6  Common Problems and Solutions
******************************************

6.1 DASP related
============================

6.1.1 How to configure Pymatgen ?
-----------------------------------------
* ``Pymatgen`` will be installed automatically when the user installs DASP. The pseudo-potential library in ``Pymatgen`` must be set to the 2003 version for VASP, otherwise, the results calculated by DASP program will be unreliable. However, the version of pseudopotential can be set arbitrarily in the DASP calculation. That is, ``potcar_path`` in ``dasp.in`` can correspond to any version of VASP pseudopotential.

6.1.2 How to place the sxdefectalign script
--------------------------------------------
* Users can download ``sxdefectalign.bz2`` by themselves from the website: https://sxrepo.mpie.de/projects/sphinx-add-ons/files<https://sxrepo.mpie.de/projects/sphinx-add-ons/files, and make the program ``sxdefectalign`` executable, namely ``chmod +x sxdefectalign`` , and add its path to the environment variables.

6.1.3 What is the role of file ``redo.in`` in PREPARE and DEC calculations?
------------------------------------------------------------------------------
* Users can modify the parameters of ``INCAR`` when they encounter errors in VASP, and then write the calculation directory to the file ``redo.in`` under the dec directory, namely ``dec/redo.in`` , and run the program again.

6.1.4 Hard disk is full and program cannot execute right.
----------------------------------------------------------
* The wavefunction and charge density will be output by default in the DASP calculation (users can manually modify ``INCAR`` after PREPARE calculation and choose not to write), therefore, large disk space is required. If the storage space is used up, the task system may crash.

6.1.5 Can :guilabel:`vasp_path_tsc` and :guilabel:`vasp_path_dec` be set to the same?
---------------------------------------------------------------------------------------
* These two parameters can be set to the same. But we suggest :guilabel:`vasp_path_dec` uses the ``vasp_gam`` version because DASP does not support multiple k points to do defect calculations currently (namely band filling correction is not included).

6.1.6 Can spin polarization and spin-orbit coupling be considered?
-----------------------------------------------------------------------
*  ``INCAR`` generated by DASP by default includes spin-polarization calculation, namely :guilabel:`ISPIN=2` . If it is necessary to do the calculation including spin-orbit coupling, the path of VASP in ``dasp.in`` need to be changed to the ``vasp_ncl`` version, and modify the relevant parameters in ``INCAR`` after the PREPARE is completed.




6.2 Relevant issues in VASP calculation
=============================================
6.2.1 VASP errors: issues related to :guilabel:`SYMPREC`
------------------------------------------------------------
* Solution: Increase :guilabel:`SYMPREC` in ``INCAR`` and execute again. Users can manually set :guilabel:`SYMPREC` in ``INCAR`` after the PREPARE module is completed to avoid such errors.

6.2.2 The electronic step does not converge when using :guilabel:`level=2 or 3` to do HSE self-consistent calculation.
----------------------------------------------------------------------------------------------------------------------------
* Solution: Change the algorithm, such as the :guilabel:`ALGO` tag in ``INCAR`` could select :guilabel:`Damped or All` (do not mixed with Normal). Or increase :guilabel:`SIGMA` , this parameter should be reset after converged, and calculated again with reading the converged wavefunction.

6.2.3 The electronic step does not converge when doing relaxation1, relaxation2, or static calculation in the first stage of TSC.
------------------------------------------------------------------------------------------------------------------------------------------------
* Solution: In the first stage of TSC, the parameters in ``INCAR`` used for DASP calculation (PBE) are completely consistent with MP database, so they may not converge in a few cases, in which only need to change the algorithm and set :guilabel:`ALGO=Fast or Normal` in ``INCAR`` .

6.2.4 What to do when the calculation stops but does not converge during structure relaxation in the DEC module?
-----------------------------------------------------------------------------------------------------------------
* Solution: Generally, copy CONTCAR to POSCAR and rerun the calculation.