The ABINIT Users Mailing List ( CLOSED )

[Matthieu Verstraete <mjv500@york.ac.uk>]

Hello Nishimatsu-san,

I like the world domination. Definitely what Xavier had in mind.

It should be possible to feed the tetrahedron-generating routine the full 
kpoint grid and no symmetries (a la VASP) but I suspect the integrals 
could be done explicitly for a specific m value. There is a simple 
analytical relation between psi(k,G) and psi(S k, S G). Xavier had some 
notes on the way this was done in Abinit, and they should be in doc 
somewhere.

Careful: the tetrahedron stuff (and the rest too) can scale badly wrt 
number of kpoints, so the symmetry reduction is quite nice. The parallel 
execution could also be a bit thorny if the kpt are distributed 
differently than in the GS calculation itself. Please try to stay as 
modular as possible (with all due respect:).

Cheers and Good Luck

Matthieu

--
================================================================
Dr. Matthieu Verstraete                 mailto:mjv500@york.ac.uk
Dept. of Physics, University of York,     tel: +44 1904 43 22 08
Heslington, YO10 5DD York, United Kingdom fax: +44 1904 43 22 14