nooperation

A no-operation, commonly called a no-op or NOP, is an instruction or sequence that does not change the observable state of a running program. It advances the program counter and consumes processor time, but it does not modify registers or memory in a way that affects subsequent computation.

In many instruction sets there is a dedicated NOP instruction; in others a sequence of instructions with no net effect serves as a NOP. For example, x86 traditionally uses the single-byte 0x90 as a NOP, while longer, multi-byte NOPs are created from specific byte sequences such as 0F 1F 00 or 66 0F 1F 44 00 00. Other architectures differ: ARM commonly uses a move or identity operation such as MOV r0, r0, MIPS uses SLL $0, $0, 0, and RISC-V uses ADDI x0, x0, 0. The important point is that the exact encoding varies by architecture.

Common uses include padding or timing control, aligning subsequent instructions for performance or branch prediction, and patching or hot-patching code without altering behavior. NOPs can also serve as placeholders for breakpoints or for adjusting an instruction cache or pipeline during low-level debugging and optimization.

In higher-level software, the term no-op is used for functions or methods that perform no action. A no-op function may serve as a default callback, a placeholder in an interface, or an identity operation in functional programming.

Although a NOP does not change program state, it is not completely free of effect: it consumes clock cycles, and in some contexts may affect power use, timing, or memory alignment. The concept is foundational in computer architecture and software engineering.