Programsko ostvarenje masivno paralelnog algoritma PSO

undergraduate thesis

• Title (english): Massively parallel PSO implementation
• Author: Joško Vrsalović
• Mentor: Domagoj Jakobović (mentor)
• Committee member: Domagoj Jakobović (predsjednik povjerenstva)
• Committee member: Josip Jukić (član povjerenstva)
• Committee member: Filip Karlo Došilović (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2023-07-05, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Svrha rada bila je istražiti i razviti poznati, optimizacijski algoritam PSO globalne topologije
za primjenu na heterogenom računalnom sklopovlju. Cilj je bio iskoristiti visoki stupanj
paralelizacije koje nudi današnja oprema za grafičku obradu podataka.
Istražene su prednosti i ograničenja algoritma, a za GPU implementaciju proučena je potreba
za zaštitom kritičnog odsječka pod pretpostavkom da bi ona mogla narušiti brzinu izvođenja.
Dani su uvid u moguće primjene i detaljan primjer korištenja za prepoznavanje objekta na
slici. Konačno, predstavljeni su rezultati usporedbe između CPU i GPU implementacija u
kategorijama brzine i sposobnosti konvergencije te brzine izvođenja u ovisnosti o
parametrima algoritma.

Abstract (english)

The purpose of this project was to explore and develop widely known, optimization
algorithm PSO using global topology for applications on heterogeneous computing
hardware. Goal was to utilize high degree of parallelization offered by today’s graphic
processing equipment.
Advantages and limitations of algorithms were explored, and with GPU implementations,
the need for critical section protection was studied under hypothesis that mechanism could
slow down the execution.
Overview of possible applications and detailed example of object recognition within image
is presented. Finally, results of convergence and performance tests were measured between
CPU and GPU with respect to algorithm parameters.

• Keywords: Optimizacija rojem čestica; Paralelno programiranje; CUDA; Traženje objekta na slici; Izračunavanje s visokim performansama
• Keywords (english): Particle swarm optimization; Parallel programming; CUDA; Image object search; High performance computing
• Language: croatian
• URN:NBN: urn:nbn:hr:168:545986
• Study programme: Title: Computing; Study programme type: university; Study level: undergraduate; Academic / professional title: sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva (sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-10-11 10:57:29