Spurious gradient warning from basis-set history?
Irikura, Karl K. Dr. (Fed)
Dear Molpro experts,
I get a warning like this when I run DFT harmonic frequencies using two basis sets in sequence:
Gradient norm at reference geometry: 0.20962D-02
?WARNING: GRADIENT NORM NOT ZERO!
It occurs for the second basis set. If I switch the order of the basis sets, it still occurs for the second basis set. It occurs also for non-Pople basis sets. It also occurs if the input program is written without loops or variables. However, the actual harmonic frequencies are close (within 0.5 cm-1) to the results from a simple job with one basis set. So wonder if the gradient value that triggers the warning is incorrect. I could ignore it, but it also triggers a failure condition in SURF/VSCF.
Thanks for any suggestions, as always!
Karl
***,HOCl harmonic vibs, DFT, basis-set memory giving spurious gradient warning?
memory,1,G;
angstrom;
geometry={
O
H 1 oh
Cl 1 ocl 2 a
}
oh=1.
ocl=1.7
a=103.
FN='BLYP'
$BASIS_SETS=[6-31g*,6-311g**]
DO J=1,#BASIS_SETS
basis=BASIS_SETS(J)
df-ks,$FN,gridgrad=1
optg
show,FN,BASIS
! harmonic vibrational frequencies
frequencies,symm=auto
ENDDO
---
basis=6-31g*
df-ks,blyp,gridgrad=1
optg
frequencies,symm=auto
text,SECOND BASIS SET
basis=6-311g**
df-ks,blyp,gridgrad=1
optg
frequencies,symm=auto
P.S. This is not the kind of calculation that I typically do, but I have been tasked with preparing a Molpro tutorial for people who use Gaussian. The tutorial will emphasize unique features of Molpro such as program control, VSCF/VCI, and multireference theories.
-------------------------------------------------------------------
Dr. Karl K. Irikura
National Institute of Standards and Technology
100 Bureau Drive, mail stop 8320
Gaithersburg, MD 20899-8320 USA
voice: 301-975-2510 fax: 301-975-3670
email: karl.i...@nist.gov
http://www.nist.gov/mml/csd/informatics_research/index.cfm
-------------------------------------------------------------------
Hans-Joachim Werner
that is indeed strange and looks like a bug. But, as you say, the frequencies are correct. If one does it with tight optg convergence they agree within 0.01 cm-1 or so if done separately or in one run. Apparently it prints/checks a wrong gradient. It also happens with HF rather than ks. I’ll investigate it.
Best regards
Joachim
P.S: Thanks for advertising Molpro! For classes, you might find our graphical user interface gmolpro useful. We are using it here for teaching all the time. The simple guided input is restricted to single reference methods, but one can also run any imported inputs and look at orbitals etc.
> 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/SA1PR09MB868816257461E70021FFD0B8EF1B9%40SA1PR09MB8688.namprd09.prod.outlook.com.
Hans-Joachim Werner
I found and fixed the problem in our master version. We will make it available as soon as possible. Only the printed gradient was wrong.
Best regards
Joachim
---
Prof. Dr. Hans-Joachim Werner
Institut für Theoretische Chemie
Universität Stuttgart
Pfaffenwaldring 55
70569 Stuttgart, Germany
e-mail: wer...@theochem.uni-stuttgart.de
> Am 06.07.2021 um 17:49 schrieb 'Irikura, Karl K. Dr. (Fed)' via molpro-user <molpr...@googlegroups.com>:
>