Deep Modeling for Molecular Simulation 2023
Hands-on tutorials for the in-person workshop, July 11 – 14, 2023
https://github.com/CSIprinceton/workshop-july-2023/tree/main/
Deep Modeling for Molecular Simulation 2022
Hands-on tutorials for the in-person workshop, July 7 – 8, 2022
https://github.com/CSIprinceton/workshop-july-2023/tree/main/
Molecular Simulation with Machine Learning 2020
Hands-on tutorials for the on-line workshop, July 13 -1 4, 2020
https://github.com/CSIprinceton/workshop-july-2020/tree/master/
DeePMD-kit
Aims
The goal of Deep Potential is to employ deep learning techniques and realize an interatomic potential energy model that is general, accurate, computationally efficient and scalable. The key component is to respect the extensive and symmetry-invariant properties of a potential energy model by assigning a local reference frame and a local environment to each atom. Each environment contains a finite number of atoms, whose local coordinates are arranged in a symmetry preserving way. These local coordinates are then transformed, through a sub-network, to a so-called atomic energy. Summing up all the atomic energies gives the potential energy of the system.
Objectives
By completing this tutorial, a user will be able to:
1. Use deep learning techniques to obtain an interatomic potential model from an ab-initio data set.
2. Use the deep potential to perform MD simulations.
Introduction
A. Prepare data (detail: https://github.com/deepmodeling/deepmd-kit#prepare-data )
B. Train a model (detail: https://github.com/deepmodeling/deepmd-kit#train-a-model )
C. Freeze the model (detail: https://github.com/deepmodeling/deepmd-kit#freeze-and-test-a-model )
D. MD runs with the model (detail: https://github.com/deepmodeling/deepmd-kit#run-md-with-lammps )
Further resources:
https://github.com/deepmodeling/deepmd-kit
CP Module of Quantum ESPRESSO (QE)
Aims
This tutorial aims at introducing Car-Parrinello (CP) molecular dynamics (MD) simulation using QE. It uses a prototypical example of an isolated water molecule in a cubic periodic cell.
Objectives
By completing this tutorial, a user will be able to:
1. Perform ground state calculation using second-order damped CP dynamics.
2. Run dynamical trajectories for the nuclei starting from the ground state calculation obtained in step 1.
Introduction
A. Preparation (using BASH, GIT, GNU compiler collection, and MPICH)
Compile cp.x binary
$ git clone git@gitlab.com:QEF/q-e.git; cd q-e; git checkout qe-6.4; ./configure; make cp; cd ..
Obtain example
$ git clone git@gitlab.com:kosinyj/csi-tutorial-qe-cp.git
B. Ground state calculation
$ cd single-water-molecule-cp-input-examples/01_emin
PBE with maximally-localized Wannier functions:
$ ./run_01.sh
PBE0 with maximally-localized Wannier functions:
$ ./run_02.sh
C. Car-Parrinello molecular dynamics simulation (NVT)
$ cd ../02_nvt
Single water molecule (300 K) single Nose-Hoover particle for 20 steps
$ ./run_01.sh
Further resources:
http://www.quantum-espresso.org/resources/users-manual
http://www.quantum-espresso.org/resources/tutorials