The ABINIT Users Mailing List ( CLOSED )

["Javier Lopez" <sklnw@hotmail.com>]

Dear Xavier Gonze,
Thank you for your quick reply!
I have tried the phonon calculation
with Abinit's linear response theory
implementation, but it's a heavy
calculation  for our machines. On the
other hand, the approximate hessian is
already calculated and so far it looks
like it's correct enough for what I want.
I would like to know the hessian units
to verify my results (and then
maybe change to linear response theory!).

Javier Lopez
Dpto Fisica Fundamental II
Universidad de La Laguna.


> From: Xavier Gonze <gonze@pcpm.ucl.ac.be>
> Reply-To: forum@abinit.org
> To: forum@abinit.org
> Subject: Re: [abinit-forum] Hessian units?
> Date: Wed, 19 Feb 2003 20:36:03 +0100
>
> Dear Javier Lopez,
>
> Javier Lopez wrote:
> > Dear ABINIT users,
> > In the log file of some relaxation processes (ionmov not 0),
> > the inverse hessian is written at the end of the log file. I think I 
> > understand the way it is written (first all the atomic coordinates, then 
> > the strain components) but I'm not sure about the units.
> > Are fractional coordinates used or rather cartesian coordinates?
> > (I have in mind extracting information about some gamma-point
> > frequencies from this matrix).
>
> If you have in mind extracting gamma-point frequencies, I suggest
> strongly that you use the linear-response part of ABINIT : the
> Hessian matrix is quite approximate, and useful only for the
> purpose of speeding the convergence ! On the contrary,
> the linear-response part of ABINIT is quite accurate and well-tested,
> please see the Tutorial number 5 ...
>
> Good continuation,
> Xavier Gonze


_________________________________________________________________
MSN. Más Útil Cada Día  http://www.msn.es/intmap/