[molpro-user] LCCSD(T) is more time and memory consuming than CCSD(T)

Leonid Shirkov leonid.shirkov at gmail.com
Fri Feb 23 14:53:01 CET 2018

Dear Tatiana and prof. Werner,

thank you for your explanations and forgive my ignorance on that
topic. My goal is big molecules and testing LCCSD(T) for such small
molecules wasn't a very good idea.

Best regards,

On Mon, Feb 19, 2018 at 8:00 AM, Hans-Joachim Werner
<werner at theochem.uni-stuttgart.de> wrote:
> Of course LCCSD is slower for small molecules such as H2O. Local approximation reduce the cost scaling with the number of correlated electrons, and only above a certain cross-over point the local methods become faster. Likely, this cross-over point is shifted to larger molecules when very large basis sets are used. For accurate calculations I recommend to use the F12 methods.
> Joachim Werner

More information about the Molpro-user mailing list