← Back to team overview

yade-dev team mailing list archive

Re: NewtonIntegrator::exactAsphericalRot

 


I just had the idea that we could selectively call
leapfrogSphericalRotate for bodies that have
inertia[0]==inertia[1]==interia[2] and leapfrogAsphericalRotate for the
rest; the exactAsphericalRot could be removed then.

The impact should be negligible and all non-spherical bodies would be
moved "exactly" without changing any options.


Hi, it sounds correct. Though, I'm not opposed but I'm always a bit afraid of introducing additional operations when they are not strictly necessary. The DEM algorithm is quite simple in itself. In one sense, adding negligible costs here and there to something that is itself negligible will result in a "significant" impact. But as long as people include the negligible stuff one by one, nobody will never notice, we could just end with a slow code, with no idea were to look for optimizations. I'm wondering.


Bruno




Follow ups

References