CompTIA A+ Cert Guide: Motherboards, Processors, and Adapter Cards

This chapter is from the book 

Processors and CPUs

To do well on A+ Certification exams, you must understand the major types of processors available for recent systems, their technologies, how to install them, and how to troubleshoot them.

Overview of Processor Differences

Although Intel and AMD processors share two common architectures, x86 (used for 32-bit processors and for 64-bit processors running in 32-bit mode) and x64 (an extension of x86 that enables larger files, larger memory sizes, and more complex programs), these processor families differ in many ways from each other, including:

• Different processor sockets
• Different types of microcode
• Differences in dual-core, triple-core, and quad-core designs (two or more processor cores help run multiple programs and programs with multiple execution threads more efficiently)
• Cache sizes (cache memory stores a copy of recently-read memory locations to help improve system performance; L1 cache is in the processor core; L2 and L3 cache are in the processor module but outside the core)
• Performance versus clock speed

Intel Processors

Intel processors developed from 2000 to the present include the following product families:

• Pentium III
• Pentium 4
• Pentium D
• Celeron
• Core 2 Duo
• Core 2 Quad
• Core i3
• Core i5
• Core i7

The Pentium III processor was the last Intel processor produced in both a slot-based and socket-based design. Slot-based versions use Slot 1, the same slot design used by the Pentium II and slot-based Celeron processors. Socketed versions use Socket 370, which is mechanically the same as the socket used by the first socketed Celeron processors. However, some early Socket 370 motherboards are not electrically compatible with the Pentium III.

The Pentium 4 replaced the Pentium III and ran at much higher clock speeds. Early versions used Socket 423, a socket used by no other Intel processor. Most Pentium 4 designs used Socket 478, while late-model Pentium 4 designs used Socket 775, which is also used by current Intel processors. The different sockets used by the Pentium 4 were necessary because of substantial design changes throughout the processor's lifespan, including the introduction of 64-bit extensions (x64).

The Pentium 4's successor was the Pentium D, which is essentially two Pentium 4 processor cores built into a single physical processor. Although it used the same Socket 775 as late-model Pentium 4 processors, it required support from different chipsets because data was transferred between processor cores via the Memory Controller Hub (North Bridge) component. The Pentium D was Intel's first dual-core processor. The Pentium Extreme Edition is a faster version of the Pentium D designed for gaming or other high-performance tasks. The Pentium D and Pentium Extreme Edition both support x64 extensions, as does the Core 2 Duo.

The Pentium D was replaced by the Core and Core 2 families of processors. The Core and Core 2 families use processor architectures that emphasize real-world performance over clock speed. The first Core 2 processors were the Core 2 Duo (featuring two processor cores), followed by the Core 2 Quad models (with four processor cores). Although Core 2 processors run at much slower clock speeds than the fastest Pentium 4 or Pentium D processors, they perform much better in real-world operations. Core processors are single-core, while Core Duo and Core 2 Duo are dual-core. Core, Core Solo, and Core Duo processors are x86 (32-bit), while Core 2 Duo, Quad, and Extreme processors are x64 (64-bit).

The most recent processors in the Core family include the Core i7, Core i5, and Core i3, all of which support x64 (64-bit) processing. The Core i7 features quad core or six core designs with Intel HT Technology (hyperthreading, which supports two processor threads per core), Intel VT-x hardware-assisted virtualization, and Intel Turbo Boost overclocking. Core i5 is a simplified version of the Core i7, with only a few dual-core models supporting HT Technology (quad-core Core i5 does not support HT Technology); however, all Core i5 desktop processors include VT-x and Turbo Boost and some also include integrated graphics. Core i3 processors are dual-core with support for HT Technology, VT-x, and integrated graphics, but lack Turbo Boost. Note that mobile processors with these same model numbers differ in some details.

Celeron is actually a brand name rather than a specific processor design. Celeron processors have been based on the Pentium II, Pentium III, Pentium 4, and Core 2 processors. However, they feature lower clock speeds, slower front side bus speeds (the clock speed of the memory bus), and smaller L2 caches, making them less powerful (and less expensive) processors than the designs they're based on. Very few Celeron models support x64 extensions.

Because most Intel processor families have gone through many changes during their lifespans, specific models are sometimes referred to by their code names. In an attempt to make it easier to understand the performance and feature differences of models in a particular processor family, Intel has assigned processor numbers to recent versions of the Pentium 4, as well as all more recent processors.

Table 3-5 provides a brief summary of Intel desktop processors produced from 1998 to mid 2010. For additional details, see Upgrading and Repairing PCs, 19th Edition by Scott Mueller (Que Publishing).

