EN 
Українська
Polski
Español
Deutsch
日本語
Italiano
Português (BR)
Français
العربية
Nederlands
中文
한국어Memristor Crossbar Array Simulator (2D)
2D schematic memristor crossbar array: apply row voltages to a 4x4 grid of memristive synapses and watch Ohm's law and Kirchhoff's current law compute a matrix-vector product in one step, versus a sequential von Neumann loop.
This 2D companion draws the same 4×4 memristor crossbar as a flat top-down schematic: row wires carry the input voltages, column wires collect the output, and a colored square at every intersection shows each memristor's stored conductance. Pressing Compute — crossbar applies Ohm's law and Kirchhoff's current law to every cross-point at once, so the full matrix-vector product appears in a single physical step; pressing Compute — von Neumann instead walks through the same 16 multiply-accumulates one at a time, the way a conventional CPU shuttling weights from separate memory would, so the step counter and elapsed-time readout make the in-memory-computing speed-up directly visible.
2D schematic memristor crossbar array: row voltages meet a 4x4 grid of memristive synapses, Ohm's law and Kirchhoff's current law compute the matrix-vector product, and a von Neumann mode replays the same arithmetic sequentially for comparison.
2D · HTML5 Canvas 2D · 60 FPS target · runs fully client-side, no install
A 4x4 grid of memristive synapses wired at every row/column intersection. Row voltages you set are multiplied by each memristor's conductance (Ohm's law) and summed on each column wire (Kirchhoff's current law), computing a full matrix-vector product in one physical step.
Crossbar mode computes all 16 multiply-accumulates at once, the way real analog in-memory hardware does. Von Neumann mode performs the identical arithmetic sequentially, one multiply-accumulate at a time, mimicking a conventional CPU shuttling weights from separate memory — the step counter and elapsed time make the difference directly visible.