Re： The calculation about VACF

LAMMPS LAMMPS Mailing List Mirror
1

Hi Axel,

Thank you for your reply.
Actually,I have seen the example in the manual. But I think there is no averaging between Vx(0) *Vx(t) and Vx(1)*Vx(1+t) simultaneously. That’s what really confusing me.

2

Hi Axel,
    Thank you for your reply.
    Actually,I have seen the example in the manual. But I think there is no averaging between Vx(0) *Vx(t) and Vx(1)*Vx(1+t) simultaneously. That's what really confusing me.

do you have any proof for this statement? does the example code *not*
produce a correct VACF for a known example?

axel.

3

If you mean compute a VACF averaged over many
starting times, then no, compute vacf does not do that.
It would require LAMMPS to store a large number
of starting time velocities for each atom, which could
be inefficient. But we may implement it as an option
at some point. Ditto for compute msd.

Steve

4

Hi Steve and Axel,

Thank you for your reply,which helps me a lot.
Sincerely,
Ray