Unintrusive aging analysis based on offline learning
Carregando...
Data
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Federal de Minas Gerais
Descrição
Tipo
Artigo de evento
Título alternativo
Primeiro orientador
Membros da banca
Resumo
Runtime aging analysis of integrated circuits enables adaptive approaches in order to enhance the system's life time and permits the user to be aware of critical states. Common approaches utilize sensors that are integrated invasively into critical paths or report experienced aging. This work presents a lightweight supportive technique that correlates environmental and internal conditions with learned data in order to predict the actual wear-out of the system. Simulation results indicate the feasibility of the approach with prediction errors below 10%.
Abstract
Assunto
Aprendizado do computador, Máquinas elétricas
Palavras-chave
Aging , Temperature sensors , Stress , Temperature measurement , Monitoring , Integrated circuit modeling, Machine Learning, Reliability , Remaining Useful Lifetime , NBTI , TDDB, Offline Learning , Prediction Error , Critical Conditions , Critical Path , System Lifetime , General Linear Model , Inverter , Strategies In Order , Field Of Systems , Supply Voltage , Stress Sensor , Frequency Scale , Active Switches , Age Profile , Hot Electrons , Voltage Stress , Mean Time To Failure , Voltage Scaling , Remaining Useful Life , Principal Idea , Supply Temperature
Citação
Departamento
Curso
Endereço externo
https://ieeexplore.ieee.org/document/8244453