Probabilistic guards: A mechanism for increasing the granularity of work-stealing programs

We propose probabilistic guards and analyze their performance. To reduce the total task division cost, probabilistic guards can prevent thief workers from stealing small tasks from victim workers probabilistically. In this study, we have implemented probabilistic guards on a work-stealing framework...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Parallel computing Ročník 82; s. 19 - 36
Hlavní autoři: Yoritaka, Hiroshi, Matsui, Ken, Yasugi, Masahiro, Hiraishi, Tasuku, Umatani, Seiji
Médium: Journal Article
Jazyk:angličtina
Vydáno: Elsevier B.V 01.02.2019
Témata:
ISSN:0167-8191, 1872-7336
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Abstract We propose probabilistic guards and analyze their performance. To reduce the total task division cost, probabilistic guards can prevent thief workers from stealing small tasks from victim workers probabilistically. In this study, we have implemented probabilistic guards on a work-stealing framework called Tascell. Without an upper limit to the number of repeated probabilistically prevented steal attempts, a thief may repeat an unbounded number of probabilistically prevented steal attempts until success if a victim uses a probabilistic guard that rejects steal attempts with a non-zero probability. We measured the actual numbers of repeated attempts until success, and evaluated the performance of probabilistic guards with various upper limits. In this paper, we also propose virtual probabilistic guards that act as probabilistic guards without repeating probabilistically prevented steal attempts. Virtual probabilistic guards exhibit superior performance compared to probabilistic guards. Our evaluation is based on parallelized “highly serial” force calculation in a shared memory environment and five Tascell programs in a distributed memory environment.
AbstractList We propose probabilistic guards and analyze their performance. To reduce the total task division cost, probabilistic guards can prevent thief workers from stealing small tasks from victim workers probabilistically. In this study, we have implemented probabilistic guards on a work-stealing framework called Tascell. Without an upper limit to the number of repeated probabilistically prevented steal attempts, a thief may repeat an unbounded number of probabilistically prevented steal attempts until success if a victim uses a probabilistic guard that rejects steal attempts with a non-zero probability. We measured the actual numbers of repeated attempts until success, and evaluated the performance of probabilistic guards with various upper limits. In this paper, we also propose virtual probabilistic guards that act as probabilistic guards without repeating probabilistically prevented steal attempts. Virtual probabilistic guards exhibit superior performance compared to probabilistic guards. Our evaluation is based on parallelized “highly serial” force calculation in a shared memory environment and five Tascell programs in a distributed memory environment.
Author Matsui, Ken
Umatani, Seiji
Hiraishi, Tasuku
Yasugi, Masahiro
Yoritaka, Hiroshi
Author_xml – sequence: 1
  givenname: Hiroshi
  surname: Yoritaka
  fullname: Yoritaka, Hiroshi
  organization: Kyushu Institute of Technology, Iizuka, Fukuoka 820-8502, Japan
– sequence: 2
  givenname: Ken
  surname: Matsui
  fullname: Matsui, Ken
  organization: Kyoto University, Kyoto 606-8501, Japan
– sequence: 3
  givenname: Masahiro
  surname: Yasugi
  fullname: Yasugi, Masahiro
  email: yasugi@ai.kyutech.ac.jp
  organization: Kyushu Institute of Technology, Iizuka, Fukuoka 820-8502, Japan
– sequence: 4
  givenname: Tasuku
  surname: Hiraishi
  fullname: Hiraishi, Tasuku
  organization: Kyoto University, Kyoto 606-8501, Japan
– sequence: 5
  givenname: Seiji
  surname: Umatani
  fullname: Umatani, Seiji
  organization: Kyoto University, Kyoto 606-8501, Japan
BookMark eNp9kM1OAyEUhYmpiW31CdzwAjNCmQJj4qJp_Eua6EI3bggDl5Y6Aw1MNX17p9a1q7O457s5-SZoFGIAhK4pKSmh_GZb7nQysZwRKkvCS0LYGRpTKWaFYIyP0HhoiULSml6gSc5bQgivJBmjj9cUG9341ufeG7ze62TzLV7gDsxGB5877GLCPpgEOvuwxv0G8DrpsG918v0BR4e_Y_oscg-6PRZ2KQ73Ll-ic6fbDFd_OUXvD_dvy6di9fL4vFysCsNo3ReOWKmpEKLhALISFQXCtJSc2ZmVVFQOGmdA1rqBOeNW1IIaLbgQThtL5myK2OmvSTHnBE7tku90OihK1FGP2qpfPeqoRxGuBj0DdXeiYJj25SGpbDwEA9YnML2y0f_L_wDwcnMm
Cites_doi 10.1016/0021-9991(90)90232-P
10.1038/324446a0
10.1109/71.86103
ContentType Journal Article
Copyright 2018 Elsevier B.V.
