I have couple of questions about msi2imp tool.
-
CVFF forcefield uses simple geometric rule to calculate non-bonded parameters for dissimilar pairs. However, “pair_modify mix geometric” is not used in any of the sample input files in the “test” directory. I assume this is because data files in this directory include only one molecule and not considering the mixing rule will not cause a major inaccuracy. I just want to confirm that mixing rule should be considered for big systems which include multiple molecules.
-
I noticed that msi2lmp code uses the harmonic form for the bond stretching term while the default in Discover program is the Morse potential. I know that the Morse form is computationally more expensive than the harmonic form. Is there any other reason for choosing harmonic term as the default?
Thank you for your help.
MK.
I have couple of questions about msi2imp tool.
1) CVFF forcefield uses simple geometric rule to calculate non-bonded
parameters for dissimilar pairs. However, "pair_modify mix geometric" is
not used in any of the sample input files in the "test" directory. I assume
this is because data files in this directory include only one molecule and
not considering the mixing rule will not cause a major inaccuracy. I just
want to confirm that mixing rule should be considered for big systems which
include multiple molecules.
two comments to that. 1) the files in the "test" directory are tests to
validate the functioning of msi2lmp; nothing more. they are _by no means_
reference examples for correct simulations. for testing the conversion, any
mixing rules are irrelevant. 2) mixing *always* matters as long as you have
more than one atom type and do not explicitly provide coefficients for all
pairs of atom type. if you go by that token, all class2 tests are wrong as
well.
2) I noticed that msi2lmp code uses the harmonic form for the bond
stretching term while the default in Discover program is the Morse
potential. I know that the Morse form is computationally more expensive
than the harmonic form. Is there any other reason for choosing harmonic
term as the default?
you would have to ask the original author(s) of msi2lmp about the reasons
for their choices. fact is, msi2lmp only supports a subset of the force
field functionality of MS toolchain and many of those choices are
hardcoded, not just the choice for bonded interactions. the main
functionality of msi2lmp has been mostly unchanged for more than ten years
now and only bugfixes, performance improvements and a few LAMMPS-related
features were added.
axel.