AMBER Archive (2009)

Subject: Re: [AMBER] RE: same result

From: zgong.hust (zgong.hust_at_gmail.com)
Date: Thu Jan 08 2009 - 02:49:29 CST

Dear sir:
         It's very kind of you to answer my questions, I have learned a lot!
Thank you very much!


2009-01-08



zgong.hust



发件人： Ross Walker
发送时间： 2009-01-08 01:00:30
收件人： amber_at_ambermd.org
抄送：
主题： [AMBER] RE: same result
 
Dear Zhou,
I am sending this to the AMBER mailing list which is the best place for
these sorts of questions.
The reason you get the exact same trajectories is because you ran the exact
same simulation. Often this doesn't give the same results, especially if
running in parallel or on different machines due to round off differences.
However, run on the exact same machine in serial it should indeed give the
same results.
To obtain different trajectories you need to specify different random number
seeds - see the ig keyword in the manual. E.g. ig=12345654321 etc.
Alternatively you could start with slightly different starting structures.
In all cases you will need to check that the simulations have run long
enough for the trajectories to uncorrelated.
Good luck,
Ross

From: zgong.hust [mailto:zgong.hust_at_gmail.com]
Sent: Wednesday, January 07, 2009 5:16 AM
To: Ross Walker
Subject: same result

Dear sir
       I have used sander to do the molecular dynamics, but I met a problem
 I have do the 8 simulations with the same mdin and same start structure.
However, I got the 8 same results and 8 same trajectories, exactly same!
Why this happened? What should I do to get the different trajectories?

my mdin file is like this

  &cntrl
   imin = 0, ntb =0,
   igb = 1, ntpr = 20000, ntwx = 20000,
   ntt = 3,gamma_ln = 1.0,
   tempi = 300.0, temp0 = 300.0,
   nstlim = 100000000, dt = 0.001,
   cut = 999
  /
2009-01-07
  _____
zgong.hust
_______________________________________________
AMBER mailing list
AMBER_at_ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber

_______________________________________________
AMBER mailing list
AMBER_at_ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber