KR102644456B1 - Local installation of low-resolution version of climate prediction software - Google Patents

Abstract

A method for local installation of low-resolution GloSea6 is disclosed. A method of local installation of low-resolution climate prediction software according to an embodiment includes configuring a community model for low-resolution climate prediction software; And it may include installing low-resolution climate prediction software on a local electronic device using the community model for the configured low-resolution climate prediction software.

Description

LOCAL INSTALLATION OF LOW-RESOLUTION VERSION OF CLIMATE PREDICTION SOFTWARE}

The description below relates to local installation techniques for climate prediction software.

Figure 1 is a diagram for explaining GloSea. GloSea is a seasonal forecasting system developed and operated by the Met Office. GloSea is an ensemble forecasting system built on a high-resolution version of the Met Office climate prediction model. It is a climate model that combines the atmosphere, land, ocean, sea, and ice of the HadGEM3 model, and the horizontal resolution has been increased from the existing version of GloSea. Each forecast requires initial ocean, sea ice, land, and atmospheric conditions, with the latter using analysis fields also used for short-range weather forecasts. Ocean and sea ice initial conditions are generated using 3DVAR data assimilation developed in the multi-institution NEMOVAR project.

GloSea is structured in such a way that it is difficult to quickly commercialize technology developed over a long period of time through R&D projects. Because the systems used by the Meteorological Administration and R&D are different, significant manpower and time must be reinvested in commercialization. Additionally, there is no standardized development platform in GloSea. Because data processing, diagnosis/analysis, and improvement experiment processes are performed individually, it is difficult to integrate research results. Additionally, modeling work for intra-seasonal scale prediction requires relatively more computational resources compared to weather prediction.

By applying parallel processing and distributed computing techniques, it is possible to provide a local installation method and system for low-resolution climate prediction software that is portable to heterogeneous types.

A method of local installation of low-resolution climate prediction software performed by a local installation system includes constructing a community model for low-resolution climate prediction software; And it may include installing low-resolution climate prediction software on a local electronic device using the community model for the configured low-resolution climate prediction software.

The configuring step includes constructing a grid of the LM model and the MM model using the source code of the community model, and the atmospheric information of the LM model has a resolution of N48 and a grid size of 2.5°x3.75°. , In the ocean information, the resolution is set to eORCA1 and the grid size is set to ~1.0°, in the atmospheric information of the MM model, the resolution is N96, the grid size is set to 1.88°x1.25°, and in the ocean information, the resolution is set to eORCA1 and the grid size is set to ~1.0°. It can be set to 1.0°.

The configuring step may include configuring an interface for model execution of the community model by setting a virtualization-based standard environment.

The installing step may include determining whether an own server exists in the electronic device as a virtual machine is installed in the electronic device.

The installing step may include installing Vagrant provided from a service provider that provides the source code of the community model when the electronic device does not have its own server.

The installation step includes installing a task setting management and execution tool (Rose), a task control management tool (Cylc), and a system to assist in software installation and management (FCM) when the electronic device has its own server. can do.

The installing step may include installing a library for local execution of low-resolution climate prediction software.

The installing step may include installing a user suite for the low-resolution climate prediction software by installing the library.

The low-resolution climate prediction software may include GloSea6.

It may include a computer program stored in a computer-readable recording medium to execute a local installation method on the local installation system.

A local installation system of low-resolution climate prediction software includes at least one processor configured to execute computer-readable instructions included in a memory, wherein the at least one processor configures a community model for low-resolution climate prediction software, and the configured Using the community model for low-resolution climate prediction software, low-resolution climate prediction software can be installed on local electronic devices.

The at least one processor may determine whether the electronic device has its own server as a virtual machine is installed in the electronic device.

If the electronic device does not have its own server, the at least one processor installs Vagrant provided from a service provider that provides the source code of the community model, and if the electronic device has its own server, In this case, you can install the Job Settings Management and Execution Tool (Rose), the Job Control Management Tool (Cylc), and a system to help you install and manage software (FCM).

