class GPAW(Calculator, PAW):
"""This is the ASE-calculator frontend for doing a PAW calculation."""
implemented_properties = [
'energy', 'forces', 'stress', 'dipole', 'magmom', 'magmoms'
]
default_parameters = {
'mode': 'fd',
'xc': 'LDA',
'occupations': None,
'poissonsolver': None,
'h': None, # Angstrom
'gpts': None,
'kpts': [(0.0, 0.0, 0.0)],
'nbands': None,
'charge': 0,
'setups': {},
'basis': {},
'spinpol': None,
'fixdensity': False,
'filter': None,
'mixer': None,
'eigensolver': None,
'background_charge': None,
'external': None,
'random': False,
'hund': False,
'maxiter': 333,
'idiotproof': True,
'symmetry': {
'point_group': True,
'time_reversal': True,
'symmorphic': True,
'tolerance': 1e-7
},
'convergence': {
'energy': 0.0005, # eV / electron
'density': 1.0e-4,
'eigenstates': 4.0e-8, # eV^2
'bands': 'occupied',
'forces': np.inf
}, # eV / Ang
'dtype': None, # Deprecated
'width': None, # Deprecated
'verbose': 0
}
default_parallel = {
'kpt': None,
'domain': gpaw.parsize_domain,
'band': gpaw.parsize_bands,
'order': 'kdb',
'stridebands': False,
'augment_grids': gpaw.augment_grids,
'sl_auto': False,
'sl_default': gpaw.sl_default,
'sl_diagonalize': gpaw.sl_diagonalize,
'sl_inverse_cholesky': gpaw.sl_inverse_cholesky,
'sl_lcao': gpaw.sl_lcao,
'sl_lrtddft': gpaw.sl_lrtddft,
'buffer_size': gpaw.buffer_size
}
def __init__(self,
restart=None,
ignore_bad_restart_file=False,
label=None,
atoms=None,
timer=None,
communicator=None,
txt='-',
parallel=None,
**kwargs):
self.parallel = dict(self.default_parallel)
if parallel:
self.parallel.update(parallel)
if timer is None:
self.timer = Timer()
else:
self.timer = timer
self.scf = None
self.wfs = None
self.occupations = None
self.density = None
self.hamiltonian = None
self.observers = [] # XXX move to self.scf
self.initialized = False
self.world = communicator
if self.world is None:
self.world = mpi.world
elif not hasattr(self.world, 'new_communicator'):
self.world = mpi.world.new_communicator(np.asarray(self.world))
self.log = GPAWLogger(world=self.world)
self.log.fd = txt
self.reader = None
Calculator.__init__(self, restart, ignore_bad_restart_file, label,
atoms, **kwargs)
def __del__(self):
self.timer.write(self.log.fd)
if self.reader is not None:
self.reader.close()
def write(self, filename, mode=''):
self.log('Writing to {0} (mode={1!r})\n'.format(filename, mode))
writer = Writer(filename, self.world)
self._write(writer, mode)
writer.close()
self.world.barrier()
def _write(self, writer, mode):
from ase.io.trajectory import write_atoms
writer.write(version=1,
gpaw_version=gpaw.__version__,
ha=Ha,
bohr=Bohr)
write_atoms(writer.child('atoms'), self.atoms)
writer.child('results').write(**self.results)
writer.child('parameters').write(**self.todict())
self.density.write(writer.child('density'))
self.hamiltonian.write(writer.child('hamiltonian'))
self.occupations.write(writer.child('occupations'))
self.scf.write(writer.child('scf'))
self.wfs.write(writer.child('wave_functions'), mode == 'all')
return writer
def read(self, filename):
from ase.io.trajectory import read_atoms
self.log('Reading from {0}'.format(filename))
self.reader = reader = Reader(filename)
self.atoms = read_atoms(reader.atoms)
res = reader.results
self.results = dict((key, res.get(key)) for key in res.keys())
if self.results:
self.log('Read {0}'.format(', '.join(sorted(self.results))))
self.log('Reading input parameters:')
self.parameters = self.get_default_parameters()
dct = {}
for key, value in reader.parameters.asdict().items():
if (isinstance(value, dict)
and isinstance(self.parameters[key], dict)):
self.parameters[key].update(value)
else:
self.parameters[key] = value
dct[key] = self.parameters[key]
self.log.print_dict(dct)
self.log()
self.initialize(reading=True)
self.density.read(reader)
self.hamiltonian.read(reader)
self.occupations.read(reader)
self.scf.read(reader)
self.wfs.read(reader)
# We need to do this in a better way: XXX
from gpaw.utilities.partition import AtomPartition
atom_partition = AtomPartition(self.wfs.gd.comm,
np.zeros(len(self.atoms), dtype=int))
self.wfs.atom_partition = atom_partition
self.density.atom_partition = atom_partition
self.hamiltonian.atom_partition = atom_partition
spos_ac = self.atoms.get_scaled_positions() % 1.0
rank_a = self.density.gd.get_ranks_from_positions(spos_ac)
new_atom_partition = AtomPartition(self.density.gd.comm, rank_a)
for obj in [self.density, self.hamiltonian]:
obj.set_positions_without_ruining_everything(
spos_ac, new_atom_partition)
self.hamiltonian.xc.read(reader)
if self.hamiltonian.xc.name == 'GLLBSC':
# XXX GLLB: See lcao/tdgllbsc.py test
self.occupations.calculate(self.wfs)
return reader
def check_state(self, atoms, tol=1e-15):
system_changes = Calculator.check_state(self, atoms, tol)
if 'positions' not in system_changes:
if self.hamiltonian:
if self.hamiltonian.vext:
if self.hamiltonian.vext.vext_g is None:
# QMMM atoms have moved:
system_changes.append('positions')
return system_changes
def calculate(self,
atoms=None,
properties=['energy'],
system_changes=['cell']):
"""Calculate things."""
Calculator.calculate(self, atoms)
atoms = self.atoms
if system_changes:
self.log('System changes:', ', '.join(system_changes), '\n')
if system_changes == ['positions']:
# Only positions have changed:
self.density.reset()
else:
# Drastic changes:
self.wfs = None
self.occupations = None
self.density = None
self.hamiltonian = None
self.scf = None
self.initialize(atoms)
self.set_positions(atoms)
if not self.initialized:
self.initialize(atoms)
self.set_positions(atoms)
if not (self.wfs.positions_set and self.hamiltonian.positions_set):
self.set_positions(atoms)
if not self.scf.converged:
print_cell(self.wfs.gd, self.atoms.pbc, self.log)
with self.timer('SCF-cycle'):
self.scf.run(self.wfs, self.hamiltonian, self.density,
self.occupations, self.log, self.call_observers)
self.log('\nConverged after {0} iterations.\n'.format(
self.scf.niter))
e_free = self.hamiltonian.e_total_free
e_extrapolated = self.hamiltonian.e_total_extrapolated
self.results['energy'] = e_extrapolated * Ha
self.results['free_energy'] = e_free * Ha
if not self.atoms.pbc.all():
dipole_v = self.density.calculate_dipole_moment() * Bohr
self.log(
'Dipole moment: ({0:.6f}, {1:.6f}, {2:.6f}) |e|*Ang\n'.
format(*dipole_v))
self.results['dipole'] = dipole_v
if self.wfs.nspins == 2:
magmom = self.occupations.magmom
magmom_a = self.density.estimate_magnetic_moments(total=magmom)
self.log('Total magnetic moment: %f' % magmom)
self.log('Local magnetic moments:')
symbols = self.atoms.get_chemical_symbols()
for a, mom in enumerate(magmom_a):
self.log('{0:4} {1:2} {2:.6f}'.format(a, symbols[a], mom))
self.log()
self.results['magmom'] = self.occupations.magmom
self.results['magmoms'] = magmom_a
self.summary()
self.call_observers(self.scf.niter, final=True)
if 'forces' in properties:
with self.timer('Forces'):
F_av = calculate_forces(self.wfs, self.density,
self.hamiltonian, self.log)
self.results['forces'] = F_av * (Ha / Bohr)
if 'stress' in properties:
with self.timer('Stress'):
try:
stress = calculate_stress(self).flat[[0, 4, 8, 5, 2, 1]]
except NotImplementedError:
# Our ASE Calculator base class will raise
# PropertyNotImplementedError for us.
pass
else:
self.results['stress'] = stress * (Ha / Bohr**3)
def summary(self):
self.hamiltonian.summary(self.occupations.fermilevel, self.log)
self.density.summary(self.atoms, self.occupations.magmom, self.log)
self.occupations.summary(self.log)
self.wfs.summary(self.log)
self.log.fd.flush()
def set(self, **kwargs):
"""Change parameters for calculator.
Examples::
calc.set(xc='PBE')
calc.set(nbands=20, kpts=(4, 1, 1))
"""
changed_parameters = Calculator.set(self, **kwargs)
for key in ['setups', 'basis']:
if key in changed_parameters:
dct = changed_parameters[key]
if isinstance(dct, dict) and None in dct:
dct['default'] = dct.pop(None)
warnings.warn('Please use {key}={dct}'.format(key=key,
dct=dct))
# We need to handle txt early in order to get logging up and running:
if 'txt' in changed_parameters:
self.log.fd = changed_parameters.pop('txt')
if not changed_parameters:
return {}
self.initialized = False
self.scf = None
self.results = {}
self.log('Input parameters:')
self.log.print_dict(changed_parameters)
self.log()
for key in changed_parameters:
if key in ['eigensolver', 'convergence'] and self.wfs:
self.wfs.set_eigensolver(None)
if key in [
'mixer', 'verbose', 'txt', 'hund', 'random', 'eigensolver',
'idiotproof'
]:
continue
if key in ['convergence', 'fixdensity', 'maxiter']:
continue
# More drastic changes:
if self.wfs:
self.wfs.set_orthonormalized(False)
if key in ['external', 'xc', 'poissonsolver']:
self.hamiltonian = None
elif key in ['occupations', 'width']:
pass
elif key in ['charge', 'background_charge']:
self.hamiltonian = None
self.density = None
self.wfs = None
elif key in ['kpts', 'nbands', 'symmetry']:
self.wfs = None
elif key in ['h', 'gpts', 'setups', 'spinpol', 'dtype', 'mode']:
self.density = None
self.hamiltonian = None
self.wfs = None
elif key in ['basis']:
self.wfs = None
else:
raise TypeError('Unknown keyword argument: "%s"' % key)
def initialize_positions(self, atoms=None):
"""Update the positions of the atoms."""
self.log('Initializing position-dependent things.\n')
if atoms is None:
atoms = self.atoms
else:
# Save the state of the atoms:
self.atoms = atoms.copy()
mpi.synchronize_atoms(atoms, self.world)
spos_ac = atoms.get_scaled_positions() % 1.0
rank_a = self.wfs.gd.get_ranks_from_positions(spos_ac)
atom_partition = AtomPartition(self.wfs.gd.comm, rank_a, name='gd')
self.wfs.set_positions(spos_ac, atom_partition)
self.density.set_positions(spos_ac, atom_partition)
self.hamiltonian.set_positions(spos_ac, atom_partition)
return spos_ac
def set_positions(self, atoms=None):
"""Update the positions of the atoms and initialize wave functions."""
spos_ac = self.initialize_positions(atoms)
nlcao, nrand = self.wfs.initialize(self.density, self.hamiltonian,
spos_ac)
if nlcao + nrand:
self.log('Creating initial wave functions:')
if nlcao:
self.log(' ', plural(nlcao, 'band'), 'from LCAO basis set')
if nrand:
self.log(' ', plural(nrand, 'band'), 'from random numbers')
self.log()
self.wfs.eigensolver.reset()
self.scf.reset()
print_positions(self.atoms, self.log)
def initialize(self, atoms=None, reading=False):
"""Inexpensive initialization."""
