Research on Task Scheduling Model for Multi-core Processor

Multi-core processor task scheduling has been proved to be an NP-hard problem, and the existing solutions mostly use heuristic algorithms. Among these heuristic algorithms, the saving algorithm is an algorithm with higher scheduling success rate. However, the algorithm does not take into account the parallelism between tasks, just use the deadline for the task to judge, did not make full use of the constraints of shared resources. In the case of larger correlation between tasks and more mutually exclusive resources, it is expensive to access mutually exclusive resources between cores, which make it easy for tasks to wait for a long time for mutually exclusive resources and reduce the utilization rate of the cores. The dispatched success rate decreased. In response to this problem, this paper introduces the concept of task relevance, using task relevance to determine the size of the relevance between tasks, sharing the number of mutually exclusive resources, so as to task-related tasks as much as possible to a scheduling Kernel to reduce the overhead of mutually exclusive access to related tasks. Experiments show that the improved algorithm has less exclusive access overhead than the original algorithm under the condition of more shared resources.