Performance modeling of heterogeneous HW platforms

The push towards automated and connected driving functionalities mandates the use of heterogeneous hardware platforms in order to provide required computational resources. For these platforms, established methods for performance modeling industry are no longer effective or adequate. In this paper, we explore detailed problem mapping a prototypical autonomous application on Nvidia Tegra X2 platform while considering different constraints application, including end-to-end latencies event chains spanning CPU GPU boundaries. With given use-case platform, propose concepts Amalthea, capturing architectural aspects also execution structure application. These models can be fed into appropriate tools predict properties. We proposed above Workshop Analysis Tools Methodologies Embedded Real-time Systems (WATERS) Industrial Challenge 2019 response, academicians came up with solutions. evaluate solutions summarize all approaches. lesson learned from challenge is then used improve simplifying assumptions made our original formulation discuss future extensions.