[gmx-users] Question about 'mdrun -nosum'

[Thomas Schlesier]

Dear all,
i have a small question regarding the '-nosum' option of 'mdrun'.
The manual states:

For a global thermostat and/or barostat the temperature and/or pressure 
will also only be updated every nstlist steps. With this option the 
energy file will not contain averages and fluctuations over all 
integration steps.

Second sentence is clear to me. But the first sentence give me some 
thoughts.
I think this would introduce some errors, but what is about the 
magnitude of these?
I would expected that the errors are in the same order of the errors in 
the forces due to the neighborlist search (error due to the fact that 
one uses 'nstlist=5' instead of 'nstlist=1') and would be more or less 
negilegible, if one doesn't use a very larger 'nstlist'-value.

But to be on the save side i wanted to ask.

Greetings
Thomas