Roughness of Microarchitectural Design Topologies and Its Implications for Optimization

Recent advances in statistical inference and machine learning close the divide between simulation and classical optimization, thereby enabling more rigorous and robust micro-architectural studies. To most effectively utilize these now computationally tractable techniques, the authors characterize design topology roughness and leverage this characterization to guide their usage of analysis and optimization methods. In particular, they compute roughness metrics that require high-order derivatives and multi-dimensional integrals of design metrics, such as performance and power.