DAS: Dynamic Adaptive Scheduling for Energy-Efficient Heterogeneous SoCs

Domain-specific systems-on-chip (DSSoCs) aim at bridging the gap between application-specific integrated circuits (ASICs) and general-purpose processors. Traditional operating system (OS) schedulers can undermine potential of DSSoCs since their execution times be orders magnitude larger than time task itself. To address this problem, we propose a dynamic adaptive scheduling (DAS) framework that combines benefits fast (low-overhead) scheduler slow (sophisticated, high-performance but high-overhead) scheduler. Experiments with five real-world streaming applications show DAS consistently outperforms both schedulers. For 40 different workloads, achieves on average 1.29x speedup 45% lower EDP compared to sophisticated low data rates 1.28x 37% when workload complexity increases.