The at least one processor may install a library for local execution of low-resolution climate prediction software.

The at least one processor may install a user's suite for the low-resolution climate prediction software by installing the library.

By applying parallel processing and distributed computing techniques, the performance of climate prediction software that can be ported to heterogeneous systems can be improved.

Simplified versions of climate prediction software can be installed locally to stimulate research and increase usability for researchers.

Figure 1 is a diagram for explaining GloSea.
Figure 2 is a diagram for explaining the general operation of the GloSea community model according to an embodiment.
Figure 3 is a diagram for explaining the operation of a community model for low-resolution climate prediction software according to an embodiment.
Figures 4 to 7 are diagrams for explaining the operation of the community model for the low-resolution climate prediction software of Figure 3.
Figures 8 and 9 are diagrams for explaining the directory file structure of a community model for climate prediction software according to an embodiment.
Figure 10 is a flowchart for explaining a method of local installation of low-resolution climate prediction software, according to one embodiment.

Hereinafter, embodiments will be described in detail with reference to the accompanying drawings.

In the embodiment, an operation for locally installing a simplified version of climate prediction software will be described in order to promote research and increase researchers' usability. Hereinafter, GloSea will be described as an example of climate prediction software.

Figure 2 is a diagram for explaining the general operation of the GloSea community model according to an embodiment. GloSea Community Model means a simplified version of GloSea.

GloSea6 simplified version refers to a version in which various system and environmental settings of GloSea6 have been adjusted for testing on Intel (x64)-based Linux operating systems. This simplified version of GloSea6 will be described as the Community Model for Climate Prediction Software.

Locally installed systems can construct community models for climate prediction software. The local installation system can configure an interface for model execution of the community model by setting up a virtualization-based standard environment.

The local installation system can construct a grid of the LM model and MM model using the source code of the community model. For example, the source code of the community model provided by the British Meteorological Administration can be used by the Korea Meteorological Administration. In the atmospheric information of the LM model, the resolution is set to N48 and the grid size is 2.5°x3.75°. In the ocean information, the resolution is set to eORCA1 and the grid size is set to ~1.0°. In the atmospheric information of the MM model, the resolution is set to N96 and the grid size is set to ~1.0°. In ocean information, the resolution can be set to 1.88°x1.25°, eORCA1, and the grid size can be set to ~1.0°.

A locally installed system can test the next operational climate prediction model (Glosea6)-based coupled system. For example, atmosphere (GA 8.2), land (GL 8.2), ocean (GO6), sea ice (GSI8), and joiner (OASIS-mct) could be used. Specifically, UM 11.3, JULES 5.4, NEMO 3.6, CICE 5.0, and OASIS-mct 4.0 can be used. In resource allocation, for the LL model, it can be set to UM (4 there is.

The local installation system is an interface for performing community models and can configure virtualization-based Rose/Cylc systems, container-based virtualization (Docket) application, patches, and package incremental update systems.

The local installation system can set up a standard environment based on virtualization. Accordingly, consistency of operating system, compiler vendor, and version can be secured. For example, OS is Ubuntu 18.x, CentOS 7.x, Kernel is 3.0 or higher, compiler is gcc, gfortran 7.x or higher, openmpi 3.1.4 or higher from GNU, icc, ifort 20.x or higher, openmpi from Intel. Environments such as 3.1.4 or higher, Python 2.7.x, 3.8, required libraries, and utility installations such as Shumlib, GCOM, and Mule can be configured. Additionally, a virtualization-based standard environment can be defined and distributed. For example, Vagrant, VMWare standard image creation, container-based virtualization (Docker), Patch, or package incremental update system can be applied.

For this community model, providers perform weather forecasts and weather-related research using GloSea on heterogeneous systems, developers develop applications using weather data, installation data, and the Korea Meteorological Administration API, and users obtain necessary weather information. .

