一种基于离散Hopfield神经网络的RTOS功耗优化方法

RTOS(Real-Time Operating System,实时操作系统)是SoC(System-on-a-Chip,系统芯片或片上系统)的一个重要组成部分,其功耗一般约占整个系统功耗30～40%的比例,而基于软/硬件划分的RTOS功耗优化方法(简称RTOS-Power划分)能够明显地减少SoC的功耗.因此,文中首先引入了RTOS-Power划分问题的一个新模型,这有助于理解RTOS-Power划分的本质.然后,提出了一种基于离散Hopfield神经网络的RTOS-Power划分方法,重新定义了神经网络的神经元表示、能量函数、运行方程和系数.最后,对该方法进行了仿真实验,并同遗传算法和蚂蚁算法进行了性能比较.实验结果表明:该文提出的方法能够以相对较小的代价(FPGA开销小于4K个可编程逻辑块)取得高达60%的功耗节省,同时,与纯软件实现的RTOS相比,系统性能也得到了相应的提高.

A Power Optimization Approach to Real-Time Operating Systems Based on Discrete Hopfield Neural Networks

The RTOS (Real-Time Operating System) is a critical component in the SoC (System-on-a-Chip), which consumes the 30~40% of total system energy in average. Power optimization based on hardware-software partitioning of a RTOS (RTOS-Power partitioning) can significantly reduce the energy consumption of a SoC. This paper presents a new model for RTOS-Power partitioning, which helps in understanding the essence of the RTOS-Power partitioning techniques. A discrete Hopfield neural network approach for implementing the RTOS-Power partitioning is proposed, where a novel neuron expression, energy function, operating equation and coefficients of the neural network are redefined. Simulations are carried out with comparison to generic algorithm and ant algorithm. Experimental results demonstrate that the proposed method can achieve higher energy savings up to 60% at relatively low costs of less than 4K PLBs while increasing the performance compared to the SoC-RTOS realized purely in software.