0

I'm trying to animate the activity of neurons in a large scale spiking neural network. I'm trying to make it as efficient as possible because it has to scale well and run smoothly for atleast 30,000 objects and preferably up to millions, from the Python API.

My efforts up to now have focussed on creating the neurons efficiently and I've ended up with:

  • 1 icosphere mesh
  • 1 mesh linked to 1 Emission material
  • n objects with their own position linked to the mesh

For the animation I've found out about relevant 2 techniques:

  • Keyframes
  • Nodes

But I haven't been able to find a way where I can do it for the objects individually since they share their mesh and material. It seems like for both of these I would have to create at least n materials or nodes so that I can animate the intensity of each object independently.

So the big question: Are there better ways to animate a property of a large group of objects, ideally where I can use 1 shared resource, passing it a vector with a value for each object?

Edit: Is a frame_change handler as proposed here efficient? Is swapping out the materials of my n objects computationally efficient?

0 Answers0