dihedral_style hybrid

Hi,

I am trying to simulate a mixture of three molecule types using LAMMPS (version: 09Oct2012). There are two types of dihedrals involved: opls and charmm. The simulations for pure components work perfectly but I am having issues when I try to simulate the mixture while using the dihedral_style hybrid command. I think I am doing everything right (specifying the dihedral_style in the input file and the dihedral coeffs in the data file) but I am getting the following error: ERROR: Dihedral coeff for hybrid has invalid style (dihedral_hybrid.cpp:261)

It would be really helpful if you can give me some suggestion.

INPUT file command: dihedral_style hybrid charmm opls

DATA:

Dihedral Coeffs

1 charmm 0.916 1 0 0.0 # charmm UA_TF2N.pdb S N S CF3
2 charmm -1.045 2 180 0.0 # charmm UA_TF2N.pdb S N S CF3
3 charmm -0.382 3 0 0.0 # charmm UA_TF2N.pdb S N S CF3
4 charmm 0.800 3 0 0.0 # charmm UA_TF2N.pdb S N S O
5 charmm 12 2 180 0.0 # charmm UA_HMIM.pdb CW CW NA CR
6 charmm 2 2 180 0.0 # charmm UA_HMIM.pdb H4 CW NA CR
7 charmm .017 1 0 0.0 # charmm UA_HMIM.pdb CT2 CN2 NA CR
8 charmm .015 2 180 0.0 # charmm UA_HMIM.pdb CT2 CN2 NA CR
9 charmm -.212 3 180 0.0 # charmm UA_HMIM.pdb CT2 CN2 NA CR
10 charmm 2 2 180 0.0 # charmm UA_HMIM.pdb CW CW NA CN2
11 charmm 12 2 180 0.0 # charmm UA_HMIM.pdb NA CR NA CW
12 charmm 1.5 2 180 0.0 # charmm UA_HMIM.pdb H5 CR NA CW
13 charmm 2 2 180 0.0 # charmm UA_HMIM.pdb CW CW NA CN3
14 charmm .193 1 0 0.0 # charmm UA_HMIM.pdb CT2 CN2 NA CW
15 charmm .220 2 180 0.0 # charmm UA_HMIM.pdb CT2 CN2 NA CW
16 charmm .009 3 0 0.0 # charmm UA_HMIM.pdb CT2 CN2 NA CW
17 charmm 2 2 180 0.0 # charmm UA_HMIM.pdb NA CR NA CN3
18 charmm 1.5 2 180 0.0 # charmm UA_HMIM.pdb H4 CW CW NA
19 charmm 12 2 180 0.0 # charmm UA_HMIM.pdb NA CW CW NA
20 charmm-3.5 1 0 0.0 # charmm UA_HMIM.pdb NA CN2 CT2 CT2
21 charmm 1.6 2 180 0.0 # charmm UA_HMIM.pdb NA CN2 CT2 CT2
22 charmm 1.6 3 0 0.0 # charmm UA_HMIM.pdb NA CN2 CT2 CT2
23 charmm 1.5 2 180 0.0 # charmm UA_HMIM.pdb H5 CR NA CN2
24 charmm 1.5 2 180 0.0 # charmm UA_HMIM.pdb H5 CR NA CN3
25 charmm 1.5 2 180 0.0 # charmm UA_HMIM.pdb H4 CW CW H4
26 charmm 1.5 2 180 0.0 # charmm UA_HMIM.pdb H4 CW NA CN3
27 charmm 1.5 2 180 0.0 # charmm UA_HMIM.pdb H4 CW NA CN2
28 charmm 2 2 180 0.0 # charmm UA_HMIM.pdb NA CR NA CN2
29 charmm -3.5 1 0 0.0 # charmm UA_HMIM.pdb CN2 CT2 CT2 CT2
30 charmm 1.6 2 180 0.0 # charmm UA_HMIM.pdb CN2 CT2 CT2 CT2
31 charmm 1.6 3 0 0.0 # charmm UA_HMIM.pdb CN2 CT2 CT2 CT2
32 charmm -3.5 1 0 0.0 # charmm UA_HMIM.pdb CT2 CT2 CT2 CT2
33 charmm 1.6 2 180 0.0 # charmm UA_HMIM.pdb CT2 CT2 CT2 CT2
34 charmm 1.6 3 0 0.0 # charmm UA_HMIM.pdb CT2 CT2 CT2 CT2
35 charmm -3.5 1 0 0.0 # charmm UA_HMIM.pdb CT2 CT2 CT2 CT3
36 charmm 1.6 2 180 0.0 # charmm UA_HMIM.pdb CT2 CT2 CT2 CT3
37 charmm 1.6 3 0 0.0 # charmm UA_HMIM.pdb CT2 CT2 CT2 CT3
38 opls 2.8828 -0.6508 2.2185 0.0 #
39 opls 0.0 -1.0 4.0 0.0 #
40 opls 2.8828 -0.6508 2.2185 0.0 #

I attached my input file and data file in an earlier mail which has been with the moderator for a while due to large size. You can also have access to it once (if) it gets approved.

Sincerely,

Hi,
   I am trying to simulate a mixture of three molecule types using LAMMPS
(version: 09Oct2012). There are two types of dihedrals involved: opls and
charmm. The simulations for pure components work perfectly but I am having
issues when I try to simulate the mixture while using the dihedral_style
hybrid command. I think I am doing everything right (specifying the
dihedral_style in the input file and the dihedral coeffs in the data file)
but I am getting the following error: ERROR: Dihedral coeff for hybrid has
invalid style (dihedral_hybrid.cpp:261)

It would be really helpful if you can give me some suggestion.

it is unlikely that somebody will look into your issue unless you
confirm that you get the same error with the current version.

INPUT file command: dihedral_style hybrid charmm opls

DATA:

[...]

I attached my input file and data file in an earlier mail which has been
with the moderator for a while due to large size. You can also have access
to it once (if) it gets approved.

same as above. it is unlikely that somebody will look at huge input
files. the issue you describe is independent from the problem set
size, so a *much* smarter approach would be to construct a small test
system with only a small number of different style dihedrals and debug
that first. that would also reduce the time required for debugging and
makes it possible to run with a serial executable and use all kinds of
debugging and testing tricks (which usually slow down a calculation).

axel.