self.log('Initialize ...\n')
if atoms is None:
atoms = self.atoms
else:
# Save the state of the atoms:
self.atoms = atoms.copy()
par = self.parameters
natoms = len(atoms)
cell_cv = atoms.get_cell() / Bohr
number_of_lattice_vectors = cell_cv.any(axis=1).sum()
if number_of_lattice_vectors < 3:
raise ValueError(
'GPAW requires 3 lattice vectors. Your system has {0}.'.
format(number_of_lattice_vectors))
pbc_c = atoms.get_pbc()
assert len(pbc_c) == 3
magmom_a = atoms.get_initial_magnetic_moments()
mpi.synchronize_atoms(atoms, self.world)
# Generate new xc functional only when it is reset by set
# XXX sounds like this should use the _changed_keywords dictionary.
if self.hamiltonian is None or self.hamiltonian.xc is None:
if isinstance(par.xc, basestring):
xc = XC(par.xc)
else:
xc = par.xc
else:
xc = self.hamiltonian.xc
mode = par.mode
if isinstance(mode, basestring):
mode = {'name': mode}
if isinstance(mode, dict):
mode = create_wave_function_mode(**mode)
if par.dtype == complex:
warnings.warn('Use mode={0}(..., force_complex_dtype=True) '
'instead of dtype=complex'.format(mode.name.upper()))
mode.force_complex_dtype = True
del par['dtype']
par.mode = mode
if xc.orbital_dependent and mode.name == 'lcao':
raise ValueError('LCAO mode does not support '
'orbital-dependent XC functionals.')
realspace = (mode.name != 'pw' and mode.interpolation != 'fft')
if not realspace:
pbc_c = np.ones(3, bool)
self.create_setups(mode, xc)
magnetic = magmom_a.any()
spinpol = par.spinpol
if par.hund:
if natoms != 1:
raise ValueError('hund=True arg only valid for single atoms!')
spinpol = True
magmom_a[0] = self.setups[0].get_hunds_rule_moment(par.charge)
if spinpol is None:
spinpol = magnetic
elif magnetic and not spinpol:
raise ValueError('Non-zero initial magnetic moment for a ' +
'spin-paired calculation!')
nspins = 1 + int(spinpol)
if spinpol:
self.log('Spin-polarized calculation.')
self.log('Magnetic moment: {0:.6f}\n'.format(magmom_a.sum()))
else:
self.log('Spin-paired calculation\n')
if isinstance(par.background_charge, dict):
background = create_background_charge(**par.background_charge)
else:
background = par.background_charge
nao = self.setups.nao
nvalence = self.setups.nvalence - par.charge
if par.background_charge is not None:
nvalence += background.charge
M = abs(magmom_a.sum())
nbands = par.nbands
orbital_free = any(setup.orbital_free for setup in self.setups)
if orbital_free:
nbands = 1
if isinstance(nbands, basestring):
if nbands[-1] == '%':
basebands = int(nvalence + M + 0.5) // 2
nbands = int((float(nbands[:-1]) / 100) * basebands)
else:
raise ValueError('Integer expected: Only use a string '
'if giving a percentage of occupied bands')
if nbands is None:
nbands = 0
for setup in self.setups:
nbands_from_atom = setup.get_default_nbands()
# Any obscure setup errors?
if nbands_from_atom < -(-setup.Nv // 2):
raise ValueError('Bad setup: This setup requests %d'
' bands but has %d electrons.' %
(nbands_from_atom, setup.Nv))
nbands += nbands_from_atom
nbands = min(nao, nbands)
elif nbands > nao and mode.name == 'lcao':
raise ValueError('Too many bands for LCAO calculation: '
'%d bands and only %d atomic orbitals!' %
(nbands, nao))
if nvalence < 0:
raise ValueError(
'Charge %f is not possible - not enough valence electrons' %
par.charge)
if nbands <= 0:
nbands = int(nvalence + M + 0.5) // 2 + (-nbands)
if nvalence > 2 * nbands and not orbital_free:
raise ValueError('Too few bands! Electrons: %f, bands: %d' %
(nvalence, nbands))
self.create_occupations(magmom_a.sum(), orbital_free)
if self.scf is None:
self.create_scf(nvalence, mode)
self.create_symmetry(magmom_a, cell_cv)
if not self.wfs:
self.create_wave_functions(mode, realspace, nspins, nbands, nao,
nvalence, self.setups, magmom_a,
cell_cv, pbc_c)
else:
self.wfs.set_setups(self.setups)
if not self.wfs.eigensolver:
self.create_eigensolver(xc, nbands, mode)
if self.density is None and not reading:
assert not par.fixdensity, 'No density to fix!'
olddens = None
if (self.density is not None and
(self.density.gd.parsize_c != self.wfs.gd.parsize_c).any()):
# Domain decomposition has changed, so we need to
# reinitialize density and hamiltonian:
if par.fixdensity:
olddens = self.density
self.density = None
self.hamiltonian = None
if self.density is None:
self.create_density(realspace, mode, background)
# XXXXXXXXXX if setups change, then setups.core_charge may change.
# But that parameter was supplied in Density constructor!
# This surely is a bug!
self.density.initialize(self.setups, self.timer, magmom_a, par.hund)
self.density.set_mixer(par.mixer)
if self.density.mixer.driver.name == 'dummy' or par.fixdensity:
self.log('No density mixing\n')
else:
self.log(self.density.mixer, '\n')
self.density.fixed = par.fixdensity
self.density.log = self.log
if olddens is not None:
self.density.initialize_from_other_density(olddens,
self.wfs.kptband_comm)
if self.hamiltonian is None:
self.create_hamiltonian(realspace, mode, xc)
xc.initialize(self.density, self.hamiltonian, self.wfs,
self.occupations)
if xc.name == 'GLLBSC' and olddens is not None:
xc.heeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeelp(olddens)
self.print_memory_estimate(maxdepth=memory_estimate_depth + 1)
print_parallelization_details(self.wfs, self.density, self.log)
self.log('Number of atoms:', natoms)
self.log('Number of atomic orbitals:', self.wfs.setups.nao)
if self.nbands_parallelization_adjustment != 0:
self.log(
'Adjusting number of bands by %+d to match parallelization' %
self.nbands_parallelization_adjustment)
self.log('Number of bands in calculation:', self.wfs.bd.nbands)
self.log('Bands to converge: ', end='')
n = par.convergence.get('bands', 'occupied')
if n == 'occupied':
self.log('occupied states only')
elif n == 'all':
self.log('all')
else:
self.log('%d lowest bands' % n)
self.log('Number of valence electrons:', self.wfs.nvalence)
self.log(flush=True)
self.timer.print_info(self)
if dry_run:
self.dry_run()
if (realspace and self.hamiltonian.poisson.get_description()
== 'FDTD+TDDFT'):
self.hamiltonian.poisson.set_density(self.density)
self.hamiltonian.poisson.print_messages(self.log)
self.log.fd.flush()
self.initialized = True
self.log('... initialized\n')
def create_setups(self, mode, xc):
if self.parameters.filter is None and mode.name != 'pw':
gamma = 1.6
N_c = self.parameters.get('gpts')
if N_c is None:
h = (self.parameters.h or 0.2) / Bohr
else:
icell_vc = np.linalg.inv(self.atoms.cell)
h = ((icell_vc**2).sum(0)**-0.5 / N_c).max() / Bohr
def filter(rgd, rcut, f_r, l=0):
gcut = np.pi / h - 2 / rcut / gamma
f_r[:] = rgd.filter(f_r, rcut * gamma, gcut, l)
else:
filter = self.parameters.filter
Z_a = self.atoms.get_atomic_numbers()
self.setups = Setups(Z_a, self.parameters.setups,
self.parameters.basis, xc, filter, self.world)
self.log(self.setups)
def create_grid_descriptor(self, N_c, cell_cv, pbc_c, domain_comm,
parsize_domain):
return GridDescriptor(N_c, cell_cv, pbc_c, domain_comm, parsize_domain)
def create_occupations(self, magmom, orbital_free):
occ = self.parameters.occupations
if occ is None:
if orbital_free:
occ = {'name': 'orbital-free'}
else:
width = self.parameters.width
if width is not None:
warnings.warn(
'Please use occupations=FermiDirac({0})'.format(width))
elif self.atoms.pbc.any():
width = 0.1 # eV
else:
width = 0.0
occ = {'name': 'fermi-dirac', 'width': width}
if isinstance(occ, dict):
occ = create_occupation_number_object(**occ)
if self.parameters.fixdensity:
occ.fixed_fermilevel = True
if self.occupations:
occ.fermilevel = self.occupations.fermilevel
self.occupations = occ
# If occupation numbers are changed, and we have wave functions,
# recalculate the occupation numbers
if self.wfs is not None:
self.occupations.calculate(self.wfs)
self.occupations.magmom = magmom
self.log(self.occupations)
def create_scf(self, nvalence, mode):
if mode.name == 'lcao':
niter_fixdensity = 0
else:
niter_fixdensity = 2
nv = max(nvalence, 1)
cc = self.parameters.convergence
self.scf = SCFLoop(
cc.get('eigenstates', 4.0e-8) / Ha**2 * nv,
cc.get('energy', 0.0005) / Ha * nv,
cc.get('density', 1.0e-4) * nv,
cc.get('forces', np.inf) / (Ha / Bohr), self.parameters.maxiter,
niter_fixdensity, nv)
self.log(self.scf)
def create_symmetry(self, magmom_a, cell_cv):
symm = self.parameters.symmetry
if symm == 'off':
symm = {'point_group': False, 'time_reversal': False}
m_a = magmom_a.round(decimals=3) # round off
id_a = list(zip(self.setups.id_a, m_a))
self.symmetry = Symmetry(id_a, cell_cv, self.atoms.pbc, **symm)
self.symmetry.analyze(self.atoms.get_scaled_positions())
self.setups.set_symmetry(self.symmetry)
def create_eigensolver(self, xc, nbands, mode):
# Number of bands to converge:
nbands_converge = self.parameters.convergence.get('bands', 'occupied')
if nbands_converge == 'all':
nbands_converge = nbands
elif nbands_converge != 'occupied':
assert isinstance(nbands_converge, int)
if nbands_converge < 0:
nbands_converge += nbands
eigensolver = get_eigensolver(self.parameters.eigensolver, mode,
self.parameters.convergence)
eigensolver.nbands_converge = nbands_converge
# XXX Eigensolver class doesn't define an nbands_converge property
if isinstance(xc, SIC):
eigensolver.blocksize = 1
self.wfs.set_eigensolver(eigensolver)
self.log(self.wfs.eigensolver, '\n')
def create_density(self, realspace, mode, background):
gd = self.wfs.gd
big_gd = gd.new_descriptor(comm=self.world)
# Check whether grid is too small. 8 is smallest admissible.
# (we decide this by how difficult it is to make the tests pass)
# (Actually it depends on stencils! But let the user deal with it)
N_c = big_gd.get_size_of_global_array(pad=True)
too_small = np.any(N_c / big_gd.parsize_c < 8)
if self.parallel['augment_grids'] and not too_small:
aux_gd = big_gd
else:
aux_gd = gd
redistributor = GridRedistributor(self.world, self.wfs.kptband_comm,
gd, aux_gd)
# Construct grid descriptor for fine grids for densities
# and potentials:
finegd = aux_gd.refine()
kwargs = dict(gd=gd,
finegd=finegd,
nspins=self.wfs.nspins,
charge=self.parameters.charge +
self.wfs.setups.core_charge,
redistributor=redistributor,
background_charge=background)
if realspace:
self.density = RealSpaceDensity(stencil=mode.interpolation,
**kwargs)
else:
self.density = pw.ReciprocalSpaceDensity(**kwargs)
self.log(self.density, '\n')
def create_hamiltonian(self, realspace, mode, xc):
dens = self.density
kwargs = dict(gd=dens.gd,
finegd=dens.finegd,
nspins=dens.nspins,
setups=dens.setups,
timer=self.timer,
xc=xc,
world=self.world,
redistributor=dens.redistributor,
vext=self.parameters.external,
psolver=self.parameters.poissonsolver)
if realspace:
self.hamiltonian = RealSpaceHamiltonian(stencil=mode.interpolation,
**kwargs)
xc.set_grid_descriptor(self.hamiltonian.finegd) # XXX
else:
self.hamiltonian = pw.ReciprocalSpaceHamiltonian(
pd2=dens.pd2, pd3=dens.pd3, realpbc_c=self.atoms.pbc, **kwargs)
xc.set_grid_descriptor(dens.xc_redistributor.aux_gd) # XXX
self.log(self.hamiltonian, '\n')
def create_wave_functions(self, mode, realspace, nspins, nbands, nao,
nvalence, setups, magmom_a, cell_cv, pbc_c):
par = self.parameters
bzkpts_kc = kpts2ndarray(par.kpts, self.atoms)
kd = KPointDescriptor(bzkpts_kc, nspins)
self.timer.start('Set symmetry')
kd.set_symmetry(self.atoms, self.symmetry, comm=self.world)
self.timer.stop('Set symmetry')
self.log(kd)
parallelization = mpi.Parallelization(self.world, nspins * kd.nibzkpts)
parsize_kpt = self.parallel['kpt']
parsize_domain = self.parallel['domain']
parsize_bands = self.parallel['band']
ndomains = None
if parsize_domain is not None:
ndomains = np.prod(parsize_domain)
if mode.name == 'pw':
if ndomains is not None and ndomains > 1:
raise ValueError('Planewave mode does not support '
'domain decomposition.')
