CPUs Buying Guide

Buyers' Guide to CPUs Contents

• What is a CPU?
• Tracing an instruction
• L1/L2/L3 Cache
• Nanometres
• Clock cycle speed
• Front side bus (FSB)
• Overclocking
• The numbers game: Intel Vs AMD
• Sockets and slots
• Dual core processors
• 64-bit processors
• AMD
• Intel
• Mobile Processors
• Budget lines: Celeron Vs Sempron

------------------------------------------------------------------------------------

Computers are meant to reduce the complexity in our lives. However, if you're looking to upgrade your current PC or buy an entirely new system, you can quickly become confused by the terminology, especially when it comes to the CPU-- the heart of any computer system. This buying guide will take you through everything you need to know to make an informed CPU buying choice.

These days, it's not just a matter of deciding to buy a particular CPU at a particular price point. You've got to take into consideration the motherboard types that an individual CPU will work with, the additional capabilities that you're looking for in a processor, and the types of software you're likely to run on your machine. It's much better to make the right choice and get a PC that's ideal for your chosen applications than end up with something that sounds good on paper, but can't run the core tasks you need.
 

What is a CPU?

CPU stands for Central Processing Unit, and even after many years of refinements in PC technology, the CPU is still the core of your computer's operations. It controls the flow of data throughout your entire PC (hence the central part). The processing part is equally vital, as the role of the CPU is to manipulate -- or process -- the data that passes through your computer, reading it from storage devices, changing it as required, and then writing it out to storage or display devices.

At a physical level, a CPU is comprised of millions of microscopic transistors which are etched onto a layer of silicon via chemical and lithographic processes. Transistors themselves are exceptionally simple devices that store binary (on/off) values, and it's from these on/off states that more complicated processes can be performed. To give you an idea of the complexity of a modern transistor, a CPU such as AMD's Athlon 64 FX-55 is comprised of just under 106 million transistors. For a dual-core chip like Intel's Pentium Extreme Edition 840, that transistor count jumps to around 230 million.

In order to perform any useful data task, a CPU needs a few basic components. While the exact details of how each processor does this varies quite widely depending on the architecture of that particular model, the basics remain the same. A CPU needs to be able to get data in, transform it meaningfully based on the instructions it is given, and then send that data to a suitable repository. The repository might be a storage medium like a hard drive, a visual display like a monitor, or even the CPU itself for further calculations.

At the simplest level, the CPU needs only four elements to perform its data operations: instructions, an instruction pointer, some registers, and the arithmetic logic unit.

The instruction pointer tells the CPU where in memory the instruction it needs to run is located.

Registers are temporary storage locations on the CPU. A register holds data waiting to be processed by an instruction, or data that has already been processed (say, the result of adding two numbers together).

The arithmetic logic unit, or ALU, acts as the CPU's calculator, performing the maths and logic functions dictated by the instructions.

As well as these core elements, the CPU has some additional parts that help those core elements do their jobs:

• The instruction fetch grabs instructions from RAM or an area of memory located on the CPU;
• The instruction decoder takes the instruction from the fetch and translates it so the CPU understands. It then determines what steps are necessary to accomplish that instruction;
• The control unit manages and coordinates all the operations of the chip. It lets the ALU know when to calculate, tells the fetch when to grab a value, and tells the decoder when to translate the value into an instruction;

Tracing an instruction

Here's how a CPU does its thing: The instruction pointer directs the instruction fetch to a spot in memory that holds an instruction. The fetch grabs the instruction and hands it to the instruction decoder, which examines the instruction and determines what steps are necessary to accomplish that instruction. (An instruction can comprise numerous steps that need to be accomplished in a specific order.)

The ALU then performs the work the instruction calls for: it adds, subtracts, or otherwise manipulates the data. After the CPU interprets and performs an instruction, the control unit tells the fetch to grab the next instruction in memory.

This process continues -- instruction after instruction, at a dizzying pace -- to create the results you see on your monitor.

Both of the major CPU manufacturers, AMD and Intel, have introduced proprietary technologies designed to optimise efficiency of CPUs at particular tasks and with particular software. It's not strictly necessary to understand where each and every bit that passes through the CPU goes in order to make an informed CPU buying decision, but it is important to understand the basics of each technological innovation in order to properly evaluate CPU performance.

