Author Topic: jobbsse with rimp2 employing ecps (TM 5.91)  (Read 10735 times)

Stefan

  • Jr. Member
  • **
  • Posts: 15
  • Karma: +0/-0
jobbsse with rimp2 employing ecps (TM 5.91)
« on: June 06, 2008, 01:52:51 pm »
Hello,

Unfortunately, I have again a problem with the jobbsse script. I want to determine the bsse-error between two molecules including iodine. I employed the def2-QZVPP basis set and the corresponding cbasis. The first 3 steps (Energy calculation dimer, part A and B with the ghostbasis, respectively) do work. Regardless, the calculation of the monomer energy without the ghostbasis does not work. It seems, that the wrong control file is created since the entry for the ecp seems strange:
Quote
ecp   =^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@q^F^@^@^@^@^@^@^@^@^@^@ \
All bsse calculations without ecps do work. TM is version 5.91.

Best Regards, Stefan


Arnim

  • Developers
  • Sr. Member
  • *
  • Posts: 218
  • Karma: +0/-0
Re: jobbsse with rimp2 employing ecps (TM 5.91)
« Reply #1 on: June 11, 2008, 12:54:24 pm »
Hi,

in TM 5.10 jobbsse the option -setup is included, which allows to
check and change the inputs for the ghost, monomer, and supermolecule
calculations in the subdirectories before actually running the job.
This helps a lot in such cases.

Cheers,
Arnim

christof.haettig

  • Global Moderator
  • Sr. Member
  • *****
  • Posts: 258
  • Karma: +0/-0
    • Hattig's Group at the RUB
Re: jobbsse with rimp2 employing ecps (TM 5.91)
« Reply #2 on: June 16, 2008, 01:02:07 pm »
If I remember correctly, there have been several problems with ECP's in jobbsse runs for TM V5.9.1.
I recommend to update to TM V5.10 for doing such calculations. In this version jobbsse definetly works with ECPs.

Christof

Stefan

  • Jr. Member
  • **
  • Posts: 15
  • Karma: +0/-0
Re: jobbsse with rimp2 employing ecps (TM 5.91)
« Reply #3 on: September 18, 2008, 12:52:25 pm »
Hello,

I have updated to TM 5.10. Unfortunately, jobbsse does not work again. I have employed the -setup option and it seems like jobbsse does not assign the ghost atoms. However, if I assign the ghost atoms in the control file in 1_ghost and 2_ghost, reduce the occupied orbitals and the frozen core orbitals, it does work.

Best regards, Stefan

Arnim

  • Developers
  • Sr. Member
  • *
  • Posts: 218
  • Karma: +0/-0
Re: jobbsse with rimp2 employing ecps (TM 5.91)
« Reply #4 on: September 18, 2008, 09:36:27 pm »
Hi,

did you define the fragments in 'define'?

In $TURBODIR/TURBOTEST/jobbsse/short are some examples, how working control and coord file should look like.

Cheers,
Arnim

Stefan

  • Jr. Member
  • **
  • Posts: 15
  • Karma: +0/-0
Re: jobbsse with rimp2 employing ecps (TM 5.91)
« Reply #5 on: November 06, 2008, 05:51:48 pm »
Yes, I have employed define to assign the fragments.

After employing jobbsse -setup, the control file in 1_ghosts for Xe2 is:
Quote
....
$coord    file=coord
$user-defined bonds    file=coord
$frag    file=coord
$atoms
xe 1-2                                                                         \
   basis =xe def2-TZVPP                                                        \
   ecp   =xe def2-ecp                                                          \
   cbas  =xe def2-TZVPP
$basis    file=basis
$ecp    file=basis
$rundimensions
....
 

For He2 it is:
Quote
...
$coord    file=coord
$user-defined bonds    file=coord
$frag    file=coord
$atoms
he 1                                                                           \
   basis =he def2-TZVPP                                                        \
   cbas  =he def2-TZVPP
he 2                                                                           \
   basis =he def2-TZVPP                                                        \
   cbas  =he def2-TZVPP                                                        \
   charge=  0.00000000
$basis    file=basis
$rundimensions
...

It can be seen, that the nuclear charge of one fragment is not zero after employing define. Furthermore, the number of electrons is the same as for the dimer. Additionally, jobbsse -setup show several errors:
Code: [Select]
/home/stefan/prog/Turbomole/bin/i786-pc-linux-gnu
convgrep will be taken out of the TURBODIR directory
 hostname is         Niando

 bsseenergy(Niando) : TURBOMOLE V5-10 14 Jan 2008 at 14:05:36
 Copyright (C) 2008 University of Karlsruhe


    2008-11-06 17:46:03.542


 ..............................................................................
               this is a HF calculation (modules dscf/grad)
 ..............................................................................
 prepare runfile: supermol/runcalcs
 prepare runfile: 1_ghosts/runcalcs
 prepare runfile: 2_ghosts/runcalcs
 prepare runfile: monomer1/runcalcs
 prepare runfile: monomer2/runcalcs
 prepare runfile: supermol/runcalcs


    ------------------------------------------------------------------------
         total  cpu-time :   0.01 seconds
         total wall-time :   0.02 seconds
    ------------------------------------------------------------------------


   ****  bsseenergy : all done  ****

   2008-11-06 17:46:03.556

 bsseenergy ended normally
forrtl: severe (24): end-of-file during read, unit 5, file stdin
Image              PC        Routine            Line        Source
define             086368B8  Unknown               Unknown  Unknown
define             08634461  Unknown               Unknown  Unknown
define             08607257  Unknown               Unknown  Unknown
define             085D6356  Unknown               Unknown  Unknown
define             085EEBC1  Unknown               Unknown  Unknown
define             080593C2  Unknown               Unknown  Unknown
define             08049E1E  MAIN__.J                  430  define.f
define             08048296  Unknown               Unknown  Unknown
define             086457C1  Unknown               Unknown  Unknown
define             08048171  Unknown               Unknown  Unknown
forrtl: severe (24): end-of-file during read, unit 5, file stdin
Image              PC        Routine            Line        Source
define             086368B8  Unknown               Unknown  Unknown
define             08634461  Unknown               Unknown  Unknown
define             08607257  Unknown               Unknown  Unknown
define             085D6356  Unknown               Unknown  Unknown
define             085EEBC1  Unknown               Unknown  Unknown
define             080593C2  Unknown               Unknown  Unknown
define             08049E1E  MAIN__.J                  430  define.f
define             08048296  Unknown               Unknown  Unknown
define             086457C1  Unknown               Unknown  Unknown
define             08048171  Unknown               Unknown  Unknown
 define ended normally
 define ended normally
 define ended normally

Initial setup of input files for ghost, monomer, and dimer calculations
done... Please check carefully all input files and restart /home/stefan/prog/Turbomole/scripts/jobbsse
with the setup option


Best regards, Stefan
« Last Edit: November 06, 2008, 05:55:15 pm by Stefan »

christof.haettig

  • Global Moderator
  • Sr. Member
  • *****
  • Posts: 258
  • Karma: +0/-0
    • Hattig's Group at the RUB
Re: jobbsse with rimp2 employing ecps (TM 5.91)
« Reply #6 on: January 11, 2009, 12:39:24 pm »
Dear Stefan,

can you send me your complete input?

Christof