ndomains = 1
parallelization.set(kpt=parsize_kpt,
domain=ndomains,
band=parsize_bands)
comms = parallelization.build_communicators()
domain_comm = comms['d']
kpt_comm = comms['k']
band_comm = comms['b']
kptband_comm = comms['D']
domainband_comm = comms['K']
self.comms = comms
if par.gpts is not None:
if par.h is not None:
raise ValueError("""You can't use both "gpts" and "h"!""")
N_c = np.array(par.gpts)
else:
h = par.h
if h is not None:
h /= Bohr
N_c = get_number_of_grid_points(cell_cv, h, mode, realspace,
kd.symmetry)
self.symmetry.check_grid(N_c)
kd.set_communicator(kpt_comm)
parstride_bands = self.parallel['stridebands']
# Unfortunately we need to remember that we adjusted the
# number of bands so we can print a warning if it differs
# from the number specified by the user. (The number can
# be inferred from the input parameters, but it's tricky
# because we allow negative numbers)
self.nbands_parallelization_adjustment = -nbands % band_comm.size
nbands += self.nbands_parallelization_adjustment
bd = BandDescriptor(nbands, band_comm, parstride_bands)
# Construct grid descriptor for coarse grids for wave functions:
gd = self.create_grid_descriptor(N_c, cell_cv, pbc_c, domain_comm,
parsize_domain)
if hasattr(self, 'time') or mode.force_complex_dtype:
dtype = complex
else:
if kd.gamma:
dtype = float
else:
dtype = complex
wfs_kwargs = dict(gd=gd,
nvalence=nvalence,
setups=setups,
bd=bd,
dtype=dtype,
world=self.world,
kd=kd,
kptband_comm=kptband_comm,
timer=self.timer)
if self.parallel['sl_auto']:
# Choose scalapack parallelization automatically
for key, val in self.parallel.items():
if (key.startswith('sl_') and key != 'sl_auto'
and val is not None):
raise ValueError("Cannot use 'sl_auto' together "
"with '%s'" % key)
max_scalapack_cpus = bd.comm.size * gd.comm.size
nprow = max_scalapack_cpus
npcol = 1
# Get a sort of reasonable number of columns/rows
while npcol < nprow and nprow % 2 == 0:
npcol *= 2
nprow //= 2
assert npcol * nprow == max_scalapack_cpus
# ScaLAPACK creates trouble if there aren't at least a few
# whole blocks; choose block size so there will always be
# several blocks. This will crash for small test systems,
# but so will ScaLAPACK in any case
blocksize = min(-(-nbands // 4), 64)
sl_default = (nprow, npcol, blocksize)
else:
sl_default = self.parallel['sl_default']
if mode.name == 'lcao':
# Layouts used for general diagonalizer
sl_lcao = self.parallel['sl_lcao']
if sl_lcao is None:
sl_lcao = sl_default
lcaoksl = get_KohnSham_layouts(sl_lcao,
'lcao',
gd,
bd,
domainband_comm,
dtype,
nao=nao,
timer=self.timer)
self.wfs = mode(lcaoksl, **wfs_kwargs)
elif mode.name == 'fd' or mode.name == 'pw':
# buffer_size keyword only relevant for fdpw
buffer_size = self.parallel['buffer_size']
# Layouts used for diagonalizer
sl_diagonalize = self.parallel['sl_diagonalize']
if sl_diagonalize is None:
sl_diagonalize = sl_default
diagksl = get_KohnSham_layouts(
sl_diagonalize,
'fd', # XXX
# choice of key 'fd' not so nice
gd,
bd,
domainband_comm,
dtype,
buffer_size=buffer_size,
timer=self.timer)
# Layouts used for orthonormalizer
sl_inverse_cholesky = self.parallel['sl_inverse_cholesky']
if sl_inverse_cholesky is None:
sl_inverse_cholesky = sl_default
if sl_inverse_cholesky != sl_diagonalize:
message = 'sl_inverse_cholesky != sl_diagonalize ' \
'is not implemented.'
raise NotImplementedError(message)
orthoksl = get_KohnSham_layouts(sl_inverse_cholesky,
'fd',
gd,
bd,
domainband_comm,
dtype,
buffer_size=buffer_size,
timer=self.timer)
# Use (at most) all available LCAO for initialization
lcaonbands = min(nbands, nao // band_comm.size * band_comm.size)
try:
lcaobd = BandDescriptor(lcaonbands, band_comm, parstride_bands)
except RuntimeError:
initksl = None
else:
# Layouts used for general diagonalizer
# (LCAO initialization)
sl_lcao = self.parallel['sl_lcao']
if sl_lcao is None:
sl_lcao = sl_default
initksl = get_KohnSham_layouts(sl_lcao,
'lcao',
gd,
lcaobd,
domainband_comm,
dtype,
nao=nao,
timer=self.timer)
self.wfs = mode(diagksl, orthoksl, initksl, **wfs_kwargs)
else:
self.wfs = mode(self, **wfs_kwargs)
self.log(self.wfs, '\n')
def dry_run(self):
# Can be overridden like in gpaw.atom.atompaw
print_cell(self.wfs.gd, self.atoms.pbc, self.log)
print_positions(self.atoms, self.log)
self.log.fd.flush()
raise SystemExit
def initialize(self, atoms=None):
"""Inexpensive initialization."""
if atoms is None:
atoms = self.atoms
else:
# Save the state of the atoms:
self.atoms = atoms.copy()
par = self.input_parameters
world = par.communicator
if world is None:
world = mpi.world
elif hasattr(world, 'new_communicator'):
# Check for whether object has correct type already
#
# Using isinstance() is complicated because of all the
# combinations, serial/parallel/debug...
pass
else:
# world should be a list of ranks:
world = mpi.world.new_communicator(np.asarray(world))
self.wfs.world = world
if 'txt' in self._changed_keywords:
self.set_txt(par.txt)
self.verbose = par.verbose
natoms = len(atoms)
cell_cv = atoms.get_cell() / Bohr
pbc_c = atoms.get_pbc()
Z_a = atoms.get_atomic_numbers()
magmom_av = atoms.get_initial_magnetic_moments()
self.check_atoms()
# Generate new xc functional only when it is reset by set
# XXX sounds like this should use the _changed_keywords dictionary.
if self.hamiltonian is None or self.hamiltonian.xc is None:
if isinstance(par.xc, str):
xc = XC(par.xc)
else:
xc = par.xc
else:
xc = self.hamiltonian.xc
mode = par.mode
if mode == 'fd':
mode = FD()
elif mode == 'pw':
mode = pw.PW()
elif mode == 'lcao':
mode = LCAO()
else:
assert hasattr(mode, 'name'), str(mode)
if xc.orbital_dependent and mode.name == 'lcao':
raise NotImplementedError('LCAO mode does not support '
'orbital-dependent XC functionals.')
if par.realspace is None:
realspace = (mode.name != 'pw')
else:
realspace = par.realspace
if mode.name == 'pw':
assert not realspace
if par.filter is None and mode.name != 'pw':
gamma = 1.6
if par.gpts is not None:
h = ((np.linalg.inv(cell_cv)**2).sum(0)**-0.5 / par.gpts).max()
else:
h = (par.h or 0.2) / Bohr
def filter(rgd, rcut, f_r, l=0):
gcut = np.pi / h - 2 / rcut / gamma
f_r[:] = rgd.filter(f_r, rcut * gamma, gcut, l)
else:
filter = par.filter
setups = Setups(Z_a, par.setups, par.basis, par.lmax, xc, filter,
world)
if magmom_av.ndim == 1:
collinear = True
magmom_av, magmom_a = np.zeros((natoms, 3)), magmom_av
magmom_av[:, 2] = magmom_a
else:
collinear = False
magnetic = magmom_av.any()
spinpol = par.spinpol
if par.hund:
if natoms != 1:
raise ValueError('hund=True arg only valid for single atoms!')
spinpol = True
magmom_av[0] = (0, 0, setups[0].get_hunds_rule_moment(par.charge))
if spinpol is None:
spinpol = magnetic
elif magnetic and not spinpol:
raise ValueError('Non-zero initial magnetic moment for a ' +
'spin-paired calculation!')
if collinear:
nspins = 1 + int(spinpol)
ncomp = 1
else:
nspins = 1
ncomp = 2
if par.usesymm != 'default':
warnings.warn('Use "symmetry" keyword instead of ' +
'"usesymm" keyword')
par.symmetry = usesymm2symmetry(par.usesymm)
symm = par.symmetry
if symm == 'off':
symm = {'point_group': False, 'time_reversal': False}
bzkpts_kc = kpts2ndarray(par.kpts, self.atoms)
kd = KPointDescriptor(bzkpts_kc, nspins, collinear)
m_av = magmom_av.round(decimals=3) # round off
id_a = zip(setups.id_a, *m_av.T)
symmetry = Symmetry(id_a, cell_cv, atoms.pbc, **symm)
kd.set_symmetry(atoms, symmetry, comm=world)
setups.set_symmetry(symmetry)
if par.gpts is not None:
N_c = np.array(par.gpts)
else:
h = par.h
if h is not None:
h /= Bohr
N_c = get_number_of_grid_points(cell_cv, h, mode, realspace,
kd.symmetry)
symmetry.check_grid(N_c)
width = par.width
if width is None:
if pbc_c.any():
width = 0.1 # eV
else:
width = 0.0
else:
assert par.occupations is None
if hasattr(self, 'time') or par.dtype == complex:
dtype = complex
else:
if kd.gamma:
dtype = float
else:
dtype = complex
nao = setups.nao
nvalence = setups.nvalence - par.charge
M_v = magmom_av.sum(0)
M = np.dot(M_v, M_v)**0.5
nbands = par.nbands
orbital_free = any(setup.orbital_free for setup in setups)
if orbital_free:
nbands = 1
if isinstance(nbands, basestring):
if nbands[-1] == '%':
basebands = int(nvalence + M + 0.5) // 2
nbands = int((float(nbands[:-1]) / 100) * basebands)
else:
raise ValueError('Integer Expected: Only use a string '
'if giving a percentage of occupied bands')
if nbands is None:
nbands = 0
for setup in setups:
nbands_from_atom = setup.get_default_nbands()
# Any obscure setup errors?
if nbands_from_atom < -(-setup.Nv // 2):
raise ValueError('Bad setup: This setup requests %d'
' bands but has %d electrons.' %
(nbands_from_atom, setup.Nv))
nbands += nbands_from_atom
nbands = min(nao, nbands)
elif nbands > nao and mode.name == 'lcao':
raise ValueError('Too many bands for LCAO calculation: '
'%d bands and only %d atomic orbitals!' %
(nbands, nao))
if nvalence < 0:
raise ValueError(
'Charge %f is not possible - not enough valence electrons' %
par.charge)
if nbands <= 0:
nbands = int(nvalence + M + 0.5) // 2 + (-nbands)
if nvalence > 2 * nbands and not orbital_free:
raise ValueError('Too few bands! Electrons: %f, bands: %d' %
(nvalence, nbands))
nbands *= ncomp
if par.width is not None:
self.text('**NOTE**: please start using '
'occupations=FermiDirac(width).')
