Supercalculation platform resource use monitoring method

Abstract

The invention provides a supercalculation platform resource use monitoring method. The method comprises the steps of determining the number of user task processes which are being executed on a currentcalculation resource; judging whether the number of the user processes being executed is greater than or equal to 2 or not; If so, determining whether the scheduling system allocates the current calculation resource or not, If not, determining that all the user task processes which are being executed on the current calculation resource are the user task processes which are submitted by mistake.

Description

technical field [0001] The invention relates to the field of supercomputing, in particular to a method for monitoring the use of supercomputing platform resources. Background technique [0002] Supercomputing platforms have been widely used in various industries. When users submit tasks to the supercomputing platform, they will apply for the required computing resources, including CPU and GPU, etc., and the scheduling system of the supercomputing platform will allocate computing resources for the task. Under reasonable circumstances, the computing resources are occupied and used by the task process submitted by the assigned user. However, there are unreasonable situations in the actual situation, such as conflicts caused by unreasonable submission of task processes by other users to the computing resource. These unreasonable situations require the operation and maintenance personnel of the supercomputing platform to regularly check and solve them. In the prior art, the che...

AI Technical Summary

Problems solved by technology

However, there are unreasonable situations in the actual situation, such as other users unreasonably submitting task processes to this computing resource, resulting in conflicts and other issues

These unreasonable situations require the operation and maintenance personnel of the supercomputing platform to regularly investigate and solve them. In the existing technology, the investigation of these problems mainly relies on manual operations based on random logic, which is very inefficient.

Images