Table 3-5. Intel Desktop Processors from Pentium III through Core i7

Processor	Code Names	Clock Speed Range	FSB Speed	Processor Socket or Slot	L2 Cache Sizes	Based On or Notes
Pentium III	Katmai, Coppermine, Coppermine-T, Tualatin	450MHz–1.3GHz	100MHz, 133MHz	Slot 1, Socket 370	256KB or 512KB	—
Celeron	Coppermine-128, Tualatin 256	533MHz–1.4GHz	66MHz, 100MHz	Slot 1, Socket 370	128KB, 256KB	Pentium III
Pentium 4	Willamette, Northwood, Prescott, Cedar Mill	1.4GHz–3.8GHz	400MHz, 533MHz, 800MHz	Socket 423, Socket 478, Socket 775	256KB, 512KB, 1MB, 2MB	—
Pentium 4 Extreme Edition	Gallatin, Prescott 2M	3.2GHz–3.733GHz	800MHz	Socket 775	512KB+2MB L3 or 2MB	Pentium 4 Prescott
Celeron	Willamette-128, Northwood-128	1.7GHz–2.8GHz	400MHz	Socket 478	128KB	Pentium 4 Willamette, Northwood
Celeron D	Prescott-256, Cedar Mill-512	2.13GHz–3.6GHz	533MHz	Socket 478, Socket 775	256KB, 512KB	Pentium 4 Prescott, Cedar Mill
Pentium D	Smithfield	2.66GHz–3.66GHz	533MHz, 800MHz	Socket 775	1MBx2 or 2MB x2	Dual-core version of Pentium 4 Prescott
Pentium Extreme Edition	Smithfield	3.73GHz	800MHz	Socket 775	2MBx2	Pentium 4 Prescott
Core 2 Duo	Conroe, Wolfdale, Allendale	1.80GHz–3.33GHz	800MHz, 1066MHz, 1333MHz	Socket 775	2MB, 4MB, 6MB	Dual-core version of Core (notebook processor)
Core 2 Extreme	Conroe XE	2.93GHz	1066MHz	Socket 775	4MB	Core 2 Duo Conroe
Celeron	Conroe L	1.2–2.2GHz	800MHz	Socket 775	512KB	Single-core version of Core 2 Duo Conroe
Celeron	Allendale-512	1.6–2.4GHz	800MHz	Socket 775	512KB	Core 2 Duo Allendale
Core 2 Quad	Kentsfield	2.4–2.6GHz	1066MHz	Socket 775	4MBx2	Two Core 2 Duo Conroe cores
Core 2 Quad	Yorkfield	2.26–3.0GHz	1333MHz	Socket 775	3MBx2, 6MBx2	Integrated quad-core design
Core 2 Extreme	Kentsfield XE	2.66–3.0GHz	1066MHz, 1333MHz	Socket 775	4MBx2	Core 2 Quad Kentsfield
Core 2 Extreme	Yorkfield XE	3.0–3.2GHz	1333MHz, 1600MHz	LGA-771	6MBx2	Core 2 Quad Yorkfield
Core i3	Clarkdale	2.93–3.33GHz	1066MHz, 1333MHz	FCLGA-1156	4MB	Clarkdale used for Core i3, Core i5
Core i5	Clarkdale	2.4–3.46GHz	1066MHz, 1333MHz	FCLGA-1156	4MB	Clarkdale used for Core i3, Core i5
Core i7	Lynnfield, Bloomfield, Gulftown	2.66–3.2GHz	800MHz– 1066MHz; 1066MHz– 1333MHz	FCLGA-1156	8MB–12MB	Gulftown is a six-core processor; others quad-core
Socket 775 is also referred to as LGA-775 because the socket contains leads that connect with solder balls on the bottom of the processor.
FCLGA-1156 is also known as LGA-1156
The processor code names and performance in this table are effective as of mid 2010.

AMD Processors

AMD processors contemporary with the Intel Pentium III and its successors include the following processor families as of mid 2010:

• Athlon
• Duron
• Athlon XP
• Sempron
• Athlon 64
• Athlon 64 FX
• Athlon 64 X2
• Phenom X3
• Phenom X4
• Phenom II X2
• Phenom II X3
• Phenom II X4
• Phenom II x6

The Athlon processor was the first (and last) AMD processor produced in a slot-based design. It uses Slot A, which physically resembled Slot 1 used by Intel Pentium II and Pentium III models, but was completely different in its pinout. Later versions of the Athlon switched to Socket A, a 462-pin socket, which was also used by the Duron, Athlon XP, and Socket A versions of the Sempron.

The Athlon XP replaced the Athlon, and featured higher clock speeds and larger L2 cache. The lower-performance counterpart of the Athlon and Athlon XP was the Duron, which featured a smaller L2 cache and slower FSB speed.

The Athlon XP design was used for the Socket A versions of the Sempron when AMD moved to 64-bit processing with the introduction of the Athlon 64, AMD's first x64 64-bit desktop processor.

