Online Machine Learning for Energy-Aware Multicore Real-Time Embedded Systems

Abstract : In this paper, we present an Online Learning Artificial Neural Network (ANN) model that is able to predict the performance of tasks in lower frequency levels and safely optimize real-time embedded systems' power saving operations. The proposed ANN model is supported by feature selection, which provides the most relevant variables to describe shared resource contention in the selected multicore architecture. The variables are used at runtime to produce a performance trace that encompasses sufficient information for the ANN model to predict the impact of a frequency change on the performance of tasks. A migration heuristic encompassing a weighted activity vector is combined with the ANN model to dynamically adjust frequencies and also to trigger task migrations among cores, enabling further optimization by solving resource contentions and balancing the load among cores. The proposed solution achieved energy-savings of 24.97% on average when compared to the run-to-end approach, and it did it without compromising the criticality of any single task. The overhead incurred in terms of execution time was 0.1791% on average. Each prediction added 15.3585us on average and each retraining cycle triggered at frequency adjustments was never larger than 100us.

 EXISTING SYSTEM :

 ? We have considerably extended our previous work, which only gives some experimental findings but without theoretical analysis, to reduce the power consumption. The proposed technique differs from the existing ones in three aspects. ? However, existing work assumes the parameters of tasks are predefined, thus it is hard to directly apply these methods to solve the problem in IoT environment. ? We introduce the architecture of processing prototype scheduler, which can handle both user transactions and update transactions, while supporting real-time validity requirements of data processing in IoT system.

 DISADVANTAGE :

 ? The optimization techniques like Genetic, Particle Swarm, Stochastic Evolution, Ant, and others algorithms are used to determine the optimal solution of a problem. ? Recent researches in energy aware scheduling uses the optimization techniques to solve the Bin-packing and scheduling problem. ? The recent challenge is how to determine the tasks with harmonic relationship and their execution time are probabilistic rather than deterministic, which is considered an NP problem. ? The problem of partitioning tasks is considered as an NP- problem. So it is transformed into the binpacking problem or optimization techniques.

 PROPOSED SYSTEM :

 • The performance of the proposed methods is evaluated by simulation experiments. Compared to the state-ofthe-art, our methods offer remarkable energy savings. • Extensive simulation results show the proposed methods offer an improved energy saving performance compared to the state-of-the-art schemes. • For multicore solution, we propose a mapping scheme called TCBM, which outperforms the state-of-the art methods in terms of energy savings. • We also conduct theoretical analysis to confirm the reliability of the proposed method.

 ADVANTAGE :

 ? The power dissipation for processor is divided into static power and dynamic power. DVFS deals with enhancing the performance of system by decreasing the supply voltage leading to decrease the dynamic power and enhances the overall energy dissipation. ? DPM is used to determine a specific point in which the processor is switched into sleep mode to decrease the leakage current, leading to decrease in its static power. ? Many algorithms are used to decrease the energy consumption under a specific temperature threshold. Some use the meta-heuristic search algorithm and other use the linear programming method (ILP) to schedule tasks under thermal constraints. Both require a huge computation.