Yubo "Paul" Yang
Yubo "Paul" Yang
@prckent from a preliminary look, I don't think the MCCoords abstraction poses any difficulty at all. For true 2D, one can: 1. fill the z-component of all vectors in `walker_deltas.positions`...
Is there any update on this issue? I still get 11 and 0 as the traces for the 1RDMs of up and down electrons using Josh's example above. ``` QMCPACK...
@jtkrogel this run was performed using complex build. The overlap matrix is within 1e-4 of being the identity. [feo-dm_basis-ovlp.zip](https://github.com/QMCPACK/qmcpack/files/5265108/feo-dm_basis-ovlp.zip)
Here are the means of the up and down DMs [feo-mew-1rdm-dats.zip](https://github.com/QMCPACK/qmcpack/files/5268874/feo-mew-1rdm-dats.zip) and the outputs from `plot_rdm.py` for a quick peak (first up then down). The down DM is certainly wrong....
@jtkrogel I tested out the commit from #1021. The RDMs still look the same as above so I did not attach them. ``` Last git commit: 0b5eab33ec8fa981d44fc51fcb7652fc289d6a4c Last git commit...
In `DensityMatrix1B.cpp::evaluate_loop`, `int ij = nindex + s * basis_size2;` is inside the `ns` loop. This looks suspicious.
NVM, moving `int ij = nindex + s * basis_size2;` outside creates seg. faults... Is the "loop" evaluator worth fixing?
Right, this is a hacky way to combine two non-orthogonal determinants. Can we add a flag to circumvent the table method? Since all the SPOs are evaluated, we just need...
Without the energy density estimator, TraceManager does run without crashing in parallel. The thing that's crashing #4527 is triggered by `write="yes"`.