This project aims to develop a robust Scientific Machine Learning pipeline that leverages Kelvin-Helmholtz simulation data using the PIConGPU PIConGPU framework as input to train a predictive model. The Kelvin-Helmholtz instability is a key phenomenon in fluid dynamics, being able to record the relevant particle and radiation data outputted by the simulation. This then is passed on and transformed using Snakemake (Python based) to orchestrate the whole pipelining process.
This project was realized in collaboration with the HZDR (Helmholtz-Zentrum Dresden-Rossendorf) research lab and TU Dresden (Technische Universität Dresden).
- Technologies and ConceptsSnakemake, Python, PIConGPU, SLURM, HPC Cluster
- GithubSnakemake Pipeline Project link
Used resources
