Prevođenje tokovnog programa za heterogenu ugradbenu platformu

master's thesis

• Title (english): Stream Program Compilation for Heterogeneous Embedded Platform
• Author: Ivan Pilat
• Mentor: Josip Knezović (mentor)
• Committee member: Josip Knezović (predsjednik povjerenstva)
• Committee member: Danko Basch (član povjerenstva)
• Committee member: Hrvoje Mlinarić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2014-07-09, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Ovaj rad iznosi pregled paradigme toka podataka pogodne za eksplicitno iskazivanje paralelizma, kao i detaljnu analizu StreamIt-a, jezika zasnovanog na toku podataka. Heterogene računalne arhitekture su prepoznate kao brze i energetski efikasne, ali i da iziskuju zahtjevan postupak razvoja koji mora raspodijeliti izračun na različite arhitekture nejednakih mogućnosti.
Glavni doprinos ovog rada je implementacija postupka prevođenja StreamIt-a za heterogenu platformu koja se sastoji od višejezgrenog procesora i rekonfigurabilne logike. Implementacija je ispitana na Xilinx Zynq uređaju; izvedena je usporedba performansi i energetske efikasnosti pokretanjem ispitnih programa na Zynq ARM procesorskom sustavu, Zynq programabilnoj logici sa sučeljem prema procesorskom sustavu, te na Intel x86-64 sustavu.

Abstract (english)

This thesis gives an overview of the streaming dataflow paradigm capable of explicitly exposing parallelism, and a detailed analysis of StreamIt, a language based on dataflow. Heterogeneous computing architectures are recognized for their performance and energy efficiency, at the cost of a difficult development process requiring mapping computation to different architectures with varying capabilities.
The main contribution of this work is an implementation of StreamIt compilation flow for a heterogeneous platform consisting of a multicore CPU and reconfigurable logic. The implementation was tested on a Xilinx Zynq device; performance and energy efficiency was compared running benchmarks on the Zynq ARM processing system, Zynq programmable logic interfacing with the processing system, and an Intel x86-64 system.

• Keywords: heterogeno računarstvo; ugradbeni sustavi; FPGA; paralelizacija; integracija programskih paradigmi; model toka podataka
• Keywords (english): heterogeneous computing; embedded systems; FPGA; parallelization; programming paradigm integration; streaming dataflow model
• Language: english
• URN:NBN: urn:nbn:hr:168:689611
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2018-02-06 11:17:57