3
$\begingroup$

I'm trying to set up a gradient calculation for psi4. In my input.dat file, I have tried the following keywords:

set basis cc-pVTZ
gradient('CCSD(T)')

which is not working. Does anyone have any recommendations on what key words to use at the end of my input file to run a gradient calculation with a CBS? Any suggestions welcome.

$\endgroup$
  • $\begingroup$ What do you mean "with a complete basis set"? $\endgroup$ – LordStryker Sep 13 '19 at 20:03
  • 1
    $\begingroup$ Can you give the entire input file and the error? It could be dying because you never tell it to do an RHF calculation. I can't remember if Psi4 has a problem with that or if it will just assume that's what you want for a reference. Also, the syntax for a direct call to the gradient is here: psicode.org/psi4manual/1.0.0/opt.html You might need to type something like: G, wfn = gradient('ccsd(t)', return_wfn=True) wfn.gradient().print_out() np.array(G) This just mimics the manual example in which you take back the wavefunction object and print the associated gradient. $\endgroup$ – jheindel Sep 13 '19 at 21:14
  • 2
    $\begingroup$ A triple zeta basis is far from complete, especially for coupled cluster methods. The entire input file would be necessary to reproduce your specific issue, and recommend further steps. $\endgroup$ – Martin - マーチン Sep 13 '19 at 23:54