To improve calculation performance, chip makers placed another arithmetic logic unit in the CPU. Theoretically, this meant twice as much processing could be done at once. Using multiple ALUs is like having two people lay a kitchen floor instead of one: the job gets finished more quickly.

In addition to multiple ALUs, Intel integrated the floating point unit (FPU) into the CPU. The FPU handles extremely large and extremely small numbers (those with many decimal places). While the FPU handles those types of calculations, the ALU is free to do something else at the same time, further enhancing performance.

AMD and Intel also sped up instruction processing by pipelining the instructions, or running them nearly parallel to each other. The execution of an instruction requires many separate steps -- fetching and decoding an instruction, for example. Originally the CPU had to complete an entire instruction before beginning the next one. Now, discrete circuits handle the separate steps. Once an instruction has moved from the first step to the second step, the transistors that performed the first step are free to start working on the next instruction, therefore speeding up the process. It's like climbing stairs: as soon as your feet leave a stair, the person behind you can use it.

Other additions that enhance performance include branch prediction (guessing which jump the program is likely to take); speculative execution, executing the predicted branch ahead of time; and out-of-order completion, the ability to complete a series of instructions out of a program's normal order.

L1/L2/L3 Cache

Whenever your CPU has to fetch data, a bottleneck can arise if the data is in a relatively distant location -- say, for example, sitting in external memory that has to be first accessed, scanned and then read. From this problem arose the practice of adding memory directly to the processor itself, creating a small storage area for commonly used data. If the CPU doesn't have to travel outside of itself to get data, it can get on with the business of number crunching much more effectively. This memory is generally referred to as a cache.

Cache within a CPU is split up into levels, which relate normally both to the size of the cache and its access speed. Level 1 (L1) cache is normally small but fast memory, which acts as the first place for the CPU when it goes to look for data. If a desired bit of data is found in the L1 cache, then you'll see a quick result; otherwise the CPU will seek out the Level 2 (L2) cache. L2 caches are normally larger in size than L1 caches, and there's a solid advantage to this design theory. With a small L1 cache, checking for data is relatively quick; it's only for data outside the L1 cache that you need to look further. A larger L2 cache gives you a sizeable reservoir for data in a slower to scan environment, but because it's not the first place the CPU will check, you get a good trade-off between latency (the time taken to perform an operation) and the success rate of searching the cache itself.

There's also a third level of cache. While historically L3 cache has been the province of server CPUs, it's starting to become a feature of some consumer desktop processors -- Intel's P4 Extreme Edition processors, for example, feature L3 cache. As you can expect L3 cache to be hit least, it can afford to be larger -- in the case of the P4 Extreme Edition it has 2MB of cache memory, four times as much as it has L2 cache. As the entire cache is situated on the CPU die itself, it's still considerably quicker for the CPU to search the cache levels sequentially than to refer to external storage such as physical memory or hard drives.

Nanometres

A common measure of the complexity of a CPU is how many individual transistors fit into the CPU itself. Getting billions of transistors into a package that's affordable and works properly is a significant challenge, and manufacturers spend billions of dollars on improving the process of laying silicon transistors down. In simple terms, the shorter the spacing in which you can successfully lay down transistors, the more you can pack into a smaller space, and the more chips you can cut out of a single silicon wafer. This has the effect of increasing capacity while lowering overall prices.

Silicon chips are etched using ultraviolet light, and the wavelength of the light -- and effective size of the transistors -- is measured in nanometres, or a billionth of a metre. Current high-end CPUs are manufactured on a 90 nanometre process, although many chips are still on the market that use a 130nm process. You may also process measurement referred to in microns -- a micron is a millionth of a metre where a nanometre is a billionth, so a 90nm process is also a .09 micron one.

On the subject of processes, manufacturers will often use the same basic architecture and micron processes across an entire family of CPU types, and this is commonly referred to as using the same "core". Vendors typically assign these cores with codenames during development, and while they don't tend to go to market with these codenames, you'll often see them used in discussions of processors from both vendors. Intel, for example, has a large quantity of processors built on its Prescott core, all with a similar architectural configuration. Prescott cores are built on a 90nm process, feature 31-stage pipelines and SSE3 instruction sets; even the company's dual-core processors are built on a variant of the Prescott core.

