Re: [AMBER] Error during frames generation from production run

From: Jason Swails <>
Date: Thu, 12 May 2016 16:10:38 -0400

To summarize what Bill and Dan have said:

On Thu, May 12, 2016 at 4:18 AM, Sneha Rai <> wrote:

> Dear Amber users,
> I have used the following input file for 20 ns production run.

​The below output file specifies only 10 ns.

> The output
> file indicates that run is complete without any errors. However, when I
> load mdcrd file of production run in VMD, only 1367 frames were loaded and
> I was expecting 4000 frames. Following is my input file for production run,
> kindly suggest possible solution for this problem.

​Try reading the mdcrd file in with cpptraj. How many frames does it
suggest exist? If it disagrees with VMD, the likely culprit is that you
picked the wrong trajectory type (e.g., you didn't pick the version with
box when you should have). Another possibility is that a particle drifted
too far away and corrupted your whole mdcrd file (you can determine if this
happened by looking for * characters in your mdcrd file).

Some suggestions for future simulations:

- use ioutfm=1 to specify NetCDF trajectories (VMD can read these). This
is actually the default in Amber 16, since they are better than ASCII
trajectories in just about every way
​- use iwrap=1 to wrap coordinates back to the central unit cell (this will
prevent coordinates from getting too large in very long simulations). As
long as you use ioutfm=1, this isn't necessary unless you run *very* long
simulations (microsecond+ time scales)

> Production
> &cntrl
> imin=0,
> ntx=5,
> irest=1,
> nstlim=10000000,
> dt=0.001,
> ntf=2,
> ntc=2,
> temp0=300.0,
> ntpr=2500,
> ntwx=2500,
> ntwr=2500,
> cut=8.0,
> ntb=2,
> ntp=1,
> ntt=3,
> gamma_ln=2.0,
> /


Jason M. Swails
AMBER mailing list
Received on Thu May 12 2016 - 13:30:04 PDT
Custom Search