AngstromCube

Official Resources

• Source Repository: https://github.com/real-space/AngstromCube
• Documentation: https://github.com/real-space/AngstromCube (README)
• License: Open Source (Check repository for specific license)

Overview

AngstromCube is a high-performance Density Functional Theory (DFT) code designed for large-scale all-electron calculations. It employs a real-space finite-difference formulation, allowing for efficient parallelization and O(N) linear scaling. The code is specifically optimized for modern hardware, featuring native GPU acceleration to handle systems with thousands of atoms.

Scientific domain: Materials science, large-scale nanostructures, all-electron precision Target user community: Researchers needing high-accuracy all-electron calculations for large systems, HPC users

Theoretical Methods

• Density Functional Theory (DFT)
• Real-Space Finite Difference method
• All-Electron methodology (no pseudopotentials)
• Linear-scaling O(N) algorithms
• Real-time Time-Dependent DFT (RT-TDDFT) capability
• Chebyshev filtering for eigensolvers

Capabilities

• Ground-state electronic structure
• Large-scale system simulations (thousands of atoms)
• Massively parallel execution (MPI + GPU)
• All-electron precision for core states
• Efficient scale-up on supercomputers

Key Strengths

Real-Space Grid

• Systematic convergence via grid spacing
• No basis set superposition error (BSSE)
• Flexible boundary conditions

Performance

• GPU-accelerated (CUDA)
• Linear scaling for large systems
• High parallel efficiency

Inputs & Outputs

• Input: Configuration files (typically text/script based), atomic coordinates
• Output: Electronic density, total energy, forces, wavefunctions

Interfaces & Ecosystem

• Python Integration:
  • Native Python interface available for scripting and workflow control.
  • Can be integrated with ASE via custom calculators (not native yet).
• Data formats:
  • Supports standard cube file outputs for density/potentials.
  • Compatible with visualization tools like VESTA and VMD via standard formats.

Advanced Features

• Real-Time TDDFT:
  • Simulation of electron dynamics in real-time.
  • Suitable for strong field physics and optical response.
• Chebyshev Filtering:
  • Efficient eigensolver for interior eigenvalues.
  • Accelerates convergence for large basis sets.

Performance Characteristics

• Speed: High throughput on GPU systems.
• Scaling: O(N) linear scaling with system size.
• Parallelization: Hybrid MPI + GPU offloading.

Community and Support

• Development: Active on GitHub.
• Issues: Bug tracking via GitHub Issues.

Computational Cost

• Scaling: O(N) for large systems, reducing the cubic scaling bottleneck of traditional DFT.
• Hardware: Optimized for GPU architectures, reducing time-to-solution.

Verification & Sources

Primary sources:

1. GitHub Repository: https://github.com/real-space/AngstromCube
2. "AngstromCube: A parallel and GPU-accelerated code for Real-Space All-Electron Linear-Scaling Density Functional Theory"

Confidence: VERIFIED Status: Active Development Note: "AngstromCube" also refers to the cubic angstrom unit; ensure searches target the software repository.