Tech-Hounds.com

Because gamers play games, not benchmarks




Building Your Own PC - Part 1

Since building your own PC is a broad subject, we've separated it into three parts. This first part can be considered as an introduction to various parts that make up every PC. Since there's a lot of them, this part is quite long. Most information contain in this part is very basic, so most of you already familiar with PC peripherals and internals can skip this part and go on to Part 2.

A Little Background

Before we begin, let's take a look inside your PC or more accurately, inside your CPU. After all, if we're planning to muck around inside it, we should know at least what the components / parts are and what they look like. Let's begin with the basics.

Processors

The brain of the PC, processors have come a long way since the early 80's. First generation PCs used a 4 MHz processor and just 640 KBs of RAM (Random Access Memory). Now, we're reaching speeds up to 4 GHz and memory up to 4 GBs. That's a thousand time faster and nearly eight thousand times larger.

Today, there are two major manufacturers of PC processors: AMD and Intel. There are other manufacturers, but they (VIA and Transmeta) only cover a small percentage of the PC market. This includes both desktop and notebook PCs to servers and workstations. For the sake of brevity, we'll limit our discussion to desktop processors by AMD and Intel.

Sockets

While basically they handle the same task, AMD and Intel use different designs and shapes for their processors. Thus their processor require different sockets. So, you couldn't put an AMD processor in a motherboard built for Intel processors and vice versa. Currently, Intel uses two sockets for their desktop processors: socket 478 and 775, while AMD uses socket 462 (otherwise known as socket A), socket 754 and socket 939. These guys sure like their sockets and numbers, huh? These numbers are actually the number of connectors (in the form of pins or pads) the processors have. Again, you could only use a socket 754 processor motherboard equipped with a socket 754 and nowhere else. That goes for every processor and their respective socket. Why the reasons for many sockets or connectors? Well, as processor becomes faster and faster, they need more connectors - either for electricity or data. Thankfully, there will be no changes in sockets for the next two or three years (at least we hope so).

Socket 462 / A from AMD and socket 478 from Intel are the oldest of the bunch. Right now, these sockets are being replaced by newer ones, such as socket 754 / 939 for AMD processors and socket 775 for Intel processors. So if you're concerned about future upgrades and don't want to change your motherboard, it's best to pick processors with the newer sockets, either socket 939 for AMD and 775 for Intel. Unfortunately, these processors are not that much faster than the old ones. So for some of us, it makes much more sense choosing the old ones for now since they're cheaper and already fast enough for most applications. You could always upgrade both the processor and motherboard later - usually with more reasonable prices than what they're currently charging now. This is what makes choosing a processor for your PC so difficult. In times like this, I'd only recommend the newer processors (socket 939 and 775) if you're more concerned with upgrades - basically if you use your PC for games. Those using their PC for office work and multimedia should go for the least expensive option - socket 462 / A or 478. Upgrading your processor won't do much anyway.

Performance: Not Just Pure Speed

If you look at the processors available today regardless of sockets, you'll see that they also come on different speeds. These speeds range from 2 to 4 GHz (or equivalent). Contrary to what most people believe, speed is not the only factor indicating processor's performance. If it is, a 4 GHz processor should be twice as fast as a 2 GHz processor. This is not the case. When we talk about performance, we're not talking about the speed of the processor, but the speed at which the processor process data. Remember that data also have to be sent back and forth between the processor and other components. With faster data transfer, the processor will more likely reach its optimal performance. This concept is what we call bandwidth. There are also internal differences between processors. Some processors process more data per MHz at the cost of speed, while some works at a higher speed (MHz) at the cost of processing power (less data processed).

Let's take a deeper look inside a processor and see what determine its performance. We will use an analogy: a car's engine. Some engines use high RPM with four cylinders to deliver power, while some use low RPM with six or even eight cylinders. Fuel economy aside, when working at the same RPM, the six / eight cylinder engine will deliver more power than the four cylinder engine. Processor work in a similar manner. AMD's engineers use a design for their Athlon processor with three, big pipelines (think of this as the cylinders for your car's engine), while Intel uses two big pipelines and one small one. This doesn't mean Intel's engineers are dumb. On the contrary they're very smart and they explicitly choose to do so. Why? The Pentium 4 was made to reach high clock speeds. Using less big pipelines means they could free up the space to make them longer and longer pipelines means higher speed - MHz. This is an overly simplistic example, but not that far from the truth. This example show how the two manufacturers use different approach and design for their processor. AMD is more concerned with doing more per cycle, while Intel is going for more cycle at the expense of fewer instructions per cycle. Congratulations, you've just learn one of the most important aspects of processor performance: IPC (Instructions per Cycle), which literally means how much instructions can be done per cycle (1 Hz).

Now look at the benchmarks, an AMD Athlon 64 processor working at 1,8 GHz could process data as fast or faster than an Intel Pentium 4 processor working at 3 GHz. That particular AMD processor is said to be 'equivalent' to that Intel processor. Remember this fact when you're contemplating whether to buy an AMD processor or an Intel processor. As of 2005, both AMD and Intel use a name scheme for their processors that omits the processor's speed. AMD have been using this scheme for awhile (since 2002 with their Athlon XP processor).

Bandwidth: Bus Speed and Cache

Let's talk about the other part of the performance equation: bandwidth. The processor transfers data back and forth to other components such as the memory (for storing temporary data) and the hard drive (for reading / saving permanent data - like your files). These components don't work at the same speed as the processor. For example, a PC3200 DDR-SDRAM memory module works at 200 MHz. An Athlon 64 motherboard also works at 200 MHz (remember a hard drive is connected to the motherboard which handles all data transfers to your processor). So, the processor is actually working faster than the other components (9 times 200 MHz for an Athlon 64 running at 1,8 GHz). This speed (200 MHz) is what we call the bus speed or Front Side Bus (FSB) speed to some.

One of the reasons AMD and Intel uses different sockets is related to how the bus (or connection) for each processor works. AMD uses a double data rate bus (DDR bus for short, don't confuse this with DDR-SDRAM) which means that a bus working at 200 MHz effectively transfers data at double the rate (400 MHz). So, when describing an AMD bus, people use words like '200 MHz FSB (effective 400 MHz)' or something similar. When speaking of bus speed on AMD processors, a processor using 100 MHz bus is effectively transferring data at 200 MHz and so on (266 MHz is 133 MHz, 400 MHz is 200 MHz). Intel chooses a slightly different route. In Intel's processors, a bus working at 200 MHz transfers data four times as much with a technique they call QuadPump. Hence the term '800 MHz QuadPumped bus'. This is one of the reasons why Intel processors are faster when processing multimedia data than AMD processors. At the same bus speed, the Intel processor theoretically transfers more data. When speaking of bus speed on Intel processors, a processor using 400 MHz bus is effectively using bus speed of 100 MHz and so on (533 MHz is 133 MHz, 800 MHz is 200 MHz, 1066 MHz is 266 MHz).