Copyright_xml – notice: 2018 Elsevier B.V.
DBID AAYXX
CITATION
DOI 10.1016/j.parco.2018.06.003
DatabaseName CrossRef
DatabaseTitle CrossRef
DatabaseTitleList
DeliveryMethod fulltext_linktorsrc
Discipline Computer Science
EISSN 1872-7336
EndPage 36
ExternalDocumentID 10_1016_j_parco_2018_06_003
S0167819118301881
GroupedDBID --K
--M
-~X
.DC
.~1
0R~
123
1B1
1~.
1~5
29O
4.4
457
4G.
5VS
6OB
7-5
71M
8P~
9JN
AACTN
AAEDT
AAEDW
AAIAV
AAIKJ
AAKOC
AALRI
AAOAW
AAQFI
AAQXK
AAXUO
AAYFN
ABBOA
ABEFU
ABFNM
ABJNI
ABMAC
ABXDB
ABYKQ
ACDAQ
ACGFS
ACNNM
ACRLP
ACZNC
ADBBV
ADEZE
ADJOM
ADMUD
ADTZH
AEBSH
AECPX
AEKER
AENEX
AFKWA
AFTJW
AGHFR
AGUBO
AGYEJ
AHHHB
AHJVU
AHZHX
AIALX
AIEXJ
AIKHN
AITUG
AJBFU
AJOXV
ALMA_UNASSIGNED_HOLDINGS
AMFUW
AMRAJ
AOUOD
ASPBG
AVWKF
AXJTR
AZFZN
BJAXD
BKOJK
BLXMC
CS3
DU5
EBS
EFJIC
EFLBG
EJD
EO8
EO9
EP2
EP3
F5P
FDB
FEDTE
FGOYB
FIRID
FNPLU
FYGXN
G-Q
G8K
GBLVA
GBOLZ
HLZ
HVGLF
HZ~
H~9
IHE
J1W
JJJVA
KOM
LG9
M41
MO0
N9A
O-L
O9-
OAUVE
OZT
P-8
P-9
P2P
PC.
