NOs

[shivupa]

@amandadumi and I spent some time today and put the start of natural orbtials into the code.

We used this paper as a reference https://arxiv.org/abs/1311.6244

This should have each MPI process accumulate the contributions to the MO 1RDM for the local set of determinants.

Some points that need discussion/thinking

• There is no AO information so we won't be able to put these back into the AO basis
• Does anyone with more MPI knowledge know how we should efficiently combine each MPI process's local copy of the RDM? They are all numpy arrays

Still a WIP but let us know if anything looks egregiously wrong

[TApplencourt]

Does anyone with more MPI knowledge know how we should efficiently combine each MPI process's local copy of the RDM? They are all numpy arrays

I think you are locking to accumulate https://www.mpi-forum.org/docs/mpi-2.2/mpi22-report/node237.htm

natural orbtials into the code.

Sorry, I don't follow... Do you mean that you added the cabalities to compute natural orbital? Sorry for the dumb question

[shivupa]

Yup sorry the wording was confusing. This pull request adds the ability to

• compute the MO basis 1 particle reduced density matrix
• diagonalize to form MO basis Natural orbitals
• (pending) back transform to AO basis NOs (once we have AO information in the code FCIDUMP doesn't provide this)

Thank you we will work on doing an accumulate