[CP2K-user] Questions about LS_SCF calculation

[ma...@gmail.com]

Hi,

I'm a new user of cp2k v8.2. Recently I'm interested in the LS_SCF method 
in cp2k because it seems pretty mystical of conducting DFT calculations for 
a million atoms. I'm wondering:
(1) Compared with the conventional DFT, will LS_SCF lose a lot accuracy for 
calculations such as geometry optimization and formation energies for 
defects?
(2) There are four PURIFICATION_METHOD 
<https://manual.cp2k.org/trunk/CP2K_INPUT/FORCE_EVAL/DFT/LS_SCF.html#list_PURIFICATION_METHOD> for 
LS_SCF. The default one is "SIGN", but it seems PEXSI is also an advanced 
method. I'm wondering could someone please provide some suggestions about 
which one is relatively better to use in cp2k v8.2?

Thanks,
Hongyang
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20210622/6447d6e1/attachment.htm>