Q38
R2-
RIG
ROL
RPZ
SBC
SCC
SDF
SDG
SDP
SES
SEW
SPC
SPCBC
SST
SSV
SSZ
T5K
WH7
WUQ
XPP
ZMT
~G-
9DU
AATTM
AAXKI
AAYWO
AAYXX
ABDPE
ABWVN
ACLOT
ACRPL
ACVFH
ADCNI
ADNMO
AEIPS
AEUPX
AFJKZ
AFPUW
AGQPQ
AIGII
AIIUN
AKBMS
AKRWK
AKYEP
ANKPU
APXCP
CITATION
EFKBS
~HD
ID FETCH-LOGICAL-c319t-f0d8a1777b6ee84741e03a8863d2d8174febfce89abe536d7971ca7677facd053
ISICitedReferencesCount 1
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000461410600003&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN 0167-8191
IngestDate Sat Nov 29 04:06:56 EST 2025
Fri Feb 23 02:29:26 EST 2024
IsPeerReviewed true
IsScholarly true
Keywords Work stealing
Probability
Barnes-Hut algorithm
Concurrency
Many-core
Parallel programming language
Language English
LinkModel OpenURL
MergedId FETCHMERGED-LOGICAL-c319t-f0d8a1777b6ee84741e03a8863d2d8174febfce89abe536d7971ca7677facd053
PageCount 18
ParticipantIDs crossref_primary_10_1016_j_parco_2018_06_003
elsevier_sciencedirect_doi_10_1016_j_parco_2018_06_003
PublicationCentury 2000
PublicationDate February 2019
2019-02-00
PublicationDateYYYYMMDD 2019-02-01
PublicationDate_xml – month: 02
  year: 2019
  text: February 2019
PublicationDecade 2010
PublicationTitle Parallel computing
PublicationYear 2019
Publisher Elsevier B.V
Publisher_xml – name: Elsevier B.V
References Mohr, Kranz, Halstead (bib0004) 1991; 2
Yasugi, Hiraishi, Umatani, Yuasa (bib0007) 2012; 20
J.E. Barnes, Treecode Guide
Halstead (bib0003) 1990
Barnes, Hut (bib0006) 1986; 324
Frigo, Leiserson, Randall (bib0001) 1998
Feeley (bib0015) 1993
Yoritaka, Matsui, Yasugi, Hiraishi, Umatani (bib0010) 2016
Janjic, Hammond (bib0021) 2010
.
Breuel (bib0014) 1998
Taura, Tabata, Yonezawa (bib0017) 1999
IBM Research, X10: Performance and Productivity at Scale.
Wang, Cui, Duan, Lu, Feng, Yew (bib0020) 2010
Olivier, Huan, Liu, Prins, Dinan, Sadayappan, Tseng (bib0013) 2007
Frigo, Halpern, Leiserson, Lewin-Berlin (bib0009) 2009
Intel Corporation, Intel Threading Building Block Reference Manual, 2007.
Hiraishi, Yasugi, Umatani, Yuasa (bib0005) 2009
Barnes (bib0012) 1990; 87
Intel Corporation, Intel Cilk Plus.
Leiserson, Schardl, Suksompong (bib0018) 2015
Duran, Corbalán, Ayguadé (bib0019) 2008
Frigo (10.1016/j.parco.2018.06.003_bib0009) 2009
10.1016/j.parco.2018.06.003_bib0011
Feeley (10.1016/j.parco.2018.06.003_bib0015) 1993
Yasugi (10.1016/j.parco.2018.06.003_bib0007) 2012; 20
Barnes (10.1016/j.parco.2018.06.003_bib0012) 1990; 87
Leiserson (10.1016/j.parco.2018.06.003_bib0018) 2015
Halstead (10.1016/j.parco.2018.06.003_bib0003) 1990
Hiraishi (10.1016/j.parco.2018.06.003_bib0005) 2009
Olivier (10.1016/j.parco.2018.06.003_bib0013) 2007
Frigo (10.1016/j.parco.2018.06.003_bib0001) 1998
Barnes (10.1016/j.parco.2018.06.003_bib0006) 1986; 324
Janjic (10.1016/j.parco.2018.06.003_bib0021) 2010
Yoritaka (10.1016/j.parco.2018.06.003_bib0010) 2016
Breuel (10.1016/j.parco.2018.06.003_bib0014) 1998
10.1016/j.parco.2018.06.003_bib0016
Taura (10.1016/j.parco.2018.06.003_bib0017) 1999
Duran (10.1016/j.parco.2018.06.003_bib0019) 2008
Wang (10.1016/j.parco.2018.06.003_bib0020) 2010
10.1016/j.parco.2018.06.003_bib0008
Mohr (10.1016/j.parco.2018.06.003_bib0004) 1991; 2
10.1016/j.parco.2018.06.003_bib0002
References_xml – volume: 20
  start-page: 128
  year: 2012
  end-page: 139
  ident: bib0007
  article-title: Parallel graph traversals using work-stealing frameworks for many-core platforms
  publication-title: J. Inf. Process.
