Task Scheduling Strategy for Heterogeneous Multicore Systems

For heterogeneous computing systems, various types of processor cores cause system performance degradation due to uneven load. In addition, the inability of multitasking to match the appropriate processor core is also an urgent problem. This article proposes a swarm intelligence task scheduling strategy based on the genetic algorithm (GA) for high-performance heterogeneous multicore processors. In order to avoid the falling into local optimal solutions, we employ an adaptive mutation and injection strategy in the algorithm design. This swarm intelligence solution detects the computing capacities of different cores by processing specified tasks beforehand, and then an appropriate solution will be explored by introducing an adaptive mutation GA. Our technique aims to execute various types of tasks on heterogeneous processing cores for optimal performance. Experimental results show that this scheduling strategy can reduce the additional overhead and improve parallel computing efficiency and system performance.