Ok, thank you. I forgot to mention I reviewed your previous message and 
also the code and found no bug. :)

This question was not concerned with overhead but more with "what do I 
need to do in my contact laws now?".

I'm still wondering how dry friction + capillary law will interact now, 
since ElasticContactLaw will delete interactions that the capillary law 
still needs (they are two different engines working consecutively on 
each interaction, one is not distant, the other is).

It looks like it works in current svn though.

Bruno


Václav Šmilauer a écrit :
>> Just one question : does it means all interaction laws (even not 
>> distant, like dry friction) have to request deletion now?
>>     
>
> [copy of previous  reply via private mail]
>
> Yes, they do have to request deletion; in itself, it is not 
> related to distant/non-distant, since in parallelized interaction 
> processing, removing interactions from container from within the loop is
> not safe and might crash (it does actually if you tried (I did)). I did 
> no measurements, but I am convinced that the overhead is close to zero 
> (a std::list of pair of body ids, and breaking an interaction happens 
> relatively "rarely", i.e. by far not at every interation).
>
> Vaclav
>
>
>
> _______________________________________________
> Mailing list: https://launchpad.net/~yade-dev
> Post to     : [email protected]
> Unsubscribe : https://launchpad.net/~yade-dev
> More help   : https://help.launchpad.net/ListHelp
>
>   


-- 
 
_______________
Chareyre Bruno
Maitre de conference

Grenoble INP
Laboratoire 3SR - bureau E145
BP 53 - 38041, Grenoble cedex 9 - France
Tél : 33 4 56 52 86 21
Fax : 33 4 76 82 70 43
________________


_______________________________________________
Mailing list: https://launchpad.net/~yade-dev
Post to     : [email protected]
Unsubscribe : https://launchpad.net/~yade-dev
More help   : https://help.launchpad.net/ListHelp
_______________________________________________
yade-dev mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/yade-dev

Reply via email to