Microinstruction

A microinstruction is a low-level control command used in a microprogrammed computer to govern the processor’s datapath during the execution of a machine instruction. It encodes the control signals and sequencing steps that drive micro-operations such as register transfers, arithmetic logic unit operations, and memory access, serving as an intermediate layer between architectural instructions and hardware.

In a microprogrammed control unit, microinstructions are stored in a control store (typically ROM or RAM). The processor executes a macroinstruction by issuing a sequence of microinstructions that implement its behavior. Each microinstruction specifies the control signals for one microcycle and may include fields for sequencing to the next microinstruction, conditional branches, or micro-branching based on run-time conditions.

Two common organization styles are horizontal and vertical microcode. Horizontal microcode uses wide words that directly enable many control signals in parallel, enabling rich, explicit control but requiring more storage. Vertical microcode uses compact, encoded words interpreted by a microdecoder to generate the necessary control signals, saving space at the cost of additional decoding logic. Hybrid schemes combine elements of both approaches.

Microinstructions provide several advantages, including easier implementation of complex instruction sets, flexibility to fix bugs or add instructions via firmware updates, and finer-grained control of parallelism in the datapath. Disadvantages include additional storage and decoding hardware, potential performance overhead compared to hardwired control, and added complexity in design and verification.

Historically and in many modern CPUs, microcode remains a practical means to implement and maintain a processor’s instruction set, especially for complex or frequently updated instructions, for compatibility, and for applying errata fixes.