Job scheduling method and system based on high response ratio, terminal and storage medium

Abstract

The invention discloses a job scheduling method and system based on a high response ratio, a terminal and a storage medium. An administrator terminal sets software and hardware basic information, a job queue, a user permission and a user weight factor; the user weight factor comprises a user weight factor dynamic value, a user weight factor quantity value and user weight factor cooling time; the user terminal submits the job after setting whether to use the user weight factor; and the user terminal enters a job queuing stage, the job priority of the user terminal is determined by a job scheduling method, and jobs are run according to the priority from high to low. According to the method, under the condition of non-manual intervention of software and hardware resources, the system automatically judges and operates current emergency important jobs, the problem that the user jobs cannot obtain computing resources relatively fairly due to fixed priorities during full-load operation of the system is solved, the scheduling success rate is increased, and the task fairness is improved.

Description

technical field [0001] The invention belongs to the field of computer application technology, and in particular relates to a job scheduling method, system, terminal and storage medium based on a high response ratio. Background technique [0002] In the current high-performance computing cluster environment, there are always some tasks that need to be run first. According to the traditional sorting method, the principle of fairness can be taken into account to a certain extent, but it cannot reflect the priority of user jobs. The traditional adjustment method can only manually intervene in the job priority order by the administrator, which is inefficient and extremely labor-intensive. [0003] The main process of the existing high response ratio priority scheduling algorithm in job scheduling is as follows: [0004] In order to simplify the description process, this method defines the minimum resource unit of a high-performance computer as a computer node, rather than severa...

AI Technical Summary

Problems solved by technology

But correspondingly, its disadvantage is to increase the response ratio calculation system overhead

In addition, according to this sorting method, the principle of fairness can indeed be taken into account to a certain extent, but because the processing job response is obtained by waiting time, it cannot reflect the priority of user jobs. When urgent tasks need to be run first, You can only manually intervene in the job priority order, resulting in low efficiency and affecting job processing efficiency

Images