This repository contains a collection of scripts and Python code developed by Efstratios Nikidis, a PhD student at the Computational Solid State Physics and Material Chemistry Lab (COSSPHY) at Aristotle University of Thessaloniki.
The scripts and Python code in this repository are used for various tasks related to GPU management, LAMMPS simulations, potential training, and file conversion. Each script or Python code is described in detail in its respective README file.
Details -> README
gpu_intro.sh: Activates a Conda environment, checks the version of the CUDA compiler, displays the status of NVIDIA GPUs, and counts the number of available GPUs.
nvtop.sh: Continuously displays the memory usage, total memory, percentage of memory used, and GPU utilization.
run_lammps.sh: Runs a LAMMPS simulation using a combination of MPI processes and OpenMP threads.
setup_lammps_aws.sh: Sets up the necessary tools for LAMMPS compilation on AWS EC2.
potential_train.sh: Trains a potential using the NequIP package on a CUDA-enabled node on AWS.
- converter.py: A Python code that modifies a given input file and writes the modified content to a new file. Details -> README
This directory contains scripts and resources related to elastic LAMMPS simulations. Details -> README
For any questions or concerns, please contact Efstratios Nikidis at enikidis@physcis.auth.gr.