Stand-alone QimPy calculation from YAML input file

Typical usage:

mpirun [mpi-options] python -m qimpy.dft -i INPUT_FILE [qimpy-options]

Command-line parameters (obtained using python -m qimpy.run -h):

python -m qimpy.dft (-h | -v | -i FILE) [-o FILE] [-c C] [-p Pr Pk Pb]
                       [-n] [-d] [-m FILE] [-V]

Run a QimPy calculation from an input file

optional arguments:
-h, --help

show this help message and exit

-v, --version

print version information and quit

-i FILE, --input-file FILE

input file in YAML format

-o FILE, --output-file FILE

output file (stdout if unspecified)

-c C, --cores C

number of cores per process (overridden by SLURM)

-p Pr Pk Pb, –process-grid Pr Pk Pb

dimensions of process grid: replicas x kpoints x bands/basis, whose product must match process count; any -1 will be set to distribute available tasks for that dimension most equally. Default: -1 -1 -1 implies all dimensions set automatically.

-n, --dry-run

quit after initialization (to check input file)

-d, --no-append

overwrite output file instead of appending

-m FILE, --mpi-log FILE

file prefix for debug logs from other MPI processes

-V, --verbose

print extra information in log for debugging

Note that qimpy must be installed to the python path for these to work in any directory. For development, run python setup.py develop –user in the root directory of the source repository to make the above usage possible without instaling from pip/conda.