Hi,
I’m using the fix ave/atom command to average positions of all atoms over a certain period of time. I was wondering if there’s a way of randomly choosing the values over which the average is performed ? For example, is it possible to use a variable defined by the random function to specify Nfreq in the fix ave/atom ? In that case, would the variable be evaluated before determining each timestep contributing to the average, or only once before the fix starts ? The reason I’m looking for a way to do this is to perform the average over values that are not correlated in time. Thank you,
Paule
Hi,
I'm using the fix ave/atom command to average positions of all atoms over a
certain period of time. I was wondering if there's a way of randomly
choosing the values over which the average is performed ? For example, is it
possible to use a variable defined by the random function to specify Nfreq
in the fix ave/atom ? In that case, would the variable be evaluated before
determining each timestep contributing to the average, or only once before
the fix starts ? The reason I'm looking for a way to do this is to perform
the average over values that are not correlated in time. Thank you,
first of all, you will have to change the sources to make that work.
it is a non-trivial change, since "nfreq" has to be known ahead of
time, so that the data accumulation can be done. also it has to be
commensurate with the choices of "nrepeat" and "nevery" *and* passed
to a suitable dump command for synchronized output.
axel.
The reason I'm looking for a way to do this is to perform the average over values that are not correlated in time.
Not clear why picking random times would be a good way to do this. I think
you simply want to set Nfreq big enough to get uncorrelated snapshots.
Steve