An-Najah Staff

A model for OS services in a distributed RSoCs is presented along with its formal description. The model provides the ability of self-adaptation at runtime. It allows the OS service to change its execution resource needs to accommodate the applications demands. A runtime algorithm is developed to be used with this novel model. The algorithm uses runtime profiling information to manage three resources (area, time, and power). In a relaxed resource distributed RSoCs, the algorithm finds an efficient execution configuration for any requested OS service. This configuration can be reconfigured at runtime to adapt to any resource change and the reconfiguration avoids as much losses as possible. The algorithm has a linear running time and performs multi-criteria optimization.