Pople basis sets in NWChem 6.6

From NWChem

Viewed 353 times, With a total of 4 Posts
Jump to: navigation, search

Just Got Here
Threads 1
Posts 3
Did something change in the basis set reference files going from NWChem 6.5 to 6.6? I was using NWChem on two different systems (one at 6.5 and the other at 6.6) and noticed differing results for the same calculation. Looking at the basis set files in src/basis/libraries, I find the 6-31gs file in 6.5 declares the basis set as CARTESIAN but in 6.6 the basis set is declared SPHERICAL.

Did I miss something here? Thanks.

Forum Vet
Threads 8
Posts 1384
I am not aware of any change in the basis set between 6.5 and 6.6.
Could you please post the input file that is causing the problem?
Thanks, Edo

Just Got Here
Threads 1
Posts 3
Edo,

My difference was actually using an auxiliary code I have written that grabs basis set information from the src/basis/libraries files, and the python script I used was originally written about 2007. I have been relying on the "cartesian" or "spherical" tags included in those files for each atomic basis, and noticed they have changed for some basis sets between 6.5 and 6.6.

The particular changes I noted were for the 6-31G* basis sets for C and S, where the unused cartesian/spherical tag changed from "cartesian" in 6.5 (denoting the actual implementation of polarization functions) to "spherical" in 6.6 (denoting who knows what). Looking at current NWChem documentation, I see the current version does not use that tag and changing it in the libraries/ files makes no difference on direct nwchem calculations, although I am curious why the text files for the basis sets were changed. Or if they were changed intentionally, why not just remove the tag so that the current text files do not incorrectly label some of the basis sets with the wrong polarization tag.


Thanks anyway, John Mintmire

Forum Vet
Threads 8
Posts 1384
John,
Thanks for clarifying.
I have just realized that we indeed changed the NWChem builtin libraries for 6-31g* from cartesian to spherical in 6.6.

Just Got Here
Threads 1
Posts 3
That seems odd to change the 6-31g* basis sets, especially considering that the bse.pnl.gov basis set exchange still reports the 6-31g* basis set as using 6-component d-functions.


Forum >> BSE: Basis Set Exchange >> General BSE Topics



Who's here now Members 0 Guests 0 Bots/Crawler 1


AWC's: 2.5.10 MediaWiki - Stand Alone Forum Extension
Forum theme style by: AWC