search

Andrew-S-Rosen / mof_screen

High-throughput DFT of MOFs using ASE/VASP
MIT License
25 stars 11 forks source link

Fix NUPDOWN in SPE #6

Closed Andrew-S-Rosen closed 5 years ago

Andrew-S-Rosen commented 5 years ago

When doing the final SPE calculation starting from the previous high-accuracy relaxation, it can sometimes be hard to converge the SCF. Fixing NUPDOWN to the value obtained from the high-accuracy relaxation will resolve this issue in most cases and should be implemented if the first attempt fails. Will need to sum the magmoms and round to nearest integer but should not proceed if rounded value differs by some threshold (e.g. greater than 0.05 mu_B).

Andrew-S-Rosen commented 5 years ago

This has been implemented in https://github.com/arosen93/mof_screen/commit/48030fbe729e54157350d491d5345876a876dfc9