AMBER Archive (2002)

Subject: Amber7, sander, restart, blast up

From: Jianhui Wu (wujih_at_BRI.NRC.CA)
Date: Wed Jun 05 2002 - 12:07:51 CDT


Dear amber user,

I run MD with sander in amber 7. Here is my setting:
*************************************************
 &cntrl
   ntc=2, ntf=2,
   cut=16.0, igb=2, saltcon=0.2, gbsa=1,
   ntpr=50, ntwx=500,
   nstlim = 1000000, dt=0.002,
   ntt=1, tempi=300.0, temp0=300.0, tautp=2.0,
   ntx=1, irest=0, ntb=0, ntx=5, irest=1,
   nscm = 1000,
*****************************************************

The job stoped at 1456ps due to machince crashed. I restarted the job
by setting

 ntx=5, irest=1,

In the next 200ps, the energies are similar to those
before the machine crashed. However, when I extract the coordinates (pdb
files) from the trajectory, I am surprised to find that the structure is
fine until 1456 ps and blast up right after the job restarted. I save a
frame per 1ps. At 1457ps, some atoms (less then 10) are in unusual
positions. At 1458ps, the whole structure blast up (totally mess).

(1) Did I do something wrong by setting ntx=5? (initial run, it is 1)

(2) How come the energies still reasonable but the struture blast up?

(3) Any way to restart the MD run from 1456ps again? The trouble is that
I did not backup the origial .rst file.

(4)It seems I have to extract the structure at 1456ps and run a new MD
from that. This means I have to run equivalence again? I intend to
merge the new trajectory file into the 1-1456ps one. Is this the best way
to do it?

Thanks a lot for your help,

Jian Hui

Lady Davis Institute