[Pw_forum] question about press_conv_thr

[Stefano de Gironcoli]

dear Jeffrey McMahon,

the bulk modulus is defined as K = V dP/dV, compressibility is its inverse.
hence a given absolute tolerance on the pressure gives a  relative 
tolerance on the volume determination proportional to the 
compressibility of the material.
I think that for "normal pressures" where bulk moduli (or 
compressibilities)  do not  change by order of magnitude as you increase 
the pressure then the absolute threshold is a better criterion. If you 
deal with ultra-high pressure and the bulk modulus becomes proportional 
to the external pressure you may be right and a relative criterion may 
be what you actually want.
Anyway it is something that you can tune on input on the basis of an 
estimated value of compressibility and the relative accuracy on volumes 
that you desire.
best,
  stefano

Jeffrey McMahon wrote:
> Dear all,
>
> Which is more important for a vc-relax calculation with external pressure,
> the absolute or relative magnitude of press_conv_thr? I am thinking the
> latter, because if I were to perform a vc-relax at a pressure of 2000 kbar
> it would seem as if the default value of press_conv_thr=0.5 is fine.
> However, if the pressure were 20000 kbar then press_conv_thr=0.5 seems
> rather small. So at very high pressures is a relative convergence value,
> such as 1e-4*P (e.g., 2 at 20000 kbar), a more appropriate quantity? Thanks
> in advance.
>
> Jeffrey M. McMahon
>
> The Institute for Condensed Matter Theory
> Department of Physics
> University of Illinois at Urbana-Champaign
> Urbana, IL 61801-3080
>
>   
> ------------------------------------------------------------------------
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum
>