Jonathon Misiewicz
Jonathon Misiewicz
The input described in [this forum post](http://forum.psicode.org/t/cisd-quadrupoles/2464/9?u=jmisiewicz) segfaults during the integral transform on current master. (Confirmed on my laptop, after removing threading instructions and reducing memory to 4GB.) The reporter...
Pair energies computed by `fnocc` have some issues: 1. They're not computed consistently across methods. DF-CCSD has them, but non-DF methods do not, for some reason. 2. `fnocc` computes spin-summed...
After a recent [forum report](http://forum.psicode.org/t/sapt0-d3mbj-error-writing-to-file/2474/6), @zachglick had the excellent suggestion of having _some_ way to warn the user when the user doesn't set the scratch space. What exactly this should...
Carried forward from [forum topic](http://forum.psicode.org/t/cisd-quadrupoles/2464/6?u=jmisiewicz). `properties('cisd')` should probably save property names as `CISD` in addition to `CI`. In this case, it's probably also worth saving as `CI UNRELAXED` and `CISD...
We need to add an example of a `properties` call to the "Psi4 Tutorial" and/or "Structuring an Input File" sections of the documentation.
Add threading to the nshell^2 loops of [this function](https://github.com/psi4/psi4/blob/c18efeda5af93d965314366774e401e7a0c76453/psi4/src/psi4/libmints/twobody.cc#L285). Carried forward from [other](https://github.com/psi4/psi4/issues/2559#issuecomment-1111344355) [discussion](https://github.com/psi4/psi4/issues/2559#issuecomment-1111350116).
The `cc` module [forcibly stops users](https://github.com/psi4/psi4/blob/master/psi4/driver/procrouting/proc.py#L3149-L3152) from requesting tight convergence out of `cclambda`, so as not to [break tests](https://github.com/psi4/psi4archive/commit/2b2eb6c5a710dde180b1ef7a050012069afeae8d). That's bad. Deal with #2519 first.
The `cc` module [has](https://github.com/psi4/psi4/blob/master/psi4/src/psi4/cc/ccenergy/diis.cc) [a](https://github.com/psi4/psi4/blob/master/psi4/src/psi4/cc/ccenergy/diis_RHF.cc) [habit](https://github.com/psi4/psi4/blob/master/psi4/src/psi4/cc/ccenergy/diis_ROHF.cc) [of](https://github.com/psi4/psi4/blob/master/psi4/src/psi4/cc/ccenergy/diis_UHF.cc) [rewriting](https://github.com/psi4/psi4/blob/master/psi4/src/psi4/cc/cclambda/diis.cc) [DIIS](https://github.com/psi4/psi4/blob/master/psi4/src/psi4/cc/ccresponse/diis.cc) when they really should [just use `libdiis`.](https://github.com/psi4/psi4/blob/master/psi4/src/psi4/libdiis/diismanager.h) In addition to code cleanup, I predict an important practical benefit. The central code...
Currently, if `cclambda` fails to converge, it will continue with the computation, with several zeros that shouldn't be there. It should _at the very least_ warn the user that it...
Consider the following input: ``` molecule h2o { 1 2 O H 1 1 H 1 1 2 104.5 } set { reference uhf basis cc-pVDZ freeze_core true roots_per_irrep [0,...