Task scheduling method and device based on heterogeneous processor, and electronic equipment

Abstract

The embodiment of the invention provides a task scheduling method and device based on a heterogeneous processor, and electronic equipment. The method comprises the following steps that: a main processor in the electronic equipment monitors the current CPU (Central Processing Unit) load of each heterogeneous slave processor, and identifies at least one source processor of which the current CPU load is greater than a first set threshold value corresponding to the current CPU load from each heterogeneous slave processor; at least one scheduling module is determined in each algorithm module of the source processor by aiming at each source processor; according to a corresponding relationship between the algorithm module stored in a module attribute table and at least one heterogeneous slave processor, a target processor scheduled by the scheduling module is obtained by aiming at each scheduling module; and connection is established between the source processor and the target processor, the scheduling module is scheduled to the target processor to enable the source processor and the target processor to jointly finish a task corresponding to the source processor. By use of the embodiment, the efficiency and the success rate of task scheduling can be improved.

Description

technical field [0001] The present invention relates to the field of computer technology, in particular to a heterogeneous processor-based task scheduling method, device and electronic equipment. Background technique [0002] In the current data analysis (such as video analysis, image analysis, etc.) system, it is necessary to use different algorithms for the same data analysis, and the implementation methods of different algorithms are very different, which is more suitable for heterogeneous processors with completely different instruction sets. Finish. In order to fully and evenly utilize the resources of various heterogeneous processors, task scheduling between such heterogeneous processors is very important. [0003] Existing task scheduling methods between heterogeneous processors mainly use feedback or prediction methods to schedule processing threads. Specifically, when the load of the source processor is too large, for the thread that needs to be scheduled in the s...

AI Technical Summary

Problems solved by technology

[0004] Therefore, the existing task scheduling method between heterogeneous processors has a complicated task scheduling process and low scheduling efficiency

Moreover, scheduling with threads as the smallest unit increases the difficulty of task scheduling, which can easily cause task scheduling failures

Images