Overview

mlx_atomistic is an experimental Apple Silicon-native atomistic simulation runtime built on MLX and Metal. It targets Python 3.13 through uv and uses MLX for local GPU execution on Apple Silicon.

Runtime Boundary

mlx_atomistic is the primary trajectory and electronic-structure runtime in this repository. OpenMM, LAMMPS, and source trees under vendors/ are reference or validation surfaces. They never replace the MLX runtime path.

Two Scales, One Runtime

• Density Functional Theory provides a legacy Gamma-point teaching and reference path plus a separate periodic PBE-GTH implementation with Monkhorst-Pack integration, block-Davidson solves, frozen-density bands, and periodic forces. Verified claims remain limited to published silicon, carbon, and magnesium-oxide workloads.
• Molecular Mechanics provides Lennard-Jones and Coulomb nonbonded work, bonded force fields, bounded Particle Mesh Ewald, constrained fixed-cell constant-particle-number, volume, and energy dynamics, and Langevin constant-particle-number, volume, and temperature dynamics.

Install the Alpha

uv run --no-project --python 3.13 --with mlx-atomistic \
  python -c "import mlx_atomistic as ma; print(ma.__version__)"

Checkout extras remain optional: prep for topology import, viz for visualization, and notebook for Jupyter.

uv venv --python 3.13
uv sync --extra notebook --extra prep --extra viz
uv run python -m ipykernel install --user \
  --name mlx-atomistic --display-name "mlx-atomistic"
uv run jupyter lab

If a sandbox cannot use the home cache, select a writable cache:

UV_CACHE_DIR=/tmp/mlx-atomistic-uv-cache \
  uv sync --extra notebook --extra prep --extra viz

Documentation Map

• Runtime boundaries
• Testing
• Molecular Mechanics
• Production MD
• MD acceleration
• Experimental Metal interaction engine
• DFT foundations
• DFT production core
• DFT roadmap
• Validation and performance
• Benchmarks
• Release checklist