Allegro

Official Resources

• Homepage: https://github.com/mir-group/allegro
• Documentation: https://github.com/mir-group/allegro
• Source Repository: https://github.com/mir-group/allegro
• License: MIT License

Overview

Allegro is a strictly local equivariant deep learning interatomic potential. It is built on the same principles as NequIP (E(3)-equivariance) but is designed to be strictly local (no message passing beyond a cutoff) and massively parallel. This allows it to scale to extremely large systems (millions of atoms) while maintaining high accuracy.

Scientific domain: Machine learning potentials, large-scale MD
Target user community: Researchers simulating very large systems (proteins, cracks, grain boundaries)

Capabilities (CRITICAL)

• Strict Locality: Interactions are strictly limited to a cutoff radius, enabling efficient parallelization.
• Equivariance: Uses tensor products for high accuracy.
• Scalability: Linear scaling with number of atoms, excellent strong scaling on GPUs.
• LAMMPS: Integration for large-scale MD.

Sources: Allegro GitHub, Nat. Commun. 14, 2038 (2023)

Inputs & Outputs

• Input formats: Training data (extxyz)
• Output data types: PyTorch models

Interfaces & Ecosystem

• NequIP: Share the same codebase/infrastructure.
• LAMMPS: Primary deployment target.

Workflow and Usage

1. Train model using nequip-train with Allegro config.
2. Deploy for LAMMPS.
3. Run large-scale MD.

Performance Characteristics

• Faster than message-passing networks for large systems.
• High parallel efficiency.

Application Areas

• Large-scale fracture simulations
• Biomacromolecules
• Electrolytes

Community and Support

• Developed by Kozinsky Group (Harvard)

Verification & Sources

Primary sources:

1. GitHub: https://github.com/mir-group/allegro
2. Publication: A. Musaelian et al., Nat. Commun. 14, 2038 (2023)

Confidence: VERIFIED

Verification status: ✅ VERIFIED

• Website: ACTIVE
• Documentation: AVAILABLE
• Source: OPEN (GitHub)
• Development: ACTIVE
• Applications: Large-scale equivariant MD