Lets talk about Mantaflow

I really hope we will get the solver nodes and mantaflow will be deprecated once for good. What a debacle it is :slight_smile:

2 Likes

He can always come back through GSOC for that gpu acceleration, that way itā€™s a win win situation.

3 Likes

ā€œWe arenā€™t going to work on the most broken aspects of blenderā€ :sweat_smile:

Paradoxically I think this is good. Almost all physics in Blender right now are more or less a joke.

Having unified underlying framework for all future solvers is a step in right direction IMHO. It can provide consistent interface with standardized datablocks (meshes, volumes, particles, curves, etc). Also this framework can potentially enable solver interactions. Something thatā€™s not always easy or possible with stand-alone physics systems.

2 Likes

Well yeah, its mostly the timeline that is the problem. Real time mode, Blender 101, particles, and physics rework has been on the to do list since Blenders dev fund was sub 10,000 dollars per month (6-7 years?)

1 Like

Oh, thatā€™s quite a waiting game. I jumped here with 2.8, so canā€™t share that perspective. But I had my battles with Mantaflow and cloth sims.

On the bright side weā€™ve got pointclouds recently. So hopefully this is part of a bigger plan.

Rendering points is pretty magical so that is nice :smiley:

Now imagine if you can combine them with granular (friction), wind and gravity solvers.

check this out ! Taichi_elements (MPM solver for Blender)

4 Likes