FuriosaAI Case Study on Advanced In-Chip Monitoring

This case study details how FuriosaAI, a semiconductor company, enhanced its AI processor performance through advanced in-chip monitoring using proteanTecs' analytics platform. The document outlines the integration of on-chip agents that provided parametric visibility into production, which helped Furiosa maintain its competitive edge in the semiconductor market. It describes two key use cases: timing margin characterization and voltage fluctuations monitoring. The first use case highlights how proteanTecs’ Margin Agents and machine learning algorithms offered insights into performance variations across different chip sections, allowing Furiosa to ensure optimal performance per watt (PPW). The second use case focuses on the identification of voltage fluctuations and thermal conditions through integrated sensors, which contributed to stable performance under varying workloads. The study emphasizes the importance of deep data analytics in achieving superior visibility during the New Product Introduction (NPI) process, ultimately leading to accelerated mass production readiness.