[molpro-user] PES calculation with AM1 couplings - record error

[Dinu, Dennis Florian]

Hello Fabian,

do you just want to perform a calculation for H2O? From my point of view your approach using Density-Fitting and Local Methods for such a small and symmetric system is a little unfortunate. If you own a more or less up to date computer you can do this calculation with conventional CCSD(T)/VTZ with up to 3D sub-potentials within a couple of hours.

Concerning your error: I could reproduce it. The SAVE, START combination just messes things up. Somehow Molpro wants to read in record 1650 MOLCAS from FILE 1, but this record is not there. If you can renounce the SAVE,START combination use the input below.

greets
D. Dinu


-----------
***,TITLE
memory,488,m

file,1,surfdebug.int
file,2,surfdebug.wf

orient,mass
angstrom
geometry={
   O,	  0.02252435720000,    -0.00256116440000,     -0.06269948230000
   H,	  0.50132694220000,    -0.75770407450000,     -0.43062909720000
   H,	  0.50412480060000,	0.74347113890000,     -0.44527202040000
}

mass,iso
! GEOMOPT and HESSIAN
basis=cc-pvtz
{df-hf
start,atden
accu,14}
df-lccsd(t)
optg
freq

! PES calculation
label1
abinitio
basis = cc-pvtz
int
{df-hf
start,atden
accu,14}
df-lccsd(t)
goto,label3

label2
semi,am1
int
rhf

label3
{surf,start1D=label1,sym=no,ndim=2,ngrid=8
vmult,start2D=label2,Multi=1
repar
disk,save=5600.2,where=home,dump='surfdebug.pot'}

! VSCF and VCI at PES
vscf,thermo=1
vci,thermo=1

________________________________________
Von: Molpro-user [molpro-user-bounces at molpro.net]" im Auftrag von "Fabian Berger [bergerfa at hu-berlin.de]
Gesendet: Freitag, 19. Januar 2018 14:38
An: molpro-user at molpro.net
Betreff: [molpro-user] PES calculation with AM1 couplings - record error

  Hello everyone,
  I have some problems with PES calculations within the SURF program of
  molpro. I found a related topic
  (http://www.molpro.net/pipermail/molpro-user/2014-April/005974.html)
  but it does not solve my problems. I got the same record error as
  described in the link.

  I want to do a PES calculation with df-l-ccsd(t) for 1D and AM1 for 2D
  couplings. If I use a DFT method instead of the semi-empirical method
  everything works fine.
  I performed a hessian calculation and a single point df-lccsd(t)
  calculation and saved the domains in record 1891.2 and everything in
  the file surfdebug.wf.

  Thank you in advance!

  Best regards,
  Fabian Berger

  ********************************************
  The error message in .out is:

  *** Long output written to logfile /scratch/fabe/207475/molpro_vscf.log
***
   Dump file changed to: surfdebug.pot
   Any existing file with the same name will be overwritten...
   Calculating 1D potential
   ? Error
   ? Record not found
   ? The problem occurs in readm

  The error message in .log is:

  FILE 1 RECORD    1650 OFFSET=          0. NOT FOUND
   Records on file 1
   IREC   NAME  TYPE        OFFSET    LENGTH   IMPLEMENTATION   EXT
  PREV   PARENT  MPP_STATE
     1     500  VAR          4096.   112002.         df          0
  0      0      1

     ...
  *******************************************

  My input is shown below:

  ***,TITLE
  memory,488,m

  file,2,surfdebug.wf

  angstrom
  geometry={
     O,     0.02252435720000,    -0.00256116440000,     -0.06269948230000
     H,     0.50132694220000,    -0.75770407450000,     -0.43062909720000
     H,     0.50412480060000,     0.74347113890000,     -0.44527202040000
  }

  ! PES calculation
  label1
  {df-hf, basis=cc-pvtz
  start,atden
  accu,14}
  {df-lccsd(t), basis=cc-pvtz ,start=1891.2}
  goto,label3

  label2
  semi,am1
  int
  rhf
  goto,label3

  label3
  {surf,start1D=label1,sym=no,ndim=2,ngrid=8
  vmult,start2D=label2
  repar
  disk,save=5600.2,where=home,dump='surfdebug.pot'}


  ! VSCF and VCI at PES
  vscf,thermo=1
  vci,thermo=1

------
_______________________________________________
Molpro-user mailing list
Molpro-user at molpro.net
http://www.molpro.net/mailman/listinfo/molpro-user