[molpro-user] FCI: "Workspace exhausted" error

Julien TOULOUSE toulouse at lct.jussieu.fr
Wed Dec 22 19:38:58 GMT 2004


Dear all,

I came across a "Workspace exhausted" error in the FCI program while writing the file FCIDUMP:
 
----------------------------------------------------------
In Procedure: fctra2
        At Line: 969

      Statement: Unformatted WRITE
           Unit: 1
   Connected To: FCIDUMP
           Form: Formatted
         Access: Sequential
Records Read   : 0
Records Written: 1534905

End of diagnostics
-----------------------------------------------------------

Here is the input:

------------------------------------------------------------
***, He in the dimer basis
memory,80,M
basis=avqz
r=3.80
geometry={nosym;
          He1;
          He2,He1,r}
dummy,He2
ks,lda
 orbitals,3500.3
fci
  core,0
  orbit,3500.3;
  dump;
exit
---------------------------------------------------------

And the corresponding FCI output before the error:

---------------------------------------------------------
PROGRAM * FCI (Full CI)     Author: P.J. Knowles, 1984

 ***  Initialisation ***

 Molecular orbitals read from record     3500.3  Type=RHF/CANONICAL (state 1.1)

 Frozen orbitals:              0( 0 0 0 0 0 0 0 0)
 Active orbitals:             92(92 0 0 0 0 0 0 0)
 Active electrons:             2
 Spin quantum number:          0.0
 Orbital pairs:                  4278       0       0       0       0       0
    0       0
 Strings:                          92       0       0       0       0       0
    0       0
                                   92       0       0       0       0       0
    0       0
 Determinants:                   8464       0       0       0       0       0
    0       0
 Load integrals    11.6
 Transform integrals    16.2
--------------------------------------------------------------------------


Is there a way to increase the workspace? I have the develop module so I can modify and recompile the program. I tried to increase the variable "maxbfn" in "common/maxbfn" but it didn't solve the problem.

Thanks in advance,

Julien Toulouse

Laboratoire de Chimie Theorique
Universite Pierre et Marie Curie
4 place Jussieu, 75252 Paris cedex 05, France.



More information about the Molpro-user mailing list