[gmx-developers] bonded interaction -- single molecule in parallel gmx?

[Anton Feenstra]

David van der Spoel wrote:
> On Tue, 2003-07-29 at 12:11, Fan Hao wrote:
> 
>>Hi, All:
>>        I am trying to parallelize my implicit solvent model in gmx
>>but I met a problem from bonded interaction. I use LINCS for the
>>constraint in my sd simulation The system is only a protein molecule so
>>when I did "grompp -np 2 ..., gmx try to split the molecule between the
>>shake border. 
[...]
> 
> how about using only constraints for the H atoms?
> 
> there is no other simple solution.

Since LINCS is rather fast, couldn't you do it on all nodes that the
molecule is on? Or would they not give identical solutions then?

(Of course this will in the end kill your dcaling due to Amdahls law...)


-- 
Groetjes,

Anton
  _____________ _______________________________________________________
|             |                                                       |
|  _   _  ___,| K. Anton Feenstra                                     |
| / \ / \'| | | Dept. of Pharmacochem. - Vrije Universiteit Amsterdam |
|(   |   )| | | De Boelelaan 1083 - 1081 HV Amsterdam - Netherlands   |
| \_/ \_/ | | | Tel: +31 20 44 47608 - Fax: +31 20 44 47610           |
|             | Feenstra at chem.vu.nl - www.chem.vu.nl/~feenstra/       |
|             | "If You See Me Getting High, Knock Me Down"           |
|             | (Red Hot Chili Peppers)                               |
|_____________|_______________________________________________________|