[gmx-users] Different results with and without constraints for pr coupling.

[A.Rzepiela]

Dear all

The system, 75 hexadecane molecules, when simulated with 
Parrinello-Rahman (isotropic), Nose-Hoover and option   "constraints 
  = none" is freezing. When constraints are changed to "constraints 
    = all-bonds"  (lincs), this is not observed ( temp is above 
melting point ). This behavior does not depend on  tau_p value ( 
checked for 0.5 and 2 ) and is probably not connected to nose-hoover. 
When simulation is carried with nose-hoover and berendsen pressure 
coupling, freezing with all-bonds option is not observed . Also when 
berendsen temp. and pressure  coupling  is applied system is not 
freezing.
Starting structure was equilibrated with berendsen coupling, freezing 
starts after few hundred picoseconds. After that average pressure and 
temperature is correct, but fluctuations are very high. Bonded 
potential energy increases a few times, the rest drop down.
Simulations were done with gromacs3.3.1, 3.3.3  and CVS version

What can be the reason for different behavior with and without 
constraints ?

Andrzej Rzepiela