CompTIA A_ Certification All-In-One Exam Guide, Seventh Edition - Michael Meyers [73]
* * *
NOTE Clock multiplying first surfaced during the reign of the Intel 80486 CPUs. The first clock multipliers exactly doubled the clock speed, resulting in the term clock doubling. This term is used interchangeably with clock multiplying, even though modern CPUs multiply far more than just times two.
All modern CPUs are clock multipliers. So in reality, every CPU now has two clock speeds: the speed that it runs internally and the speed that it runs when talking on the address bus and the external data bus. Multipliers run from 2× up to almost 30×! Multipliers do not have to be whole numbers. You can find a CPU with a multiplier of 6.5× just as easily as you would find one with a multiplier of 7×. A late-generation Pentium would have an external speed of 66 MHz multiplied by 4.5× for an internal speed of 300 MHz. The Intel Pentium 4 3.06-GHz CPU runs at an external speed of 133 MHz with a 23× multiplier to make—yes, you’ve got it—3.06 GHz. Without the invention of multiplying, modern CPUs would be nowhere near their current blazing speeds.
The clock speed and the multiplier on Pentium CPU systems had to be manually configured via jumpers or DIP switches on the motherboard (Figure 5-32). Today’s CPUs actually report to the motherboard through a function called CPUID (CPU identifier), and the speed and multiplier are set automatically.
Figure 5-32 DIP switch on an old motherboard
For years, users pushed for faster and faster CPU clock speeds, because clock speed was considered the most important way to differentiate one CPU from another. By 2003, advancements in caching, pipelining, and many other internal aspects of the CPU made clock speed an inaccurate way to compare one CPU to another. CPU makers give their processors model numbers—nothing more than marketing names—to tell one processor from another. The Intel Core Duo T2300, for example, actually runs 1.66 GHz (166 MHz external speed with a 10× multiplier). If you want to know the speed of a particular processor, you must go to the CPU maker’s Web site or other source.
CPU Voltages
In the simplest sense, a CPU is a collection of transistors, tiny electrical switches that enable the CPU to handle the binary code that makes up programs. Transistors, like other electrical devices, require a set voltage to run properly. Give a transistor too much and you fry it, too little and it doesn’t work. For the first ten years of the personal computer, CPUs ran on 5 volts of electricity, just like every other circuit on the motherboard. To increase the complexity and capability with new generations of CPUs, microprocessor developers simply increased the number of transistors. But eventually they altered this strategy to increase the efficiency of the CPUs and keep the size down to something reasonable.
Intel and AMD discovered that by reducing the amount of voltage used, you could reduce the size of the transistors and cram more of them into the same space. Intel released the Pentium, for example, that required only 3.3 volts. AMD responded with its versions of the Pentium-class CPUs with even lower voltages.
Motherboard manufacturers had to scramble to adapt to the changing CPU landscape by creating motherboards that could handle multiple voltages of CPUs. All of the logic circuits still ran at 5 volts, so manufacturers started installing a voltage regulator module (VRM) that damped down voltages specifically for the CPUs.
Because the new and improved motherboards handled many CPU voltages, initially techs had to install a VRM specific to the CPU. As manufacturers got better at the game and built VRMs into the motherboards, techs just had to change jumpers or flip switches rather than install a VRM (Figure 5-33).
Figure 5-33 Volt regulator module
Getting