4 ms·

Can there be a 3D equivalent?

• VRay 6y ago
  should be possible.. I got something similar working in 3d a few years ago, and I'm just a mild-mannered code monkey who almost flunked out of the relevant courses on computational differential equations in college. Mine was sloppy cpu-based C# code and not this butter-smooth webGL, but mathematically it's "relatively straightforward" so long as you aren't doing anything too tricky with the boundaries
  • anoncow 6y ago
    A simulation of 3D space that I came across is what piqued my interest. https://youtu.be/xdZNwlzaUxI https://youtu.be/xdZNwlzaUxI
• slavik81 6y ago
  Mathematically, the extension to 3D is trivial. Computationally, the difference is enormous. You're moving from O(n²) cells required to fill a space to O(n³). Rendering also becomes significantly more complex, as you wouldn't be able to see anything interesting if it remained fully opaque.
• doctorhandshake 6y ago
  Vincent Houzé is an artist who works with 3D fluid sims quite a bit. These run in real-time and are interactive. He’s a wizard. https://www.instagram.com/p/B8Z2ZUYndxL/?igshid=15jc8xisc1714 https://www.instagram.com/p/B8Z2ZUYndxL/?igshid=15jc8xisc171...
• gdubs 6y ago
  Yes. Typically these types of GPU based fluid solvers use texture maps to store velocity and pressure information. (See NVIDIA’s GPU Gems paper on fluids.) Most GPUs support 3D texture maps, and changing something like this to support 3D should be fairly trivial. From what I remember, the calculations for each dimension are identical.