I'm trying to use NBO7 in the standalone version on an ORCA job, by generating molden file from ORCA via orca_2mkl and then using molden2aim to obtain the 47 input file for NBO. On small toy jobs (like H2O) this works, but on larger jobs (say, 30 atoms, quadruple zeta basis) I get the following error from gennbo: "MRSEP: LW=49770 is greater than 33*M=46926", which in the source code seems to be linked to the 'dsyevr' diagonalization subroutine from LAPACK. NBO7 is using Intel MKL as LAPACK driver and I have no control over the effects (this is a precompiled version from a HPC center). Has anyone come across a similar error?
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science
- Materials
- Inorganic Materials
- Metals
- Alloys
- Steel
More Maciej Smiechowski's questions See All
Similar questions and discussions