Machine learning opportunities and applications in SoC design
Bauli Yan · 2018
As the complexity of integrated circuits (IC) grows significantly with process evolution, design integration and new emphasis in safety applications, the controllability of design flow iterations is getting critical, particularly for collaborations that span technology domains in global sites. It would greatly help risk management if we have a precise and comprehensive prediction process built upon a machine learning framework with historical data. In this talk, we identify some typical challenges in IC design and demonstrate the corresponding practices concerning power, performance, test, and project schedule optimization.