Cache-Partitioning for COTS Multi-coreArchitecture

Multi-core architectures present challenges to execute real-time applications. Concurrently executing real-time tasks on different cores, can produce negative impact on each others execution times due to the shared resources e.g. shared last-level cache, shared memory bus etc. Shared last-level cach...

Full description

Bibliographic Details
Main Author: Konstantopoulos, Konstantinos
Format: Others
Language:English
Published: Mälardalens högskola, Inbyggda system 2017
Online Access:http://urn.kb.se/resolve?urn=urn:nbn:se:mdh:diva-34844
id ndltd-UPSALLA1-oai-DiVA.org-mdh-34844
record_format oai_dc
spelling ndltd-UPSALLA1-oai-DiVA.org-mdh-348442017-02-18T05:28:57ZCache-Partitioning for COTS Multi-coreArchitectureengKonstantopoulos, KonstantinosMälardalens högskola, Inbyggda system2017Multi-core architectures present challenges to execute real-time applications. Concurrently executing real-time tasks on different cores, can produce negative impact on each others execution times due to the shared resources e.g. shared last-level cache, shared memory bus etc. Shared last-level cache is a resource of cache pollution that negatively impacts the execution times of tasks. In this thesis project we provided support for last-level cache partitioning through the mechanism of page coloring. Using page coloring, the last level cache is divided in multiple partitions. Tasks are allowed to access their own partition only, thus achieving cache isolation. The implementation is done in Linux OS. The aim of the thesis is to incorporate this work with the previous implementation of Multi-Resource Server to achieve task isolation. Student thesisinfo:eu-repo/semantics/bachelorThesistexthttp://urn.kb.se/resolve?urn=urn:nbn:se:mdh:diva-34844application/pdfinfo:eu-repo/semantics/openAccess
collection NDLTD
language English
format Others
sources NDLTD
description Multi-core architectures present challenges to execute real-time applications. Concurrently executing real-time tasks on different cores, can produce negative impact on each others execution times due to the shared resources e.g. shared last-level cache, shared memory bus etc. Shared last-level cache is a resource of cache pollution that negatively impacts the execution times of tasks. In this thesis project we provided support for last-level cache partitioning through the mechanism of page coloring. Using page coloring, the last level cache is divided in multiple partitions. Tasks are allowed to access their own partition only, thus achieving cache isolation. The implementation is done in Linux OS. The aim of the thesis is to incorporate this work with the previous implementation of Multi-Resource Server to achieve task isolation.
author Konstantopoulos, Konstantinos
spellingShingle Konstantopoulos, Konstantinos
Cache-Partitioning for COTS Multi-coreArchitecture
author_facet Konstantopoulos, Konstantinos
author_sort Konstantopoulos, Konstantinos
title Cache-Partitioning for COTS Multi-coreArchitecture
title_short Cache-Partitioning for COTS Multi-coreArchitecture
title_full Cache-Partitioning for COTS Multi-coreArchitecture
title_fullStr Cache-Partitioning for COTS Multi-coreArchitecture
title_full_unstemmed Cache-Partitioning for COTS Multi-coreArchitecture
title_sort cache-partitioning for cots multi-corearchitecture
publisher Mälardalens högskola, Inbyggda system
publishDate 2017
url http://urn.kb.se/resolve?urn=urn:nbn:se:mdh:diva-34844
work_keys_str_mv AT konstantopouloskonstantinos cachepartitioningforcotsmulticorearchitecture
_version_ 1718415229002448896