gpaw

Generate and manage GPAW Python-based DFT calculations. Use when the user requests GPAW, Python DFT, real-space grid DFT, or LCAO-DFT with ASE integration.

SKILL.md

GitHub

GPAW (Python DFT)

When to Use

  • User explicitly requests GPAW
  • User wants tight ASE integration (optimize with ASE, calculate with GPAW)
  • User needs real-space grid, LCAO, or plane-wave modes in a single code
  • User wants Python-scripted DFT workflows (no input files, pure Python)

Prerequisites

  1. GPAW + ASE installed on HPC (gpaw --version, python -c "import gpaw")
  2. PAW datasets installed (gpaw install-data)
  3. Structure loaded in viewer — verify with catgo_view(action="get_state")

Workflow Steps

1. Verify structure

catgo_view(action="get_state")

2. Create workflow

catgo_workflow_engine(action="create", params={"name": "GPAW PBE relaxation"})

3. Add GPAW task via shell script

CatGo does not yet have a native GPAW engine. Use task_type: "shell" with a Python script.

catgo_workflow_engine(action="add_task", params={
  "workflow_id": "wf_xxx",
  "task_type": "shell",
  "name": "gpaw_relax",
  "command": "python gpaw_relax.py",
  "input_files": {
    "gpaw_relax.py": "<script content>",
    "structure.json": "<pymatgen dict>"
  },
  "system_name": "TiO2_relax"
})

When a @register_engine("gpaw") is added to CatGo, use task_type: "geo_opt" with software: "gpaw" instead.

Script Template — SCF

from ase.io import read
from gpaw import GPAW, PW

atoms = read('structure.json')

calc = GPAW(
    mode=PW(500),            # Plane-wave mode, 500 eV cutoff
    xc='PBE',
    kpts={'density': 3.0},   # ~0.03 A^-1 k-point density
    txt='gpaw_scf.txt',
    occupations={'name': 'fermi-dirac', 'width': 0.05},
    convergence={'energy': 1e-5},
)

atoms.calc = calc
energy = atoms.get_potential_energy()
print(f'Total energy: {energy:.6f} eV')

Script Template — Relaxation

from ase.io import read, write
from ase.optimize import BFGS
from ase.constraints import FixAtoms
from gpaw import GPAW, PW

atoms = read('structure.json')

# Freeze bottom layers for slabs
c = FixAtoms(indices=[i for i, a in enumerate(atoms)
                      if a.position[2] < atoms.cell[2][2] * 0.4])
atoms.set_constraint(c)

calc = GPAW(
    mode=PW(500),
    xc='PBE',
    kpts={'density': 3.0},
    txt='gpaw_relax.txt',
    convergence={'energy': 1e-5},
)
atoms.calc = calc

opt = BFGS(atoms, trajectory='relax.traj', logfile='relax.log')
opt.run(fmax=0.02)

write('CONTCAR.vasp', atoms)

Parameter Guidance

Parameter Typical value Notes
mode PW(500) Plane-wave cutoff in eV; PW(600) for accurate forces
mode LCAO(dzp) LCAO mode for large systems (1000+ atoms)
xc 'PBE' Also: 'RPBE', 'BEEF-vdW', 'mBEEF'
kpts {'density': 3.0} Auto k-mesh; higher = denser
convergence {'energy': 1e-5} In eV; tighten for phonon calcs
occupations fermi-dirac, 0.05 Smearing width in eV
parallel {'domain': 2, 'band': 2} Domain decomposition for MPI

Calculation Modes

Mode Best for Speed
PW (plane-wave) Accurate bulk/surface Moderate
LCAO Large systems, screening Fast
FD (finite-difference) Real-space, nanostructures Slow but flexible

Common Pitfalls

  1. Forgetting txt parameter — without it, GPAW writes no log and debugging is impossible
  2. LCAO basis not installed — run gpaw install-data with --basis flag
  3. Memory for large PW calculations — GPAW PW mode stores wavefunctions in memory; use LCAO for >500 atoms
  4. No restart file — add calc.write('checkpoint.gpw') after SCF for restart capability
  5. Parallel decomposition mismatch — domain * band * kpt must equal total MPI ranks
  6. Slab k-points — use kpts={'size': (N, N, 1)} to avoid k-points along vacuum direction