Philip Rideout, a graphics engineer, writes about 3D rendering, mesh generation, SVG, procedural terrain and OpenGL.
HN commenters praise his SVG knot diagrams and OpenGL tutorials, and point to his octasphere article as a method that gives nicer meshes.
I recently got tip on a 3D version of round rectangles, called octasphere: https://prideout.net/blog/octasphere/
It's not a direct-to-screen rendering technique like the article, rather a way to procedurally construct the 3D mesh based on few parameters (W,H,D, radius, # of subdivisions). Just like rectangles in 2D, boxes in 3D are too simple and have harsh sharp edges. Adding rounded edges makes them nicer to look at. Additionally, the octasphere method gives you capsules and spheres and everything in-between.
Edit: Oh look - there's a front page article that talks about this in-depth: https://wwwtyro.net/2021/09/24/rounded-box.html
Performance for this kind of particle effect will be even better once Webgl2 is widely available. Transform feedback (or other hacks) let you manipulate the particle positions on the videocard itself. There's a great article about them here: http://prideout.net/blog/?p=67
How is it feasible to have a .svg file that has links to other .SVG files? https://prideout.net/blog/svg_knots/knottable_v1.svg. Has this capacity been around for some time?
Computing distance fields with a jump flooding approach in O(n log n) is neat, but there is a O(n) algorithm, which might be faster https://prideout.net/blog/distance_fields/
This is the book that inspired me to create a WebGL knot gallery (https://prideout.net/knotgl/) which I eventually rewrote using Filament (https://prideout.net/knotess/).
Also http://prideout.net/ has some very nice tutorials on advanced OpenGL topics (tesselation for example).