According to embodiments, local installation of low-resolution climate prediction software may provide optimal scalability for a simplified version of the climate prediction software. By enabling optimization for various compilers, the usability of community models for climate prediction software can be increased. In addition, it can be optimized for multiple computational platforms, making it possible to use it as a community model that can be used for joint research by other institutions with different environments.

According to the embodiment, preparation and technology for a high-resolution climate prediction software measurement prediction system can be accumulated. In preparation for operating high-resolution climate prediction models in the future, optimization technologies for compilers, calculation platforms, and parallel environments can be accumulated in advance. It is possible to proactively study the efficient computational performance of high-resolution climate prediction models with very large amounts of computational data.

Figure 3 is a diagram for explaining the operation of a community model for low-resolution climate prediction software according to an embodiment.

The low-resolution GloSea6 system may include a distribution server (KMA Glair) 310, a cluster server 320, and an electronic device (Local PC Client) 330.

The distribution server 310 distributes source information related to GloSea6 provided by the Korea Meteorological Administration (KMA Glair). In order to download the latest GloSea to the distribution server 310, you must register the cluster server IP address. By registering the IP address of the cluster server in the distribution server 310, the user's account information can be applied for and registered. Accordingly, only users with registered user account information can access the distribution server.

The cluster server 320 is a server configured to locally install low-resolution Glosea6. At this time, a local installation system can be operated that configures a community model for low-resolution GloSea6 on a cluster server and installs low-resolution climate prediction software on local electronic devices using the configured community model for low-resolution GloSea6. Cluster server (320) includes 1) GNU Compiler, 2) FCM, Rose, Cylc, 3) HDF5, NetCDF (+Parellel), 4) Eccode, Zlib, 5) GCOM, Shumlib, 6) OASIS3-mct, 7) CTLDATA , Input Data, etc. can be installed and set.

The electronic device 330 may be a fixed terminal implemented as a computer device or a mobile terminal. Examples of the electronic devices 330 include smart phones, mobile phones, navigation devices, computers, laptops, digital broadcasting terminals, Personal Digital Assistants (PDAs), Portable Multimedia Players (PMPs), and tablet PCs. For example, the electronic device 330 may communicate with other terminals and/or servers 300 and 310 through a network using a wireless or wired communication method. To the electronic device 330, 1) Virtual Machine (VMware), 2) Wget, 3) bash shell for Window (ex git), 4) cmdt Setting(HOST: /etc/hosts, ROSE: rose config, CYLC: globar. rc) can be installed and set up.

Data can be transmitted and received between the cluster server 320 and the distribution server 310 by being connected through FCM. Here, FCM is an open source system that helps install and manage software, and performs Fortran build system, Subversion version management, and Trac project management. GloSea src may be requested from the cluster server 320 to the distribution server 310. The latest GloSea src may be provided from the distribution server 310 to the cluster server 320. GloSea src may be updated from the distribution server 310 to the cluster server 320. Cluster server 320 can update GloSea src on its own. Accordingly, the current modeling performance environment and the joint research environment model program source can be synchronized.

The cluster server 320 and the electronic device 330 can be connected via SSH to transmit and receive data. GloSea execution may be requested from the electronic device 330 to the cluster server 320 using the Cylc workflow engine. The cluster server 320 may return the execution result file to the electronic device 330. The cluster server 320 can re-execute the GloSea result file with the electronic device 330.

In detail, referring to FIG. 4 , the task (suite) that the user wants to perform can be executed on the electronic device 330 through a task setting management and execution tool (rose). The electronic device 330 may transmit a work (suite) file that the user wants to perform to the cluster server 320. rose is a framework for managing, configuring, and executing suites, and uses the Cylc engine to execute suites. Cylc performs workflow configuration and monitoring. Cylc serves as a task scheduler, monitoring, and task control management tool.

Referring to FIG. 5, the cluster server 320 can read a task (suite) file that the user wants to perform transmitted from the electronic device 330 and operate Cylc. The cluster server 320 may operate a log on the electronic device 330. The cluster server 320 may request fcm_make from the distribution server 310. In this case, fcm_make may be a task workflowed by Cylc.

