No Cover Image

Journal article 145 views

POETS: An Event-driven Approach to Dissipative Particle Dynamics: Implementing a Massively Compute-intensive Problem on a Novel Hard/Software Architecture

Andrew D. Brown Orcid Logo, Jonathan R. Beaumont Orcid Logo, David B. Thomas Orcid Logo, Julian C. Shillcock Orcid Logo, Matthew F. Naylor Orcid Logo, Graeme M. Bragg Orcid Logo, Mark L. Vousden Orcid Logo, Simon W. Moore Orcid Logo, Shane Fleming

ACM Transactions on Parallel Computing, Volume: 10, Issue: 2, Pages: 1 - 32

Swansea University Author: Shane Fleming

Full text not available from this repository: check for access using links below.

Check full text

DOI (Published version): 10.1145/3580372

Abstract

HPC clusters have become ever more expensive, both in terms of capital cost and energy consumption; some estimates suggest that competitive installations at the end of the next decade will require their own power station. One way around this looming problem is to design bespoke computing engines, bu...

Full description

Published in: ACM Transactions on Parallel Computing
ISSN: 2329-4949 2329-4957
Published: Association for Computing Machinery (ACM) 2023
Online Access: Check full text

URI: https://cronfa.swan.ac.uk/Record/cronfa69820
first_indexed 2025-06-26T12:37:27Z
last_indexed 2025-07-18T04:59:25Z
id cronfa69820
recordtype SURis
fullrecord <?xml version="1.0"?><rfc1807><datestamp>2025-07-17T16:03:43.1260243</datestamp><bib-version>v2</bib-version><id>69820</id><entry>2025-06-26</entry><title>POETS: An Event-driven Approach to Dissipative Particle Dynamics: Implementing a Massively Compute-intensive Problem on a Novel Hard/Software Architecture</title><swanseaauthors><author><sid>fe23ad3ebacc194b4f4c480fdde55b95</sid><firstname>Shane</firstname><surname>Fleming</surname><name>Shane Fleming</name><active>true</active><ethesisStudent>false</ethesisStudent></author></swanseaauthors><date>2025-06-26</date><deptcode>MACS</deptcode><abstract>HPC clusters have become ever more expensive, both in terms of capital cost and energy consumption; some estimates suggest that competitive installations at the end of the next decade will require their own power station. One way around this looming problem is to design bespoke computing engines, but while the performance benefits are good, the design costs are huge and cannot easily be amortized. Partially Ordered Event Triggered System (POETS)&#x2014;the focus of this article&#x2014;seeks to exploit a middle way: The architecture is tuned to a specific algorithmic pattern but, within that constraint, is fully programmable. POETS software is quasi-imperative: The user defines a set of sequential event handlers, defines the topology of a (typically large) concurrent ensemble of these, and lets them interact. The &#x201C;solution&#x201D; may be exfiltrated from the emergent behaviour of the ensemble. In this article, we describe (briefly) the architecture, and an example computational chemistry application, dissipative particle dynamics (DPD). The DPD algorithm is traditionally implemented using parallel computational techniques, but we re-cast it as a concurrent compute problem that is then ideally suited to POETS. Our prototype system is realised on a cluster of 48 FPGAs providing 50K concurrent hardware threads, and we report performance speedups of over two orders of magnitude better than a single thread baseline comparator and scaling behaviour that is almost constant. The results are validated against a &#x201C;conventional&#x201D; implementation.</abstract><type>Journal Article</type><journal>ACM Transactions on Parallel Computing</journal><volume>10</volume><journalNumber>2</journalNumber><paginationStart>1</paginationStart><paginationEnd>32</paginationEnd><publisher>Association for Computing Machinery (ACM)</publisher><placeOfPublication/><isbnPrint/><isbnElectronic/><issnPrint>2329-4949</issnPrint><issnElectronic>2329-4957</issnElectronic><keywords>Parallel architectures, event-driven processing, massively micro-parallel systems</keywords><publishedDay>20</publishedDay><publishedMonth>6</publishedMonth><publishedYear>2023</publishedYear><publishedDate>2023-06-20</publishedDate><doi>10.1145/3580372</doi><url/><notes/><college>COLLEGE