The better method to do antiferromagnetic calculation is : spin-polarized wthout afm option.

From the wien mailing list :[Wien] Please help for AFM calculation

Madhav Ghimire Sat, 5 May 2012 18:44:25 +0900

Dear all,
     I am trying to perform the AFM calculations on iridate compounds a
number of times by following the faq provided in
http://www.wien2k.at/reg_user/faq/afm.html.  For that, I selected the afm
calculation from w2web and edit the case.inst (as guided in wien2k
userguide for NiO case). I flipped the spin for Ir and equal occupancy for
O atoms as shown below.
Ir
Xe 5
4, 3,3.0  N
4, 3,3.0  N
4,-4,4.0  N
4,-4,4.0  N
5, 2,2.0  N
5, 2,2.0  N
5,-3,1.0  N  <=== spin flipped
5,-3,2.0  N  <=== spin flipped
6,-1,1.0  N
6,-1,1.0  N
Na
Ne 1
3,-1,1.0  N
3,-1,0.0  N
......
......
......
O
He 3
2,-1,1.0  N
2,-1,1.0  N
2, 1,1.0  N
2, 1,1.0  N
2,-2,1.0  N  <=== equal occup for up/dn
2,-2,1.0  N  <=== equal occup for up/dn
O
He 3
2,-1,1.0  N
2,-1,1.0  N
2, 1,1.0  N
2, 1,1.0  N
2,-2,1.0  N  <=== equal occup for up/dn
2,-2,1.0  N  <=== equal occup for up/dn

<=== indicates the region where I flipped the spin/equal occup for up/dn in
case.inst.
After saving this, I continue with lstart upto spin-polarized mode. Then
selected antiferromagnetic calculation>>
continued with >> yes continue with afminput
   >> x afminput
   >> specified TRANSLATION VECTOR as 0.0, 0.0, 0.66 for Ir.
While executing the command, I got the following information:

case.struct_supergroup present
 The super and subgroups are KLASSENGLEICH
 You must specify a translation vector which transforms the
 spin-up into the spin-dn atom (e.g.: 0.5,0.5,0.5 for AFM bcc Cr)!
rrot not found
0.007u 0.000s 0:00.01 0.0%      0+0k 8+0io 0pf+0w

Here, I found that rrot not found.
Due to this, case.inclmcopy_st is completely empty and shows an error in
scf cycles as

>   stop error: the required input file naioafm.inclmcopy for the next step 
> could not be found

Could anyone help me to identify where I did mistake and how I have to
generate the case.struct_supergroup. I will be very thankful and glad to
you.
Thanks in advance
-- 
M. P. Ghimire

Reply by  Stefaan Cottenier

 Mind the 'or' at the end of step 5. The most straightforward procedure 
is to go up to step 5 only (i.e. use runsp_lapw, not runafm_lapw).

Step 6 can save you one half of the computation time, but is not so 
straightforward. If you run into trouble here, just go for the longer 
and easier procedure, and stop after step 5.

Stefaan