The ABINIT Users Mailing List ( CLOSED )

["Anglade Pierre-Matthieu" <anglade@gmail.com>]

Hi,
following you hints, I've had a look in csigme and there were a lot of
memory which was never deallocated. I'm currently testing the fix. Yet
I'm still not able to reproduce you computation and the bug you've
encountered.  I hope that you will be able to help me:
- first run with input epsilon1.in runs perfect
-second run with epsilon2.in fails on  reading the _KSS file
I get the following output:
"""
testing Kohn-Sham structure file: to_DS1_KSS
At line 188 of file ../../../src/01managempi/hdr_io.F90 (Unit 23 "to_DS1_KSS")
"""
I suppose that I'm doing something wrong ? (I have copyed the _KSS
file of the first run into to_DS1_KSS )

PMA

On 7/19/06, Deyu Lu <dylu@ucdavis.edu> wrote:
> Anglade and Fabien:
>         Setting stack memory to unlimited does help this issue. I
> managed to finish one test with this setting. After some debugging,
> I found that static variables in routine csigme.F90 consumed a lot of
> memory in my case. Under a default 10 M stack memory, the job always
> crashed. Tough I haven't figured out why my previous test failed even
> with "ulimit -s unlimited", I believe these jobs can be finished on
> machines with enough memory.
>
>        Best
>        Deyu
>
> On Tue, 2006-07-18 at 07:23 +0200, Anglade Pierre-Matthieu wrote:
> > On 7/18/06, Deyu Lu <dylu@ucdavis.edu> wrote:
> > > Anglade:
> > >         I haven't done that yet. I plan to submit the job later to the
> > > teragrid supercomputer center and see how it works.
> > >
> > >         Thanks
> > >         Deyu
> >
> > If abinit is failing this is going to spoil you some computing credit.
> > Moreover once you have your data file computed abinit is failing
> > pretty fast on the second input doesn't it ? May be you don't need a
> > supercomputer to get there ? At least I would advise you to try to run
> > a debugger on your case to know what is going wrong before hand.
> >
> > regards
> >
> > PMA
> >
> >
> > >


--
Pierre-Matthieu Anglade