[Pw_forum] about matching grid's

Axel Kohlmeyer akohlmey at cmm.chem.upenn.edu
Mon Sep 8 21:11:04 CEST 2008


On Mon, 8 Sep 2008, Ezad Shojaee wrote:

ES> 
ES> Hi
ES> i have a technical question, maybe so trivial. 

ES> Imagine our unit cell consists of two different atoms which their 
ES> pseudo file are constructed in different # of points as a radial 

there is no relation between the grid for the pseudopotentials and
the selection of the grids for density and wavefunction.

ES> mesh(coarser near their nucleus ..) Is matching the grid points for 
ES> the scf calculation in the cell a problem? How does the code do 
ES> this? I mean how does the code choose the grid for scf and what is 
ES> it's relation to the atomic grids?

the grid for wavefunction is chosen based on the wavefunction cutoff,
box size and the fourier transform library. for computational efficiency, 
you want the smallest possible number of grid points that allows to 
represent the highest frequency plane wave determined through the 
wavefunction cutoff. this grid is compared to a list of grids supported 
grids by FFT library and the smallest grid that has equal or larger
number of grid points is chosen. similar for the density cutoff
(if not predetermined by ecutwfc*4, which is always supported...)

cheers,
   axel.


ES> Thanx
ES> _________________________________________________________________
ES> Invite your mail contacts to join your friends list with Windows Live Spaces. It's easy!
ES> http://spaces.live.com/spacesapi.aspx?wx_action=create&wx_url=/friends.aspx&mkt=en-us

-- 
=======================================================================
Axel Kohlmeyer   akohlmey at cmm.chem.upenn.edu   http://www.cmm.upenn.edu
   Center for Molecular Modeling   --   University of Pennsylvania
Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323
tel: 1-215-898-1582,  fax: 1-215-573-6233,  office-tel: 1-215-898-5425
=======================================================================
If you make something idiot-proof, the universe creates a better idiot.


More information about the Pw_forum mailing list