I think, Bruno is right. There will unlikely be a speedup. But you can try anyway.
Anton On Tue, Jan 17, 2012 at 7:23 AM, Klaus Thoeni <klaus.tho...@gmail.com> wrote: > Hi Bruno, > > On Mon, 16 Jan 2012 11:20:57 PM Bruno Chareyre wrote: >> >> It would work but it would be ugly to include WireMat.hpp in the >> >> collider's code. >> > >> > I agree but I would need a quick solution. >> >> The the sort of quick solution that you will never commit? Then try it. >> My bet is that, unfortunately, it will not give a very big speedup. >> The big task of the collider is to sort bounds, and it has to be done >> for all your elements anyway, since they can interact with other bodies. > > Quick and dirty ;-) But I prefer you solution anyway. > >> >> I'm thinking we could have a global "selfInteractionMask", so that if >> >> mask1==mask2 && !(mask1 & selfInteractionMask), we skip. >> > >> > Please, can you give some more details? And how and where would this >> > "selfInteractionMask" be implemented? >> >> Collider.cpp:13, this is where masks are compared in bool maycollide(). >> You could test bodies masks with a self-interaction mask there, the new >> mask could be a member of the collider. >> Maybe try the first method before deciding if it's worth the additional >> test. > > I will try to implement it in the next couple of days. So you think there will > be no speed-up? Any other suggestion how I can get some speed-up? > > Thanks > > > _______________________________________________ > Mailing list: https://launchpad.net/~yade-dev > Post to : yade-dev@lists.launchpad.net > Unsubscribe : https://launchpad.net/~yade-dev > More help : https://help.launchpad.net/ListHelp _______________________________________________ Mailing list: https://launchpad.net/~yade-dev Post to : yade-dev@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-dev More help : https://help.launchpad.net/ListHelp