Method and device for determining tasks to be migrated based on cache perception

Abstract

The invention discloses a method for determining tasks to be migrated based on cache perception. The method comprises the steps that a source processor core and a target processor core are determined according to loads of each processor core; the number of times of cache invalidation and the number of executed orders of each task in the source processor core and the target processor core are monitored to obtain the number of times of cache invalidation of thousands of orders of each task in the source processor core and the target processor core; the average number of times of cache invalidation of thousands of orders of the source processor core and the average number of times of cache invalidation of thousands of orders of the target processor core are obtained; the tasks needing to be migrated from the source processor core to the target processor core are determined according to the average number of times of cache invalidation of thousands of orders of the source processor core and the average number of times of cache invalidation of thousands of orders of the target processor core. According to the method for determining the tasks to be migrated, an operating system can perceive the behavior of programs, and more reasonable tasks can be selected when the tasks are migrated. The invention further discloses a device for determining the tasks to be migrated based on cache perception.

Description

technical field [0001] The present invention relates to the technical field of computer science, in particular to a method and system for determining tasks to be migrated based on cache awareness. Background technique [0002] Task scheduling is one of the core functions of the operating system. The quality of task scheduling directly affects the performance, fairness, and real-time performance of the program. For an operating system with only a single processor core, task scheduling only needs to solve the problem of switching between different tasks. For an operating system with multiple processor cores, in addition to scheduling the switching between different tasks, it is also necessary to deal with the allocation of multiple tasks on multiple processor cores and the allocation of tasks among multiple processor cores. Migration process to ensure load balancing among multiple processor cores. [0003] In an operating system with multiple processor cores, multiple tasks ...

AI Technical Summary

Problems solved by technology

[0003] In an operating system with multiple processor cores, multiple tasks need to compete for many shared resources such as shared cache, memory controller, and memory bus. Different tasks have different requirements for resources. If the above resources are not considered, it will cause some resources (such as shared cache, shared memory access, etc.) to compete while the resources of other processor cores are not fully utilized, thus adversely affecting the performance of the entire system

[0004] In the processing of task scheduling in the prior art, the default is that when migrating tasks, the operating system selects the tasks that are allowed to migrate from the tail of the highest priority linked list containing tasks in order of priority from high to low, and does not analyze the tasks to be migrated. Therefore, the performance and quality of service of the entire system cannot be guaranteed. If the task to be migrated is not suitable for running on the target processor core, the performance of the system will become very bad

Images