[molpro-user] CI vectors, single vs. multiple state calcs

[Seth Olsen]

Hi Terry hope you're well. I don't have a straight answer for you but I do know that when issues like this come up in the forums the answer almost always contains the term 'Pspace'.  Use one of the gprint options to print out the results of the initial pspace diagonalization. As I understand it, this result is the initial guess to the subsequent iterative CI algorithm. It may be that the Pspace is being chosen differently. This doesn't explain much: if the problem is very sensitive to the initial guess then the answer is going to be the same as always: use another active space. However, if you ensure the initial problem is the same in both cases then this may solve the immediate problem. -Seth


Sent from my iPad

On 19/03/2011, at 0:01, "Terry Frankcombe" <tjf at rsc.anu.edu.au> wrote:

> Hi folks
> 
> I'm doing what one might describe as CAS-CI in a (14,9) active space of
> a large system, using Multi.  The orbitals come from a previous CASSCF
> calculation.  I don't understand the results I get for doing a 2 state
> calculation.
> 
> The two commands (in two different jobs, all other aspects being
> identical) are:
> 
> {CASSCF; occ 67; closed 58; WF 130 1 2; dont orbital}
> I{CASSCF; occ 67; closed 58; WF 130 1 2; state 2; dont orbital}
> 
> The jobs run fine.  The ground state energy is identical in each case
> (and identical to the energy from the CASSCF calculation that the
> orbitals come from, as expected).  However, for the CI vectors I get:
> 
> CI vector
> =========
> 
> 222222aa0      0.9713361
> 222220aa2     -0.2062039
> 
> for the single state calc and
> 
> CI vector
> =========
> 
> 222222aa0      0.8989227   0.2567884
> 22222a2a0     -0.2150793   0.6329933
> 2222220aa     -0.2237106   0.4707357
> 22222a0a2      0.0532983  -0.2444110
> 2222202aa      0.0534625  -0.2372961
> 222222a0a      0.0722705  -0.2227077
> 22222abaa     -0.1444124  -0.1890428
> 222220aa2     -0.1682664  -0.0702741
> 22222aa20      0.0920977  -0.1629586
> 22222baaa      0.0944303   0.1467891
> 22222aaab      0.0683397   0.1165439
> 22222aa02     -0.0352542   0.0965585
> 22222a02a      0.0491501   0.0911501
> 222220a2a     -0.0008757   0.0870140
> 22222aaba     -0.0183576  -0.0742902
> 22222a20a     -0.0611334  -0.0626342
> 2222a22a0      0.0012538   0.0506651
> 
> for the 2 state calc.
> 
> Why are the two CI vectors for the ground state so different?
> 
> As I understand things, the CI vector is essentially the eigenvector of
> a matrix which should be the same in both cases.  The energy calculated
> for the second state is well separated.  Can anyone help me understand
> what is going on here?
> 
> Regards
> Terry
> 
> 
> -- 
> Dr. Terry Frankcombe
> Research School of Chemistry, Australian National University
> Ph: (+61) 0417 163 509    Skype: terry.frankcombe
> 
> _______________________________________________
> Molpro-user mailing list
> Molpro-user at molpro.net
> http://www.molpro.net/mailman/listinfo/molpro-user