S5-2.2 Basic Computer Organization & Control

Standard computer-architecture theory and the published 8086 architecture — written September 2026

What this is and why it exists

A calculator does one fixed job. A computer does any job you describe. The difference is one idea, and this topic is built around it.

Instructions live in the same memory as data. That is the stored program idea. Change the contents of memory and you change what the machine does, without changing any wiring.

The vocabulary

  • Instruction code — the bit pattern that names an operation and where its operands are.
  • Stored program — instructions held in ordinary memory alongside data.
  • Register — a small, very fast store inside the processor.
  • Program counter — the register holding the address of the next instruction.
  • Instruction cycle — fetch, decode, execute, and repeat.
  • Interrupt — a signal that makes the processor pause its program and run a service routine.
  • Hardwired control — control built from fixed logic gates.
  • Microprogrammed control — control implemented as a small program held in its own memory.
  • Control store — the memory that holds those microinstructions.
  • Address sequencing — working out where the next microinstruction lives.

The mental model

A machine has a small set of registers with fixed jobs. One holds the address of the next instruction. One holds the instruction being worked on. One or more hold operands and results. The instruction set is the list of things the machine can be asked to do with them. Both lists are short and concrete, and the rest of the topic assumes them.

The instruction cycle is the heartbeat. Fetch the instruction whose address the program counter holds. Advance the program counter. Decode the instruction to find out what it wants. Execute it, which may mean fetching operands from memory. Then repeat. Follow one real instruction all the way round once, on paper, and the abstractions become concrete.

Interrupts bend that loop. Something outside the program needs attention. At the end of the current instruction the processor saves where it was. It then jumps to a service routine and returns afterwards. Follow the same cycle with an interrupt inserted. You will see why the saved state includes the flags as well as the address.

Timing and control is the part that is dry alone and necessary for what follows. Something must produce the sequence of small actions each instruction needs, in the right order, at the right moments. There are two ways to build it, and the choice between them is the classic trade of this subject.

Hardwired control builds that sequence out of fixed logic. A state machine drives the control signals directly. It is fast, because the signals come straight from gates. It is awkward to change, because changing it means changing the hardware.

Microprogrammed control does the same job in software. Each machine instruction is carried out by a short program of microinstructions, held in a control store inside the processor. Each microinstruction says which control signals to assert this cycle. It is slower, because every step costs a control store read. It is cheap to change, because a change is a new microprogram rather than new wiring.

That is the same speed against flexibility axis that runs through the whole subject. Notice it here and you will recognise it in caches, in instruction sets, and in input and output.

Two details finish the microprogrammed design. Address sequencing decides where the next microinstruction comes from. It may be the next address in order, an address taken from the instruction being carried out, or an address chosen by a condition. Microinstruction format decides how wide each microinstruction is. A wide format names every control signal separately and runs fast. A narrow format encodes signals into fields and needs decoding, which saves control store and costs time. That is the same trade one level down.

What you should now be able to explain or do

State the stored program idea and say why it makes a machine general purpose. Trace one instruction round the fetch, decode and execute cycle. Show where an interrupt enters that cycle and what has to be saved. Compare hardwired and microprogrammed control on speed and on changeability. Explain address sequencing and the trade between wide and narrow microinstruction formats.

Check yourself

Its instructions sit in ordinary memory beside data. Changing memory changes the machine's behaviour without changing any wiring.

The return address and the processor flags. Without the flags, a conditional test made before the interrupt could give a different answer after it.

Its control signals come straight from logic gates. A microprogrammed unit must read a microinstruction from its control store first.

Where the next microinstruction comes from: the next address in order, an address derived from the current machine instruction, or one chosen by a condition.

A smaller control store is gained. The encoded fields must now be decoded, which costs time, so the machine runs more slowly.

Go deeper

Back to Basic Computer Organization & Control: work through the checklist