Bus speed and cache sizes of different processors

Bus Speed
 L2 Cache Size
Pentium 4
 100, 133, 200, 266 MHz
 512 KB, 1 MB, 2 MB
Celeron
 100, 133, 200 MHz 256 KB
Athlon 64
 200 MHz 512 KB, 1 MB
Athlon XP 100, 133, 166, 200 MHz 256 KB, 512 KB
Sempron
 100, 133, 166 MHz 192 KB
Duron 100, 133, 166 MHz 192 KB

Both AMD and Intel makes processor not just using different speeds, but also different bus speeds (although they're using the same socket). For example, Intel's Celeron uses a bus speed of 100 or 133 MHz while their Pentium 4 uses 133 or 200 MHz. AMD does the same thing with their Athlon XPs and Duron / Sempron processors. Remember that a higher bus speed means higher bandwidth and higher bandwidth means the processor is more likely to reach its optimal or maximum performance. The other difference between these processors is cache size. Cache is small amount of memory inside your processor that acts as a buffer for your PC's main memory (RAM). Cache really helps out the processor in the bandwidth department. Since cache runs at the same speed as the processor and located inside of it, your processor can get the data much faster than if the data is stored in RAM. The larger the cache, the higher the possibility the processor can find the data in the cache. Intel's Celeron processors uses a smaller cache (256 KB) compared to their Pentium 4s (512 or 1024 KB). AMD's Duron and Sempron processor also uses a smaller cache of 128 KB, while their Athlon XP and Athlon 64 uses 512 KB or 1024 KB. Not surprisingly, these lesser bandwidth processors (with lower bus speed and smaller cache) are priced lower than the higher bandwidth ones.

Bandwidth: AMD's Unique Solution

Beginning with Athlon 64, AMD choose a different approach to bandwidth. Instead of using the memory controller outside the processor (in the north bridge chipset connected through the bus), AMD integrates the memory controller into processor. Just like cache, the memory controller runs at the same speed so overall data transfer are quicker.

There are several interesting points that make this move a very 'smart' solution. Since the main memory now has a direct line to the processor, the processor don't have to use the bus to transfer data when accessing memory. So in effect, AMD's Athlon 64 didn't just transfer data faster from memory, but also from other components (by using the bandwidth that's previously used to access data in RAM). They also solve most memory compatibility and performance problems. Data access is still faster, even with slower memory. If you look at the various Athlon 64 benchmarks, you'll notice that the performance stays relatively the same regardless of the motherboard and chipset. The downside to this solution is that to support newer types of memory, AMD will have to revise the memory controller. This means you have to replace your still working processor. But don't worry too much, these processors still delivers very high performance with the current memory standard - Double Data Rate (DDR) SDRAM.

The 64 bit Evolution

If you haven't been following modern processors developments, you might be wondering what the '64' in Athlon 64 stands for. Quite simply, it indicates that Athlon 64 is a 64 bit processor. In fact, Athlon 64 is the first 64 bit desktop processor for the PC that's fully compatible with current 32 bit operating systems and applications. This may not seem much, but it really is.

What's with the bit and numbers? Well, a lot actually. In case you don't know, processors (and all PCs for that matter) handle data as combinations of 1s and 0s. Each one of these 1s and 0s are called bit(s). Using more bit, we can store more data. For example, 8 of these bits (together they formed a byte) can hold any number from 0 to 255. If we add more bits to this, we can use it to store even greater numbers. So a 32 bit data can hold any number from 0 to 4.228.250.625 and a 64 bit data can hold any number from 0 to 17.878.103.347.812.890.625. Well, that's quite a lot. A 64 bit processor is capable of processing a lot more range of data than a 32 bit processor. Remember that we're talking about the range of data and not the data itself, a 64 bit processor will not be twice as fast as a 32 bit processor.

Well, if that's good than why didn't anybody do it sooner? Well, they actually did, but earlier 64 bit processor uses a totally different architecture from PC processors. PC processors are CISC processors, while 64 bit processors are RISC processors. They handle different types of data but more importantly, they use different instructions that are not compatible with each other. The only way to run 32 bit PC processor's applications and operating systems on these 64 bit processors is through a process called emulation. Emulation is slow, tricky and problematic. Slow because the 64 bit processor must translate all the 32 bit instructions and reformat the 32 bit data into a format that it can understand and process. Tricky and problematic because a 32 bit application expects to see a 32 bit processor so the 64 bit processor must 'fool' that application. It will also still be 32 bit application and can not take full advantage of the 64 bit processor's features. Simply put, we never want to emulate anything and emulation should always be a last resort.

AMD's engineer figured out a way to make 64 bit processors that still run current 32 bit PC applications and operating systems without emulation. Quite simply, they made their Athlon 64 processors share the same features with their 32 bit processors, but can switch to 64 bit mode if the software tells it to do so. To really take full advantage of 64 bit computing on Athlon 64, you have to use a 64 bit operating system, drivers and application, but fortunately that's not far down the road. A 64 bit version of Windows is scheduled to launch in April 2005 and 64 bit versions of your favorite applications are soon to follow. The 64 bit drivers are being finished and should offer the same (or faster) performance as the 32 bit ones. Since you can run your 32 bit applications just as well (and more importantly, just as fast), there's no reason not to go 64 bit.

As of 2004, Intel have also supports this 64 bit evolution. Their new processors, while using a slightly different instruction set, will still be compatible with 64 bit version of Windows - and that means it's compatible with AMD's Athlon 64 as well, software wise.

Remember these facts when choosing a processor:

Memory

Just like any other components in your PC, memory modules come in a variety of speed, sizes, types and forms. Memory size / capacity range from 4 MBs to a whopping 1 GBs per module, but for the most part you will using 256 MB or 512 MB modules. Of course, having more memory is always good, but too much is not always a good thing. Most users won't see any performance improvement above 512 MB, since only the most demanding games and applications uses more memory than 512 MB. Knowing this, you might be tempted to purchase a single 512 MB module, but it's actually better to use two 256 MB module. Why? In most cases, a 512 MB module works slower than a 256 MB module. So if you're really want performance, you'll get a little more by paying a little more - two 256 MB modules cost a little over one 512 MB module.

