localhost info

[Klaus Thoeni]

When running a simulation, which has already some pre-history (e.g. load a simulation which has done already 100000 iterations and run some more simulations), in batch mode some information in localhost info are wrong. I guess the average iterations per seconds include the iterations from the loaded state. Therefore, the value is far to high and the estimate for the finishing time/date is as well wrong. When running without batch mode the controller is showing the correct value.

[Anton Gladky (gladky-anton)]

I hope r3025 will fix the bug.

[Klaus Thoeni (klaus.thoeni)]

Hi Anton,

well done! The information shown by the localhost are fine now. The only thing
which is still strange is that the estimated day/time when finished is wrong
even if the iter/sec is more or less constant. It gives dates like 20 March
whereas the simulation is finished in 5 minutes. Towards the end of the
simulation (90%) it's getting more realistic. The same problem exists for
simulations starting from the beginning.

Not sure how the time/date is estimated. But since value iter/sec is correct
it shouldn't behave so strange.

Thanks,
Klaus

On Tue, 14 Feb 2012 08:07:57 AM Anton Gladky wrote:
> I hope r3025 will fix the bug.

[Anton Gladky (gladky-anton)]

Hi Klaus,

it should be ok in r3026. There was a typo in a formulation.

Thanks for pointing that out.

Anton

2012/2/14 Klaus Thoeni <email address hidden>:
> Hi Anton,
>
> well done! The information shown by the localhost are fine now. The only thing
> which is still strange is that the estimated day/time when finished is wrong
> even if the iter/sec is more or less constant. It gives dates like 20 March
> whereas the simulation is finished in 5 minutes. Towards the end of the
> simulation (90%) it's getting more realistic. The same problem exists for
> simulations starting from the beginning.
>
> Not sure how the time/date is estimated. But since value iter/sec is correct
> it shouldn't behave so strange.
>
> Thanks,
> Klaus
>
> On Tue, 14 Feb 2012 08:07:57 AM Anton Gladky wrote:
>> I hope r3025 will fix the bug.
>
> --
> You received this bug notification because you are subscribed to Yade.
> https://bugs.launchpad.net/bugs/931263
>
> Title:
>  localhost info
>
> Status in Yet Another Dynamic Engine:
>  Fix Released
>
> Bug description:
>  When running a simulation, which has already some pre-history (e.g.
>  load a simulation which has done already 100000 iterations and run
>  some more simulations), in batch mode some information in localhost
>  info are wrong. I guess the average iterations per seconds include the
>  iterations from the loaded state. Therefore, the value is far to high
>  and the estimate for the finishing time/date is as well wrong. When
>  running without batch mode the controller is showing the correct
>  value.
>
> To manage notifications about this bug go to:
> https://bugs.launchpad.net/yade/+bug/931263/+subscriptions

[Klaus Thoeni (klaus.thoeni)]

Yes, thanks Anton!

On Tue, 14 Feb 2012 05:42:19 PM Anton Gladky wrote:
> Hi Klaus,
>
> it should be ok in r3026. There was a typo in a formulation.
>
> Thanks for pointing that out.
>
> Anton
>
> 2012/2/14 Klaus Thoeni <email address hidden>:
> > Hi Anton,
> >
> > well done! The information shown by the localhost are fine now. The only
> > thing which is still strange is that the estimated day/time when
> > finished is wrong even if the iter/sec is more or less constant. It
> > gives dates like 20 March whereas the simulation is finished in 5
> > minutes. Towards the end of the simulation (90%) it's getting more
> > realistic. The same problem exists for simulations starting from the
> > beginning.
> >
> > Not sure how the time/date is estimated. But since value iter/sec is
> > correct it shouldn't behave so strange.
> >
> > Thanks,
> > Klaus
> >
> > On Tue, 14 Feb 2012 08:07:57 AM Anton Gladky wrote:
> >> I hope r3025 will fix the bug.
> >
> > --
> > You received this bug notification because you are subscribed to Yade.
> > https://bugs.launchpad.net/bugs/931263
> >
> > Title:
> > localhost info
> >
> > Status in Yet Another Dynamic Engine:
> > Fix Released
> >
> > Bug description:
> > When running a simulation, which has already some pre-history (e.g.
> > load a simulation which has done already 100000 iterations and run
> > some more simulations), in batch mode some information in localhost
> > info are wrong. I guess the average iterations per seconds include the
> > iterations from the loaded state. Therefore, the value is far to high
> > and the estimate for the finishing time/date is as well wrong. When
> > running without batch mode the controller is showing the correct
> > value.
> >
> > To manage notifications about this bug go to:
> > https://bugs.launchpad.net/yade/+bug/931263/+subscriptions