|
|
|
|
| 5:08:44 AM PDT - Sun, Sep 18th 2016 |
|
I am trying to perform a CCSD(T) calculation using DFT wfs as starting guess (since HF has too much spin contamination for the system I am interested). The system is quite large, so I need to reduce the memory footprint. I follow the suggestion in the documentation regarding 2EMET,
However, I get an error like "nwchem tce_energy invalid 2emet "
would anyone please suggest a method to reduce the memory footprint of a CCSD calculation with DFT initial wavefunctions?
|
|
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Dr. O. Baris Malcioglu,
University of Liege,
Bât. B5 Physique de la matière condensée
allée du 6 Août 17
4000 Liège 1
Belgique
|
|
|
|
|
| 7:44:09 PM PDT - Mon, Sep 19th 2016 |
|
Hi Baris.
Most likely you are using unrestricted wave function as it spin contaminant.
As described in docs http://nwchemgit.github.io/index.php/Release66:TCE#2EMET_--_alternative_storage_of_two-elect... there are three algorithms permitted for UHF, you can try any of them.
Or write your input file if you have already tried to do this.
Best, Vladimir.
|
|
|
|
|
|
|
|
|
| 2:58:58 PM PDT - Wed, Sep 21st 2016 |
|
CCSD with a DFT reference is not necessary a well-defined theory, particularly for systems with significant open-shell character, so you should consider replacing NWChem with a random number generator 
|
|
|