The CPU -- a small, nondescript square of plastic (silicon, actually) -- runs the entire show inside your computer system. The central processor performs the calculations and issues the orders that coordinate all other components and processes of your machine.

The speed of a CPU -- commonly called clock speed -- is measured in megahertz (MHz). This number designates how fast a CPU can process information. One Megahertz is one million clock cycles per second. Thus, a 400 MHz processor will have twice as many clock cycles per second as a 200 MHz processor, but this doesn't necessarily mean it is twice as FAST

CPUs work in cycles. Depending on the power and internal design of the chip, a CPU can accomplish a variety of tasks within each cycle.

Many CPUs have a "math co-processor," or FPU (which stands for "floating point unit"). The FPU is the main processor's companion, a chip that's designed to handle complicated arithmetic. "Floating point" means any number with decimals. CPUs that don't have a math co-processor can also do floating point calculations, but nowhere near as quickly.

The cache: high-speed storage

As they work, CPUs store instructions and other data in temporary memory called RAM (Random Access Memory). However, modern processors run so swiftly that communicating with conventional RAM creates bottlenecks that slow down the entire system.

CPUs route around this traffic jam by using high-speed temporary memory areas called cache. Also referred to as primary cache, L1 Cache is a very small amount of memory that is installed directly onto the CPU. This provides easy and extremely fast access to the most commonly accessed data.

When first developed, the secondary cache linked to the CPU through circuitry on the motherboard. But the CPU could still process instructions faster than the data could be moved.

Pentium versus AMD: two flavors of CPU

Two CPUs dominate the desktop computer market. Most machines sold today use a variant of Intel's Pentium processor. "Pentium" is an Intel high-performance microprocessor introduced in 1993, also called P5, 586, or 80586. The name Pentium refers to the fact that it is the fifth microprocessor in the 80x86 series. It is about twice as fast as the 486.

A large number of machines use processors made by AMD, Advanced Micro Devices, Inc. AMD manufactures integrated circuits and is based in Sunnyvale , California . It is the second-largest supplier of x86 compatible processors, and a leading supplier of non-volatile flash memory. It was founded in 1969 by a group of defectors from Fairchild Semiconductor. AMD is one of Intel’s main competitors. AMD processors are rated highly and are lower priced than Intel processors.

The Apple Macintosh computer and its clones use versions of the PowerPC chip.

PowerPC is a RISC microprocessor architecture created by the 1991 Apple-IBM-Motorola alliance, known as AIM. Originally intended for workstations, PowerPC CPUs have since become popular embedded and high-performance processors as well. PowerPC was the cornerstone of AIM's PReP and Common Hardware Reference Platform initiatives in the 1990s, but the architecture found the most success in the personal computer market in Apple's Power Macintosh line from 1997 - 2005.

Technically, CPUs come in two main designs, or architectures: CISC (Complex Instruction-Set Computing) and RISC (Reduced Instruction-Set Computing).

CISC chip instructions often involve many individual execution steps and thus many clock cycles. The Intel 80x86 processors are the most common CISC devices.

RISC chips favor a smaller and simpler set of instructions that all take about the same amount of time to execute. Most types of modern microprocessors are RISCs, for instance ARM, DEC Alpha, SPARC, MIPS, and PowerPC.

The difference is one of computing philosophy. CISC chips like the Pentium sacrifice some raw speed in favor of having a complete set of built-in instructions on the chip.

Practically speaking, RISCs are faster at tasks that involve serious number crunching -- such as computer-aided design, high-resolution graphics or video processing -- while CISCs have an edge in applications such as database management. But all CPUs are so fast today that you don't have to worry about the architecture unless you're specializing in one of these applications.

Some computers use parallel processing to get the job done. This refers to a computing method that can only be performed on systems containing two or more processors operating simultaneously. Parallel processing uses several processors, all working on different aspects of the same program at the same time, in order to share the computational load. Multiprocessor computers are generally used as servers that manage computer networks and large Web sites.