This report aims to comprehensively examine metrics for evaluating the NEUROPULS Horizon Europe project (Grant Agreement n° 101070238) accelerator and their simulation counterparts and the security of the phase-change material PUFs provide. The report provides an in-depth analysis of the key performance indicators, methodologies, and tools utilized in assessing the efficiency and efficacy of the revolutionary computing platforms proposed in the project. By defining a standardized set of metrics and evaluation practices, this document aims to foster cross-comparisons, facilitate advancements, and guide researchers, developers, and industry stakeholders in harnessing the true capabilities of neuromorphic computing.
This deliverable explores the challenges of benchmarking photonic chips, considering speed, power efficiency, and scalability factors. Additionally, we emphasize the importance of establishing a robust framework for comparing simulation results with physical implementations, enabling researchers and engineers to gain meaningful insights into the capabilities and limitations of these cutting-edge technologies.
Through thoroughly examining established evaluation metrics and emerging standards, this document aims to provide a comprehensive guide for researchers, developers, and industry professionals engaged in evaluating and benchmarking photonic chips and their simulation models. By addressing the unique challenges and opportunities in this field, we strive to contribute to the ongoing dialogue surrounding the advancement of photonic computing and foster a standardized approach to benchmarking that ensures meaningful and reliable assessments.