Types of Memory

In general, there are three types of memory available today: the old standard SDRAM, the current popular choice DDR-SDRAM and the newer DDR2-SDRAM. If your PC is still using a Pentium II / III or a very old Athlon / Duron, you're probably using SDRAM. Since this type of memory is not in production anymore (or produced in very small numbers), they are very limited. In fact, while they are slower than DDR-SDRAM, SDRAM are more expensive. So, if you're planning to upgrade your SDRAM memory, don't. It's actually will be much cheaper (memory wise) to upgrade to DDR-SDRAM. You will need a motherboard that supports DDR-SDRAM since these memory standards all use different slots - you must install a SDRAM module in a SDRAM DIMM slot and this is also true for DDR-SDRAM and DDR2-SDRAM. Some motherboards come equipped with both slots. If you have one of these, only use one type of memory - you can't mix both.


 SDRAM
 DDR-SDRAM DDR2-SDRAM
100 MHz
 PC100 PC1600/DDR200
133 MHz PC133 PC2100/DDR266
166 MHz
PC2700/DDR333
200 MHz
PC3200/DDR400 DDR2-400
233 MHz

DDR2-533
266 MHz

DDR2-667


DDR2-SDRAM arrives on the scene alongside Pentium 4's new 775 socket. Most new motherboard use this new memory standard requiring a new slot type. Unfortunately, the first generation of DDR2-SDRAM doesn't offer any performance improvement over the current PC3200 DDR-SDRAM. Only the fastest DDR2-SDRAM rated at 533 and 667 MHz offer real improvements. Since DDR2-SDRAM is only used on the new socket 775 Pentium 4s, it would be better to choose these faster DDR2 memory modules, especially if you're going to use the 1066 FSB Pentium 4. Remember, these modules don't come cheap, so if you're concerned about budget, it might be better to choose the equivalent Athlon 64 processor, since they still uses DDR-SDRAM memory.

It's All in the Speed and Timing

This paragraph may interest those who want to know more about memory performance. Just like your processor, memory has a rated working speed. That's why we have ratings such as PC100, PC133 for SDRAM, PC1600, PC2100, PC2700 and PC3200 for DDR-SDRAM, and DDR2-400, DDR2-533 and DDR2-667 for DDR2-SDRAM. These numbers indicate what their rated speeds are, in SDRAM and DDR2-SDRAM it’s in MHz. For DDR-SDRAM, these numbers actually signifies the estimated memory bandwidth in MBs, so in terms of speed, a PC1600 memory module works at 100 MHz, while PC2100 works at 133 MHz, PC 2700 at 166 MHz and PC3200 at 200 MHz.

Once again, speed in not everything. There is also what's called timing to consider. Every memory works very much the same, they require several steps when accessing data to actually retrieving it when the processor requests it. First, they must look for the data and then retrieves them. The time to retrieve (and write) data is very quick and generally doesn't vary that much from memory to memory. It's the looking part that interest most people and rightly so since this affects much of the memory's performance. Timing in general is measured in latency - how much cycle the memory will spend looking for data. In short, the less latency, the faster the memory. As memory speed goes up, latency also increases. But wait! This means it is actually running slower! You're right, but this penalty is usually hidden or offset by the increase in speed. In fact, this is the main reason why a PC3200 DDR-SDRAM in most cases can outperform a DDR2-400 - the DDR2 module has a higher latency but works at the same speed. You're only going to see any performance increase over PC3200 on DDR2-533 and DDR2-667 modules.

Two Modules is better than one

Newer processors and motherboards use a dual channel memory controller. These processors and motherboards retrieve and write data to two memory modules at the same time, thus in theory doubling the performance. AMD uses dual channel memory controller in all of their socket 939 processors. For socket A / 462 AMD processors, you need to use NVIDIA NForce 2. Intel have used dual channel memory controller since their Granite Bay chipset for socket 478 and you could find the same controller on motherboards using Intel 865/875 chipsets and of course, the latest Intel 915/925 chipsets for socket 775 processors. Other third party chipset manufacturers such as SiS and VIA also provide dual channel memory controller equipped chipset for Intel's Pentium 4. To fully utilize the capabilities of dual channel memory controller, you will need two identical memory modules installed in different channels. Of course, this means they must have the same rated speed and timing. This is another reason why I recommend using two 256 MB modules instead of one 512 MB module. We could even have more performance this way.

So in retrospect, this is what you should consider when choosing your PC's memory:

Graphics card

Next to the processor, the graphics card is the hottest part of a PC - in both sense of the word. As PC games become more mainstream, graphics card have receive more and more attention. You could easily find graphics card that cost more than your processor or even more than your whole PC! Of course, we're talking about add-on cards here, not integrated solutions.

Integrated Graphics: Basic and Cheap

An integrated graphics card means that the graphics card is built into the motherboard. This integrated solution is much cheaper than an add-on solution, but not necessarily a better one. To integrate the graphics card into the motherboard, some compromise must be made. For one, these integrated graphics solutions are usually slower than their add-on counterparts. They also use part of your main memory (RAM), so that means less memory for your operating system and applications. Integrated solutions are more suitable for PCs that's not geared for gaming - only for office work or multimedia.

Some motherboards with integrated graphics still allow you to use an add-on card since they come with an expansion slot. So, if you're unsure whether or not you'll be using your PC just for work or games, go with this solution. If you're sure you only want a PC that can only be used to work, you can forgo the add-on connector or any other expansion slots for that matter. Just make sure the motherboard comes with all the peripherals you'll ever need (sound card, Ethernet adapter / network controller, graphics card).

Add-on Graphics: High to Low

Add-on card is where most of the action (and attention) is. You need a powerful graphics card to play all those eye candy games. These add-on cards usually come in two flavors: an AGP or PCI Express card. AGP is the old standard while PCI Express is the newer one. Be careful though, a PCI Express card is not always faster than their AGP counterpart. If they have the same chip, the PCI Express will not be faster. If you're concerned with future upgrades, PCI Express is the way to go. AGP is still here because it's cheaper and has more installed user base.

For the most part, your option will either be an ATI or NVIDIA powered graphics card. ATI and NVIDIA are the two largest graphics chip maker in the world. Add-on card manufacturers built cards using their chips, and these add-ons are what you, the end-user, buy. Just like any other manufacturer, ATI and NVIDIA have several products to offer, ranging from the high end (most performance, highest price) to the entry level (least performance, lowest price). The sweet spot is in the middle, what's usually called the mainstream series. These products offer the best bang for the buck. Each series may include several products, so in total, you might be looking at 6 (or more) possible products from each manufacturer.


