Research on Compensation Policy Scheduling Algorithm for Multi-core Processors

Abstract

With rapid development of processors, multi-core processors occupied the main market due to their superior performance and power characteristics. Threads run concurrently on a multi-core processor and share the processor’s the second-level (L2) cache. Cache allocation and sharing is critical to the effective utilization of multi-core processors.  Unfair CPU cache allocation can easily delay the critical task and cause serious problems, such as thread starvation, priority inversion, and inadequate CPU accounting. To solve these problems, this paper proposed a compensation policy scheduling algorithm which can decrease system variability and improve system performance effectively. The experiment results show that the system variability adopting compensation policy scheduling algorithm is lower than that using conventional scheduling algorithms. Consequently, compensation policy scheduling algorithm improves the stability of system and applications effectively on multi-core processors.