search

shirtsgroup / InterMol

Conversion tool for molecular simulations
MIT License
188 stars 54 forks source link

DESMOND->LAMMPS energies are not working that well #114

Closed mrshirts closed 9 years ago

mrshirts commented 10 years ago

Is there something overall weird going on? Let's figure out.

mrshirts commented 10 years ago

Responding to Christoph-

Right now I'm scratching my head over why the LAMMPS example no longer converts correctly to GROMACS (this was an issue before I started reorganizing things). The files seem to convert perfectly from visual inspection but the energies don't agree. I think the committed example has both harmonic and morse bonds which work for L2L but the bond terms disagree for L2G. However, the nonbonded interactions are also creating a discrepancy now...

Can you verify that there was a version that did work? Either you can try to pick the out the problem, or I can take a look? This is just the dihedral term part, right?

ctk3b commented 10 years ago

This is unrelated to dihedrals. This little example just has angles and I was using it to implement reading of multiple bondtypes in one file (mixing morse and harmonic bonds in this case). For some reason, the non-bonded terms appear to be off.

I'll have to take a closer look tomorrow or late tonight and hopefully find the version where it did work.

ctk3b commented 9 years ago

Let's see if this comes up again after the re-write.