The ABINIT Users Mailing List ( CLOSED )

[St Druid <tq02ksu@gmail.com>]

Dear Luo,

Thank you for your reply, the criterion tolvrs=1.0d-18 can expedite
the convergence in phonon's  nscf calculation.

It is known that the first DATASETs in both phonon and el-phon are
generate a ground-state ware function. but the tolerant criterion
tolwfr is even harder to reach. I just wondering what is the insight
in the different tolerant criterion.  Can I use tolvrs=1.0d-18 in
el-phon 's ground-state calculation.

On Fri, Jun 5, 2009 at 1:53 AM, <6671011@163.com> wrote:
> Dear Dong,
>
> Why do you think tolvrs=1.0d-18 can guarantee a fully relaxed structure?  
> It is
> only a tolerance for the SCF; you need to use a low tolmxf, the direct
> criterion for geometrical optimization, instead.
>
> Many causes can impove the SCF convergence, like tsmear.  You'd better 
> attach
> your input file.
>
>
> Gunagfu Luo
>
>
> --------------------------------------------------------------
> Dear abinit users:
>
> I perform el-phon calculations using abinit 5.73 with GGA FHI pseudos.
>
> I am sure that the structure is reasonable, and fully relaxed using
> tight criteria tolvrs 1.0d-18.
> but when i am performing el-phon calculation, the gs calculation with
> tight criteria tolwfr 1.0d-16( even low to 1.0d-22) can not converge.
>
> I am wondering that what's wrong with the calculation. and why is the
> gs criteria tolwfr 1.0d-16 in el-phon, but tolvrs 1.0d-18 is phonon
> calculation.
>
> Regards.
>
> Thanks in advance.
>
>
> --
> B.W. Dong
> State Key Laboratory of Superhard Materials,
> Jilin University, Changchun, China. 130012
>
> tq02ksu@gmail.com [ dbw07@mails.jlu.edu.cn ]
>



-- 
B.W. Dong
State Key Laboratory of Superhard Materials,
Jilin University, Changchun, China. 130012

tq02ksu@gmail.com [ dbw07@mails.jlu.edu.cn ]