AMBER Archive (2005)

Subject: AMBER: Error: vlimit exceeded for step

From: Hwankyu Lee (leehk_at_umich.edu)
Date: Mon Jun 06 2005 - 08:55:57 CDT


Dear Amber-users,

I got an error like below at 580ps point. After energy minimization
with and without restraints of solutes, I heated up this system
(~80,000 atoms including water) from 0K to 298K for the first 10 or
20ps with NVT. When I heated up the system, I used weak restraints or
did not use restraints of solutes because I got an advice, which is
that restraints of solute was too big leading to SHAKE failure. Then, I
started equilibration runs without any restraints with NPT.

Last time, when I got this error at the beginning of the MD running, I
posted this problem here. At that time, I got some advice and follow
those.
What I did was that

1) Whenever this error happens, I changed timestep from 0.002ps to
0.0016ps, and then from 0.0016ps to 0.001ps.

2) When I heated up the system, I tried using weak restraints of
solutes and also tried using no restraints of solutes.

Although I decreased timestep and tried using with weak restraints or
without restraints of solutes, I still have this error. I kind of
surprised that this happens at 580ps, not beginning of the simulations.
   Could you help me handle this problem? I really appreciate your
help in advance.

This is an error at the point of 580ps.
------------------------------------------------------------------------
-------
  NSTEP = 93750 TIME(PS) = 580.000 TEMP(K) = 298.19 PRESS =
    33.8
  Etot = -220356.2840 EKtot = 49808.2308 EPtot =
-270164.5148
  BOND = 877.7503 ANGLE = 2787.6343 DIHED =
1489.3510
  1-4 NB = 854.2081 1-4 EEL = -10081.5554 VDWAALS =
33965.8149
  EELEC = -300057.7180 EHBOND = 0.0000 RESTRAINT =
  0.0000
  EKCMT = 23054.6737 VIRIAL = 22448.7378 VOLUME =
830196.8591
                                                     Density =
  1.0082
  Ewald error estimate: 0.1015E-03
   
------------------------------------------------------------------------
------

  vlimit exceeded for step 93807 ; vmax = 23.91630977621595
  vlimit exceeded for step 93808 ; vmax = 40.75498760131579

      Coordinate resetting cannot be accomplished,
      deviation is too large
      iter_cnt, my_bond_idx, i and j are : 3 1 5140 5139
------------------------------------------------------------------------
---------
This is my input script.
-------------------------------
   imin = 0, irest = 1, ntx = 7,
   ntb = 2, pres0 = 1.0, ntp = 1,
   taup = 5.0,
   cut = 9, ntr = 0,
   ntc = 2, ntf = 2,
   tempi = 298, temp0 = 298,
   ntt = 3, gamma_ln = 1.0,
   nstlim = 5000000, dt = 0.001,
   ntpr = 1000, ntwx = 1000, ntwr = 10000
---------------------------------------

best,
Hwankyu.

On May 24, 2005, at 2:12 PM, Thomas E. Cheatham, III wrote:

>
>> I put wrong info about 20ps point. here, I attach 20ps energies again.
>> I really apologize for my mistake.
>> In 20ps, it has 299.88, which looks like reasonable temp.
>> could you look at this one more time?
>> -------------------------------
>> check COM velocity, temp: 0.002858 0.01(Removed)
>>
>> NSTEP = 10000 TIME(PS) = 20.000 TEMP(K) = 299.88
>> PRESS =
>> 0.0
>> Etot = -201648.0152 EKtot = 50211.0109 EPtot =
>> -251859.0262
>> BOND = 898.2060 ANGLE = 2473.6478 DIHED =
>> 1351.3450
>> 1-4 NB = 826.9863 1-4 EEL = -10156.4761 VDWAALS =
>> 32172.4335
>> EELEC = -280549.9268 EHBOND = 0.0000 RESTRAINT =
>> 1124.7582
>
> If this is correct, and you are indeed running with restraints (as
> pointed
> out on this reflector many times if you search the archive), likely you
> have a force constant that is too large leading to SHAKE failure. Try
> using a force constant for the restraints of less than 5 kcal/mol.
> Note
> that 1000 kcal/mol of restraint energy is a lot so make sure that you
> are
> restraining to what you want to be restraining too...
>
>
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to amber_at_scripps.edu
> To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu
>

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu