Theory benchmark status summary (2020-10-28)
- Hyesu Jang
Concerns on the current level of theory
We do not know how well DZVP performs for charged molecules, and for properties other than conformational energies.
DZVP results are less accurate than literature for medium/large systems largely due to lack of DZVP-specific dispersion parameters and three-body term.
The need of inclusion of diffuse function: diffuse functions are for describing anions or dipole moments and for accurate modeling of intra- and inter-molecular bonding.
Benchmarking set generation
https://openforcefield.atlassian.net/l/c/ux1aNe1G
Performance tests of DZVP and Ahlrichs family
Psi4 performance benchmark for a single gradient calculation of FGG114 on i7-6850K 1 core
Dataset submission status
method | available in QCArchive? | dataset name |
|---|---|---|
B3LYP-D3(BJ)/DZVP | O |
|
B3LYP-D3(BJ)/def2-TZVP | O |
|
B3LYP-D3(BJ)/def2-TZVPP | O |
|
B3LYP-D3(BJ)/def2-TZVPD | O |
|
B3LYP-D3(BJ)/def2-TZVPPD | submitted |
|
B3LYP-D3(BJ)/6-31+G** | submitted |
|
Reference (MP2/heavy-aug-cc-pVTZ) | X (now running performance test locally) | OpenFF Theory Benchmarking Set MP2 heavy-aug-cc-pVTZ |
Reference and other combinations | ||
Questions
other functionals/ basis sets to test?
Is MP2/heavy-aug-cc-pVTZ reasonable for reference? (too expensive?)