ARM Architecture: Origins, Design & Key Facts

ARM is a family of reduced-instruction-set computer (RISC) architectures that dominate mobile and embedded devices, from smartphones to microcontrollers. Its key facts are its energy-efficient design, its origins as a British company spun out of Acorn Computers in 1990, and its licensing model where ARM designs chips but does not manufacture them. This article explains how ARM works, why it became so widespread, and what makes it different from x86.

Origins of ARM

ARM began in the 1980s at Acorn Computers in Cambridge, UK. The team wanted a processor for its upcoming personal computers but found existing chips too slow or too expensive. They designed a 32-bit RISC processor, the ARM1, which first worked in 1985. In 1990, Acorn spun off the design team into a new company called Advanced RISC Machines (ARM) Ltd., partnering with Apple and VLSI Technology. The first commercial chip, ARM2, was used in Acorn’s Archimedes computers and later in Apple’s Newton PDA.

How ARM Architecture Works

ARM uses a load-store architecture: only load and store instructions access memory; all other operations work on registers. This simplifies the instruction set and enables a deeply pipelined, low-power design. Key features include:

Fixed-length 32-bit instructions (in classic ARM) or 16-bit Thumb instructions for higher code density.
Conditional execution on most instructions, reducing branch penalties.
Many registers (typically 16 general-purpose) to minimize memory traffic.
Optional SIMD (NEON) and floating-point units for multimedia.

Modern ARM cores (like Cortex-A series) are superscalar, out-of-order, and use advanced power management. The architecture also supports TrustZone for security and big.LITTLE (heterogeneous computing) to combine high-performance and high-efficiency cores.

ARM vs x86: Key Differences

The main rival is Intel/AMD’s x86, a CISC (complex instruction set) architecture. Differences:

Instruction complexity: x86 has variable-length, complex instructions; ARM has simple, fixed-length (mostly) instructions.
Power consumption: ARM is designed for low power, making it ideal for battery devices; x86 historically prioritizes raw performance.
Licensing: ARM licenses its designs; x86 is proprietary and only Intel/AMD make it (plus a few others under license).
Ecosystem: x86 runs Windows and most desktop software; ARM runs iOS, Android, and increasingly Windows on ARM.

In recent years, Apple’s M-series chips (based on ARM) have shown that ARM can also deliver high performance for laptops and desktops, blurring the line.

Common Misconceptions

One frequent myth is that ARM is a single chip. In reality, ARM is an architecture; companies like Qualcomm, Apple, Samsung, and Nvidia design their own cores based on ARM’s instruction set. Another misconception is that ARM is slower than x86. While raw clockspeed may be lower, ARM’s efficiency means it can perform well while using far less power. Also, ARM does not manufacture chips; it only sells designs and licenses. The actual fabrication is done by foundries like TSMC and Samsung.

Why ARM Dominates Mobile

ARM’s success in smartphones and tablets comes from its energy efficiency. A typical ARM core uses a fraction of the power of an x86 core, extending battery life. The licensing model also allows companies to customize cores for specific needs (e.g., Apple’s performance cores). Additionally, ARM’s small die size and low cost make it ideal for mass-market devices. Today, over 200 billion ARM-based chips have been shipped, covering everything from sensors to servers.

How to Choose Between ARM and x86

For most consumers, the choice is already made by the device. If you need a laptop or desktop that runs legacy Windows software or heavy gaming, x86 is safer. If you want a lightweight, fanless laptop with long battery life, ARM (like Apple M-series or Snapdragon-based Windows laptops) is excellent. For embedded projects (Raspberry Pi, Arduino), ARM is the standard. For servers, ARM is emerging as a power-efficient option for cloud workloads, but x86 still has broader software support.

Frequently Asked Questions

Is ARM a RISC or CISC architecture?

ARM is a RISC (Reduced Instruction Set Computer) architecture. It uses a small, highly optimized set of instructions, typically one instruction per clock cycle (though modern cores can do more). This contrasts with CISC like x86, which has complex, multi-cycle instructions.

Who owns ARM?

ARM Holdings is owned by SoftBank Group (as of 2024). The company does not manufacture chips; it designs the architecture and licenses it to other companies. The headquarters is in Cambridge, UK.

Can ARM run Windows?

Yes, Windows 11 runs on ARM processors, and Windows on ARM can emulate x86 applications (though with some performance overhead). Apple’s M-series Macs run macOS natively on ARM. Many Linux distributions also support ARM.

What is the difference between ARMv8 and ARMv9?

ARMv8 introduced 64-bit support (AArch64) and is used in most modern phones. ARMv9 adds enhanced security (e.g., Memory Tagging Extension), better AI processing, and improved performance. Both are backward compatible with 32-bit ARMv7 code (in some cases).

Why are ARM chips so power-efficient?

ARM’s RISC design reduces complexity, allowing simpler hardware and lower voltage. Additionally, ARM cores often use fewer transistors and have advanced power gating. The architecture also allows for heterogeneous computing (big.LITTLE) where low-power cores handle simple tasks.

Can I upgrade my ARM processor?

In most devices (phones, tablets), the processor is soldered and cannot be upgraded. In some laptops (e.g., Framework, some Windows on ARM), the mainboard may be replaceable, but the CPU itself is not. For embedded systems, you can replace the whole board.

Method: the standard formula for this calculation, rounded as each result states. See how the calculators work or report an error.

More cards in this drawer

The whole Numbers and Powers drawer →