yade-dev team mailing list archive

[Burak ER <burak.er@xxxxxxxxxx>]

Dear yade-dev,

I see that I have sent multiple messages as I thought that is I saved
them as drafts.

Sorry for that.

Burak ER

-------- Yönlendirilmiş İleti --------
Gönderen: Burak ER <burak.er@xxxxxxxxxx>
Alıcı: bruno.chareyre@xxxxxxxxxxx
Bilgi: yade-dev@xxxxxxxxxxxxxxxxxxx
Konu: Re: [Yade-dev] "About using adaptive integration algorithms"
Tarih: Mon, 30 Sep 2013 18:14:54 +0300

Hi Bruno,

What are the things to save within "almost everything" except body and
interaction lists and their corresponding instances? Also, your
saveTmp()and reload() suggestion works with the serialization behind and
saving to disk or memory?

I should say that interaction deletion and creation in sub steps is the
main problem, therefore, the change of the structure of the scene. At
this moment I dont think interaction forces are not conservative.

Lastly, is it enough to keep only the interactions and body states
before integration? If its what are these?

Pzt, 2013-09-30 tarihinde 16:32 +0200 saatinde, Bruno Chareyre yazdı:
> Hi,
> Not an easy question.
> I guess that you have to save almost everything, since substeps will
> change positions and contact forces.
> Scene contains everything (exception for local variables and functions
> defined in your script). It is saved for instance if you type
> O.saveTmp(). And reloaded with O.reload(). You could probably use that
> for a start. There are probably more efficient ways but it is hard to
> imagine without looking at the details of odeint.
> 
> Bruno
> 
> 
> _______________________________________________
> Mailing list: https://launchpad.net/~yade-dev
> Post to     : yade-dev@xxxxxxxxxxxxxxxxxxx
> Unsubscribe : https://launchpad.net/~yade-dev
> More help   : https://help.launchpad.net/ListHelp

-- 
Burak ER <burak.er@xxxxxxxxxx>

-- 
BTÜ Makine Mühendisliği Bölümü 
Araştırma Görevlisi