[Wannier] How wannier functions are stored

[xiaochuan Ge]

Dear all,

I kindly ask someone to explain me how indeed wannier functions (WFs) are
stored in wannier90.

First please let me elaborate my own understanding. Though WFs are not
periodical, in practice using finite number of k-points (say nk) to
generate the WFs leads to a periodic length of nk*cell. So in order to
represent WFs with plane waves, the G-mesh must be nk times finer than the
one for bloch wavefunction. Is this really how WFs are represented in
wannier90? In addition, if one has to do some calculation with WFs, does he
has to deal with a vector nk times larger than the wavefunction? It would
be very appreciated if someone could summarize different strategies to
represent and manipulate WFs in the code and give me some reference for
understanding the details.

Thank you very much in advance.

===================
Dr. Xiaochuan Ge (Giovanni)
Center for Functional Nanomaterials
Brookhaven national laboratory
===================
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/wannier/attachments/20141230/af94c273/attachment.html>