Referring to FIG. 6, the distribution server 310 may respond regarding the fcm_make task. The distribution server 310 may deliver the GloSea source to the cluster server 320.

Referring to FIG. 7, Cylc may be operated in the cluster server 320, and the resulting data may be transmitted to the electronic device 330.

Figures 8 and 9 are diagrams for explaining the directory file structure of a community model for climate prediction software according to an embodiment.

Figures 8 and 9 show the directory structure of the cluster server. By default, the home directory location of a Linux user is /home/$USER, but since the cluster server is /home/users/$USER, you can confirm that a shortcut is created. You can.

Figure 10 is a flowchart for explaining a method of local installation of low-resolution climate prediction software, according to one embodiment.

The local installation system can install a virtual machine on an electronic device (1110). A virtual machine can be installed on an electronic device. A virtual machine is software that can be used on multiple operating systems on one device. It is a workflow engine that predefines package installation, user environment configuration, etc.

As the local installation system installs the virtual machine on the electronic device, it can determine whether the electronic device has its own server (1120). If the electronic device does not have its own server, the local installation system can install Vagrant provided from a service provider that provides the source code of the community model (1130). At this time, when installing Bagrant, a task setting management and execution tool (Rose), a task control management tool (Cylc), and a system to assist with software installation and management (FCM) can be installed manually. In detail, it can be installed in the following order: Vagrant installation, metomi installation, FCM, Rose, and Cylc installation in VirtualBox installed on the electronic device. Centos 7 can be installed by cd TI/utils/install-um-libs/virtualbox, yum localinstall -y./VirtualBox-6.1-6.1.22_144080_el7-1.x86_64.rpm. metomi-vms is a project for Vagrant virtual machines with FCM + Rose + Cylc installed. By default, Ubuntu 18.04 can be installed, and CentOS 7 can be installed depending on the user.

If the electronic device has its own server, the local installation system can install the task settings management and execution tool (Rose), the task control management tool (Cylc), and the system to assist with software installation and management (FCM) (1140). If you proceed with your own server, you cannot use the Rose/Cylc library image file provided by Vagrant, so you must install FCM, Rose, and Cylc manually. FCM (Flexible Configuration Management system) is a system that simultaneously supports a development history (version) management system (Subversion) and a Fortran-based compilation program (Make). Run the Install-fcm-rose-cylc.sh file to install all versions of FCM, Rose, and Cylc. GloSea6 is optimized for FCM 2019.05.0, but higher versions can also be installed. Rose can be installed with the recommended version 2019.01.2 (ROSE_VERSION=2019.01.2, ln -sf /opt/ROSE/rose-$ROSE_VERSION /opt/rose, ln -sf /usr/local/bin/rose /usr/ local/bin/rosie). Cylc can be installed with CYLC_VERSION=7.8.3.

The local installation system may install libraries for local execution of low-resolution climate prediction software (1150). The local installation system is Install GCOM, Install Shumlib, Install NetCDF and related libraries, Install ZLIB-1.2.11 (as root), Install HDF5, Install HDF5-1.10.3-parallel (as root), Install NetCDF-C-4.7. 3 (as root), Install NetCDF-Fortran-4.4.4 (as root), Install Eccodes and related libraries, Install Cmake-3.20.2 (as root), Install Eccodes-2.9.2 (as root), Install OASIS3- mct and related libraries, Install parallel-netcdf-1.8.1 (as root), Install GPTL, Install libtool-2.4.6 (as root), Install gptl-8.0.3 (as root), Install pio-2.5.4 ( You can install Install pio-2.5.4 (as root), Install OASIS3-MCT 4.0 (not as root), and Install CTLDATA.

