David},
   address = {Neuveden},
   booktitle = {Algorithms and Architectures for Parallel Processing},
   doi = {http://dx.doi.org/10.1007/978-3-319-03859-9_24},
   editor = {Kołodziej, Joanna and Martino, Beniamino and Talia, Domenico and Xiong, Kaiqi},
   keywords = {deconvolution; gpu; multi-gpu; parallelization; implementation; algorithm; em-mle; richardson-lucy; ictm; wiener},
   howpublished = {tištěná verze "print"},
   language = {eng},
   location = {Neuveden},
   isbn = {978-3-319-03858-2},
   pages = {279-290},
   publisher = {Springer International Publishing},
   title = {Deconvolution of huge 3-D images: Parallelization strategies on a multi-GPU system},
   url = {http://dx.doi.org/10.1007/978-3-319-03859-9_24},
   year = {2013}
}

TY  - JOUR
ID  - 1124393
AU  - Karas, Pavel - Kuderjavý, Michal - Svoboda, David
PY  - 2013
TI  - Deconvolution of huge 3-D images: Parallelization strategies on a multi-GPU system
PB  - Springer International Publishing
CY  - Neuveden
SN  - 9783319038582
KW  - deconvolution
KW  - gpu
KW  - multi-gpu
KW  - parallelization
KW  - implementation
KW  - algorithm
KW  - em-mle
KW  - richardson-lucy
KW  - ictm
KW  - wiener
UR  - http://dx.doi.org/10.1007/978-3-319-03859-9_24
L2  - http://dx.doi.org/10.1007/978-3-319-03859-9_24
N2  - In this paper, we discuss strategies to parallelize selected deconvolution methods on a multi-GPU system. We provide a comparison of several approaches to split the deconvolution into subtasks while keeping the amount of costly data transfers as low as possible, and propose own implementation of three deconvolution methods which achieves up to 65x speedup over the CPU one. In the experimental part, we analyse how the individual stages of the computation contribute to the overall computation time as well as how the multi-GPU implementation scales in various setups. Finally, we identify bottlenecks of the system.
ER  -

KARAS, Pavel, Michal KUDERJAVÝ and David SVOBODA. Deconvolution of huge 3-D images: Parallelization strategies on a multi-GPU system. In Kołodziej, Joanna and Martino, Beniamino and Talia, Domenico and Xiong, Kaiqi. \textit{Algorithms and Architectures for Parallel Processing}. Neuveden: Springer International Publishing, 2013, p.~279-290. ISBN~978-3-319-03858-2. Available from: https://dx.doi.org/10.1007/978-3-319-03859-9\_{}24.

Detailed Information on Publication Record