Yes, it is possible, but I think reasonable results could be achieved with less computationally expensive solutions
Honestly, with my current idea Iām interested in making a hyper realistic footprint effect which requires this type of ability. Maybe try a slanted hill?
See: https://forums.unrealengine.com/development-discussion/content-creation/1384773-sand-physics-idea
some slanted planes could be observed at some point here: https://www.youtube.com/watch?v=hpuEdXn_M0Q&t=50m35s
Just on a larger scale with smaller particles. They seem a bit big. But that is why I would love to get my hands on this XD
In those videos all particles were always simulated. For sand some optimization can be done. Itās somewhere in our future plan.
Hey lubor_, what is the current status of your fluid plugin? When do you think it might be ready for release or have you changed directions with it at all?
we will release a testable demo this summer
@lubor_ I have a quick question for you. Is this simulation numerically stable and converges in a numerical sense? In other words, is this just a deceptively real looking simulation or is it actually usable for scientific purposes?
When will this be on the marketplace?
Thatās questionable. First, it is not even known whether for a there exist a smooth and bounded solution to Navier stokes equations exists - it is one of million dollar unsolved millennium prize problems ( NavierāStokes existence and smoothness - Wikipedia ). However, under fixed time step and discretization of the space into grid or particles standard solvers are be proven to be stable, but it is not clear whether it corresponds to real-world solution.
Our solver approximates one of the solvers using the complete information about the system, and thus given infinite amount of training data and infinite capacity of the training model, it should converge to the solution, predicted by the solver. However, these assumptions can hardly be met:). So it depends on the application, whether it is good enough. For many engineering problems (rivers, floods, rain, ā¦) it might be, as there is much more uncertainty in the initial conditions or incomplete measurements. If you are interested in a very specific problem, it would work well, if it is trained only on that kind of data.
Would it be possible to simulate deformable terrain, like sand or loamy dirt using Physics Forest?
Iām attempting to create an enduro/trail riding game and I was considering implementing a heightmap based terrain with a layer of deformable terrain on top. The heightmap terrain would serve as an extreme base, think compact dirt that cannot be interacted with, but the physics particle system on top, the loamy, or muddy, or sandy dirt, can be deformed depending on parameters that simulate the material
Is this a possible/realistic usage of this technology?
I guess that kind of application will be one of the best use cases, and thus we planned to do a simple scene simulating exactly that.
We released the demo. Download at: Downloads ā Physics Forests
Requirements: Windows 64bit, CUDA-compatible graphics card (preferably GTX1080)
Wow, really impressive! Awesome job 
the demo above is now available also on (compiled on Ubuntu 16)
I am so ready to buy this.
Will it be on Unreal Market place soon? Can I get my hands on it before it hits the market? Iāll pay you up to 300$ just to beta test it and be able to use it before you release it here.
Iām not making a game so my requirments for speed optimizations are not the same as everybody else.
I am using Unreal to make real time visualisations and Iāve got 1080Ti with 11 GB memory, machine with 32 GB and 6 cores right now and I plan to get 2080 in few months and ThreadRipper 2 as well.
Can we talk about beta access? I am eager to try this ASAP!
Iāll be using it for project featuring these bad boys among others, so once I start showing it to the world youāll get **** load of exposure as well and hopefully many more sales
Iām really sold on this plugin, so please get in touch!
I only need watterfalls, if you got them working in Unreal Plugin, lets talk
I need this plugin, how do I get it? I would gladly beta test and provide feedback. Where do I send my money?!
Do you think the plugin can achieve jelly-like fluids? or even mimic an objectās shape to create soft tires for vehicles and whatnot?
About a year ago we decided to postpone the development of the plugin mainly for 2 reasons:
- we were not happy about the quality of simulations
- there were some performance issues ( Update vertex buffer with data from GPU - Rendering - Epic Developer Community Forums )
Meanwhile the simulations became much better (this might not be obvious from videos, but it would be, if we published old version of demo) and also that data transfer problem seems resolvable. So a few days ago we have resumed the development of the plugin(s). I canāt promise any date though.
current version does not support elasto-plastic fluids, but it should be possible, we just havenāt done it yet