Learning-based BTI stress estimation and mitigation in multi-core processor systems

Abstract With the increasing demand of designing a reliable processing devices, issue CMOS ageing is jeopardising industry digital devices. Many studies has been cover this area for modelling behaviour at device level or developing sensors on-line delay detection system level. However, we are presenting method to estimate stresses (e.g. Temperature, Ageing Stress Activity) rather than (performance degradation) itself. The purpose estimating stress optimise utilisation with minimisation stress. In multicore processors, existence more one source higher single core processor but optimisation space as well along temperature and power optimisation. paper, have modelled from application using machine learning techniques train data extracted high workloads ( e.g. parsec splash2 benchmarks) on four cores Xeon. model able 0.1% error proactively reduce by 50%.