Clock Cycle Speed

In order to correctly synchronise the actions of a CPU, a metronome-like clock sends deliberately timed pulses throughout the CPU. This is referred to as the clock cycle, and is typically measured in gigahertz. The higher the GHz rating, the faster the clock speed of the CPU. In the past, clock speed was a good raw indicator of CPU performance, but it's increasingly just becoming part of the picture, rather than the picture itself.

For the consumer, things were much simpler five to ten years ago, when the main metric for CPU performance was the clock cycle speed. With the advent of extension technologies and performance upgrades such as Hyperthreading, 64-bit processors that can run 32-bit applications and the move to dedicated graphics processing units, the clock speed now means less to the overall performance of the CPU than it used to. Put simply, CPUs of today aren't just being measured by the number of operation cycles they can perform in a given second -- the classic CPU speed rating -- but by what operations they can perform within those cycles as well.

Both Intel and AMD have invested heavily in maximising the output of each and every clock cycle, and in doing so, they've made the actual clock speed less important. This is especially true in the mobile market, where battery power is a much more pressing concern than just about any other factor.

Front side bus (FSB)

The Front Side Bus (FSB) is the primary connection point between the CPU and the primary chipset on the motherboard. Motherboard specifications will list a FSB frequency, measured in MHz (like CPU clock cycles). The combination of FSB speed and a processor's internal multiplier determines the final speed of the CPU itself. As with processors themselves, FSB speeds have grown astronomically in recent times, with modern motherboards supporting FSB speeds of anywhere between 533MHz and 1066MHz.

Overclocking

Overclocking is a popular pastime for PC enthusiasts, and involves squeezing additional performance out of a CPU by altering the frequency at which it is run. Clock speed is a combination of the Front Side Bus speed of the motherboard and the clock multiplier of the CPU. By altering either the FSB or clock multiplier, you can, in theory, up the overall frequency of a given processor. To do this, you need overclocking to be supported on both the processor and the motherboard.

The most common way to stop users overclocking is to lock the clock multiplier of a given CPU. Manufacturers have played around with how difficult it is to alter clock multipliers on their CPUs, and don't officially condone the idea, although it's worth noting that the current high-end Pentium 4 Extreme Edition and Athlon 64 FX processors come with no lock on the clock multiplier.

In CPU manufacturing, manufacturers first test each CPU at the theoretical maximum speed it's capable of. If it passes, it's sold as a premium-priced, high speed chip. If it fails, it'll be tested at lower speed, and sometimes with specific functional operations disabled, until it's either running in a stable form -- and sold as a lower speed chip -- or junked altogether. From the manufacturer's point of view, the fewer chips they have to junk, the more profitable a given run of CPUs will be.

Because of this manufacturing economy, there is scope for overclocking chips, but be careful. No CPU manufacturer offers warranty support for overclocked chips, as there's a distinct danger of overheating a chip by running it faster than it's been safely clocked at. After all, if it ran well at that higher speed it would be sold to market as a higher speed chip. Overclocking enthusiasts try to get around the increased heat output by changing the cooling arrangements within a PC to dissipate the extra heat. This can range from as simple as putting on a slightly larger fan than the vendor supplied CPU fan all the way up to water-cooled semi-frozen PCs running at much higher speeds than would seem sensible.

The numbers game: Intel Vs AMD

In marketing their respective CPUs, both Intel and AMD play games that can be confusing to the first-time CPU buyer. Both use numbering schemes which are built around the same central premise, namely that describing a processor in terms only of its clock speed is misleading in that it doesn't paint a complete picture of the processor's performance when you take into consideration its other features.

AMD was the first to essentially drop processor speeds from its CPU descriptors. It now gives all of its processors model numbers that are intended to indicate relative software performance within the AMD family of processors. So an AMD Sempron 3300+, for example, isn't a processor with a 3.3GHz clock speed -- it's actually running at 2GHz, but in AMD's ordering of things, that's where it sits.

The exception to this rule is the Athlon 64 FX series, which simply have two digit oddly sequenced numbers to differentiate them. The first Athlon 64 FX chip was the FX-51, with the next being the FX-53, then the FX-55. According to AMD the lack of a relative model number within the FX series is because the higher-end customers for Athlon 64 FX CPUs are more likely to benchmark the processors themselves in order to sort out any performance confusion issues.

