About fix ave/correlation command

Dear all,

I am wondering the result difference of two commands (c_flux is compute heat/flux in three directions):

  1. fix JJ1 all ave/correlate 10 50000 200000 c_flux[1] c_flux[2] c_flux[3] type auto file J0Jt.running.dat1 ave running
    run 1000000

  2. fix JJ2 all ave/correlate 10 50000 1000000 c_flux[1] c_flux[2] c_flux[3] type auto file J0Jt.running.dat2 ave running
    run 1000000

J0Jt.running.dat1 will include 6 replicas (timestep 0, 200000, 400000, 600000, 800000, 1000000) of heat current autocorrelation function (HCACF), and dat2 will include 2 replicas (timestep 0 and 1000000). I am using running average option. The last replica at timestep 1000000 of two data files will be the same or not? I think it will be the same after looking into the fix_ave_correlate.cpp, but I am not so sure. Actually, I got different results from method 1 and my own autocorrelation postprocess (similar to method 2) from printed heat flux data. I haven’t test method 2 from LAMMPS yet. Please confirm it if you have some experiences or you know this. Thanks.

Best regards,

I haven't test method 2 from LAMMPS yet.

Why don't you run it and see? If you think it
does something different than the doc page
explains, then post a Q.