compute_msd
This keyword computes the self-diffusion coefficient (SDC) from the mean-square displacement (MSD) function. If this keyword appears in a run, the MSD function will be computed and each Cartesian component of the SDC is calculated as one half of the time derivative of the corresponding MSD component.
Syntax
compute_msd <sample_interval> <Nc> [<option> ...]
Parameters
sample_interval: Positive integer giving the number of time steps between position samplesNc: Positive integer giving the number of correlation points, including zero correlation time
Each optional argument can appear once, in any order:
group <group_method> <group>selects one group.all_groups <group_method>computes separate results for every group in the selected grouping method, for example one group per molecule. It cannot be combined withgroup.save_every <interval>writes intermediate results tomsd_step[step].out.intervalis a positive integer in MD time steps.
A copy is written only on a step that is a multiple of both sample_interval and interval, after at least Nc position samples have been collected.
The first copy is therefore written at the first such step at or after Nc * sample_interval.
Choose interval to be a multiple of sample_interval to obtain the requested spacing between saved copies.
Step numbers are counted from the beginning of the current run.
Each copy contains the averages accumulated so far; saving does not reset the accumulation.
Output
The results will be written to msd.out output file.
Examples
An example of this function is:
compute_msd 5 200 group 1 1
This samples positions every 5 steps for group 1 in grouping method 1, using 200 correlation points including zero time.
To compute the MSD for all groups in group method 1 and save a copy of the MSD every 100 000 steps, one can write:
compute_msd 5 200 all_groups 1 save_every 100000
In this example, sufficient position samples are available at step 1000, but the first intermediate file is msd_step100000.out at step 100000.
The final results are also appended to msd.out at the end of the run.
Notes
None.