The Athlon 64 family initially used Socket 754, but because the memory controller is built into the processor, rather than into the North Bridge as on conventional processors, it was necessary to develop a new Socket 939 to support dual-channel memory.

The Athlon 64 FX is a faster performance–oriented version of the Athlon 64. Initial versions were based on the Opteron workstation and server processor, and thus used Socket 940. Later versions used Socket 939 and its successor, Socket AM2.

AMD's first dual-core processor was the Athlon 64 X2, which uses a design that permits both processor cores to communicate directly with each other, rather than using the North Bridge (Memory Controller Hub) as in the Intel Pentium D. This enabled upgrades from Socket 939 Athlon 64 to the X2 version after performing a BIOS upgrade.

AMD's economy version of the Athlon 64 is also called the Sempron, various versions of which have used Socket 754 and Socket 939.

AMD's Phenom series is based on the AMD K10 processor architecture, and all Phenoms include multiple processor cores that are built as a single unit. Phenom II is an improved version of Phenom, featuring a smaller process, more cache, and better cache management. The Phenom II series uses a more efficient socket and increases the total possible amount of processor cores to 6. Processor speeds are also increased in this series. These processors use a more powerful chipset. The chipset is the main controller of the motherboard. When selecting an AMD processor, the motherboard's chipset should be taken into account to ensure compatibility.

AMD's Athlon II series is available in dual-core, triple-core, quad-core, and six-core versions (X2, X3, X4, and X6).

Because most AMD processor families have gone through many changes during their lifespans, specific models are sometimes referred to by their code names.

Table 3-6 provides a brief summary of AMD desktop processors produced over the last decade. For additional details, see Upgrading and Repairing PCs, 19th Edition.

Table 3-6. AMD Desktop Processors from Athlon through Phenom II

Processor	Code Names	Clock Speed Range	FSB Speed	Processor Socket or Slot	L2 Cache Sizes	Based On
Athlon	K7, K75, Thunderbird	500MHz–1.4GHz	200–266MHz	Slot A, Socket A (aka Socket 462)	256–512KB	—
Athlon XP	Palomino, Thoroughbred, Thorton	1.333–2.2GHz	266–400MHz	Socket A	256–512KB	—
Duron	Spitfire, Morgan, Applebred, Appaloosa	550MHz–1.8GHz	200–266MHz	Socket A	64KB	Athlon Socket A
Sempron	Thorton, Barton	1.5–2.2GHz	166–200MHz	Socket A	256KB (Thorton), 512KB (Barton)	Athlon XP
Sempron	Paris, Palermo	1.4–2.0GHz	800MHz–1GHz	Socket 754	128–256KB	Athlon 64 (Socket 754 versions)
Athlon 64	ClawHammer, Newcastle, San Diego, Venice, Orleans	1–2.6GHz	800MHz–1GHz	Socket 754, Socket 939, Socket 940, Socket AM2	512KB–1MB	—
Sempron	Palermo	1.8–2.0GHz	800MHz	Socket 939	128KB–256KB	Athlon 64
Athlon 64 FX	See Athlon 64 code names; also Windsor	2.2–2.8GHz	800MHz–1GHz	Socket 939, Socket 940	1MB	Athlon 64
Sempron	Manila, Sparta	1.6–2.3GHz	800MHz	Socket AM2	128-256-512KB	Athlon 64
Athlon 64 X2	Manchester, Toledo, Windsor, Brisbane	1.9–3.2GHz	1GHz	Socket 939 (Manchester, Toledo), Socket AM2 (Windsor, Brisbane)	256KBx2; 512KBx2; 1MBx2	Dual-core version of Athlon 64
Athlon 64 FX	Toledo, Windsor	2.0–3.2GHz	1GHz	Socket 939 (Toledo), Socket AM2 (Brisbane)	1MBx2	Dual-core version of Athlon 64 FX
Phenom X4	Agena	1.8–2.6 GHz	1.6-2GHz	Socket AM2	512KBx4 + 2MB L3	K10 microarchitecture
Phenom X3	Toliman	2.1–2.5GHz	1.6-1.8GHz	Socket AM2	512KBx3 + 2MB L3	K10 microarchitecture
Phenom X2	Kuma	2.3–2.8GHz	1.8GHz	Socket AM2+	512KBx2 + 2MB L3	K10 microarchitecture
Athlon X2	Kuma	2.3–2.8GHz	1.8GHz	Socket AM2+	512KBx2 + 2MB L3	Phenom X2
Phenom II X2	Callisto	2.8–3.3GHz	2GHz, 2.2GHz	Socket AM3	512KBx2 + 6MB L3	Deneb with two cores disabled
Phenom II X3	Heka	2.4–3.2GHz	2GHz	Socket AM3	512KBx3 +6MB L3	Deneb with one core disabled
Phenom II X4	Deneb	2.5–3.5GHz	1.8GHz, 2GHz	Socket AM2+, AM3	512KBx4 + 4MB or 6MB L3	DDR3 memory supported on Socket AM3 only
Phenom II X6	Thuban	2.7–3.2GHz	2GHz	Socket AM3	512KBx6 +6MB L3	Includes Turbo Core overclock support
Athlon II	Regor	1.8–2.0GHz	1.8GHz	Socket AM3	1MB	Single-core version of Regor (X2)
Athlon II	Regor (X2)	1.6–3.3GHz	2GHz	Socket AM3	512KBx2 or 1024KBx2	Phenom II without L3 cache
Athlon II	Rana (X3)	2.2–3.2GHz	2GHz	Socket AM3	512KBx3	Phenom II without L3 cache
Athlon II	Propus (X4)	2.2–3.1GHz	2GHz	Socket AM3	512KBx4	Phenom II without L3 cache
The processor code names and performance in this table are effective as of mid 2010.

