The ABINIT Users Mailing List ( CLOSED )

[Bryce Meredig <bmeredig@gmail.com>]

Matthieu--

Thanks for your reply. I was doing optcell=1, volume-only relaxation. I didn't know that the trace was involved; I guess it's more complicated than just the elements of the stress tensor...

Bryce

On Wed, Jul 29, 2009 at 1:34 PM, matthieu verstraete <matthieu.jean.verstraete@gmail.com> wrote:

which optcell have you used? Check the source file
src/67_common/fconv.F90 for more specifics on how the fmax is
calculated in each case. Your forces are clearly converged, so it's
the stress, eventually subtracting the trace or something like that,
which gives fmax (3.5201E-04)

Matthieu

On Wed, Jul 29, 2009 at 7:41 PM, <bmeredig@gmail.com> wrote:
> Hi all,
>
> Hunted around here for a while but couldn't find exactly what I'm looking
> for--apologies in advance if this is a very simple question.
>
> I ran a BFGS minimization of a cell, which converged after 9 steps. The code
> printed
>
> At Broyd/MD step   9, gradients are converged :
>  max grad (force/stress) = 3.5201E-04 < tolmxf= 5.0000E-04 ha/bohr (free
> atoms)
>
> I was under the impression that the convergence criterion (here 3.5201E-04)
> ought to relate to the largest single Cartesian force component on any of the
> atoms, or the largest stress component multiplied by strfact (the default,
> 100). Yet my forces are
>
> cartesian forces (hartree/bohr) at end:
>    1      0.00000022196000    -0.00000026184815    -0.00000036891432
>    2      0.00000014872891    -0.00000019322115     0.00000038078365
>    3      0.00004108870399     0.00007711205062     0.00000161583881
>    4     -0.00004129394759    -0.00007690312473    -0.00000162123890
>    5     -0.00005079422183    -0.00008924105679     0.00000990463720
>    6      0.00005062877651     0.00008948720020    -0.00000991110645
>  frms,max,avg= 4.5074698E-05 8.9487200E-05  -1.466E-08  1.384E-08 -4.976E-10
> h/b
>
> and my stresses are
>
>  Cartesian components of stress tensor (hartree/bohr^3)
>  sigma(1 1)=  2.33910376E-05  sigma(3 2)=  3.67303937E-06
>  sigma(2 2)=  1.91354379E-05  sigma(3 1)=  1.59285145E-06
>  sigma(3 3)= -5.30868121E-05  sigma(2 1)= -5.32246843E-05
>
> Could someone kindly explain the origin of the 3.5201E-04 that caused the code
> to reach convergence?
>
> Thanks much,
> Bryce
>

--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Dr. Matthieu Verstraete

European Theoretical Spectroscopy Facility (ETSF)
Dpto. Fisica de Materiales,
U. del Pais Vasco,
Centro Joxe Mari Korta, Av. de Tolosa, 72,   Phone: +34-943018393
E-20018 Donostia-San Sebastian, Spain        Fax  : +34-943018390

Mail : matthieu.jean.verstraete@gmail.com
http://www-users.york.ac.uk/~mjv500