 Vertex Pipes
 Pixel Pipes
 Bus Width
 Core Speed
 Memory Speed
ATI




RADEON X850 Series:
RADEON X850 XT Platinum Edition,
X850 XT,
X850 PRO 		6
 Up to 16
 256-bit , 256 MB GDDR3

RADEON X800 Series:
RADEON X800 XT Platinum Edition,
X800 XT, X800 XL,
X800
 6
 Up to 16
 256-bit , 256 MB GDDR3

RADEON X700 Series:
RADEON X700 XT,
X700 PRO,
X700LE
 6
 4 or 8
 64-bit or 128-bit, 64MB, 128MB, or 256MB DDR1, DDR2 or DDR3

RADEON X300 Series:
RADEON X300,
X300 SE,
X300 SE HyperMemory
 2
 4
 64-bit or 128-bit, 64MB, 128MB, or 256MB DDR1







NVIDIA
 Vertices per Second
 Pixels / textures per clock


GeForce 6800 Series:
GeForce 6800 Ultra, 6800 GT, and 6800
 600 Million
 16 / 16 (6.4 billion texels/sec)
 256-bit, 256 MB Bandwidth 33.6 GB/sec
1050 MHz
GeForce 6600 Series:
GeForce 6600 GT and 6600
 375 million
 8 / 8 (4.0 billion texels/sec)
 128-bit, 256 MB, Bandwidth 14.4 to 16.0 GB/sec
900 / 1000 MHz
GeForce 6200 Series:
GeForce 6200, GeForce 6200 with TurboCache supporting 128 MB / 256 MB
 225 to 263 million
 4 / 4 (1.2 to 1.4 billion texels/sec)
 128-bit or 64-bit, 128 MB (for 6200), 32 / 64 MB (for 6200 TurboCache)
Bandwidth 12.4 to 13.6 GB/sec


Confusing, isn't it? This is just the current generation of products available on the market. For simplicity reasons, let's talk about the series and not the individual products. Like we said before, the main difference between these series is performance. ATI's X300 and NVIDIA's 6200 series are entry level cards, X700 and 6600 series are mainstream products and X800/850 and 6800 series are high-end products. Naturally, they also vary in price the same way.

For those of you that are going to use your PC for serious gaming fun, choose either the mainstream or high-end products. Entry level products offer abysmal performance with newer games - these are better suited for casual gamers or office work and general multimedia use. A note on high end products, these cards are very expensive and only available in limited numbers. They will work best when coupled with the fastest processors available, making your PC even more expensive. Only choose these high-end products if you really don't want to play games in resolution below 1280 x 960 or 1280 x 1024 (on at least a 19' CRT monitor or a 17' LCD), with all the graphical features and effects set to their highest.

That said, buying a high-end graphics card (often at twice the price of a mainstream card) also allows you to enjoy high frame rates on current and newer games for at least two years. Mainstream graphics card can supply frame rates fast enough for a single year of gaming, but you might have to choose a lower resolution or tone down some features / effects for the next batch of games in the second year. They effectively become low end cards in two years time. Of course, this will also happen to your high end graphics card (eventually) although longer.

ATI or NVIDIA

This matter is very subjective. On the current products (ATI's Xx00 series and NVIDIA's 6x00 series), the display and image quality is not that different, most people won't even see the difference. As for feature, both camps have cards that offer just the basic package to those with TV tuner and capture / record capabilities. Of course, more features means higher price. Performance wise, each product in their respective market segments (mainstream, high-end) have their highs and lows, but all are enough for comfortable gaming. What you should be more concerned with is price and availability: choose the card(s) that's available (in stock) and offer the best combination of price and performance for you.

Graphics Performance

Now, let's talk about performance, one of the most 'interesting' aspects of any graphics card. Since many of us buy graphics card to play games, we want the highest possible performance from our graphics card (within our reasonable budget of course). Some of you may think that measuring graphics cards performance should be easy enough, just choose the graphics card that provides high enough frame rates when we play games. Unfortunately, it's not as cut and dry as that.
The Processor Connection
It's true that a faster graphics card will display (or render) images faster, but there are many other factors to a graphics card performance. For one, a graphics card will only render images of anything when it's told to, usually by your processor. That's why we need to have the fastest processor if we want to use the fastest graphics card. If the processor is too slow, your super-duper-fastest-graphics-card-on-the-planet will have to wait for instructions, doing nothing in the meanwhile. Needless to say, this is not what we want. Then of course these instructions must be sent to the graphics card itself, which is why we need a fast connection for the graphics card. In fact, this is the sole reason why we moved from ordinary PCI connectors to AGP (Accelerated Graphics Port for those who don't know). This is also the reason why graphics card have the fastest connector available in PCI Express, the x16 slot.
Inside your Graphics Card
Let's take a deeper look at your graphics card. Of course, we want our graphics card to be fast, but what does that mean? Basically, we need our graphics card to be able to process data faster. So in a way, your graphics card is like a mini PC, complete with its own processor and memory. Like we talked about before when talking about processors, pure speed isn't everything. The same things that affects processor performance also applies here such as IPC (Instruction Per Cycle) and bandwidth, the two other major factors making up the performance formula.



To be able to display those scenes in your favorite games, your graphics card uses two kinds of data: triangles / polygons and textures. Of course, one measure of a graphics card performance is how much polygons and texture can the graphics card process per second. These are usually stated as MPolys / sec (for triangles / polygons) and MPixels / sec (for textures). If a graphics card can process one polygon and one pixel per cycle, running it at 500 MHz would mean it can process up to 500 MPolys / sec and 500 MPixels / sec. Now remember, this is an overly simplistic example and graphics card are anything but simple. This number will increase if the graphics card can do more per cycle, in fact this is what modern graphics card do. They use several processing units within one chip or one card (just like a processor's pipeline - only this is basically a graphics pipeline), some dedicated to processing polygons, others for textures and then just running the entire card as fast as possible. Units dedicated for polygons are called vertex units or pipelines, while units for textures are called pixel units or pipelines.

Here comes the other part of the equation: bandwidth. Faster graphics card can process much more data, so it's only natural that we increase the amount of data. Why? More data means more detail (both polygon and texture) and that's why we are seeing more and more memory on graphics cards, the latest being 512 MB on a single graphics card. By contrast, before 3D graphics accelerator became the norm, 8 MB was enough. A graphics card with that much RAM would be able to store roughly 512 MB of polygon and texture data, but if you only have a 128 MB graphics card, the additional 384 MB of data must be placed in your PC's RAM, which is slower. So, in this case you must choose a lower detail setting that will use less data (and less detail). There's also transfer speed to consider, we want the data stored be transfered as fast possible to the graphics chip. There are two ways to do this: use the fastest memory possible (speed and timing) and use the widest channel possible (usually stated in bits), much like Intel and AMD did when they move from single to dual channel memory controller. Graphics card manufacturers opt to do both, as you can see from the memory clock speed and bits in technical specifications.

