Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • harvard1
  • 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
Speeding Up Value at Risk Calculations Using Accelerators
KTH, School of Information and Communication Technology (ICT).
2014 (English)Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
Abstract [en]

Calculating Value at Risk (VaR) can be a time consuming task. Therefore it is of interest to find a way to parallelize this calculation to increase performance. Based on a system built in Java, which hardware is best suited for these calculations?

This thesis aims to find which kind of processing unit that gives optimal performance when calculating scenario based VaR. First the differences of the CPU, GPU and coprocessor is examined in a theoretical study. Then multiple versions of a parallel VaR algorithm are implemented for a CPU, a GPU and a coprocessor trying to make use of the findings from the theoretical study.

The performance and ease of programming for each version is evaluated and analyzed. By running performance tests it is found that the CPU was the winner when coming to performance while running the chosen VaR algorithm and problem sizes.

Abstract [sv]

Att beräkna Value at Risk (VaR) kan vara tidskrävande. Därför är det instressant att finna möjligheter att parallelisera och snabba upp dessa beräkningar för att förbättra prestandan. Men vilken hårdvara är bäst lämpad för dessa beräkningar?

Detta arbete syftar till att för ett system skrivet i Java hitta vilken typ av beräkningsenhet som ger optimal prestanda vid scenariobaserade VaR beräkningar. Först gjordes en teoretisk undersökning av CPUn, GPUn och en coprocessor. Flera versioner av en parallel VaR algoritm implementeras för en CPU, GPU och en coprocessor där resultaten från undersökningen utnyttjas.

Prestandan samt enkelheten att programmera varje version utvärderas och analyseras. De utförda prestanda testerna visar att vinnaren vad gäller prestanda är CPUn för den valda VaR algoritmen och de testade problemstorlekarna.

Place, publisher, year, edition, pages
2014. , 90 p.
Series
TRITA-ICT-EX, 2014:154
National Category
Computer and Information Science
Identifiers
URN: urn:nbn:se:kth:diva-177384OAI: oai:DiVA.org:kth-177384DiVA: diva2:872595
Examiners
Available from: 2015-11-19 Created: 2015-11-19 Last updated: 2017-08-03Bibliographically approved

Open Access in DiVA

fulltext(3385 kB)10 downloads
File information
File name FULLTEXT01.pdfFile size 3385 kBChecksum SHA-512
c56f1bda8468622656369f40739966e5c57b44fb0bd13d97f7d26fb3fd7bbd740a1001e50061e62b8430b0df3e7ebedff8c388384e2e0606ad0d7525c9cfb3a3
Type fulltextMimetype application/pdf

By organisation
School of Information and Communication Technology (ICT)
Computer and Information Science

Search outside of DiVA

GoogleGoogle Scholar
Total: 10 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: 161 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • harvard1
  • 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