Task processing method, device and equipment

Abstract

The present application provides a task processing method, device, and equipment, the method including: receiving a first job message sent by a client and matched with a first resource management system, the first job message includes a first resource use command, and a first resource Use the parameter value and job script information corresponding to the command; the first resource use command matches the command line template of the first resource management system; if the HPC device uses the second resource management system to perform the task, then convert the first job message into the same as the first resource management system The second job message matched by the second resource management system, the second job message includes the second resource use command, the parameter value corresponding to the second resource use command and job script information; the second resource use command and the command of the second resource management system Perform template matching; send the second job message to the HPC device. Through the technical solution of the present application, the HPC device can execute the user's job message to meet the requirements of different users for using the resource management system.

Description

technical field [0001] The present application relates to the technical field of communications, and in particular, to a task processing method, device and equipment. Background technique [0002] The main purpose of HPC (High Performance Computing, high-performance computing) equipment is to increase the computing speed and reach the computing speed of trillions per second. The requirements for processors, memory bandwidth, computing methods, system I / O, storage, etc. All are relatively high, and each link will affect the computing speed of HPC equipment. HPC equipment mainly solves the calculation of large-scale scientific problems and the processing of massive data, such as scientific research, weather forecast, computational simulation, military research, biopharmaceuticals, gene sequencing, image processing, etc. [0003] HPC equipment can use resource management systems such as SLURM (Simple Linux Utility for Resource Management), LSF (Load Sharing Facility, load shar...

AI Technical Summary

Problems solved by technology

If the user sends a job message that matches the LSF resource management system to the HPC device, the HPC device cannot execute the job message, that is, the execution of the job message fails

[0004] However, different users are very different in their familiarity with the resource management system. If the user only knows the job information that matches the LSF resource management system, then the HPC device cannot execute the user's job information when the HPC device adopts the SLURM resource management system. , causing the job message execution to fail

Images