[molpro-user] Too many records - reservem

[Gabriel Freitas]

Dear prof. Joachim,

Thank you very much for the response.

I followed the recomendation, and the reported error isn't occuring
anymore. But some trajectories are dying with another error message, like:

 ERROR WRITING        32768 WORDS AT OFFSET  466681080. TO FILE 5
 IMPLEMENTATION=df   FILE HANDLE=  1030  IERR=******
 ? Error
 ? I/O error
 ? The problem occurs in writew

This example occured after ~ 120 timesteps. I used [ data,truncate,5 ].
Some other trajectories, using exactly the same options, didn't die.

Is there any way to avoid it?

Kind regards,

Gabriel




2013/6/30 Hans-Joachim Werner <werner at theochem.uni-stuttgart.de>

> This has nothing to do with the file size but only with the number of
> records on a file.
> You can use data,truncate to truncate or erase a file after each step of
> your calculation.
> For example
>
> data,truncate,,1
>
> erases File 2. But note that you might then loose restart information.
> At the end of each job step a list of records ist printed, and you should
> look at this in order to decide after how many records the file can be
> truncated.
> For more details please see manual.
>
> Best regards
> Joachim Werner
>
>
> Am 29.06.2013 um 16:08 schrieb Gabriel Freitas <gabrielnfreitas at gmail.com
> >:
>
> > Dear Molpro developers / users,
> >
> > I'm running single state, RHF-QCISD(T) AIMS trajectories in serial
> Molpro 2012 (non-trial version), and my jobs end prematurely with the
> following message (always on the 251th timestep of the job):
> >
> >  More than                   512  records on file                     2
> >  ? Error
> >  ? Too many records
> >  ? The problem occurs in reservem
> >
> > It happens even though I export a large enough tmp directory (export
> TMPDIR=<tmpdir>) and adds it with -d option (molpro -d <tmpdir> job). The
> drive of this directory has 83 GB free.
> >
> > How can I avoid it? Can I "flush" the memory corresponding to the
> previous steps, regarding the information I need is printed in the AIMS
> outputs?
> >
> > Best regards,
> >
> > Gabriel
> > _______________________________________________
> > Molpro-user mailing list
> > Molpro-user at molpro.net
> > http://www.molpro.net/mailman/listinfo/molpro-user
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.molpro.net/pipermail/molpro-user/attachments/20130703/1c30aeb1/attachment.html>