– volume: 324
  start-page: 446
  year: 1986
  end-page: 449
  ident: bib0006
  article-title: A hierarchical
  publication-title: Nature
– reference: J.E. Barnes, Treecode Guide,
– start-page: 36:1
  year: 2008
  end-page: 36:11
  ident: bib0019
  article-title: An adaptive cut-off for task parallelism
  publication-title: Proceedings of the 2008 ACM/IEEE Conference on Supercomputing, SC ’08
– reference: IBM Research, X10: Performance and Productivity at Scale.
– reference: Intel Corporation, Intel Cilk Plus.
– start-page: 266
  year: 2010
  end-page: 277
  ident: bib0020
  article-title: An adaptive task creation strategy for work-stealing scheduling
  publication-title: Proceedings of the 8th Annual IEEE/ACM International Symposium on Code Generation and Optimization, CGO ’10
– start-page: 212
  year: 1998
  end-page: 223
  ident: bib0001
  article-title: The implementation of the Cilk-5 multithreaded language
  publication-title: Proc. of the ACM SIGPLAN Conf. PLDI
– year: 1998
  ident: bib0014
  article-title: Lexical closures for C++
  publication-title: Usenix Proceedings, C++ Conference
– start-page: 235
  year: 2007
  end-page: 250
  ident: bib0013
  article-title: UTS: an unbalanced tree search benchmark
  publication-title: Proceedings of the 19th International Conference on Languages and Compilers for Parallel Computing
– volume: 87
  start-page: 161
  year: 1990
  end-page: 170
  ident: bib0012
  article-title: A modified tree code: don’t laugh; it runs
  publication-title: J. Comput. Phys.
– start-page: 2
  year: 1990
  end-page: 57
  ident: bib0003
  article-title: New ideas in parallel lisp: language design, implementation, and programming tools
  publication-title: Parallel Lisp: Languages and Systems
– start-page: 1
  year: 2015
  end-page: 18
  ident: bib0018
  article-title: Upper bounds on number of steals in rooted trees
  publication-title: Theory Comput. Syst.
– reference: .
– volume: 2
  start-page: 264
  year: 1991
  end-page: 280
  ident: bib0004
  article-title: Lazy task creation: a technique for increasing the granularity of parallel programs
  publication-title: IEEE Trans. Parallel Distrib. Syst.
– start-page: 123
  year: 2010
  end-page: 134
  ident: bib0021
  article-title: Granularity-aware work-stealing for computationally-uniform grids
  publication-title: Proceedings of the 2010 10th IEEE/ACM International Conference on Cluster, Cloud and Grid Computing, CCGRID ’10
– reference: Intel Corporation, Intel Threading Building Block Reference Manual, 2007.
– start-page: 79
  year: 2009
  end-page: 90
  ident: bib0009
  article-title: Reducers and other Cilk++ hyperobjects
  publication-title: Proceedings of the Twenty-first Annual Symposium on Parallelism in Algorithms and Architectures
– start-page: 60
  year: 1999
  end-page: 71
  ident: bib0017
  article-title: StackThreads/MP: integrating futures into calling standards
  publication-title: Proc. of ACM SIGPLAN Symposium PPoPP
– start-page: 171
  year: 2016
  end-page: 180
  ident: bib0010
  article-title: Extending a work-stealing framework with probabilistic guards
  publication-title: Proceedings of the 45th International Conference on Parallel Processing Workshops (ICPPW 2016) (Ninth International Workshop on Parallel Programming Models and Systems Software for High-End Computing P2S2 2016, held in conjunction with ICPP 2016)
– start-page: 94
  year: 1993
  end-page: 107
  ident: bib0015
  article-title: A message passing implementation of lazy task creation
  publication-title: Proc. of the Intl. Workshop on Parallel Symbolic Computing: Languages, Systems, and Applications, number 748 in Lecture Notes in Computer Science
– start-page: 55
  year: 2009
  end-page: 64
  ident: bib0005
  article-title: Backtracking-based load balancing
  publication-title: Proc. of the 14th ACM SIGPLAN Symposium PPoPP
– volume: 87
  start-page: 161
  year: 1990
  ident: 10.1016/j.parco.2018.06.003_bib0012
  article-title: A modified tree code: don’t laugh; it runs
  publication-title: J. Comput. Phys.
