[gmx-users] restarting mdrun

[Mark Abraham]

On 17/02/10 02:52, Chandan Choudhury wrote:
> Hello gmxusers !!!
> I want to continue my simulation, from where the point where it stopped.
> The gmxcheck on .trr and cpt files are as:
>
> $gmxcheck -f state.cpt
>
> Checking file state.cpt
>
> # Atoms  186864
> Last frame         -1 time 9362.500
>
>
> Item        #frames Timestep (ps)
> Step             1
> Time             1
> Lambda           1
> Coords           1
> Velocities       1
> Forces           0
> Box              1
>
> $gmxcheck -f 5-10.trr
>
> Checking file 5-10.trr
> trn version: GMX_trn_file (single precision)
> Reading frame       0 time 5169.500
> # Atoms  186864
> Reading frame    8000 time 9169.500
>
>
> Item        #frames Timestep (ps)
> Step          8395    0.5
> Time          8395    0.5
> Lambda        8395    0.5
> Coords        8395    0.5
> Velocities    8395    0.5
> Forces           0
> Box           8395    0.5
>
>
> There is a mismatch of time in both the outputs.
>
> The date of modification of both the files as shown are:
>
> Feb 15 19:55 state.cpt
>
> $date -r 5-10.trr
> Mon Feb 15 19:59:56 IST 2010
>
> The cpt file shows time greater than the trr file. But the time of
> modification of cpt file is less than trr file.

If the simulation stopped catastrophically, then the timestamp could 
easily differ, especially if the .cpt is opened, written to and closed 
periodically, and the .trr is open continuously.

> Cannot understand the,
> how to restart the mdrun. And the size of the trr file is 36GB. Can I
> use the append flag in mdrun for gmx 4.0.5.

Search the webpage for suggestions.

Mark