CompTIA A+ 220-701 and 220-702 Exam Cram: The CPU

This chapter is from the book 

The central processing unit, or CPU, is quite often referred to as the "brain" of the computer. Today's CPUs are like superbrains! A typical CPU today runs at 3GHz or higher, use two or more cores, and some can easily process 50 billion operations per second. That's a good deal more than we would have seen just 5 years ago. Some mornings I have trouble processing the thought need coffee! Of course we know that the human brain is much more sophisticated and functional than a CPU, but the CPU wins out when it comes to sheer calculating power.

You might hear the CPU referred to as a microprocessor, which technically it is. It's a much smaller version of the processors that were used 50 years ago. And although microprocessor might be a more accurate term, it has become more acceptable to refer to it as CPU, which this chapter does. However, you also see CPU manufacturers such as Intel refer to them as processors, so for all intents and purposes, the three terms mean the same thing. Keep in mind that a computer has other processors used by video cards and elsewhere, but know that the CPU is the main processor.

This chapter discusses some CPU technologies and cooling methods and talks about the models of CPUs offered by Intel and AMD. Afterward, the chapter demonstrates how to install and troubleshoot the CPU.

CPU 101

The CPU is often the most-expensive component in the computer; it's also one of, if not the, most important. The CPU's main function is to execute instructions or programs. Its speed, or clock rate, is measured in Hertz. For example, at 2.66GHz, a CPU operates at 2.66 billion cycles per second; we speak more to this concept in a moment. But although the speed of the CPU might be important, other factors should also play into your decision when choosing a CPU, including the chipset on the motherboard, CPU technology, and the brand of CPU. Chapter 2, "Motherboards," covers chipsets, but let's go ahead and talk about the various CPU technologies and brands of CPUs now.

CPU Technology

CPU technology is a key factor when considering a CPU. It all comes back to the motherboard; the CPU must be compatible with the motherboard in a number of ways. It is important to think about the speed (clock rate) of the CPU you want to use and whether that speed can be supported by the motherboard, and if the CPU fits in the motherboard's socket. Also, a decision has to be made as to whether to use a 32-bit or 64-bit CPU, and choose either a single-core or multi-core CPU; this will be based off the motherboard and the type of operating system you plan to install. Getting deeper into the technical side of the CPU, you might want to know the amount of cache included with the CPU, and the amount of power it requires.

Clock Rate

The clock rate is the frequency (or speed) of a component. It is rated in cycles per second and measured in hertz (Hz). For all practical purposes, the term clock rate is the same as the more commonly used term: clock speed.

Components are sold to consumers with a maximum clock rate, but they don't always run at that maximum number. To explain, let me use a car analogy. The CPU is often called the "engine" of the computer, like a car engine. Well, your car's speedometer might go up to 120MPH, but you'll probably never drive at that maximum—for a variety of reasons! When it comes to CPUs, the stated clock rate is the maximum clock rate, and the CPU usually runs at a speed less than that; in fact, it can run at any speed below the maximum.

Now, we're all familiar with speeds such as 2.4GHz, 3.0GHz, or 3.2GHz. But what is the basis of these speeds? Speed can be broken down into three categories that are interrelated:

• Motherboard four clock speed: The base clock speed of the motherboard. Also referred to as the system bus speed, this speed is generated by a quartz oscillating crystal soldered directly to the motherboard. For example, the base clock speed on the motherboard used in Chapter 2 is 333MHz.
• External clock speed: This is the speed of the front side bus (FSB), which connects the CPU to the Memory Controller Hub (northbridge) on the motherboard. This is usually variable and depends on the CPU you install. In addition, it is determined from the base clock speed of the motherboard. For example, our motherboard's maximum external clock speed (or FSB) is 1333MHz. Simply put, this means that it is transferring four times the amount of data per cycle as compared to the original base clock speed. 333 MHz x 4 = 1,333MHz.
• Internal clock speed: This is the internal speed of the CPU. For this book I purchased the Intel Q8400 CPU that is rated at 2.66GHz. The CPU uses an internal multiplier that is also based off the motherboard base clock. The multiplier for this CPU is 8. The math is as follows: base clock speed x multiplier = internal clock speed. In our example, that would be 333MHz x 8 = 2.66GHz. Our motherboard can support faster CPUs also, for example, the Intel Q9650 that has an internal clock speed of 3.00GHz. This means that it has a multiplier of 9 (3.00GHz / 333MHz = 9). Some motherboards allow for overclocking (not ours), which enables the user to increase the multiplier within the BIOS, thereby increasing the internal clock speed of the CPU. This could possibly cause damage to the system, analogous to blowing the engine of a car when attempting to run a 10 second ¼ mile. So approach overclocking with caution.

