[Sax-project-users] GW cost

[marsamos at democritos.it]

Dear user, GW calculation is parameter free (not as DFT calculations) 
so it is obviously more computationally demanding. The memory depends 
on wavefunction cutoff (size of file states), the cutoff polarizability 
and the number of k points. the file states growth linearly with the 
number of kpoints as well as the size of polarizability. For example, 
for a supercell of 11*11*11 Ang^3 a  cutoff for the polarizability of 6 
Ry, a cutoff for wavefunctions of 70 Ry and 800 bands you need around 2 
Gb per proc (with 16 proc) for doing the whole GW calculation.

bests

Layla


Quoting kazempoor at ph.iut.ac.ir:

> Dear All
> I am doubting about How expense is the GW approximation for supercell 
> calculation. I mean the memory is required is proportional to which 
> parameters( for example kpoint* ecut* polarizability) or not?
> thanks a lot
>
> Ali Kazempour
> Physics Department, Isfahn University of Technology
> _______________________________________________
> Sax-project-users mailing list
> Sax-project-users at qe-forge.org
> http://qe-forge.org/mailman/listinfo/sax-project-users
>



-- 
Dr. L. Martin-Samos
tel. +39 040 3787 429
CNR-DEMOCRITOS and
International School for Advanced Studies (ISAS-SISSA)
via Beirut 2-4
34151 Trieste Italy

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.