GPAW (TDDFT)

Official Resources

• Homepage: https://wiki.fysik.dtu.dk/gpaw/
• Documentation: https://wiki.fysik.dtu.dk/gpaw/documentation/documentation.html
• Source Repository: https://gitlab.com/gpaw/gpaw
• License: GNU General Public License v3.0 or later

Overview

GPAW is a density-functional theory Python code based on the projector-augmented wave (PAW) method. It combines the efficiency of real-space grids with the accuracy of plane-wave methods and integrates seamlessly with the Atomic Simulation Environment (ASE). GPAW is particularly strong for large-scale calculations, time-dependent DFT, and as a platform for method development.

Scientific domain: Real-space DFT, PAW method, Python-based calculations
Target user community: Researchers needing flexible, programmable DFT calculations, method developers

Theoretical Methods

• Kohn-Sham DFT (LDA, GGA, meta-GGA)
• Projector augmented wave (PAW) method
• Real-space grid representation
• Plane-wave mode available
• Finite-difference mode
• LCAO (linear combination of atomic orbitals) mode
• Plane-wave mode
• Linear combination of atomic orbitals (LCAO)
• Time-Dependent DFT (TDDFT)
• GW approximation
• Bethe-Salpeter Equation (BSE)
• Random Phase Approximation (RPA)
• Exact exchange and hybrid functionals
• Non-collinear magnetism and spin-orbit coupling

Capabilities (CRITICAL)

• Ground-state electronic structure calculations
• Geometry optimization and structure relaxation
• Molecular dynamics (NVE, NVT, NPT)
• Band structure and density of states
• Total energy calculations
• Forces and stress tensors
• Optical absorption spectra via TDDFT
• Quasiparticle energies via GW
• Optical spectra via BSE
• Linear response calculations (phonons, dielectric)
• Magnetic properties (moments, anisotropies)
• STM image simulation
• Electric field calculations
• Bader charge analysis
• Python scripting for workflows
• Real-time TDDFT propagation
• Delta-SCF for excited states
• Implicit solvation models

Sources: Official GPAW documentation, cited in 7/7 source lists

Inputs & Outputs

• Input formats:

  • Python scripts (native interface)
  • ASE Atoms objects
  • Standard structure files via ASE (CIF, XYZ, POSCAR, etc.)
  • GPAW restart files (.gpw)

• Output data types:

  • GPAW binary files (.gpw) with wavefunctions
  • Text output with energies, forces
  • Cube files for densities and orbitals
  • Eigenvalues and DOS
  • Trajectories in ASE format

Interfaces & Ecosystem

• Framework integrations:

  • ASE - native integration (GPAW built on ASE)
  • pymatgen - structure conversion
  • AiiDA - workflow automation possible
  • Fireworks - workflow integration

• Built-in tools:

  • TDDFT module for optical spectra
  • GW module for quasiparticles
  • BSE module for excitonic effects
  • Linear response for phonons
  • Response function calculations

• Post-processing:

  • gpaw-tools for common analysis
  • Python-based custom analysis
  • Integration with matplotlib, numpy, scipy

Limitations & Known Constraints

• Python overhead: Slower than pure Fortran/C++ codes for routine calculations
• Real-space grids: Require careful convergence testing for grid spacing
• Memory: Can be memory-intensive for large systems with fine grids
• Pseudopotentials: Limited to PAW datasets included with GPAW
• Parallelization: MPI + OpenMP supported but efficiency varies
• LCAO mode: Basis set quality depends on available atomic orbital sets
• GW/BSE: Computationally expensive; limited to smaller systems
• Documentation: Good but assumes Python/ASE familiarity
• Installation: Dependencies can be complex (libxc, BLAS/LAPACK, FFTW)

Computational Cost

• LCAO Mode: Very fast ($O(N)$), comparable to SIESTA.
• FD/PW Mode: Slower than VASP/QE for small systems due to Python overhead, but scales well for large grids.
• Memory: Real-space grids can consume significant RAM if grid spacing ($h$) is very small (< 0.15 Å).

Comparison with Other Codes

• vs VASP/QE: GPAW offers greater flexibility (3 modes: PW, FD, LCAO) and full Python scripting, but pure compute speed in PW mode is often lower than optimized Fortran codes.
• vs ASE: GPAW is the "reference" calculator for ASE and has the tightest integration.

Best Practices

• Mode Selection: Use LCAO for quick screening, then Finite Difference (FD) or Plane Wave (PW) for final high-accuracy numbers.
• Grid Spacing: Standard is $h=0.18-0.20$ Å. Refine to $0.15$ Å for hard potentials.
• Parallelization: Parallelize over k-points first, then domains.

Community and Support

• Mailing List: gpaw-users list is very active.
• Development: Hosted on GitLab; easy to contribute Python code.

Verification & Sources

Primary sources:

1. Official website: https://wiki.fysik.dtu.dk/gpaw/
2. Documentation: https://wiki.fysik.dtu.dk/gpaw/documentation/
3. GitLab repository: https://gitlab.com/gpaw/gpaw
4. J. Phys.: Condens. Matter 22, 253202 (2010) - GPAW code paper
5. J. Chem. Phys. 152, 124101 (2020) - GPAW LCAO developments

Secondary sources:

1. ASE documentation (GPAW calculator)
2. GPAW tutorials and workshops
3. Published applications using GPAW
4. Confirmed in 7/7 source lists (claude, g, gr, k, m, q, z)

Confidence: CONFIRMED - Appears in all 7 independent source lists

Verification status: ✅ VERIFIED

• Official homepage: ACCESSIBLE
• Documentation: COMPREHENSIVE and ACCESSIBLE
• Source code: OPEN (GitLab)
• Community support: Active (mailing list, GitLab issues)
• Academic citations: >1,000 (main papers)
• Active development: Regular releases