Microarchitecture Definition

(microarchitecture, uarch or arch). In computing, a microarchitecture is a description of a computer’s electrical circuitry, central processing unit, or digital signal processor that is sufficient to fully describe the operation of the hardware.

In academia, the term «computer organization» is often used, while in the computing industry, the term microarchitecture is often used. Both words are synonyms.

The microarchitecture and the instruction set architecture make up the computer architecture.

Microarchitecture vs. instruction set architecture

The microarchitecture must be distinguished from the instruction set architecture. The latter is an abstract image of a computer system as it would be seen by a machine language programmer, and includes the instruction set, memory address modes, processor registers, and address and data formats.

Microarchitecture, on the other hand, is lower level, more concrete. It shows the constituent parts of the system and how they interconnect and interoperate to implement the architecture specification.

Different machines may have the same instruction set architecture, and thus be able to run the same programs, yet they may have different microarchitectures.

These different microarchitectures (along with advances in semiconductor manufacturing technologies) are what allow new generations of processors to achieve better levels of performance compared to previous generations. In theory, a single microarchitecture (especially if it includes microcode) could be used to implement two different instruction sets, by programming two different control stores (the control store stores the CPU microprogram).

Representation of a microarchitecture

The microarchitecture of a machine is usually represented using a block diagram that describes the interconnections between registers, buses, and functional blocks of the machine. This description includes the number of execution units, the type of the execution units (such as floating point, integer, SIMD, etc.), the nature of the pipelining, the cache layout, and peripheral support.

The physical schematic of the circuit, the hardware constructions, and other physical details are called the implementation of that microarchitecture. Two machines can have the same microarchitecture, and thus the same block diagram, but different hardware implementations.