Problem (bug?) in the extended Lagrangian metadynamics with the Langevin thermostat

[Sal]

Dear CP2K developers/users,
I am using extended Lagrangian metadynamics and it looks like there is 
something wrong with how the hills are spanned/written when I try to use 
Langevin thermostat for the metadynamics.
The problem is that when Langevin is turned on hills are spanned twice. I 
have the following lines in my output (without anything between them). It 
seems like it's adding 2 hills instead of 1:

 METADYN| Hills number (17) added.


 METADYN| Hills number (18) added.



In the restart file under the        &SPAWNED_HILLS_POS     directive it 
prints triples of the positions. Something like this:

       &SPAWNED_HILLS_POS
             1.3038877512362206E+00
             1.2265614190873904E+00
             1.2265614190873904E+00
             1.2265614190873904E+00
             1.2604256138213792E+00
             1.2604256138213792E+00
             1.2604256138213792E+00

 
None of this happens when I turn off the Langevin and use the default 
(velocity re-scaling) thermostat. 
This happens across versions and with the QUICKSTEP module.
So, I'm wondering if this is how it should be. If not, is this just an I/O 
issue or the dynamics is affected too?
I have attached all the files for you to check.

Regards,
Saleh




-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20170327/3051db51/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: xmtd.tar.gz
Type: application/octet-stream
Size: 3607811 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20170327/3051db51/attachment.obj>