Here's a list of some internal factors that influence a graphics card's performance:
So, where do we start looking then? Look back at what we've discussed before: high-end, mainstream and level entry cards differ not only in price, but also performance. So you can be sure that these internal factors will vary between each series. Usually, the fastest card will have the fastest clock both for the graphics card chip itself and the memory. The chip will also be able to process more data per cycle. To feed enough data, they will use a wider bus (coupled with a high enough memory clock) to provide massive bandwidth. For mainstream card, one or several of these aspects will be lower (ie. lower bus width or lower clock or lower data processed per cycle). And of course, the entry level card will have the lowest specification (ie. lowest bus width or lowest data processed per cycle).

Looking back at the graphics card technical specification from both manufacturers, we see that both of them use the same decisions when differentiating their products. To differentiate mainstream and high-end products, they lower the data processed per cycle by lowering the number of 'pipelines' and also lowering the bus width from 256 bits to 128 bits or even 64 bits. They also offer several products in each series using different clocks for both core and memory or even the number of pipelines. So based on this specification, the products from both ATI and NVIDIA are evenly matched. We just have to find what graphics card will meet our need for speed.
It's In The Games
Now, we have some idea of what determines the performance of a graphics card. But unfortunately, there is no way of actually knowing for certain how these cards will perform while playing games except by actually measuring their performance during gameplay or a running a benchmark that simulates it. And since there are external factors affecting performance, this complicates things further. But all is not lost, many print and online publications provide benchmarks of several games and applications, either using the actual game or a specialized synthetic benchmark designed to measure performance.

Here are several things you have to consider when looking at the game's benchmark results.

Storage: Hard drives

Every PC needs storage to read / save data, the operating system and games / applications. For the most part, you will be using hard drive(s), although you periodically backup your data to CD / DVD-R / Ws or install applications from CD / DVD-ROMs. So, hard drives are not only a critical component of your PC, it is also one that affects your PCs performance, though indirectly. Why? Since the PC only use the hard drive when it needs to get or save something, getting a bigger or faster hard drive won't have much effect on your frame rates. However, they will take less time when loading maps or levels.

At the moment, hard drives capacity ranges from 30 to 500 GBs. While this may seem unnecessary for some, these hard drives are here for the duration. Look at some modern games, its not that uncommon to find games that's available in DVDs. Games are getting larger as more and more detail are put into it. More and more people are using their PCs not only to play games, but store MP3 songs, DivX videos and other types of media that will quickly gobble up space even on the biggest hard drives. As the number of files, their capacity and size gets larger, seeking a particular file or file segment may take longer so we need a high performance hard drive and storage sub system. That's why hard drive performance is more important now than ever, despite the little impact they have on gaming performance.

Up to 2004, PC hard drives and CD / DVD-ROMs have been using the same standard interface. You may know this interface as IDE or Enhanced IDE (EIDE) or even the newer versions of it such as Ultra ATA or ATA-33 / 66 / 100 / 133. All these standards basically use the same connector, so newer ones are backward compatible with older drives. This is one of the reasons IDE have been here like forever. The other reasons are its price (both the controller and the storage devices are cheap), availability and user install base. Every motherboard since the days of Intel 486 and AMD 5x86 processors in 1993 have an onboard IDE controller. While there are other interfaces such as SCSI or the more exotic Fibre Channel, they're usually reserved for professional and corporate users. These devices are more expensive, although they're offer more performance and features.

Out with the Old, In with the New

Since it's very old, IDE is not very suitable for modern PC users needs anymore. Physically, the connectors takes too much space, the cables too wide by today's standards and quite often they hinder airflow. An IDE controller can only support no more than four devices - by using two channels, two devices per channel. To use two device in one channel, you need to configure them properly, which for some might be confusing or too much of a hassle.

That's why the industry is moving on to a new standard, Serial ATA or S-ATA for short. This new standard uses a new, smaller connector and cable while at the same time providing higher bandwidth than ATA-133. They're also easier to install, since you don't have to configure anything. Compared to the first generation of S-ATA (S-ATA 150), S-ATA 250 compliant controllers offers features such as hot-swap / hot-plug and native command queuing 'borrowed' from SCSI's into the consumer market.

S-ATA: The Real Deal

Unfortunately, only S-ATA 250 controllers offer support for four devices and they have just recently arrived on the market (note the same maximum number of devices with IDE). S-ATA hard drives are also more expensive than their IDE counterparts, but in general they don't offer much performance improvement. To use extra feature of S-ATA 250 like hot-swap and native command queuing, you must use them with S-ATA 250 storage devices (hard drives and CD/DVD-ROMs). That means first generation S-ATA hard drives won't be able to utilize this feature. Furthermore, for most users with one or two hard drives, S-ATA 150's bandwidth of 150 MB / sec or S-ATA 250's 250 MB / sec won't be much utilized. Why? Most hard drives available now (regardless of interface) offer a maximum transfer rate of 60 to 70 MB / sec, so only when your PC is accessing two hard drive at the same time (such as in the case in a RAID array) does the bandwidth really do any good. S-ATA's extra bandwidth only really comes to play when you use more than two devices.

Granted, some of you might be thinking of using four hard drives on a S-ATA 250 controller. But before you do, think carefully. Sure, using four hard drives with a maximum transfer rate of 60 to 70 MB / sec each will deliver a total of 240 to 280 MB / sec maximum transfer rate. But if the S-ATA 250 controller is still connected through a PCI interface, that bandwidth would go to waste - PCI still transfers data at 133 MB / sec! Make sure it's connected through a PCI Express x1 (266 MB / sec) or PCI-X which is an extension of PCI targeted for professional workstation and servers.

Knowing this, why should anyone invest in S-ATA peripherals? Well, since newer motherboards come equipped with them, there's no additional cost (except for the hard drives) of using S-ATA peripherals and devices. The transition from IDE to S-ATA will not happen overnight, so think of it as a long term investment. If one day manufacturers stops supporting IDE, you would be happy that you made this decision now rather than later. Believe me, that day will eventually come - probably the same day S-ATA hard drives becomes cheaper than IDE drives.

On to the Drives

Let's talk about hard drives in general. Capacity is all the rage these days, as you may have noticed. You will also undoubtedly notice that the hard drive's dimensions itself haven't changed much, most come as a 3,5' drive. That's because, manufacturers have been able to squeeze more and more capacity from a single data platter than ever before. So, while the platter dimension stays the same, data density is much higher on modern drives. If users need more capacity, they put in a second platter to effectively double the capacity.

