Frozen orbitals in TCE


Just Got Here
I'd like to ask if anyone knows how to freeze arbitrary orbitals when running calculations (e.g. CCSDT) with the TCE?
I have successfully done this with (e.g.) MP2 in NWChem with an input such as:

MP2
VECTORS SWAP 1 8 2 7
FREEZE 6
END
TASK MP2 ENERGY

In this case, only 2 orbitals are correlated (since there are 8 in this calculation) but instead of them being the highest energy orbitals (7 & 8) I have swapped 1 with 8 and 2 with 7, so that in fact the two core orbitals are correlated instead.

Obviously, this is a weird example, but my question is if this kind of functionality is available for the TCE, because apparently the TCE does not recognize the VECTORS directive.

Is there a workaround?

Forum Vet
This would require a two step process. Within the SCF you can reorder the orbitals by using "vectors swap" and zero iterations. Now you have the orbitals in the order you want and you can do the normal freeze.

Alternatively, in the contrib/mov2asc directory we have a piece of code that allows you to write the movecs to an ASCII file. You can manually swap the ones you want, and then convert back to an movecs file with asc2mov.

Hope this helps.

Bert



Quote:Markaw Mar 20th 8:24 pm
I'd like to ask if anyone knows how to freeze arbitrary orbitals when running calculations (e.g. CCSDT) with the TCE?
I have successfully done this with (e.g.) MP2 in NWChem with an input such as:

MP2
VECTORS SWAP 1 8 2 7
FREEZE 6
END
TASK MP2 ENERGY

In this case, only 2 orbitals are correlated (since there are 8 in this calculation) but instead of them being the highest energy orbitals (7 & 8) I have swapped 1 with 8 and 2 with 7, so that in fact the two core orbitals are correlated instead.

Obviously, this is a weird example, but my question is if this kind of functionality is available for the TCE, because apparently the TCE does not recognize the VECTORS directive.

Is there a workaround?


Forum >> NWChem's corner >> Running NWChem