yade-dev team mailing list archive

[Klaus Thoeni <klaus.thoeni@xxxxxxxxx>]

Hi Bruno,

it is purely for speed up! The model already has the functionality in the IP2 
functor, so creation of new interaction is omitted. However, I want to avoid 
new contact detection since I am running examples with huge meshes (200000 
particles), and this I guess is Collider related.

And regarding the groupMask, If the limitation is 32 or 64 bodies this is not 
an option in my case. Anyone who can confirm this?

I think the thematic is similar to a clump. A clump of particle is a group 
clumped together and contact detection in between the particle of this group 
is never done. How does the collider handle this? 

Something else I had in mined, the collider could check the material of two 
bodies and if it is of type WireMat then do nothing. What do you think?

Thanks,

Klaus


On Thu, 12 Jan 2012 06:32:24 AM Bruno Chareyre wrote:
> Klaus, do you ask that for speedup or is it really a functionality problem?
> In the later case, you could use the same idea as for setCohesionNow
> (https://www.yade-dem.org/doc/yade.wrapper.html?highlight=setcohesionnow#ya
> de.wrapper.Ip2_CohFrictMat_CohFrictMat_CohFrictPhys.setCohesionNow). The
> difference is that the CohFrict functor keeps assigning other
> parameters (e.g. stiffness) when the flag is false.
> In your case you could just return.
> 
> It would still waste a bit of time to create useless interaction
> geometries. Probably not a significant loss, but it would be better if
> the mask logic could handle this.
> If I'm not wrong, bitmasks can only be used to hide about 32 (or 64)
> bodies from each other currently.
> 
> 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