But if you look at spindle speed, hard drives haven't really changed that much. For desktop users, hard drive's spindle speed are either 5.400 or 7.200 rpm. Only Western Digital currently offers 10.000 rpm S-ATA hard drives for desktop users. There are faster hard drives (15.000 rpm), but they're usually targeted for professional workstation and server users, and they're only available on SCSI or Fibre Channel. You may be asking why haven't everyone move to 10.000 rpm hard drives. First and foremost, cost. While you do get the better performance due to lower access time with these drives, you lose much in capacity. The first 10.000 rpm WD Raptors came with only 36 GBs. Now, compare that to the 80 GBs capacity which is the norm for 7.200 rpm hard drives at the time. Plus, these 7.200 rpm drives actually cost less. The second factor is surprisingly, performance. While a 10.000 rpm is faster, it's not that much faster and in reality 7.200 rpm hard drives are fast enough for most users.

insert table: 5400 rpm, 7200 rpm, 10000 rpm (capacities, access time, cache size)

This doesn't apply to differences between 5.400 and 7.200 rpm hard drives. First, 7.200 rpm hard drives are faster due to the lower access time (around 10-12 ms compared to 15-17 ms on 5.400 drives and 7-9 ms on 10.000 rpm drives). Second, they're available in the same capacity since manufacturers use the same platter for both 5.400 and 7.200 hard drives. Third, they only cost a fraction more than at the same capacity. So you could honestly say that 7.200 rpm hard drives are the best bang for the buck solution in storage.

Hard drive buying tips:

Storage: Optical drives

Most of you have either a CD-ROM or DVD-ROM installed in your PC. This is the primary source for installing games and application, playing back videos from DVD/VCD or audio from an audio CD. In fact, for most of us, it's the only way for us to install Windows. So, needless to say, a CD/DVD-ROM is very important. Now for those of you less technically inclined, a DVD-ROM will also read ordinary CDs, so you don't have to buy a separate drive just to read CD-ROMs.

The only reason you probably want to buy another drive is you want to make your own CDs or DVDs. An ordinary CD/DVD-ROM drive can't write data onto a CD or DVD media, so you'll need either a CD-RW or a DVD-RW drive. Again, a CD-RW drive will only be able to write to regular CD-R (R for Recordable) or CD-RW (RW for ReWritable), while DVD-RW drive can write to CD-R, CD-RW, DVD-R and DVD-RW media. So if you can spare the money, get a DVD-RW since they support reading and writing all those medias. Fortunately, DVD-RW drives have become more affordable.

Before you go and buy a DVD-RW drive, you must know that there are two competing standards of DVD recordable media, the DVD-R (minus) and the DVD+R (plus). Of course, this also applies to the rewritable version that media (DVD-RW and DVD+RW. Confusing isn't? Thankfully, manufacturers now have drives that supports both standards, usually called DVD-/+RW. So, buy a drive that supports both standards. As a plus, the newer drives also support dual layer recording. Dual layer recording allows you to use dual layer media that holds twice the capacity of single layer media.

The Next Battle: HD-DVD and Blu-Ray

Trying to come up with even bigger capacities, the industry is proposing two new standards: HD-DVD and Blu-Ray. Don't worry, these devices will still be able to read all current DVD formats. However, they will also support a new media that will hold up to 40 GBs of data. Unfortunately, these new standards are not compatible with each other, so first generation drives supporting either one of these format most likely will not be able to read the other. Hopefully, the next generation of these drives will (eventually) support both formats, just like the case with the DVD recordable media. In the meantime, most people will find a DVD writer is enough since it's way more affordable and DVD-ROMs have a larger install base than either HD-DVD and Blu-Ray put together.

The Media Debacle

Choosing a recordable media can be a major pain. There are a variety of media grades, with the same variety in quality and longevity. After years and years of writing both CDs and DVDs, I think it's safe to say that no single media will last forever. Even the best grade media will only stay readable for one or two years if you use them regularly. So the best policy is redundancy: make two copies of your data - one for storage and one for access. After one year, repeat the process for the same data. Sure, you will probably end up with a stack of media containing the same set of data, but you won't lose your data.

The same policy also applies to backup, although there are some additional measures you may want to take. First of all, combine two sets of backup (full and incremental) so you won't lose (much) if anything bad happens. Depending on how often your files change, set a full backup every week (often) or every month (less often). Between those week (or months), do incremental backups, either daily or weekly. For backups of systems, use rewritable media so you can reuse the media several times. For data backups and archives, use recordable media - they can only be used one time so use them for finalized data (data that will not change) if possible.

Remember this if you're thinking about getting a CD/DVD-ROM

Motherboard

All the components that made up every PC are connected to the motherboard, so it's no surprise many magazines and websites regularly test and review these boards. Though we have come a long way since the early days, a motherboard's design haven't changed that much. Here, you will find a processor socket, several DIMM slots for the memory modules, several expansion slots for add-on cards and I / 0 (input / output) ports for the mouse, keyboard, floppy, hard drive(s) and CD / DVD-ROM. You might also find some integrated components such as LAN / network adapter, graphics card, additional storage controller and of course, sound cards.

We have learned that the motherboard is running at a fraction the speed of the processor. Let's rephrase this: it is actually the chipset on the motherboard we're talking about, not the entire motherboard. Chipsets acts as a traffic controller of sorts in a busy street, passing on data and data request from one component to the other. In the old days, chipsets comes as a pair of chipsets: a north bridge and a south bridge. Advancements in technology have made it possible to put them into one package, so not every motherboard has two chipsets.

Bus speed: Processor and Memory

Since processor comes in all kinds of speed and uses several bus speed, your motherboard's chipset have to support them. If not, you would have to run the processor at a bus speed the chipset supports. For example, if you're using a Pentium 4 with a 800 MHz FSB on a motherboard that only supports up to 533 MHz FSB, you will have to use the 533 MHz FSB. Of course, you'll end up with a slower PC since your processor will run not only on a slower bus but also slower overall speed as well. Why? Remember that your processor runs at a certain multiplier so a Pentium 4 2.4 GHz that uses 800 MHz FSB (2.4 GHz is 16 times the actual 200 MHz FSB) will only run 1.6 GHz when using 533 MHz FSB (the FSB is actually running at 133 MHz). So pick your processor and motherboard wisely.

Chipset manufacturers continually update their chipsets to keep up with processor and bus speed changes. Unfortunately, you can't change the chipsets on your motherboard. If you want to use the new chipsets, for example to use the new bus speed, you will have to buy a new motherboard. This is what happens when you 'upgrade' your PC, where you change the processor and the motherboard as well. You don't have to change the motherboard if the motherboard still supports the bus speed of your new processor. So, again remember before you buy the motherboard, check what processors and bus speed does the motherboard supports. Make sure it supports your processor and (at least) that processor's bus speed.