32-Bit Versus 64-Bit

The bulk of today's CPUs are 64-bit; it's a type of CPU architecture that incorporates registers that are 64 bits wide. These registers, or temporary storage areas, allow the CPU to work with and process 64-bit data types and provide support for up to one-terabyte of platform address space. 64-bit CPUs have been available for PCs since 2003. Examples of 64-bit CPUs include the AMD Phenom and Intel Duo Core CPUs.

The predecessor to the 64-bit CPU was the 32-bit CPU. Intel started developing well-known 32-bit CPUs as early as 1985 with the 386DX CPU (which ran at a whopping 33MHZ!), and AMD did likewise in 1991 with the Am386. A 32-bit CPU can't support nearly as much address space as a 64-bit CPU; 32-bit is limited to 4GB. Most editions of Windows are available in both 32-bit and 64-bit versions.

You will probably still see 32-bit technologies (such as the Pentium 4) in the field; however, due to applications' ever-increasing need for resources, these older CPUs continue to diminish, whereas 64-bit technologies (such as Core 2 Duo) will become more prevalent.

You might hear of the terms x86 and x64. x86 refers to older CPU names that ended in an 86—for example, the 80386 (shortened to just 386), 486, or 586 CPU and so on. Generally, when people use the term x86, they refer to 32-bit CPUs that enable 4GB of address space. x64 (or x86-64) refers to newer 64-bit CPUs that are a superset of the x86 architecture. This technology can run 64-bit software and 32-bit software and can address a maximum of 1TB.

Windows Vista and Windows XP come in 64-bit and 32-bit versions so that users from both generations of computers can run the software efficiently. Windows 2000 Professional was designed for 32-bit CPUs only.

Sockets

The socket is the electrical interface between the CPU and the motherboard. It attaches directly to the motherboard and houses the CPU. It also physically supports the CPU and heat sink and enables for easy replacement of the CPU.

The socket is either made of plastic or metal, with metal contacts for connectivity to each of the pins/lands of the CPU. A metal lever (retaining arm) locks the CPU in place. Figure 3.1 shows an example of an unlocked socket.

Historically the socket has been considered a ZIF, short for zero insertion force. This means that the CPU should connect easily into the socket, with no pressure or force involved during the installation. Installing the CPU into these ZIF sockets is kind of like moving a planchette over a Ouija board until the CPU falls into place! Today's newer Land Grid Array (LGA) sockets require you to place the CPU into the socket housing, but it still doesn't require much force at all. The socket will have many pin inserts, or lands (on newer sockets), for the CPU to connect to. Pin 1 can be found in one of the corners and can be identified by one or more missing pins or pinholes depending on the type of socket. This helps you to orient the CPU, which also has the missing pin(s), or an arrow, in the corresponding corner. Here are two types of sockets you should know for the exam:

• PGA: Pin Grid Array sockets accept CPUs that have pins covering the majority of their underside. The pins on the CPU are placed in the pinholes of the socket, and the CPU is locked into place by a retaining arm. PGA has been in use since the late '80s, and is still in use on some motherboards today, but is quickly giving way to LGA.
• LGA: Land Grid Array sockets use lands that protrude out and touch the CPU's contact points. This newer type of socket (also known as Socket T) offers better power distribution and less chance to damage the CPU compared to PGA. LGA has been used since the later versions of Pentium 4 and is commonly used today.

The CPU and socket must be compatible. For example, the motherboard we use has an LGA775 CPU socket, which is common but not the only socket that Intel uses on its motherboards. The Q8400 CPU we use is designed to fit into the LGA775 socket, and several other CPUs are capable of fitting into this socket as well, but not all. For example some of Intel's Extreme CPUs are packaged differently and might need a different socket, such as the LGA771, which means a different motherboard must be used. Common sockets used by AMD are the Socket AM2 and AM2+.