  doi: 10.1016/0021-9991(90)90232-P
– start-page: 1
  year: 2015
  ident: 10.1016/j.parco.2018.06.003_bib0018
  article-title: Upper bounds on number of steals in rooted trees
  publication-title: Theory Comput. Syst.
– start-page: 171
  year: 2016
  ident: 10.1016/j.parco.2018.06.003_bib0010
  article-title: Extending a work-stealing framework with probabilistic guards
– start-page: 235
  year: 2007
  ident: 10.1016/j.parco.2018.06.003_bib0013
  article-title: UTS: an unbalanced tree search benchmark
– start-page: 2
  year: 1990
  ident: 10.1016/j.parco.2018.06.003_bib0003
  article-title: New ideas in parallel lisp: language design, implementation, and programming tools
– start-page: 79
  year: 2009
  ident: 10.1016/j.parco.2018.06.003_bib0009
  article-title: Reducers and other Cilk++ hyperobjects
– ident: 10.1016/j.parco.2018.06.003_bib0002
– volume: 324
  start-page: 446
  year: 1986
  ident: 10.1016/j.parco.2018.06.003_bib0006
  article-title: A hierarchical O(Nlog N) force-calculation algorithm
  publication-title: Nature
  doi: 10.1038/324446a0
– volume: 2
  start-page: 264
  issue: 3
  year: 1991
  ident: 10.1016/j.parco.2018.06.003_bib0004
  article-title: Lazy task creation: a technique for increasing the granularity of parallel programs
  publication-title: IEEE Trans. Parallel Distrib. Syst.
  doi: 10.1109/71.86103
– volume: 20
  start-page: 128
  issue: 1
  year: 2012
  ident: 10.1016/j.parco.2018.06.003_bib0007
  article-title: Parallel graph traversals using work-stealing frameworks for many-core platforms
  publication-title: J. Inf. Process.
– ident: 10.1016/j.parco.2018.06.003_bib0008
– start-page: 36:1
  year: 2008
  ident: 10.1016/j.parco.2018.06.003_bib0019
  article-title: An adaptive cut-off for task parallelism
– start-page: 123
  year: 2010
  ident: 10.1016/j.parco.2018.06.003_bib0021
  article-title: Granularity-aware work-stealing for computationally-uniform grids
– start-page: 60
  year: 1999
  ident: 10.1016/j.parco.2018.06.003_bib0017
  article-title: StackThreads/MP: integrating futures into calling standards
– year: 1998
  ident: 10.1016/j.parco.2018.06.003_bib0014
  article-title: Lexical closures for C++
– start-page: 266
  year: 2010
  ident: 10.1016/j.parco.2018.06.003_bib0020
  article-title: An adaptive task creation strategy for work-stealing scheduling
– ident: 10.1016/j.parco.2018.06.003_bib0016
– start-page: 94
  year: 1993
  ident: 10.1016/j.parco.2018.06.003_bib0015
  article-title: A message passing implementation of lazy task creation
– start-page: 212
  year: 1998
  ident: 10.1016/j.parco.2018.06.003_bib0001
  article-title: The implementation of the Cilk-5 multithreaded language
– start-page: 55
  year: 2009
  ident: 10.1016/j.parco.2018.06.003_bib0005
  article-title: Backtracking-based load balancing
– ident: 10.1016/j.parco.2018.06.003_bib0011
SSID ssj0006480
Score 2.194713
Snippet We propose probabilistic guards and analyze their performance. To reduce the total task division cost, probabilistic guards can prevent thief workers from...
