Integracija akceleratora za duboko učenje u heterogeni računalni sustav zasnovan na procesoru arhitekture RISC-V

master's thesis

• Title (english): Deep Learning Accelerator Integration in Heterogeneous Computing RISC-V System
• Author: Dorian Brmbota
• Mentor: Josip Knezović (mentor)
• Committee member: Josip Knezović (predsjednik povjerenstva)
• Committee member: Hrvoje Mlinarić (član povjerenstva)
• Committee member: Daniel Hofman (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2020-07-17, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Veliki dio IT industrije je posvećen istraživanju algoritama dubokog učenja. Računalno su složeni i zahtjevaju veliku količinu memorijskih resursa. Dok se treniranje dubokih neurnoskih mreža izvodi na snažnim radnim stanicama, samo zaključivanje modela se često vrši na ugradbenim, energetski učinkovitim uređajima. NVIDIA Deep Learning Accelerator(NVDLA) je podesiv akceleratora otvorenog koda, idealan za integraciju u ugradbena, energetski učinkvita, system-on-chip(SoC) računala. Ovaj rad se bavi integracijom NVDLA u PULPissimo - sustav baziran na RISC-V.

Abstract (english)

Great part of IT industry is dedicated to researching Deep Learning algorithms. They are computationally intensive and requires large amount of memory. While training of Deep Neural Networks can be done on powerful workstations, the inference often needs to be done on some embedded, low-power devices. NVIDIA Deep Learning Accelerator(NVDLA) is an open source, low power, configurable accelerator that can be integrated into embedded system-on-chip(SoC) devices. This thesis describes integration of NVDLA into PULPissimo - a RISC-V architecture based system.

• Keywords: RISC-V; PULP; NVDLA; SoC; duboko učenje; ASIC
• Keywords (english): RISC-V; PULP; NVDLA; SoC; Hardware Design; Deep Learning; ASIC
• Language: english
• URN:NBN: urn:nbn:hr:168:952803
• 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: 2023-01-13 11:53:58