[Wien] AFM calculations: runsp_lapw and runafm_lapw

[Stefaan Cottenier]

>  According to the usersguide section 4.5.4 (antiferromagnetic calculations), I see the steps to perform AFM. But I see some 
>people don’t do like that. They don’t setup case.struct_supergroup. They just flip the spin in case.inst, and perform runsp_lapw 
>with spin polarization directly.
>
>  What are the differences between them?
>
>  
>
The first method your describe is twice as fast as the second (no need
to run lapw1 -up and -dn separately). But it is (a bit) more difficult
to set up, and perhaps not bug-free (but that's easy to test by doing a
few more regular runsp iterations after convergence to check if it not
changes any further).

If you have a small, fast case: just use normal runsp. If you have a big
case where saving half of the computational time does matter: use the
AFM version.

Stefaan


Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm