Software and HPC Platforms

We apply a combination of open-source and licensed software, machine learning toolkits, and custom thermodynamic databases to support its integrated experimental and computational workflows:

Open-source software:

• CP2K, LAMMPS, VMD for molecular dynamics simulations, data analysis and virtualisation

• PHREEQC, GEM-Selektor for geochemical modelling and speciation analysis and mineral properties

Licensed software:

• VASP – for quantum mechanical (QM) calculations, ab initio molecular dynamics (AIMD), and machine learning molecular dynamics (MLMD)

• CRYSTAL – for QM calculations of crystalline materials

• Geochemist’s Workbench (GWB) and HCh – for advanced geochemical modelling and thermodynamic property calculations

Machine learning platforms

• DeePMD-kit – for developing deep learning-based interatomic potentials to enable accurate and efficient MD simulations

• AI2-kit – for integrating artificial intelligence with ab initio methods to accelerate computational chemistry and materials research

Self-maintained thermodynamic database

• Developed and maintained in GWB and HCh formats

• Transferable to PHREEQC and other widely used modelling platforms

Compute infrastructure

We run at scale on Pawsey (Setonix) and NCI (Gadi) with CSIRO HPC support, spanning CPU/GPU nodes and production schedulers (Slurm/PBS). We match problem scale to architecture—from rapid screening to long AIMD campaigns.