How To Install
This page will explain how to install all the packages you need for this project step by step.
Requirement
This package requires a C++ compiler”. Environments under intel 19.1.2 were tested
Install libraly code (MolStrucClassifier)
Copy the repository
You can copy the repository by typing the following command:
git clone https://github.com/rogalj/MolStrucClassifier.git
If you already installed the repository, you can update it by typing:
git pull
If you are an editor, check the following command as well.
If you make any changes and want to add directories/files to the GitHub repository, you can upload them by typing:
git add name_of_folder/file
The following command can check the status of the local repository. This command will show all the files that you added, deleted or modified which are not yet reflected to the local repository.
git status
When you upload the directories/files, you can also add a comment as following:
git commit -m "WHATEVER THE COMMENT YOU WANT"
This process will add to your local repository. If you want to add them to the global repository, type:
git push
At this point, your work will be on the GitHub repository.
If you made a mistake and added an unnecessary file, type:
git reset
to cancel the command you were working on.
Install LAMMPS code
In order to run the libraly programs in LAMMPS, you need to have the LAMMPS code. If you don’t need to connect with LAMMPS, you can skip this section. If you already have LAMMPS, make sure you connect the libraly and have the appropriate packages shown in Step 4.
Step 1
Go to the directory where you want to store the lammps code. I use mylammps as an example. Run the following command at mylammps directory to install LAMMPS from github.
Run the following command
git clone -b release https://github.com/lammps/lammps.git mylammps
If you want to add files (such as fix, compute files), add them in the src directory. You can also add directories in the src directory.
Step 2
The Next step is to get an appropriate module into your system:
module load intel/19.1.2
module load openmpi/intel/4.1.1
module load gcc/10.2.0
We also tested with Intel 17.0.0.098 and OpenMPI 2.1.6.
After you load modules, update files in the src directory by typing:
make mpi
Step 3
The next step is to run the LAMMPS code using symbolic links and soft links. (which enable you to run the program from different directories)
On GREENE, people usually make a link in a bin directory. Let’s make a bin directory if you don’t have one. (first go to /home/username/bin).
ln -s /home/username/mylammps/src/lmp_mpi lammps
Make another softlink that connects the LAMMPS code and the libraly files. Make a file in the libraly directory like Dafed_DK.inc. In this file you should have two things to connect the file. When you make this file, make sure to change the path to the appropriate name for your case.
DAFED_LOAD= /home/drk354/MolStrucClassifier/MolStrucClassifier/molvec_lib/ver2.0/libcv_nn.so -ldl
DAFED_DEPENDENCIES = /home/drk354/MolStrucClassifier/MolStrucClassifier/molvec_lib/ver2.0/libcv_nn.so
After this, go to the parent directory and type the following commands:
ln -s /home/drk354/MolStrucClassifier/MolStrucClassifier/molvec_lib/ver2.0/Dafed_DK.inc Dafed.inc
ln -s /home/drk354/MolStrucClassifier/MolStrucClassifier/molvec_lib/ver2.0/dafed_link.h Dafed.h
Step 4
You are almost there! The next step is to install some additional LAMMPS packages at the source directory. You will need the following packages:
name_of_package |
|---|
KSPACE |
MANYBODY |
MOLECULE |
EXTRA-MOLECULE |
If you use the old LAMMPS from 2019 or before, you also need to add package called ‘OPT’.
The command you need to type to get these additional packages is as folows:
make yes-name_of_package
You can check the status from typing:
make pi
Open Makefile.package and add the following to connect the source directory to the libraly directories:
PKG_LIB = $(DAFED_LOAD)
Open Makefile.package.setting and add the following line at the end of the file:
include ../../Dafed.inc
Install moltemplate
In order to run the libraly programs in LAMMPS, you need to have specific data file called .data. If you cannot find the dataset online, check the following section to create one. If you have a .data file with missing critical information (such as bonds, angles, dihedrals, improper angles information), you can use moltemplate to fill these. If you do not have .data file, refer to How to link with LAMMPS and run for more information.
Step 1
Run the following command to install Moltemplate:
git clone https://github.com/jewettaij/moltemplate moltemplate
Step 2
To create a symbolic link, navigate to the bin directory (/home/username/bin), as shown in Step 3 of Installing LAMMPS.
git clone https://github.com/jewettaij/moltemplate moltemplate
Step 3
Python3.7.2 or above is required to use for this program. On Greene, we have python 3.8.6 and you have to load this module as follows:
module load python/intel/3.8.6
Step 4
Navigate to the Moltemplate directory and finish the installation:
pip3 install .
Install Ogre
Ogre is a computational tool to create a surface with a vacuum region. In this project, I used this code to create surfaces.
Step 1
If you want to download the code, go to Ogre webpage and choose Ogre2.0.
Step 2
After downloading the program, make some minor changes if the code does not work. Go to /Ogre-master/ogre/utils and remove the following two lines:
from ogre.utils.surface_energy import convergence_plots
from ogre.utils.unique_planes import UniquePlanes
Next, go to /Ogre-master/ogre/utils and remove the following two lines:
from ogre.utils.surface_energy import convergence_plots
from ogre.utils.unique_planes import UniquePlanes
Go to /Ogre-master/ibslib/io and open check.py and change line 22 from:
extension2format = ase_extension2format`
to:
extension2format = deepcopy(ase_extension2format)
Using NYU GREENE
We will use NYU’s high-performance computer, GREENE, as an example. If you do not have access to the account, visit the NYU HPC account page to request access or download the appropriate softwares on your computer.