[gmx-users] errors depending on (lack of) solvent

[Anton Feenstra]

gianluca santarossa wrote:
> Sorry, but I didn't undersand. Is there a way to split a simulation of a
> single molecule on, for istance, 4 machines, without using MPI?

No, communication between CPU's is done via MPI, always. But, the limitation
is in solving (distance) constraints (e.g., for constant bondlengths).

> My simulations consist of dynamics of quite small molecules ( 100 or 200
> residues )  for a long time. Do you think that parallel computation
> would be unuseful in my case?

If it is in vacuum, you can get to 100's of ns's in a week or so on a
single CPU. I don't see what you might gain by adding a few factors
in speed to that. If you have (explicit) water, you have many more
molecules and you can go up to 8-16 CPU's (depending on the amoount
of water, and the type of machine you have and the hard- and software
setup of it).


-- 
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)                               |
|_____________|_______________________________________________________|