Processor Sockets and Packaging

Most processors listed in the previous sections use some form of the pin grid array (PGA) package, in which pins on the bottom of the processor plug into holes in the processor socket. The exceptions include slot-mounted processors (Slot 1 and Slot A) and the current LGA and FCLGA sockets, which use a different type of processor package called the land grid array (LGA). LGA packaging uses gold pads on the bottom of the processor package to connect with raised leads in the processor socket.

Figure 3-21 compares processor packages and sockets to each other.

CPU Technologies

Processor technologies in the following sections might be used by AMD only, by Intel only, or by both vendors. These technologies are used to help distinguish different processors from each other in terms of performance or features.

Hyperthreading (HT Technology)

Hyperthreading (HT Technology) is a technology developed by Intel for processing two execution threads within a single processor. Essentially, when HT Technology is enabled in the system BIOS and the processor is running a multithreaded application, the processor is emulating two physical processors. The Pentium 4 was the first desktop processor to support HT Technology, which Intel first developed for its Xeon workstation and server processor family.

Pentium 4 processors with processor numbers all support HT Technology, as do older models with 800MHz FSB and a clock speed of 3.06GHz or higher. HT Technology is also incorporated in a number of more recent dual-core, quad-core, and six-core processors in the Core 2, Core i5, and i7 series to further improve the execution of multithreaded applications.

Dual-Core and Multi-Core

Two or more physical processors in a system enable it to perform much faster when multitasking or running multithreaded applications. However, systems with multiple processors are very expensive to produce and some operating systems cannot work with multiple processors. Dual core processors, which combine two processor cores into a single physical processor, provide virtually all of the benefits of two physical processors, and are lower in cost and work with any operating system that supports traditional single-core processors.

The first dual-core desktop processors were introduced by Intel (Pentium D) and AMD (Athlon 64 X2) in 2005. Athlon 64 X2's processor cores communicate directly with each other, enabling systems running single-core Athlon 64 processors to swap processors after a simple BIOS upgrade. The Pentium D, on the other hand, required new chipsets to support it. Core 2 Duo, Core i3, and some versions of the Core i5 represent major current dual-core processor families. Like the AMD Athlon 64 X2 and newer AMD dual-core processors, these processors' cores communicate directly with each other.

Both Intel and AMD have released processors that include more than two cores. Intel's Core 2 Quad, Core i7, and some versions of the Core 2 Extreme contain four or more processor cores, while AMD's Phenom and Phenom II are available in versions with two, three, four, or more processor cores.

Processor Throttling

Processors do not need to run at full speed when they have little, or no, work to perform. By slowing down—or throttling—the processor's clock speed when the workload is light, the processor runs cooler, the system uses less energy, and—in the case of mobile systems—the computer enjoys a longer battery life. Throttling, sometimes referred to as thermal throttling, can also take place when a processor gets too hot for the computer's cooling system to work properly.

Intel uses the terms SpeedStep or Enhanced SpeedStep for its throttling technologies. AMD uses the term Cool'n'Quiet for its throttling technology.

Microcode (MMX)

All Intel and AMD processors in current use include various types of microcode instructions for boosting multimedia performance. The first processor to include this type of microcode was the Pentium MMX, which included 57 new instructions (known as MMX) for working with multimedia. MMX was the first example of what is known as single instruction, multiple data (SIMD) capability.

Later Intel processors included enhanced versions of MMX known as SSE (MMX+70 additional instructions, introduced with the Pentium III), SSE2 (MMX+SSE+144 new instructions, introduced with the Pentium 4), SSE3 (MMX+SSE+SSE2+13 new instructions, introduced with the Pentium 4 Prescott), and, most recently, SSSE3 (MMX+SSE+SSE2+SSE3+32 new instructions, introduced with the Core 2 Duo). The SSE4 instruction set, which adds 51 new instructions, was introduced with the introduction of 45nm processor technology in the Penryn versions of the Core 2 Duo and subsequent processors. SSE4.1 is a subset of SSE4, containing 47 instructions. SSE4.2 includes the seven remaining instructions and was introduced with the Core i7. The term "HD Boost" refers to SSE4 support.

