[gmx-users] Pressure coupling with non-rectangular box

[Erik Lindahl]

Justin MacCallum wrote:
> Hi,
> 
> I'm currently setting up a simulation of a membrane protien.  I want to 
> use a non-rectangular box to minimize the amount of lipid and water I 
> need.  I'm currently using a box with angles of 90 90 and 60 degrees. My 
> problem is that I don't know what pressure coupling option to use. I 
> would like to maintain the box with those angles and just scale the box 
> vectors.  I think anisotropic pressure coupling scales the X Y and Z 
> size of the box, leading to distortion.  Am I wrong about this?

If I remember correctly we tried to be smart and have the pressure 
coupling scale the actual box vectors (in contrast to x,y,z components) 
when you use anisotropic or semiisotropic scaling with all off-diagonal 
elements of the compressibility set to zero, but I might be wrong - test 
it and see what happens :-)

Cheers,

Erik