[gmx-users] Charge groups, PME and ions [SOLVED]

[Andreas Kring]

David van der Spoel wrote:
> Andreas Kring wrote:
>> Hello all.
>>
>> I have read the manual and searched the archives, but I am still a 
>> little confused so here goes:
>>
>> I am going to perform an MD simulation of a molecule consisting of 38 
>> atoms and with a total charge of -4e (water as the solvent). For this 
>> molecule it is not possible to make charge groups with a total charge 
>> of zero (or and integer number).
>>
>> Is this best solved by using a separate charge group for each atom?...
>> and is this allowed since the charge of such a charge group is not zero?
>>
>> According to the manual, the charge groups are used to make neighbor 
>> lists... but these are only important with respect to the 
>> computational speed as long as PME is used to deal with electrostatic 
>> interactions? Or have I misunderstood this?
> 
> For such a small molecule you can have separate charge groups for each 
> atom when using PME. Efficiency gains will be minimal when grouping 
> atoms into charge groups.
> 

Great! Thanks a lot.

/Andreas