[molpro-user] Molpro-user Digest, Vol 101, Issue 8

[Peterson, Kirk]

Nike,

my feeling is that both occ and closed are directives meant to be specific to a particular program, so it should also follow the program directive (rhf, multi, etc) and be enclosed in braces.

In your CCSD(T) calculations the default in Molpro is the frozen core approximation while in CFour the default is to correlate everything.  Add a core directive to your CCSD(T) line in order to reproduce CFour.  The RHF difference is probably due to slight variances in integral thresholds, default convergence thresholds, etc.

best,

-Kirk

On Dec 21, 2016, at 2:06 PM, Nike Dattani <dattani.nike at gmail.com<mailto:dattani.nike at gmail.com>> wrote:

Dear Jacky, and Kirk,
Both of your methods worked!

In summary, using MOLPRO 2012:

==========================
geometry={angstrom;Li1;Li2,Li1,3.065}
basis = cc-pVDZ
wf,charge=0,spin=2,sym=1
occ,   3,0,0,0,1,0,0,0
closed,1,0,0,0,1,0,0,0
core
rhf
==========================

Gives: RHF STATE 1.1 Energy =  -12.635662909597

If I wrap up the RHF part in brackets:

==========================
 {rhf
 wf,charge=0,spin=2,sym=1
 occ,3,0,0,0,1,0,0,0
 closed,1,0,0,0,1,0,0,0
 core
 }
==========================

I get: RHF STATE 1.1 Energy =  -14.821433808867

Even if I wrap up everything but the "occ" part:

==========================
wf,charge=0,spin=2,sym=1
closed,1,0,0,0,1,0,0,0
core

{rhf
occ,3,0,0,0,1,0,0,0
}
==========================

I get the correct energy and occupations.

It appears that the SCF program simply does not remember the "occ" cards defined before the RHF command, even though it DOES remember charge, spin, sym, and closed.

Is this enough to be considered a bug?
Or was "occ" never designed to work the same way as "closed" ?

Does my first input work in MOLPRO 2015 ?

==========================

Another puzzle is that even with Kirk's input format:

angstrom
geometry={Li1;Li2,Li1,3.065}
basis = aug-cc-pCV5Z
{rhf;occ,3,0,0,0,1,0,0,0
closed,1,0,0,0,1,0,0,0
wf,6,1,2
}

The occupations are correct, but the RHF energy is few cm-1 higher than CFOUR,
and the CCSD(T) energy is 0.1 Hartree higher than CFOUR at aug-cc-pCV5Z level:

RHF__CFOUR = -14.824,356
RHF_MOLPRO = -14.824,341

CCSD(T)__CFOUR = -14.91920357
CCSD(T)_MOLPRO = -14.82979310

Respectfully yours,
Nike Dattani








_______________________________________________
Molpro-user mailing list
Molpro-user at molpro.net<mailto:Molpro-user at molpro.net>
https://urldefense.proofpoint.com/v2/url?u=http-3A__www.molpro.net_mailman_listinfo_molpro-2Duser&d=DgICAg&c=C3yme8gMkxg_ihJNXS06ZyWk4EJm8LdrrvxQb-Je7sw&r=mqaXWsZ0ghQR-pXvPLjY_g&m=nKB6s_VY1hxErFWZX62H2BV3olJdRjU2FciTLartrI8&s=3s1c9ccFbDfO9k0dH5G2dphjtHrSXmXgyBr62yi_jE8&e=

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.molpro.net/pipermail/molpro-user/attachments/20161222/f8c06e88/attachment.html>