Time between dump files

[Daniel Price]

Hi Christophe, J-F,

There is an option to automatically adjust dtmax downwards if it is going to take more than 12 hours of wall time to get an output file.

James just pushed a fix so that this also promotes small dumps to full dumps to avoid the problem you encountered. I’ve also suggested we make this optional (on by default, but able to be turned off in the input file).

Cheers,

Daniel

> On 16 Jan 2018, at 10:07 am, Christophe Pinte <christop...@univ-grenoble-alpes.fr> wrote:
>
> Hi Daniel,
>
> I received this email from JF, and I have a quick question regarding this point :
>
> "However, the time step must be getting smaller and smaller because of dust accumulating in a denser and narrower disk, so that phantom decided to divide the time between dumps by 2, then to do it again sometimes later… which means that full dumps no longer coincide with complete orbits!"
>
> is that correct ? I thought the time between dumps was always constant.
>
> Cheers,
> C.
>

[James Wurster]

Hello

The parameter dtwallmax in the .in file now has a new meaning.  In the previous versions of Phantom, if the time between dumps was greater than dtwallmax, then all subsequent dumps would become fulldumps.  

This parameter has been re-tasked such that, if the time between dumps is greater than dtwallmax, then all subsequent dumps will be full dumps *and* dtmax will be decreased such that the next dump should be produced in less than dtwallmax time.  The default value is dtwallmax = 12h, but the option can be turned off by setting dtwallmax = 0.

The hard-coded feature that automatically decreased dtmax if the dumps were longer than 12h apart has been removed.

This change was committed to the repository on 17 Jan 2017 in commit bda34ba.

Cheers,

James

On 17 January 2018 at 00:55, 'Daniel Price' via Phantom users <phantoms...@googlegroups.com> wrote:

Hi Christophe, J-F,

 There is an option to automatically adjust dtmax downwards if it is going to take more than 12 hours of wall time to get an output file.

James just pushed a fix so that this also promotes small dumps to full dumps to avoid the problem you encountered. I’ve also suggested we make this optional (on by default, but able to be turned off in the input file).

Cheers,

Daniel

> On 16 Jan 2018, at 10:07 am, Christophe Pinte <christophe.pinte@univ-grenoble-alpes.fr> wrote:
>
> Hi Daniel,
>
> I received this email from JF, and I have a quick question regarding this point :
>
> "However, the time step must be getting smaller and smaller because of dust accumulating in a denser and narrower disk, so that phantom decided to divide the time between dumps by 2, then to do it again sometimes later… which means that full dumps no longer coincide with complete orbits!"
>
> is that correct ? I thought the time between dumps was always constant.
>
> Cheers,
> C.
>

--
You received this message because you are subscribed to the Google Groups "Phantom users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to phantomsph-users+unsubscribe@googlegroups.com.
To post to this group, send an email to phantomsph-users@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/phantomsph-users/D4B8C2C4-652F-4E45-81B7-347EB898ECCE%40monash.edu.
For more options, visit https://groups.google.com/d/optout.