SourceID crossref
elsevier
SourceType Index Database
Publisher
StartPage 19
SubjectTerms Barnes-Hut algorithm
Concurrency
Many-core
Parallel programming language
Probability
Work stealing
Title Probabilistic guards: A mechanism for increasing the granularity of work-stealing programs
URI https://dx.doi.org/10.1016/j.parco.2018.06.003
Volume 82
WOSCitedRecordID wos000461410600003&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
journalDatabaseRights – providerCode: PRVESC
  databaseName: Elsevier SD Freedom Collection Journals 2021
  customDbUrl:
  eissn: 1872-7336
  dateEnd: 99991231
  omitProxy: false
  ssIdentifier: ssj0006480
  issn: 0167-8191
  databaseCode: AIEXJ
  dateStart: 19950101
  isFulltext: true
  titleUrlDefault: https://www.sciencedirect.com
  providerName: Elsevier
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwtV07b9swECZcp0OXNn2hadOCQ7dWhRTJIpnNCFKkHQIPLmB0ESiRdJSHYugR5Af0h-f4kpSkCJqhi2DIEm3rPt4dz9_HQ-iziKSKEsmCSIZhkCQ5DQAlLGCzPKIhLzjj3DSbIMfHdLVii8nkj9fCXJ2TqqLX12zzX00N58DYWjr7CHP3g8IJeA1GhyOYHY7_ZPhFDVNUU171Dsxf1hoCjdWfX0gt89VdMZRhoeuEsfFyqTUELU1JdRQNzdcKNAKcXN2wuJpxKrvgte7Dcm5Y6V3rQ6DxILricGbS0qMSwvBJORS-26YrnSCov5433bq0yqGGn8AdQ0G81u2hzXtLuOqsG1cptDCqZ3y4wiU4ZL04HHte23bIuU7nOW0Qtpui3HPvttJw-m0DyNDKzchuvhrGQzTz_-DfCXI99dCz2k4zM0imB8kMty9-grb2yIzRKdqa_zhc_ewjepqYDnz9b_C7Vxme4L3v8vcMZ5S1LLfRc7fcwHMLk5doIqtX6IVv5YGdZ3-Nft9CDbao2cdz3GMGA2bwgBkMmMEjzOBLhW9hBnvMvEG_vh8uD44C13YjKMAft4EKBeURISRPpYTkJYFpHHNK01jsCQorWCVzVUjKeC5ncSoII1HBSUqI4oUAp_4WTavLSr5DmAiVspxDjit4oginEUz_VCkFiSJLhdhBX_2jyjZ2d5XsAQPtoNQ_zswliDbxywAgD934_nGf8wE9GzC8i6Zt3cmP6Glx1ZZN_cmh4wZrF41a
linkProvider Elsevier
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%3Ajournal&rft.genre=article&rft.atitle=Probabilistic+guards%3A+A+mechanism+for+increasing+the+granularity+of+work-stealing+programs&rft.jtitle=Parallel+computing&rft.au=Yoritaka%2C+Hiroshi&rft.au=Matsui%2C+Ken&rft.au=Yasugi%2C+Masahiro&rft.au=Hiraishi%2C+Tasuku&rft.date=2019-02-01&rft.issn=0167-8191&rft.volume=82&rft.spage=19&rft.epage=36&rft_id=info:doi/10.1016%2Fj.parco.2018.06.003&rft.externalDBID=n%2Fa&rft.externalDocID=10_1016_j_parco_2018_06_003
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=0167-8191&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=0167-8191&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=0167-8191&client=summon