Self interaction correction scheme (SIC)

[Andrei Buin]

I have question regarding the SIC schemes. 

I'm using one which is proposed by VandeVondele  and Michiel Sprik with 
a=0.2 and b=0(exchange term).

My question is, based on the formalism it should not matter what method i'm 
using PZ or Mauri (i.e. EXPLICIT_ORBITALS, MAURI_SPZ, MAURI_US), it should 
all give the same answer as long as i'm using only scaled Hartree term in 
compensating SIC.

However problem comes as i start using EXPLICIT_ORBITALS, it slows down 
things tremendously, and i cant get data out of it in reasonable time 
(large system). I expect to be the case since it does SIC subtraction for 
each orbital. I'm worried that i will get a different result.

My question is, should i expect different results based on different SIC 
schemes as long as i'm using Hartree scaled constant only, in other words 
am i understanding things correctly ?


With Best Regards, Andrei Buin.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20150429/cf1b338a/attachment.htm>