[QE-users] Primitive and conventional unit cell

Stefano Baroni baroni at sissa.it
Tue Sep 11 14:06:37 CEST 2018


Carina: two different calculations done with conventional and primitive cells will give _exactly_ the same results if performed with an equivalent number of k points. Conventional cells may have non orthogonal axes, which complicates a bit coding, and using the output of fields (such as the charge density ir SCF potential). Other than this, as said, the two calculations, if properly done, are totally equivalent, but for the compute time that is larger for conventional cells. Why some insist in wasting their computer budget is a mystery for me as well. SB

--
Stefano Baroni, Trieste -- swift message written and sent on the go

> On 11 Sep 2018, at 13:52, Carina Backtorp <carinabacktorp at hotmail.com> wrote:
> 
> Dear All,
> 
> I have seen many times in different papers that people use a conventional unit cell instead of the primitive unit cell when they do calculations.
> This means that they commonly use a structure with many more atoms than required for a primitive unit cell.
> I realize that there may be a symmetry difference between the conventional unit cell compared to the primitive.
> As people do in this way, which require much more computational efforts that should be of importance.
> 
> I wonder if someone might explain to me why I should not use a primitive unit cell.
> Will using a primitive unit cell give the wrong result? Will the calculated structure be wrong?
> I would be very grateful, if you could help me understanding the thinking about the computational technique regarding this.
>  
> Kind regards,
> Carina
> _______________________________________________
> users mailing list
> users at lists.quantum-espresso.org
> https://lists.quantum-espresso.org/mailman/listinfo/users
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20180911/825b0086/attachment.html>


More information about the users mailing list