You also have to consider the memory as well. Remember, for the processor to run at its optimal performance, we have to use memory modules that run as fast or faster than your processor's bus speed. In addition to that, if you're going to use than one memory module, check how many banks of memory modules the motherboard (chipsets) can use when using the specified memory module. Remember, banks don't necessarily equal slots. You don't have to do this for Athlon 64 processors and motherboard, since the memory controller is integrated in the processor. Again, try to use single sided (single bank) memory modules when possible.

Expansion Slots for Add-On Cards

Expansion slots are just that, they make it possible for you to expand the capabilities and features of your PC with add-on cards. There are several types add-on cards, the most common are graphics cards, sound cards, network / LAN adapter, internal fax / modem adapter, storage controllers either RAID IDE or RAID SCSI, and TV tuner / video capture cards. There are others such as additional USB or Firewire ports, Wi-Fi adapter, satellite broadcast receiver and so many more. Nowadays, expansion slots come in two types: AGP (used only for graphics card) / PCI (for all others) and PCI Express. Check your motherboard's manual to see how many of either are available on the motherboard. Needless to say, they uses different physical connectors so you can only put PCI add-on cards on PCI slots and the same with PCI Express.

Having lots of slots can be good, but not necessarily so. If you don't use many expansion cards or are not planning to use them, you don't need more than two expansion slots. Sound cards and network adapter are usually integrated into the motherboard these days. Your motherboard may also come with IDE or SATA controller that's RAID capable. Most users only use the expansion slots either for an internal fax/modem, TV tuner / video capture cards or sound cards if they want to use a separate sound card. You might ask why use an additional sound card if the motherboard already comes with one? First of all, integrated sounds card vary much in quality so many users that want or need better audio quality still opt to use an add-on solution. Second, there is usually a little performance hit with integrated sound cards when playing games. With an add-on sound card that's geared for gaming, this performance hit is less and often they support additional features that the integrated solution lacks.

If you're not going to use expansion slots, you can opt to choose a motherboard that has all the features integrated. These are usually cheaper and come in a smaller factor form. Just remember to have at least two expansion slots on the motherboard: one for graphics card (AGP or PCI Express x16 slot) and an additional PCI or PCI Express x1 or x4 slot.

When choosing a motherboard, keep these things in mind:

Ergonomics and Features

With all the craze of integrating more features into motherboards these days, the first thing manufacturers tend to compromise is space. This means installing and removing cables and other removable components can be very troublesome. On smaller and cheaper motherboards, you have to install the memory first and then the graphics card. Uninstalling them have to be done in reverse, and since these two are usually placed near the processor's socket, fan and power supply cables can get in the way. Sometimes IDE and floppy connectors are placed too close to other components, making installation and uninstallation more difficult.

Sacrificing ergonomics and cramming everything into so little space is not only making the installation and uninstallation process difficult, but can also hinder airflow when the PC's case is closed. If airflow is hindered, the cooler air outside can't get in and change the hot air inside. This will eventually lead to heat issues such as overheating or even worse, permanent damage. Needless to say, we don't want any of this to happen.

Fortunately, the ATX and BTX form factor standard have address much of these problems. Of course, using smaller boards still complicates things, but usually there are ways we can get around that. A motherboard's layout will affect much of its ergonomy, so choose a motherboard that uses a good layout.



Pay attention to these areas:

BIOS

Of course, when talking about a motherboard's features, we are not only talking about hardware, but also the motherboard's BIOS. BIOS (short for Basic Input Output System) is software that allows you to turn on / off features and enhancements and set values to be used for those features and enhancements. All motherboard comes with their own BIOS, since they're usually flashed into a BIOS chip on your motherboard. BIOS are often updated, so it's a good idea to check the motherboard's manufacturer website to see whether or not a newer BIOS for the motherboard exist. You can see what's changed or what bugs are fixed in the readme file accompanying the new BIOS.

Motherboard uses BIOS from different manufacturers, so the menus and options shown may be different from motherboard to motherboard. Check the motherboard's manual, there's usually a short explanation about the BIOS and the settings that you can change. Some will have a default and recommended value. Just remember, if you're going to use integrated peripherals, you have to turn them off through the BIOS - usually in the 'Integrated Peripherals' sub menu.

Some manufacturers make it a habit to hide some options and features in the BIOS. They did it because they're afraid end users will try to tweak the BIOS and unknowingly cause a system instability or crash. However don't be afraid, if you changed something and it makes your PC unstable, just reset the BIOS by clearing the CMOS (using the clear CMOS jumper) or bypassing the BIOS values stored in CMOS by pressing and holding the 'Ins' key when booting up.

Ergonomy wise, some features we want and need is shown in the BIOS. In a perfect world, these settings should be on all motherboards. Sadly, we're not in a perfect world, so check whether or not your motherboard's BIOS shows these settings:

In the PC Health Status:

In the Advanced BIOS Features:
If the above settings exist, chances are you could monitor them using additional hardware or software. Monitoring software can be configured to display an alert and take immediate action when these values hit a certain point. So for example, we could minimize the damage to your processor by shutting your PC down when the processor's temperature gets over 50 degrees Celsius and / or the ambient temperature goes over 40 degrees. You can do the same thing to power supply rails (we'll talk later in the power supply section about power supply rails and voltages). S.M.A.R.T allows you to monitor the health of your hard drives so you would be warned when your hard drives are beginning to fail. These features are important. There are also several additional settings for advanced users. This settings will allow you to wrench more performance from the motherboard, but if you're not careful or know what you're doing, they may cause system instabilities. It's nice to have them, so you could have more control over your motherboard.

In the Advanced Chipset Features:

In the Frequency Control:

Power supply

Choosing a power supply can be a daunting task. Looking at the total output (in watts) is not enough. A PC's power supply unit transfer electricity to components in your PC through three different lines - we usually call them rails. These rails are the 3.3 volt, 5 volt and 12 volt rails. Electricity from the 12 volt rail is usually used for hard drives and CD/DVD-ROMs, the 5 volt rail is used to supply power to the processor and the 3.3 volt rail is used to power add-on and integrated components. However, there are some exceptions. Athlon 64 and Pentium 4 use an additional ATX 12 volt cable, drawing extra electricity from the 12 volt rail to help power the processor. Some high-end graphics card also comes equipped with an additional power connector, so they can also draw electricity from the 12 volt rail.