if par.fixmom:
self.text('**NOTE**: please start using '
'occupations=FermiDirac(width, fixmagmom=True).')
if self.occupations is None:
if par.occupations is None:
# Create object for occupation numbers:
if orbital_free:
width = 0.0 # even for PBC
self.occupations = occupations.TFOccupations(
width, par.fixmom)
else:
self.occupations = occupations.FermiDirac(
width, par.fixmom)
else:
self.occupations = par.occupations
# If occupation numbers are changed, and we have wave functions,
# recalculate the occupation numbers
if self.wfs is not None and not isinstance(self.wfs,
EmptyWaveFunctions):
self.occupations.calculate(self.wfs)
self.occupations.magmom = M_v[2]
cc = par.convergence
if mode.name == 'lcao':
niter_fixdensity = 0
else:
niter_fixdensity = None
if self.scf is None:
force_crit = cc['forces']
if force_crit is not None:
force_crit /= Hartree / Bohr
self.scf = SCFLoop(cc['eigenstates'] / Hartree**2 * nvalence,
cc['energy'] / Hartree * max(nvalence, 1),
cc['density'] * nvalence, par.maxiter,
par.fixdensity, niter_fixdensity, force_crit)
parsize_kpt = par.parallel['kpt']
parsize_domain = par.parallel['domain']
parsize_bands = par.parallel['band']
if not realspace:
pbc_c = np.ones(3, bool)
if not self.wfs:
if parsize_domain == 'domain only': # XXX this was silly!
parsize_domain = world.size
parallelization = mpi.Parallelization(world, nspins * kd.nibzkpts)
ndomains = None
if parsize_domain is not None:
ndomains = np.prod(parsize_domain)
if mode.name == 'pw':
if ndomains > 1:
raise ValueError('Planewave mode does not support '
'domain decomposition.')
ndomains = 1
parallelization.set(kpt=parsize_kpt,
domain=ndomains,
band=parsize_bands)
comms = parallelization.build_communicators()
domain_comm = comms['d']
kpt_comm = comms['k']
band_comm = comms['b']
kptband_comm = comms['D']
domainband_comm = comms['K']
self.comms = comms
kd.set_communicator(kpt_comm)
parstride_bands = par.parallel['stridebands']
# Unfortunately we need to remember that we adjusted the
# number of bands so we can print a warning if it differs
# from the number specified by the user. (The number can
# be inferred from the input parameters, but it's tricky
# because we allow negative numbers)
self.nbands_parallelization_adjustment = -nbands % band_comm.size
nbands += self.nbands_parallelization_adjustment
# I would like to give the following error message, but apparently
# there are cases, e.g. gpaw/test/gw_ppa.py, which involve
# nbands > nao and are supposed to work that way.
#if nbands > nao:
# raise ValueError('Number of bands %d adjusted for band '
# 'parallelization %d exceeds number of atomic '
# 'orbitals %d. This problem can be fixed '
# 'by reducing the number of bands a bit.'
# % (nbands, band_comm.size, nao))
bd = BandDescriptor(nbands, band_comm, parstride_bands)
if (self.density is not None
and self.density.gd.comm.size != domain_comm.size):
# Domain decomposition has changed, so we need to
# reinitialize density and hamiltonian:
if par.fixdensity:
raise RuntimeError(
'Density reinitialization conflict ' +
'with "fixdensity" - specify domain decomposition.')
self.density = None
self.hamiltonian = None
# Construct grid descriptor for coarse grids for wave functions:
gd = self.grid_descriptor_class(N_c, cell_cv, pbc_c, domain_comm,
parsize_domain)
# do k-point analysis here? XXX
args = (gd, nvalence, setups, bd, dtype, world, kd, kptband_comm,
self.timer)
if par.parallel['sl_auto']:
# Choose scalapack parallelization automatically
for key, val in par.parallel.items():
if (key.startswith('sl_') and key != 'sl_auto'
and val is not None):
raise ValueError("Cannot use 'sl_auto' together "
"with '%s'" % key)
max_scalapack_cpus = bd.comm.size * gd.comm.size
nprow = max_scalapack_cpus
npcol = 1
# Get a sort of reasonable number of columns/rows
while npcol < nprow and nprow % 2 == 0:
npcol *= 2
nprow //= 2
assert npcol * nprow == max_scalapack_cpus
# ScaLAPACK creates trouble if there aren't at least a few
# whole blocks; choose block size so there will always be
# several blocks. This will crash for small test systems,
# but so will ScaLAPACK in any case
blocksize = min(-(-nbands // 4), 64)
sl_default = (nprow, npcol, blocksize)
else:
sl_default = par.parallel['sl_default']
if mode.name == 'lcao':
# Layouts used for general diagonalizer
sl_lcao = par.parallel['sl_lcao']
if sl_lcao is None:
sl_lcao = sl_default
lcaoksl = get_KohnSham_layouts(sl_lcao,
'lcao',
gd,
bd,
domainband_comm,
dtype,
nao=nao,
timer=self.timer)
self.wfs = mode(collinear, lcaoksl, *args)
elif mode.name == 'fd' or mode.name == 'pw':
# buffer_size keyword only relevant for fdpw
buffer_size = par.parallel['buffer_size']
# Layouts used for diagonalizer
sl_diagonalize = par.parallel['sl_diagonalize']
if sl_diagonalize is None:
sl_diagonalize = sl_default
diagksl = get_KohnSham_layouts(
sl_diagonalize,
'fd', # XXX
# choice of key 'fd' not so nice
gd,
bd,
domainband_comm,
dtype,
buffer_size=buffer_size,
timer=self.timer)
# Layouts used for orthonormalizer
sl_inverse_cholesky = par.parallel['sl_inverse_cholesky']
if sl_inverse_cholesky is None:
sl_inverse_cholesky = sl_default
if sl_inverse_cholesky != sl_diagonalize:
message = 'sl_inverse_cholesky != sl_diagonalize ' \
'is not implemented.'
raise NotImplementedError(message)
orthoksl = get_KohnSham_layouts(sl_inverse_cholesky,
'fd',
gd,
bd,
domainband_comm,
dtype,
buffer_size=buffer_size,
timer=self.timer)
# Use (at most) all available LCAO for initialization
lcaonbands = min(nbands, nao)
try:
lcaobd = BandDescriptor(lcaonbands, band_comm,
parstride_bands)
except RuntimeError:
initksl = None
else:
# Layouts used for general diagonalizer
# (LCAO initialization)
sl_lcao = par.parallel['sl_lcao']
if sl_lcao is None:
sl_lcao = sl_default
initksl = get_KohnSham_layouts(sl_lcao,
'lcao',
gd,
lcaobd,
domainband_comm,
dtype,
nao=nao,
timer=self.timer)
if hasattr(self, 'time'):
assert mode.name == 'fd'
from gpaw.tddft import TimeDependentWaveFunctions
self.wfs = TimeDependentWaveFunctions(
par.stencils[0], diagksl, orthoksl, initksl, gd,
nvalence, setups, bd, world, kd, kptband_comm,
self.timer)
elif mode.name == 'fd':
self.wfs = mode(par.stencils[0], diagksl, orthoksl,
initksl, *args)
else:
assert mode.name == 'pw'
self.wfs = mode(diagksl, orthoksl, initksl, *args)
else:
self.wfs = mode(self, *args)
else:
self.wfs.set_setups(setups)
if not self.wfs.eigensolver:
# Number of bands to converge:
nbands_converge = cc['bands']
if nbands_converge == 'all':
nbands_converge = nbands
elif nbands_converge != 'occupied':
assert isinstance(nbands_converge, int)
if nbands_converge < 0:
nbands_converge += nbands
eigensolver = get_eigensolver(par.eigensolver, mode,
par.convergence)
eigensolver.nbands_converge = nbands_converge
# XXX Eigensolver class doesn't define an nbands_converge property
if isinstance(xc, SIC):
eigensolver.blocksize = 1
self.wfs.set_eigensolver(eigensolver)
if self.density is None:
gd = self.wfs.gd
if par.stencils[1] != 9:
# Construct grid descriptor for fine grids for densities
# and potentials:
finegd = gd.refine()
else:
# Special case (use only coarse grid):
finegd = gd
if realspace:
self.density = RealSpaceDensity(
gd, finegd, nspins, par.charge + setups.core_charge,
collinear, par.stencils[1])
else:
self.density = pw.ReciprocalSpaceDensity(
gd, finegd, nspins, par.charge + setups.core_charge,
collinear)
self.density.initialize(setups, self.timer, magmom_av, par.hund)
self.density.set_mixer(par.mixer)
if self.hamiltonian is None:
gd, finegd = self.density.gd, self.density.finegd
if realspace:
self.hamiltonian = RealSpaceHamiltonian(
gd, finegd, nspins, setups, self.timer, xc, world,
self.wfs.kptband_comm, par.external, collinear,
par.poissonsolver, par.stencils[1])
else:
self.hamiltonian = pw.ReciprocalSpaceHamiltonian(
gd, finegd, self.density.pd2, self.density.pd3, nspins,
setups, self.timer, xc, world, self.wfs.kptband_comm,
par.external, collinear)
xc.initialize(self.density, self.hamiltonian, self.wfs,
self.occupations)
self.text()
self.print_memory_estimate(self.txt, maxdepth=memory_estimate_depth)
self.txt.flush()
self.timer.print_info(self)
if dry_run:
self.dry_run()
if realspace and \
self.hamiltonian.poisson.get_description() == 'FDTD+TDDFT':
self.hamiltonian.poisson.set_density(self.density)
self.hamiltonian.poisson.print_messages(self.text)
self.txt.flush()
self.initialized = True
self._changed_keywords.clear()
class GPAW(PAW, Calculator):
"""This is the ASE-calculator frontend for doing a PAW calculation."""
implemented_properties = [
'energy', 'forces', 'stress', 'dipole', 'magmom', 'magmoms'
]
default_parameters = {
'mode': 'fd',
'xc': 'LDA',
'occupations': None,
'poissonsolver': None,
'h': None, # Angstrom
'gpts': None,
'kpts': [(0.0, 0.0, 0.0)],
'nbands': None,
'charge': 0,
'setups': {},
'basis': {},
'spinpol': None,
'fixdensity': False,
'filter': None,
'mixer': None,
'eigensolver': None,
'background_charge': None,
'experimental': {
'reuse_wfs_method': 'paw',
'niter_fixdensity': 0,
'magmoms': None,
'soc': None,
'kpt_refine': None
},
'external': None,
'random': False,
'hund': False,
'maxiter': 333,
'idiotproof': True,
'symmetry': {
'point_group': True,
'time_reversal': True,
'symmorphic': True,
'tolerance': 1e-7,
'do_not_symmetrize_the_density': False
},
'convergence': {
'energy': 0.0005, # eV / electron
'density': 1.0e-4,
'eigenstates': 4.0e-8, # eV^2
'bands': 'occupied',
'forces': np.inf
}, # eV / Ang
'dtype': None, # Deprecated
'width': None, # Deprecated
'verbose': 0
}
default_parallel = {
'kpt': None,
'domain': gpaw.parsize_domain,
'band': gpaw.parsize_bands,
'order': 'kdb',
'stridebands': False,
'augment_grids': gpaw.augment_grids,
'sl_auto': False,
'sl_default': gpaw.sl_default,
'sl_diagonalize': gpaw.sl_diagonalize,
'sl_inverse_cholesky': gpaw.sl_inverse_cholesky,
'sl_lcao': gpaw.sl_lcao,
'sl_lrtddft': gpaw.sl_lrtddft,
'use_elpa': False,
'elpasolver': '2stage',
'buffer_size': gpaw.buffer_size
}
def __init__(self,
restart=None,
ignore_bad_restart_file=False,
label=None,
atoms=None,
timer=None,
communicator=None,
txt='-',
parallel=None,
**kwargs):
self.parallel = dict(self.default_parallel)
if parallel:
for key in parallel:
if key not in self.default_parallel:
allowed = ', '.join(list(self.default_parallel.keys()))
raise TypeError('Unexpected keyword "{}" in "parallel" '
'dictionary. Must be one of: {}'.format(
key, allowed))
self.parallel.update(parallel)
if timer is None:
self.timer = Timer()
else:
self.timer = timer
self.scf = None
self.wfs = None
self.occupations = None
self.density = None
self.hamiltonian = None
self.spos_ac = None # XXX store this in some better way.
