> For optimized build I get 20% speedup: from 127 iter/sec to 152 iter/sec for > 1000 spheres. It's great! > I am thinking about not providing Vector3r& from BexContainer, but rather have functions forceAdd(id,const Vector3r&), forceGet, forceSet. That would allow us to experiment with per-body locks in BexContainer without changes to the interface. This will be necessary if constitutive laws ever run in parallelized.
Oh, BTW, InteractionVecMap works, right? I got some crashes, but I will investigate further (it was before removing persistentInteraction and the containers had different types). Its great advantage WRT parallelization is that it has operator[uint], which is usable by openMP. You said once it was even faster than InteractionVecSet, right? Best, 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
