[Wannier] R: Issue with convergence

[Carlo Motta]

Dear Giovanni,

thank you very much for your kind reply.
Would you like to have a look at the wannier band attached? It is "almost" the same as the pw one. Would you call it entangled? It is actually separated from the lower ones, although it falls in an energy range where also other bands are present. Your comment would be very helpful for me!

As for the XSF, I attached it to this email (I had to cut out some NaN lines, they are all the same...). I can't really understand why is it behaving so, it used to work with other systems.

Many thanks again,

Carlo
--

Carlo Motta

PostDoc Researcher
School of Physics and CRANN,
Trinity College Dublin,
Dublin 2, IRELAND
http://sites.google.com/site/carlomottaphd
________________________________________
Da: Wannier [wannier-bounces at quantum-espresso.org] per conto di Giovanni Pizzi [giovanni.pizzi at epfl.ch]
Inviato: martedì 6 ottobre 2015 11.14
A: wannier at quantum-espresso.org
Oggetto: Re: [Wannier] Issue with convergence

Dear Carlo,
from the file you sent, it seems that after getting to 1e-14, the delta does not go increase again, but it actually drops to zero.
1e-14 is already extremely tight, and I think that numerically you cannot go lower than this due to the finite precision of float numbers.
Note also that the syntax you are using is in practice asking only for a s orbital on a single Pb atom, check if this is what you want (I don't think that the l=0:l=1 syntax is valid, even if Wannier90 is not complaining). Check also that the single band you are considering is actually disentangled from the other ones.

Finally, I don't know why the XSF contains NaN, it is difficult to say why it happens with the info you provided.

Best,
Giovanni



On 10/06/2015 11:48 AM, Carlo Motta wrote:

Dear Wannier90 users,

I write to ask for some help.
I am trying to wannierize the valence band of a perovskite material, I tried with many initial centers but it seems that I am not able to converge the calculation.
I reach a value of Delta~1E-14 but then it increases and does not converge. I think that in order to have a good result I should reach something on the order of E-20.

However, even if I stop the calculation once the E-14 Delta has been reached (by tuning num_iter), the XSF file contains all NaN, which is worrying me.

Could anyone give me a hint?

Thank you in advance,


--

Carlo Motta

PostDoc Researcher
School of Physics and CRANN,
Trinity College Dublin,
Dublin 2, IRELAND
http://sites.google.com/site/carlomottaphd



_______________________________________________
Wannier mailing list
Wannier at quantum-espresso.org<mailto:Wannier at quantum-espresso.org>
http://mailman.qe-forge.org/cgi-bin/mailman/listinfo/wannier




--
Giovanni Pizzi
Post-doctoral Research Scientist
EPFL STI IMX THEOS
MXC 340 (Bâtiment MXC)
Station 12
CH-1015 Lausanne (Switzerland)
Phone: +41 21 69 31124