Problem with Jobex related to Concheck

[Workingmarc]

Hi, I am a newcomer in the utilization of Turbomole (I intend to use it via DFT in oder to get a rough estimation of activation energies of reaction which are not accessible experimentally) , but I have an huge problem with Jobex (and alo jobex-ri).

I want to carry out geometry optimizations of molecules, so I think that Jobex is the most appropriate choice.

However, whenever I launch it, I always get the same error:


hammer@linux-6udr:~/Desktop/TURBOMOLE/Exercises/Water1> jobex
convgrep will be taken out of the TURBODIR directory
FORTRAN STOP  dscf ended normally
FORTRAN STOP  grad ended normally
FORTRAN STOP  relax ended normally
start test_relax
FORTRAN STOP  dscf ended normally


Input/Output Error 137: Value not recognized

   In Procedure: main program
        At Line: 45

      Statement: List-Directed READ
           Unit: 1
   Connected To: tmp.conv
           Form: Formatted
         Access: Sequential
Records Read   : 3
Records Written: 0

Current I/O Buffer:

1 6 \c
    !


End of diagnostics

error in /home/hammer/Desktop/TURBOMOLE/scripts/jobex : /home/hammer/Desktop/TURBOMOLE/bin/i786-pc-linux-gnu/convcheck failed


The problem clearly stems from Concheck, a call of it alone leads to similar results:

hammer@linux-6udr:~/Desktop/TURBOMOLE/Exercises/Water1> convcheck


Output Error 137: Value not recognized

   In Procedure: main program
        At Line: 45

      Statement: List-Directed READ
           Unit: 1
   Connected To: tmp.conv
           Form: Formatted
         Access: Sequential
Records Read   : 3
Records Written: 0

Current I/O Buffer:

1 6 \c

End of diagnostics

I have no ideas at all of the origin of such a problem.


I must say that the other components of Jobex (namely grad and relax) work normally and provide me with the outputs.

I am also perfectly able to carry out Grad and relax successfully.


Now, I would have two questions for the experimented users in this forum :

1) How could I correct the error in concheck ?

2) Otherwise, is it possible to carry out geometry optimization without convcheck, i.e just with grad and relax ?


I thank you very much ahead of time for your answers !

[uwe]

Hi,

this seems to be quite an old Turbomole version.

See:

http://www.turbo-forum.com/index.php?topic=166.0

Uwe