problem of natmmx

Posted on : April 29, 2016 (Fri) 18:15:40

by MOUNA

Hello dear Mr. Hisazuki
I always work with the code \"cpa2002v009c\"2015 version ; I am trying to study ZnO;
but the problem that does not work if specx NATM = 8;
The valid specx worth natmmx = 4, I tried to increase this value to 8; but it did not work.
Thank you tell me where is the problem
By MOUNA

[Re:03] problem of natmmx

Posted on : May 21, 2016 (Sat) 16:52:01

by Administrator

Dear Mouna,

ncmpmx is the maximum number of components that system contains. For example, if it is bcc Fe_1-x Co_x random alloys, the number of atoms in the primitive unit-cell is 1 but the number of the component atoms is 2 (Fe and Co). Then for such calculation natmmx=1 and ncmpmx=2 is the required minimum size. Larger is not the problem. msizmx is the maximum size required for the Green\'s function. Suppose that you have N atoms in the unit cell. For each atom, calculate m=(l_max+1)^2. Here l_max is the angular momentum you take into account. Normally l_max=2 (include up to d-scattering) or 3 (include up to f-scattering). Then some over m for all those atoms in the unit cell. The result gives the minimum size you need to perform the calculation. Therefore, if it is bcc Fe and you take up to d-scattering, msizmx=9. If it is hcp Co, msizmx=18.

[Re:02] problem of natmmx

Posted on : May 20, 2016 (Fri) 20:21:59

by togocha

dear administrator,
can you explain me how you had found that ncmpmx=72, msizmx=128??

[Re:01] problem of natmmx

Posted on : April 29, 2016 (Fri) 18:50:34

by Administrator

You also have to increase msizmx in specx. For example if you take into account up to f-scattering for all 8 atoms, msizmx=128, i.e. mxlmx**2*natmx, is the minimum size you need:
parameter
& (natmmx=8, ncmpmx=72, msizmx=128, mxlmx=4, nk1x=550, nk3x=300,
& msex=201, ngmx=15, nrpmx=650, ngpmx=650, npmx=350, msr=400)

On the other hand, if you need only up to d-scattering, msizmx=72 would be enough.