this looks great!

I can't send images in DMs. I got this with 150994944 dots/particles, 12 Newton method steps, and work on random Newton method seeds positions. It takes less than a second to render.

Thanks. I've got this project, "variograph", which similarly renders a lot of dots from input parameters. I just put a function to find roots of one of your posts' polynomials. I guess it's very broken so far but it runs in less than 1 second on GPU to make this. It stops at 8 newton method steps.

How long does it take to render the image?

I've got this video from 2022, experiment with inertia looking similar to some of your results: www.youtube.com/watch?v=uIYE...

Not a fluid simulation, using the physarum algorithm ("36 Points" version by mxsage, described in my article) with nice parameters

new gallery/webpage for this project:
bleuje.com/variograph-g...

Very cool, I might try things with it

I'm not sure doing complex things with different simulation parameters par agent is a good idea, from my experience. It'a also not used in 36 Points. But would be interested if it's successful on your side.
On my side I would like to get closer to 36 Points with different deposit.

happy to see you experimenting with spatial variations of parameters. Do you plan to use a more complex simulation algorithm with more parameters?

nice, I don't have good knowledge of what people usually do to implement physarum. I feel like most people don't use compute shaders and don't do atomic add. I went into the atomic add just from trying to implement the Jeff Jones paper myself.

nice, how are you doing the deposit? drawn dots, or atomic counters like me?
Also that's a large resolution you've got!

please try stuff and report results :)

my pleasure!

that's so nice to hear! thanks
by then it'll probably have more stuff but I should still try keep it concise. Thinking about pressing buttons to show more parts/details or videos, so that the article doesn't appear to be longer

thank you!