Install GCOM is a communication layer that provides an interface for message passing with a multi-processor machine. Since GCOM supports MPI, you must first install MPICH and then install GCOM. The recommended GCOM version for the low-resolution combined model is 6.8, and the MPICH version optimized for GCOM 6.8 is 3.1.4, and all versions of GCOM 6.8 or higher are supported (Install MPICH-3.1.4 (as root), Install GCOM 6.8 (not as root)).

Install Shumlib is a collection of utilities used in UM. Install Shumlib-2019.10.1 (not as root) can be installed (fcm co fcm:shumlib/tags/2019.10.1@HEAD shumlib-2019.10.1 (2019.10.1)). GCOM and Shumlib are essential libraries for installing UM.

Install NetCDF and related libraries in the following order: ZLIB, HDF5, NetCDF, and NetCDF-Fortran. Installation via yum is also possible, but errors may occur during the process or errors may occur during compilation of low-resolution combined models even after installation is complete, so it is recommended to compile the source code directly.

Install HDF5 requires both serial and parallel versions depending on the user environment, so both versions must be installed. If you receive it from the web or install a version without _modified in the file name, a warning or error will occur during the compilation process, so you must use the version with _modified in the file name.

Install Eccodes and related libraries are installed in the order of Cmake and Eccodes.

Install OASIS3-mct and related libraries in the following order: PnetCDF, GPTL, Parallel-IO, and OASIS3-mct. Installation of PnetCDF to Parallel-IO is not required, so you can install only OASIS3-mct.

Install CTLDATA installs cd roses, Fcm, fcm bc --bob --type DEV::USER um.xm_tr_vn11.3_LowGS6, fcm:um.xm_br/dev/nikebagf/um.xm_tr_vn11.3_LowGS6@67.

The local installation system can install the user's suite for GloSea6 by installing the library (1160). A low-resolution combined model (GloSea6) can be developed based on the u-av133 suite, which is a pilot version of the GloSea6-GC3.2 model. The British Meteorological Administration develops models in the form of a Rose/Cylc-based suite. Among these suites, the suite provided by the British Meteorological Administration through the roses project in MOSRS (Met Office Science Repository System) begins with u, followed by two English letters and three numbers. The serial number of the digit is attached. The u-av133 suite is divided into forecast, hindcast, and monthly mean tasks, and the low-resolution combined model applies only the forecast part and is registered as branch r3316_N96eORCA1_kmt-ab447 of the kmt-ab347 suite managed by the Korea Meteorological Administration. For example, <run the following to download the suite. A message such as “Only personal PCs, laptops, or servers that have requested APCC firewall permission are allowed to access.” may be displayed.

By locally installing a community model for low-resolution climate prediction software, the following effects can be expected: Collaborative research can be conducted by securing the versatility of a community model (simplified version) for low-resolution climate prediction software. Currently, climate models are developing into Earth system models with a large amount of calculations due to the advancement of modeling technology, and there is a trend to develop the technology through joint research by various organizations. A simplified version with a relatively small amount of calculation can be used as a community model to stimulate joint research by multiple model research institutes.

Additionally, Korea's seasonal forecasting system can be upgraded. Optimized operation of seasonal and intra-seasonal prediction systems using community models for climate prediction software can efficiently produce highly reliable prediction results. Highly accurate intra-seasonal climate prediction results can have a positive effect on planning for national resources such as disaster, energy, and food, and on social and economic policy decisions.

The device described above may be implemented with hardware components, software components, and/or a combination of hardware components and software components. For example, devices and components described in embodiments may include, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA), etc. , may be implemented using one or more general-purpose or special-purpose computers, such as a programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions. A processing device may execute an operating system (OS) and one or more software applications that run on the operating system. Additionally, a processing device may access, store, manipulate, process, and generate data in response to the execution of software. For ease of understanding, a single processing device may be described as being used; however, those skilled in the art will understand that a processing device includes multiple processing elements and/or multiple types of processing elements. It can be seen that it may include. For example, a processing device may include a plurality of processors or one processor and one controller. Additionally, other processing configurations, such as parallel processors, are possible.

