[molpro-user] P86 functional difference between Molpro and Gaussian09

[Susi Lehtola]

On 10/03/2014 04:22 PM, Jason Goodpaster wrote:
> Dear All,
>
> I am trying to determine what is the difference between P86 in Molpro
> and Gaussian09.  On a simple Ne atom in the 6-31g basis for B88 ONLY I
> have,
>
> Molpro:    -128.496410793009
> Gassian09: -128.496410693

FWIW, I ran this with my code that uses libxc to evaluate the functionals
   ERKALE:    -128.49641120
as well as Q-Chem
   Q-Chem:    -128.4964105792

so a very small difference, less than 1 uHa.

> Therefore B88 is the same.  But, for BP86 in both codes I get
>
> Molpro:     -128.894112927843
> Gaussian09: -128.887877179
   ERKALE:     -128.88777766
   Q-Chem:     -128.8878692238

> Anyone know why?

Maybe there's a bug in the Molpro implementation?
-- 
-----------------------------------------------------------------------
Mr. Susi Lehtola, PhD             Chemist Postdoctoral Fellow
susi.lehtola at alumni.helsinki.fi   Lawrence Berkeley National Laboratory
http://www.helsinki.fi/~jzlehtol  USA
-----------------------------------------------------------------------