self.observers = [] # XXX move to self.scf
self.initialized = False
self.world = communicator
if self.world is None:
self.world = mpi.world
elif not hasattr(self.world, 'new_communicator'):
self.world = mpi.world.new_communicator(np.asarray(self.world))
self.log = GPAWLogger(world=self.world)
self.log.fd = txt
self.reader = None
Calculator.__init__(self, restart, ignore_bad_restart_file, label,
atoms, **kwargs)
def __del__(self):
# Write timings and close reader if necessary.
# If we crashed in the constructor (e.g. a bad keyword), we may not
# have the normally expected attributes:
if hasattr(self, 'timer'):
self.timer.write(self.log.fd)
if hasattr(self, 'reader') and self.reader is not None:
self.reader.close()
def write(self, filename, mode=''):
self.log('Writing to {} (mode={!r})\n'.format(filename, mode))
writer = Writer(filename, self.world)
self._write(writer, mode)
writer.close()
self.world.barrier()
def _write(self, writer, mode):
from ase.io.trajectory import write_atoms
writer.write(version=1,
gpaw_version=gpaw.__version__,
ha=Ha,
bohr=Bohr)
write_atoms(writer.child('atoms'), self.atoms)
writer.child('results').write(**self.results)
writer.child('parameters').write(**self.todict())
self.density.write(writer.child('density'))
self.hamiltonian.write(writer.child('hamiltonian'))
self.occupations.write(writer.child('occupations'))
self.scf.write(writer.child('scf'))
self.wfs.write(writer.child('wave_functions'), mode == 'all')
return writer
def _set_atoms(self, atoms):
check_atoms_too_close(atoms)
self.atoms = atoms
# GPAW works in terms of the scaled positions. We want to
# extract the scaled positions in only one place, and that is
# here. No other place may recalculate them, or we might end up
# with rounding errors and inconsistencies.
self.spos_ac = atoms.get_scaled_positions() % 1.0
def read(self, filename):
from ase.io.trajectory import read_atoms
self.log('Reading from {}'.format(filename))
self.reader = reader = Reader(filename)
atoms = read_atoms(reader.atoms)
self._set_atoms(atoms)
res = reader.results
self.results = dict((key, res.get(key)) for key in res.keys())
if self.results:
self.log('Read {}'.format(', '.join(sorted(self.results))))
self.log('Reading input parameters:')
# XXX param
self.parameters = self.get_default_parameters()
dct = {}
for key, value in reader.parameters.asdict().items():
if (isinstance(value, dict)
and isinstance(self.parameters[key], dict)):
self.parameters[key].update(value)
else:
self.parameters[key] = value
dct[key] = self.parameters[key]
self.log.print_dict(dct)
self.log()
self.initialize(reading=True)
self.density.read(reader)
self.hamiltonian.read(reader)
self.occupations.read(reader)
self.scf.read(reader)
self.wfs.read(reader)
# We need to do this in a better way: XXX
from gpaw.utilities.partition import AtomPartition
atom_partition = AtomPartition(self.wfs.gd.comm,
np.zeros(len(self.atoms), dtype=int))
self.wfs.atom_partition = atom_partition
self.density.atom_partition = atom_partition
self.hamiltonian.atom_partition = atom_partition
rank_a = self.density.gd.get_ranks_from_positions(self.spos_ac)
new_atom_partition = AtomPartition(self.density.gd.comm, rank_a)
for obj in [self.density, self.hamiltonian]:
obj.set_positions_without_ruining_everything(
self.spos_ac, new_atom_partition)
self.hamiltonian.xc.read(reader)
if self.hamiltonian.xc.name == 'GLLBSC':
# XXX GLLB: See test/lcaotddft/gllbsc.py
self.occupations.calculate(self.wfs)
return reader
def check_state(self, atoms, tol=1e-15):
system_changes = Calculator.check_state(self, atoms, tol)
if 'positions' not in system_changes:
if self.hamiltonian:
if self.hamiltonian.vext:
if self.hamiltonian.vext.vext_g is None:
# QMMM atoms have moved:
system_changes.append('positions')
return system_changes
def calculate(self,
atoms=None,
properties=['energy'],
system_changes=['cell']):
"""Calculate things."""
Calculator.calculate(self, atoms)
atoms = self.atoms
if system_changes:
self.log('System changes:', ', '.join(system_changes), '\n')
if system_changes == ['positions']:
# Only positions have changed:
self.density.reset()
else:
# Drastic changes:
self.wfs = None
self.occupations = None
self.density = None
self.hamiltonian = None
self.scf = None
self.initialize(atoms)
self.set_positions(atoms)
if not self.initialized:
self.initialize(atoms)
self.set_positions(atoms)
if not (self.wfs.positions_set and self.hamiltonian.positions_set):
self.set_positions(atoms)
if not self.scf.converged:
print_cell(self.wfs.gd, self.atoms.pbc, self.log)
with self.timer('SCF-cycle'):
self.scf.run(self.wfs, self.hamiltonian, self.density,
self.occupations, self.log, self.call_observers)
self.log('\nConverged after {} iterations.\n'.format(
self.scf.niter))
e_free = self.hamiltonian.e_total_free
e_extrapolated = self.hamiltonian.e_total_extrapolated
self.results['energy'] = e_extrapolated * Ha
self.results['free_energy'] = e_free * Ha
dipole_v = self.density.calculate_dipole_moment() * Bohr
self.log(
'Dipole moment: ({:.6f}, {:.6f}, {:.6f}) |e|*Ang\n'.format(
*dipole_v))
self.results['dipole'] = dipole_v
if self.wfs.nspins == 2 or not self.density.collinear:
totmom_v, magmom_av = self.density.estimate_magnetic_moments()
self.log(
'Total magnetic moment: ({:.6f}, {:.6f}, {:.6f})'.format(
*totmom_v))
self.log('Local magnetic moments:')
symbols = self.atoms.get_chemical_symbols()
for a, mom_v in enumerate(magmom_av):
self.log('{:4} {:2} ({:9.6f}, {:9.6f}, {:9.6f})'.format(
a, symbols[a], *mom_v))
self.log()
self.results['magmom'] = self.occupations.magmom
self.results['magmoms'] = magmom_av[:, 2].copy()
self.summary()
self.call_observers(self.scf.niter, final=True)
if 'forces' in properties:
with self.timer('Forces'):
F_av = calculate_forces(self.wfs, self.density,
self.hamiltonian, self.log)
self.results['forces'] = F_av * (Ha / Bohr)
if 'stress' in properties:
with self.timer('Stress'):
try:
stress = calculate_stress(self).flat[[0, 4, 8, 5, 2, 1]]
except NotImplementedError:
# Our ASE Calculator base class will raise
# PropertyNotImplementedError for us.
pass
else:
self.results['stress'] = stress * (Ha / Bohr**3)
def summary(self):
efermi = self.occupations.fermilevel
self.hamiltonian.summary(efermi, self.log)
self.density.summary(self.atoms, self.occupations.magmom, self.log)
self.occupations.summary(self.log)
self.wfs.summary(self.log)
try:
bandgap(self, output=self.log.fd, efermi=efermi * Ha)
except ValueError:
pass
self.log.fd.flush()
def set(self, **kwargs):
"""Change parameters for calculator.
Examples::
calc.set(xc='PBE')
calc.set(nbands=20, kpts=(4, 1, 1))
"""
# Verify that keys are consistent with default ones.
for key in kwargs:
if key != 'txt' and key not in self.default_parameters:
raise TypeError('Unknown GPAW parameter: {}'.format(key))
if key in ['convergence', 'symmetry', 'experimental'
] and isinstance(kwargs[key], dict):
# For values that are dictionaries, verify subkeys, too.
default_dict = self.default_parameters[key]
for subkey in kwargs[key]:
if subkey not in default_dict:
allowed = ', '.join(list(default_dict.keys()))
raise TypeError('Unknown subkeyword "{}" of keyword '
'"{}". Must be one of: {}'.format(
subkey, key, allowed))
changed_parameters = Calculator.set(self, **kwargs)
for key in ['setups', 'basis']:
if key in changed_parameters:
dct = changed_parameters[key]
if isinstance(dct, dict) and None in dct:
dct['default'] = dct.pop(None)
warnings.warn('Please use {key}={dct}'.format(key=key,
dct=dct))
# We need to handle txt early in order to get logging up and running:
if 'txt' in changed_parameters:
self.log.fd = changed_parameters.pop('txt')
if not changed_parameters:
return {}
self.initialized = False
self.scf = None
self.results = {}
self.log('Input parameters:')
self.log.print_dict(changed_parameters)
self.log()
for key in changed_parameters:
if key in ['eigensolver', 'convergence'] and self.wfs:
self.wfs.set_eigensolver(None)
if key in [
'mixer', 'verbose', 'txt', 'hund', 'random', 'eigensolver',
'idiotproof'
]:
continue
if key in ['convergence', 'fixdensity', 'maxiter']:
continue
# Check nested arguments
if key in ['experimental']:
changed_parameters2 = changed_parameters[key]
for key2 in changed_parameters2:
if key2 in ['kpt_refine', 'magmoms', 'soc']:
self.wfs = None
elif key2 in ['reuse_wfs_method', 'niter_fixdensity']:
continue
else:
raise TypeError('Unknown keyword argument:', key2)
continue
# More drastic changes:
if self.wfs:
self.wfs.set_orthonormalized(False)
if key in ['external', 'xc', 'poissonsolver']:
self.hamiltonian = None
elif key in ['occupations', 'width']:
pass
elif key in ['charge', 'background_charge']:
self.hamiltonian = None
self.density = None
self.wfs = None
elif key in ['kpts', 'nbands', 'symmetry']:
self.wfs = None
elif key in ['h', 'gpts', 'setups', 'spinpol', 'dtype', 'mode']:
self.density = None
self.hamiltonian = None
self.wfs = None
elif key in ['basis']:
self.wfs = None
else:
raise TypeError('Unknown keyword argument: "%s"' % key)
def initialize_positions(self, atoms=None):
"""Update the positions of the atoms."""
self.log('Initializing position-dependent things.\n')
if atoms is None:
atoms = self.atoms
else:
atoms = atoms.copy()
self._set_atoms(atoms)
mpi.synchronize_atoms(atoms, self.world)
rank_a = self.wfs.gd.get_ranks_from_positions(self.spos_ac)
atom_partition = AtomPartition(self.wfs.gd.comm, rank_a, name='gd')
self.wfs.set_positions(self.spos_ac, atom_partition)
self.density.set_positions(self.spos_ac, atom_partition)
self.hamiltonian.set_positions(self.spos_ac, atom_partition)
def set_positions(self, atoms=None):
"""Update the positions of the atoms and initialize wave functions."""