Intel's take on processor numbering removes the core clock speed from the equation altogether. Instead, what you get is a three digit number that designates the intended purpose of the processor, be it mobile, desktop, budget or high performance, along with some indication of where it stands relative to other processors within the same processor family. Number differences don't always represent speed ratings, but can indicate the presence or absence of other features such as expanded cache size or Hyperthreading.

As an example, the Intel Celeron D 345J and Intel Celeron D 345 differ only in their socket form factor, for example, while the Pentium 4 530 and Pentium 4 540 differ in clock speed, with speeds of 3.0GHz and 3.2GHz respectively.

Sockets and Slots

The packaging that CPUs come in has evolved over time, as have approaches to how the CPUs 'slot' into the motherboard. The socket type on your motherboard of choice pretty much determines the processor that can run on it -- you can't fit an Athlon 64 FX chip on a motherboard designed for a Pentium 4 Extreme Edition Chip. Chipsets and processors are inextricably linked, and the practical upshot of this is that if you're in the market for a new CPU, unless it shares a very recent history with existing processor models, you'll probably be in the market for a new motherboard and supporting chipset as well.

You may also need to consider purchasing new memory, as that's also a factor that changes over time. With the move towards high speed DDR2 memory (and beyond), you might not be able to take the 512MB of memory from your old system and install it in your new one; if you're lucky all it'll do is slow down system performance, but it's more likely that it'll be totally incompatible.

Intel sets the pace for its processor chipsets, and does manufacture motherboards as well, although a number of third party manufacturers also licence the chipsets to use on motherboards. The onboard chipset also works in combination with the CPU to run certain CPU-specific tasks, so if you try to match up an Intel processor that supports hyperthreading to a chipset that's never heard of it, the results won't be pretty.

Intel's current desktop processor lines are transitioning over to the pinless LGA 775 package, although some older lines still use Socket 478-based motherboards. On the AMD side of the fence there's a move towards Socket 939 packaging, although there are still plenty of models that use the older 462-pin Socket A or Socket 754 configurations. Early Athlon 64 FX chips used a Socket 940 chipset that required more expensive registered memory. Both Intel and AMD provide Web resources to help you identify the correct matching chipset for each of their processors, and any decent vendor should be able to do likewise. (Intel: http://indigo.intel.com/mbsg/ AMD: http://www.amd.com/us-en/Processors/TechnicalResources/0,,30_182_869,00.html).

Dual core processors

CPU makers have struggled in recent years to keep upping the productivity of CPUs without adversely affecting either their power requirements or heat output. Put simply, the more power a CPU uses the more heat it's likely to produce, which makes it costlier to run -- and requires ever larger power supplies -- as well as having a limiting effect on the lifespan of the processor itself, as greater heat will burn out a processor much more quickly.

Both AMD and Intel are combating the problem with dual core CPUs. These are CPUs that contain not one, but two core processors, with dedicated cache memory for each processor. Sticking two processors into a PC isn't entirely new, although it's largely been the domain of more enterprise-based computing to date. However, if you're running software that can take advantage of multiple processors, you should see improved performance with a dual-core CPU.

Likewise, if you're a fan of running several heavy-duty processor applications simultaneously -- say, running a virus scan while simultaneously encoding video and perhaps doing some light Web surfing -- you should see benefits from a dual-core approach. Like 64-bit processors, and indeed many of the additional technologies such as hyperthreading and multimedia extensions, however, dual core processors will really hit their stride and be a worthwhile buy once more applications are specifically written with multiple cores in mind, as that's when software and hardware will be able to work together most effectively.

Intel's dual core CPUs were codenamed "Smithfield", and the first of these is now marketed as the Pentium Processor Extreme Edition 840. The 840 features two processor cores (3.2GHz or 3.46GHz) with 2MB of L2 cache -- 1MB per core. Both cores feature Intel's hyperthreading technology, as well as Extended Memory 64 Technology. Successors to the first generation of dual core CPU technology are expected late in 2006, with the codenames of Conroe for desktop processors and Merom for the notebook models.

