<div dir="ltr">Hi all, My name is Sam and I am currently studying for a PhD at Cardiff University. I've just joined this group and am still relatively new to computational chemistry, so please be kind! :-) I'm trying to run some MD simulations on the solid phases of water using the TIP4P/Ice model in cp2k. Using the TIP4P example in the Fist directory, I have managed to get a working MD run in the NVE ensemble. I have two questions regarding this issue: 1) I've noticed that the input file from the test directories requires that the virtual site has its coordinates defined in the xyz/pdb file. However, most of the coordinate files I have for liquid water and the solid polymorphs of water contain coordinates for three sites (ie the two Hydrogens and the one Oxygen) only. Is it possible to generate the virtual site 'on the fly' in cp2k, or is there an easy way of generating four-site water structures? 2) I would like to run these simulations using the NPT ensemble with a flexible cell, in order to see how pressure affects the dynamics. However, on running the TIP4P water example on cp2k, I get the following error: *** 14:51:23 ERRORL2 in constraint:shake_roll_control processor 0 *** *** err=-1000 UNIMPLEMENTED, Virtual Site Constraint/Restraint not *** *** implemented for SHAKE_ROLL! *** It is clear from this that the Shake algorithm doesn't like the constraints forced upon the virtual site, however it seems to me that these are imperative for keeping the charge located near the oxygen. Is there any way around this? ..and please don't say, just use TIP3P!! :-) Cheers, Sam </div>