umu.sePublications
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
CONGESTION-CONTROLLED AUTOTUNING OF OPENMP PROGRAMS
Umeå University, Faculty of Science and Technology, Department of Computing Science.
2019 (English)Independent thesis Basic level (degree of Bachelor), 10 credits / 15 HE creditsStudent thesis
Abstract [en]

Parallelisation is becoming more and more important as the single core performance increase is stagnating while the amount of cores is increasing with every new generation of hardware. ŒThe traditional approach of manual parallelisation has an alternative in parallel frameworks, such as OpenMP, which can simplify the creation of parallel code. Optimising this code can, however, be cumbersome and difficult. Automating the optimisation or tuning of parallel code and computations is a very interesting alternative to manually optimising algorithms and programs. Previous work has shown that intricate systems can eff‚ectively autotune parallel programs with potentially the same eff‚ectiveness as human experts. ThŒis study suggests using an approach with the main algorithm used inspired from the congestion control algorithms from computer networks, namely AIMD. By applying the algorithm on top of an OpenMP program the parallel parameters such as grain size can be controlled. TheŒe simpli€ed algorithm is shown to be able to achieve a 19% speedup compared to a naive static parallel implementation

Place, publisher, year, edition, pages
2019. , p. 22
Series
UMNAD ; 1183
National Category
Engineering and Technology
Identifiers
URN: urn:nbn:se:umu:diva-163181OAI: oai:DiVA.org:umu-163181DiVA, id: diva2:1349831
Educational program
Bachelor of Science Programme in Computing Science
Supervisors
Examiners
Available from: 2019-09-10 Created: 2019-09-10 Last updated: 2019-09-10Bibliographically approved

Open Access in DiVA

fulltext(388 kB)4 downloads
File information
File name FULLTEXT01.pdfFile size 388 kBChecksum SHA-512
ecb574121aa88046bd8de334b68e5e8c6db1371223499109ec74e9ea5104e9dd0d61f889d9427c6d4e8ec383e81dfd3b8f879b887759f359f2afce0e4e96e045
Type fulltextMimetype application/pdf

By organisation
Department of Computing Science
Engineering and Technology

Search outside of DiVA

GoogleGoogle Scholar
Total: 4 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

urn-nbn

Altmetric score

urn-nbn
Total: 14 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf