Malmö University Publications
Change search
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
Performance impacts of profiling multi-threaded applications with flexible analysis tools
Malmö högskola, Faculty of Technology and Society (TS).
2014 (English)Independent thesis Basic level (degree of Bachelor), 10 credits / 15 HE creditsStudent thesis
Abstract [sv]

Syftet med denna studie var att undersöka hur användningen av profilers påverkade prestandan hos flertrådade applikationer. Studien genomfördes inom ramen för ett projekt med Edument AB som skapat en av de profilers som undersökts i studien. Mer specifikt så syftade studien till att ta reda på vilken ytterligare CPU-tid och RAM som användes och hur mycket längre tid det tog att exekvera en flertrådad applikation vid användningen av profilers. Uppsatsens hypotes var att valet av data som varje profiler registrerade skulle påverka prestanda, och att de som registrerade fler detaljer om applikationen skulle ha störst påverkan. Fem profilers valdes ut med olika egenskaper och funktioner och en testapplikation skrevs för att simulera en flertrådad applikation. En minimalistisk applikation skrevs också och användes för att registrera varje profilers påverkan på RAM och CPU, samt hur testapplikationen påverkades av profilern. Alltså, varje profiler har var för sig blivit profilerad för att försäkra att samma data konsekvent samlats in. Resultaten visade att valet av profiler kan ha stor påverkan på den profilerade applikationens prestanda. Användningen av en av profilerna ledde till att testapplikationen tog 513% så lång tid att exekvera och lade också till 1400% ytterligare RAM-användning. Efter en analys av insamlade data verkade det finnas ett samband mellan de funktioner som varje profiler erbjöd och påverkan på applikationens prestanda vilket stämmer överens med uppsatsens hypotes.

Abstract [en]

The aim of this study was to investigate how the use of profilers impacted the performance of multi-threaded applications. This study was undertaken in the framework of a project for Edument AB, creators of one of the profilers investigated in this study. Specifically, the study aimed to find out what additional overhead can result in profiling multi-threaded applications, with regards to CPU use, memory use, and additional time taken to complete execution. The paper hypothesised that the selection of data that each profiler recorded would impact the performance, with those that recorded a very large amount of detail about the application being profiled would introduce a higher overhead. A selection of five profilers was made, each profiler offering a range of features and functionality, and a test application was written to simulate a multi-threaded application. A very lightweight application, referred to as the monitoring application, was also written which recorded the overhead each profiler used, along with the overhead of the test application and the time taken to run. Essentially, each profiler was itself profiled by the monitoring application, ensuring consistent overhead data was gathered for each profiler. The results of the study showed that the choice of profiler can have a substantial impact on the performance of the application being profiled. One profiler resulted in execution of the test application taking 513% as much time to run, and adding an overhead of 1400% memory use. After analysing this data, there appeared to be a correlation between what features the profilers recorded and the overhead, consistent with the hypothesis put forward at the beginning of the study.

Place, publisher, year, edition, pages
Malmö högskola/Teknik och samhälle , 2014. , p. 34
National Category
Engineering and Technology
Identifiers
URN: urn:nbn:se:mau:diva-20723Local ID: 17390OAI: oai:DiVA.org:mau-20723DiVA, id: diva2:1480603
Educational program
TS Data- och telekom - Teknik, ekonomi och ledarskap, TELMah
Available from: 2020-10-27 Created: 2020-10-27 Last updated: 2022-06-27Bibliographically approved

Open Access in DiVA

fulltext(291 kB)366 downloads
File information
File name FULLTEXT01.pdfFile size 291 kBChecksum SHA-512
01afae1ff6005ab7b0de284d9f59257d09cfa1a88ffa6439bc0348369a5cccd0dd42dc9e1c9a4f81f3773726f464e9fc3218dc5b256929a1a0602f9fb37a0d9b
Type fulltextMimetype application/pdf

By organisation
Faculty of Technology and Society (TS)
Engineering and Technology

Search outside of DiVA

GoogleGoogle Scholar
Total: 366 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: 85 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