AMD's initial foray into dual core processors came with its enterprise-level processor offerings, but the company also has offerings in the desktop space, known as the AMD Athlon 64 X2 series. X2 CPUs are dual-core 64-bit desktop processors. According to AMD, the X2 family currently ships with ratings numbers of 4200+, 4400+, 4600+ and 4800+ -- although it should be born in mind that these are AMD's performance ratings, not clock cycle ratings. AMD's pitching the Athlon 64 X2 at the digital media market, while holding to the mantra that the Athlon 64 FX is the best performing processor in the world for single-threaded games.
 

64-bit processors - the next generation?

There's a limitation on existing 32-bit processors (which are found in most of today's PCs). A 32-bit processor can, at most, address 4 gigabytes of data. For many purposes, that's fine, but once you step into more demanding applications, be they gaming, video processing or even working with large databases, it becomes a limiting factor that slows system performance. As main memory in desktop PCs creeps higher and higher, we're getting closer and closer to that 4GB barrier.

A 64-bit processor isn't hampered by those problems, as its increased addressing space of 64-bits versus the current 32-bits allows it to address up to a potential 16 exabytes of memory -- that's 16 billion gigabytes, quite a jump from the previous generation. Having a 64-bit data path also doubles the amount of data that can be fetched at any one time, increasing data throughput within the processor itself.

The earliest 64-bit processors were server-based affairs designed explicitly for 64-bit operating systems, but the current crop of 64-bit processors, first seen in the consumer space with AMD's Athlon 64 line of processors. Consumer-level 64-bit processors like the Athlon 64 can handle both current 32-bit operating systems and software, as well as true 64-bit software designed more specifically for a 64-bit processor. So if you purchase a 64-bit processor, you're still going to get solid performance from existing applications, as well as a touch of future-proofing when 64-bit OSes and applications become more solidly mainstream. For the moment, the best work you can get out of a 64-bit processor is in the gaming arena, as the processor-intensive nature of gaming -- combined with several gaming engines which have been or are being optimised for 64-bit computing -- give gamers a solid edge.

AMD

AMD was the first processor manufacturer to offer a 64-bit CPU with the launch of the Athlon 64 processor. However, the vendor's quick to point out that it's a technology that's suited to today's applications as well as any future 64-bit-only applications. AMD hasn't abandoned the 32-bit CPU market entirely, as its Sempron line continues to offer pure 32-bit processors for the budget market.

AMD's approach to 64-bit processing on the desktop has been a largely inclusive one, as both its Athlon FX and Athlon 64 processors support multiple operating modes that will run 32-bit operating systems and applications, true 64-bit operating systems and applications and even 32-bit applications on 64-bit operating systems. It's a sign of how well AMD's taken this market that Intel's competing 64-bit x86 technology, known as Extended Memory 64 Technology more or less mirrors the AMD model.

There's an element of future-proofing in buying a 64-bit processor, as you're readying yourself for a full commercial release of a widespread 64-bit Windows OS -- at the time of writing a trial version of 64-bit Windows XP is all that's available in the Windows world, although several Linux variants support true 64-bit computing, and it's expected that Microsoft's upcoming Vista (previously codenamed Longhorn) OS will ship in a native 64-bit variant.
 

Athlon 64 vs Athlon 64 FX

AMD currently sells two differing types of 64-bit processor, aimed at different market segments and with somewhat different internal architecture and packaging. When initially launched, the Athlon 64 FX line was the premium brand version of the Athlon 64, with a dual channel memory controller built into the chip, double the L2 cache memory and a 939 pin socket. Newer Athlon 64s are built to an identical standard though, using Socket 939, with similar levels of cache and integrated dual channel memory controllers. AMD's keeping the FX line as premium chips primarily through higher clock speeds, and by keeping the clock multiplier on the FX line unlocked, which makes them a popular choice with overclockers.

AMD's 64-bit processors shine in the gaming arena, partly because of elements such as the CPU's memory controller being integrated onto the CPU itself, reducing memory latency, and also in single threaded games that can take advantage of the additional memory registers within the Athlon 64 and Athlon 64FX. A number of popular game engines have also been significantly rewritten to take advantage of a 64-bit processor.
 

What is Cool N' Quiet?

