General Purpose Task-Dependence Management Hardware for Task-Based Dataflow Programming Models

Task-based programming models such as OpenMP, IntelTBB and OmpSs offer the possibility of expressing dependences among tasks to drive their execution at runtime. Managing these dependences introduces noticeable overheads when targeting fine-grained tasks, diminishing the potential speedups or even i...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Proceedings - IEEE International Parallel and Distributed Processing Symposium S. 244 - 253
Hauptverfasser: Xubin Tan, Bosch, Jaume, Vidal, Miquel, Alvarez, Carlos, Jimenez-Gonzalez, Daniel, Ayguade, Eduard, Valero, Mateo
Format: Tagungsbericht
Sprache:Englisch
Veröffentlicht: IEEE 01.05.2017
Schlagworte:
Discrete cosine transforms
Energy saving
Fine- grained task parallelism
Hardware
Instruction sets
Nested tasks
Parallel processing
Picos++ Hardware Accelerator
Programming
Runtime
Task-based dataflow programming models
ISSN:1530-2075
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Abstract Task-based programming models such as OpenMP, IntelTBB and OmpSs offer the possibility of expressing dependences among tasks to drive their execution at runtime. Managing these dependences introduces noticeable overheads when targeting fine-grained tasks, diminishing the potential speedups or even introducing performance losses. To overcome this drawback, we present a general purpose hardware accelerator, Picos++, to manage the inter-task dependences efficiently in both time and energy. Our design also includes a novel nested task support. To this end, a new hardware/software co-design is presented to overcome the fact that nested tasks with dependences could result in system deadlocks due to the limited amount of resources in hardware task dependence managers. In this paper we describe a detailed implementation of this design and evaluate a parallel task-based programming model using Picos++ in a Linux embedded system with two ARM Cortex-A9 and a FPGA. The scalability and energy consumption of the real system implemented have been studied and compared against a software runtime. Even in a system limited to 2 threads, using Picos++ results in more than 1.8x speedup and 40% of energy savings in the most demanding parallelizations of real benchmarks. As a matter of fact, a hardware task dependence manager should be able to achieve much higher speedup and provide more energy savings with more threads.
AbstractList Task-based programming models such as OpenMP, IntelTBB and OmpSs offer the possibility of expressing dependences among tasks to drive their execution at runtime. Managing these dependences introduces noticeable overheads when targeting fine-grained tasks, diminishing the potential speedups or even introducing performance losses. To overcome this drawback, we present a general purpose hardware accelerator, Picos++, to manage the inter-task dependences efficiently in both time and energy. Our design also includes a novel nested task support. To this end, a new hardware/software co-design is presented to overcome the fact that nested tasks with dependences could result in system deadlocks due to the limited amount of resources in hardware task dependence managers. In this paper we describe a detailed implementation of this design and evaluate a parallel task-based programming model using Picos++ in a Linux embedded system with two ARM Cortex-A9 and a FPGA. The scalability and energy consumption of the real system implemented have been studied and compared against a software runtime. Even in a system limited to 2 threads, using Picos++ results in more than 1.8x speedup and 40% of energy savings in the most demanding parallelizations of real benchmarks. As a matter of fact, a hardware task dependence manager should be able to achieve much higher speedup and provide more energy savings with more threads.
