[Pw_forum] relaxation problem on IBM bluegene machine

[pushpa raghani]

On Tue, Jan 6, 2009 at 5:53 AM, Paolo Giannozzi <giannozz at democritos.it>wrote:

>
> On Jan 5, 2009, at 23:50 , pushpa raghani wrote:
>
> > I am doing slab relaxation (surface) calculations on IBM bluegene
> > machine.
> > For certain cell sizes, the calculations go very fine but for some
> > others, it
> > stops after doing  the first scf run and then calculating forces
> > and the new
> > atomic positions, ie, it stops while 'Writing output data file
> > pwscf.save'. The
> > status shows that the calculation is running but it doesn't do
> > anything. Could
> > somebody tell me why does it this happen and how can it be fixed?
>
> short answer: no, nobody can. You have to discover what happens exactly
> and under which circumstances: if this is blue-gene specific or not,


It is bluegene specific. If I do the same calculation on another cluster, it
works fine but it has lesser number of processors, so the calculation is
very slow.

>
> if it is
> reproducible or erratic,


It is reproducible, I tried to change the number of processors, change the
symmetries of the system (which anyway I had to do).
Some more info: I am doing calculations on FCC(100) surface with Hubbard U.
When I use c(2x3) cell (with U and/or without U), everything goes fine. When
I use c(2x4) cell without U (just spin polarized calculation), then also
there is no such problem, it writes all the data in the directory
(pwscf.save) and finishes smoothly. The problem arises when I do LDA+U for
c(2x4) cell. Then for some reason it stops (not really stops) while writing
pwscf.save!

what happens if you change options (try to keep
> data into memory instead of saving them to file), etc. etc..


This could work but how to do that? Should I comment out the line in the
program where it writes the data to pwscf.save or there is some simple
switch to do that?

It looks
> to me
> like a problem in the file system, or in its usage, rather than in
> quantum
> espresso itself.


As I said, it all works fine for other cell sizes and also without U. Then
it doesn't seem to be the problem with file system?

>
>
> > I am using 3.2.3 version.
>
> you should use more recent versions. They better support bluegene
> machines
> and have a better scalability. Not sure this will do anything to fix
> your problem,
> though


I will also try with the new version.

Many Thanks,
Pushpa
Stanford Universiy

>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20090106/387fecf3/attachment.html>