No Cover Image

Journal article 424 views 114 downloads

Accelerating magnetic induction tomography‐based imaging through heterogeneous parallel computing

David W. Walker, Stephan C. Kramer, Fabian R. A. Biebl, Paul Ledger, Malcolm Brown

Concurrency and Computation: Practice and Experience, Volume: 31, Issue: 17, Start page: e5265

Swansea University Author: Paul Ledger

Check full text

DOI (Published version): 10.1002/cpe.5265

Abstract

Magnetic Induction Tomography (MIT) is a non‐invasive imaging technique, which has applications in both industrial and clinical settings. In essence, it is capable of reconstructing the electromagnetic parameters of an object from measurements made on its surface. With the exploitation of parallelis...

Full description

Published in: Concurrency and Computation: Practice and Experience
ISSN: 1532-0626 1532-0634
Published: 2019
Online Access: Check full text

URI: https://cronfa.swan.ac.uk/Record/cronfa49975
Tags: Add Tag
No Tags, Be the first to tag this record!
first_indexed 2019-04-15T09:29:09Z
last_indexed 2020-12-10T04:03:06Z
id cronfa49975
recordtype SURis
fullrecord <?xml version="1.0"?><rfc1807><datestamp>2020-12-09T11:31:56.3899339</datestamp><bib-version>v2</bib-version><id>49975</id><entry>2019-04-12</entry><title>Accelerating magnetic induction tomography&#x2010;based imaging through heterogeneous parallel computing</title><swanseaauthors><author><sid>068dd31af167bcda33878951b2a01e97</sid><firstname>Paul</firstname><surname>Ledger</surname><name>Paul Ledger</name><active>true</active><ethesisStudent>false</ethesisStudent></author></swanseaauthors><date>2019-04-12</date><deptcode>FGSEN</deptcode><abstract>Magnetic Induction Tomography (MIT) is a non&#x2010;invasive imaging technique, which has applications in both industrial and clinical settings. In essence, it is capable of reconstructing the electromagnetic parameters of an object from measurements made on its surface. With the exploitation of parallelism, it is possible to achieve high quality inexpensive MIT images for biomedical applications on clinically relevant time scales. In this paper we investigate the performance of different parallel implementations of the forward eddy current problem, which is the main computational component of the inverse problem through which measured voltages are converted into images. We show that a heterogeneous parallel method that exploits multiple CPUs and GPUs can provide a high level of parallel scaling, leading to considerably improved runtimes. We also show how multiple GPUs can be used in conjunction with deal.II, a widely&#x2010;used open source finite element library.</abstract><type>Journal Article</type><journal>Concurrency and Computation: Practice and Experience</journal><volume>31</volume><journalNumber>17</journalNumber><paginationStart>e5265</paginationStart><paginationEnd/><publisher/><placeOfPublication/><isbnPrint/><isbnElectronic/><issnPrint>1532-0626</issnPrint><issnElectronic>1532-0634</issnElectronic><keywords/><publishedDay>31</publishedDay><publishedMonth>12</publishedMonth><publishedYear>2019</publishedYear><publishedDate>2019-12-31</publishedDate><doi>10.1002/cpe.5265</doi><url/><notes/><college>COLLEGE NANME</college><department>Science and Engineering - Faculty</department><CollegeCode>COLLEGE CODE</CollegeCode><DepartmentCode>FGSEN</DepartmentCode><institution>Swansea University</institution><apcterm/><lastEdited>2020-12-09T11:31:56.3899339</lastEdited><Created>2019-04-12T09:23:06.2495450</Created><path><level id="1">College of Engineering</level><level id="2">Engineering</level></path><authors><author><firstname>David W.</firstname><surname>Walker</surname><order>1</order></author><author><firstname>Stephan C.</firstname><surname>Kramer</surname><order>2</order></author><author><firstname>Fabian R. A.</firstname><surname>Biebl</surname><order>3</order></author><author><firstname>Paul</firstname><surname>Ledger</surname><order>4</order></author><author><firstname>Malcolm</firstname><surname>Brown</surname><order>5</order></author></authors><documents><document><filename>0049975-12042019145926.pdf</filename><originalFilename>walker2019.pdf</originalFilename><uploaded>2019-04-12T14:59:26.9300000</uploaded><type>Output</type><contentLength>957345</contentLength><contentType>application/pdf</contentType><version>Accepted Manuscript</version><cronfaStatus>true</cronfaStatus><embargoDate>2020-04-11T00:00:00.0000000</embargoDate><copyrightCorrect>true</copyrightCorrect><language>eng</language></document></documents><OutputDurs/></rfc1807>
