[lammps-users] CG method with initial velocity.

Hi, all:

I was trying to use CG method to relax a structure. My starting configuration is the last snapshot from a MD run.

I know CG method shouldn't do anything with the initial velocity. But looks like LAMMPS read in velocities and used them.

What I observed is the velocities remains constant during CG relaxation, and total energy simply fluctuates instead of droping.

Does this mean a bug in LAMMPS?

Thanks,

The minimizers ignore velocity. They should simply be
preserved after the minimization is done.

Steve

Unfortunately, in my test, the structrure doesn't seem to reduce in energy if there is initial velocity.

If I set the initial velocity to zero, then the minimizer works as expected.

Y. Lin

Is the *potential* energy any different if the initial
velocity is nonzero?

This doesn't sound right ... can you post an input script
for as simple/small a problem as will reproduce the
problem?

Thanks, Steve