CPU Cache

Several types of cache are used in computers, but CPU cache is a special high-speed memory that reduces the time the CPU takes to access data. By using high-speed static RAM (SRAM) and because the cache is often located directly on, or even in the CPU, CPU cache can be faster than accessing information from dynamic RAM (DRAM) sticks. However, it will be limited in storage capacity when compared to DRAM. Cache is divided into levels:

• Level 1: L1 cache is built in to the CPU and gives fast access to the most frequently used data. This level cache is the first one accessed by the CPU and is usually found in small amounts. However, it is the fastest cache to be found, offering the lowest latency of any of the types of cache. One of the reasons for this is that it resides within the CPU core. Our Q8400 CPU has 4 x 32KB of L1 cache; 32KB for each core. You can find more information about multi-core technology later in this chapter.
• Level 2: L2 cache can be built on to the CPU or placed on a separate chip on the motherboard. L2 cache is accessed after L1 cache, and it serves the CPU with less frequently used data in comparison to L1 but still more frequently used than DRAM data. L2 cache feeds the L1 cache, which in turn feeds the CPU. L2 is not as fast as L1 cache but is superior to DRAM sticks. Today's CPUs have the L2 cache directly on-die, and the cache takes up the majority of the CPU's real estate. The Q8400 CPU we use for our build has a total of 4MB L2 cache.

  Exam Alert

  Know the difference between L1 and L2 cache for the exam.

• Level 3: L3 cache comes in the largest capacities of the three types of cache and has the most latency; therefore, it is the slowest. If the CPU can't find what it needs in L1, it moves to L2 and finally to L3. Or you could think of it this way: L3 cache feeds L2 cache, which feeds L1 cache, which in turn feeds the CPU with data. If the CPU can't find the data it is seeking, it moves on to the DRAM sticks. L3 cache could be on-die or on-board, but most of today's CPUs (if they use it at all) have it on-die. Newer AMD CPUs utilize a large amount of L3 cache, but most Intel CPUs do not use it, although this could obviously change in the future.

Generally, the more cache the better. The less the CPU needs to access DRAM, the faster it can calculate data.

Hyper-Threading

Intel's Hyper-Threading (HT) enables a single CPU to accept and calculate two independent sets of instructions simultaneously, simulating two CPUs. The technology was designed so that single CPUs can compete better with true multi-CPU systems but without the cost involved. In an HT environment, only one CPU is present, but the operating system sees two virtual CPUs and divides the workload, or threads, between the two.

Hyper-Threading began during the Pentium 4 days, but is not used in Intel's Core 2 CPUs. However, in 2009 it made a return with the Core i7 CPU.

Multi-Core Technologies

Whereas HT technology simulates multiple CPUs, multi-core CPUs physically contain two or more actual processor cores, in one CPU package. These newer CPUs can have 2, 4, or even 8 cores, each acting as a single entity, but in many cases sharing the CPU cache. This enables for more-efficient processing of data. Not only is less heat generated, but also a 1.8GHz dual-core CPU can process more data per second than a 3.6GHz single-core CPU.

Current examples of multi-core CPUs include Intel's Core 2 Duo, Core 2 Quad, and Core 2 Extreme, and AMD's X2 and Phenom CPUs. Intel's new i7 Core CPUs combine multi-core technology with Hyper-Threading enabling for as many as eight simultaneous threads in a single CPU package. It just goes on and on!

Power Consumption

Power consumption of CPUs is normally rated in watts. For example, the Q8400 is rated as a 95 watt-hour CPU. This rating is known as thermal design point (TDP), and it signifies the maximum power that the computer's cooling system needs to dissipate heat generated by the CPU. This doesn't mean that it always uses that much power, but it should play into your decision when planning what power supply to use and what kind of cooling system. For more information on power supplies, see Chapter 5, "Power." One hundred watts, or thereabouts, is a common amount for multi-core CPUs. They are more efficient than their predecessor single-core CPUs, such as the Pentium D that could use as much as 215 watts.

Because we are talking electricity, another important factor is voltage. CPUs are associated with a voltage range; for example, the Q8400 ranges from 0.86V—1.28V. It is important to monitor the voltage that is received by the CPU; you can do this in the BIOS. If the CPU goes beyond the specified voltage range for any extended length of time, it will damage the CPU. This becomes especially important for overclockers.

Brands of CPUs

For the average user, it doesn't matter too much which CPU you go with. However, for the developer, gamer, video editor, or musician, it can make or break your computer's performance. Although the CompTIA A+ objectives cover only Intel and AMD (Advanced Micro Devices), you should be aware that there are others in the market. Intel and AMD dominate the PC and laptop arena, but other companies such as VIA have made great inroads into niche markets and are moving deeper into the laptop/mobile markets as well. CPU manufacturers use the make/model system. For example, the CPU we use is the Intel (make) Core 2 Quad Q8400 Yorkfield (model).

