Scaling Molecular Dynamics with ab initio Accuracy to 149 Nanoseconds per Day

Physical phenomena such as chemical reactions, bond breaking, and phase transition require molecular dynamics (MD) simulation with ab initio accuracy ranging from milliseconds to microseconds. However, previous state-of-the-art neural network based MD packages such as DeePMD-kit can only reach 4.7 n...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:SC24: International Conference for High Performance Computing, Networking, Storage and Analysis S. 1 - 15
Hauptverfasser: Li, Jianxiong, Li, Boyang, Guo, Zhuoqiang, Li, Mingzhen, Li, Enji, Liu, Lijun, Yuan, Guojun, Wang, Zhan, Tan, Guangming, Jia, Weile
Format: Tagungsbericht
Sprache:Englisch
Veröffentlicht: IEEE 17.11.2024
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Abstract Physical phenomena such as chemical reactions, bond breaking, and phase transition require molecular dynamics (MD) simulation with ab initio accuracy ranging from milliseconds to microseconds. However, previous state-of-the-art neural network based MD packages such as DeePMD-kit can only reach 4.7 nanoseconds per day on the Fugaku supercomputer. In this paper, we present a novel node-based parallelization scheme to reduce communication by 81%, then optimize the computationally intensive kernels with sve-gemm and mixed precision. Finally, we implement intra-node load balance to further improve the scalability. Numerical results on the Fugaku supercomputer show that our work has significantly improved the time-to-solution of the DeePMD-kit by a factor of 31.7 x, reaching 149 nanoseconds per day on 12,000 computing nodes. This work has opened the door for millisecond simulation with ab initio accuracy within one week for the first time.
AbstractList Physical phenomena such as chemical reactions, bond breaking, and phase transition require molecular dynamics (MD) simulation with ab initio accuracy ranging from milliseconds to microseconds. However, previous state-of-the-art neural network based MD packages such as DeePMD-kit can only reach 4.7 nanoseconds per day on the Fugaku supercomputer. In this paper, we present a novel node-based parallelization scheme to reduce communication by 81%, then optimize the computationally intensive kernels with sve-gemm and mixed precision. Finally, we implement intra-node load balance to further improve the scalability. Numerical results on the Fugaku supercomputer show that our work has significantly improved the time-to-solution of the DeePMD-kit by a factor of 31.7 x, reaching 149 nanoseconds per day on 12,000 computing nodes. This work has opened the door for millisecond simulation with ab initio accuracy within one week for the first time.
Author Li, Boyang
Yuan, Guojun
Li, Mingzhen
Li, Enji
Jia, Weile
Li, Jianxiong
Guo, Zhuoqiang
Wang, Zhan
Liu, Lijun
Tan, Guangming
Author_xml – sequence: 1
  givenname: Jianxiong
  surname: Li
  fullname: Li, Jianxiong
  email: lijianxiong20g@ict.ac.cn
  organization: Institute of Computing Technology,Chinese Academy of Sciences,State Key Lab of Processors
– sequence: 2
  givenname: Boyang
  surname: Li
  fullname: Li, Boyang
  email: liboyang22s@ict.ac.cn
  organization: Institute of Computing Technology,Chinese Academy of Sciences,State Key Lab of Processors
– sequence: 3
  givenname: Zhuoqiang
  surname: Guo
  fullname: Guo, Zhuoqiang
  email: guozhuoqiang20z@ict.ac.cn
  organization: Institute of Computing Technology,Chinese Academy of Sciences,State Key Lab of Processors
– sequence: 4
  givenname: Mingzhen
  surname: Li
  fullname: Li, Mingzhen
  email: limingzhen@ict.ac.cn
  organization: Institute of Computing Technology,Chinese Academy of Sciences,State Key Lab of Processors
– sequence: 5
