[molpro-user] Large freq calculation fails

Alan Chien alandc at umich.edu
Mon Jan 18 14:54:48 CET 2016


I'm trying to run a very large frequency calculation involving 207 atoms
(910 elec) in a 6-31G basis.

In parallel, the program stall after entering the HESSIAN program, and
eventually gets killed by the queuing system due to one of the parallel
threads dying. The last line of output is : "Number of electrons=910
Singlet     Space symmetry=1      Wavefunction type: CLOSED"

In serial, the program hangs during the initial integral portion of MOLPRO,
with the last line of output being "Memory used in sort:     112.55 MW". I
manually ended this run after about 1.5 weeks at this stage. Using the
DIRECT keyword fails at the RHF stage.

The obvious culprit may be memory space, but I've given it a large chunk of
it (8000MW) and normally MOLPRO outputs memory error msgs if that is the
case. Any recommendations for running this large calculation would be

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.molpro.net/pipermail/molpro-user/attachments/20160118/02f1f64f/attachment.html>

More information about the Molpro-user mailing list