A demo kitchen for the fetch · decode · execute cycle
Pick a dish. Watch every instruction get pulled from the pantry, read by the chef, and cooked to completion — one order ticket at a time.
Hover or focus any station above to see its computer-science name, kitchen name, and real purpose.
Why the kitchen actually works as a model
Von Neumann architecture is the basic blueprint almost every computer still follows today. Its core idea is simple but powerful: a single memory holds both the program's instructions and the data those instructions work on, and one processor walks through that memory, step by step, fetching an instruction, figuring out what it means, and carrying it out — before moving to the next one.
That's exactly what the Warehouse in this simulation is doing. Its shelves hold two different kinds of things side by side: numbered recipe steps like "Addr 101 — Boil Carrot," and the raw ingredients those steps act on. Nothing on the shelf marks one as "code" and the other as "data" — the chef decides which is which only by when he reads it. That's the stored-program concept in miniature: instructions and data sharing one memory, one address space.
The rest of the cycle follows the same one-track logic. The Instruction Counter holds exactly one address. The Notepad holds exactly one address. The Fridge holds exactly one instruction. The chef can only be doing one thing at a time — so the kitchen, like a real von Neumann CPU, works through the recipe strictly in order: fetch, decode, execute, repeat, one instruction per lap.
That single-file design is also where the famous von Neumann bottleneck comes from. Because there's only one servant making one trip at a time between the Warehouse and the Fridge, the chef can never fetch the next instruction until the current one is fully handled — the trip itself becomes the limiting factor, no matter how fast the stove can cook. Real CPUs hit this same wall: no matter how fast the ALU is, it's only ever as fast as memory can feed it new instructions and data.