deterministic 3d rendering
the cybergraph is a world. open cyb and you walk inside it.
every particle occupies one place in space. every cyberlink traces a line of flowing light. focus glows. heat zooms. diffusion flows. the world has physics — the same physics that computes cyberank.
one graph, one world
every neuron running the tri-kernel on the same graph sees the same world. positions match. shapes match. colors match. light matches. the math converges, and the world appears.
what you see
- shape from crystal-type: spheres for entities, helices for processes, tori for patterns, scrolls for articles
- color from crystal-domain: blue for the protocol, violet for the formal sciences, green for life, amber for society
- size from crystal-size: atoms, enzymes, bridges, articles, deeps
- light from focus: high-cyberank particles burn bright; the total light stays conserved
- motion from diffusion: currents of attention flow along every cyberlink
- scale from heat: pull the camera back and crystal-domain regions merge into one glow; lean in to see the threads
real physics
the Laplacian of the graph is the geometry of the world. springs place every particle where its topology demands. heat is the zoom. diffusion is the flow. focus is the conserved light. the operators that rank knowledge are the operators that shape space.
one frame budget at every scale
backend: honeycrisp — aruminium for Metal raster, rane for ANE inference, acpu for AMX eigensolve, unimem for zero-copy buffers. cost is per pixel rather than per particle. the same frame budget holds from 10⁶ to 10¹⁵ particles.
spec
full specification: render (R-1.0) at cyb/render/specs/render.md
discover all concepts