A dynamic resource allocation optimization method for compilation subsystems in power IoT environment

Abstract Affected by the resource allocation skew problem, which leads to the low utilization rate of resources, we propose to study the optimization method of dynamic resource allocation of compiler subsystem under the environment of electric power IoT (Internet of Things). The skewness willingness function is constructed to reflect the imbalance degree of the compiler subsystem partitioning in the power IoT environment, and at the same time, quantitative indexes are provided for evaluating the effectiveness of the compiler subsystem partitioning algorithm in the power IoT environment. In the resource allocation stage, combined with the compiler subsystem resource tilt allowable threshold, the dynamic allocation of resources is carried out based on the existing compiler tasks in accordance with the priority, with the goal of minimizing the additional waiting time (tilt). In the test results, the compiler subsystem is tested for the distribution of η EE under different types of requests is always within 0.25 J/bit. Compared with the control group, it has obvious advantages in processing performance and stability.