Intel Versus AMD

Intel and AMD are both good companies that make quality products, which leads to great competition. Which is better? In all honestly, it varies and depends on how you use the CPU. You can find advocates for both (albeit subjective advocates), and the scales are constantly tipping back and forth. On any given day, a specific Intel CPU might outperform AMD, and 3 months later, a different AMD CPU will outperform an Intel. It's been that way for years now. Table 3.1 and Table 3.2 give a synopsis of currently offered CPUs by the two manufacturers, with the latest at the top and the oldest at the bottom. All these are 64-bit CPUs.

Table 3.1. Comparison of Intel CPUs (as of July, 2009)

Intel CPU	Cores	Speed	L2 Cache	Bus Speed (FSB)
Core i7 Extreme	4	3.29–3.33GHz	8MB	—
Core i7	4	2.66–3.06GHz	8MB	—
Core 2 Extreme	4	2.66–3.2GHz	4–12MB	1066–1600MHz
Core 2 Quad	4	2.4–3.0GHz	4–12MB	1066–1333MHz
Core 2 Duo	2	1.8–3.33GHz	2–6MB	800–1333MHz

Table 3.2. Comparison of AMD CPUs (as of July, 2009)

AMD CPU	Cores	Speed	L2 Cache	L3 Cache
Phenom II
X2, X3, X4	2–4	2.4–3.1GHz	512KB	Maximum of 6144KB
Phenom X4	4	1.8–2.6GHz	512KB	2048KB
Phenom X3	3	1.9–2.5GHz	512KB	2048KB
Athlon X2 and II X2	2	1.9–3.1GHz	512KB	—

Whatever CPU you choose, make sure that you get a compatible motherboard. A few things to watch for are compatibility with the FSB (if applicable), chipset, socket type, and voltage. However, Intel and AMD have tools on their websites that make it easy for you to find compatible motherboards.

Cooling

Now that we know a CPU can effectively use as much electricity as a light bulb, we can understand why it gets so hot. Hundreds of millions of transistors are hammering away in these powerhouses, so we need to keep it and other devices in the computer cool. This is done in a few ways as outlined in this section.

Heat Sinks

The heat sink is a block of metal made to sit right on top of the CPU, with metal fins stretching away from the CPU. It uses conduction to direct heat away from the CPU and out through the fins. With passive heat sinks, that's all there is to it. But with active heat sinks, a fan is attached to the top of the heat sink. The fan plugs into the motherboard for power and usually blows air into the heat sink and toward the CPU helping to dissipate heat through the heat sink fins. More powerful aftermarket CPU fans can be installed as well; just make sure that your power supply can handle the increased power requirements. In today's motherboards the chipset's northbridge and southbridge have passive heat sinks, but all new CPUs come with active heat sinks. Traditionally heat sinks have been made of aluminum, but now you also see copper heat sinks used due to their superior conductivity.

Thermal Compound

The CPU cap and the bottom of the heat sink have slight imperfections in the metal. The best heat dissipation from CPU to heat sink would occur if the metal faces on each were completely and perfectly straight and flat, but you would find that only in a platinum-iridium alloy. So, to fill the tiny gaps and imperfections, thermal compound (aka thermal interface material or TIM) is used. One example of thermal compound is Arctic Silver, available online and at various electronics stores. Now, if this is a new installation, thermal compound is probably not needed. Most new CPUs' heat sinks have factory applied thermal compound that spreads and fills the gaps automatically after you install the heat sink and boot the computer. However, if you need to remove the heat sink for any reason, for example to clean it, thermal compound should be applied to the CPU cap before re-installing the heat sink, or installing a new heat sink. To do this, first clean any old thermal compound off of the CPU cap and the heat sink with TIM remover such as Akasa TIM-Clean. Then, clean a credit card with isopropyl alcohol or denatured alcohol. Next, apply a small amount of thermal compound to the center of the CPU cap. (This is the top of the installed CPU. You don't want to get any thermal compound on the actual CPU or motherboard.) With the credit card, spread the thermal compound carefully so that that you end up with a thin layer. Finally, install the heat sink. Try to do so in one shot without jostling the heat sink excessively.

