CSVR timecon value

[Ada Sedova]

Hi,

A question about the TIMECON value used in the CSVR thermostat. First, I'm assuming this is the same as tau in his papers. Is this correct? Second, according to a slide deck by Bussi found online:

http://scc.acad.bg/ncsa/articles/library/Library2016_Supercomputers-at-Work/CP2K_Publications/MolecularDynamics.pdf,

production-run values for TIMECON should be 1-100 fs. The default value, according the the CP2K manual, however, is 1000. I'm wondering why this is the default, would this be equivalent to no thermostat? What you get when you use values over 100? I tried 300 and I don't really notice a huge difference vs. 100.

Thanks,

AS