Cool N' Quiet is AMD's proprietary CPU throttling technology that dynamically changes the CPU multiplier and voltage depending on the workload you're placing on it. This means that if you're merely pottering around the web, it'll scale back the processor speed and power consumption, but if you then launch into some heavy duty video encoding, it'll ramp back up. AMD claims that it's capable of reducing power consumption by up to 40 watts. From an end-user standpoint, it's very similar to Intel's mobile Speedstep technology, although AMD's chosen to implement it on desktop processors. AMD's mobile equivalent for Cool N' Quiet, which more closely mirrors Intel's Speedstep, is called AMD PowerNow!.
 

What is Enhanced Virus Protection?

Enhanced Virus Protection is a security technology implemented in AMD's Athlon 64/FX lines, and some of its Sempron lines. It works in concert with Windows XP Service Pack 2 to stop certain applications from running into a situation where by overflowing memory buffers, system security is compromised.
 

What is Hypertransport?

Hypertransport is a technology originally known as Lightning Data Transport (LDT), jointly developed by AMD and other partner companies. It's an extremely high-speed system bus that's currently used by AMD and other HyperTransport Consortium members, such as Apple. Within AMD's architectures, hypertransport provides a 16-bit internal link running at up to 2000MHz, capable of shuffling up to 14.4GB/sec between the processor and the main system.

Intel

Pentium 4 vs Pentium 4 Extreme Edition

The Pentium 4 has been a solidly performing chip for Intel, but it's not the company's boutique chip for power users. That title falls to the Pentium 4 Extreme Edition. The first Extreme edition CPUs were based upon a modified version of the Gallatin core used in Intel's Xeon processor line, although newer Extreme Edition chips are being built around the Prescott core used in existing Pentium 4 CPUs. Over time, it's likely that many of the features found in existing Extreme Edition processors will become part of the regular Pentium 4 lineup.

The advantages of the Prescott-based Extreme Edition chips come from a larger quantity of cache memory, with 2MB of L3 cache on most of the Extreme Edition line. They all have Hyperthreading enabled, and the higher clock speed Extreme Edition processors feature an enhanced FSB speed of 1066MHz. While it's shifted away from the server-like Gallatin cores, Intel's maintaining the Extreme Edition's premiere place by using it as the launch pad for a number of emerging CPU technologies. Newer Extreme Edition chips feature Intel's EM64T 64-Bit application support, and SSE-3 instruction sets, as well as ever-increasing clock speeds and the company's first foray into desktop dual core chips, the Pentium Processor Extreme Edition 840, which features two cores (3.2GHz or 3.46GHz) on an 800MHz FSB.
 

What is hyperthreading?

Hyperthreading is a technology that's becoming increasingly common across Intel's entire processor family, although some of its offerings in the lower budget and mobile spaces still omit the technology. It's a technology that allows a single CPU to process multiple threads simultaneously, giving the processor a theoretical application speed boost. It shouldn't be confused with the newer, dual-core processors that actually do feature multiple processors in the same CPU core, and can thus perform true multi-threaded computing. Hyperthreading relies on having software that can intelligently use the Hyperthreading resources, and while plenty of applications -- including Windows XP itself -- are multithreaded applications, in most real-world testing hyperthreading hasn't really been seen to generate much extra oomph.
 

What is NetBurst MicroArchitecture?

NetBurst MicroArchitecture is Intel's catch-all name for a group of technologies found in its Pentium 4 processor lines. These technologies include an Execution Trace Cache, an out-of-order core, and a Rapid Execution Engine. More specifics on how NetBurst MicroArchitecture can be found at Intel's Web site,but in short, they're all technologies designed to lower the latency of a given CPU and thereby increase data efficiency.
 

What is SSE3?

What is Speedstep?

Speedstep is Intel's power-saving and heat-reduction technology. Originally only available on mobile processors, where energy consumption is a key concern, it's also implemented on a number of Intel's desktop processors. On the desktop it's got a dual role to play, reducing power needs as well as lowering the heat output of some of Intel's hotter processors. Speedstep has undergone several revisions in its lifetime, but the basic operation from an end-user perspective has remained unchanged, as it throttles back processor frequency in response to user needs, so that when your system is idling, it's also not burning itself out.

