Hi Troels,
Just so you know, this is a problem I have tried to tackle before, but
did not find a successful solution. It's a general problem with an
MPI-based set up. It may be possible to have the slave nodes return
progressive meta information back to the master to allow for this, but
I believe that in the end I decided that it was not worth the time
investment to come up with a solution which did not slow down the
calculations.
Cheers,
Edward
On 11 June 2015 at 13:20, Troels E. Linnet
<NO-REPLY.INVALID-ADDRESS@xxxxxxx> wrote:
Update of bug #23618 (project relax):
Status: None => Invalid
Assigned to: None => tlinnet
Open/Closed: Open => Closed
_______________________________________________________
Follow-up Comment #7:
This problem is maybe more related to missing printout of progress in for
multi core monte-carlo simulations.
If 500 montecarlo simulations is spread out on 10 processors, each job gets
50
MC to handle.
But the processors does not report back, before all 50 MC is completed, and
this can take some time.
Therefore, it will for the user seems like that relax is "stuck", when using
10 processors.
For a uni-processor setup, each MC simulation is handled, and reported back
when completed. This gives a better "feeling with the process".
_______________________________________________________
Reply to this item at:
<http://gna.org/bugs/?23618>
_______________________________________________
Message sent via/by Gna!
http://gna.org/
_______________________________________________
relax (http://www.nmr-relax.com)
This is the relax-devel mailing list
relax-devel@xxxxxxx
To unsubscribe from this list, get a password
reminder, or change your subscription options,
visit the list information page at
https://mail.gna.org/listinfo/relax-devel
Re: [bug #23618] queuing system for multi processors is not well designed.