NANME</college><department>Mathematics and Computer Science School</department><CollegeCode>COLLEGE CODE</CollegeCode><DepartmentCode>MACS</DepartmentCode><institution>Swansea University</institution><apcterm>Not Required</apcterm><funders>EPSRC Grant EP/N031768/1; Swiss Governments ETH Board of the Swiss Federal Institutes of Technology</funders><projectreference/><lastEdited>2025-07-17T16:03:43.1260243</lastEdited><Created>2025-06-26T11:46:28.1729999</Created><path><level id="1">Faculty of Science and Engineering</level><level id="2">School of Mathematics and Computer Science - Computer Science</level></path><authors><author><firstname>Andrew D.</firstname><surname>Brown</surname><orcid>0000-0002-0700-9433</orcid><order>1</order></author><author><firstname>Jonathan R.</firstname><surname>Beaumont</surname><orcid>0000-0002-1217-8725</orcid><order>2</order></author><author><firstname>David B.</firstname><surname>Thomas</surname><orcid>0000-0002-9671-0917</orcid><order>3</order></author><author><firstname>Julian C.</firstname><surname>Shillcock</surname><orcid>0000-0002-7885-735X</orcid><order>4</order></author><author><firstname>Matthew F.</firstname><surname>Naylor</surname><orcid>0000-0001-9827-8497</orcid><order>5</order></author><author><firstname>Graeme M.</firstname><surname>Bragg</surname><orcid>0000-0002-5201-7977</orcid><order>6</order></author><author><firstname>Mark L.</firstname><surname>Vousden</surname><orcid>0000-0002-6552-5831</orcid><order>7</order></author><author><firstname>Simon W.</firstname><surname>Moore</surname><orcid>0000-0002-2806-495X</orcid><order>8</order></author><author><firstname>Shane</firstname><surname>Fleming</surname><order>9</order></author></authors><documents/><OutputDurs/></rfc1807>
spelling 2025-07-17T16:03:43.1260243 v2 69820 2025-06-26 POETS: An Event-driven Approach to Dissipative Particle Dynamics: Implementing a Massively Compute-intensive Problem on a Novel Hard/Software Architecture fe23ad3ebacc194b4f4c480fdde55b95 Shane Fleming Shane Fleming true false 2025-06-26 MACS HPC clusters have become ever more expensive, both in terms of capital cost and energy consumption; some estimates suggest that competitive installations at the end of the next decade will require their own power station. One way around this looming problem is to design bespoke computing engines, but while the performance benefits are good, the design costs are huge and cannot easily be amortized. Partially Ordered Event Triggered System (POETS)—the focus of this article—seeks to exploit a middle way: The architecture is tuned to a specific algorithmic pattern but, within that constraint, is fully programmable. POETS software is quasi-imperative: The user defines a set of sequential event handlers, defines the topology of a (typically large) concurrent ensemble of these, and lets them interact. The “solution” may be exfiltrated from the emergent behaviour of the ensemble. In this article, we describe (briefly) the architecture, and an example computational chemistry application, dissipative particle dynamics (DPD). The DPD algorithm is traditionally implemented using parallel computational techniques, but we re-cast it as a concurrent compute problem that is then ideally suited to POETS. Our prototype system is realised on a cluster of 48 FPGAs providing 50K concurrent hardware threads, and we report performance speedups of over two orders of magnitude better than a single thread baseline comparator and scaling behaviour that is almost constant. The results are validated against a “conventional” implementation. Journal Article ACM Transactions on Parallel Computing 10 2 1 32 Association for Computing Machinery (ACM) 2329-4949 2329-4957 Parallel architectures, event-driven processing, massively micro-parallel systems 20 6 2023 2023-06-20 10.1145/3580372 COLLEGE NANME Mathematics and Computer Science School COLLEGE CODE MACS Swansea University Not Required EPSRC Grant EP/N031768/1; Swiss Governments ETH Board of the Swiss Federal Institutes of Technology 2025-07-17T16:03:43.1260243 2025-06-26T11:46:28.1729999 Faculty of Science and Engineering School of Mathematics and Computer Science - Computer Science Andrew D. Brown 0000-0002-0700-9433 1 Jonathan R. Beaumont 0000-0002-1217-8725 2 David B. Thomas 0000-0002-9671-0917 3 Julian C. Shillcock 0000-0002-7885-735X 4 Matthew F. Naylor 0000-0001-9827-8497 5 Graeme M. Bragg 0000-0002-5201-7977 6 Mark L. Vousden 0000-0002-6552-5831 7 Simon W. Moore 0000-0002-2806-495X 8 Shane Fleming 9