Mobile Processors

Mobile processors differ from their desktop processors in a number of significant areas. The most telling is power consumption, as it's safe to say that a desktop PC should never run out of power, whereas any notebook user will have experienced a flat battery more than once in their notebook usage. It's a tougher market to describe in terms of shopping for a new or replacement CPU, however, as no manufacturer currently sells a notebook with a user-swappable CPU. If you've got a lot of guts, determination and no care for your warranty it is possible to change CPUs, but given the way that battery technologies degrade over time and especially the dropping prices of notebooks, it's not a viable option for most consumers. It does, however, pay to know what's going into each vendors mobile CPU families.
 

AMD

AMD spreads its mobile family quite wide, incorporating lower-cost Mobile Sempron processors, higher end Mobile Athlon 64 processors and Turion 64 mobile processors. Its Sempron and Mobile Athlon 64 models are pitched towards the budget and desktop buyers respectively, while the Turion 64 lines are AMD's equivalent to Intel's Centrino platform, matching an AMD processor with a set list of components , although it's not as strict a list as with the Centrino platform. Turion 64 processors use a challenging naming convention, wherein chips are labelled with a two letter and two number nomenclature. The two numbers indicate relative performance, while the letters indicate relative power consumption. Higher numbers and letters closer to the end of the alphabet indicate better performance and lower battery consumption.
 

Intel

What is Centrino?

Centrino isn't just a processor; it's Intel's integrated approach to mobile notebook technology, encompassing a specific chipset, wireless network connectivity and low-power CPUs. In order for a vendor notebook to qualify for Centrino branding, all three components must be present within the notebook.

The current Centrino package, codenamed "Sonoma", comprises Intel's mobile 915 Express chipset with either integrated or separate graphics chips running with either a 400Mhz or 533MHz FSB. Sonoma-based Centrino notebooks feature 90nm Pentium-M CPUs based around Intel's Dothan core with 2MB of L2 cache. Sonoma CPU speeds vary from the 1.2GHz Pentium-M 753 Ultra Low Voltage chip with a 400MHz FSB to the 2.26GHz Pentium M 780 with a 533MHz FSB.

It's not the CPU itself that makes Centrino a wireless solution - that's down to wireless-enabled chips on the 915 Express chipset, along with a wireless antenna, normally built into the back of the screen of most notebooks. The part that the CPU plays here is to run as well as possible with as little an impact on power consumption as possible. That's because wireless networking is power-thirsty stuff, so by reducing the power needs of the CPU, you gain precious extra working time with your notebook.

It's worth noting that while Intel's heavily invested in selling Centrino as the wireless solution, it's merely using wireless standards to enable Centrino-badged notebooks to communicate, via a wireless chip and antenna built into every Centrino notebook. Any notebook with a wireless card can manage the same communication feats as a Centrino notebook, albeit without the potential power and cost-savings of the Centrino package.

While notebooks don't really feature upgradeable processors -- unless you don't like warranty support very much -- the platform approach of Centrino, which has worked remarkably well for Intel is going to be duplicated on the company's consumer desktop lines, emphasising the role of combining specific chipsets with specific processors around a core computing task.

Budget Lines: Celeron Vs Sempron

If your computing needs don't stretch very far, or your wallet is under tight constraints, you may want to consider a more budget-friendly CPU. Both AMD and Intel offer budget processor lines. Intel's longstanding brand for its budget processors has been Celeron, although the internal architecture of Celerons has changed dramatically since the first Celerons hit the market. AMD has dumped one budget line, Duron, experimented with just selling cheap Athlon XP processors and now sells its budget processors under the Sempron brand.

Desktop Sempron processors come in Socket A (older) and Socket 754 (newer) variants, meaning that it's possible to upgrade processors without necessarily changing motherboards, as long as your Sempron-compatible motherboard will support the feature sets of Athlon XP and older Athlon 64 processors respectively. Intel's current line of desktop Celeron D processors suit Socket478 or LGA775 motherboards, giving a potential upgrade path to Pentium 4 and Pentium 4 Extreme Edition lines.

While budget processors perform well enough for those with no need for really CPU intensive tasks such as video rendering or gaming, one thing to be careful of is to check prices on the bottom of the rung processors regularly when you go to buy a budget CPU. The market for CPUs is exceptionally fierce, and pricing cuts on processor lines is a regular occurrence, and this can often have the effect of stripping away much of the price difference between a budget and regular processor, especially if your vendor of choice is selling off older stock.

Your department store source to buy cheap used and new items online