gomc-wsu / gomc_examples Goto Github PK

Since the MEMC move is complicated, and the manual does not explain the setup as clearly as it could, we need to have multiple examples of how to setup the MEMC move.

Implementation of data collection and sending data back to our server is next priority and will be implemented before next release.

@msoroush

In my meeting with Dr. Potoff today, we briefly discussed using MBAR with GCMC. We need to validate that MBAR is giving the same results as histogram reweighting for the case where epsilon and sigma do not change. Unfortunately, our histogram reweighting code appears to have a bug in it. Specifically, the VLE properties we obtain have systematic deviations from the values reported in the literature.

For example, we simulated hexane using the Mie 16-6 potential with GOMC.

These are the histograms we obtained for the different chemical potentials and temperatures:

As you can see here, our densities do not agree perfectly with the values reported by Potoff et al. (especially at higher T):

The strange thing is, that our higher temperature results get worse as we include histograms from the lower temperature simulations. For example, this is what we get at the highest temperature if we only include the histograms at the "bridge":

This suggests that we are not weighting the histograms properly.

Dr. Potoff thought that we could do a couple of things to resolve this issue:

1. You send me your histogram reweighting code
2. I send you my raw simulation output files for hexane and you run your code to see what results you get
3. You send me your raw simulation output for some system (he thought you should have the data for the noble gases readily available) and I use my code to see if I get the same results
4. I send you my histogram reweighting code

I think we should try options 1, 2, and 3 before we consider 4.

If you could send me your histogram reweighting code and/or simulation output files that would be great.

Here are my hexane GOMC output files for the number of particles and energies (first row has Temperature, Chemical potential, and box dimensions)

hexane_GCMC_output.zip

Thanks

We may want to consider adding the NEMTMC since it is several variables to set.

The current SPC/E water NPT example starts from a rather bad initial configuration with a number of significant overlaps. This causes the simulation to quickly "blow up." Increasing the initial box lengths to 24 A (was 23 A) seems to fix the problem. Running more extensive testing now.