Just got into your videos after watching parts one and part two of this series. Your explanations at the beginning are very helpful, and the code is laid out in an easy-to-read way and seems really adaptable, which is amazing. Very impressed
Unity tells me I cant use "new" as a keyword in the FillPopulationWithRandomValues() funktion in this line: newpopulation[StartingIndex] = new NNet();. Can anyone help please?
Hi, I saw your post in reddit about an reliable input recorder (unity) and that you were about to create a tutorial. Sadly I didn´t see one on your channel. Do you consider to do one, it really would help me with my final exam :)))
How we can enable this fluid to flow? actually i am working on a project "Fluid dynamics Experiment demonstration" where I am trying to create a environment where we can simulate fluid dynamics experiments like demonstration pf "water turbulence" or "Calculation of fluid flow". for that I need a realistic fluid simulation something similar to this video. Is there someone who can help me with this and guide me?
@batchprogrammer108 Ok, I am sorry for asking you on this video but on your nerual network video, is it possible so the "AI" gets a reward if it copleates it the track and if it goes around the track faster it gets a bigger award?
@batchprogrammer108 I finished the build but is it possible so it gets a reward if it copleates it and if it goes around the track faster it gets a bigger award?
I am thinking of creating a game in the style of Royal match, essentially a match-3 game. I would like to incorporate a fluid structure in some levels to navigate through the board. Can I use this system, or do I need to use a different structure?
Hey! I think SPH might be overkill for what you're trying to do, here's another one of my GitHub repositories which is using a simpler version of fluid simulations which just uses the Navier Stokes equation, it's a lot more efficient and can be run on bigger grids github.com/AJTech2002/Realtime-Fluid-Simulation This is the paper I referenced for that repo: graphics.cs.cmu.edu/nsp/course/15-464/Fall09/papers/StamFluidforGames.pdf Hope this helps!
Absolutely! On my GitHub I actually have a branch where I turned the water into lego pieces for fun and it was much cheaper to render as well. Would love to see what you do with it, so please leave a comment with some examples if you end up using it :D
Hi! I have an URP project and i don't know how can i adapt the shaders in the project to work properly. I tried to upgrade the shaders but apparently this can't be done in custom shaders. Can someone tell me where can i learn to do this? Or if someone can provide an URP version of the shaders it would be very helpful.
Another optimization you could try is creating a 3D dispatch size and thread group and use shared memory. This way all threads would access data from shared memory rather than global memory and improve memory bandwidth. You might need to play around with the thread group size to get better performance.
You can but something like this would run much faster on a compute shader, especially since in future you'll need to take this data to the GPU so that you can render it to look like actual water. Since this is already on the GPU it will be much more performant :)
A lot of us game devs who like making super good niche things typically spend an outrageous amount of time on just one thing you know. Like I've been making a voxel engine for my game forever
Don't put yourself down man! I spent months researching and getting my head around this, I felt really dumb because I couldn't understand it for ages, it's just a niche topic but making a whole game is much harder. Good luck on your game 😁
Hey everyone hope you enjoyed this follow-up video on fluid simulations, if you have any questions or comments leave them down below. If you're interested in another part to this series let me know as well!
I've added optimization into the GitHub on this branch (github.com/AJTech2002/Smoothed-Particle-Hydrodynamics/tree/feat/optimization) incase you're interested, currently working on the video :)
Hey, thanks for this tutorial that was insane ! 11:36 nazi pro tip : if particleMass is constant, then 1f / particleMass is a constant too force * (1f / particleMass) will be faster to compute :P
