[molpro-user] Too many active orbitals in CASPT2 Calculation?

[Kirk Peterson]

Brian,

in the "normal" rs2 program, the closed orbitals are not treated any  
differently than
the active ones. So you're hitting the usual 16 active orbital limit  
for a 32-bit integer implementation
of molpro.  You'll want to use the internally contracted version of  
caspt2, which you can
invoke with "rs2c".  This will respect the 30 closed orbitals that  
you have specified, leaving only
4 active in the present (4,4) active space.

-Kirk


On Apr 5, 2006, at 6:19 PM, Brian Vest wrote:

> I have been trying to run a CASPT2 calculation with a (12,14) active
> space.  For debugging purposes, I simply limited the active space to
> (4,4).  Here's the input file:
>
> ***,CrF2 Dimer
> memory,300000000
> gprint,basis,orbital=5,civector
> geomtyp=xyz
> geometry={
> 6
> CrF2 Dimer
> Cr,       -1.537948,    0.147900,    0.004020
> F,      -0.000004  ,  1.417027  ,  0.005595
> Cr,        1.537951,    0.147912,    0.004031
> F,       0.000001  , -1.074047  ,  0.051993
> F,      -3.193075  , -0.565898  , -0.039522
> F,       3.193071  , -0.565912  , -0.039535
> }
> basis=LanL2DZ
> uhf;wf,64,1,0,0;occ,32;closed,32
> option,iptyp=diis,accura=13,ipnit=2,ipstep=1,maxdis=100,maxit=500
> save,2200.2
> multi;start,2200.2;occ,34;closed,30;maxiter, 
> 100;accuracy,energy=1.0D-08
> !iprint,micro
> rs2
>
> The CASSCF converged nicely, but when the RS2 part of the program  
> began,
> I received this error:
>
> 1PROGRAM * RS2 (Multireference RS Perturbation Theory)     Authors:
> H.-J. Werner (1993), P. Celani (1998)
>
>
>  TOO MANY ACTIVE ORBITALS: 22 THIS VERSION ALLOWS 16
>
>  ERROR EXIT
>  CURRENT STACK:      CIINP  CIPRO  MAIN
>
>
>
> ********************************************************************** 
> ************************************************************
>  DATASETS  * FILE   NREC   LENGTH (MB)   RECORD NAMES
>               1      21       64.21       600      500       
> 700      960
> 900      950      970     1000     1100     1400
>                                         BASINP     VAR     GEOM
> ABASIS   SYMINP    ZMAT    AOBASIS   BASIS      S        T
>                                          1410     1420     1200      
> 1210
> 1080     1600      129     1650     1300     1700
>                                            V       ECP       
> H0       H01
> AOSYM     SMH      P2S    MOLCAS    ERIS     OPER
>                                          1380
>                                          JKOP
>
>               2       4        0.67       700     1000     2200      
> 2140
>                                          GEOM     BASIS     UHF
> MCSCF
>
>  PROGRAMS   *        TOTAL       RS2     MULTI   UHF-SCF       INT
>  CPU TIMES  *       190.43      0.00    143.55     40.81      5.97
>  REAL TIME  *       222.04 SEC
>  DISK USED  *       331.12 MB
>
> ********************************************************************** 
> ************************************************************
>
>  GA ERROR termxy
>
> I thought there might be a problem with the a RS2 calculation with
> pseudopotentials, so I reran the same calculation without the
> pseudopotentials in the LanL2DZ basis set and received the same error.
> Any ideas what's going on?
>