Change search
ReferencesLink to record
Permanent link

Direct link
Hybrid Acceleration of a Molecular DynamicsSimulation Using Short-Ranged Potentials
KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Numerical Analysis, NA.
2013 (English)Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesisAlternative title
Hybrid acceleration av en molekylärdynamisksimulation för potentialer med kort räckvidd (Swedish)
Abstract [en]

Molecular dynamics simulations are a very usefultool to study the behavior and interaction of atoms and molecules in chemicaland bio-molecular systems. With the fast rising complexity of such simulationshybrid systems with both, multi-core processors (CPUs) and multiple graphics processingunits (GPUs), become more and more popular. To obtain an optimal performance thisthesis presents and evaluates two different hybrid algorithms, employing allavailable compute capacity from CPUs and GPUs. The presented algorithms can beapplied for short-range force calculations in arbitrary molecular dynamicssimulations

Abstract [sv]

Molekylärdynamiska simulationer är ett mycket lämpligt verktyg för att studera beteendet och växelspelet av atomer och molekyler inom kemiska och biomolekylära system. Hybrida datorsystem som innehåller såväl processorer med multipla kärnor (CPU:er) som flera grafikprocessorer (GPU:er) blir populära i växande utsträckning. Vi undersöker två olika hybrida algoritmer som använder den hela beräkningsförmågan av CPU:er och GPU:er för att få ut den optimala beräkningsprestandan. De presenterade algoritmerna kan användas i godtyckliga molekylärdynamiska simulationer med potentialer av kort räckvidd.

Place, publisher, year, edition, pages
2013. , 96 p.
TRITA-MAT-E, 2013:36
National Category
Computational Mathematics
URN: urn:nbn:se:kth:diva-124239OAI: diva2:633887
Subject / course
Numerical Analysis
Educational program
Master of Science - Scientific Computing
Physics, Chemistry, Mathematics
Available from: 2013-06-27 Created: 2013-06-27 Last updated: 2014-02-26Bibliographically approved

Open Access in DiVA

fulltext(1240 kB)335 downloads
File information
File name FULLTEXT01.pdfFile size 1240 kBChecksum SHA-512
Type fulltextMimetype application/pdf

By organisation
Numerical Analysis, NA
Computational Mathematics

Search outside of DiVA

GoogleGoogle Scholar
Total: 335 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

Total: 532 hits
ReferencesLink to record
Permanent link

Direct link