Neuromorphic vs Clocked Processing: Energy-per-Spike Simulator

Sensor Event Stream

Event rate 2.0 Hz/node

Spike threshold 3

Clocked Array

Clock sample rate 20 Hz

Energy Meters

1.0×

Clocked / Neuromorphic energy

Neuromorphic total

0 pJ

Clocked total

0 pJ

Active fraction now

Elapsed time

0.0 s

How it works

Two 8×8 arrays of processing nodes receive the same sparse sensor-event stream — each node independently sees random events at the chosen event rate (a Poisson-like arrival process), just as a real IoT/edge sensor grid would.

Left grid — event-driven (neuromorphic). Each node integrates incoming events into a membrane potential and only fires — a spike — once it crosses the threshold, exactly like a leaky-integrate-and-fire neuron:

Firing rate ≈ (Σ inputs) / Threshold
E_neuromorphic(t) = N_spikes(t) × E_spike

Energy is spent only at the moment of a spike. A quiet sensor costs almost nothing.

Right grid — clocked (traditional DSP). Every node is sampled and processed on every tick of a fixed clock, whether or not anything happened at that node — the standard synchronous digital-logic model:

E_clocked(t) = N_nodes × f_clock × t × E_sample

Because the clocked array pays a fixed energy bill every cycle regardless of activity, raising the clock rate (for lower latency) or raising the spike threshold (for a quieter neuromorphic array) both push the energy ratio up — try pushing the clock slider high and the threshold slider high together to see an order-of-magnitude gap open up, matching why real neuromorphic edge chips report power savings of 10–100× over always-on digital signal processing for sparse, bursty sensor data.

Event rate — how often real events arrive at each sensor node (shared by both arrays).
Spike threshold — how much integrated input a neuromorphic node needs before it fires; higher = sparser, cheaper spiking.
Clock sample rate — how often the clocked array polls every node regardless of activity.

Neuromorphic — fires on spike only Clocked — samples every tick

⚠ Couldn't load the 3D engineThree.js failed to load from the CDN. Check your connection and reload.