Really impressive - browser-based hardware emulation is a brutal WebAssembly performance problem. Curious what you're doing for the rendering layer: are the circuit visualizations on canvas/WebGL, or DOM-based? I've been hitting similar render-loop problems with a deck.gl mapping project (realtime vessel positions, thousands of points updating simultaneously) and the pattern I found that helped most was batching state updates on a fixed 2s interval before diffing the render layer - prevents the WebGL context from recomposing faster than the data actually changes meaningfully. Does Velxio have a similar throttle, or does it redraw on every emulation tick?