Peter Kraus

`dict_func["name"]` should be capitalised properly to look nice in printing. The `dashparams[disp_type][func].lower()` keys are matched against all aliases (including `dict_func["name"].lower()`), if there is a match with any of them, all...

An updated script that works with current master: ``` from qcengine.programs.dftd3 import get_dispersion_aliases from qcengine.programs.dftd3 import dashcoeff from psi4.driver.procrouting.dft import dft_builder from psi4.driver import procedures import psi4 for dtype in...

- [x] The documentation generates a lot of warnings about hardcoded links when built, see [here](https://github.com/psi4/psi4/runs/5024225771?check_suite_focus=true#step:9:5287)

- [ ] clean-up of the cbs module, including more comments and reducing clutter, in #2132

I'd prefer to keep the blurb simple, and if at all possible, separate tests from docs. Why not go with something like: ``` @loriab : Your PR is modifying Psi4's...

Could you post the output of `conda list`, and try again by first deleting `objdir`, then running the `psi4-path-advisor` command without the `--gcc` option, and then `make`-ing again?

Step 1 would be to try some of the tests. If you self-compiled, then `ctest -L smoke -j N` in `$objdir` will run N simultaneous tests from the "smoke" test...

Try with `set guess core`; it seems to work here. The calculation of the gradients is still painfully slow, though. With `set guess sad` (which is default), you're right, the...

@TermeHansen Did my PM on the forum reach you?

Hi, it is not implemented at the moment. I think the main problem is the averaging of the MP2 extrapolated contributions (`-PS3`), which is not currently implemented. If you were...