I’m trying to learn the fix colvars for free energy calculations, so I started with a very simple system in which there is only one particle I’m trying to pull it with a spring along y direction (from y=8.55 to y=1.0), basically using SMD. Attached is the input file and colvars config file of the systems I’m trying to simulate.
input file = in.a : I’m using “atom_style sphere” and I’m getting following error:
ERROR: Cannot use fix colvars for atoms with rmass attribute
question: is this because of the atom_style that I’m using? If so, is there a way that I can use fix colvars with sphere atom_style?
input file = in.b : I tried doing the simulation but using atom_style atomic to see if the problem is resolved, this time I got the colvars is reporting the initialization part but right after that I get:
Exit code -5 signaled from [machine]
MPI process terminated unexpectedly
I'm trying to learn the fix colvars for free energy calculations, so I
started with a very simple system in which there is only one particle I'm
trying to pull it with a spring along y direction (from y=8.55 to y=1.0),
basically using SMD. Attached is the input file and colvars config file of
the systems I'm trying to simulate.
1. input file = in.a : I'm using "atom_style sphere" and I'm getting
following error:
ERROR: Cannot use fix colvars for atoms with rmass attribute
question: is this because of the atom_style that I'm using? If so, is there
a way that I can use fix colvars with sphere atom_style?
yes and no. this is because the colvars library currently assumes that
particles are point particles with the mass being defined through its
type. with atom style sphere, however, you have extended particles
where the mass can be set per atom. colvars does not support this
(currently). however, you still have the option to set per type masses
(using the mass command), but they will be (mostly) ignored in LAMMPS.
then you could comment out the check for the rmass attribute in the
fix colvars, recompile and it will run. of course, this will only work
correctly for as long as the per type mass is consistent with the per
atom mass.
2. input file = in.b : I tried doing the simulation but using atom_style
atomic to see if the problem is resolved, this time I got the colvars is
reporting the initialization part but right after that I get:
this is due to a bug. you can work around it by adding:
atom_modify map array