Author Vidal, Miquel
Bosch, Jaume
Xubin Tan
Ayguade, Eduard
Alvarez, Carlos
Jimenez-Gonzalez, Daniel
Valero, Mateo
Author_xml – sequence: 1
  surname: Xubin Tan
  fullname: Xubin Tan
  email: xubin.tan@bsc.es
  organization: Barcelona Supercomput. Center, Univ. Politεcnica de Catalunya, Barcelona, Spain
– sequence: 2
  givenname: Jaume
  surname: Bosch
  fullname: Bosch, Jaume
  email: jbosch@bsc.es
  organization: Barcelona Supercomput. Center, Univ. Politεcnica de Catalunya, Barcelona, Spain
– sequence: 3
  givenname: Miquel
  surname: Vidal
  fullname: Vidal, Miquel
  email: mvidal@bsc.es
  organization: Barcelona Supercomput. Center, Univ. Politεcnica de Catalunya, Barcelona, Spain
– sequence: 4
  givenname: Carlos
  surname: Alvarez
  fullname: Alvarez, Carlos
  email: eduard@bsc.es
  organization: Barcelona Supercomput. Center, Univ. Politεcnica de Catalunya, Barcelona, Spain
– sequence: 5
  givenname: Daniel
  surname: Jimenez-Gonzalez
  fullname: Jimenez-Gonzalez, Daniel
  email: djimenez@ac.upc.edu
  organization: Barcelona Supercomput. Center, Univ. Politεcnica de Catalunya, Barcelona, Spain
– sequence: 6
  givenname: Eduard
  surname: Ayguade
  fullname: Ayguade, Eduard
  organization: Barcelona Supercomput. Center, Univ. Politεcnica de Catalunya, Barcelona, Spain
– sequence: 7
  givenname: Mateo
  surname: Valero
  fullname: Valero, Mateo
  email: mateo.valero@bsc.es
  organization: Barcelona Supercomput. Center, Univ. Politεcnica de Catalunya, Barcelona, Spain
BookMark eNotj8FOAjEURavRRECWrtz0Bwb7Ou20XSookECcRNaSx_QNQWc6pB1D_HtJcHVzk3NPcofsJnSBGHsAMQEQ7mlZzsqPiRRgJspesbEzFnRui9yBKq7Z4FxEJoXRd2yY0pcQUuTKDdjnnAJFbHj5E49dIr7B9J3N6EjBU6iIrzHgnloKPV9g9CeMxOsuXrgXTOT5DHusm-7Ey9jtI7btIez5uvPUpHt2W2OTaPyfI7Z5e91MF9nqfb6cPq-ySoLpM-WNqnIrrawdeKkVkCChKwHF-ciuxt2uqr1V4LSXlSu8Pc9Ik5amACHzEXu8aA9EtD3GQ4vxd2tcYQBU_gciaFRK
CODEN IEEPAD
ContentType Conference Proceeding
DBID 6IE
6IL
CBEJK
RIE
RIL
DOI 10.1109/IPDPS.2017.48
DatabaseName IEEE Electronic Library (IEL) Conference Proceedings
IEEE Proceedings Order Plan All Online (POP All Online) 1998-present by volume
IEEE Xplore All Conference Proceedings
IEEE Xplore
IEEE Proceedings Order Plans (POP All) 1998-Present
DatabaseTitleList
Database_xml – sequence: 1
  dbid: RIE
  name: IEEE/IET Electronic Library (IEL) (UW System Shared)
  url: https://ieeexplore.ieee.org/
  sourceTypes: Publisher
DeliveryMethod fulltext_linktorsrc
Discipline Computer Science
EISBN 9781538639146
1538639149
EISSN 1530-2075
EndPage 253
ExternalDocumentID 7967114
Genre orig-research
GroupedDBID 29O
6IE
6IF
6IH
6IK
6IL
6IN
AAJGR
AAWTH
ABLEC
ADZIZ
ALMA_UNASSIGNED_HOLDINGS
BEFXN
BFFAM
BGNUA
BKEBE
BPEOZ
CBEJK
CHZPO
IEGSK
IPLJI
OCL
RIE
RIL
ID FETCH-LOGICAL-c217t-4d74c38282f91d2541e0e05c016978bfabbcfd84195d2c96d8217e5e52761023
IEDL.DBID RIE
ISICitedReferencesCount 9
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000427044800026&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
IngestDate Wed Aug 27 02:19:53 EDT 2025
IsDoiOpenAccess false
IsOpenAccess true
IsPeerReviewed false
IsScholarly false
Language English
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-c217t-4d74c38282f91d2541e0e05c016978bfabbcfd84195d2c96d8217e5e52761023
OpenAccessLink https://recercat.cat/handle/2072/291863
PageCount 10
ParticipantIDs ieee_primary_7967114
PublicationCentury 2000
PublicationDate 2017-May
PublicationDateYYYYMMDD 2017-05-01
PublicationDate_xml – month: 05
  year: 2017
  text: 2017-May
PublicationDecade 2010
PublicationTitle Proceedings - IEEE International Parallel and Distributed Processing Symposium
PublicationTitleAbbrev IPDPS
PublicationYear 2017
Publisher IEEE
Publisher_xml – name: IEEE
SSID ssj0020349
ssib030101683
Score 1.6920213
Snippet Task-based programming models such as OpenMP, IntelTBB and OmpSs offer the possibility of expressing dependences among tasks to drive their execution at...
