[Wien] equavalent atoms and memory/speed

[Peter Blaha]

> I been wondering about this. Does one gain or loose in speed or in memory when using equavalent atoms. I have a slab now that has 70 atoms. I have 35 lots of 2 pairs of equavalent atoms. Am I gaining by this or am I better of labeling them so that I have 70 individual atoms. What are the pros and cons between the two cases.

Yes, you can gain a lot.
When you have 70 "individual" atoms, you definitely have only P1 symmetry.
This means:
i) You definitely use the "complex" version (speed factor of almost 4)
ii) You need a "full" k-mesh, (speed-factor X, ~ number of symm.op.)
iii) scf-convergence is more difficult, because it takes longer until
     equivalent atoms are really equivalent (instead of "forcing" them
     to be equivalent).

I strongly recommend to "use" the symmetry of your system.

                                      P.Blaha
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-15671             FAX: +43-1-58801-15698
Email: blaha at theochem.tuwien.ac.at    WWW: http://info.tuwien.ac.at/theochem/
--------------------------------------------------------------------------