If you are unwilling to go under the hood of the program, it should be
possible to oscillate the field in jumps of width N steps, but with
the overhead of re-setting the fix every N steps. The idea would be to
(in quasi input that I have not checked).
If you are unwilling to go under the hood of the program, it should be
possible to oscillate the field in jumps of width N steps, but with
the overhead of re-setting the fix every N steps. The idea would be to
there is more overhead to this. as with each run statement there is
some overhead, e.g. building neighbor lists and many other steps,
that will seriously slow this approach down. however, in this case
one might get away with reducing the run overhead by using the
"pre no" option, but that would have to be very carefully checked.
in any case, the necessary change to the source code is
rather trivial and should be doable even by somebody that is
not an expert programmer. whoever is able to write and understand
this kind of lammps input script code, should be able to write
the required c++ code. 
any lammps users around that want to try their submitting
their first contribution to the project?
cheers,
axel.