QECalc Wrapper for Quantum Espresso
Download

QECalc Ranking & Summary

QECalc Tags

Python library wrapper Quantum Quantum Espresso Espresso wrapper Espresso

QECalc Description

Wrapper for Quantum Espresso qecalc is a Python wrapper for Quantum Espresso.Quantum Espresso (QE) is widely used GNU distributed open source ab initio package for plane wave Density Functional Theory (DFT) and molecular dynamics calculations. Often users need to go beyond basic capabilities of an ab initio program and use its outputs for more advanced tasks. Some examples: * convergence studies of a property of interest with respect to ranging values of different input parameters * Various optimization and minimization problems * Plotting and data processingQECalc is a set of Quantum Espresso launchers and input/ouput parsers available organized under single API. Its primary goal is to use its classes to streamline user’s work flow, offer new functionality and provide the machinery to build new features using numpy, scipy, and matplotlib. One such example can be the class Converger from qecalc/converger.py which can be used to converge such properties as ‘total energy’, ‘geometry’, and ‘single phonon’ with respect to any iteratable variable of PW config file. More examples can be seen in examples directory.InstallationCurrently there is no easy_install option for qecalc. You should have svn client installed and go through the following steps: 1. Go to your installation dir ($INSDIR), for example, ~/apps and type: svn co svn://svn@danse.us/AbInitio/espresso/qecalc qecalc project tree will be created 2. Add qecalc to your PYTHONPATH variable: export PYTHONPATH=$INSDIR/qecalc:$PYTHONPATH 3. This module also depends on diffpy.Structure package. Make sure setuptools is installed and type: easy_install diffpy.StructureUsageIt is essential the user knows how to use Quantum Espresso for the basic tasks. Excellent place to start is the Quantum Espresso wiki page. It is important to check that QE input files lead to satisfactory results before using them in automated manner.In order to run python scripts with Quantum Espresso, one needs to provide all the appropriate config files (scf.in for total energy or geometry optimization; additionally ph.in and dynmat.in for single phonon; or ph.in, q2r.in and matdyn.in for multi phonon calculation; etc) and place config.ini into working dir, which specifies parallel environment of your task as well as all the relevant input and output files. An example of config.ini is located in qecalc directory. All its sections do not need to be populated, only the parameters needed for a specific task. If some of the parameters are missing, default values will be used. The default values are located in qecalc/qecalc/settings.pyBefore the run, check that all the pseudopotentials from the pw config file are available and your output dir existsts (e.g. temp/ ). Also make sure Quantum Espresso is in your $PATH environment variable.Execute your python script which uses qecalc API from your working dir.See examples directory as well as API documentation for more detailsExamplesPWCalcPWCalc consists of one task, launching pw.x. Before running the example, one needs to create config.ini file in the current dir as well as scf.in input file for pw.x. Example of config.ini is provided below:# parallelization parameters# if this section is empty - serial mode is usedparaPrefix: mpiexec -n 8paraPostfix: -npool 8#useTorque: True#paraPrefix: mpirun --mca btl openib,sm,self#paraPostfix: -npool 900# this string will be passed to qsub, -d workingDir -V are already there:#torqueResourceList: -l nodes=16:ppn=12 -N MyJobName -j oe# pwscf input/output filespwscfInput: scf.inpwscfOutput: scf.outlookupProperty() goes through the all hte output files of given qalc:# PWCalcfrom qecalc.pwcalc import PWCalcpwcalc = PWCalc('config.ini')pwcalc.launch()pwcalc.lookupProperty('total energy')pwcalc.lookupProperty('total energy', withUnits = True)pwcalc.lookupProperty('stress', withUnits = True)pwcalc.lookupProperty('forces', withUnits = True)MultiPhononCalcconfig.ini, pw.x, ph.x, q2r.x, and matdyn.x input files should be in the current dir. config.ini should have additional sections corresponding to additional tasks:#ph.x input/ouput, relevant to all phonon calculations:phInput: ph.inphOutput: ph.out#dynmat.x input/output files relevant to single phonon calculationdynmatInput: dynmat.indynmatOutput: dynmat.out# input/output files relevant to multiple phonon calculationq2rInput: q2r.inq2rOutput: q2r.out# input/output files relevant to multiple phonon calculationmatdynInput: matdyn.inmatdynOutput: matdyn.outmatdynModes: matdyn.modesmatdynFreqs: matdyn.freqmatdynfldos: matdyn.phdosIn the following example it is also assumed outputs are laready there after a succesfull run:from qecalc.multiphononcalc import MultiPhononCalcmphon = MultiPhononCalc('config.ini')for task in mphon.taskList: task.output.parse()mphon.lookupProperty('total energy', withUnits = True)# this will output out qpoints, frequencies and eigen modesmphon.lookupProperty('multi phonon', withUnits = True)mphon.dispersion.launch('M', 'Gamma', 'A','L', 50, 50, 50)mphon.dispersion.plot()ConvergerClass converger will converge a value with respect to k-points or different parameters in ‘system’ namelist of pw.x input file. Currently, value can be ‘total energy’, ‘fermi energy’ or ‘single phonon’:from qecalc.converger import Convergeropt = Converger('config.ini','total energy', tolerance = 0.1)ecut = opt.converge(what = 'ecutwfc', startValue = 18, step = 4)conv_thr = opt.converge(what = 'conv_thr', startValue = 1e-4, multiply = 0.1) Requirements: · Python

QECalc Related Software