I previously ran Electron-phonon interaction calculations using the TDepES module in ABINIT version 8.10.3. Now that I’ve upgraded to version 10.2.3, I’m encountering some issues.
- Issue with iomode setting (iomode = 1 and 3)
While running the teph_tdep_legacy_4 tutorial, when iomode is set to 3 the DDB.nc files (and not DDB files) are being generated. The post-processing script requires DDB files.
In version 10.2.3, we have to perform two separate runs to generate all the necessary output files:
First run: Set iomode = 3 to generate the .nc files. [*EIGR2D.nc, *EIG.nc, *GKK.nc, etc. are generated in this run]
Second run: Set iomode = 1 (the default) to generate the DDB files required for postprocessing. [*DDB, *EIG.nc, *GKK.nc etc. are generated in this run]
Using this approach and combining the relevant output files, we are able to perform the post-processing operations and generate the temperature.txt files for Carbon polytypes (diamond and hexagonal (lonsdaleite)).
This requirement of two separate runs doubles the computational cost and time.
Q1. We would like to know if there is any tag which enables to generate all the required output files in a single run?
- Correction in temperature_final.py when using ieigr2rf = 2 and 3
We are also interested in performing sum over states calculations using the tag ieig2rf = 2 and 3. These runs do not generate the *GKK.nc files in their output. The post-processing calculations are thus done using the type 3 mode (static AHC calculation without control on active space). The type 3 calculation do not require *.GKK.nc files. However, the current temperature_final.py is asking for the GKK files even in the type 3 calculations (line no. 316 and 408).
We have commented out the above lines in the temperature_final.py file and generated the temperature.txt file.
Q2. We would like to know if it is the right way to do this?
- Variation in values across different versions of ABINIT
While executing the teph_tdep_legacy_4, for sum over states computations, discrepancies were observed in the values obtained from different versions of ABINIT (v.8.10.3 and v.10.2.3) when setting the tag ieig2rf to 2 and 3.
For instance, in the case of the Carbon diamond structure (for nband=nvalence(4)), the ZPR values of the VBM (at Gamma) are 482.4 meV and -9.4 meV for ieig2rf values of 2 and 3, respectively in the present version, v.10.2.3,. In contrast, in the earlier version, v.8.10.3, the corresponding ZPR values were 484.8 meV and 475.4 meV for ieig2rf values of 2 and 3, respectively.
Q3. We would like to know if this variation in the upgraded version (v.10.2.3) is correct or is it an artefact of us commenting out lines 316 and 408 (which called for the *GKK.nc files) in temperature_final.py file.
- Missing rf_mods.py script
The postprocessing script, temperature_final.py, requires the script rf_mods.py, but this file is not present in the current version of ABINIT (v.10.2.3). Additionally, temperature_final.py is still calling the rf_final.py script, which is causing confusion if the file structure has changed in the new version or not.
Q4. We doubt if the rf_mods.py script is either deprecated or hasn’t been properly included in the current version.
Looking forward for clarifications and suggestions.
Thanks
Shilpa