About rheological simulation of large open granular systems

[JoeyGu3512]

Dear Newton Physics developers and community,

I am trying to build a simulation of a large open granular system with non-spherical particles.
I tried multiple CPU-only discrete element software packages and it turns out that they are.... slow....

As I just came across this nice GPU-accelerated engine, I am a bit not sure what it is capable of.

I wonder if particles and physics objects can be safely removed during a simulation?
Is there something like a deletion region that removes physics objects that enters it and records outflow events?
Or is it possible to build something like this?

Also, I wonder how well the engine handles large number of non-spherical particles?
Is it possible to have large granular simulations like the MPM examples but with non-spherical particles?

A thousand thanks for help!!!!

Best wishes,
Joey