Method and device for managing tasks of many-core system

Abstract

The invention provides a method and a device for managing tasks of a many-core system, and relates to the field of computers. The method includes adding task descriptor pointers of processes into task descriptor items corresponding to the processes if the processes are required to be added into task linked lists, and adding the task descriptor items into the task linked lists; searching the task descriptor items corresponding to the processes if the processes are required to be deleted, and removing the task descriptor items from the task linked lists; removing available priority index registers with the highest priority from priority index register queues when new tasks are required to run by processor cores. The method and the device in an embodiment of the invention have the advantages that the tasks of the system can be scheduled and managed by the aid of the device for managing the tasks in the many-core system, and accordingly the process management efficiency of the many-core system can be improved; the method and the device are used for scheduling and managing the tasks of the system by the aid of the device for managing the tasks.

Description

technical field [0001] The invention relates to the field of computers, in particular to a task management method and device for many-core systems. Background technique [0002] In the future processor manufacturing process, the number of processor cores integrated on a single chip will increase. This kind of processor integrating many processor cores is generally called a many-core processor (Many Core). In the environment of many-core processors, the task scheduling of the operating system still adopts the simple operating system software task scheduling method, and the efficiency bottleneck of the task scheduling itself will become the bottleneck of the entire system; if the hardware scheduling method is completely adopted, it will bring many problems. The complexity of the large hardware system increases the difficulty of hardware design and verification, and may cause the system to be completely incompatible with existing systems. [0003] In the prior art, taking the ...

AI Technical Summary

Problems solved by technology

[0006] The algorithm of load balancing is very complicated, and its execution time will increase with the increase of the number of processors. In future many-core processors, the time required for runqueue to perform load balancing will reduce the operating efficiency of the system, and whenever there is When the CPU calls the scheduler() function, the system will lock the runqueue, so that each processor needs to wait for the runqueue resource

Therefore, the system's frequent waiting for runqueue resources will also reduce the efficiency of system process management

Images