  givenname: Enji
  surname: Li
  fullname: Li, Enji
  email: lienji23s@ict.ac.cn
  organization: Institute of Computing Technology,Chinese Academy of Sciences,State Key Lab of Processors
– sequence: 6
  givenname: Lijun
  surname: Liu
  fullname: Liu, Lijun
  email: liu@mech.eng.osaka-u.ac.jp
  organization: Osaka University,Graduate School of Engineering,Department of Mechanical Engineering
– sequence: 7
  givenname: Guojun
  surname: Yuan
  fullname: Yuan, Guojun
  email: yuanguojun@ncic.ac.cn
  organization: Institute of Computing Technology,Chinese Academy of Sciences,State Key Lab of Processors
– sequence: 8
  givenname: Zhan
  surname: Wang
  fullname: Wang, Zhan
  email: wangzhan@ncic.ac.cn
  organization: Institute of Computing Technology,Chinese Academy of Sciences,State Key Lab of Processors
– sequence: 9
  givenname: Guangming
  surname: Tan
  fullname: Tan, Guangming
  email: tgm@ict.ac.cn
  organization: Institute of Computing Technology,Chinese Academy of Sciences,State Key Lab of Processors
– sequence: 10
  givenname: Weile
  surname: Jia
  fullname: Jia, Weile
  email: jiaweile@ict.ac.cn
  organization: Institute of Computing Technology,Chinese Academy of Sciences,State Key Lab of Processors
BookMark eNotj8tOwzAQRY0EElD6A4iFfyBlJnYy8bIqT6mFRWFdjd0JWEqdKkmF8vdEgtVdnaNzr9V5apModYuwQAR3v11ZtFAucsjtAgBMeabmjlxlCjBF7pAu1bzvo4eCyJABc6U228BNTF960zYSTg13-mFMfIih1z9x-NbsdUxxiK1ehnDqOIx6aDVap984tb2ENu17fZSJ4_FGXdTc9DL_35n6fHr8WL1k6_fn19VynXFe2CEr0Qs5qcgTeM4DAaHdF6GUqhZf09QrBq1HCCXXzgBUzrEBNBXmBYmZqbs_bxSR3bGLB-7GHQI5g9OtX4jtTVU
CODEN IEEPAD
ContentType Conference Proceeding
DBID 6IE
6IL
CBEJK
RIE
RIL
DOI 10.1109/SC41406.2024.00036
DatabaseName IEEE Electronic Library (IEL) Conference Proceedings
IEEE Xplore POP ALL
IEEE Xplore All Conference Proceedings
IEEE Electronic Library (IEL)
IEEE Proceedings Order Plans (POP All) 1998-Present
DatabaseTitleList
Database_xml – sequence: 1
  dbid: RIE
  name: IEEE Electronic Library (IEL)
  url: https://ieeexplore.ieee.org/
  sourceTypes: Publisher
DeliveryMethod fulltext_linktorsrc
EISBN 9798350352917
EndPage 15
ExternalDocumentID 10793130
Genre orig-research
GrantInformation_xml – fundername: National Postdoctoral Program for Innovative Talents
  funderid: 10.13039/501100012152
– fundername: RIKEN
  funderid: 10.13039/501100006264
– fundername: Chinese Academy of Sciences
  funderid: 10.13039/501100002367
– fundername: National Key Research and Development Program of China
  funderid: 10.13039/501100012166
– fundername: National Science Foundation
  funderid: 10.13039/100000001
GroupedDBID 6IE
6IL
ACM
ALMA_UNASSIGNED_HOLDINGS
APO
CBEJK
LHSKQ
RIE
RIL
ID FETCH-LOGICAL-a254t-61be79e87b70ba2c70714d5c6e8febf7350e314b10c6af9300899a301381257e3
IEDL.DBID RIE
ISICitedReferencesCount 0
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001414891300012&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
IngestDate Wed Jan 01 06:01:57 EST 2025
IsPeerReviewed false
IsScholarly false
Language English
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-a254t-61be79e87b70ba2c70714d5c6e8febf7350e314b10c6af9300899a301381257e3
PageCount 15
ParticipantIDs ieee_primary_10793130
PublicationCentury 2000
PublicationDate 2024-Nov.-17
PublicationDateYYYYMMDD 2024-11-17
PublicationDate_xml – month: 11
  year: 2024
  text: 2024-Nov.-17
  day: 17
PublicationDecade 2020
PublicationTitle SC24: International Conference for High Performance Computing, Networking, Storage and Analysis
PublicationTitleAbbrev SC
PublicationYear 2024
Publisher IEEE
Publisher_xml – name: IEEE
SSID ssib057737303
Score 1.9122721
Snippet Physical phenomena such as chemical reactions, bond breaking, and phase transition require molecular dynamics (MD) simulation with ab initio accuracy ranging...
