Hi, On Mon, Aug 21, 2017 at 10:02:07PM +0200, Georg wrote: > Package: psi4 > Version: 1:1.0-1+b1 > Severity: normal > > Dear Maintainer, > * What led up to the situation? > Running a KS/CC-PVQZ H2O computation with geometry optimisation > * What exactly did you do (or not do) that was effective (or > ineffective)? > Applying the cc-pvqz basis set. Smaller sets (e. g. cc-pvtz or cc-pvdz) were > ok.
Thanks for the report. I can reproduce it here with the example /usr/share/doc/psi4/examples/tu3-h2o-opt.dat and replacing "set basis cc-pVDZ" with "set basis cc-pVQZ". I agree that a geometry optimization of water at quadruple zeta level is reasonable and this shouldn't crash. On the other hand, I think quintuple zeta (i.e. cc-pV5Z) is not really general-purpose anymore and having to recompile things would possibly be acceptable. We should document this better, however. This will needs to be changed in libint-dev though I believe, as we changed psi4 not to use the internal copy of libint. Michael