Models: Sqw: phonons: support bigDFT

McStasMcXtrace / iFit

a simple library to analyze data (with McCode and Phonons/DFT hooks). :warning: this project has been moved to https://gitlab.com/soleil-data-treatment/soleil-software-projects/remote-desktop

http://ifit.mccode.org

Other

5 stars 5 forks source link

Models: Sqw: phonons: support bigDFT #124

Closed farhi closed 6 years ago

farhi commented 7 years ago

Get bigDFT from http://bigdft.org/Wiki/index.php?title=BigDFT_website

Currently version 1.8

farhi commented 7 years ago

bigDFT ASE/iFit interface specifications:

sqw_phonons requirements:

atoms
potentials
kpoints
xc
nbands
diagonalization
occupations/smearing
nsteps
toldfe (scf E tol)

farhi commented 7 years ago

ASE Calculator snippets:

ASE features:

implemented_properties = ['energy', 'forces', 'stress', 'magmom']

bigDFT yaml keywords: <see http://bigdft.org/Wiki/index.php?title=First_runs_with_BigDFT>

positions -> yaml: posinp: positions OR write a xyz file: ase.io.write(file, atoms, format='xyz')
potentials -> yaml: psppar.
xc -> yaml: psppar. {Pseudopotential XC: 11} allow: PBE=11, PBEsol, LDA=1
xc -> yaml: dft: ixc
kpoints -> yaml: kpt: method: MPgrid
occupations -> yaml: mix: occopt
nsteps -> yaml: mix: itrpmax OR dft: itermax

bigDFT output: in log.yaml

energy -> "Energy (Hartree)"
forces -> "Atomic Forces (Ha/Bohr):"
stress -> tensor ???
electric dipole -> "Electric Dipole Moment"
magmom -> ???

Help for keys:

$ $BIGDFT_ROOT/bigdft-tool -a wizard

farhi commented 7 years ago

bigDFT installation:

go into sandbox:

source ../build/install/bin/bigdftvars.sh 
command start: OMP_NUM_THREADS=1 mpirun -np 8 $BIGDFT_ROOT/bigdft -l no
export PKG_CONFIG_PATH=/net/serhom.ill.fr/volumes/data/home/cs/farhi/Programs/bigdft/build/install/lib/pkgconfig/
sudo apt-get install flex bison bzr qbzr gfortran libxc1 libxc-dev liblapack-dev libfftw3-dev
../Installer.py build -f farhimacpro
../Installer.py make -f farhimacpro
compile: make install -j FCFLAGS='-O2 -fPIC -g'

performance analysis:

ipython:
In [1]: from futile import Time
In [2]: tt=Time.TimeData('time.yaml')
In [3]: tt.show()

farhi commented 7 years ago

bigDFT test:

cd ~/Programs/bigdft/build/bigdft/tests/DFT/cubic/C-test
OMP_NUM_THREADS=1 mpirun -np 4 $BIGDFT_ROOT/bigdft

must run with log=yes (no '-l no' in command) then:

**ipython**:
import yaml
stream = file('log.yaml','r')
y=yaml.load(stream)
y['Atomic Forces (Ha/Bohr)']    -> FORCES per atom
y['Energy (Hartree)']           -> energy