[molpro-user] default fitting basis for F12

[Andy May]

Boris,

The two inputs are not equivalent. The simple input:

basis=aug-cc-pvqz

sets new default values for all basis sets, whereas:

basis={
Cl=aug-cc-pvqz
}

only sets the orbital basis set for Cl atom. The simple way to achieve
what you had before is:

basis={
default,aug-cc-pvqz
}

or the complicated way, which shows exactly what is being changed, is:

basis={
Cl=aug-cc-pvqz
}
rhf
rmp2-F12,ri_basis=vqz,df_basis_exch=vqz

Best wishes,

Andy

On 02/03/10 07:57, Boris Averkiev wrote:
> I have question about density fitting basis sets for F12 methods I calculated Cl atom.
> 
> When I specified basis one way
> geomtyp=zmat
> geometry={ Cl }
> basis=aug-cc-pvqz
> rhf
> rmp2-F12
> 
> MOLPRO used the following basis sets:
> 
>                           Density fitting integral evaluation by F. R. Manby, 2003,2007
>                                       
> Basis set CC-PVQZ/JKFIT generated.      Number of basis functions:   148 
> Basis set AUG-CC-PVQZ/MP2FIT generated. Number of basis functions:   194
> ...
> 
> AO(A)-basis ORBITAL           loaded. Number of functions:      84
> RI(R)-basis CC-PVQZ/JKFIT     loaded. Number of functions:     148
> DF-basis AUG-CC-PVQZ/MP2FIT   loaded. Number of functions:     194
> 
> When I specified the same basis set a little bit different
> geomtyp=zmat
> geometry={ Cl }
> basis={
> Cl=aug-cc-pvqz
> }
> rhf
> rmp2-F12
> 
> MOLPRO used different density fitting basis sets.
> 
> Density fitting integral evaluation by F. R. Manby, 2003,2007
> 
> Basis set JKFIT generated for CL=AUG-CC-PVQZ. Number of basis functions:   184
> Basis set JKFIT2 generated for CL=AUG-CC-PVQZ. Number of basis functions:   184
> Basis set MP2FIT generated for CL=AUG-CC-PVQZ. Number of basis functions:   194
> ...
> 
> AO(A)-basis ORBITAL           loaded. Number of functions:      84
> RI(R)-basis JKFIT2            loaded. Number of functions:     184
> DF-basis MP2FIT               loaded. Number of functions:     194
> 
> 
> Is it suppose to be like this? Does it mean that when I use F12 methods with standard basis sets like aug-cc-pvqz (and I don't need to explicitly define fitting basis) it would be better to specify basis set the second way?
> 
> Thank you very much.
> 
> 
> Boris
> 
> 
>       
> _______________________________________________
> Molpro-user mailing list
> Molpro-user at molpro.net
> http://www.molpro.net/mailman/listinfo/molpro-user