AMD also provides multimedia-optimized microcode in its processors, starting with 3DNow! (introduced by the K6, which was roughly equivalent to the Pentium MMX). However, AMD's version differs in details from Intel's, offering 21 new instructions. The AMD Athlon introduced 3DNow! Enhanced (3DNow!+24 new instructions), while the Athlon XP introduced 3DNow! Professional (3DNow!+Enhanced+51). 3DNow! Professional is equivalent to Intel's SSE. Starting with the Athlon 64 family, AMD now supports SSE2, and it added SSE3 support to the Athlon 64 X2 and newer versions of the Athlon 64 family. AMD also supports four SSE4 instructions as well as two SSE instructions known as SSE4a.

Overclocking

Overclocking refers to the practice of running a processor or other components, such as memory or the video card's graphics processing unit (GPU) at speeds higher than normal. Overclocking methods used for processors include increasing the clock multiplier or running the front side bus (FSB) at faster speeds than normal. These changes are performed by altering the normal settings in the system BIOS setup for the processor's configuration. Figure 3-22 is a typical BIOS processor configuration screen.

Most processors feature locked clock multipliers. That is, the clock multiplier frequency cannot be changed. In such cases, the only way to overclock the processor is to increase the front side bus speed, which is the speed at which the processor communicates with system memory. Increasing the FSB speed can lead to greater system instability than changing the clock multipliers.

Some processors from Intel and AMD feature unlocked clock multipliers, so that the user can choose the best method for overclocking the system. Overclocked processors and other components run hotter than normal, so techniques such as using additional cooling fans, replacing standard active heat sinks with models that feature greater cooling, and adjusting processor voltages are often used to help maintain system stability at faster speeds.

Intel's Core i7, Core i5, and AMD's Phenom II series support automatic overclocking according to processor load. Intel refers to this feature as Turbo Boost, while AMD's term is Turbo Core.

Cache

Cache memory, as mentioned previously, improves system performance by enabling the processor to reuse recently retrieved memory locations without needing to fetch them from main memory. Processors from AMD and Intel feature at least two levels of cache:

• Level 1 (L1) cache is built into the processor core. L1 cache is relatively small (8KB–64KB). When the processor needs to access memory it checks the contents of L1 cache first.
• Level 2 (L2) cache is also built into the processor. On older slot-mounted processors, L2 cache was external to the processor die, and ran at slower speeds than the processor. On socketed processors, L2 cache is built into the processor die. If the processor does not find the desired memory locations in L1 cache, it checks L2 cache next.
• Level 3 (L3) cache is found on some very high-performance processors from Intel (such as the Core i7 series) and on several high-performance and mid-level processors from AMD. L3 is also built into the processor die. On systems with L3 cache, the processor checks L3 cache after checking L1 and L2 caches.

If cache memory does not contain the desired information, the processor retrieves the desired information from main memory, and stores copies of that information in its cache memory (L1 and L2, or L1, L2, and L3). Processors with larger L2 caches (or L2 and L3 caches) perform most tasks much more quickly than processors that have smaller L2 caches for two reasons. Cache memory is faster than main memory, and the processor checks cache memory for needed information before checking main memory.

VRM

Starting with Socket 7 versions of the Intel Pentium, processors have not received their power directly from the power supply. Instead, a device called a voltage regulator module (VRM) has been used to reduce 5V or 12V DC power from the power supply to the appropriate power requested by the processor through its voltage identification (VID) logic.

Although some motherboards feature a removable VRM, most motherboards use a built-in VRM that is located next to the processor socket, as shown in Figure 3-23.

Speed (Real Versus Actual): Clock Speed Versus Performance

A common measurement of processor performance has been clock speed. However, clock speed can be misleading. For example, the Intel Core 2 Duo and AMD Athlon 64 X2 processors perform computing tasks much more quickly than the Pentium D, even though the Pentium D runs at a much higher clock speed.

To determine the actual performance of a processor, you should use benchmark tests such as Futuremark's SYSmark, PCMark, and 3DMark.

32-bit Versus 64-bit

Processors developed before the AMD Athlon 64 were designed only for 32-bit operating systems and applications. 32-bit software cannot access more than 4GB of RAM (in fact, 32-bit Windows programs can use only 3.25GB of RAM), which makes working with large data files difficult, as only a portion of a file larger than the maximum memory size can be loaded into memory at one time.

