ORCA RT-TOOLS : a suite of tools to support real-time analysis on manycore systems

• Ano de defesa: 2024
• Autor(a) principal: Domingues, Anderson Roberto Pinheiro
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Pontifícia Universidade Católica do Rio Grande do Sul; Escola Politécnica; Brasil; PUCRS; Programa de Pós-Graduação em Ciência da Computação
• Palavras-chave em Português: Network-on-chip; Real-time Systems; Manycores; Redes Intrachip; Sistemas de Tempo-real; CIENCIA DA COMPUTACAO::TEORIA DA COMPUTACAO
• Link de acesso: https://tede2.pucrs.br/tede2/handle/tede/11636

The COVID-19 pandemic allowed plenty of companies to enter the market offering solutions to help fight the SARS-CoV-2 virus, including robots for UV-C disinfection and autonomous vehicles. With the demand for computational resources growing in the criticalmission systems domain, the community began exploring resources from manycore platforms, which have a potential for parallelism compared to the single- and multi-core platforms currently used. However, critical-mission systems need support for several non-functional requirements for communication and computation, including real-time (RT). Over the years, different research communities have addressed RT guarantees for communication and computation. On the one hand, research on RT computing has advanced significantly for singleand multi-core platforms. On the other hand, RT communication research developed the socalled real-time intra-chip networks (RT-NoCs). As a result, the literature lacks an approach that tackles both perspectives of a system, mostly neglecting the joint analysis of computing and communication operations synchronization. This Thesis presents an approach to RT analysis in NoC-based manycores, which tackles computation and communication jointly. We used our approach to guarantee the RT requirements for computation and communication on a system without an RT-NoC, using a low-cost, open-source NoC. We validate our approach for synthetic and benchmark applications in an RTL simulation environment.