Query for the cell vector matrix

[Unmesh Mondal]

Hello,

I want to use I-Pi(v2) with cp2k6.1.0 for  PIGLET simulation of a crystal. The crystal structure of my system is triclinic which does not have an upper triangular cell vector matrix. Can I- Pi be used for such a system?  Additionally, can someone direct me to an example for the same? 

Thanks

Unmesh

[Michele Ceriotti]

Of course. You can map *all* crystal structures to an upper triangular matrix, it is just a convention. Now, doing so can be a bit of a pain, but you can have a look at 

ipi/utils/io/backends/io_xyz.py

where it does the conversion from a generic cell matrix (GENH) to the internal upper-triangular representation. 

Note however it might be better to also update the input of CP2K to reflect this change, as dramatic changes of the cell require deep reinitialization of the code,

and that tends to be a dark area for the various i-pi interface implementations.