Joshua Goings

@pwborthwick That makes sense. It would be best to toggle real/complex for the overall calculation, because there are some cases we want everything complex. Singlet O2 is a great example...

@pwborthwick Unless you do some tricks to reduce the dimension, the full RPA response matrix is non-Hermitian, and can in principle give complex eigenvalues. See, for example, Eq(31) [here](https://joshuagoings.com/2013/05/03/derivation-of-time-dependent-hartree-fock-tdhf-equations/). So...

Hey thanks @AlexB67! These comments and suggestions are great. You are right, with regards to complex, all the integrals are done in real arithmetic. The complex comes afterward when we...

Excellent @AlexB67! Impressive timing, though I'm not surprised. I'll gladly take any help I can get :)

No, I had not planned on implementing CCSD(T) at the moment. It certainly could be done by adding a new routine in `postscf.py`. I doubt it would be very fast,...

Check that the SCF and correlation energy match. I found that the nuclear repulsion energy was wrong (go figure) so I fixed that on the site. That explains most of...

I will need to see the code and some tests then. The case of HeH+ works for the values hard-coded, but unfortunately I cannot test it on anything else since...

Okay, I ran the code. I agree something is not right, but I need some time to figure out where the error lies. It may be in the transition to...

I managed to track down the bugs in that blog post. Thanks @jl2922 for pointing them out. The error was in a couple permutations for constructing the T2. One messed...

Sorry to hear that. I just ran a fresh compilation and had no issues. ``` $ python sample-input.py E(SCF) = -74.942079897820 in 9 iterations Convergence: FPS-SPF = 8.703089748378866e-11 RMS(P) =...