self.initialize_positions(atoms)
nlcao, nrand = self.wfs.initialize(self.density, self.hamiltonian,
self.spos_ac)
if nlcao + nrand:
self.log('Creating initial wave functions:')
if nlcao:
self.log(' ', plural(nlcao, 'band'), 'from LCAO basis set')
if nrand:
self.log(' ', plural(nrand, 'band'), 'from random numbers')
self.log()
self.wfs.eigensolver.reset()
self.scf.reset()
print_positions(self.atoms, self.log, self.density.magmom_av)
def initialize(self, atoms=None, reading=False):
"""Inexpensive initialization."""
self.log('Initialize ...\n')
if atoms is None:
atoms = self.atoms
else:
atoms = atoms.copy()
self._set_atoms(atoms)
par = self.parameters
natoms = len(atoms)
cell_cv = atoms.get_cell() / Bohr
number_of_lattice_vectors = cell_cv.any(axis=1).sum()
if number_of_lattice_vectors < 3:
raise ValueError(
'GPAW requires 3 lattice vectors. Your system has {}.'.format(
number_of_lattice_vectors))
pbc_c = atoms.get_pbc()
assert len(pbc_c) == 3
magmom_a = atoms.get_initial_magnetic_moments()
if par.experimental.get('magmoms') is not None:
magmom_av = np.array(par.experimental['magmoms'], float)
collinear = False
else:
magmom_av = np.zeros((natoms, 3))
magmom_av[:, 2] = magmom_a
collinear = True
mpi.synchronize_atoms(atoms, self.world)
# Generate new xc functional only when it is reset by set
# XXX sounds like this should use the _changed_keywords dictionary.
if self.hamiltonian is None or self.hamiltonian.xc is None:
if isinstance(par.xc, (basestring, dict)):
xc = XC(par.xc, collinear=collinear, atoms=atoms)
else:
xc = par.xc
else:
xc = self.hamiltonian.xc
mode = par.mode
if isinstance(mode, basestring):
mode = {'name': mode}
if isinstance(mode, dict):
mode = create_wave_function_mode(**mode)
if par.dtype == complex:
warnings.warn('Use mode={}(..., force_complex_dtype=True) '
'instead of dtype=complex'.format(mode.name.upper()))
mode.force_complex_dtype = True
del par['dtype']
par.mode = mode
if xc.orbital_dependent and mode.name == 'lcao':
raise ValueError('LCAO mode does not support '
'orbital-dependent XC functionals.')
realspace = (mode.name != 'pw' and mode.interpolation != 'fft')
if not realspace:
pbc_c = np.ones(3, bool)
self.create_setups(mode, xc)
if par.hund:
if natoms != 1:
raise ValueError('hund=True arg only valid for single atoms!')
spinpol = True
magmom_av[0, 2] = self.setups[0].get_hunds_rule_moment(par.charge)
if collinear:
magnetic = magmom_av.any()
spinpol = par.spinpol
if spinpol is None:
spinpol = magnetic
elif magnetic and not spinpol:
raise ValueError('Non-zero initial magnetic moment for a ' +
'spin-paired calculation!')
nspins = 1 + int(spinpol)
if spinpol:
self.log('Spin-polarized calculation.')
self.log('Magnetic moment: {:.6f}\n'.format(magmom_av.sum()))
else:
self.log('Spin-paired calculation\n')
else:
nspins = 1
self.log('Non-collinear calculation.')
self.log('Magnetic moment: ({:.6f}, {:.6f}, {:.6f})\n'.format(
*magmom_av.sum(0)))
if isinstance(par.background_charge, dict):
background = create_background_charge(**par.background_charge)
else:
background = par.background_charge
nao = self.setups.nao
nvalence = self.setups.nvalence - par.charge
if par.background_charge is not None:
nvalence += background.charge
M = np.linalg.norm(magmom_av.sum(0))
nbands = par.nbands
orbital_free = any(setup.orbital_free for setup in self.setups)
if orbital_free:
nbands = 1
if isinstance(nbands, basestring):
if nbands == 'nao':
nbands = nao
elif nbands[-1] == '%':
basebands = (nvalence + M) / 2
nbands = int(np.ceil(float(nbands[:-1]) / 100 * basebands))
else:
raise ValueError('Integer expected: Only use a string '
'if giving a percentage of occupied bands')
if nbands is None:
# Number of bound partial waves:
nbandsmax = sum(setup.get_default_nbands()
for setup in self.setups)
nbands = int(np.ceil((1.2 * (nvalence + M) / 2))) + 4
if nbands > nbandsmax:
nbands = nbandsmax
if mode.name == 'lcao' and nbands > nao:
nbands = nao
elif nbands <= 0:
nbands = max(1, int(nvalence + M + 0.5) // 2 + (-nbands))
if nbands > nao and mode.name == 'lcao':
raise ValueError('Too many bands for LCAO calculation: '
'%d bands and only %d atomic orbitals!' %
(nbands, nao))
if nvalence < 0:
raise ValueError(
'Charge %f is not possible - not enough valence electrons' %
par.charge)
if nvalence > 2 * nbands and not orbital_free:
raise ValueError('Too few bands! Electrons: %f, bands: %d' %
(nvalence, nbands))
self.create_occupations(magmom_av[:, 2].sum(), orbital_free)
if self.scf is None:
self.create_scf(nvalence, mode)
self.create_symmetry(magmom_av, cell_cv)
if not collinear:
nbands *= 2
if not self.wfs:
self.create_wave_functions(mode, realspace, nspins, collinear,
nbands, nao, nvalence, self.setups,
cell_cv, pbc_c)
else:
self.wfs.set_setups(self.setups)
if not self.wfs.eigensolver:
self.create_eigensolver(xc, nbands, mode)
if self.density is None and not reading:
assert not par.fixdensity, 'No density to fix!'
olddens = None
if (self.density is not None and
(self.density.gd.parsize_c != self.wfs.gd.parsize_c).any()):
# Domain decomposition has changed, so we need to
# reinitialize density and hamiltonian:
if par.fixdensity:
olddens = self.density
self.density = None
self.hamiltonian = None
if self.density is None:
self.create_density(realspace, mode, background)
# XXXXXXXXXX if setups change, then setups.core_charge may change.
# But that parameter was supplied in Density constructor!
# This surely is a bug!
self.density.initialize(self.setups, self.timer, magmom_av, par.hund)
self.density.set_mixer(par.mixer)
if self.density.mixer.driver.name == 'dummy' or par.fixdensity:
self.log('No density mixing\n')
else:
self.log(self.density.mixer, '\n')
self.density.fixed = par.fixdensity
self.density.log = self.log
if olddens is not None:
self.density.initialize_from_other_density(olddens,
self.wfs.kptband_comm)
if self.hamiltonian is None:
self.create_hamiltonian(realspace, mode, xc)
xc.initialize(self.density, self.hamiltonian, self.wfs,
self.occupations)
description = xc.get_description()
if description is not None:
self.log('XC parameters: {}\n'.format('\n '.join(
description.splitlines())))
if xc.name == 'GLLBSC' and olddens is not None:
xc.heeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeelp(olddens)
self.print_memory_estimate(maxdepth=memory_estimate_depth + 1)
print_parallelization_details(self.wfs, self.hamiltonian, self.log)
self.log('Number of atoms:', natoms)
self.log('Number of atomic orbitals:', self.wfs.setups.nao)
self.log('Number of bands in calculation:', self.wfs.bd.nbands)
self.log('Bands to converge: ', end='')
n = par.convergence.get('bands', 'occupied')
if n == 'occupied':
self.log('occupied states only')
elif n == 'all':
self.log('all')
else:
self.log('%d lowest bands' % n)
self.log('Number of valence electrons:', self.wfs.nvalence)
self.log(flush=True)
self.timer.print_info(self)
if dry_run:
self.dry_run()
if (realspace and self.hamiltonian.poisson.get_description()
== 'FDTD+TDDFT'):
self.hamiltonian.poisson.set_density(self.density)
self.hamiltonian.poisson.print_messages(self.log)
self.log.fd.flush()
self.initialized = True
self.log('... initialized\n')
def create_setups(self, mode, xc):
if self.parameters.filter is None and mode.name != 'pw':
gamma = 1.6
N_c = self.parameters.get('gpts')
if N_c is None:
h = (self.parameters.h or 0.2) / Bohr
else:
icell_vc = np.linalg.inv(self.atoms.cell)
h = ((icell_vc**2).sum(0)**-0.5 / N_c).max() / Bohr
def filter(rgd, rcut, f_r, l=0):
gcut = np.pi / h - 2 / rcut / gamma
ftmp = rgd.filter(f_r, rcut * gamma, gcut, l)
f_r[:] = ftmp[:len(f_r)]
else:
filter = self.parameters.filter
Z_a = self.atoms.get_atomic_numbers()
self.setups = Setups(Z_a, self.parameters.setups,
self.parameters.basis, xc, filter, self.world)
self.log(self.setups)
def create_grid_descriptor(self, N_c, cell_cv, pbc_c, domain_comm,
parsize_domain):
return GridDescriptor(N_c, cell_cv, pbc_c, domain_comm, parsize_domain)
def create_occupations(self, magmom, orbital_free):
occ = self.parameters.occupations
if occ is None:
if orbital_free:
occ = {'name': 'orbital-free'}
else:
width = self.parameters.width
if width is not None:
warnings.warn(
'Please use occupations=FermiDirac({})'.format(width))
elif self.atoms.pbc.any():
width = 0.1 # eV
else:
width = 0.0
occ = {'name': 'fermi-dirac', 'width': width}
if isinstance(occ, dict):
occ = create_occupation_number_object(**occ)
if self.parameters.fixdensity:
occ.fixed_fermilevel = True
if self.occupations:
occ.fermilevel = self.occupations.fermilevel
self.occupations = occ
# If occupation numbers are changed, and we have wave functions,
# recalculate the occupation numbers
if self.wfs is not None:
self.occupations.calculate(self.wfs)
self.occupations.magmom = magmom
self.log(self.occupations)
def create_scf(self, nvalence, mode):
# if mode.name == 'lcao':
# niter_fixdensity = 0
# else:
# niter_fixdensity = 2
nv = max(nvalence, 1)
cc = self.parameters.convergence
self.scf = SCFLoop(
cc.get('eigenstates', 4.0e-8) / Ha**2 * nv,
cc.get('energy', 0.0005) / Ha * nv,
cc.get('density', 1.0e-4) * nv,
cc.get('forces', np.inf) / (Ha / Bohr),
self.parameters.maxiter,
# XXX make sure niter_fixdensity value is *always* set from default
# Subdictionary defaults seem to not be set when user provides
# e.g. {}. We should change that so it works like the ordinary
# parameters.
self.parameters.experimental.get('niter_fixdensity', 0),
nv)
self.log(self.scf)
def create_symmetry(self, magmom_av, cell_cv):
symm = self.parameters.symmetry
if symm == 'off':
symm = {'point_group': False, 'time_reversal': False}
m_av = magmom_av.round(decimals=3) # round off
id_a = [id + tuple(m_v) for id, m_v in zip(self.setups.id_a, m_av)]
self.symmetry = Symmetry(id_a, cell_cv, self.atoms.pbc, **symm)
self.symmetry.analyze(self.spos_ac)
self.setups.set_symmetry(self.symmetry)
def create_eigensolver(self, xc, nbands, mode):
# Number of bands to converge:
nbands_converge = self.parameters.convergence.get('bands', 'occupied')
if nbands_converge == 'all':
nbands_converge = nbands
elif nbands_converge != 'occupied':
assert isinstance(nbands_converge, int)
if nbands_converge < 0:
nbands_converge += nbands
eigensolver = get_eigensolver(self.parameters.eigensolver, mode,
self.parameters.convergence)
eigensolver.nbands_converge = nbands_converge
# XXX Eigensolver class doesn't define an nbands_converge property
if isinstance(xc, SIC):
eigensolver.blocksize = 1
self.wfs.set_eigensolver(eigensolver)
self.log('Eigensolver\n ', self.wfs.eigensolver, '\n')
def create_density(self, realspace, mode, background):
gd = self.wfs.gd
big_gd = gd.new_descriptor(comm=self.world)