The Athlon 64 was the first desktop processor to support 64-bit extensions to the 32-bit x86 architecture. These 64-bit extensions, commonly known as x64, enable processors to use more than 4GB of RAM and run 64-bit operating systems, but maintain full compatibility with 32-bit operating systems and applications.

Late-model Pentium 4 processors from Intel also support x64, as do subsequent processors such as the Pentium 4 Extreme Edition, Pentium D, Pentium Extreme Edition, Core 2 Duo, Core 2 Quad, Core 2 Extreme, Core i3, Core i5, and Core i7. Subsequent AMD processors including the Athlon X2, Athlon II, Phenom, and Phenom II also support x64. Most processors made today support x64 operation.

Choosing the Best Processor for the Job

If you are buying or building a new system, you have free rein in the choice of a processor to build the system around. This section describes important considerations.

Performance

If you need a system that can handle high-resolution graphics and video, and can perform heavy-duty number crunching, get the fastest dual-core or multi-core processor you can afford. However, if your requirements are less extreme, you can save money for your clients by opting for a processor from the same family with slower clock speed or less cache memory.

Thermal Issues

Many processor models are available in two or more versions that differ in their thermal requirements; that is, the type of active heat sink necessary to cool them and the amount of power (in watts) needed to operate them. This figure is often referred to as Max TDP (maximum thermal design power). In a mid-tower or full tower system, these considerations might be less important than in a micro-tower or small form factor system, or a system that might need to run as quietly as possible.

32-bit Versus 64-bit (x64) Compatibility

Unless you are trying to build the least-expensive system possible, you will find it difficult to find 32-bit only processors today. However, if you are repurposing existing systems, you might need to determine which systems include processors with support for 64-bit operation, and which support only 32-bit operation.

Other Processor Features

Processor features such as NX (no execute, which provides hardware-based protection against some types of viruses and malware) and hardware-based virtualization (which enables a single processor to be split into multiple virtual machines with little or no slowdown) are also important to consider in business environments. Check the specification sheets provided by processor vendors to determine the exact features supported by a particular processor.

Installing Processors

Processors are one of the most expensive components found in any computer. Because a processor can fail, or more likely, might need to be replaced with a faster model, knowing how to install and remove processors is important. On the A+ Certification exams, you should be prepared to answer questions related to the safe removal and replacement of socketed processors.

The methods used for CPU removal vary according to two factors: the processor type and the socket/slot type.

As you saw in Tables 3-5 and 3-6, most recent processors are socketed. Before the development of the ZIF socket, the processor was held in place by tension on the chip's legs, pins, or leads. Thus, to remove these chips, you must pull the chip out of the socket. Because the chip's legs, pins, or leads are fragile, special tools are strongly recommended for removing chips that are not mounted in ZIF sockets.

Before removing and installing any CPU or other internal component, be sure to review and follow the ESD precautions discussed in Chapter 17.

Removing the Heat Sink

ZIF sockets are used on almost all desktop systems using Pentium III-class or newer socketed processors (except for processors using LGA sockets). They allow easy installation and removal of the processor.

What makes ZIF sockets easy to work with? They have a lever that, when released, loosens a clamp that holds the processor in place.

If the processor has a removable heat sink, fan, or thermal duct that is attached to the motherboard, you must remove these components before you can remove the processor.

Heat sinks used on Socket 370 and Socket A processors have a spring-loaded clip on one side and a fixed lug on the other side. To release this clip, press down on it using a screwdriver, as shown in Figure 3-24.

Most newer processors use heat sinks that are attached to a frame around the processor or are mounted through the motherboard. To release these heat sinks, you might need to flip up a lever on one side of the heat sink or release the locking pins. Figure 3-25 illustrates a typical installation on an Athlon 64 processor, and Figure 3-26 illustrates the components of a typical heat sink for LGA 775 processors.

BTX systems use a horizontally mounted thermal module that is equipped with a fan. The thermal module also helps cool other components such as the motherboard chipset and memory. Figure 3-27 illustrates a typical thermal module installed on a motherboard. Note that the front of the thermal module extends below the edge of the motherboard to provide cooling for both top and bottom.

To remove a thermal module from a BTX motherboard, follow these steps:

• Step 1.Remove the screws that attach the module to the retention bracket on the underside of the motherboard.
• Step 2.Disconnect the thermal module's fan power lead.
• Step 3.Lift the thermal module off the processor.

Be careful when removing head sinks or thermal modules. Be careful not to drop the heat sink or thermal module on the CPU or on the motherboard. Heat sinks and thermal modules are bulky and heavy and can easily damage the expensive parts of the your computer.

Removing the Processor

After removing the heat sink, follow these instructions to complete the processor removal process.

