On Sat, Feb 9, 2013 at 12:30 AM, Sajeewa Pemasinghe <sajeewasp.gmail.com>wrote:
> Yeah it had been a problem with compiler mismatch. Now it is running
> fine.One last question. The contents of the rem.log is NOT dynamically
> updated
> right? Will the contents be written to it when thw whole calculation is
> over? coz although there is a em.log file, currently it is empty.
>
This depends on the file system (and perhaps operating system as well --
I'm not sure of everything that is responsible for file writing behavior).
Many file writes are buffered -- that is, data is written to a chunk of
memory and stored until one of two things happens:
1) The buffer fills up so the buffer empties into the file.
2) The buffer is forcibly flushed to the file. Most languages have a
"flush" command to do this (although perhaps unreliably). Closing the file
flushes the buffer as well.
The file will be closed (and the buffer) flushed at the end of the file,
but for long runs the buffer should fill up and flush at some point during
the calculation. Note, though, if your simulation is killed before it
finishes you may lose some data that was stored in the buffer and never
flushed.
HTH,
Jason
>
> Thank you
>
> Sajeewa Dewage
>
> On Fri, Feb 8, 2013 at 4:37 PM, Sajeewa Pemasinghe <sajeewasp.gmail.com
> >wrote:
>
> > Thank you very much. Now that I know where the problem lies, I will keep
> > on working on that.
> >
> >
> > On Fri, Feb 8, 2013 at 4:17 PM, Ross Walker <ross.rosswalker.co.uk>
> wrote:
> >
> >> >know why it fails now. If there is a mismatching in MPIs above MD
> >> >calculations should also have given errors right?
> >>
> >> Not necessarily - they would just run on 1 core. Have you checked the
> >> output of your regular MD runs to see how many nodes sander reports it
> is
> >> using?
> >>
> >> All the best
> >> Ross
> >>
> >> >
> >>
> >>
> >>
> >> _______________________________________________
> >> AMBER mailing list
> >> AMBER.ambermd.org
> >> http://lists.ambermd.org/mailman/listinfo/amber
> >>
> >
> >
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
--
Jason M. Swails
Quantum Theory Project,
University of Florida
Ph.D. Candidate
352-392-4032
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sat Feb 09 2013 - 09:00:03 PST
