[Pw_forum] How to select unit cell for optimization calculation?

[Youssef]

Dear Max,

I think that this rule is a myth,  for the simple reason that more 
ecutwfc=more plane waves = more energy, In addition to that , the more 
system is big the bigger is the energy (and the energy jumps vs 
ecutwfc).
I suggest you to rerun the calculations with tstress = .true.  &     
tprnfor = .true. , then you look after on the impact of ecutwfc upon 
force and stress tensor.

By plotting your data, I suggest that ecutwfc threshold is laying 
between 50 to 150 Ry.

On Fri, Jul 10, 2015 at 1:02 AM, max <aquiles011 at gmail.com> wrote:
> Dear Youssef and Giovanni,
> 
> Thank you very much, I have doubt about convergence energy.
> 
> I run scf  calculations for ecutwfc from 10 to 250 Ry using 
> ecutrho=4*ecutwfc and I get the values bellow ecutwfc vs etotal 
> (values in Ry). I read that the convergence will be achieved when the 
> difference between two adjacent scf calculation is smaller than 0.1 
> meV [1] or 1 mRy [2], but this is for unit cell or for atoms (divide 
> the energy difference between the number of atoms in unit cell)?
> 
> Other question, is normal that the convergence value for ecutwfc be 
> more than 200 Ry?
>  
> ecutwfc etotal
> 10    -941.67085713
> 15    -1005.07732548
> 20    -1038.00240543
> 25    -1058.8736733
> 30    -1073.25016374
> 35    -1082.62081233
> 40    -1089.10688841
> 45    -1093.8583202
> 50    -1097.63582121
> 55    -1100.63775846
> 60    -1103.0243432
> 65    -1105.15423136
> 70    -1107.08545327
> 75    -1108.79020201
> 80    -1110.27166937
> 85    -1111.59023819
> 90    -1112.768963
> 100    -1114.66472444
> 110    -1116.05760199
> 120    -1117.06611139
> 130    -1117.76952606
> 140    -1118.25442342
> 150    -1118.58413784
> 160    -1118.80480982
> 170    -1118.95194442
> 180    -1119.0489105
> 190    -1119.11208313
> 200    -1119.15329918
> 210    -1119.17990393
> 220    -1119.19703143
> 230    -1119.20790385
> 240    -1119.21481705
> 250    -1119.21915626
> 
> [1] https://www.mail-archive.com/pw_forum%40pwscf.org/msg11093.html
> [2] http://qe-forge.org/pipermail/pw_forum/2013-February/101122.html
> 
> On Tue, Jul 7, 2015 at 7:35 PM, Youssef <aharbil at gmail.com> wrote:
>> Dear Max,
>> 
>> You can use the following 14 atoms coordinates for spinel AB2O4 
>> (fd-3m origin choice 2), replace "x" with oxygen parameter from 32e 
>> wykoff site (0.2627 for your case)
>> 
>> A	 0.125	 0.125	 0.125
>> A	 0.875	 0.375	 0.375
>> B	 0.5	 0.5	 0.5
>> B	 0.25	 0.75	 0
>> B	 0.75	 0	 0.25
>> B	 0	 0.25	 0.75
>> O	 (x)	 (x)	 (x)
>> O	 (1.25-x)	 (x+0.5)	 (1.75-x)
>> O	 (x+0.75)	 (x+0.25)	 (1.5-x)
>> O	 (x+0.25)	 (1.5-x)	 (1.75+x)
>> O	 (1.75-x)	 (1.25-x)	 (x+0.5)
>> O	 (x+0.5)	 (1.75-x)	 (1.25-x)
>> O	 (1-x)	 (1-x)	 (1-x)
>> O	 (1.5-x)	 (x+0.75)	 (x+0.25)
>> 
>> 
>> to be used with  ibarv = 2 and ATOMIC_POSITIONS alat (not crystal) , 
>> I didn't test it in QE, please confirm if it works!
>> 
>> Youssef Aharbil,
>> Laboratory of Physics and Chemistry of Materials
>> Faculty of sciences Ben msik, Casablanca
>> Morocco
> 
> 
> 
> -- 
> El éxito solo es para los que se arriesgan por sus objetivos.
> 
>      Máximo Ramírez Mateo
>           Desarrollador de Software
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20150710/e63fd161/attachment.html>