title POETS: An Event-driven Approach to Dissipative Particle Dynamics: Implementing a Massively Compute-intensive Problem on a Novel Hard/Software Architecture
spellingShingle POETS: An Event-driven Approach to Dissipative Particle Dynamics: Implementing a Massively Compute-intensive Problem on a Novel Hard/Software Architecture
Shane Fleming
title_short POETS: An Event-driven Approach to Dissipative Particle Dynamics: Implementing a Massively Compute-intensive Problem on a Novel Hard/Software Architecture
title_full POETS: An Event-driven Approach to Dissipative Particle Dynamics: Implementing a Massively Compute-intensive Problem on a Novel Hard/Software Architecture
title_fullStr POETS: An Event-driven Approach to Dissipative Particle Dynamics: Implementing a Massively Compute-intensive Problem on a Novel Hard/Software Architecture
title_full_unstemmed POETS: An Event-driven Approach to Dissipative Particle Dynamics: Implementing a Massively Compute-intensive Problem on a Novel Hard/Software Architecture
title_sort POETS: An Event-driven Approach to Dissipative Particle Dynamics: Implementing a Massively Compute-intensive Problem on a Novel Hard/Software Architecture
author_id_str_mv fe23ad3ebacc194b4f4c480fdde55b95
author_id_fullname_str_mv fe23ad3ebacc194b4f4c480fdde55b95_***_Shane Fleming
author Shane Fleming
author2 Andrew D. Brown
Jonathan R. Beaumont
David B. Thomas
Julian C. Shillcock
Matthew F. Naylor
Graeme M. Bragg
Mark L. Vousden
Simon W. Moore
Shane Fleming
format Journal article
container_title ACM Transactions on Parallel Computing
container_volume 10
container_issue 2
container_start_page 1
publishDate 2023
institution Swansea University
issn 2329-4949
2329-4957
doi_str_mv 10.1145/3580372
publisher Association for Computing Machinery (ACM)
college_str Faculty of Science and Engineering
hierarchytype
hierarchy_top_id facultyofscienceandengineering
hierarchy_top_title Faculty of Science and Engineering
hierarchy_parent_id facultyofscienceandengineering
hierarchy_parent_title Faculty of Science and Engineering
department_str School of Mathematics and Computer Science - Computer Science{{{_:::_}}}Faculty of Science and Engineering{{{_:::_}}}School of Mathematics and Computer Science - Computer Science
document_store_str 0
active_str 0
description HPC clusters have become ever more expensive, both in terms of capital cost and energy consumption; some estimates suggest that competitive installations at the end of the next decade will require their own power station. One way around this looming problem is to design bespoke computing engines, but while the performance benefits are good, the design costs are huge and cannot easily be amortized. Partially Ordered Event Triggered System (POETS)—the focus of this article—seeks to exploit a middle way: The architecture is tuned to a specific algorithmic pattern but, within that constraint, is fully programmable. POETS software is quasi-imperative: The user defines a set of sequential event handlers, defines the topology of a (typically large) concurrent ensemble of these, and lets them interact. The “solution” may be exfiltrated from the emergent behaviour of the ensemble. In this article, we describe (briefly) the architecture, and an example computational chemistry application, dissipative particle dynamics (DPD). The DPD algorithm is traditionally implemented using parallel computational techniques, but we re-cast it as a concurrent compute problem that is then ideally suited to POETS. Our prototype system is realised on a cluster of 48 FPGAs providing 50K concurrent hardware threads, and we report performance speedups of over two orders of magnitude better than a single thread baseline comparator and scaling behaviour that is almost constant. The results are validated against a “conventional” implementation.
published_date 2023-06-20T05:27:58Z
_version_ 1851369642555080704
score 11.089572

Similar Items