[gmx-users] problem regarding "mdrun" and "g_hbond"

[Mark Abraham]

> tc-grps             =  Protein  SOL   NA+
> tau_t               =  0.1      0.1   0.1
> ref_t               =  523      523   523

This is a bad idea and we see it regularly on this mailing list. Coupling
a single atom to a thermal bath is asking for trouble... what is the
temperature of a single atom? Do you really want to be adding/taking
random amounts of energy all the time?

It's possible this explains your problem when you move to higher
temperatures, but it makes everything else you've done suspect as well. A
better approach is to couple to Protein and Non-Protein groups.

Developers... how about a warning in grompp when a group has less than 10
atoms or some such?

Mark