A task scheduling method, device, equipment and medium

Abstract

The present application discloses a task scheduling method, device, device, and medium. The method includes: acquiring a task to be executed, and dividing the task to be executed into a plurality of subtasks to be executed that are executed in parallel; determining each of the tasks to be executed The number of shared variables between subtasks; based on the number of shared variables, each of the subtasks to be executed is grouped to obtain different task groups to be executed; each of the task groups to be executed is assigned to the corresponding CPU, and is Variables in the to-be-executed task allocate storage space so that the CPU executes the to-be-executed subtasks in the assigned to-be-executed task group. In this way, subtasks with higher coupling degree can be allocated to the same CPU, avoiding the blocking problem of subtasks, improving task execution efficiency, and further improving the data processing efficiency and performance of the entire computer system.

Description

technical field [0001] The present application relates to the technical field of multi-core computers, and in particular, to a task scheduling method, apparatus, device, and medium. Background technique [0002] Multi-core processors are the current mainstream commercial CPU (Central Processing Unit, central processing unit) architecture. Multi-core processors integrate multiple processor cores into one chip, and all processor cores can access all physical memory. , easy to manage, so it has been widely used. Multi-core programming often adopts multi-task mode design, and assigns tasks to each processor core for execution in parallel. The typical programming model is shared memory programming model. [0003] see figure 1 As shown, in the UMA (Unified Memory Architecture, unified memory architecture) architecture, each CPU implements consistent storage, and each CPU can access memory through a memory controller. and see figure 2 As shown, different processors on the same...

AI Technical Summary

Problems solved by technology

However, in this allocation mode, the operating system scheduling module only schedules subtasks mechanically, which will lead to frequent access between subtasks running inside different CPUs when subtasks with a high degree of coupling are allocated to different CPUs. The memory operation leads to the blocking problem of subtasks, which leads to low task execution efficiency, which in turn reduces the data processing efficiency of the entire computer system and reduces the performance of the computer system.

Images