The Instruction Cycle

Back to Computer Architecture
Arjit Sharma
01 Min

The processor doesn’t see loops, if-statements, or functions it just sees binary instructions, one at a time. The instruction cycle is the fundamental loop the CPU follows to process those instructions.

The 3 Main Steps

  1. Fetch
    • The Program Counter (PC) holds the address of the next instruction.
    • The CPU sends that address to memory via the address bus and fetches the instruction into the Instruction Register (IR) via the data bus.
  2. Decode
    • The Control Unit decodes the binary instruction (e.g., ADD? JUMP?).
    • It figures out what needs to happen, like which registers to use or whether memory should be read or written.
  3. Execute
    • The appropriate parts of the CPU are activated:
      • The ALU may perform a calculation.
      • A value may be loaded from memory or stored.
      • The Program Counter is updated to point to the next instruction.

Note - This cycle repeats billions of times per second on a modern CPU.

Example -

Lets say CPU has to execute the following assembly instruction

ADD R1, R2, R3;
  • Fetch: Get the binary version of ADD R1, R2, R3 from memory
  • Decode: Recognize it’s an ADD operation; identify source and destination registers
  • Execute: ALU adds values of R2 and R3; result is stored in R1

Program Counter and Flow Control

The Program Counter usually increases by one instruction each cycle. But some instructions (like JUMP or CALL) modify it directly, allowing for loops, conditions, and function calls (the foundation of structured programming)

Timing and Clock Cycles

Each stage of the instruction cycle is tied to clock cycles. Simple instructions may take 1–2 cycles, while complex ones (like memory access or multiplication) may take more. Modern CPUs pipeline these stages, overlapping them to improve throughput.

Next →