• Step 1.Disconnect the active heat sink (if included) from its power source and lift the assembly away.
• Step 2.Push the lever on the ZIF socket slightly to the outside of the socket to release it.
• Step 3.Lift the end of the lever until it is vertical (see Figure 3-28). This releases the clamping mechanism on the processor's pins.
  

  Figure 3-28 After the heat sink fan is disconnected from power (left) to reveal the processor (center), the lever on the ZIF socket (right) can be lifted to release the processor.

• Step 4.Grasp the processor on opposite sides, making sure not to touch the pins, and remove it from the socket. Put it into antistatic packaging.

The process of removing an LGA-based processor is a bit different:

• Step 1.Disconnect the active heat sink (if included) from its power source and lift the assembly away.
• Step 2.Lift the locking lever to release the load plate, which holds the processor in place.
• Step 3.Carefully lift the processor away and place in into antistatic packaging.

Be careful when removing the processor and when unlocking any sockets. These components are very delicate. Think of yourself as a watchmaker when dealing with these parts!

Installing a New Processor

Before installing a new processor, verify that the processor you plan to install is supported by the motherboard. Even though a particular combination of processor and motherboard might use the same socket, issues such as BIOS, voltage, memory support, or chipset considerations can prevent some processors from working on particular motherboards. You can destroy a processor or motherboard if you install a processor not suitable for a particular motherboard.

After verifying compatibility by checking the system or processor manual (and installing any BIOS updates required for processor compatibility), check a PGA-type processor for bent pins, and the socket of an LGA processor for bent leads. Correct these problems before continuing.

To insert a PGA-type CPU into a ZIF socket, find the corner of the chip that is marked as pin 1 (usually with a dot or triangle). The underside of some chips might be marked with a line pointing toward pin 1. Then follow these steps:

• Step 1.Line up the pin 1 corner with the corner of the socket also indicated as pin 1 (look for an arrow or other marking on the motherboard). If you put the chip in with pin 1 aligned with the wrong corner and apply the power, you will destroy the chip.
• Step 2.Make sure the lever on the ZIF socket is vertical; insert the CPU into the socket and verify that the pins are fitting into the correct socket holes.
• Step 3.Lower the lever to the horizontal position and snap it into place to secure the CPU.
• Step 4.Before attaching the heat sink or fan, determine if the heat sink has a thermal pad (also called a phase-change pad) or if you need to apply thermal compound to the processor core (refer to Figure 3-27). Remove the protective tape from the thermal pad or apply thermal compound as needed. Attach the heat sink or fan. You must use some type of thermal compound between the processor and the bottom of the heat sink.
• Step 5.Attach the heat sink to the processor as directed by the processor vendor (for heat sinks supplied with the processor) or heat sink vendor (for aftermarket heat sinks). In some cases, you might need to attach mounting hardware to the motherboard before you can attach the heat sink.
• Step 6.If you are installing an active heat sink (a heat sink with a fan), plug the fan into the appropriate connector on the motherboard.

To insert an LGA processor, locate the notches on each side of the processor. These correspond with key tabs in the processor socket. Then follow these steps:

• Step 1.Make sure the load plate assembly is completely open. It has a plastic cover that can be removed at the end of Step 5.
• Step 2.Line up the notches in the processor with the key tabs in the processor socket. This assures that the processor's Pin 1 is properly aligned with the socket.
• Step 3.Lower the processor into place, making sure the metal heat spreader plate faces up and the gold pads face down. Do not drop the processor, as the lands in the processor socket could be damaged.
• Step 4.Push down the load plate and close the load plate assembly cam lever.
• Step 5.Lock the lever in place on the side of the socket. Remove the plastic cover and save it for future use.
• Step 6.Before attaching the heat sink or fan, determine if the heat sink has a thermal pad (also called a phase-change pad) or if you need to apply thermal compound to the processor core (refer to Figure 3-27). Remove the protective tape from the thermal pad or apply thermal compound as needed. Attach the heat sink or fan. You must use some type of thermal compound between the processor and the bottom of the heat sink.
• Step 7.Attach the heat sink to the processor as directed by the processor vendor (for heat sinks supplied with the processor) or heat sink vendor (for aftermarket heat sinks). In some cases, you might need to attach mounting hardware to the motherboard before you can attach the heat sink.
• Step 8.If you are installing an active heat sink (a heat sink with a fan), plug the fan into the appropriate connector on the motherboard.

Check the processor installation by booting the computer and by checking the speed of the processor in the BIOS and in Windows.

Slot-Type CPU (early Pentium III, early AMD Athlon, and Others)

You won't see many slot-type CPUs anymore, but if you need to install one on a motherboard, make sure the motherboard has a retention mechanism attached. If the motherboard doesn't have one, you will need to remove the motherboard from the case to attach a retention mechanism if it is not already attached.