spelling 2020-12-09T11:31:56.3899339 v2 49975 2019-04-12 Accelerating magnetic induction tomography‐based imaging through heterogeneous parallel computing 068dd31af167bcda33878951b2a01e97 Paul Ledger Paul Ledger true false 2019-04-12 FGSEN Magnetic Induction Tomography (MIT) is a non‐invasive imaging technique, which has applications in both industrial and clinical settings. In essence, it is capable of reconstructing the electromagnetic parameters of an object from measurements made on its surface. With the exploitation of parallelism, it is possible to achieve high quality inexpensive MIT images for biomedical applications on clinically relevant time scales. In this paper we investigate the performance of different parallel implementations of the forward eddy current problem, which is the main computational component of the inverse problem through which measured voltages are converted into images. We show that a heterogeneous parallel method that exploits multiple CPUs and GPUs can provide a high level of parallel scaling, leading to considerably improved runtimes. We also show how multiple GPUs can be used in conjunction with deal.II, a widely‐used open source finite element library. Journal Article Concurrency and Computation: Practice and Experience 31 17 e5265 1532-0626 1532-0634 31 12 2019 2019-12-31 10.1002/cpe.5265 COLLEGE NANME Science and Engineering - Faculty COLLEGE CODE FGSEN Swansea University 2020-12-09T11:31:56.3899339 2019-04-12T09:23:06.2495450 College of Engineering Engineering David W. Walker 1 Stephan C. Kramer 2 Fabian R. A. Biebl 3 Paul Ledger 4 Malcolm Brown 5 0049975-12042019145926.pdf walker2019.pdf 2019-04-12T14:59:26.9300000 Output 957345 application/pdf Accepted Manuscript true 2020-04-11T00:00:00.0000000 true eng
title Accelerating magnetic induction tomography‐based imaging through heterogeneous parallel computing
spellingShingle Accelerating magnetic induction tomography‐based imaging through heterogeneous parallel computing
Paul Ledger
title_short Accelerating magnetic induction tomography‐based imaging through heterogeneous parallel computing
title_full Accelerating magnetic induction tomography‐based imaging through heterogeneous parallel computing
title_fullStr Accelerating magnetic induction tomography‐based imaging through heterogeneous parallel computing
title_full_unstemmed Accelerating magnetic induction tomography‐based imaging through heterogeneous parallel computing
title_sort Accelerating magnetic induction tomography‐based imaging through heterogeneous parallel computing
author_id_str_mv 068dd31af167bcda33878951b2a01e97
author_id_fullname_str_mv 068dd31af167bcda33878951b2a01e97_***_Paul Ledger
author Paul Ledger
author2 David W. Walker
Stephan C. Kramer
Fabian R. A. Biebl
Paul Ledger
Malcolm Brown
format Journal article
container_title Concurrency and Computation: Practice and Experience
container_volume 31
container_issue 17
container_start_page e5265
publishDate 2019
institution Swansea University
issn 1532-0626
1532-0634
doi_str_mv 10.1002/cpe.5265
college_str College of Engineering
hierarchytype
hierarchy_top_id collegeofengineering
hierarchy_top_title College of Engineering
hierarchy_parent_id collegeofengineering
hierarchy_parent_title College of Engineering
department_str Engineering{{{_:::_}}}College of Engineering{{{_:::_}}}Engineering
document_store_str 1
active_str 0
description Magnetic Induction Tomography (MIT) is a non‐invasive imaging technique, which has applications in both industrial and clinical settings. In essence, it is capable of reconstructing the electromagnetic parameters of an object from measurements made on its surface. With the exploitation of parallelism, it is possible to achieve high quality inexpensive MIT images for biomedical applications on clinically relevant time scales. In this paper we investigate the performance of different parallel implementations of the forward eddy current problem, which is the main computational component of the inverse problem through which measured voltages are converted into images. We show that a heterogeneous parallel method that exploits multiple CPUs and GPUs can provide a high level of parallel scaling, leading to considerably improved runtimes. We also show how multiple GPUs can be used in conjunction with deal.II, a widely‐used open source finite element library.
published_date 2019-12-31T04:03:17Z
_version_ 1737027168217595904
score 10.916305