Fans

Case fans are also needed to get the heat out of the case. The power supply has a built-in fan that is adequate for lesser systems. However, multi-core systems should have at least one extra exhaust fan mounted to the back of the case, and many cases today come with one for this purpose. An additional fan on the front of the case can be used as an intake of cool air. If you aren't sure which way the fan blows, connect its power cable to the computer but don't mount it; then hold a piece of paper against the fan. The side that pulls the paper toward it should be the side facing the front of the computer when it is mounted. Some cases come with fans that are mounted to the top, which is also ingenious because heat rises. Another thing to consider is where the heat goes after it leaves the case. If the computer is in an enclosed area, the heat will have a hard time escaping and might end up back in the computer. Make sure there is air flow around the computer case. I have seen some people point the front of their computer toward an AC vent in the summer and even use special exhaust fans (such as bathroom fans) that butt up against the power supply or secondary exhaust fan on the case and lead hot air directly out of the house, but I digress.

Another possibility is a solution Intel developed called the Chassis Air Guide system, which is essentially a hollow tube that leads from the side of the case to the CPU, guiding cool room ambient air toward the CPU. For more information on the Intel Chassis Air Guide and Intel's Thermally Advantaged Tested Chassis list, see the following link: http://www.intel.com/go/chassis/. Of course, three or four fans can make a decent amount of noise, and they still might not be enough for the most powerful computers, especially the overclocked ones, which leads us to our next option.

Liquid Cooling Systems

Although still uncommon, liquid cooled systems are looked at as more of a viable option than they would have been 5 or 10 years ago. And newer water cooling kits can be used to not only cool the CPU, but also the chipset, hard drives, video cards, and more. A kit usually comes with a CPU water block, pump, radiator/fan, PVC tubing, and of course, coolant. The advantages are improved heat dissipation (if installed properly), higher overclocking rates, and support for the latest, hottest CPUs. The disadvantage as you can guess is the risk of a leak that can damage components. Due to the complexity of the installation, and the fact that most computers do not need this level of heat dissipation, liquid cooling is usually employed only by enthusiasts.

Cram Quiz

Answer these questions. The answers follow the last question. If you cannot answer these questions correctly, consider reading this section again until you can.

1. Which of these is the speed of the CPU?

   	A.	External clock speed
   	B.	FSB
   	C.	Internal clock speed
   	D.	System bus speed

2. Which of the following are 64-bit CPUs? (Select all that apply.)

   	A.	Core 2 Duo
   	B.	Phenom II
   	C.	Pentium III
   	D.	Celeron

3. Which is the fastest cache memory?

   	A.	L2
   	B.	L3
   	C.	HTTP
   	D.	L1

4. What does Hyper-Threading do?

   	A.	It gives you multiple cores within the CPU.
   	B.	It enables for four simultaneous threads to be processed by one CPU core.
   	C.	It enables for two simultaneous threads to be processed by one CPU core.
   	D.	It is a high-speed connection from the CPU to RAM.

5. What seals the tiny gaps between the CPU cap and the heat sink?

   	A.	Thermal jelly
   	B.	Peanut butter and jelly
   	C.	3-in-1 house oil
   	D.	Thermal compound

6. What is the amount of power required to cool the computer?

   	A.	FSB
   	B.	TDP
   	C.	MMX
   	D.	TDK

7. Which kind of socket incorporates "lands" to ensure connectivity to a CPU?

   	A.	PGA
   	B.	Chipset
   	C.	LGA
   	D.	Copper

Cram Quiz Answers

1. C. The internal clock speed is the speed of the CPU, for example 2.4GHz. The external clock speed is the speed of the FSB, and the system bus speed (base clock) is what the internal clock speed is based off. An example of a base clock system bus speed would be 333MHz.

2. A and B. Intel's Core 2 Duo and AMD's Phenom II are both 64-bit CPUs. The Pentium III and Celeron are 32-bit CPUs.

3. D. L1 is the fastest cache memory and is located within the CPU's core.

4. C. Hyper-Threading allows for an operating system to send two simultaneous threads to be processed by a single CPU core. The OS views the CPU core as two virtual processors. Multiple cores would infer multi-core technology that means that there are two physical processing cores within the CPU package. The high-speed connection used by AMD from the CPU to RAM is Hyper-Transport.

5. D. Thermal compound is used to seal the small gaps between the CPU and heat sink. Did I ever tell you about the time I found grape jelly inside a customer's computer?

6. B. TDP (Thermal design point) is the amount of power required to cool a computer and is linked directly to the amount of heat a CPU creates.

7. C. LGA (Land Grid Array) is the type of socket that uses "lands" to connect the socket to the CPU. PGA sockets have pinholes that make for connectivity to the CPU's copper pins.