This is a generic n body simulation, I did it to investigate using svelte, typescript and threejs. All three of these were new to me before I started so this code is likely ugly etc. Considering the very few resources for svelte cubed this could also be usefull in that regard for someone stumbling across it.
I used the Barnes-Hut algorithm, a recursive n log(n) algorithm that subdivides the space and places every particle into a tree structure that keeps track of the space in which the particle is located. This allows for optimization by considering far away particles as a single entity with mass and position calculated for the space said particles is contained in. You can read more on that at wikipedia. Something potentially intersting in the version of this algorithm that I made is that the total space of a node is dynamically updated as particles are inserted into said node all nodes the particle moves through when it is placed in the tree. From what I have seen in other implimentations a fixed space is usually subdivided. The difference being that without any further work this algorithm works for sets of particles with arbitary volumes.
The project is used mostly to investigate some front end technologies with which I am not familiar. As I have said before I used Svelte as the front end framework with typescript for and threejs for the 3d rendering of the scene. Although later I switched over to using svelte-cubed to interact with three-js.
If you want to investigate the code it is under src/Gsim.svelte, that is the main component that implements the algorithm and makes the scene. App.svelte just makes the rest of the html page.
It should be as simple as cloning the repo and doing a npm i, 'npm run dev' and visiting 127.0.0.1:5000.