To remove a slot-type CPU, follow these steps:

• Step 1.Push down on the retainers at each end of the CPU to release the CPU from the retention mechanism.
• Step 2.Disconnect the power lead to the CPU fan (if present).
• Step 3.Remove the CPU and fan/heat sink from the retention mechanism. The CPU slides straight up from the slot.

To attach a slot-type CPU, follow these steps

• Step 1.Attach the CPU retention mechanism to the motherboard. Leave the foam backing on the bottom of the motherboard while pushing the supports into place. Lift up the motherboard and secure the retention mechanism with the screws supplied.

  Some motherboards are shipped with the retention mechanism already installed, so this step might not apply to you. If the retention mechanism is folded against the motherboard, unfold it so the supports stand straight up.

• Step 2.Attach the fan and heat sink to the CPU if it is not already attached; some CPUs have a factory-attached heat sink/fan, whereas others require you to add it in the field.
• Step 3.Match the pinouts on the bottom of the CPU to the motherboard's slot; note that the slot has two sides of unequal length, making it easy to match the slot with the CPU.
• Step 4.Insert the CPU into the retention mechanism; push down until the retaining clips lock the CPU into place. Figure 3-29 shows the CPU in place.
  

  Figure 3-29 A Slot 1–based Celeron CPU after installation. The heat sink and fan are attached to the rear of the CPU.

• Step 5.Connect the power lead from the fan (if present) to the motherboard or drive power connector as directed.

Troubleshooting Processors

Keeping the processor running reliably is vital to correct system operation. This section focuses on some common problems and solutions.

System Runs Slower Than Rated Speed

A system running slower than its rated speed might do so because of processor throttling due to overheating, less than optimal settings in the Windows Power Options in Control Panel, or because of incorrect BIOS timing.

Overheating of the Processor or System

A system that overheats will stop operating, and with some older processors serious damage can result. Most processors today are fitted with active heat sinks that contain a fan. If the fan stops working, the process will overheat.

Fan Failure

Heat sink fans don't have to stop turning to fail; if they turn more slowly than they are specified to run, they can cause processor overheating.

Fan failures can be caused by dirt in the fan, worn-out bearings, or a bad connection to the motherboard or drive-cable power. In most cases, it's better to replace the heat sink fan than to try to clean it. If you must clean it, follow these steps:

• Step 1.Remove the heat sink from the CPU.
• Step 2.Place it on a surface covered with old newspapers or waste paper.
• Step 3.Blow it out with compressed air.

Before reattaching the heat sink, clean the old thermal material from the processor and the heat sink and reapply a small amount of thermal material to the top center of the processor cap. For specific thermal material installation recommendations for a particular processor, check the processor manufacturer's website.

If you opt for a replacement fan, improve reliability and life by specifying a ball-bearing fan rather than the typical (and cheap) sleeve-bearing units. Overheating can also be caused by a dirty power supply or case fan, or by missing slot covers. Clean or replace the fans, and replace the slot covers. Don't overlook cleaning out the inside of the case, because a dirty case interior will eventually clog other components due to the system's airflow.

Incorrect Heat Sink for Processor Type/Speed

If the processor overheats and the heat sink is properly attached and the fan is running, make sure the heat sink is designed for the processor type and speed in use. Heat sinks made for lower speed processors might not provide adequate cooling for faster processors, which often run at higher temperatures.

Use the heat sink provided by the processor vendor, or, if you are using a separately purchased heat sink, make sure the heat sink is designed for the processor type and speed in use.

The hardware monitor feature in the system BIOS can warn of overheating or fan failure. This is most effective if the motherboard or system vendor's monitoring software is also installed so you can be warned of problems while Windows is running.

Windows Power Options in Control Panel

Computers which are configured to use power settings other than High Performance will run more slowly at times to help save power and reduce heat. Systems using settings other than High Performance might also go into sleep mode more quickly, which can reduce system responsiveness. For maximum performance, use the High Performance power management setting (known as power scheme in some versions of Windows). Note that some older laptop computers use a special keystroke to activate or manage proprietary power management software.

Underclocked System

Some systems revert to a "fail-safe" setting in which the CPU frequency and/or clock multiplier default to low-speed settings if the system fails to boot properly or is shut off before starting. Check the system speed reported on the System properties sheet in Windows XP/Vista/7 or the CPU frequency/multiplier values in the BIOS. If these values are incorrect, set the CPU frequency and multiplier values according to the processor manufacturer's guidelines. See Chapter 4 for details.

If the system is configured to automatically detect the correct values for CPU frequency and clock multiplier but will not report the correct speed, the system might need a BIOS upgrade to properly support the processor, or you might be using a remarked processor (one that has had its original model number and technical information altered to make it appear as if it's a faster processor).

Processor Failure

If the processor is not locked into place, you will not be able to attach the heat sink. Never run the system if the processor is not properly installed, including heat sink installation.