# Check whether grid is too small. 8 is smallest admissible.
# (we decide this by how difficult it is to make the tests pass)
# (Actually it depends on stencils! But let the user deal with it)
N_c = big_gd.get_size_of_global_array(pad=True)
too_small = np.any(N_c / big_gd.parsize_c < 8)
if (self.parallel['augment_grids'] and not too_small
and mode.name != 'pw'):
aux_gd = big_gd
else:
aux_gd = gd
redistributor = GridRedistributor(self.world, self.wfs.kptband_comm,
gd, aux_gd)
# Construct grid descriptor for fine grids for densities
# and potentials:
finegd = aux_gd.refine()
kwargs = dict(gd=gd,
finegd=finegd,
nspins=self.wfs.nspins,
collinear=self.wfs.collinear,
charge=self.parameters.charge +
self.wfs.setups.core_charge,
redistributor=redistributor,
background_charge=background)
if realspace:
self.density = RealSpaceDensity(stencil=mode.interpolation,
**kwargs)
else:
self.density = pw.ReciprocalSpaceDensity(**kwargs)
self.log(self.density, '\n')
def create_hamiltonian(self, realspace, mode, xc):
dens = self.density
kwargs = dict(gd=dens.gd,
finegd=dens.finegd,
nspins=dens.nspins,
collinear=dens.collinear,
setups=dens.setups,
timer=self.timer,
xc=xc,
world=self.world,
redistributor=dens.redistributor,
vext=self.parameters.external,
psolver=self.parameters.poissonsolver)
if realspace:
self.hamiltonian = RealSpaceHamiltonian(stencil=mode.interpolation,
**kwargs)
xc.set_grid_descriptor(self.hamiltonian.finegd)
else:
# This code will work if dens.redistributor uses
# ordinary density.gd as aux_gd
gd = dens.finegd
xc_redist = None
if self.parallel['augment_grids']:
from gpaw.grid_descriptor import BadGridError
try:
aux_gd = gd.new_descriptor(comm=self.world)
except BadGridError as err:
import warnings
warnings.warn('Ignoring augment_grids: {}'.format(err))
else:
bcast_comm = dens.redistributor.broadcast_comm
xc_redist = GridRedistributor(self.world, bcast_comm, gd,
aux_gd)
self.hamiltonian = pw.ReciprocalSpaceHamiltonian(
pd2=dens.pd2,
pd3=dens.pd3,
realpbc_c=self.atoms.pbc,
xc_redistributor=xc_redist,
**kwargs)
xc.set_grid_descriptor(self.hamiltonian.xc_gd)
self.hamiltonian.soc = self.parameters.experimental.get('soc')
self.log(self.hamiltonian, '\n')
def create_kpoint_descriptor(self, nspins):
par = self.parameters
bzkpts_kc = kpts2ndarray(par.kpts, self.atoms)
kpt_refine = par.experimental.get('kpt_refine')
if kpt_refine is None:
kd = KPointDescriptor(bzkpts_kc, nspins)
self.timer.start('Set symmetry')
kd.set_symmetry(self.atoms, self.symmetry, comm=self.world)
self.timer.stop('Set symmetry')
else:
self.timer.start('Set k-point refinement')
kd = create_kpoint_descriptor_with_refinement(kpt_refine,
bzkpts_kc,
nspins,
self.atoms,
self.symmetry,
comm=self.world,
timer=self.timer)
self.timer.stop('Set k-point refinement')
# Update quantities which might have changed, if symmetry
# was changed
self.symmetry = kd.symmetry
self.setups.set_symmetry(kd.symmetry)
self.log(kd)
return kd
def create_wave_functions(self, mode, realspace, nspins, collinear, nbands,
nao, nvalence, setups, cell_cv, pbc_c):
par = self.parameters
kd = self.create_kpoint_descriptor(nspins)
parallelization = mpi.Parallelization(self.world, nspins * kd.nibzkpts)
parsize_kpt = self.parallel['kpt']
parsize_domain = self.parallel['domain']
parsize_bands = self.parallel['band']
ndomains = None
if parsize_domain is not None:
ndomains = np.prod(parsize_domain)
parallelization.set(kpt=parsize_kpt,
domain=ndomains,
band=parsize_bands)
comms = parallelization.build_communicators()
domain_comm = comms['d']
kpt_comm = comms['k']
band_comm = comms['b']
kptband_comm = comms['D']
domainband_comm = comms['K']
self.comms = comms
if par.gpts is not None:
if par.h is not None:
raise ValueError("""You can't use both "gpts" and "h"!""")
N_c = np.array(par.gpts)
else:
h = par.h
if h is not None:
h /= Bohr
N_c = get_number_of_grid_points(cell_cv, h, mode, realspace,
kd.symmetry)
self.symmetry.check_grid(N_c)
kd.set_communicator(kpt_comm)
parstride_bands = self.parallel['stridebands']
bd = BandDescriptor(nbands, band_comm, parstride_bands)
# Construct grid descriptor for coarse grids for wave functions:
gd = self.create_grid_descriptor(N_c, cell_cv, pbc_c, domain_comm,
parsize_domain)
if hasattr(self, 'time') or mode.force_complex_dtype or not collinear:
dtype = complex
else:
if kd.gamma:
dtype = float
else:
dtype = complex
wfs_kwargs = dict(gd=gd,
nvalence=nvalence,
setups=setups,
bd=bd,
dtype=dtype,
world=self.world,
kd=kd,
kptband_comm=kptband_comm,
timer=self.timer)
if self.parallel['sl_auto']:
# Choose scalapack parallelization automatically
for key, val in self.parallel.items():
if (key.startswith('sl_') and key != 'sl_auto'
and val is not None):
raise ValueError("Cannot use 'sl_auto' together "
"with '%s'" % key)
max_scalapack_cpus = bd.comm.size * gd.comm.size
sl_default = suggest_blocking(nbands, max_scalapack_cpus)
else:
sl_default = self.parallel['sl_default']
if mode.name == 'lcao':
assert collinear
# Layouts used for general diagonalizer
sl_lcao = self.parallel['sl_lcao']
if sl_lcao is None:
sl_lcao = sl_default
elpasolver = None
if self.parallel['use_elpa']:
elpasolver = self.parallel['elpasolver']
lcaoksl = get_KohnSham_layouts(sl_lcao,
'lcao',
gd,
bd,
domainband_comm,
dtype,
nao=nao,
timer=self.timer,
elpasolver=elpasolver)
self.wfs = mode(lcaoksl, **wfs_kwargs)
elif mode.name == 'fd' or mode.name == 'pw':
# Use (at most) all available LCAO for initialization
lcaonbands = min(nbands, nao)
try:
lcaobd = BandDescriptor(lcaonbands, band_comm, parstride_bands)
except RuntimeError:
initksl = None
else:
# Layouts used for general diagonalizer
# (LCAO initialization)
sl_lcao = self.parallel['sl_lcao']
if sl_lcao is None:
sl_lcao = sl_default
initksl = get_KohnSham_layouts(sl_lcao,
'lcao',
gd,
lcaobd,
domainband_comm,
dtype,
nao=nao,
timer=self.timer)
reuse_wfs_method = par.experimental.get('reuse_wfs_method', 'paw')
sl = (domainband_comm, ) + (self.parallel['sl_diagonalize']
or sl_default or (1, 1, None))
self.wfs = mode(sl,
initksl,
reuse_wfs_method=reuse_wfs_method,
collinear=collinear,
**wfs_kwargs)
else:
self.wfs = mode(self, collinear=collinear, **wfs_kwargs)
self.log(self.wfs, '\n')
def dry_run(self):
# Can be overridden like in gpaw.atom.atompaw
print_cell(self.wfs.gd, self.atoms.pbc, self.log)
print_positions(self.atoms, self.log, self.density.magmom_av)
self.log.fd.flush()
# Write timing info now before the interpreter shuts down:
self.__del__()
# Disable timing output during shut-down:
del self.timer
raise SystemExit
def initialize(self, atoms=None):
"""Inexpensive initialization."""
if atoms is None:
atoms = self.atoms
else:
# Save the state of the atoms:
self.atoms = atoms.copy()
par = self.input_parameters
world = par.communicator
if world is None:
world = mpi.world
elif hasattr(world, 'new_communicator'):
# Check for whether object has correct type already
#
# Using isinstance() is complicated because of all the
# combinations, serial/parallel/debug...
pass
else:
# world should be a list of ranks:
world = mpi.world.new_communicator(np.asarray(world))
self.wfs.world = world
if 'txt' in self._changed_keywords:
self.set_txt(par.txt)
self.verbose = par.verbose
natoms = len(atoms)
cell_cv = atoms.get_cell() / Bohr
pbc_c = atoms.get_pbc()
Z_a = atoms.get_atomic_numbers()
magmom_av = atoms.get_initial_magnetic_moments()
self.check_atoms()
# Generate new xc functional only when it is reset by set
# XXX sounds like this should use the _changed_keywords dictionary.
if self.hamiltonian is None or self.hamiltonian.xc is None:
if isinstance(par.xc, str):
xc = XC(par.xc)
else:
xc = par.xc
else:
xc = self.hamiltonian.xc
mode = par.mode
if mode == 'fd':
mode = FD()
elif mode == 'pw':
mode = pw.PW()
elif mode == 'lcao':
mode = LCAO()
else:
assert hasattr(mode, 'name'), str(mode)
if xc.orbital_dependent and mode.name == 'lcao':
raise NotImplementedError('LCAO mode does not support '
'orbital-dependent XC functionals.')
if par.realspace is None:
realspace = (mode.name != 'pw')
else:
realspace = par.realspace
if mode.name == 'pw':
assert not realspace
if par.filter is None and mode.name != 'pw':
gamma = 1.6
if par.gpts is not None:
h = ((np.linalg.inv(cell_cv)**2).sum(0)**-0.5
/ par.gpts).max()
else:
h = (par.h or 0.2) / Bohr
def filter(rgd, rcut, f_r, l=0):
gcut = np.pi / h - 2 / rcut / gamma
f_r[:] = rgd.filter(f_r, rcut * gamma, gcut, l)
else:
filter = par.filter
setups = Setups(Z_a, par.setups, par.basis, par.lmax, xc,
filter, world)
if magmom_av.ndim == 1:
collinear = True
magmom_av, magmom_a = np.zeros((natoms, 3)), magmom_av
magmom_av[:, 2] = magmom_a
else:
collinear = False
magnetic = magmom_av.any()
spinpol = par.spinpol
if par.hund:
if natoms != 1:
raise ValueError('hund=True arg only valid for single atoms!')
spinpol = True
magmom_av[0] = (0, 0, setups[0].get_hunds_rule_moment(par.charge))
if spinpol is None:
spinpol = magnetic
elif magnetic and not spinpol:
raise ValueError('Non-zero initial magnetic moment for a ' +
'spin-paired calculation!')
if collinear:
nspins = 1 + int(spinpol)
ncomp = 1
else:
nspins = 1
ncomp = 2
if par.usesymm != 'default':
warnings.warn('Use "symmetry" keyword instead of ' +
'"usesymm" keyword')
par.symmetry = usesymm2symmetry(par.usesymm)
symm = par.symmetry
if symm == 'off':
symm = {'point_group': False, 'time_reversal': False}
bzkpts_kc = kpts2ndarray(par.kpts, self.atoms)
kd = KPointDescriptor(bzkpts_kc, nspins, collinear)
m_av = magmom_av.round(decimals=3) # round off
id_a = zip(setups.id_a, *m_av.T)
symmetry = Symmetry(id_a, cell_cv, atoms.pbc, **symm)
kd.set_symmetry(atoms, symmetry, comm=world)
setups.set_symmetry(symmetry)
if par.gpts is not None:
N_c = np.array(par.gpts)
else:
h = par.h
if h is not None:
h /= Bohr
N_c = get_number_of_grid_points(cell_cv, h, mode, realspace,
kd.symmetry)
symmetry.check_grid(N_c)
width = par.width
if width is None:
if pbc_c.any():
width = 0.1 # eV
else:
width = 0.0
else:
assert par.occupations is None
if hasattr(self, 'time') or par.dtype == complex:
dtype = complex
else:
if kd.gamma:
dtype = float
else:
dtype = complex
nao = setups.nao
nvalence = setups.nvalence - par.charge
M_v = magmom_av.sum(0)
M = np.dot(M_v, M_v) ** 0.5
nbands = par.nbands
orbital_free = any(setup.orbital_free for setup in setups)
if orbital_free:
nbands = 1
if isinstance(nbands, basestring):
if nbands[-1] == '%':
basebands = int(nvalence + M + 0.5) // 2
nbands = int((float(nbands[:-1]) / 100) * basebands)
else:
raise ValueError('Integer Expected: Only use a string '
'if giving a percentage of occupied bands')
if nbands is None:
nbands = 0
for setup in setups:
nbands_from_atom = setup.get_default_nbands()
# Any obscure setup errors?
if nbands_from_atom < -(-setup.Nv // 2):
raise ValueError('Bad setup: This setup requests %d'
' bands but has %d electrons.'
