eljost/pysisyphus: Bugfixes, Refactoring for more consistency
Addition Added ability to retry/rerun crashed calculations (retry_calc: 1). For now only supported by ORCA. Support for fragments: total in endopt: section. This allows optimizing the system as whole as well as optimizing the separate fragments Single geometries from multi-geometry input files can n...
| Main Authors: | , , , |
|---|---|
| Format: | Software |
| Language: | unknown |
| Published: |
2021
|
| Subjects: | |
| Online Access: | https://zenodo.org/record/5497269 https://doi.org/10.5281/zenodo.5497269 |
| Summary: | Addition Added ability to retry/rerun crashed calculations (retry_calc: 1). For now only supported by ORCA. Support for fragments: total in endopt: section. This allows optimizing the system as whole as well as optimizing the separate fragments Single geometries from multi-geometry input files can now be accessed via pythons index-notation: fn: multi_geoms.trj[1] would only load the second geometry from multi_geoms.trj Very basic support for preconditioning of position and rotation. precontr: in YAML input; runs before preopt. Still very slow for sizable systems. pr_nto keyword support for OverlapCalculator to dynamically select the number of NTO pairs in overlap calculations Bugfixes Hessian was unnecessarily recalculated in TS optimizations after COS runs Fixed Eckart projection to remove translation and rotation. Adapted from geomeTRIC Eckart projection in calculation of initial displacement from TS in IRC runs was never carried out Made parsing & storing of MO and CI coefficients in store_overlap_data() much more consistent Miscellaneous coord_type in optimizations and wavenumber of imaginary mode in IRC runs is now reported Added separate ORCA5 calculator (type: orca5) so ORCA4 and ORCA5 can be setup in parallel Refactored geometry loading via get_geoms and read_geoms in trj.py to support arbitrary kwargs All preopt/tsopt/endopt sections can now contain a geom section, e.g., to control the coord_type Requested and actual energy lowering is now reported in IRC runs Wavenumber/frequencies are now also reported when thermoanalysis is printed |
|---|