Reducing idle time in event-triggered software execution via runnable migration and DPM-Aware scheduling

Being the main driver of innovation in the automotive industry, vehicle software calls for more efficient tools for its development. Safety-critical applications are predominantly written respecting the AUTOSAR standard, where tasks comprise code-fragments called runnables, and deployed on multi-core architectures. However, task parallelization introduces a substantial number of idle intervals, as runnables have to be mapped and scheduled at design-time without violating data dependencies. In parallelized event-triggered tasks, these intervals are utilized by migrating runnables from the interrupted time-triggered tasks. We propose to also migrate runnables from consecutively scheduled tasks and additionally, to maximize their number using DPM-aware application scheduling.