AMBER Archive (2007)

Subject: Re: AMBER: &dipole input and some questions on using polarizable force fields

From: Seongeun Yang (seongeun_at_korea.ac.kr)
Date: Sat Oct 27 2007 - 00:03:36 CDT

Thanks for your really kind reply, Ross.
And I'm sorry I skipped saying some previous steps before I used that input.
I used "saveamberparmpol", of course, when saving top and crd files.

Actually, I've used "solvateoct" with nonpolarizable force fields to save computation time, with no problem.
I tested again with "solvateoct NMA TIP3PBOX buffer", and the run had no problem with buffer = 9.0.
But with buffer = 8.0, the run stopped at t = 0.4 ps.
The input I used is exactly:
--------------------------------------
 Heating to 100 K for 20 ps
 &cntrl
  imin = 0, ntx = 1,
  irest = 0, iwrap = 1,
  ntpr = 50, ntwe = 50,
  ntwr = 50, ntwx = 0,
  ntave = 10000,
  ntf = 2, ntc = 2, ntb = 2,
  cut = 8.0, nsnb = 10,
  nstlim = 10000, dt = 0.002,
  tempi = 5.0, temp0 = 100.0,
  ntt = 1, tautp = 1.0,
  ntp = 1, pres0 = 1.0, taup = 1.0,
  ipol = 1,
 /
 &ewald
  indmeth = 1,
 /
-------------------------------------
The error messages in mdout is like this:
------------------------------------------------------------
 Cutoff list exceeds largest sphere in unit cell!!
 Big problems with imaging!!
 a,b,c = 24.4937980338375 24.4937980338375
   24.4937980338375
 alpha,beta,gamma = 109.471219000000 109.471219000000
   109.471219000000
 cutlist,sphere = 10.0000000000000 9.99955167855377
-------------------------------------------------------------
I don't know exactly why, but it seems that the box dimension matters in this case.
And the small difference between two numbers in the last line above seems to be just a numerical error.

For the "&dipoles" input, I put a space on the beginning of every line in mdin, of course.
But I found that was the real cause of the failure.
There should be no first space on lines between "&dipoles" and "/" !!
Now I got the problem solved thanks to your tip, but the rule of writing input is definitely inconsistent.
Because a number of spaces on lines between "&cntrl" and "/" do not matter at all.

Thanks again.

Seongeun

----- Original Message -----
From: "Ross Walker" <ross_at_rosswalker.co.uk>
To: <amber_at_scripps.edu>
Sent: Saturday, October 27, 2007 1:28 AM
Subject: RE: AMBER: &dipole input and some questions on using polarizable force fields

> Hi Seongeun
>
> > Does the box shape matter?
> > I tried "solvateoct NMA TIP3PBOX 8.0 iso" with the following md input.
>
> It shouldn't matter sander should be able to handle all box shapes with
> polarizable force fields. I don't think you need the iso option here. In
> fact I didn't even know it existed as an option to solvateoct so I don't
> know what it is doing here. Maybe it is benign, maybe not...
>
> > --------------------------------------------
> > &cntrl
> > imin = 0, ntx = 1,
> > irest = 0, iwrap = 1,
> > ntpr = 2, ntwe = 50,
> > ntwr = 50, ntwx = 0,
> > ntave = 1000,
> > ntf = 2, ntc = 2, ntb = 2,
>
> You should probably run at least 10 to 20ps at constant volume before
> changing to constant pressure. Also you are writing a restart file here
> every 50 steps (very very often) and not writing a trajectory file at all
> (ntwx=0). Are you sure this is what you want? I would typically write a
> restart file no more often than 5,000 steps or so.
>
> > cut = 8.0, nsnb = 10,
>
> nsnb is not not used by default anymore - it is only used if you set
> nbflag=0 so you don't need to specify it.
>
> > nstlim = 10000, dt = 0.001,
> > tempi = 300.0, temp0 = 300.0,
>
> Hot starting at 300 K, rather than heating gradually can sometimes cause
> you problems.
>
> > ntt = 1, tautp = 1.0,
> > ntp = 1, pres0 = 1.0, taup = 1.0,
> > ipol = 1,
>
> Did you use saveamberparmpol to save a prmtop with polarization info?
>
> > ---------------------------------------------
> > Then the run crashed after some steps with the messages:
> > SANDER BOMB in subroutine nonbond_list
> > volume of ucell too big, too many subcells
> > list grid memory needs to be reallocated, restart sander
>
> This is strange - can you post the steps? I suspect your system has blown
> up due to possibly being saved without using saveamberparmpol, being hot
> started, being started with constant pressure and/or having a bad starting
> structure. Try writing to the output every step (ntpr=1) and to mdcrd every
> step (ntwx=1) then you can visualize the trajectory and see what is going
> wrong.
>
> > Secondly, I want to print dipole values in output and added
> > the following lines right below the above input.
> > --------------------------
> > &dipoles
> > GROUP1
> > RES 1 2
> > END
> > END
> > /
> > --------------------------
> > But I got an error message in mdout such as
> > rfree: End of file on unit 5
>
> This is strange - your input above looks good. Although perhaps you don't
> have an initial space before the &dipoles and the /. Here is input I have
> used before that works:
>
> &dipoles
> Group1 - a single water molecule
> RES 1
> END
> Group2 - A selection of 2000 H2O mols from bulk
> RES 2 2000
> END
> END
> /
>
> Note the leading space before & and /
>
> Good luck
> Ross
>
> /\
> \/
> |\oss Walker
>
> | HPC Consultant and Staff Scientist |
> | San Diego Supercomputer Center |
> | Tel: +1 858 822 0854 | EMail:- ross_at_rosswalker.co.uk |
> | http://www.rosswalker.co.uk | PGP Key available on request |
>
> Note: Electronic Mail is not secure, has no guarantee of delivery, may not
> be read every day, and should not be used for urgent or sensitive issues.
>
>
> -----------------------------------------------------------------------
> 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