SourceID ieee
SourceType Publisher
StartPage 1
SubjectTerms Accuracy
Computational efficiency
Computational Science
DeePMD-kit
Distance measurement
High performance computing
Kernel
LAMMPS
Molecular Dynamics
Neural networks
Optimization
Scalability
Supercomputers
Testing
Title Scaling Molecular Dynamics with ab initio Accuracy to 149 Nanoseconds per Day
URI https://ieeexplore.ieee.org/document/10793130
WOSCitedRecordID wos001414891300012&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV09T8MwELVoxcAEiCC-5YE14MRJbI-oUDHQqlIBdats5yJ1SaomrdR_z13aAAsDWxzJtnQ--_x89-4Yu88tHoweklAnRoQJ7olQWy1DncfOpIi2bcuQ-3xT47GezcxkT1ZvuTAA0AafwQN9tr78vPJreirDHY7ahGP3WE-pbEfW6pQnVUqitsqOGCPM43SQIHygOISYUmQLSsP8q4RKa0GGx_-c-4QFP1w8Pvm2MqfsAMozNpqiaLHJR111W_68Ky1fc3pZ5dbxBYUFVfzJ-_XK-i1vKo44heNxWtWEgvOaLwH72W3APoYv74PXcF8YIbSI5xqEew6UAa2cEs7GXhELKU99BroAVyiZCpBR4iLhM1sYKci3ZyU5JfE-o0Ces35ZlXDBeIr2Wohcx0XsE2syh_ezKHI0IFr_yF6ygGQxX-5yX8w7MVz98f-aHZG4ia0XqRvWb1ZruGWHftMs6tVdu2Jftz6T-w
linkProvider IEEE
linkToHtml http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1NT8MwDI1gIMEJEEN8kwPXQtqkS3pEg2mIbZq0gXabktSVdlmntUPav8fuVuDCgVtTKYnkOHFe7Gczdp9aPBg9qMCoRAQK90RgrJGBSSOXxIi2bcWQ--jpwcBMJslwS1avuDAAUAWfwQN9Vr78NPcreirDHY7ahGPvsr1YqUhs6Fq1-sRaS9RXWVNjRPI4aisEEBSJEFGSbEGJmH8VUalsSOfon7Mfs-YPG48Pv-3MCduB-Snrj1C42OT9ur4tf94Uly84va1y6_iMAoNy_uT9amn9mpc5R6TC8UDNC8LBacEXgP3susneOy_jdjfYlkYILCK6EgGfA52A0U4LZyOviYeUxr4FJgOXaRkLkKFyofAtmyVSkHfPSnJL4o1GgzxjjXk-h3PGY7TYQqQmyiKvbNJyeEMLQ0cDov0P7QVrkiymi032i2kthss__t-xg-6435v2XgdvV-yQRE_cvVBfs0a5XMEN2_ef5axY3lar9wWd_pdC
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Abook&rft.genre=proceeding&rft.title=SC24%3A+International+Conference+for+High+Performance+Computing%2C+Networking%2C+Storage+and+Analysis&rft.atitle=Scaling+Molecular+Dynamics+with+ab+initio+Accuracy+to+149+Nanoseconds+per+Day&rft.au=Li%2C+Jianxiong&rft.au=Li%2C+Boyang&rft.au=Guo%2C+Zhuoqiang&rft.au=Li%2C+Mingzhen&rft.date=2024-11-17&rft.pub=IEEE&rft.spage=1&rft.epage=15&rft_id=info:doi/10.1109%2FSC41406.2024.00036&rft.externalDocID=10793130