Dissertation/ Thesis
Προβολή σε EDS

Parallel sparse matrix vector multiplication techniques for shared memory architectures

Λεπτομέρειες βιβλιογραφικής εγγραφής
Τίτλος: Parallel sparse matrix vector multiplication techniques for shared memory architectures
Συγγραφείς: Başaran, Mehmet
Έτος έκδοσης: 2014
Συλλογή: Bilkent University: Institutional Repository
Θεματικοί όροι: SpMxV, Parallelization, KNC, Intel Xeon Phi, Many-core, GPU, Vectorization, SIMD, Adaptive scheduling and load balancing, Work stealing, Distributed Systems, Data Locality, QA188 .B37 2014, Sparse matrices, Sparse matrices--Data processing
Περιγραφή: Cataloged from PDF version of article. ; Includes bibliographical references leaves 97-99. ; SpMxV (Sparse matrix vector multiplication) is a kernel operation in linear solvers in which a sparse matrix is multiplied with a dense vector repeatedly. Due to random memory access patterns exhibited by SpMxV operation, hardware components such as prefetchers, CPU caches, and built in SIMD units are under-utilized. Consequently, limiting parallelization efficieny. In this study we developed; • an adaptive runtime scheduling and load balancing algorithms for shared memory systems, • a hybrid storage format to help effectively vectorize sub-matrices, • an algorithm to extract proposed hybrid sub-matrix storage format. Implemented techniques are designed to be used by both hypergraph partitioning powered and spontaneous SpMxV operations. Tests are carried out on Knights Corner (KNC) coprocessor which is an x86 based many-core architecture employing NoC (network on chip) communication subsystem. However, proposed techniques can also be implemented for GPUs (graphical processing units). ; Başaran, Mehmet
Τύπος εγγράφου: thesis
Περιγραφή αρχείου: xvi, 99 leaves, graphics; application/pdf
Γλώσσα: English
Relation: https://hdl.handle.net/11693/16868; B148387
Διαθεσιμότητα: https://hdl.handle.net/11693/16868
Rights: info:eu-repo/semantics/openAccess
Αριθμός Καταχώρησης: edsbas.BDFFFD41
Βάση Δεδομένων: BASE
FullText Text:
  Availability: 0
CustomLinks:
  – Url: https://hdl.handle.net/11693/16868#
    Name: EDS - BASE (ns324271)
    Category: fullText
    Text: View record from BASE
Header DbId: edsbas
DbLabel: BASE
An: edsbas.BDFFFD41
RelevancyScore: 769
AccessLevel: 3
PubType: Dissertation/ Thesis
PubTypeId: dissertation
PreciseRelevancyScore: 768.644287109375
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Parallel sparse matrix vector multiplication techniques for shared memory architectures
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Başaran%2C+Mehmet%22">Başaran, Mehmet</searchLink>
– Name: DatePubCY
  Label: Publication Year
  Group: Date
  Data: 2014
– Name: Subset
  Label: Collection
  Group: HoldingsInfo
  Data: Bilkent University: Institutional Repository
– Name: Subject
  Label: Subject Terms
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22SpMxV%22">SpMxV</searchLink><br /><searchLink fieldCode="DE" term="%22Parallelization%22">Parallelization</searchLink><br /><searchLink fieldCode="DE" term="%22KNC%22">KNC</searchLink><br /><searchLink fieldCode="DE" term="%22Intel+Xeon+Phi%22">Intel Xeon Phi</searchLink><br /><searchLink fieldCode="DE" term="%22Many-core%22">Many-core</searchLink><br /><searchLink fieldCode="DE" term="%22GPU%22">GPU</searchLink><br /><searchLink fieldCode="DE" term="%22Vectorization%22">Vectorization</searchLink><br /><searchLink fieldCode="DE" term="%22SIMD%22">SIMD</searchLink><br /><searchLink fieldCode="DE" term="%22Adaptive+scheduling+and+load+balancing%22">Adaptive scheduling and load balancing</searchLink><br /><searchLink fieldCode="DE" term="%22Work+stealing%22">Work stealing</searchLink><br /><searchLink fieldCode="DE" term="%22Distributed+Systems%22">Distributed Systems</searchLink><br /><searchLink fieldCode="DE" term="%22Data+Locality%22">Data Locality</searchLink><br /><searchLink fieldCode="DE" term="%22QA188+%2EB37+2014%22">QA188 .B37 2014</searchLink><br /><searchLink fieldCode="DE" term="%22Sparse+matrices%22">Sparse matrices</searchLink><br /><searchLink fieldCode="DE" term="%22Sparse+matrices--Data+processing%22">Sparse matrices--Data processing</searchLink>