Software may include a computer program, code, instructions, or a combination of one or more of these, which may configure a processing unit to operate as desired, or may be processed independently or collectively. You can command the device. Software and/or data may be used on any type of machine, component, physical device, virtual equipment, computer storage medium or device to be interpreted by or to provide instructions or data to a processing device. It can be embodied in . Software may be distributed over networked computer systems and stored or executed in a distributed manner. Software and data may be stored on one or more computer-readable recording media.

The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc., singly or in combination. Program instructions recorded on the medium may be specially designed and configured for the embodiment or may be known and available to those skilled in the art of computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic media such as floptical disks. -Includes optical media (magneto-optical media) and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, flash memory, etc. Examples of program instructions include machine language code, such as that produced by a compiler, as well as high-level language code that can be executed by a computer using an interpreter, etc.

As described above, although the embodiments have been described with limited examples and drawings, various modifications and variations can be made by those skilled in the art from the above description. For example, the described techniques are performed in a different order than the described method, and/or components of the described system, structure, device, circuit, etc. are combined or combined in a different form than the described method, or other components are used. Alternatively, appropriate results may be achieved even if substituted or substituted by an equivalent.

Therefore, other implementations, other embodiments, and equivalents of the claims also fall within the scope of the claims described below.

Claims (15)

In a method of local installation of low-resolution climate prediction software performed by a local installation system,
Constructing a community model for low-resolution climate prediction software; and
Installing low-resolution climate prediction software on a local electronic device using the community model for the low-resolution climate prediction software configured above.
Including,
The installation steps are:
Install a virtual machine on the electronic device, determine whether the electronic device has its own server as the virtual machine is installed, and if the electronic device does not have its own server, the community model Install Vagrant provided by a service provider that provides source code, and if Vagrant is installed or the electronic device has its own server, a task setting management and execution tool (Rose), a task control management tool ( Cylc) and a system to help install and manage the software (FCM), install a library for local execution of the low-resolution climate prediction software, and install the library to perform the user's actions (suite) on the low-resolution climate prediction software. ) Steps to install
Including,
The source code of the community model is provided by the service provider, and the source code of the community model provided by the service provider is distributed by the distribution server, and the cluster server is connected to the distribution server to download the latest community model. As the IP address is registered, the user's account information is applied and registered, and access to the distribution server is possible based on the registered user's account information.
Local installation method, including . According to paragraph 1,
The configuring steps are,
Constructing a grid of the community model using the source code of the community model
Local installation method, including . According to paragraph 1,
The configuring steps are,
Setting up a virtualization-based standard environment to configure an interface for model execution of the community model
Local installation method, including . delete delete delete delete delete According to paragraph 1,
A local installation method, characterized in that the climate prediction software is GloSea6. A computer program stored in a computer-readable recording medium to execute the local installation method of any one of claims 1 to 3 and 9 on the local installation system. In a local installation system of low-resolution climate prediction software,
At least one processor configured to execute computer-readable instructions contained in the memory,
The at least one processor,
Construct community models for low-resolution climate prediction software;
Including installing low-resolution climate prediction software on a local electronic device using the community model for the low-resolution climate prediction software configured above,
In the at least one processor, a virtual machine is installed in the electronic device, and as the virtual machine is installed, it is determined whether the electronic device has its own server, and if the electronic device does not have its own server. If not, install Vagrant provided by a service provider that provides the source code of the community model, and if Vagrant is installed or the electronic device has its own server, a task settings management and execution tool (Rose ), install a task control management tool (Cylc) and a system to help install and manage the software (FCM), install libraries for local execution of low-resolution climate prediction software, and install said libraries to enable local execution of said low-resolution climate prediction software. Install the user's suite for
The source code of the community model is provided by the service provider, and the source code of the community model provided by the service provider is distributed by the distribution server, and the cluster server is connected to the distribution server to download the latest community model. As the IP address is registered, the user's account information is applied and registered, and access to the distribution server is possible based on the registered user's account information.
A local installation system containing .
delete delete delete delete