SourceID ieee
SourceType Publisher
StartPage 244
SubjectTerms Discrete cosine transforms
Energy saving
Fine- grained task parallelism
Hardware
Instruction sets
Nested tasks
Parallel processing
Picos++ Hardware Accelerator
Programming
Runtime
Task-based dataflow programming models
Title General Purpose Task-Dependence Management Hardware for Task-Based Dataflow Programming Models
URI https://ieeexplore.ieee.org/document/7967114
WOSCitedRecordID wos000427044800026&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/eLvHCXMwlV09T8MwELVKxcBUoEV8ywMjbhM3tuMVSgUSqiLRoROV4w8JAQ1KWvr38Tlpy8DCFkVJHJ2TXJ7v3XsI3UjHWW4jR1SU5CRhwhFJPVjhw8QoqsDuNujMPovJJJ3NZNZCt9teGGttIJ_ZPmyGWr4p9AqWygZCchGDa_WeELzu1do8O8OglZZuv8IUdFd2mpqDp2yUvQCTS_TB6ueXk0pIJOPO_27hEPV2HXk42-aaI9Syi2PU2Vgy4OYN7aLXRkYaZz6ARWXxVFXvZNQ43fqL7OguGIr2a1Va7P9b6-PufEozeKTA0rdYw4DA3fr0I2LwTPuoemg6fpjeP5LGQoFojzWWJDEi0UOPqqiTsfFgMLaRjZgGDRaR5k7luXYmTWLJDNWSm9SfZpllVHAQdThB7UWxsKcIO-FopJw21Cc0wzwwzCkzMqWhspfwM9SFaM2_apGMeROo8793X6ADmIuaOXiJ2styZa_Qvv5evlXldZjZH-gko24
linkProvider IEEE
linkToHtml http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV3PT8IwFG4ImugJFYy_7cGjg62063pVJBCRLHEHTpKuPxIjMsNA_n37tgEevHhblm1dXre9fX3f-z6E7oQNWWp860mfph5l3HqCOLASdqmWRILdbaEzO-LjcTSZiLiG7re9MMaYgnxm2rBZ1PJ1plawVNbhIuQBuFbvMUqJX3ZrbZ6ebqGWFm2_wwSUV3aqmp1h3ItfgcvF22D288tLpUgl_cb_buIItXY9eTjeZptjVDPzE9TYmDLg6h1tordKSBrHLoRZbnAi8w-vV3nduovsCC8YyvZruTDY_bmWxz24pKZxT4Kpb7aGAYG99elGxOCaNstbKOk_JY8DrzJR8JRDG0uPak5V1-EqYkWgHRwMjG98pkCFhUeplWmqrI5oIJgmSoQ6cqcZZhjhIcg6nKL6PJubM4Qtt8SXVmniUppmDhqmhGkRkaK2R8Nz1IRoTb9KmYxpFaiLv3ffooNB8jKajobj50t0CPNS8givUH25WJlrtK--l-_54qaY5R-Xhaa1
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=Proceedings+-+IEEE+International+Parallel+and+Distributed+Processing+Symposium&rft.atitle=General+Purpose+Task-Dependence+Management+Hardware+for+Task-Based+Dataflow+Programming+Models&rft.au=Xubin+Tan&rft.au=Bosch%2C+Jaume&rft.au=Vidal%2C+Miquel&rft.au=Alvarez%2C+Carlos&rft.date=2017-05-01&rft.pub=IEEE&rft.eissn=1530-2075&rft.spage=244&rft.epage=253&rft_id=info:doi/10.1109%2FIPDPS.2017.48&rft.externalDocID=7967114