NEVPT2 calculation crashes with problem in fmove

[Jhonathan Souza]

Dear all,

I am trying to perform a energy calculation by means of NEVPT2 method, but an error crashed this calculation just begining. What do the final messages (in red) mean? There is some way to fix it?

. . . .

 Nuclear energy:                    2172.53958195
 Core energy:                      -2049.04399732
 Zeroth-order valence energy:       -110.32389505
 Zeroth-order total energy:           13.17168958
 First-order energy:               -1156.61251804
 ? Error
 ? a and b overlap, use smove
 ? The problem occurs in fmove
                                                                                          

Thanks for your consideration,

Jhonathan R Souza

[qia...@theochem.uni-stuttgart.de]

This looks like a bug in Molpro. Please provide the input file so that we can inspect.

[Hans-Joachim Werner]

Dear Jhonathan,
I cannot reproduce the problem with the current Molpro version and probably it has been fixed since version 2018.0.

I strongly suggest to install the current Molpro version 2021.1 and to use df-multi. This reduces the time for the multi calculation from your your 12 h to about 2 minutes (using 16 cores)! Without density fitting, you should use multi,so-sci for such large calculations; this is much slower than the df calculation, but still much faster than the second-order casscf (with df-multi, so-sci is the default).

In our current development version nevpt2 also works with density fitting, and again this is much faster than the conventional calculation. This will be made available as soon as possible. You can also use pno-caspt2 for fast calculations. df-rs2c also works but is still rather slow. Fully contracted canonical caspt2 can be done using

df-caspt2,contract,fclos=0

This is reasonably fast, but requires quite a lot of memory (memory,2000,m).

If you still encounter the error with molpro 2021, please let me know.

Best regards
Joachim Werner


---
Prof. Dr. Hans-Joachim Werner
Institut für Theoretische Chemie
Universität Stuttgart
Pfaffenwaldring 55
70569 Stuttgart, Germany
e-mail: wer...@theochem.uni-stuttgart.de




> Am 19.05.2021 um 13:53 schrieb Jhonathan Souza <jhonath...@gmail.com>:
>
> Hi Hans-Joachim,
>
> I am using MOLPRO version 2018.0. I am sending the input and output files as attachment in this email.
> Thanks for the attention and quick response!
>
> Best regards,
>
> Jhonathan
>
> Em qua., 19 de mai. de 2021 às 04:00, Hans-Joachim Werner <wer...@theochem.uni-stuttgart.de> escreveu:
> Dear Jhonathan,
> which Molpro version are you using? In order to check the problem, we would beed an input that shows the problem.
> Best regards
> Joachim Werner
> ---
> Prof. Dr. Hans-Joachim Werner
> Institut für Theoretische Chemie
> Universität Stuttgart
> Pfaffenwaldring 55
> 70569 Stuttgart, Germany
> e-mail: wer...@theochem.uni-stuttgart.de

> > --
> > You received this message because you are subscribed to the Google Groups "molpro-user" group.
> > To unsubscribe from this group and stop receiving emails from it, send an email to molpro-user...@googlegroups.com.
> > To view this discussion on the web, visit https://groups.google.com/d/msgid/molpro-user/f21db8d1-b39b-427f-a9fd-facbdb9b308bn%40googlegroups.com.
>
> <input_nevpt2><output_nevpt2.out>

[Jhonathan Souza]

The input file

*** p-methylphenyl-pyranoflavylium

!!! Memory setup
memory,3500,m

!!! Special files setup
file,1,2_1_4j-int_6_0_0ANC_0.int,new

!!! General setup
gprint, orbitals = 0, civec, basis
gthresh, thrprint = 0

!!! Geometry External File
geomtyp = xyz
include,../geometry_S0_1.xyz

!!! Basis set
basis = 6-311G(d,p)

!!! CASSCF Calculations
{multi

!!! Active Space definition: 12-9 | HOMO=92
occ, 95
closed, 88

!!! Wave function properties
WF,nelec = 184,symmetry = 1,spin = 0,charge = 1}

!!! Orbitals printing
{put, xml, orb_0_0ANC_0.xml; nosort; keepspherical}

!!! NEVPT2
{nevpt2}

~                                                                                                                     
~