Skip to content

README.md

Latest commit

 

History

History
81 lines (55 loc) · 4.13 KB

README.md

File metadata and controls

81 lines (55 loc) · 4.13 KB

GI_SR

This directory contains the GI_SR alloy-supercell workflow and its generated results. In practice, it is a working area for building special quasirandom structures, preparing Quantum ESPRESSO calculations, running relaxation and SCF jobs, and collecting fitted or post-processed outputs such as HDF5 summaries, plots, and unfolded band-structure data.

What is here

At the top level, this directory mixes three kinds of content:

  • workflow inputs such as base.yaml, base.cube_2x2x2.crys, input_data.json, pseudopotentials.json, U_base.json, and the ksection_*.txt files,
  • workflow code and launch helpers such as tools/, run_them_functions.sh, main_base.sh, make_phonopy.sh, and the various run_*.sh scripts,
  • generated calculation directories and analysis outputs such as dir_In.../, GI_SR.results.hdf5, GI_SR_4D.results.hdf5, all_scf/, unfold_data/, and the *.png summary plots.

Because this is both a code directory and a results directory, it is best treated as an active project workspace rather than a clean software package.

Typical workflow

The common path through this directory is:

  1. Define the alloy composition and supercell size.
  2. Generate an SQS configuration from base.yaml.
  3. Convert the selected configuration into a .crys structure file.
  4. Interpolate the Hubbard U values for that composition.
  5. Write a QE relaxation input.
  6. Run the relaxation job.
  7. Extract the relaxed crystal structure.
  8. Write and run the SCF job.
  9. Collect outputs into HDF5 files, plots, effective-mass fits, or unfolded band-structure products.

Most of the reusable logic for those steps lives in GI_SR/tools/README.md.

Important files and directories

Core inputs

  • base.yaml: sqsgenerator template used to build alloy configurations.
  • base.cube_2x2x2.crys and base2.cube_2x2x2.crys: template crystal structure files for the 2x2x2 workflows.
  • input_data.json: base QE namelist content.
  • pseudopotentials.json: pseudopotential mapping used by the QE input generators.
  • U_base.json: endpoint Hubbard U values used for interpolation.
  • a_base.json: endpoint lattice constants used when preparing structures and inputs.

Calculation directories

  • dir_In..._cube_2x2x2/: per-composition working directories, usually containing multiple calc_* subdirectories.
  • run_files/, done_run_files/, and log_files/: bookkeeping and job-tracking outputs.
  • all_scf/: collected SCF outputs.

Analysis outputs

  • GI_SR.results.hdf5, GI_SR_4D.results.hdf5, and related HDF5 files: aggregated numerical results.
  • *.bandgap.png and bandgap.png: summary plots for composition-dependent band gaps.
  • unfold_data/: unfolded band-structure data products.
  • fit_all/, fit_gen/, and fit_new/: fitting outputs and related generated data.

Workflow helpers

  • tools/: the main reusable helper scripts for structure generation, QE input writing, and extraction.
  • run_them_functions.sh: shell functions that drive the repeated relax/SCF workflow.
  • remove_all_wfc.sh: cleanup helper for scratch files across many calculations.
  • make_phonopy.sh, make_meff.sh, meff_them_functions.sh: specialized post-processing and derived-workflow helpers.

Environment assumptions

This project assumes a cluster environment with:

  • Slurm variables such as SLURM_NPROCS,
  • Quantum ESPRESSO available through the local setup scripts,
  • the local /work/i254/i254/gentles paths present,
  • Python environments and module loads configured through the helper scripts.

The setup scripts in tools/ are intended to be sourced into the current shell rather than executed directly.

Recommended entry points

If you are starting from the code rather than from an existing calculation directory, the most useful places to begin are:

  • GI_SR/tools/README.md for the reusable helper scripts,
  • run_them_functions.sh for the batch workflow functions,
  • input_data.json, U_base.json, a_base.json, and the ksection_*.txt files for the physical and QE setup,
  • one representative dir_In.../ directory if you want to see the generated layout in practice.