[gmx-users] Re: Using virtual site

라지브간디 rajiv at kaist.ac.kr
Wed May 1 15:16:41 CEST 2013


Dear
Mark, 

 


As per you suggestion, I have simulated a single CO as well as
CO-with-vsite in a box.

 


A single CO runs without any problem, BUT the CO-with-vsite
gives an same error as before:

 


Steepest
Descents:

   Tolerance (Fmax)   =
 1.00000e+03

   Number of steps    =  
     50000

Step=    0, Dmax= 1.0e-02 nm, Epot=
-3.73341e+04 Fmax= 3.38911e+05, atom= 2

Step=    1, Dmax= 1.0e-02 nm, Epot=
-4.54970e+04 Fmax= 4.95889e+05, atom= 2

Step=    2, Dmax= 1.2e-02 nm, Epot=
-6.18180e+04 Fmax= 9.34585e+05, atom= 2

Step=    3, Dmax= 1.4e-02 nm, Epot=
-1.10395e+05 Fmax= 3.03088e+06, atom= 1

Step=    5, Dmax= 8.6e-03 nm, Epot=
-2.09874e+05 Fmax= 1.09138e+07, atom= 2

Step=    9, Dmax= 1.3e-03 nm, Epot=
-2.42526e+05 Fmax= 1.45550e+07, atom= 1

Step=   11, Dmax= 7.8e-04 nm, Epot=
-2.67493e+05 Fmax= 1.76910e+07, atom= 2

Step=   14, Dmax= 2.3e-04 nm, Epot=
-2.76005e+05 Fmax= 1.88299e+07, atom= 2

Step=   16, Dmax= 1.4e-04 nm, Epot=
-2.81365e+05 Fmax= 1.95651e+07, atom= 2

Step=   19, Dmax= 4.2e-05 nm, Epot=
-2.83032e+05 Fmax= 1.97967e+07, atom= 1

Step=   23, Dmax= 6.3e-06 nm, Epot=
-2.83281e+05 Fmax= 1.98315e+07, atom= 1

Step=   26, Dmax= 1.9e-06 nm, Epot=
-2.83355e+05 Fmax= 1.98418e+07, atom= 2

Step=   28, Dmax= 1.1e-06 nm, Epot=  
       nan Fmax= 0.00000e+00, atom= 0

Stepsize too small, or no change in energy.

Converged to machine precision,

but not to the requested precision Fmax < 1000

 


Double precision normally gives you higher
accuracy.

 


writing lowest energy coordinates.

 


Back Off! I just backed up min.gro to ./#min.gro.1#

 


Steepest Descents converged to machine precision in
29 steps,

but did not reach the requested Fmax < 1000.

Potential Energy  = -2.8335531e+05

Maximum force     =  1.9841798e+07
on atom 2

Norm of force     =  1.6200760e+07

 


 


I am afraid that I am a making any mistakes of setting the
vsite? 


 


My
topology file vsite setting :

 


[ moleculetype ]


;
Name            nrexcl 

Other_chain_A       3

    

[ atoms ]

;   nr       type  resnr
residue  atom   cgnr     charge       mass
 typeB    chargeB      massB

; residue 157 CMO rtp CMO  q -1.6

     1        
 C    157    CMO      C    
 1      -0.75     12.011   ; qtot -0.75

     2        
 O    157    CMO      O    
 2      -0.85    15.9994   ; qtot -1.6

     3        
 I     157     CMO      I    
  3       1.6      0.0000

[ bonds ]

;  ai    aj funct    
       c0            c1
           c2          
 c3

    1     2     2
   gb_4

[ virtual_sites2 ]

; Vsite from        funct
  a             

 


3       1     2 
    1       0.4376625

 


 


My coordiante (.gro) file where i added the vsite xyz: Also included
the vsite atom info in atomypes as well as ffnonbonded.itp.

 


Gravel
Rubs Often Many Awfully Cauterized Sores

 
  3

 
157CMO      C    1   1.625   2.697  
0.490

 
157CMO      O    2   1.722   2.647  
0.482

 
157CMO       I     3   1.673   2.672
  0.486

 0.09660
  0.05040   0.00760

~
                     
            

 


Could you tell what mistakes I am making here ?


 


Thanks in advance.


 


Raju

 

 

On Wed, May 1, 2013 at ��������������� :



I mean simulate a single CO in a box. Then CO-with-vsite in a box.
If you 
can do that, then a bunch of problems are known to be solved. Leaping in 
the deep end with a complex system means you don't know where the trouble 
comes from when you have it. 

Mark 

On Wed, May 1, 2013 at 8:45 AM, ���������������
<rajiv at kaist.ac.kr> wrote: 

> Dear Mark, 
> 
> > You mean do the CO simulation and then add V-site upon
production run? 
> 
> Could you please more specific, what has to be done to avoid this
error? 
> 
>  > 
> On Fri, Apr 26, 2013 at 6:16 PM, ������������������������������
<rajiv at kaist.ac.kr> wrote: 
> 
> Maybe your starting configuration of atoms is too broken to be
useful. 
> Maybe you have some interaction on atom 2513 that doesn't work. 
> 
> Make a simple system of CO work, then do CO+vsite. Then apply those
lessons 
> to your complex system. 
> 
> Mark 
> 

 


More information about the gromacs.org_gmx-users mailing list