What Is a CPU?

A CPU (central processing unit) is the processor that runs a computer’s operating system and applications. Every click, calculation and network packet ends up as a stream of simple instructions, and the CPU works through them one after another at very high speed. People often just call it the processor or the chip.

In a desktop PC the CPU usually sits in a socket on the motherboard under a heatsink. In laptops, embedded boards and most industrial PCs it is soldered directly to the board, which saves space and copes better with vibration. Either way, it connects to memory, storage, graphics and I/O through the motherboard and chipset.

Today’s CPUs are rarely just a processor. One package typically holds several cores, shared cache, a memory controller, PCIe lanes, integrated graphics and, on recent Intel and AMD parts, a neural processing unit (NPU) for AI work.

Every CPU runs the same basic loop, known as the instruction cycle. It fetches the next instruction from memory, decodes it to work out what it asks for, executes it and writes the result back to a register or to memory. A clock signal keeps each stage in step, and the clock speed in gigahertz (GHz) says how many of those clock cycles happen each second.

Modern cores do not wait for one instruction to finish before starting the next. They work as a pipeline, with several instructions at different stages at once, and a superscalar core can issue several instructions in the same clock cycle. Branch prediction guesses which way the program will go next so the pipeline stays full.

Inside each core, a few main parts share the work:

A core is a complete processor in its own right, so a 6-core CPU can run six streams of work at once. Many cores also run two threads each, a technique Intel calls hyperthreading, which keeps a core busy when one thread stalls. Recent Intel processors mix large performance cores with smaller efficiency cores, as our guide to P-cores, E-cores and Intel’s hybrid architecture explains.

The instruction set matters too. Almost every current x86 and Arm CPU is 64-bit, which lets it address far more than 4 GB of memory; the difference between 32-bit and 64-bit still shows up with legacy software and drivers.

The figures are typical ranges, not limits, and they vary by generation. No single number predicts performance: a CPU with fewer, faster cores may suit one application better than one with many slower cores.

Two instruction set families cover almost every CPU you will meet. x86, from Intel and AMD, runs Windows and Linux and decades of existing industrial software. Arm designs are licensed to many chipmakers, including NXP, Qualcomm, Rockchip and NVIDIA, and appear in phones, tablets and most single board computers and edge AI modules.

Arm CPUs are usually built into a system on chip with graphics, video engines and I/O, and they tend to use less power. x86 offers the broadest software and driver support. In industrial systems the choice usually comes down to the operating system, the software you already run and the power budget.

Industrial and embedded computers often use the same core designs as office PCs, but the chips are chosen for different reasons. Clock speed matters less than how long the part stays available and how it behaves in a sealed box on a hot or cold site, which is why the processor options on our industrial PCs look different from a consumer range.

We select, fit and validate the processor, memory and cooling in every industrial PC we supply, as standard. Tell us about your software, power budget and site conditions, and we will recommend the CPU that suits them.