Git master branch¶
- In slfcfg.f (KGRN) (vol-vols) is changed to voli
- In fermis.f (KGRN) nky=400
- Local (atom) projected fixed spin approach is implemented
- In kmesh.f (KLIB) the triclinic case (IBZ=14) is corrected
- In slfcns.f (KGRN) the nlin=0 case is corrected for soft-core calculation (A. Östlin)
- In trnsfm.f (KGRN) facd is defined as real number (A. Östlin)
- Save LAT in KSTR
- Delete file *.tfm if HIGH=Y in KSTR
- Read LAT and compare with IBZ in KGRN
- Small formatting changes in KGRN
- Slope matrix storage on FOR011 is fixed (STMP=Y) in KGRN
- Total energy computed for valence+frozen_core for SOFC=Z in KGRN and KFCD
- PBEsol implemented in KLIB, KGRN (IEX=9) and KFCD
- AM05 implemented in KFCD (only for total energy)
- Local muffin-tin zero implemented in both KGRN (FIGX=4) and KFCD
- Alphacpa has been moved to the input file in KGRN
- Long filenames allowed in KGRN and KFCD (by Matti Ropo)
- Due to this change, there are new input files in KGRN.
- Information about the actual version in VERSION
- Information about the input files in README files
- Hopfield parameter in KGRN.
- Total charge versus radius in KFCD (with FPOT=Y it prints up to a large radius).
- Two corrections in KFCD for UBG=Y:
- In input.f use actual WS radii (alat*wst) insted of those specified in KGRN.
- In fcdexc.f for any r initialize figv (filv) as figv=sigma*r (filv=sigma).
Only in KGRN:Set only the dimension for thead in modules.f and initialize in 0kgrn.f
Only in KFCD:no shell cut-off is imposed for the Madelung correction for high-l
Only in KGRN:Modified two-center expansion implemented for the slope matrix (for wide bands).
Only in KGRN:Modified greenf.f and grnfcd.f for more accurate contour integrals (removing the unphysical poles from the complex plane).
- Due to this change, the self-consistent iterations are different in EMTO5.3 (and later versions) compered to EMTO5 (and former versions). However, the total energy is only slightly affected by that.
Only in KFCD:Proper one center density in KFCD
Only in KFCD:When the full potential is calculated, the TFM correction is added to the energy
THIS IMPLEMENTATION IS STILL UNDER DEVELOPMENT!
Only in KGRN:Vacuum (Va) sites are introduced, which differ from Em sites by the renormalization condition (they are considered as real empty sites, and excluded from the charge renormalization for Poisson equation).
Only in KGRN:Electronic entropy is added
Only in KGRN:New library routin for matrix inversion
Only in KGRN:For EXPAN=D and zmsh=M or m lower Taylor is used only for lower contour
- Exact FCD calculation in KGRN (using the input EF from the last iteration)
- Renormalization in KFCD from Si (NOT RINS)
- SCA density calculated and stored in KGRN
- Exact full potential calculated in KFCD
- Transition Functional Method in KFCD (not included in total energy !!)
- Fermi Surface calculation in KGRN
- Different output for DOS and FS in KGRN
- Updated space group in KGRN (before only the point group was used)
- DOS(EF) is always calculated in KGRN