[Wien] GMAX for H

[Peter Blaha]

Uusally you have to use very small atomic spheres for H. Depending on 
your system probably between 0.55-0.7 bohr.
For such small spheres one needs a large cutoff in the Fourierseries to 
expand the coulomb potential properly.

The proper value for GMAX depends on your chosen RMT for H. The smaller 
the sphere, the larger GMAX.

However, this does NOT mean that you should expand the H sphere and 
reduce the O,N,C sphere (whatever your small H-X bond is), because this 
would require a larger RKMAX, which is much more expensive than a larger 
GMAX.

Otherwise test convergence as usual: put GMAX to 15, run the scf cycle 
and save the calculation. Increase GMAX to 20, run_lapw and compare eg. 
forces (or your desired property) between these calculations.

PS: GMAX should cost time mainly in lapw0, but NOT so much in lapw2.

John Appleton schrieb:
> Dear WIEN user,
> it is recommended in the user guide that for systems with H
> gmax should be 15 to 24. I use gmax=24 lapw0 and lapw2
> take a long time. I want to use 15 but I'm not sure.
> First why large gmax for H? Second, how
> to determine gmax so that you don't waste time.
>  
> Thanks
>  
> J. A.
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien

-- 

                                       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/
--------------------------------------------------------------------------