[gmx-users] no temperature coupling

[Mark Abraham]

Cherry Y. Yates wrote:
> Dear GROMACS developers,
> 
> I would like to do a SPC-water simulation without temperature coupling. 
> I set tcouple to 300K and did Berenderson T-coupling first for 400 ps to 
> equilibrate the system. Then I copy confout.gro to conf.gro, set tcoupl 
> to no. However the Temperature of the later similation jumped from 300K 
> to 520K immediately and fluctuate around 520K instead of 300K. I wonder 
> (1) what could be wrong? 

The .mdp file you attached has gen_vel = yes, which if used for the 
second calculation will lead to regeneration of velocities and might 
explain this behaviour. You will need to supply .trr and .edr files to 
grompp for this second stage, of course.

> (2) how can I do a single simulation in which 
> first 400 ps using some temperature coupling scheme to equilibrate the 
> system, then do 100 ps without temperature coupling? I attached the 
> grompp.mdp file.

You can't do this in a single simulation, however it's straightforward 
to prepare your .mdp files in advance and have a script run a

grompp -f -c -p -o
mdrun -s
grompp -f -c -p -o -e -t
mdrun -s

sequence to deal with what you want to do.

Mark