[Pw_forum] hybrid functional error

[Peter Wang]

I am sorry I can not get you. Do you recommand me to use PBE0 instead of 
HSE06?

best,
Yun-Peng

-----原始邮件----- 
From: hannu.komsa at epfl.ch
Sent: Sunday, December 18, 2011 10:51 PM
To: pw_forum at pwscf.org
Subject: Re: [Pw_forum] hybrid functional error


Since n is always 1, and by quickly looking at the code it should be
that x>=0. Furthermore, x should be zero only if the density gradient
is zero and the screening parameter is set to zero (=screening length
infinity = PBE0). If you have done so, then you can fix the problem by
either using PBE0 or using a very small parameter value (btw., these
two approaches give slightly different results, although they should
give the same). If you haven't done so, then... ?


Hannu-Pekka Komsa
University of Helsinki

>
>> from expint : error #         1
>> bad arguments
>>
>> Where is the problem?
>
> it is where the error message says it is: in "expint" (flib/expint.f90).
> If you go inside the file, you will see that this error arises when
>    * n < 0 , or
>    * n=0,1 and x<=0, or
>    * n>1 and x<0
> The second case is the relevant one, I guess,
> Please specify which version of QE you are using: there have been
> several minor changes to the HSE routines since the first release.
>

_______________________________________________
Pw_forum mailing list
Pw_forum at pwscf.org
http://www.democritos.it/mailman/listinfo/pw_forum