Re:Re: [AMBER] Does the frequency of output affect MD simulation in AMBER10?

From: Jianing <song_jianing_hi.126.com>
Date: Tue, 13 Apr 2010 13:17:46 +0800 (CST)

Hi Bill,
Thank you very much for your reply.
I have equilibrated my system for 5ns with NPT ensemble without restraints. yes, I run in parallel. The following is my input file for equlibration:
 
villin:5ns MD
 &cntrl
  imin = 0,
  irest = 0,
  ntx = 1,
  ntb = 2,
  ntp = 1,
  pres0 = 1.0,
  taup = 2.0,
  cut = 12.0,
  ntc = 2,
  ntf = 2,
  tempi = 300.0,
  temp0 = 300.0,
  ntt = 3,
  gamma_ln = 1.0,
  nstlim = 2500000, dt = 0.002
  ntpr = 100, ntwx = 100, ntwr = 1000
/
I will rerun my system to see whether this is helpful.
Thanks in advance.
Jianing
 

 



2010-04-13 02:13:25"Bill Ross" <ross.cgl.ucsf.edu> д
>How did you equilibrate before the 2 'equivalent' runs?
>If running in parallel, runs can diverge because of different order
>of execution involving timing of interprocessor communication. Such
>divergence may be more extreme in a simulation that wasn't well-
>equilibrated first. It could be interesting to re-run one of the
>existing trajectories, and see how it compares to the original run.
>
>Bill
>
>_______________________________________________
>AMBER mailing list
>AMBER.ambermd.org
>http://lists.ambermd.org/mailman/listinfo/amber

_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Apr 12 2010 - 22:30:02 PDT
Custom Search