[gmx-users] genion II

[Pedro Alexandre de Araujo Gomes Lapido Loureiro]

Dear users,

I managed to get an answer to part of my questions in the e-mail by hugo:

">>You are probably using conjugate gradients with single precision mdrun. 
>>This can be problematic for algorithmic reasons. If you are just preparing 
>>for a MD run, conjugate gradients is overkill, and you can use steep. 
>> 
>>Conjugate gradients minimizes better, but you typically need this only 
>>when preparing for normal mode analysis, in which case you need doube 
>>precision anyway."

But if I run a MD after some steepest descent, there is a halt in the program 
due to deviation in LINCS.
For those who are following this discussion, my previous e-mail was: 

""""after using the program genion, to insert a Cl ion a box containing a 
peptide 
in water I got the following warnings/errors: 

1)After runnig the program grompp (mshion.top and mshion1-1.gro already contain 
the ion): 

"Warning: atom names in mshion.top and mshion1-1.gro don't match (Cl - OW) 
... 
(more than 20 non-matching atom names) 
Warning 7 [file "mshion.top", line 33]: 
  33145 non-matching atom names 
  atom names from mshion.top will be used 
  atom names from mshion1-1.gro will be ingnored" 

In the .top file I had already added one Cl atom and deleted one water 
molecule. 
Am I doing somethig wrong? 

2)When using the random insertion mode in the genion, later when I run a 
conjugate gradient minimization (after having runned a steepest descent without 
problems): 

"Conjugate gradients: 
   Tolerance         =  1.00000e+02 
   Starting Energy   = 1.639584000000000e+09 
   F-Norm            =  6.35435e+07 

Step 1, E-Pot = 1.6376719360e+09, F-max =  5.34807e+10 
Fatal error: ci = -2147483648 should be in 0 .. 2495 [FILE nsgrid.c, LINE 214]" 

What does it mean ? 
There was an e-mail with this kind of error in the list, but I was not able to 
find any reply. 

3)Using the electrostatic insertion mode gives the same error, when I use mdrun 
to run a conjugate gradient minimization """"

Cheers, Pedro.

--
Pedro Alexandre Lapido Loureiro
Laboratório de Física Biológica
Instituto de Biofísica
UFRJ
Brasil