% (nbands_from_atom, setup.Nv))
nbands += nbands_from_atom
nbands = min(nao, nbands)
elif nbands > nao and mode.name == 'lcao':
raise ValueError('Too many bands for LCAO calculation: '
'%d bands and only %d atomic orbitals!' %
(nbands, nao))
if nvalence < 0:
raise ValueError(
'Charge %f is not possible - not enough valence electrons' %
par.charge)
if nbands <= 0:
nbands = int(nvalence + M + 0.5) // 2 + (-nbands)
if nvalence > 2 * nbands and not orbital_free:
raise ValueError('Too few bands! Electrons: %f, bands: %d'
% (nvalence, nbands))
nbands *= ncomp
if par.width is not None:
self.text('**NOTE**: please start using '
'occupations=FermiDirac(width).')
if par.fixmom:
self.text('**NOTE**: please start using '
'occupations=FermiDirac(width, fixmagmom=True).')
if self.occupations is None:
if par.occupations is None:
# Create object for occupation numbers:
if orbital_free:
width = 0.0 # even for PBC
self.occupations = occupations.TFOccupations(width,
par.fixmom)
else:
self.occupations = occupations.FermiDirac(width,
par.fixmom)
else:
self.occupations = par.occupations
# If occupation numbers are changed, and we have wave functions,
# recalculate the occupation numbers
if self.wfs is not None and not isinstance(
self.wfs,
EmptyWaveFunctions):
self.occupations.calculate(self.wfs)
self.occupations.magmom = M_v[2]
cc = par.convergence
if mode.name == 'lcao':
niter_fixdensity = 0
else:
niter_fixdensity = None
if self.scf is None:
force_crit = cc['forces']
if force_crit is not None:
force_crit /= Hartree / Bohr
self.scf = SCFLoop(
cc['eigenstates'] / Hartree**2 * nvalence,
cc['energy'] / Hartree * max(nvalence, 1),
cc['density'] * nvalence,
par.maxiter, par.fixdensity,
niter_fixdensity,
force_crit)
parsize_kpt = par.parallel['kpt']
parsize_domain = par.parallel['domain']
parsize_bands = par.parallel['band']
if not realspace:
pbc_c = np.ones(3, bool)
if not self.wfs:
if parsize_domain == 'domain only': # XXX this was silly!
parsize_domain = world.size
parallelization = mpi.Parallelization(world,
nspins * kd.nibzkpts)
ndomains = None
if parsize_domain is not None:
ndomains = np.prod(parsize_domain)
if mode.name == 'pw':
if ndomains > 1:
raise ValueError('Planewave mode does not support '
'domain decomposition.')
ndomains = 1
parallelization.set(kpt=parsize_kpt,
domain=ndomains,
band=parsize_bands)
comms = parallelization.build_communicators()
domain_comm = comms['d']
kpt_comm = comms['k']
band_comm = comms['b']
kptband_comm = comms['D']
domainband_comm = comms['K']
self.comms = comms
kd.set_communicator(kpt_comm)
parstride_bands = par.parallel['stridebands']
# Unfortunately we need to remember that we adjusted the
# number of bands so we can print a warning if it differs
# from the number specified by the user. (The number can
# be inferred from the input parameters, but it's tricky
# because we allow negative numbers)
self.nbands_parallelization_adjustment = -nbands % band_comm.size
nbands += self.nbands_parallelization_adjustment
# I would like to give the following error message, but apparently
# there are cases, e.g. gpaw/test/gw_ppa.py, which involve
# nbands > nao and are supposed to work that way.
#if nbands > nao:
# raise ValueError('Number of bands %d adjusted for band '
# 'parallelization %d exceeds number of atomic '
# 'orbitals %d. This problem can be fixed '
# 'by reducing the number of bands a bit.'
# % (nbands, band_comm.size, nao))
bd = BandDescriptor(nbands, band_comm, parstride_bands)
if (self.density is not None and
self.density.gd.comm.size != domain_comm.size):
# Domain decomposition has changed, so we need to
# reinitialize density and hamiltonian:
if par.fixdensity:
raise RuntimeError(
'Density reinitialization conflict ' +
'with "fixdensity" - specify domain decomposition.')
self.density = None
self.hamiltonian = None
# Construct grid descriptor for coarse grids for wave functions:
gd = self.grid_descriptor_class(N_c, cell_cv, pbc_c,
domain_comm, parsize_domain)
# do k-point analysis here? XXX
args = (gd, nvalence, setups, bd, dtype, world, kd,
kptband_comm, self.timer)
if par.parallel['sl_auto']:
# Choose scalapack parallelization automatically
for key, val in par.parallel.items():
if (key.startswith('sl_') and key != 'sl_auto'
and val is not None):
raise ValueError("Cannot use 'sl_auto' together "
"with '%s'" % key)
max_scalapack_cpus = bd.comm.size * gd.comm.size
nprow = max_scalapack_cpus
npcol = 1
# Get a sort of reasonable number of columns/rows
while npcol < nprow and nprow % 2 == 0:
npcol *= 2
nprow //= 2
assert npcol * nprow == max_scalapack_cpus
# ScaLAPACK creates trouble if there aren't at least a few
# whole blocks; choose block size so there will always be
# several blocks. This will crash for small test systems,
# but so will ScaLAPACK in any case
blocksize = min(-(-nbands // 4), 64)
sl_default = (nprow, npcol, blocksize)
else:
sl_default = par.parallel['sl_default']
if mode.name == 'lcao':
# Layouts used for general diagonalizer
sl_lcao = par.parallel['sl_lcao']
if sl_lcao is None:
sl_lcao = sl_default
lcaoksl = get_KohnSham_layouts(sl_lcao, 'lcao',
gd, bd, domainband_comm, dtype,
nao=nao, timer=self.timer)
self.wfs = mode(collinear, lcaoksl, *args)
elif mode.name == 'fd' or mode.name == 'pw':
# buffer_size keyword only relevant for fdpw
buffer_size = par.parallel['buffer_size']
# Layouts used for diagonalizer
sl_diagonalize = par.parallel['sl_diagonalize']
if sl_diagonalize is None:
sl_diagonalize = sl_default
diagksl = get_KohnSham_layouts(sl_diagonalize, 'fd', # XXX
# choice of key 'fd' not so nice
gd, bd, domainband_comm, dtype,
buffer_size=buffer_size,
timer=self.timer)
# Layouts used for orthonormalizer
sl_inverse_cholesky = par.parallel['sl_inverse_cholesky']
if sl_inverse_cholesky is None:
sl_inverse_cholesky = sl_default
if sl_inverse_cholesky != sl_diagonalize:
message = 'sl_inverse_cholesky != sl_diagonalize ' \
'is not implemented.'
raise NotImplementedError(message)
orthoksl = get_KohnSham_layouts(sl_inverse_cholesky, 'fd',
gd, bd, domainband_comm, dtype,
buffer_size=buffer_size,
timer=self.timer)
# Use (at most) all available LCAO for initialization
lcaonbands = min(nbands, nao)
try:
lcaobd = BandDescriptor(lcaonbands, band_comm,
parstride_bands)
except RuntimeError:
initksl = None
else:
# Layouts used for general diagonalizer
# (LCAO initialization)
sl_lcao = par.parallel['sl_lcao']
if sl_lcao is None:
sl_lcao = sl_default
initksl = get_KohnSham_layouts(sl_lcao, 'lcao',
gd, lcaobd, domainband_comm,
dtype, nao=nao,
timer=self.timer)
if hasattr(self, 'time'):
assert mode.name == 'fd'
from gpaw.tddft import TimeDependentWaveFunctions
self.wfs = TimeDependentWaveFunctions(
par.stencils[0],
diagksl,
orthoksl,
initksl,
gd,
nvalence,
setups,
bd,
world,
kd,
kptband_comm,
self.timer)
elif mode.name == 'fd':
self.wfs = mode(par.stencils[0], diagksl,
orthoksl, initksl, *args)
else:
assert mode.name == 'pw'
self.wfs = mode(diagksl, orthoksl, initksl, *args)
else:
self.wfs = mode(self, *args)
else:
self.wfs.set_setups(setups)
if not self.wfs.eigensolver:
# Number of bands to converge:
nbands_converge = cc['bands']
if nbands_converge == 'all':
nbands_converge = nbands
elif nbands_converge != 'occupied':
assert isinstance(nbands_converge, int)
if nbands_converge < 0:
nbands_converge += nbands
eigensolver = get_eigensolver(par.eigensolver, mode,
par.convergence)
eigensolver.nbands_converge = nbands_converge
# XXX Eigensolver class doesn't define an nbands_converge property
if isinstance(xc, SIC):
eigensolver.blocksize = 1
self.wfs.set_eigensolver(eigensolver)
if self.density is None:
gd = self.wfs.gd
if par.stencils[1] != 9:
# Construct grid descriptor for fine grids for densities
# and potentials:
finegd = gd.refine()
else:
# Special case (use only coarse grid):
finegd = gd
if realspace:
self.density = RealSpaceDensity(
gd, finegd, nspins, par.charge + setups.core_charge,
collinear, par.stencils[1])
else:
self.density = pw.ReciprocalSpaceDensity(
gd, finegd, nspins, par.charge + setups.core_charge,
collinear)
self.density.initialize(setups, self.timer, magmom_av, par.hund)
self.density.set_mixer(par.mixer)
if self.hamiltonian is None:
gd, finegd = self.density.gd, self.density.finegd
if realspace:
self.hamiltonian = RealSpaceHamiltonian(
gd, finegd, nspins, setups, self.timer, xc,
world, self.wfs.kptband_comm, par.external,
collinear, par.poissonsolver, par.stencils[1])
else:
self.hamiltonian = pw.ReciprocalSpaceHamiltonian(
gd, finegd, self.density.pd2, self.density.pd3,
nspins, setups, self.timer, xc, world,
self.wfs.kptband_comm, par.external, collinear)
xc.initialize(self.density, self.hamiltonian, self.wfs,
self.occupations)
self.text()
self.print_memory_estimate(self.txt, maxdepth=memory_estimate_depth)
self.txt.flush()
self.timer.print_info(self)
if dry_run:
self.dry_run()
if realspace and \
self.hamiltonian.poisson.get_description() == 'FDTD+TDDFT':
self.hamiltonian.poisson.set_density(self.density)
self.hamiltonian.poisson.print_messages(self.text)
self.txt.flush()
self.initialized = True
self._changed_keywords.clear()
