feenstra at chem.vu.nl
Tue Mar 4 16:01:17 CET 2003
Berk Hess wrote:
> Another option is to pressure scale the center of mass of the frozen parts
> of each molecule. But this might be complicated to implement.
> Personally I would only need scaled coordinates, but maybe we should
> two options?
I can see four possibilities:
1) no scaling (current implementation)
2) 'normal scaling': frozen coordinates scale like normal ones
3) COM scaling: COM of frozen groups 'scale' along with normal coordinates.
4) group scaling: COM remains frozen, group 'size' is scaled.
I don't know if option 4 makes sense, it seems it might be useful for
some systems where a large group (protein?) should remain in a fixed
conformation and fixed location (in a bilayer), without the pressure
scaling distorting the interface between frozen and normal atoms.
I believe option 3 is what you describe above. Option 4 is probably
even worse to implement ;-)
Long-term coordinate drift might be minimized by treating the COM
of frozen groups separately. That will not eliminate the drift, but
probably postpone it till much longer timescales (?).
While we're at it, we don't yet have an option to freeze *only* the
COM of a group, do we? That might be handy also.
After reading the gmx-users discussion (between Berk & Bert), I wonder
if it isn't possible to add additional constraints for the 1/2D frozen
particles, which would not allow them to move in 1/2D during constraints
| | |
| _ _ ___,| K. Anton Feenstra |
| / \ / \'| | | Dept. of Pharmacochem. - Vrije Universiteit Amsterdam |
|( | )| | | De Boelelaan 1083 - 1081 HV Amsterdam - Netherlands |
| \_/ \_/ | | | Room P275 Tel: +31 20 44 47608 Fax: +31 20 44 47610 |
| | Feenstra at chem.vu.nl - www.chem.vu.nl/~feenstra/ |
| | "I've Got Two Turntables and a Microphone" (B. Hansen)|
More information about the gromacs.org_gmx-developers