Method and device for automatic job submission of numerical simulation parallel computing

Abstract

The invention provides a method and device for automatic task submitting in numerical simulation parallel computing. The method comprises the steps of automatically mounting a user personal space distributed to a user in a server memory space to a local client, and generating a user information packet; scanning .DAT files in the user personal space, and sorting and screening data files already existing in the user personal space; endowing priority to the screened data files, making computing waiting marks, and generating a file information packet; transmitting the file information packet and the user information packet to a parallel computing server, writing information in the file information packet, such as file states, into a database; acquiring information from the database, performing sensible resource allocation and then performing computing, and computing a submitted task. According to the method and device, the numerical simulation parallel computing efficiency can be improved, the network quality influence is lowered, and the time and human cost in the numerical simulation parallel computing process is reduced.

Description

Technical field [0001] The invention relates to the technical field of oil reservoir development information, in particular to a method and device for automatically submitting jobs for parallel calculation of numerical simulation. Background technique [0002] There are two main functions for the automatic submission of assignments for numerical simulation parallel computing. One part is located on the client side. When the user turns on the local machine, the user information is confirmed according to the user information passed by the server-side monitoring module last time, and the legal user information is packaged into a user information package. In addition, immediately scan the .DATA files in the automatically mounted personal space, sort the files based on the last time stamp passed by the server-side monitoring module, filter the files, and record the time stamp again and record it in the file information. The calculation file is given priority according to the user aut...

AI Technical Summary

Problems solved by technology

In the case of centralized configuration of hardware, users are faced with the following prominent problems: ①One-machine-one-use or special-purpose mode leads to shortage of hardware resources in a certain period of time, while a large number of advantageous resources in other times are idle, and the overall advantages of hardware resources are not reflected. In addition, sometimes due to network jitter, the user's connection to the parent process is interrupted, and the server's terminal process, that is, the user's job calculation, is terminated; ② In terms of storage resources: users wait for a long time to use network data, because users currently cannot efficiently use core area computers, storage and For network resources, only the local hard disk is used as a cache, and the actual reading and writing are transmitted through the network many times, resulting in an additional network delay each time; the data is stored in a specific storage device, and the storage device is directly connected to a dedicated server. It greatly restricts the use of data by users, which means that when software and hardware resources are available, the "floating" of data applications cannot be realized, which means that users cannot use additional computing resources, and the free area of ​​specific storage devices is difficult to implement. Used by other users, resulting in waste of storage resources

Images