Hi Mike,
In my opinion the info on grid generation is a little messy. In fact,
some time ago I needed it and ended up digging into the ATOM utility
sources in order to find out what was going on.
In a nutshell, the radial grid is generated according to
r_i = A * (exp((i-1)/B) - 1)
evaluated up to r_i < r_max.
Note however, that the ATOM utility writes 1/B, *** not B ***, in
the first few lines of the pseudo (*.psf) file.
Thus, you can figure out by yourself what the effects of changing
A or B are.
Best,
Roberto
On Tue, 1 Nov 2011, Michael Williams wrote:
Hi everyone,
I've been using SIESTA 3.0b-rev367 (or I guess it's called siesta-trunk-367 on
the website), and I was wondering if anyone has any suggestions for
(rationally) choosing values for the radial grid parameters, specifically when
using one of the VdW functionals. The systems that I'm interested in are
molecular crystals. The reason I ask is that I saw in the release notes of
SIESTA 3.1 that the 'Reparameterize.Pseudos' option has new default settings.
I read the manual, but I don't quite understand the effect that
'New.A.Parameter', 'New.B.Parameter', and 'Rmax.Radial.Grid' each specifically
have on the fineness of the radial grid. Thanks very much for any help you can
offer, and I apologize if I'm missing something obvious.
Mike
