[gmx-users] RE:RE: How to calculate virial in gromacs-4.0? (xuji)

[xuji]

Hi
Thank you for your explaination. But when I calculate the virial in all the inner force loops, 
the short range VDW and QQ force contribution to virial can be claculated straight forward.
But when I claculate the bonded force, it seems difficult. Can you explain to me how to
calculate the covalent bonds, angles and dihedral forces in the inner loop? Especially how to 
define the r and f in the formula (r x f) in angles and dihedrals?


>Gromacs uses a more complicated, but also more efficient way to
>determine the virial.
>If you write your own program, I would the much simpler direct way
>of just calculating r x f in all the force loops.
>Gromacs 4.0 still always uses shift vectors, but in parallel the graph is not used.
>
>Berk

>> Hi all
>> 
>> I want to do pressure coupling in my own program.
>> So These days I read associated codes in Gromacs_4.0 and I find that it
>> must sum the virial first. In Gromacs-3.3.3 you use the intramolecular shift(graph)
>> to sum the virial of the covalent bonds, but in Gromacs-4.0 the graph variable seems to be NULL
>> when run the mdrun with domain decomposition. Is it the true? 
>> And if it's true how do you sum the virial with the periodic boundary condition 
>> in Gromacs_4.0 with domain decomposition? Can someone explain the way of summing 
>> virial in Gromacs_4.0 in detail? Or can you recommend some resoures about summing the virial with 
>> the periodic boundary condition? 
>> 	
>> Appreciate any help in advance!
>> 
>> Best
>> 	wishes!
>>