[gmx-users] how many groups shall I use to control the temperature?

[陈功]

Dear justin,
     I have doubts that whether I should use two groups to control the temperature or just one since there are just 604 hundred Protein beads and 60000+ solvent beads. It is said that two groups are more accurate but when when i split them into two groups "Protein nonProtein", I found that fluctulation of temperature in group "Protein" are obvious, Im afraid that would have an effect on accuracy on my simulation results. So how many groups shall I use to control the temperature?
    Thanks in advance
regards,
chen