– Name: Abstract
  Label: Description
  Group: Ab
  Data: Cataloged from PDF version of article. ; Includes bibliographical references leaves 97-99. ; SpMxV (Sparse matrix vector multiplication) is a kernel operation in linear solvers in which a sparse matrix is multiplied with a dense vector repeatedly. Due to random memory access patterns exhibited by SpMxV operation, hardware components such as prefetchers, CPU caches, and built in SIMD units are under-utilized. Consequently, limiting parallelization efficieny. In this study we developed; • an adaptive runtime scheduling and load balancing algorithms for shared memory systems, • a hybrid storage format to help effectively vectorize sub-matrices, • an algorithm to extract proposed hybrid sub-matrix storage format. Implemented techniques are designed to be used by both hypergraph partitioning powered and spontaneous SpMxV operations. Tests are carried out on Knights Corner (KNC) coprocessor which is an x86 based many-core architecture employing NoC (network on chip) communication subsystem. However, proposed techniques can also be implemented for GPUs (graphical processing units). ; Başaran, Mehmet
– Name: TypeDocument
  Label: Document Type
  Group: TypDoc
  Data: thesis
– Name: Format
  Label: File Description
  Group: SrcInfo
  Data: xvi, 99 leaves, graphics; application/pdf
– Name: Language
  Label: Language
  Group: Lang
  Data: English
– Name: NoteTitleSource
  Label: Relation
  Group: SrcInfo
  Data: https://hdl.handle.net/11693/16868; B148387
– Name: URL
  Label: Availability
  Group: URL
  Data: https://hdl.handle.net/11693/16868
– Name: Copyright
  Label: Rights
  Group: Cpyrght
  Data: info:eu-repo/semantics/openAccess
– Name: AN
  Label: Accession Number
  Group: ID
  Data: edsbas.BDFFFD41
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=edsbas&AN=edsbas.BDFFFD41
RecordInfo BibRecord:
  BibEntity:
    Languages:
      – Text: English
    Subjects:
      – SubjectFull: SpMxV
        Type: general
      – SubjectFull: Parallelization
        Type: general
      – SubjectFull: KNC
        Type: general
      – SubjectFull: Intel Xeon Phi
        Type: general
      – SubjectFull: Many-core
        Type: general
      – SubjectFull: GPU
        Type: general
      – SubjectFull: Vectorization
        Type: general
      – SubjectFull: SIMD
        Type: general
      – SubjectFull: Adaptive scheduling and load balancing
        Type: general
      – SubjectFull: Work stealing
        Type: general
      – SubjectFull: Distributed Systems
        Type: general
      – SubjectFull: Data Locality
        Type: general
      – SubjectFull: QA188 .B37 2014
        Type: general
      – SubjectFull: Sparse matrices
        Type: general
      – SubjectFull: Sparse matrices--Data processing
        Type: general
    Titles:
      – TitleFull: Parallel sparse matrix vector multiplication techniques for shared memory architectures
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Başaran, Mehmet
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 01
              Type: published
              Y: 2014
          Identifiers:
            – Type: issn-locals
              Value: edsbas
            – Type: issn-locals
              Value: edsbas.oa
ResultId 1