Needless to say, you power supply must be able to supply the electricity requested by all components inside your PC. Failure to do so mean your PC may experience system crashes and instabilities. You could check whether or not your power supply is supplying enough power by monitoring the voltage values of each rail through the BIOS or by using hardware monitoring software. If a rail's value drops or exceeds 10 % of the specified value (3.3, 5 and 12 volts), your power supply is barely coping with the load. As an example, this can be a value of 10,9 (or below) volt or 13,1 volt (or above) on your power supply 12 volt rails.

Fortunately, picking a good power supply (or one that's good enough) is not that hard. First, look at the technical specification of the power supply. They usually put some info on the box or a sticker on the power supply itself. The total output should be no less than 350 watts. Second, you need to look at the watts per rail, you can get the watts by multiplying the ampere values with the volt values (20 A x 12 volt equals 240 watts). You would want the 5 and 12 volt rails to have at least 150 watts. Remember, in most cases what you see in the technical specification is the maximum output, we need the sustained output. Make allowances and subtract 10 to 15 percent to get the estimated sustained output.

An extra tip is to never connect your monitor's power cable to your PC's power supply. Connect them directly to the UPS if available. The monitor will be an additional burden on your power supply, especially when they're just turning on.

Some power supply comes with S-ATA power connectors, silent mode and additional short circuit protection. While these are good features, they came at a price. For those of you with limited funds, choose a power supply that meets your PC's requirements electrically and still fits your budget. If you have additional funds, then you can choose the more feature packed version. Personally, I'd rather spend the fund to get a small, reliable, good UPS.

As you can see, a good power supply is a necessity for any PC. But the best power supply can only do so much, you got to have a clean input electricity to it as well. Most people still think that using a voltage regulator or stabilizer is enough. I think this is the bare minimum. To ensure your PC's health and data safety, use an uninterruptible power supply unit (UPS) with an automatic voltage regulator. Not only will this ensure your PC will get clean and stable electricity, they will also provide you with some additional time when you experience a blackout / brownout.

Remember that some parts of the world use a 110 volt standard while others use 220 volts. The actual volts itself may vary, for example some countries use a range between 220 and 240 volts, while others may use 200 to 220 volts. Make sure you have configured your UPS and power supply correctly before plugging them to your wall's electric socket.

Casing

Of course, for the most part you will be using a case to store all these components. There are many form factors to choose from desktop, mini-tower to full tower enclosures. They also vary in the number of 5.25' slots, metal used and additional features. While you may want the best looking case around, there are important factors to remember when choosing a case.

Airflow and Heat Issues

Since the case is actually a container, usually box shaped, air flow becomes an important factor to consider. Without good airflow, heat will build up and temperature will rise inside the case. Your PC's components can only tolerate so much heat, so when the temperature is too much it may become unstable or worse, permanently damaged.

Even the coolest components will always generate heat, warming the air around it. That's why all cases come with some ventilation holes so that the hot air inside can be replaced by the cool air outside. If you remember your Physics 101, hot air tend to rise upwards, while cooler air tend to stay near the surface. That's why ventilation holes are place both in the bottom front and upper back area of the case. There may be additional holes in the side, but that's about it.

While the hot air will eventually go outside the case, we might want to speed things up a bit with some fans. Two well placed fans will do much to help airflow and keep the temperature from rising. One for the intake on the bottom front and another for the ??? on the upper back of the case.

Form and Build

Remember that the form also determines how much room inside the case. A bigger case can be used to store more components, but more importantly it also means there is more room for air. A mini-tower case is ideal for most users, since they feature several 5.25' drive slots and a single 3.5', room for three to five expansion slots.

A mini-tower case is also usually lighter than a full tower case. Usually, since the case material (aluminum or steel) will also factor in its weight. Aluminum cases are lighter than steel cases, at the same form factor. Steel is much tougher though, so check for any creaks and moving parts in the case. A good case should not creak or move a budge.

Ergonomy and features

Just like the motherboard, the case is actually comprised of several parts. Check how easy or difficult it is to install hard drives and floppy. Some cases can also be opened with using any tools, needless to say this greatly simplifies troubleshooting and installation. The side panel and other removable parts should also fit without you using much force. Some cases may feature a removable back panel for the motherboard, this makes the installation process easier, since you can mount the motherboard on this panel outside the case.

Having a front panel for USB, Firewire and audio ports is a plus. With it, you don't have to reach for the back of the case when you want to use them. Those with security concerns will want a locking mechanism so not everyone can open the case.

Fan and Heatsinks

Of the many components and parts inside every PC, the processor, graphics card and hard drive are usually the ones that generates much heat. The faster they are, the higher the heat. So, we can't just rely on airflow to cool them down. That's why fan and heatsinks are usually bundled with the processor and graphics card.

Some graphics card and motherboards come only with heatsinks, relying solely on good airflow. This method of cooling is known as passive cooling and is usually used when the components being cooled are already cool enough for the most part and these heatsinks are usually used as a precautionary measure. The advantage of this method is silence. If the component is quite hot, the heatsink can be quite large.

Since not every component can use large, passive heatsinks, some will use fans to directly cool the heatsink. This method is known as active cooling. Of course, the faster air blows on the heatsink, the cooler it gets. But very fast fans makes a lot of noise and not very suitable for everyday use. You can use larger, slower fans instead - they push relatively the same amount of air without having to spin very fast. On very hot processors and graphics card, you may see a combination of both large heatsink and fan, offering the best of both approaches.

Usually, heatsinks are made of aluminum. This metal quickly dissipates heat, so it's very ideal for heatsinks. There are also others made of copper, since copper absorbs heat much faster than aluminum. Some will even use both, using copper as the heatsink's base and aluminum for the fins, thus combining the advantages of both metals.

Heatsink design have also progress very far in the last couple of years. It's not that strange anymore to see heatsinks using both aluminum and copper in a variety of shapes and sizes. Manufacturers have also made use of heatpipes, using a pipe containing heat sensitive liquid to transfer heat much more quickly between parts of a heatsink. For most people, the heatsink bundled with your processor will be enough. Intel have been bundling heatsinks with their processor since the Pentium, with AMD following suit since the Athlon XP. So, always buy the official box packaged version when you're buying processors. The heatsink bundled inside is tested and guaranteed to work with your processor.

For hard drives, you can get hard drive enclosures equipped with a fan to cool them down. A more direct and less expensive approach will be to place another fan just for the hard drives and let the air through from in front of the casing. Not only will this cool the hard drives, but they will also help airflow since the case have additional holes for intake air.

Another important factor to remember is the room temperature. Since we're relying on the air to keep things cool, we have to cool the air in the room where we're going to use our PC. The best fan / heatsink and casing can not cool any component if the air is already too hot. So keep the room temperature around 25 to 30 degrees Celsius.


Go to top
Disclaimer and Privacy policy.